CONF-860932-

DE87 007690

ANS Topical Meeting on

RadiologicalAccidents —

Perspectives and

PlanningSeptember 15—17, 1986

Bethesda, Maryland

Proceedings

Date Published: March 1987

Prepared by theOak Ridge National LaboratoryOak Ridge, Tennessee 37831

operated byMartin Marietta Energy Systems, Inc.

for theU.S. DEPARTMENT OF ENERGY

under Contract No. DE-AC05-840R21400 MASTEROlSTRIBUTUJH OF THIS DOCUMENT IS

SPONSORING ORGANIZATIONS

American Nuclear Society

U.S. Department of Energy

U.S. Nuclear Regulatory Commission

PARTICIPATING ORGANIZATIONS

Federal Emergency Management Conference of Radiation ControlAgency Program Directors

U.S. Environmental Protection National Emergency ManagementAgency Association

I Iealth Physics Society U.S. Dept. of Transportation

This meeting could not have taken place without the assistance providedby the U.S. Department of Energy Office of Nuclear Safety.

DISCLAIMER

This report was prepared as an account of work sponsored by an agency of the United StatesGovernment. Neither the United States Government nor any agency thereof, nor any of theiremployees, makes any warranty, express or implied, or assumes any legal liability or responsi-bility for the accuracy, completeness, or usefulness of any information, apparatus, product, orprocess disclosed, or represents that its use would not infringe privately owned rights. Refer-ence herein to any specific commercial product, process, or service by trade name, trademark,manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recom-mendation, or favoring by the United States Government or any agency thereof. The viewsand opinions of authors expressed herein do not necessarily state or reflect those of theUnited States Government or any agency thereof.

Foreword

The increasing use of radioactive materials and the increasing public concern about possibleaccidents involving these materials has led to greater emphasis on preparing for such emergencies.The ANS Topical Meeting on Radiological Accidents—Perspectives and Emergency Planning gaveus an opportunity to review our experience with radiological accidents to determine what informa-tion from this experience could be applied to improve our preparedness for future accidents.

The meeting covered some of the most important aspects of radiological accidents. We startedby inviting several speakers to present papers dealing with radiological accident experience and thensolicited other papers on related areas.

Technical response to accidents is of primary interest to many in the nuclear community; mostof the papers submitted fell into this area. So many of these papers dealt with the use of computersin response that a session on that topic was arranged.

A very significant impact of most radiological accidents is the cost, especially the cost ofcleanup. There were papers on what is known about costs and associated current topics, such asmodification and extension of the Price-Anderson Act.

At least as important as the technical response to accidents is how society attempts to dealwith them. A session on institutional issues was included to discuss how governments and otherorganizations respond to and deal with accidents.

Medical effects of accidents are of great concern to the public. Invited papers to review theeffects of high doses of radiation as well as very low doses were included in that session.

Although the nuclear industry has an excellent safety record, this fact often does not agreewith the public perception of the industry. The final session explored the public response to and per-ception of radiological emergencies and accidents. This subject will ultimately determine the futureuse of radioactive materials in this country.

I think you will find that this volume contains an interesting set of papers. The Technical Pro-gram Committee did an excellent job of reviewing and selecting the contributions that appear here.Although the Chernobyl accident occurred long after the planning for this meeting was under way,a few papers on Chernobyl were added and many of the other authors managed to incorporate thisrecent experience. If nothing else, this accident confirmed the timeliness and appropriateness of ourtopic.

I hope that this review will help us improve our emergency planning, so that we will be evenbetter prepared to deal with any future accidents.

L. Joe Deal

General Chairman

MEETING OFFICERS

General ChairmanL. Joe DealU.S. Department of Energy

Associate General ChairmanClyde JupiterU.S. Nuclear Regulatory Commission

Technical ProgramConrad V. ChesterOak Ridge National Laboratory

ArrangementsRaymond W. DuranteSchneider Enterprises

FinancialDonald MarksberryU.S. Nuclear Regulatory Commission

RegistrationEmmanuel GlakpeHoward University

PublicationsKathy S. GantOak Ridge National Laboratory

Public InformationJoseph MarkiewiczEnergy Consultants, Inc.

David AldrichScience Applications InternationalCorporation

Joel BuchananOak Ridge National Laboratory

Robert CatlinElectric Power Research Institute

Harold J. CollinsInternational Atomic Energy Authority

Richard CuddihyLovelace Foundation

Robert W. DaviesU.S. Department of Energy

Russell DynesUniversity of Delaware

Andrew HullBrookhaven National Laboratory

TECHNICAL PROGRAM COMMITTEE

Chairman: Conrad V. ChesterOak Ridge National Laboratory

Milton LevinsonBechtel Power Company

Joseph LogsdonU.S. Environmental Protection Agency

Clarence LushbaughOak Ridge Associated Universities

Steven R. MillerU.S. Department of Energy

Robert RicksOak Ridge Associated Universities

John SorensenOak Ridge National Laboratory

Roberts. WilkersonFederal Emergency Management Agency

IV

Contents

Sponsoring and Participating Organizations ii

Foreword iii

L. Joe Deal

Meeting Officers and Technical Program Committee iv

Introductory Remarks 1Joseph Hendrie

Luncheon Presentations

The Federal Perspective 5Samuel W. Speck

The Industrial Perspective 11James A. Wick

Section 1. Accident Experience

Use of Radiological Accident Experience in EstablishingAppropriate Perspectives in Emergency Planning 19J. M. Selby, D. W. Moeller, E. J. Vallario, and J. G. Stephan

A Chronology of the Chernobyl-4 Accident 29G. Donald McPherson

Uranium Yellow Cake Accident—Wichita, Kansas 39Harold R. Borchert

Aerial Survey Efforts in the Search for Radon-Contaminated Houses in theReading Prong Area Near Boyertown, Pennsylvania 43

Raymond A. Hoover and Dennis E. Mateik

Emergency Planners, Look Back at TMI-2 47Robert L. Long

Section II. Technical Aspects

Source Terms Derived from Analyses of Hypothetical Accidents, 1950-1986 53William R. Stratton

Radiological Source Term Estimation Methods 63P. C. Owczarski, M. Y. Ballinger, J. Nfishima, and J. E. Ayer

Changing Perspectives on Severe Accident Source Terms 67R. S. Denning, P. Cybulskis, and R. DiSalvo

NRC Perspective on Severe Accident Consequence Assessment 73Thomas McKenna and James A. Martin

A History of Aerial Surveys in Response to Radiological Incidentsand Accidents 79Joel E. Jobst

Status of Aerial Survey Emergency Preparedness and Ground SupportEquipment, Calibration, and Sensitivities 85

Thomas S. Dahlstrom

v

Aerial Systems Support for Nevada Test Site Weapons Testing 91Philip K. Boyns

Mobile Robot Response to Actions Associated with the Release ofHazardous Materials 95Harvey B. Meieran

Communications Systems for Emergency Deployment Applications 101Charles A. Gladden

A Mobile Laboratory—Emergency Sample Analyses at the Accident Site 103R. H. Wilson

Developing a Comprehensive and Accountable Data BaseAfter a Radiological Accident 109Hollis A. Berry and Zolin G. Burson

A New Method for Presenting Off-Site RadiologicalMonitoring Data During Emergency Preparedness Exercises 113M. P. Moeller, G. F. Martin, E. E. Hickey, and J. D. Jamison

Enhancement of the 1985 Browns Ferry Exercise Through the Use ofSpiked Samples 119James L. McNees

Integrated Emergency Response Program at the Savannah River Plant 123Richard W. Benjamin

Savannah River Plant Emergency Response: Operations and Exercises 127Doris D. Hoel

A Mobile Laboratory for Near-Real-Timc Response to Radioactive Releases 131R. A. Sigg

Improving Emergency Response Through Field Exercises 135R. P. Addis, R. J, Kurzeja, and A. H. Weber

Savannah River Plant Remote Environmental Monitoring System 137/. F. Schubert

Aquatic Emergency Response Model at the Savannah River Plant 141David W. Hayes

Meeting NUREG-0737, II.B.3 Requirements—BackupAnalysis of Postaccident Samples 145

Todd L. Hardt, Mark J. Bradley, and Trudy E. Phillips

An Evaluation of Emergency Systems for the Monitoring, Sampling, andAnalysis of Reactor Coolant, Containment Atmosphere, and Airborne Effluents 149Andrew P. Hull, Wayne H. Knox, and John R. White

A State-of-the-Art Aporoach to Emergency Preparedness—RemoteMonitoring of Nuclear Power Plants 161James A. Blackburn and Michael C. Parker

Section III. Computer Applications

Computer-Assisted Emergency Preparedness and Response 169James W. Morentz

FEMA'S Integrated Emergency Management Information System (IEMIS) 175Robert T. Jaske and Wayne Meitzler

VI

The MAPSS Nuclear Emergency Management System 181Howard A. Price, Jr., Larry D. Sadler, and Robert W. Johnson, Jr.

A Comparison of Computerized Dose Projection Models and TheirImpact oi. Protective Action Decision Making 187Susan M. Reilly

An Importance Ranking of Various Aspects of Off-Site RadiologicalEmergency Preparedness 193John W. Hockert and Thomas F. Carter

A Goal-Oriented Functional Tree Structure for Nuclear Power PlantEmergency Preparedness 197Richard V. Calabrese and Marvin L. Roush

Results of the Early NRC Analyses of the RadiologicalMonitoring Data from the Chernobyl Accident 209Rosemary Hogan and Thomas McKenna

ARAC: A Centralized Computer-Assisted Emergency Planning, Response,and Assessment System for Atmospheric Releases of Toxic Material 215M. H. Dickerson and J. B. Knox

Digital Imagery Manipulation and Transmission for EmergencyDeployment Applications 223

C. A. Gladden and D. S. Phillipson

Image/Data Storage, Manipulation, and Recall Using Video/ComputerTechnology for Emergency Applications 225James M. Thorpe

Using MENU-TACT for Estimating Radionuclide Releases Duringa Reactor Emergency 229Andrea L Sjoreen

The MESORAD Dose Assessment Model 237R. I. Scherpelz, J. V. Ramsdell, G. F. Athey, and T. J. Bander

Accident Analysis Codes That Predict the Transport of RadiologicalAerosol Through a Fuel Cycle Facility as a Result of an Accident 241Kevin C. Greenaugh

REALM: An Expert System for Classifying Emergencies 247Robert A. Touchton, Alan Dale Gunter, and David G. Cain

Section IV. Economic Issues

Perspectives on the Economic Risks of LWR Accidents 253Lynn T. Ritchie and Richard P. Burke

Restoration of a Radiologically Contaminated Site: SAGEBRUSH IV 259JackJ. Tawil

The NRC's Rulemaking to Require Materials Licensees toBe Financially Responsible for Cleanup of Accidental Releases 265

Mary Jo Seeman

Applicability of Comprehensive Environmental Response, Compensation,and Liability Act of 1980 (CERCLA) to Releases of Radioactive Substances 271Steven R. Miller

vn

Price-Anderson—Where We've Been, Where We're Going 273Ira Dinitz

Section V. Institutional Issues

Protective Action Guides: Rationale, Interpretation, and Status 279Joe E. Logsdon

The Federal Radiological Emergency Response Plan 283Vernon Adler

Role of the Federal Radiological Monitoring and Assessment Center(FRMAC) Following a Radiological Accident 287John F. Doyle HI

Exercising the Federal Radiological Emergency Response Plan 291Kathy S. Gant, Martha V. Adler, and William F. Wolff

Emergency Response to Transportation Accidents 295Cheryl Sakenas

New Source Terms and the Implications for Emergency PlanningRequirements at Nuclear Power Plants in the United States 297

Geoffrey D. Kaiser and Michael C. Cheok

Coordination Between NRC and FEMA: Emergency Preparedness Issuesof Current Interest 303Edward M. Podolak, Jr.

Interactions Between Multiple Organizations Responding to a ReactorAccident in Switzerland 307Martin Baggenstos and Hans-Peter Isaak

Utility Perspective on Emergency Preparedness 313George J. Giangi

Improving Emergency Management Through Shared InformationProcessing—Considerations in Emergency Operations Center Design 319Richard E. DeBusk and J. Andrew Walker

The NRC Operations Center's Function 325Eric W. Weiss

Section VI. Medical Issues

Dose-Rate Models for Human Survival After Exposure to Ionizing Radiation 331Troyce D. Jones, M. D. Morris, and R. W. Young

Radiation Carcinogenesis Following Low-Dose or Low-Dose-Rate Exposures 341R. L. Ullrich

A National Emergency Medical Assistance Program forCommercial Nuclear Power Plants 345Roger E. Linnemann, M.D., and Mary Ellen Berger

Section VII. Accidents and the Public

Evacuation Behavior in Nuclear Power Plant Emergencies:An Alternative Perspective 351John H. Sorensen

Warning Human Populations of Technological Hazard 357George O. Rogers and Jiri Nehnevajsa

viii

Emergency Preparedness Training for Local Communities 363Michael J. Cooley and Karen K. Thompson

Emergency Planning, Public Information, and the Media: SomeRecent Experiences 367Steven B. Goldman

Factors Influencing Media Coverage of a Radiological Incident 373Roy K. Bernhardt and Layton J. O'Neill

Indexes

Author Affiliations and Addresses 379

Author Index 385

IX

Introductory Remarks

Joseph Hendrie

It is a pleasure to welcome you to the ANS Topical Meeting on RadiologicalAccidents—Perspectives and Emergency Planning. It is an auspicious time for such a meeting.Much credit is due to the sponsors and the participating organizations. The sponsors are theEnvironmental Sciences Division and the Washington Section of the American Nuclear Society, theU.S. Department of Energy, and the U.S. Nuclear Regulatory Commission. The participatingorganizations are the Federal Emergency Management Agency, the U.S. Environmental ProtectionAgency, the Health Physics Society, the Conference of Radiation Control Program Directors, theNational Emergency Management Association, and the U.S. Department of Transportation. I willspeak for all of them in welcoming you here.

When they began the planning for this meeting, the sponsors could hardly have anticipated thegreatly increased interest in emergency planning that grew out of the Chernobyl accident. We arevery fortunate that the timing was such that we can hear at this meeting the early results from theVienna meeting at which the Soviets explained Chernobyl. In administering organizations, you getblamed for a lot of things for which you are not responsible, but which happened at a time whenyou were in charge. That suggests that you ought to get credit for events which occur during youradministration for which you had no responsibility at all; therefore, the sponsors of this meetingought to take full credit for the timing.

Our purpose here is to review what we have learned from past accidents that involve nuclearmaterials and facilities and to use this knowledge to foster more effective emergency planning forthe future. We are going to have sessions on past accidents, medical and health consequences ofaccidents, lessons learned in the past, current thoughts on radiation dose response, and mtdicalemergency preparedness. Economic considerations in nuclear accidents will be discussed in one ses-sion, and we will have sessions on technical problems in emergency planning, institutional problemsin emergency planning and emergency response, and, finally, the role of the public in nuclear acci-dent emergency planning and the importance of public perceptions.

Nuclear technology is one of the few technologies in which an appreciation of the possiblehazards was present at the beginning and became a serious and ongoing part of the planning anddevelopment work. In most technologies, accidents begin happening after the equipment and processhave been invented, developed, and placed into service; over time a set of mitigation measures andresponse measures evolves and becomes embedded in society's practice. In nuclear technology, westarted out with an early appreciation of the hazards.

As time went on during the development of nuclear technology, there were criticality accidents,ovcrcxposurcs, and the like. The consequences of these were largely limited to on-site and local

areas. Windscalc, in the 1950s, was probably the first significant accident that had off-site conse-quences and affected the public. Other accidents followed; again, the consequences were mostly lim-ited to on-site effects.

The seminal event in this country for reactor accidents and the response to them was, ofcourse, Three Mile Island. My own experience at Three Mile Island strongly emphasizes the impor-tance of emergency planning, especially the importance of aspects of emergency planning that previ-ously had not been given a great deal of attention. The most notable of these aspects was the unex-pectedly massive public and media response to the accident, which created a tidal wave of interestand concern that washed across all of us who were trying to understand and deal with the accident,making all of our work much more difficult. Here in this country we have now a very goodappreciation of those effects. The emergency planning which has been done since Three Mile Islandhas, in my view, been an enormous step forward in preparing us should something like that everhappen again and has included measures to deal with that great public-interest and concern.

In the aftermath of the Chernobyl accident, there is now more than ever a great interest andconcern about radiological accidents and about what should be done in the event of another suchaccident. I am sure that this meeting will provide us with a chance to examine the past accidentrecord, including the recent events, to find where there might be weak points in our emergencyplanning structure and to look for places where measures now in use could be improved, made moreeffective, or made more efficient.

Luncheon Presentations

ANS Topical Meeting on Radiological Accidentf—Perspectives and Emergency Planning

The Federal PerspectiveSamuel W. Speck0

I appreciate the opportunity to speak on thenuclear power option in the post-Chernobyl era.I speak as one who until about one week ago hadresponsibility for commercial nuclear power plantoffsite emergency planning at the FederalEmergency Management Agency (FEMA) for some threeyears. I also speak as a member of the delega-tion that went to Vienna to meet with theRussians and the other delegations from aroundthe world to discuss what happened at Chernobyl.

Major disasters, whether they be Chernobyl,Bhopal, or the Mexico City earthquake, have a wayof significantly impacting the emergency planningthat comes after them. When we look back on theChernobyl incident, I suspect we are going tofind it responsible for a number of impacts onthe peaceful uses of nuclear energy.

First, the Chernobyl accident made it clearthat we will potentially be affected by the useof nuclear power elsewhere in the world, whetheror not we continue to make use of the nuclearoption in the United States. Russia made it veryclear at th» Chernobyl conference that she isputting her energy eggs in a nuclear basket. Anyincrease in electric generating capacity in theSoviet Union during the rest of this century willcome from nuclear generating facilities, at leaston the European side of the Urals, not fromfossil fuel facilities or hydroelectric facili-ties. There are several reasons for this. Onereason is that Russia wants to export her oil andgas to her Eastern European allies and also toWestern Europe in exchange for hard currency.So, nuclear power is part of the plan for theoverall economic development of the Soviet Union.Elsewhere, by the end of this century, Franceintends to obtain approximately 75% of her powerfrom the nuclear option, and Japan is moving downthe same road.

Another thing that is going to come out ofChernobyl is the internationalization of themanagement of nuclear power, not the control, butcertainly the cooperative internationalization ofthe management. We plan to participate in aconvention on the obligations of all nations fornotifying each other when there is a nuclearincident which could have transboundary impacts,I suspect that in the future there will be

aFormerly Associate Director for State and LocalPrograms, Federal Emergency Management Agency,Washington, D.C., 20472.

pressure to develop international referent levelsfor taking various actions. Given the kind ofopenness that we saw on many scientific issues atthe conference, I anticipate continued progressin the internationalization of the professionalnuclear energy community.

The Chernobyl incident caused Russia to openup to some degree. We have not seen that kind ofopenness in the past, and it is not clear that itis going to continue. Russia may remain closedin many areas, but once there is openness in oneparticular area, it is much more difficult (andit looks suspicious) to keep other areas closed.I think we saw a little of this when the Russiancruise ship sank recently; Russia was much more,open than she had previously been in similar >events.

There are a number of other issues raised bythe Chernobyl incident. One that is certainlygoing to have an impact here is the issue of thecomplexity, the difficulty, and the cost ofresponse and recovery in these incidents. Forexample, by meteorological manipulation, theSoviets were able to keep rain away from theChernobyl plant for about one month after theincident. That action protected the surroundingarea from the kind of run-off that it might havehad otherwise. On the other hand, the lack ofrain created more dust problems. Because of themovement of the dust particles in the air whenareas were decontaminated, there was a tendencyfor recontamination. Something that lessened thecontamination in one way exacerbated the contami-nation problem in another. In forests, it takesthree to four years for all the radioactiveparticles to settle to the ground, and thus thereare acute problems from fires in such areas.Here the Russians decided that perhaps the bestresponse would be to increase their forest-firefighting potential. Otherwise, a significantforest fire might well produce a new mini-Chernobyl.

There is a great d.eal that we found out, butalso an enormous amounpthat we did not find out,in regard to the health impacts of the Chernobylaccident. Milt Levinson relevantly noted thatbecause of the enormous follow-up the Russiansare doing with the people who had significantradiation exposure, the health impacts may not beas great as expected. Left unchecked, theremight be a modest increase in the rate of deathsfrom cancer in this population. If you have

sufficient follow-up, you might ultimately findlittle or no increase or even a decrease in theoverall cancer deaths. The predictions :.hat wehave been seeing have focused on the increase inthe number of deaths, not the increase in thenumber of cancers. If these people are checkedon an annual basis and cancers from all causes,not just radiation-related cancers, are detectedsoon enough, then overall cancer-related deathsmay turn out to be lower than projected becauseof this unique medical intervention.

Nevertheless, this event and the offsiteelements of it mean that, even if nuclear poweris going to be with us globally, its future isperhaps less certain in the United States. Pollstaken fairly soon after the accident indicatedthat more than three out of four Ame'icansopposed building additional nuclear power plants,and more than half favored phasing down orclosing those plants that are already in opera-tion.

In this kind of environment, the UnitedStates, perhaps more than any other country, issubject to what I would call a "hostage problem."Various elements in our society are able to keepa plant from opening or to close a plant that isopen by withholding the requisite support for theopening or for the continued operation of theplant.

Nuclear power was in trouble in the UnitedStates long before Chernobyl. It was in trouble,in part, because we developed our nuclear powercapability very rapidly. In the process, we didnot use the care in some areas in building theplants as well as they should have been built orin managing them as well as they should have beenmanaged. I emphasize in some areas.

Although the release at Three Mile Islandhas been estimated to be about one-millionth ofthe release at Chernobyl, the Three Mile Islandaccident, nevertheless, had a very significantimpact in terms of increased fears, in terms ofincreased costs due to the substantial retrofit-ting that had to be done, and in terms of the newoffsite planning requirements which were mandate?by Congress.

Those changes led to delays. Those delaysled to increased costs. As people began to loo!;at those increased costs, they began to see that,if you had economic concerns about the cost ofnuclear power, it was probably easier to use thesafety regulatory process to stop it than it wasto use the economic regulatory process. Inrecent years, we've seen many instances, at thestate and local level, of people turning to thesafety regulatory structures to try and stopnuclear power for what are essentially economicreasons and, in the process, perverting thesafety regulatory process that has been developedin the United States.

This development was possible partly becauseit came at an unusual time. There had been anoil crisis and threats'of severe oil shortages.We miscalculated. We did not appreciate theeffect the crisis would have on the traditionallinkage between the growth in the gross nationalproduct and the growth in use of electricalenergy. We did not anticipate the impact of therecession that ensued, and we began to conserveelectrical energy in ways we had not conserved in

the past. Then we had the decline in oil prices,a glut in oil, and a period in which we had avery modest excess capacity. Thus, the pressurewas off. You could play with nuclear power. Youcould slow down nuclear power without peoplebeing worried about brownouts.

In this context, then, I would like to focuson the future of nuclear power from a politicalperspective, from the perspective of the hostagethreat to nuclear power in the United States. Iwould argue that in the United States we havewhat is particularly, if not peculiarly, anAmerican problem. After Three Mile Island,Congress and the administration developed asystem for requiring offsite planning as acondition of licensing a nuclear power plant.FEMA, just being established at that time, wasgiven the responsibility of advising the NuclearRegulatory Commission (NRC) by making findings ofreasonable assurance that the public health andsafety could be protected in the event of anaccident. This was to be done Dy providingguidance, evaluating plans, doing limitedtraining, and evaluating exercises, and thenproviding findings of reasonable assurance, orthe lack thereof, to the NRC. In 1982 anotheraspect was added. The NRC might also, in theabsence of state and local plans, look at utilityplans. What it might do with those utility plansis not quite so clear.

When I refer to the hostage problem, I referparticularly to the possibility of a state or alocal government refusing to participate in theoffsite planning requisite for licensing and,thereby, keeping the plant from operating orclosing an operating plant, hence holding a planthostage. We sae a number of hostage problems,and they vary considerably in their nature.There are not too many plants that have notexperienced a little of this, when local govern-ments, more than state governments, delay theirrequired participation to see what they can getto enhance their overall emergency managementcapability. Sometimes their demands includebridges, new fire trucks, and all sorts ofthings. It is perhaps a little bit of a shake-down, but in most cases it is not carried toofar. Then there are situations where the processis simply being slowed until after the nextelection, because of concern for what the votersmight say or how an opponent might use the stateor local government's participation in offsiteplanning. Finally, there is the kind of situa-tion that we have had at Shoreham, and that wemay have in Massachusetts in respect to Seabrook.where a government entity wants to stop the plantfrom ever opening for economic, safety, orpolitical reasons, or a combination of these.

It is easy to overemphasize this problem.We have been incredibly successful, given thetremendous dispersion of power in this country,in bringing some 66 plants and a considerablylarger number of units online. We have hadsuccessful exercises at least twice at all theoperating commercial reactors, and we navecompleted the CFR 350 requirement for some two-thirds of the plants. We have developed andexercised a Federal Radiological EmergencyResponse Plan in the last couple of years, a planthat we did not really have until 1984. So we

have had considerable success. I think it iseasy to overemphasize the problems, the Shorehamsof this world, in that context.

What I have defined as a hostage problem upto this point is essentially the refusal of stateand local governments to participate in theoffsite planning process and thereby to stop aplant from coming online or to cause it to beclosed down. The other possibility is one thatwe have not seen, that the state and localgovernment will simply do a lousy job of offsiteplanning and exercising in a deliberate effort toshow how bad they are, so that FEMA and the NRCcannot make a finding of reasonable assuranceunder the conditions of such total ineptitude bythe state and local governments. When theydiscover that this approach would probably be amuch cheaper way of stopping nuclear power plantsthan the eight million dollars in legal fees onerounty has apparently paid, I suspect that wewill probably see some try this method. Indeed,some have already indicated to me that they arelooking in that direction.

The hostage problem is a particularly, ifnot peculiarly, American problem for severalreasons. In Europe you avoid this problembecause the national governments do one of twothings. They either mandate that the subnationalunits (whatever they happen to be) participate inemergency planning, or they simply tell them theyare going to assume that the subunit is going todo emergency planning and let it go at that. We,in contrast, require state and local governmentsto participate tf the NRC is gqing to be able tolicense, but .there is no way that we can mandatethat they participate. f

There are some other things that make thesituation in Europe easier. Europe has far fewerlocal government units, and their emergencyplanning zones are usually considerably smaller.Even if there were the same number of localgovernmentrunits, fewer of them would be includedin the emergency planning zone.

For all these reasons, the situation israther different in Europe than in the UnitedStates, but things are not all rosy there either.You may have a change in the national governmentand get an anti-nuclear; government, such as theBritish Labor Party and1 the German SocialDemocratic Partly, wliich both have indicated theyvery well might be.

In the 'United States, we chose to deal withthe problem of regulating the nuclear powerindustry by an enormous dispersion of power. Wedisperse power by requiring state and localgovernments to participate if plants are to belicensed and are to continue to operate. Wefurther disperse power by requiring that everytwo years a plant must go through what amounts toa relicensing process, because it must exerciseits plans and prove that they are still effec-tive. We disperse power at the administrativelevel. We disperse power within the NRC; I amnot going into that except to say that it hasmade for a somewhat complex, sometimes difficultto understand, and not always terribly well-coordinated process. And, of course, there arenine other agencies involved in emergencyplanning through the Radiological AssistanceCommittee process.

Then there is a dispersion of power withinCongress. That has created a particularlydifficult situation because, in the last coupleof years, it has been almost impossible to knowwhere Congress stands or whether the UnitedStates even has a nuclear energy policy. FEMAfound itself, on one hand, with a joint statementfrom our appropriations subcommittees in theHouse and Senate, telling the agency what itexpected.

"In particular, the Committee isconcerned about situations where Stateand local government arbitrarilyrefused to develop radiologicalemergency preparedness plans or toparticipate in the exercise or imple-mentation of such plans. The Committeedoes not believe that State and localgovernment entities should be permittedto veto the operation of commercialnuclear facilities simply by refusingto participate in the preparation,exercise, and implementation of suchplans. In that regard it is theCommittee's intention, in its review ofsuch plans, FEMA should presume thatFederal, State and local governmentswill abide by their legal duties toprotect public health and safety in anactual emergency. However, whera Stateand local participation in the exerciseor implementation of offsite plans isinadequate, the Committee intends forFEMA as a last resort to coordinate thesupplemental assistance to Federalagencies that are expected to providerequisite resources within theirauthorities."

Of course if you go to court as many times asFEMA has and other agencies have, it is in partbecause there is some question of exactly whatyour authorities are.

On the other hand, there is the Authoriza-tion Committee, chaired by another member of theHouse from the same party as the one who co-signed the statement I just read to you, indeedfrom the same state. This chairman has used thehearing process and other means (which I thinkany person who can tie his shoes and wave "bye-bye" would assume were meant to intimidate) totry to prevent the policy proposed by theAppropriations Committee from being carried out.The Congress simply cannot get its act togetherbecause power is so dispersed, not between thechambers, not even between the parties, butwithin the parties, often between members fromthe same state.

What are our options, within this context?We can follow the option of compensatory actions,using the state to take compensatory action forlocals that will not participate. That is whatthe State of New York did for Rockland County atIndian Point. That is what New Hampshire hasbeen doing for some of the communities in respectto Seabrook, and that continues to be a viableoption. There is some indication that we areseeing locals compensate for state inaction in

some other places, perhaps to some degree in Ohioand on some occasions in California.

Now we are getting utility plans. Thesewere first suggested by the House in an 1982-83NRC authorization bill as something that the NRCmight look at, but there is great conflict inCongress over just what this means. Untilrecently the issue has been hung up on "reason-able assurance," whether FEMA or the NRC couldfind a reasonable assurance that public healthand safety would be protected by the use of theutility plan when the state and locals say theywill not participate. There are the problems ofwhether the utility can carry out its plan if itdoes not have police powers and whether state andlocal governments, even if they do participate,can be said to give reasonable assurance if theyhave not exercised and have not demonstrated anyfamiliarity with the plan.

Recently, the NRC instructed the AtomicSafety and Licensing Board (ASLB) to review theutility plan associated with Long Island LightingCompany and Suffolk County. The legality of thisplan is being challenged in the courts. It gavethe ASLB two assumptions to follow. One was thatthe state and local governments would respond inthe event of an incident to protect their people,and the second was that they would use theutility plan as the best plan available to themat that time for that response. What will resultfrom that review remains to be seen, but itinvolves a considerable leap of faith and maytake us a considerable leap further in dealingwith the hostage problem. It certainly takes uscloser to the European approach, but the outcomeis unclear. How it would work if state and localgovernments just decide to uo a poor job, insteadof completely refraining from participating inemergency planning, is aiso unclear.

In conclusion, then what are our options?One, I suppose, is to continue to muddle along,satisfied by the fact that most plants have beenbrought into service and have continued to be inservice. A second approach is much greaterfederal involvement. We have seen a number ofoptions proposed here ranging from the develop-ment of some kind of federal brigade or SWAT teamwhich would be ready to go in and providedirection and control, using state and localresources, or perhaps going beyond that and usingcivil defense authorities. For example, severalcongressman have suggested developing a federalteam that would go in, perhaps in conjunctionwith nationalization of the national guard, andliterally take over. All of these things have atime factor. We know that under the FederalRadiological Emergency Response Plan, valuable asit is, it takes six to eight hours before you canget enough people there to be of much support tostate and local governments. So one of theproblems with this approach is achieving a timelyresponse with people who really understand theproblem and can marshal the local resources.

The third option is to give the state andlocal governments greater responsibility foremergency preparedness, along the lines of whatis done in Europe. Offsite planning could betaken out of the critical path for licensing sothat state and local governments could not stop anuclear power plant from starting or continuing

to run by refusing to participate. The politicalpressure would then no longer be on them to stopthe plant, but instead, political pressure wouldbe on them to do as good a job as possible ofmaking certain they were prepared in the event ofan incident at that plant. Part of this schemewould include making a determination at the timethat the construction permit is issued that anoffsite emergency plan for that area is feasible.

Finally, if you simply take the state andlocals out of the critical path and say it is upto them to protect their people, what are thefederal government's obligations in that regard?It seems to me there are a number. First of all,it would be appropriate to look to the industryfor the funding of the federal, the state, andlocal roles in terms of emergency planning. Thatis not regarded in all circles as a terribly goodidea, but I think much of the industry would besupportive of that if it were done in theappropriate way. Second, continue to improve theguidance that we are giving state and localgovernments. Frankly, FEMA has had so manyresponsibilities for trying to put out firesarising from hostage problems during the lastcouple of years, that the agency has not beenable to devote resources to improving theplanning guidance or developing the guidance thatgood offsite emergency planning demands whereguidance is not available. The third obligationis training. While we provide limited trainingat the federal level, this is a place where wecould be of increasing assistance to state andlocal governments and make them feel more secureabout meeting their obligations. One thing thatwe do not do now is provide very much technicalassistance in the development of plans and theexercising of those plans. We send the state andlocals off on their own and the utility off onits own, and after they come in with their besteffort, then we critique it. It would make a lotmope sense to be providing more technicalassistance in the development of the plan and theexercise process. And finally, in this scenario,we would expect the federal government tocontinue to evaluate the plans, evaluate theexercises, and publish the results of thoseevaluations for the public.

It seems to me that the current ad hocsystem is not working as well as it should.Congress is placing unreasonable and conflictingburdens on administrative agencies that aresimply impossible to reconcile. FEMA's resourceshave been reduced and diverted to crisis casemanagement at the cost of systems development.The threat of the hostage situation is inevitablygoing to create an environment where, if any morenuclear power plants are developed, they willprobably be sited adjacent to existing plants.This may be done even if these sites are notparticularly good places for additional powerplants, simply because there would not be theproblem of going through a whole new offsiteplanning process as would be necessary for a newlocation.

After serving as Associate Director of FEMAwith responsibility for offsite emergencyplanning for three years and since I have beenout of that position for a week, I am free togive you an jnvarnished version of the situation.

These are some of the thoughts I have as to wherewe are, how we got here, and where we should begoing, in order to preserve the nuclear option.

ANS Topical Meeting on Radiological Accidents-Perspectives and Emergency Planning

The Industrial PerspectiveJames A. Wick

My objective is to share some perceptionsabout major accident planning and response fromthe chemical industry perspective. Let's beginwith a quick test. I would like everyone in cheaudience to participate. Last summer, I gavethis same test to over 2000 chemical plantmanagers and their assistants. We were able totell a few things about them, and I was wonderingwhether people in the nuclear business are similaror different. Let's see whether we can predictsomething about yea.

Write down the very first thing that popsinto your mind. I don't want you to think aboutthefie answers. (1) Name a color. (2) Name aflower. (3) Name a piece of furniture. (4)Name an animal in a zoo. Now indicate, by a showof hands, how many answered question No. 1 with"red." About 85 or 90%. How many answeredquestion No. 2 with "rose"? Again 85 or 90%.How many answered question No. 3 with "chair"?About 85%. How many answered question No. 4with "lion" or "tiger"? The majority.

Most of you, 85 or 90%, answered as Ipredicted. You are utterly predictable as agroup, just as were the chemical plant managersI mentioned. They answered in about the sameproportions with the same answers.

How many of you answered all four as Ipredicted? Only a few. How many answered noneas I predicted? Again, maybe a half-dozen. With85-90% answering as I predicted, only a very fewanswered all or none of the questions as expected.You are still very much individuals! We have tokeep that in mind. As a group, people arepredictable when it comes to emergency responseand planning, yet we are very much individuals inour lives and in our response to this whole issue.

I want to give you another quick test. Foldyour hands for me and see which thumb is on top.There is no incorrect answer. Move your handsapart and quickly bring them back togetherwith the other thumb on top. Hard to do, isn'tIt doesn't feel right. Fold your arms. Seewhich arm is on top, then move them apart and tryto fold them the other way. It's the samephenomenon.

We are very much creatures of habit ineverything we do. I would venture to say thatyour routine in the morning for the first few

minutes after getting out of bed is a matter ofhabit. When you get up, you do the same thingsevery morning in the same sequence, such asbrushing your teeth, putting on your pants, ortaking a shower. We are creatures of habit, andwe have to take that fact into consideration aswe go forward in emergency response planning.

For the third little test, 1 would like toask you to think back for just a moment to yourfirst formal education, whether it was kinder-garten or first grade. Can you tell me yourteacher's name? Most of you answered that youcan do that. How long has it been since youthought of that person? We have tremendous graymatter potential. In fact, safety programs inboth of our industries are very much oriented toimparting information into that gray matter forrecall when the time is right. When that time isright, there may be a signal, some sort ofstimulus, that tells sc.:ieone to put on gloves orto look r"t for a particular hazard or to be sureto wear breathing apparatus in that situation.We want that information recalled at the appro-priate time. We should never underestimate thegray matter potential in people. We shouldcontinue to work with all people in an attempt tomaximize the use of their talents and skills andthe potential that has been given to all of us.

I. COMPARING OUR INDUSTRIES WITH PUBLICPERCEPTIONS

There are many similarities between you inthe nuclear industry and us in the chemicalindustry. Since the accident in Bhopal, we havehad to learn a lot. You have blazed a trail, andwe thank you for that. Our businesses are indeedalike in that we have the potential to do greatharm if we mismanage them. We are also alike inour dedication to managing our technology safely,and since the tragedy in Bhopal, we are alsounder the tremendous public and governmentalscrutiny that has developed due to a lack ofconfidence and trust in us. This fear of theunknown is commonly called "chemophobia" in thepetrochemical industry.

The nuclear industry suffers from some ofthat same kind of fear. For example, here is a

11

12

recent UPI clip from my hometown newspaper. Letme read it to you. The date line is Seabrook,New Hampshire.

"More than a dozen frightened and some'hysterical' residents reminiscent ofthe 1938 broadcast classic 'War of theWorlds' called the police during asimulated news broadcast of a meltdownat the Seabrook Power Plant, policesaid Monday. Seabrook Police Chief,Paul Cronan, said the Sunday nightradio broadcast, a political advertise-ment for an anti-Seabrook MassachusettsDemocrat, was irresponsible. TheSeabrooK police station received 15-20calls from Massachusetts and NewHampshire residents who heard the 8:30p.m. broadcast and believed a realnuclear disaster had occurred."

None of us like to see those kinds of reactions.

II. RESPONDING WITH CAER (COMMUNITY AWARENESSAND EMERGENCY RESPONSE)

We are faced with an international issue oftrust trat can only be solved at the local level.We have: taken this lesson to heart in ourindustry and embarked on raijor industry-wideinitiatives at the local level. These proactive,positive initiatives are targeted to meet realcommunity needs in information and in emergencyresponse and to build public trust throughimproved relationships and cooperation. We workvery closely with and through our trade associa-tion, the Chemical Manufacturers Association, andallied trade groups. We have been at it forabout a year with renewed vigor, and the prelimi-nary results look positive.

Our industry has discovered several of your'esources to help us along the way. We have usedconsulting firms who have cut their teeth in thenuclear industry on emergency response issues,and they have been able to u«lp us. Theyunderstand the differences between our indus-tries, and they have helped us develop guidelinesand materials for our local initiatives.

The cornerstone of this effort is our CAERinitiative, Community Awareness and EmergencyResponse. This initiative is a ten-step processof how to go about organizing and implementing abetter community emergency response plan at thelocal level. It doesn't say what the end productshould be; it only describes the process. It isliberally based on FEMA-10, the Federal EmergencyManagement Agency's contingency planning andguidelines document, which is presently beingrevised. CAER is focusing on improving localemergency response in areas where we havefacilities, not just for chemical emergencies,but for the whole family of emergencies, whetherthey be natural, technological, or nuclear. Infact, we have been successful in piggy-backing ona few of the nuclear emergency response prepared-ness systems that already exist.

We have 1500 chemical plants today that areactive in programs of the CAER type. In mycompany alone there are 100 U.S. sites, of which97 are very active in these kinds of programs.

About 35 have conducted community-wide, multi-disciplinary, multisgency drills within the lastyear.

Another major initiative has been theNational Chemical Referral and Information Center(KCRIC), an enhancement of CHEMTREC. CHEMTREChas a toll-free number for chemical transporta-tion emergencies and provides information and alink to the shipper and/or the manufacturer. Thesecond activity of NCRIC is the Chemical ReferralCenter, which has a toll-free telephone numberfor non-emergency information that anyone can use.The third NCRIC activity is CHEMNET, an indus-trial mutual-aid pact that dispatches the nearestqualified hazardous materials response team toincidents occurring anywhere in the UnitedStates. Where we didn't have facilities andteams, we have been able to hire contractors tocover tho.se localities. The last NCRIC activityis improved first-responder training materials.

Many people in my industry have onlyrecently become aware of the tremendous resourcesof agencies like the Federal Emergency ManagementAgency, which has a lot of talented people thatwe have been able to use to help us developguidelines and procedures and new materials forour plant managers. For example, they have aweek-long course on how to conduct a community -wide drill. That course was too long for most ofour plant managers to attend. So, using thecourse as a base, a contractor from the nuclearindustry, and people from the chemical industry,we developed a 24-page guide. We intentionallymade ic short, so a manager could read it in 30minutes and get an overview of how to organizeand conduct a drill in his community.

III. CITIZEN EXPECTATIONS

One of the things that we are learning iswhat citizens expect of local government. Inmajor emergencies, the area outside the fence isthe responsibility of the local government. Ifwe are to make improvements there, we have tomake improvements in cooperation with the localgovernment. Before we can do that, we have tounderstand what demands are being placed on thelocal government. We concluded that there areabout seven things that citizens expect fromtheir local government officials.

1. They must provide advance information,such as alerting signals, response capabilities,and evacuation plans. You have larger emergencyplanning zones to deal with than we have. Weusually think in terms of a mile or less. But,in alerting the community, you are way ahead ofus. You have learned some lessons the hard way,and we are learning from you.

One of the things that I tell our plantmanagers is that if they get involved in acommunity system of alerting that involvessirens, there are two things to remember: (I) putin a Cadillac, not a cheap siren system;and (2) make sure that the button that activatesit is in the hands of the local officials. Ifyou need better communication with them, put in ahotl ine to the local officials.

Here is a letter to the editor that appearedin my local newspaper which demonstrates thefirst point well.

13

"On August 22 at 11:15 p.m., my familyand I were wakened by several long,loud blasts, which eventually became acontinuous scream, from the nuclearwarning siren at the Townsend FireCompany. Since the siren is one of thesystems designed to warn residentswithin a 10-mile radius of the Salem/Hope Creek power facility of malfunc-tions at the plant, we were thoroughlyand justifiably frightened. Afterturning on the radio and getting noinformation, we made several phonecalls, all of which reassured us thatthe problem was in the warning systemand not the plant."

He goes on to indicate that the siren continuedto blow for about two hours, and he lost anight's sleep. That is not an unusual occurencewith siren systems around nuclear power-gener-ating facilities, according to the news clipsthat I read. In fact, they had some problems atthe Limerick Station, and a fellow wrote to thelocal newspaper in West Chester, Pennsylvania.He said, "Look, if these guys can't get thissiren system straightened out, what makes themthink we have confidence that they can run areactor."

False alarms produced by warning systemmalfunctions are a major reason that I have toldour plant managers, "If you are putting in asiren system, put in a Cadillac. Don't tell meit can't be done!" There are 25,000 firedepartments in the United States, and 24,000 ofthem are volunteer departments that are alertedthrough tone-activated siren systems. There areabout 1 million volunteer firefighters; if theyget v.p in the middle of the night throughinadvertent activation of their siren, they donot tolerate that very long before they makecorrections to the system. There are thousandsof remotely operated sirens that do not triggerinadvertently very often.

The whole psychology of alerting people andgetting them to react in a positive, plannedfashion can be a problem. In our business, oneof the recommended protective actions may be tostay indoors, shut the windows, and turn off thefans--that is your safe haven. The chemicalcloud will pass, the fumes will dissipate, andthe area will be safe. These actions are a saferalternative to evacuating, getting in the car,and getting to the intersection at the sams timeas everyone else. As you well know, that can bea real problem.

Behavioral scientists tell us that peoplehave a natural reaction to fight or flee whensubjected to danger such as from a clovJ of.fumes. They can't fight it so their reat.: ion isto flee. The scientists suggested that if wedevote the same amount of energy and effort totraining people to evacuate more effectively andefficiently rather than training them to stayindoors, we might have a safer situation. We hadnever thought about that. We are learning, andwe are trying to learn more each day about thosekinds of difficult questions that we face at alocal level.

2. The second thing that citizens expectlocal government to do is to assess the magnitudeof the emergency quickly and accurately. If youhave bad information, you will have two or threeunwanted results. For example, let me tell you astory that I heard at Johns Hopkins. They haddispatched a team to Bangladesh after a tsunamiwent through there several years ago; about247,000 people were estimated to have been killedin the Delta area of Bangladesh. However, as theearly word of the tremendous tragedy spread, theJapanese decided they would respond, and theysent a complete field hospital with doctors,nurses, and staff to Bangladesh. The Germansthought that was a good idea, so they alsodispatched one, and the Americans, not to beoutdone, sent two. So now Bangladesh had fourtotally equipped field hospitals. But only thefittest had survived the tsunami. The veryyoung, the very old, the very weak, and femaleswere killed in large numbers. The survivorstended to be physically fit males who had clungto a tree or something for 12 to 14 hours. Theirinjuries were really abrasions on the inside ofthe forearms and on the chest. They needed onlya little iodine. Bangladesh did not need fieldhospitals; they were sent due to poor initialassessment.

On that same emergency, incidentally, out ofone of the relief planes came a container likethe back of a tractor-trailer rig, and In thecontainer was a whole load of teddy bears from atoy manufacturer. Teddy bears are not recognizedas a toy in Banjladesh, and the children wouldnot use them. So then they had to divertresources from the emergency to repackaging thattractor-trailer load of teddy bears and shippingthem back out of the country.

Quick and accurate assessment of themagnitude of the emergency is critical. Gettingthe resources that you need and getting them inplace fast is what we arc talking about. Thatmeans having knowledgeable people, good informa-tion systems, and good communication systems.

3. The third major expectation of citizensis that we provide emergency mitigation services.These services include fire, police, emergencymedical, and shelter.

4. The public expects to be kept informed.We have an important role to play in our indus-tries . Too often the spokesman during anemergency is a university, a health department,or a government official, who may or may not knowa thing about your operation. This is the"expert" who is quoted on IV. They are great atspeculating in response to questions like "Howbad could it be?" Therefore, our industries havea role in providing spokespeople who are knowl-edgeable, truthful, and accurate, to help keepthe public and government officials informed.

5. People expect the local government torapidly restore services and utilities, even whenthose services are not their direct responsibil-ity, for example, electrical service.

6. Citizens also expect the local govern-ment to provide access to recovery services suchas family unification, counseling, insuranceclaim preparation, and relief assistance.

14

7. Last, people expect local government Coprovide Information on preventing similar majorimpacts in the future.

We have a role to play. If we know what isexpected of local government, we can identify ourvole and begin to interact in a positive fashion.

IV. DRILLS

The best way to test where we are and wherewe ought to go is to conduct drills. You in thenuclear industry are certainly not strangers tomultiagency drills. Nuclear power plant drillsare conducted to demonstrate performance; thedrill must demonstrate performance to get thechecks in the blocks in order to continue thelicense. That is a different kind of drill fromthe one we want to conduct. We have differentobjectives for our drill. Ours asks a differentquestion, which is "How can we improve thesystem?" We are trying to have it achieve thosethings that the Federal Emergency ManagementAgency has said that drills ought to do. We'vetried to keep our drill scenarios and drillimplementation positive and voluntary. Thepeople become stakeholders in building a bettersystem and are better equipped mentally, as wellas physically, to deal with future emergencies.We build on what already exists, and we make itwork for all types of major emergencies, naturalas well as technological.

V. COMMUNICATING RISK

We've also had to learn a lot about communi-cation. Perhaps this is the time for anotherkind of test. Write a definition of "how fastis fast?" Difficult question, isn't it? We haveanother version of that question in regard todumps and environmental concerns. How clean iself an when you clean up a dump vsite? How safe issafe around your facility?

The nuclear industry is in the lead when itcomes to process hazards analysis. In fact, youare very good at sophisticated techniques such asfault tree anaylsis. That is great. We learnfrom that, but how do you communicate the resultsfrom a fault tree analysis to the general public?That has been difficult for us, and I presume ithas been difficult for you as well, because youhave had a difficult time convincing people thi.you are the safe industry that you are.

So we need to work on better ways tocommunicate. We need to learn some of thetechniques of the public relations profession.For example, a sign says "Discount for Cash."Isn't that a neat concept? But what would happenif they put up a big sign that said, "Surchargefor Credit"? That's the same thing as a discountfor cash, isn't it? But the public relationsexperts would never put up a sign that said"Surcharge for Credit." We have to learn tocommunicate in positive ways, rather thannegative ways, so we can help our communitybecome more informed and make better decisions.

VI. OUTREACH AND CRISIS COMMUNICATION "MUSTS"

I would like to share with you ten things,which I call common outreach and crisis communi-cation "musts," things that you must do in yourcommunity.

1. Plan. Plan for those events and planagain. It is a process, not a product. If youdevelop a plan, put it in a 3-ring notebook, andcall that the end of the process, you are introuble. The plan has to be a living, workingtool. Once I borrowed the emergency responseplans for a nuclear powex* generating station, andI had to use a suitcase to carry them back to DuPont because there were five three-ring volumes,plus a couple of smaller books. I defy you to gothrough those volumes at 3:00 a.m. to determinewhat the appropriate procedures are. The localfire company took those iive volumes and ex-tracted about 30 pages that were meaningful tothem. Then they had a good working document anda tool they could use. The planning processincludes how you communicate with others beforeand at the time of emergency.

2. Accept Responsibility. Accept theresponsibility for a role in crisis communicationand crisis management. For instance, one of thethings we try very hard to Jo is to get cur plantmanager to talk to the loc^l officials. The plantmanager is the most credible representative of DuPont in any community where we have a facility--not a public relations official from Wilmington,not a safety official from Wilmington (thelocation of our corporate headquarters), not atechnical engineer, and not a consultant. Theplant manager and his staff are the chief liaisonto the community. They have to accept theirresponsibility for crisis communication andmanagement.

3. Take Charge. Take charge of yourfacility and establish a communications andcommand center. If you allow someone else to dothis and you allow others to take these responsi-bilities, you lose control of the situation andyou lose your opportunity for input.

4. Never Underestimate the Media. Themedia are there and they believe strongly, "Thereis a story here and we're going to get it." Youare in the driver's secit; you can provide thatstory. You need to learn how to provide it in apositive, proactive fashion. When I have gone toaccidents, I have helped provide the story, evenwhen we were not the lead agency or the leadcompany involved. We can help. We can do thisin very positive ways that come across well.

For example, holding an onscene pressconference is awful. In most media markets, youare going to have at least three TV stations,several newspapers, and a couple of radiostations represented. And then there is you, ifyou are the spokesman. You go out and give themsome details, and suddenly there are 97 millionquestions being thrown at you. The pressconference almost degenerates into a riot. Onereason for this, if you think about it and watchyour local news, is that Channel 10 will sticktheir microphone right into your face (don't

15

forget the camera is over their shoulder) and askyou a question. Channel 6 is also taping thatanswer, but they don't like to show Channel 10'smicrophone. The next thing you know Channel 6is doing the same thing, maybe even asking thesame question. This goes on and on, and itbecomes very difficult to manage.

I can think as fast as any single reporter,but there are five or eight of them out there,and I have trouble keeping up. They can get mein a trap in a hurry. So one of the things wedo,once we've given them all the straight factsin press conference fashion, is give themexclusive interviews, one on one. It takes aboutthe same amount of time to give exclusiveinterviews, and it is a lot more civil. Thatnight, when the clip comes on the 10:00 o'clocknews, it shows as an "exclusive" with the Du Pontofficial. They love the word "exclusive."Incidentally, the first three questions, maybeeven the first four or five questions, from everyreporter are roughly the same, so you answer thesame question over and over again. However, itlooks a lot better that night. It looks likeyou're more in control.

Those are the kinds of lessons that we needto learn in order to communicate with the media.If we communicate effectively with the media,then we communicate effectively with the public.Also, we can't forget our employees. We have anobligation to communicate with them as well.Although we may use a different mechanism, wemust communicate with them.

5. Anticipate the Questions and Concerns.What is really on the minds of the people? Whatare they worried about? Are they fearful fortheir child's safety or well-being? Are theyworried that the industry will fold up and leavetown? Those are things of the kind that you canunderstand only from knowing the community.

6. Grasp the "Real" Iss.ue. You have tograsp what is really bothering the people. Forexample, you have a man living near a site withgroundwater contamination problems. Is he reallyworried about the groundwater being contaminated?No, he's worried about his daughter who has beendrinking that water. So, if you want to movethat person from a highly energetic negativeposition to a negative but low-energy position,perhaps the things to do are (1) provide him withbottled water until the scientists can get Inthere and examine his well and do all the thingsthat good scientific people do and (2) pay forsome medical screening at the nearby hospital forhis daughter. It would not cost much, andsuddenly the very vocal neighbor is not so vocal,because his real concerns have been dealt with.There is still the concern about groundwatercontamination, but now we have more time, andreason and logic can prevail.

7. Be Accurate and Honest. If you try topull the wool over people's eyes, they will catchyou, and your credibility will be lost.

8. Never Speculate, nor Place Blame. Neverspeculate nor place blame during a crisiscommunication situation. Speculation Is alwaysbad business. Let's talk instead about what ishappening and be realistic and honest.

9. Be Able to Articulate Your Positions andPolicies Clearly and Simply. You must certainlyexplain your position clearly and simply as youdeal with local officials. You might have thebest statement of corporate or departmentalpolicy dealing with this kind of situation, butmy guess is that it runs about four lines long ina paragraph that covers the first half of thefirst page of the manual. That material ispretty difficult to get on the evening news or tocommunicate effectively to local officials.Communicate in straightforward layman's terms.

10. Practice. In crisis communication andcommunity outreach, you must practice. You mustpractice to be more proactive.

VII. BENEFITS

But why bother doing this at all? Becausewe can all benefit.

Our companies will benefit from proactive,positive community relations. We will have morefreedom to operate. We won't have as manylegislative and regulatory initiatives as wewould have otherwise. Our employees will beproud of our companies, our agencies, or ourdepartments as we do the right thing for ourcommunity. And certainly, customers will preferto do business with us because we are reputable,we have integrity, and we are interested in doingthe right things.

The community will benefit. If you wereintroducing risk into the community and thecommunity had no say, then there would probablybe some perception of inequitable distributionof that risk; such a perception often results intremendous backlash and negative reactionto your initiative. When you begin to work'.inthe field of emergency response, planning, andpreparedness, the community now has a'say in itsidestiny. The community will alsp benefit,whenthey receive good information and develop'a stakein the venture. Not only do your'employees havea stake in what you are trying to accomplish, butcommunity regulators and elected officials willalso have a stake in making it work. It becomesa win-win situation. Certainly, the communitywill win when the tax base has been expandedbecause the community is recognized as a goodplace to do business.

And certainly, individuals can benefit;individuals such as you and me. We will grow inour knowledge of other people and in the knowl-edge of our government, in the elected environ-ment, as well as the regulatory environment. Wein the emergency response field have an oppor-tunity to apply our skills to solving thetechnological problems that are facing oursociety. Certainly, emergency planning improvesour stake in the community in which we live.After all, we live, work, play, and raise ourfamilies in these same communities. You bet wecare. In fact, today we cannot afford not to beinvolved. We must be proactive and leaders inthe local decision-making process.

Section I

Accident Experience

Chairman: Joseph Hendrie

Brookhaven National Laboratory

ANS Topical Meeting on Radiological Accidents—Perspectives md Emergency PlannBg

Use of Radiological Accident Experience in EstablishingAppropriate Perspectives in Emergency Planning

/ . M. Selby, D. W. Moeller, E. J. Vallario, and J. G. Stephan

ABSTRACT Within a nuclear facility, an emer-gency can range from a situation that only in-volves the employees of that facility to aseries of events that have both onsite andoffsite consequences. Analyses of nuclear andnon-nuclear emergencies can provide valuableinformation on the causes of, as well theproblems encountered during emergencies. Re-ports on facility emergencies indicate that upto 90% involve human error. Such events occurmore frequently during the night shifts or onweekends. These occurrences may result fromthe absence of experienced personnel as wellas the reduced alertness of onsite personnel.Therefore, this paper emphasizes the humanelement in a review of accidents that haveoccurred at nuclear facilities including Wind-scale, SL-1, the Recuplex criticality, theWood River Junction criticality, the BrownsFerry fire, Three Mile Island, and Chernobyl.These accidents are described, and their con-sequences are evaluated. The information ob-tained from these evaluations may be usefulfor in- elusion in nuclear plant operating andtesting procedures.

INTRODUCTIONAn emergency can be defined as a serious

situation that develops suddenly and unexpect-edly and demands immediate attention. Withina nuclear facility, an emergency can rangefrom a situation involving only facility em-ployees to an event having both onsite andoffsite consequences.

The analysis of nuclear emergencies canprovide valuable insights on a generic basis,into the causes of such events. In some in-stances, nuclear emergencies can be initiatedthrough failures of equipment. However, datashow that human error is a major source ofmany industrial accidents, and yet human re-sponse after such accidents can also play amajor role in bringing accident situationsunder control.

In the chemical industry, data show that upto 90% of all accidents are induced or furthercomplicated by human error (Joksimovich 1986).In the airline industry, up, to 80% of allcrashes are thought to involve erroneous humaninteraction (Lewis 1985a). Similarly, humanerror appears to have played a major role inessentially all nuclear accidents.

Human error is not confined to the Westernnuclear industry; this fact was demonstratedby the recent Chernobyl accident in the SovietUnion. While the exact cause of the accidenthas not been reliably reported, a number ofSoviet plant officials were reportedly removedfrom their jobs for responding to the accidentimproperly.

In the preparation of this paper, theauthors have examined a range of nuclear acci-dents from the 1950s to the present that werereported in the literature, and have identi-fied a number of contributing factors whichaffected human judgment during these events.One common thread found in a large number ofaccidents is the time of occurrence. The datashow that such events, whether severe acci-dents or operational incidents, appear tooccur more frequently during off normal hours(the graveyard shift, weekends, or holidays).Accidents seldom occur during the day shiftwhen the full management team and senior oper-ations personnel are present. Because acci-dents seldom occur during the day shift, thosefacility employees most expert in coping withthe situation may not be available, and thenormal chain of command may be disrupted. Atseveral nuclear power plants, it was alsoobserved that new or less experienced tech-nicians are assigned to the back or graveyardshifts. The lack of experienced human re-sources and the pressure of an accident situ-ation can be enormous on those individuals whoare faced with making important decisions.

An in-depth review of the literature con-ducted for the U.S. Nuclear Regulatory Commis-sion (NRC) by staff members at the PacificNorthwest Laboratory (Lewis 1985b) determined

19

that human errors generally increase at nightand the chance for error is significantlygreater if the worker has already been workingfour or more hours before midnight. The high-est error rates were reported to occur between3 and 6 a.m. The most efficient work is typi-cally performed during the day. When a personis required to work at night and sleep duringthe day, both work performance and sleep werefound to be degraded.

Researchers at Harvard University (Squires1986) have reported that 30 percent of theoperators of heavy equipment fall asleep whiledriving on the night shift. During a recentannual meeting of the Academy of BehavioralMedicine Research, members of the HarvardMedical School research staff observed thatthree of the recent major disasters — ThreeMile Island, Bhopal, and Chernobyl -- all hap-pened at night. The concern expressed by be-havioral scientists was not only that loweredalertness might cause accidents at nuclearfacilities during off-hours, but that the con-current human ability to detect and correctthe problem in a timely manner might also bereduced (Squires 1986). Dr. C. A. Czeisler ofthe Harvard Medical School, Department ofMedicine, has conducted extensive and indepthstudies of the effects of rotating shift workschedules (Hearings 1983). Dr. Czeisler ob-served that workers on night or rotatingshifts experience adverse consequences becausetheir circadian rhythm, the physiological func-tions such as body temperature, hormone secre-tions, cell division, antibody formation,etc., varies over a 25-hour period.

These physiological functions are regulatedby the body's internal clock as well as by in-dicators from the environment, such as lightand temperature. According to Dr. Czeisler,whenever a worker's normal wake/ sleep sched-ule is interrupted, there is a mismatch bet-ween the body's ability and the demandsplaced upon it in the work place. Stress,gastrointestinal disorders, low morale, highrates of accidents and illness, as well as lowproductivity, result from this mismatch.Sleep-deprived shift workers often experienceinvoluntary lapses of wakefulness: while theyappear to be awake, they may actually bedrifting in and out of sleep. Because ofthese lapses they may be impaired in theirability to respond to warnina signals orlights (Czeisler 1980; Rafferty 1983).

Researchers in Sweden (Hearings 1983) havereported that train crewmen fall asleep oneout of every six hours during night runs.While asleep, the drivers maintain full pres-sure on the accelerator, and were found to beunresponsive to warning lights. Czeisler re-ported in field studies of 1500 workers at anumber of industrial facilities, that over 55%of the workers admitted to "nodding off," orfalling asleep on the job during any givenweek. Czeisler1s efforts have since focused

on identifying effective counter-measures thatwould mitigate the effects of the interrup-tions in normal sleep patterns experienced byworkers during off-normal hours.

Another common thread to most accidents isthat, no matter how well emergency scenariosare developed and emergency planning exercisesare conducted, no scenario that fits an emer-gency situation exists (Vallario and Selby1981). During an emergency situation, thereis little time to extensively consult theemergency operating plan to decide what to do;the major response will have to be based onthe quality of the training imparted to emer-gency response personnel from plan and pro-cedure implementation, as well as tnorougSiperiodic drills and exercises.

How an individual will respond to an actualemergency situation is another uncertainty.In Japan, analysis of safety evaluations per-formed at commercial installations revealedthat an appropriate response of operators toan accident could not be expected for at least10 minutes after they became aware of the emer-gency situation (Katsuda, Suehiro and Taniguchi1984). Part of this delay is due to the timerequired to effectively analyze the situation.Another portion of the delay is likely to re-sult from the reluctance of the operators toacknowledge that they have an emergency. Irre-spective of the number of emergency drills andexercises conducted, the real situation may re-veal flaws in the emergency plan and in thetempered response. Therefore, everything thateven borders on an emergency at a facilityshould be treated as an emergency.

A REVIEW OF SELECTED NUCLEAR ACCIDENTSA review of a few of the more f ami liar

nuclear incidents is included to illustratethe sequence of events, the consequences, andthe lessons learned. All these accidents be-gan during night shifts, on weekends, or holi-days, which suggests that possibly fatigueand/or the absence of more experienced per-sonnel could have been among the causes oftheseincidents.

Windscale Works of British Kuclear fuels Acci-dent - 7:25 p.m., Monday, October 7, 1957

Th"e Windscale Works Ts located on" thenorthwest coast of England. The accident atthis facility occurred during a routine main-tenance operation in the Head End TreatmentPlant. Two air-cooled graphite-moderateduranium reactors were being operated at thesite to produce plutonium. At that time, itwas known that graphite will store energy asthe result of neutron bombardment (the so-called Wigner energy) and that the controlledrelease of such energy could be used for an-nealing the graphite structure. Experienceshowed that achieving a release of all thestored energy in a single step was difficult.

21

Therefore, a second heating was usually ap-plied.

Unfortunately, at Windscale the secondheating was initiated prematurely by the oper-ator because thermocouples used for indicatingtemperature in the graphite structure showeddecreases in graphite temperatures when infact, the temperature in the graphite struc-ture was rising. This temperature rise wasnot detected by the operator because of insuf-ficient core instrumentation. The prematuresecond heating caused a uranium fuel cartridgeto fail. The fuel cartridge and graphite re-acted with air and initiated a fire.

The first indication of the accident wasprovided by an air sampler located 1/2 milefrom the plant. The instrument indicated anincrease in beta activity. When the workerssuspected a failed fuel cartridge, a remotescanning device was used to locate the cart-ridge. The scanning device jammed, and theworkers donned protective clothes, peeredthrough a plug at the face of the reactor, andfound the fuel was red hot. This discoverywas the first indication that there was a fire.By that time, the fire had been underway fortwo days. The intense heat distorted the fuelcartridges and they could not be discharged.However, the cartridges in adjacent channelswere successfully removed. By October 10, thespreading fire involved approximately 150 chan-nels in a rectangular cross section.

Although several schemes were attempted,efforts failed to extinguish the fire. Amongthese efforts was an attempt to cool thegraphite by increasing the air flow; this sim-ply caused the fire to burn better. A heatsink was needed to extinguish the fire. Thus,on the fifth day the reactor was flooded withwater and the fire was extinguished.

The reactor was extensively damaged. All35 personnel in the building at the time ofthe accident inhaled quantities of radioactivematerials and were contaminated externally.The release of radioactive materials, prim-arily iodine and noble gases, into the atmos-phere was widespread. Approximately 20,000 Ciof iodine-131 were released through a 405-ftstack. Exposure rates in the surrounding areameasured up to 4 mR/h. Radioactive gases suchas iodine-131, were transported directly toanimal feed and resulted in the contaminationof milk. The use of milk was banned within a30-mile radius of the plant, covering a totalarea of about 200 square miles. The mill, con-sumption in that area was banned for 25 days;in some of the more heavily contaminatedareas, milk consumption was banned for as manyas 44 days (Eisenbud 1973; Atomic EnergyOffice 1957).

Lessons learned from the Windscale accidentinclude:• A major nuclear accident can have worldwide

consequences, both in terms of the airbornereleases and in the public impact.

• There was insufficient temperature-indicat-ing instrumentation, which led the opera-tors into making the decision to initiate asecond Wigner release, thus initiating theaccident.

• The surveillance instrumentation used prov-ed inadequate and failed when an attemptwas made to locate and survey the potentialdamage.

• No written or sufficiently detailed in-structions, such as a Pile Operating Man-ual with special sections on Wigner re-lease, were available.

• An adequate heat sink is required to ex-tinguish a fire.

• The apparent makeshift emergency planningin England was not adequate to handlethis emergency situation.

• The countermeasures and public healthactions that had been developed for such anemergency were insufficient.

SL-1 U.S. Army factor Accident - 9:01 p.m.,Monday, January 3, 1961

Th"e Stationary Low Power Reactor (SL-1)accident at the National Reactor TestingStation in Idaho resulted in the first U.S.fatalities from such an event. The SL-: wasan U.S. Army research facility that was con-structed as a direct cycle, boiling waterreactor with a thermal capacity of 3 MW and anelectrical output of 0.8 MW.

The reactor had been used to providemilitary personnel with operating and mainte-nance experience, to obtain performance charac-teristics and fuel bum-up data, and to testcomponents for future improved reactors. Theintent was to develop small reactors like theSL-1 for use at Army installations in remoteareas such as the Antarctic.

Two weeks before the accident, the reactorhad been shut down after being operational formore than two years. Prior to shutdown, thecontrol rods in the reactor core ware stickingwith increasing frequency. The impedance wasthought to be cause by the bowing of boronstrips attached to the fuel elements or anaccumulation of dirt or corrosion.

On the day of the accident, personnel wereinserting cobalt flux measuring wires into thecoolant channels between the plates of thefuel assemblies. This task required theremoval of some of the control rod driveassemblies and control rods.

The nuclear excursion occurred when 3military personnel on the night shift attempt-ed to remove the control rods to prepare forthe reactor start up. The excursion exten-sively damaged the reactor core, did minordamage to the reactor building, and fatallyinjured the 3 people. Since none of themilitary personnel involved in the accidentsurvived, and because of the rather high expo-sure rates in the reactor building (approxi-mately 1000 R/h), it was some time before the

22

actual cause of the accident was established.In May 1961, a board of investigation

concluded that the withdrawal of the centralcontrol rod resulted in the excursion. Al-though the building was not designed for tightcontainment, it was effective in reducing thespread of gaseous fission products to offsitelocations. An exposure rate of approximately50 mR/h was observed about 3/4 mile from thefacility (Eisenbud 1973; Vallario and Selby1981; Cottrell 1962).

The lessons learned from SI-1 accidentinclude:

• The design, in which the withdrawal of asingle control rod could cause the reactorto go critical, should be avoided in thefuture.

• Repeated malfunctioning of control systemsshould result in remedial actions. Thedecision to continue operation of thereactor was made at too low a decisionlevel and without review by qualifiedsupervisory personnel.

• Potential accidents need to be anticipatedthrough safety studies and analyses on acontinuing formal basis. Source termsshould be estimated for a range of acci-dents to permit the development of plansfor realistic emergency preparedness.

• Specialized remote equipment and techniquescould be used to determine the condition ofthe reactor after the accident. Thisdemonstrated the need for the developmentof a robot that is capable of operatingwithout deteriorating or failing in a rela-tively high radiation field.

• A remotely operated retrieval system forthe nuclear accident dosimeters would bebeneficial.

• Shielded medical facilities are requiredfor adequately protecting the medical staffwho may be called upon to treat a contami-nated accident victim.

• Emergency preparedness instrumentation thatis able to assess the high radiation fieldswhich may be present during an accidentneeds to be developed. The studiesperformed after the accident indicated thatthat exposure rates in excess of 1000 R/hcould be traversed safely by humans inlifesaving situations.

•. Survey instruments must also be designed tooperate in a wide range of environments. Atthe time of the accident the outsidetemperature was -20"^.

• Contamination control needed to be improved

to avoid the spread of radioactiveparticles. Some of these particles werecarried from the plant area on the tires ofmotor vehicles.

Recuplex Nuclear Criticality - 10:59 a.m.,"STturcfay, April 7, 1962"This was the first accidental nuclear ex-

cursion that occurred at any of the productionfacilities at Hanford. The accident beganduring a cleanup operation at the plutoniumwaste recovery facility, known as Recuplex. Aconcentrated plutonium solution was accident-ally pumped from a sump, where it was in thinslab geometry and therefore subcritical, intoa 45-liter tank, where it became critical. 17

The initial pulse consisted of some 10fissions, with repeated pulses following for20 minutes. The excursion exposed three per-sonnel in the facility to gamma and neutronradiation doses of 110, 43, and 19 rem, res-pectively.

The Recuplex incident is mentioned becauseit occurred only a few months after the SL-1accident. The SL-1 accident had revealed theimportance of retrieving the data from theaccident dosimeters in order to estimate thedose received by the exposed workers. In thecase of the Recuplex event, a trained seniorstaff member at the facility performed suchretrieval. Without consultation or direction,he ran into the building, following the ini-tial pulse, to retrieve the dosimeter. In sodoing, he just barely missed a i%»cond muchlarger pulse consisting of some 10 fissions.Exposure at this level would have been fatal.In this case, a senior staff member who washighly trained in routine and emergency situ-ations did not react properly in an actualemergency. (Vallario and Selby 1981; Zanger1962; Callihan 1963.)

The lessons learned from the Recuplex acci-dent include:

• Predicting the reactions of personnel, evenseasoned "veterans," in an accidentsituation is difficult if not impossible.If tests could be developed to ascertainsuch behavior in advance, they would bevery useful.

• Robot equipment can be very useful, both ingathering information about an accident andin conducting remedial actions. Thislesson was later dramatically confirmed inthe longer-range recovery actions at ThreeIsland, Unit 2.

Wood River Junction Criticality - 6:00 p.m.,Friday, July 24, 196"4~

This cnticality accident occurred in afuel recovery plant of United Nuclear Oper-ations in Wood River Junction, Rhode Island.It was the sixth such event to occur in theU.S.

23

The accident was initiated when a tech-nician erroneously poured 11 liters of concen-trated uranyl nitrate into a tank that con-tained 4! liters of 0.54 M sodium carbonate.Five persons were in the facility at the timeof the excursion. The technician saw the typ-ical light blue flash and was knocked to thefloor. Although dazed, he ran out of thebuilding to an emergency shack some 500 feetaway. He was taken to the Rhode Island hos-pital, and died 47 hours later. Analyses ofhair and blood samples, taken from the tech-nician's body, indicated that he received morethan 14,000 rad to the head and 46,000 rad tothe pelvic area of his body.

Two supervisory personnel, who arrived onthe scene after the first excursion were ex-posed during a second, lesser excursion thatwas initiated when they shut down the power tothe stirrer that was used to mix the solution.They received approximately 50 rad of gammaradiation and 5 rad of neutron radiation.This portion of the event was similar to theRecuplex incident at Hanford (Auxier 1965).

The lessons learned from the Wood RiverJunction accident include:

• Better administrative control of fissilematerials was needed. Procedural methodsand controls had not been updated at thetime of the accident.

• Better identification of fissile materialswas essential.

• Better training of plant personnel innormal operating procedures was essential.A review by the investigating committeeindicated the need for complete writtenprocedures. Except for the plant super-intendent and two supervisors, the staff atthe facility was not experienced in thekind of operations they performed.

Browns Ferry Nuclear Power Plant Fire - 12:20-p.m, Saturday, March 22, 1$75"

Although this accident did not involve therelease of any radioactive materials, thenational media described it as a "disaster andnear disaster." The accident was initiated bya team, consisting of an electrician and anengineering aide, who were checking for leaksin the cable spreading room. The team was re-sponsible for locating and plugging holes inthe polyurethane sealant surrounding the cab-les within wall openings. A candle with anopen flame was used to indicate the presenceof drafts; the candle ignited the polyure-thane.

Unfortunately, the fire ignited in aninaccessible location. Fifteen minutes afterthe fire began, the alarm sounded. The alarmwarned people that the C02 CARDOX fire extin-guishing equipment would be operated. How-ever, because the draft was so strong, this

equipment was not effective.Ultimately, as in the Windscale accident, a

heat sink was required to extinguish the fire..Personnel of the Athens Fire Departmentrecommended the use of water to provide theheat sink and extinguish the fire. The plantpersonnel were concerned, however, that usingwater would result in shorted electrical cir-cuits, a concern also expressed by the Tennes-see Valley Authority Public Safety Officer.Not until 6:00 a.m. on March 23, J975 did theplant superintendent finally agree to usewater. The fire was extinguished about45 minutes later (Scott 1983; Moeller 1979).

The lessons learned from the Browns Ferryincident include:

• Personnel involved in the work of leak-testing, sealing and inspection need to betrained in fire control and proper firereporting procedures. They also need tohave detailed written, approved proceduresfor work activities and fire control-

• Fire fighting equipment needs to be fullyoperational at all times. A metal plateinstalled for the safety of the personnelprevented the activation of the (XL CARDOXsystem.

• Fire control practices at nuclear facili-ties at the time emphasized the precautionof not using water on electrical fires.One of the reasons for this emphasis wasthat the possibility of shorted electricalcircuits that could have been instrumentalin the loss of safety controls.

• As noted in the Windscale accident, anintense fire requires a heat sink. Whenthe hee* sink was provided by the appli-cation of water at Browns Ferry, the firewas quickly extinguished.

• Redundant control systems must be suffi-ciently independent so that they do notinterfere with each other during a fire.Control cables, for example, had been runin close proximity in metal conduits. Themetal conduits did not provide the intendedprotection; when the fire melted theinsulation of the wires, feedback throughindicator lights made safety equipmentunavailable.

• Emergency responses needed to be coor-dinated. The plant superintendent spenttoo much time in communications with theCentral Emergency Center and did not directthe efforts of the plant's emergencyresponse team and the fire fighting teams.

Three Mile Island Unit 2 Nuclear Power StationAccident - Wednesday, 4:00 a.m., March 28, 1979

Ro other accident until tTTe Chernobvl

24

event, has had such a profound impact on thepublic, the nuclear industry, and the regula-tory agencies. The accident involved ThreeMile Island (TMI) Unit ?., a 906-MW pressurizedwater reactor, located 9 miles from Harrisburg,Pennsylvania, a city of 70,000.

At the time of tne accident, Unit 2 wasoperating near full capacity. Unit 1 had beenshut down for refueling. A condensate pump inthe turbine building stopped, which also causedother pumps in the system downstream to shutdown. As a result, the reactor scrammed auto-matically. In accordance with the system'sdesign, emergency feedwater pumps came intooperation. However, during an earlier mainte-nance procedure several valves had been mis-takenly left closed, blocking the flow ofwater. Without this water, the heat and pres-sure in the primary system began to rise andthe power-operated relief valve on top of thepressurizer opened and became stuck. Waterpoured through the open valve and the pressurefell. After the pressure was reduced, the re-lief valve was designed to close and restorethe integrity of the primary system. In thecontrol room, the relief valve indicatorshowed it to be closed, when it actually re-mained open and continued to release waterfrom the primary system. The relief valve re-mained open 2 1/2 hours; during that time theoperators maintained the water in the pres-surizer at the normal level, assuming thatthis was indicative of an acceptable watervolume in the primary system. Actually, morecoolant was being lost than was being replacedand a small loss-of-coolant accident was under-way. Ultimately, water surrounding the coreboiled and the top of the reactor core was un-covered. This event resulted in seriousdamage to the fuel and the associated releasesof radioactive fission products to the primarycoolant, to the containment, and to the envi-ronment.

In the next 11 hours, the operational staffmade several attempts to reduce pressure inthe primary system. Fifteen hours afterthe accident began, forced circulation throughthe core had been established and a relativelystable plant condition was obtained.

Three hours after the accident began,radiation levels in the reactor containmentbuilding and the auxiliary building led to thedeclaration of a site emergency. During theinitial phases of the accident, the contain-ment building became contaminated with radio-active water, which was pumped over to theauxiliary building. Abnormal releases weremade to the environment through the ventil-ation system of the auxiliary building.

Dose rates inside the Unit 2 containmentranged up to thousands of rem/h; a measurementof 70 mrem/h was reported at the north gate ofthe plant. However, dose rates at most ground-level stations around the plant site followingthe accident were only 1 to 3 mrem/h (Bertini

1980; Okrent and Moeller 1981; NP.C 1979).No injuries were reported. The reactor

core was heavily damaged, resulting in exten-sive cleanup operations.

At least 14 committees reviewed the acci-dent and an ad hoc committee concluded thatthe offsite collective dose represented min-imal risk of additional health effects to theoffsite population. The President's Commis-sion pointed to the root problem as being"people-oriented," rather than related to de-ficiencies in plant design or equipment. Theweaknesses identified were not only the short-comings of individual human beings, but alsoproblems of organizational structure and com-munications among key individuals and groups.The Commission also found a preoccupation withregulations themselves, rather than with thesafety the regulations were supposed to pro-mote. The Commission especially noted thefocus on "worst case" accidents to the neglectof less consequential but more probable scen-arios (Bertini 1980).

The lessons learned from the TMI accidentencompass a wide spectrum of concerns (NRC1979). The lessons include:

• Training of THI personnel was deficient anddid not prepare the staff to deal withemergency situations. Improved plantpersonnel training was needed on a periodicbasis.

• The TMI control room lacked the applicationof good ergonomic design. Control roomsshould be designed with optimumconsideration for human factors.

• There was a need for rigorous emergencypreparedness exercises on a periodic basis.After TMI, the NRC required that licenseessubmit detailed emergency plans and asso-ciated implementating procedures, includingstate and local governmental emergencyplans for review and approval. The NRCthen determines whether the emergency pre-paredness capability provided reasonableassurance that the plant personnel couldrespond promptly and adequately to an emer-gency and provide measures to protectworkers and the public.

• While actual radioactive releases to theenvironment were negligible, the mentalstress to the population may have been themajor impact during the TMI accident. Thesiting of future nuclear power plants atmore remote sites would be required.

Chernobyl Nuclear Reactor Accident1:23 a.m., Saturday, April 26, 1986

The Chernobyl accident was first noted by asudden increase in the power of one of fourreactors located at the site. While under-going shutdown, a large emission of steam re-

25

acted with the overheated fuel to form the hy-drogen that exploded.

When the accident occurred, the reactor wnsoperating at 755 power. The explosion exten-sively damaged the reactor building; portionsof the building collapsed. The reactor wasdamaged and substantial radioactive releasesresulted. Three days later, however, person-nel circling the facility in a truck and fly-ing overhead ir a helicopter noted that ap-proximately one fourth of the exposed graphitecontinued to burn.

Based on initial information, temperaturesexceeding 9000r'F and relatively dry weatherwere thought to be responsible for transport-ing the radioactive plume high into the atmos-phere and then across the Soviet Union tonortheast Europe. Lawrence Livenr.ort NationalLaboratory scientists sucoested, however, that50K of the radioactive material was releasedin the hydrogen and steam explosion and theother 50^ was dispersed during the graphite-fire (Rippcn, Blake and Payne 1986).

It was reported that the area aroundChernobyl did not immediately receive a highexposure to radiation, probably because of thehigh plume rise. Soviet eiii&rgency plans re-quire evacuation if anticipated doses approach25 Rem. Evacuation was initiated 36 hoursafter the accident. It is not known whetherthe evacuation was delayed because of rela-tively low radiation levels or for other reas-ons. It was reported, however, that specta-tors from the town of Pripyat watched the firefrom the plant perimeter on the day of theaccident. Although dose rates reported bySoviet sources were relatively low (15 mrem/hrat the zone during the first day of the acci-dent and 0.15 mrem/hr at the zone perimeter),actual rates are not available at this time.Levels observed in Finland, Sweden and theFederal Republic of Germany were 2-40 timesthat of normal background radiation, mostlydue to deposition on the ground. Wide dif-ferences in radiation levels in some of theareas appeared to be the result of rainfall.(Wilson 198%f).

It is st'ill too early to list all the les-sons that were learned from the Chernobylaccident, but some of them are suggested hereon the basis of a report by Or. Richard Wilson,Mallinckrodt Professor of Physics of HarvardUniversity.

• The Chernobyl accident underscores the im-portance of reactor operator training,

• Emergency planning and the appropriate useof staff during emergencies is essential.Also emergency plan?, should focus less onhow to evacuate, and more on whether andwhen to evacuate the affected population.

• International cooperation is essential; a

rr.ore prompt release of data by the Sovietswould have been helpful.

• Governments should be care-fiil not tooverreact in response to an accident thatimpacto the international community. Dur-ing the Chernobyl accident the met.erologicaland radiation protection services of variouscountries responded promptly and correctly.However, politicians appeared to ignoretheir council and kept a ban en food andwater in areas were it was not needed.

SUMMARYB~ review of nuclear facility accidents

since the 1950s revealed that human error wasfound to contribute to a iripjority of the inci-dents. The review also indicated that theaccidents were caused or complicated by per-sonnel working during off-normal hours.

Contributing factors to the earlier inci-dents were perhaps deficiencies in the designof facilities or equipment involving this newtechnology. Later accidents, however, in-volved facilities and equipment that had be-come "safer," as a result of the many redun-dancies built into the operating systems tosafeguard the facility and human health. Themore recent accidents appear to be the resultof errors on the part of humans trying torespond to unusual events involving this in-creasingly complex technology. Compoundingthis problem was the fact that many of theseaccidents may have occurred when the mostqualified and best trained personnel were noton duty. In addition, decreased alertness ofkey personnel during off-normal working hoursmay have contributed to these events.

Several of the following lesspns, learnedfrom these accidents, are being or have beenimplemented to some degree:

• Nuclear accidents can have worldwide impacton the public, governmental agencies, andthe nuclear industry.

• Rotation of operating personnel to workduring off-normal hours requires carefulplanning, taking into consideration thenatural body rhythm to maximize perfor-mance.

• The response of an individual to an emer-gency cannot always be predicted. Also,operating personnel require some time torespond to an emergency. Ultimately, themajor response will be based on the quelityof the training imparted to emergencyresponse personnel.

• Emergency preparedness instrumentationcapable of assessing the high radiationfields that may be present is required.This equipment must be operable in a wide

26

range of environments.

• Remotely operated retrieval and surveil-lance equipment during emergencies is es-sential. Control rooms should be designedwith optimum consideration for human fac-tors.

• Potential accidents need to be anticipatedthrough safety studies and on continuingformal analysis.

• Regularly scheduled, rigorous emergencypreparedness exercises are needed. Theseexercises must include objective, post-exercise critiques and a willingness on thepart of facility management to correct anydeficiencies identified.

• Improved plant personnel emergency responsetraining is needed on a periodic basis.

• Emergency responses require well-directedcoordination.

• Rigid administrative control of fissilematerials is essential.

• Siting of nuclear facilities should be con-sidered with respect to population density.

REFERENCESAtomic Energy Office. 1957. Accident at Wind-

scale No. 1 Pile on 10 October, 1957.Majesty's Stationery Office, London.

Her

Auxier, J. A. 1965. "Nuclear Accident at WoodRiver Junction," Nuclear Safety, 6(3):298-301.

Bertini, H. W. 1980. Descriptions of Selected Accidents Tliat Have Occurred atNuclear Facilities. ORNL/NSlC-176, OaTRidge National Laboratory, Oak Ridge, Ten-nessee.

Biological Clocks and Shift Work Scheduling,1983: Hearings before the Subcommittee ofInvestigation and Oversight of the Committeeon Science and Technology, House of Repre-sentatives, 89th Cong., First Sess. (March23-24, 1983) (Testimony by C. A. Czeisler,Harvard Medical School). Cited in text asHearings 1983.

Callihan, D. 1963. "Accidental Nuclear Excur-sion in Recuplex Operation at Hanford inApril 1962." Nuclear Safety. 4(4):136-144.

Cottrell, W. B. 1962. "The SL-1 Incident."Nuclear Safety. 3(3):64-74.

Czeisler, C. A., E. D. Weitzman, M. C. MooreEde, J. C. Zimmerman and R. S. Knauer.1980. "Human Sleep: Its Duration and Organ-

ization Depend on Its Circadian Phase."Science. 210:1264-1267.

Eisenbud, M. 1973. Environmental Radio-activity.2nd ed. pp. 4ll-417. iScademicPress, New York.

Joksimovich, V. 1986. "An Overview of In-sights Gained and Lessons Learned from U.S.Plant- Specific PRA Studies." NuclearSafety. 27(l):15-27.

Lewis, P. M. 1985b. Shift Scheduling andOveir-time at Nuclear Power Plants. RDREG/CR-4248, PNL-5435, Pacific Northwest Labor-atory, Rich- land, Washington.

Lewis, P. M. 1985b. Shift Scheduling andOvertime: A Critical Review of trie Literature. PNL-5391, Pacific Northwest Labor-atory, Rich- land, Washington.

Matsuda, Y., K. Suehiro, and S. Taniguchi.19P4. "The Japanese Approach to NuclearPower Safety." Nuclear Safety. 25(4):451-471.

Moeller, 0. W. 1979. "Failures in Air Moni-toring, Air Cleaning, and Ventilation Sys-tems in Commercial Nuclear Power Plants;.lan. 1, 1975 to June 30, 1978)." Nuclear

20(2):176-188.

Okrent, D. and D. W. Moeller. 1981. "Impli-cations for Reactor Safety of the Accidentat Three Mile Island, Unit 2," Ann. Rev.Energy. 6:43-88.

Raftferty, J. F. 1983. "Watching the Biologi-cal Clock," Harvard Magazine. March-April1983. pp. Z6-W.

Rippon, S., E. M. Blake and J. Payne, 1986."The Chernobyl Accident." Nuclear News.29(8):87-94.

Scott, R. L. 1976. "Browns Ferry NuclearPower Plant Fire on March 22, 1975."Nuclear Safety. 17(5):592-611.

Squires, S. Sunday, June 15, 1986. "Research-ers: Nighttime is the right time for workaccidents." Reno Gazette Journal. pp.1E-3E.

U.S. Nuclear Regulatory Commission (NRC).1979. TMI-2 Lessons Learned Task ForceStatus ~R"etions. Ffi

rtanaShort-TermRecommenda£6 0578, U. S. Nuclear Regulatory

Commission, Washington, D.C.

Vallario, E. J. and J. M. Selby. 1981. "Useof Accident Experience in Developing Criter-ia for Teleoperator Equipment." Paper pre-sented at Harvard School of Public Health,

27

June 8-12, 1981, Boston, Massachusetts.

Zanger, C. N. 1962. Summary Report of Acci-dental Nuclear Excursion in Recuplex Opera-tion 234-5 Facility. Atomic Energy Coimiis-sion, Richland Operations Office, Richland,Washington.

Wilson, R., 1986. Tentative Implications ofthe Accident at Chernobyl, wri tten forIssues in Science and Technology, Fall 1986.

ANS Topical Meeting on Radiological Accidents—Perspectives and Emergency Planniag

A Chronology of the Chernobyl-4 AccidentG. Donald McPherson

Abstract - This paper presents a chronologicaldescription of the events leading to theApril 7.5, 1987, accident at the Chernobyl Unit Ureactor at Pripyat in the Soviet Union.

I. INTRODUCTION

On August 25-29, 1986, the Union of SovietSocialist Republics (USSR) presented their reporton the April 25, 1986, accident at the ChernobylUnit 4 reactor to International Atomic EnergyAgency Experts' Meeting. The accident chronologyincluded in this paper is based upon informationincluded in the working document of the USSRState Committee on the Utilization of AtomicEnergy, which was provided at that meeting, aswell as additional information supplied orally bythe Soviets at that meeting.

II. REACTOR DESCRIPTION

The Chernobyl reactors are boiling-water-cooled, graphite-moderated, pressure-tubereactors (RBM) 14 m diameter by 9 m high of 1000MWe capacity. The core is a stack of graphiteblocks penetrated by 1680 vertical coolingchannels which also contain the fuel bundles.The cooling channels are connected as twoseparate loops, each one with its own steamseparator, four circulating pumps and twofeedwater pumps. The latter return water fromthe condenser to the steam separator. Theequipment for each loop is located on one side ofthe core and is connected to channels only onthat side of the core. The pipes leaving the topof the core reactor carry the steam-water mixtureto the steam separators. From there the steamgoes to the turbines. After passing through acondenser, the steam condensate is returned tothe separators. This water is injected in almostjet-pump fashion into a downcomer leading to thecirculating pumps, which return the coolant tothe reactor.

The only union in the double circuit is thatsteam from the steam separators on the right sideand the left side may be directed to either orboth turbines. Thus, one side could feed oneturbine, and the other side, the other turbine;or the steam could be mixed to feed both. Thatis what was being done before the accident.

III. BACKGROUND

The reactor had been running at full power.Personnel at Chernobyl Unit U were preparing toshut the reactor down for scheduled maintenance.During the course of the shut-down, the operatorsplanned to perform a test of the capability touse the rotating kinetic energy of the turbo-generator to supply power during the shut-downprocess.

At the beginning of the test, the reactorpower was planned to be 700-1000 MWth (21-33Z offull power). Turbogenerator No. 8 was connectedto four of the main cooling pumps and twofeedwater pumps. The power demand of these pumpswould be equivalent to that of the emergency corecoolant system. The other four cooling pumps,other two feedwater pumps, and, of course, theother power demands of the reactor building wouldbe carried by the electrical grid. During thetest, the operators planned to close the stopvalve on the one operating turbine (No. 8) anddetermine how long the generator could carry theload; i.e., how long it could power the fourcooling pumps and two feedwater pumps. At theend of the test, the reactor power would beunchanged. Both turbines on Unit A would bestopped, and the grid would be supplying thepower to the remaining four cooling pumps and twofeedwater pumps, in addition to any other powerdemands.

Following an earlier test of the same type,the generator exciter block had been modified inan attempt to increase the time during which therun-out power would be available to the emergencyequipment. This would allow additional time forthe onsite diesel generators to start supplyingpower to the critical components.

The test was designed by an electricalengineer who was unfamiliar with the operatingcharacteristics of the RBMK reactor. Inparticular, the test designer was unaware of thepositive void reactivity coefficient of thereactor and its increasingly unstable character-istics at lower power. Although the test was tobe conducted at 22-31% power (the Soviets statethat the exact value was unimportant), the testdesigner believed that a lower value would beacceptable.

Permission had been requested to conduct thetest, but, the Soviets stated that "the plantmanagement did not get around to approving it."

29

30

The test occurred on the Saturday before MayDay. It was very early in the morning. If thetest could not be done while the reactor wasbeing shutdown for this maintenance period, itwould be another year before it could be at-tempted. Given the situation, the operatingstaff was psychologically unprepared and in-attentive to the emergency condition that wasarising.

The sequence of events that led to thedisaster is detailed in Table I.

IV. AFTER THE ACCIDENT

After early May the situation at the reactorhad largely stabilized. The short-lived fissionproducts had decayed, and the radiation exposurerate in the reactor compartments was in the"single R/h" range. Temperatures in the reactorcavity were stable, and in some cases, droppingby 0.5 °C/day. Ninety-six per cent of thereactor fuel remained in the reactor shaft andadjacent compartments. Any radioactivity stillescaping from the reactor was due to windentrainment of aerosols and did not exceed dozensof Ci/day.

The other three reactor units were contami-nated during the accident due to radioactivitybeing transported through the ventilation system.Sections of the turbine room had high radiationlevels since they were contaminated through thedestroyed roof of the third block. Afterdecontamination, dose rates in the compartmentsof Units 1-3 were 2-10 mR/day.

During the accident, large amounts of radio-active material, including graphite and fuel,were discharged over the plant grounds andreactor buildings. Contamination of the groundwas not uniform. Several methods of controllingthe contamination were used. Polymerizing agentswere applied to soil and buildings to reduce airentrainment of particles. As much as 10 cm ofsoil was removed in some areas. Concrete slabsor mounds of clean soil were placed over thecontaminated ground in places. After thesemeasures the total gamma background in the areaof the first unit was reduced to 20-30 mR/h.This radiation exposure rate was due principallyto releases from Unit 4.

The Soviet long-range plans for mothballingUnit 4 include the following:

o Building protective walls outside along theperimeter,

o Erecting concrete and metal dividers betweenthe turbine rooms for Unit 3 and those forUnit 4,

o Constructing a protective cover over theturbine room,

o Sealing off "other" compartments as needed, ando Providing for ventilation of the damaged unit.

ACKNOWLEDGMENT

The assistance of Jim McNeece, PacificNorthwest Laboratory, and Brian Sheron, U.S.Nuclear Regulatory Commission, in preparing thischronology is gratefully acknowledged.

Table I. Chronology of the accident

Date Time Power Event Comment/Motivation

April 25 0100

1305

1400

3200 Mwth

1600 Mwth

1600 Mwth

2310

0028

April 26 0100

1600 Mwth50% Power

Variable

200 Mwthvariable

Begin power descent to 700-1000 Hwth as required for test.This power level was apparently chosen by test designer;Soviets have stated that the test could have been run atzero power (i.e., immediately following a scram).

Turbogenerator #7 disconnected. Power for auxiliaries (4main cooling pumps, 2 electrical feedwater pumps, etc.)was transferred to busbars of turbogenerator #8 (TG #8).Later disengaged scram on 2 turbogenerator trips.

Emergency core cooling system (ECCS) disconnected asrequired by test procedure. This was a violation ofregulations.

Continued power reduction was delayed for 9 h on orders ofthe power grid dispatcher to meet demand. This allowed alarge xenon buildup.

Power reduction resumed.

Switch off local automatic control as permitted at lowerpower operation; because bulk power set point was presetto a low power level, the reactor power was driven down tobelow 30 Mwth.

Power stabilized at this level. Attempt to increase powerto the desired 700-1000 Mwth level was difficult (neveraccomplished) due to the low amount of operatingreactivity margin (ORM.) (Operating reactivity margin isa calculated reactivity margin based on power and controlrod distribution and positions, given in terms of numberof control rods, where 30 is the minimum allowable. Incertain situations this may be lowered to an absoluteminimum of 15.)

As part of test objective, neededequipment was to be powered fromoffsite, and TG #8 output was requiredfor test.

ECCS was disconnected so that the testcould still be rerun.

Sharp power reduction results in xenonbuildup requiring removal of controlrods to maintain criticality.

Xenon poisoning builds up greater thanthat anticipated for test; as a result,ORM was 6-8 rods, substantially below15-30 rods required.

Table I. (cont'd)

Date Time Power Event Comment/Moti vati on

0103-0107

200 Mwthvariable

Before0119:00

0119:10

200 Mwthvariable

200 Mwthvariable

0119:30 200 Mwth

0119:40

At 0103, one additional main cooling pump was placed intoservice and at 0107 the second main cooling pump wasplaced into service. This resulted in a total of 8 pumpsas the test program directed. The total flow increased to57,000 nr/h, above that allowed in the operatingregulations. Low power to flow ratio possibly reducedcore void to 10% or less. (The Soviets are concernedthese pumps may have been operating close to point ofcavitation). This resulted in a decrease in steampressure and water level in steam separators.

In order to keep low steam pressure and low water levelfrom causing reactor shutdown, the operators blocked theemergency protection signals related to these parameters.

Operator began manual replenishment of water to steamseparator.

The AR (automatic power regulating) rods moved upward tocompensate for reduction in power. Some manual rodsremoved entirely to allow insertion of AR rods.

The feedwater flow to the steam separator had increased bya factor of 3 over the balanced flow for this power level.As this colder water reached the reactor core, there was asharp drop in the steam fraction of the coolant, and acorresponding power decrease.The feedwater flows into thesteam separator at the nozzles of the downcomers leadingto the primary coolant pumps. Hence, the core inlet flowtemperature responds rapidly to any change in feedwaterflow.

Based on the planned test power level,(700-1000), Mwth, 4 main cooling pumpswere required to continue operation.Soviets stated that this mode ofoperation - low core void, very lowORM, slow power measurement responseand pumps near cavitation - is veryunstable.

By blocking these protection signals,operator is able to proceed towardstest.

Apparently anxious to establish thiswater level well within the normalrange, operator sharply increased thefeedwater flow above its nominal value.

The Soviet report implies that thismaneuver is probably not permitted.

Table I. (cont'd)

Date Time Power Event Comment/Motivation

0119:58

0121:50 200 Mwthvariable

0122:10

0122:30 200 Mwthvariable

0122:45

0123:04 200 Mwth

Turbine bypass valve (steam dump) closed.

Feedwater flow raised to 4 times the balanced flow forthis power level.

Operator reduced feedwater flow rate sharply to a level ofabout two-thirds of balanced flow.

Outlet steam quality increases,begin inserting.

Therefore automatic rods

A printout of the actual core flux monitor outputs and theposition of all the regulating rods was obtained at thistime ("Skala" system). Operator noted that ORM was about6-8 rods, far below the value where immediate reactorshutdown is required. Nonetheless, no action taken. Theflux was "practically arched" in the radial direction anddouble peaked axially with the higher peak in the topsection of the core.

Feedwater flow stops decreasing at a value 2/3 of balancedflow. Increase in steam quality stops because of pressureincrease and feedwater flow stabilization.

To begin the test the TG #8 turbine stop-valve was closedand the turbogenerator bypass valve remained closed. Thesignal for reactor shutdown on closure of both turbo-generator steam valves had been disengaged. This was inviolation of the test program and normal operatingprocedures.

This was done to raise the steampressure which was too low. However itcontinued to drop slowly until thestart of the accident (0123:40).

This caused increase in inlettemperature and compounded the eventsone minute later.

Due to approximately 20 sec transienttime from steam down to core inlet.

At this highly unusual power shape, theerror in estimating the control rodworths was extremely large, hence theerror in estimating the ORM was alsolarge.

By avoiding reactor shutdown, it wouldbe possible to repeat the test usingreactor generated steam. Soviets statethat this is the most infelicitous(awkward) moment for this test to berun.

Table I. (cont'd)

Date Time Power. Event Comment/Moti vati on

Flow rate began to fall as the 4 main cooling pumpspowered by TG #8 began to run down. Steam pressures alsobegan to increase due to removal of TG steam load and thereduction (by operator action) of the feedwater rate about1 minute before. These, factors combined to increase thecoolant void fraction (positive reactivity insertion) andconsequently the power.

0123:10 Power increase causes pressure increase which causes voidcollapse. This feeds back to decrease power and automaticrods rise.

0123:21 As four pumps coast down, the flow drops, void and powerincrease. To compensate, automatic power regulating rodsbegin to lower.

0123:31 Reactivity and power increase further because of pumpcoast down.

0123:40 320 Mwth Unit shift foreman gave order to press emergency scramincreasing button.

Due to the reactor conditions at thetime of the test, the void fraciton isbelieved to have increased many timesmore sharply than it would have atnormal power. This sharp increase invoid fraction rsults in a sharpincrease in reactivity.

Soviets stated: "When they were stillalive, the operators gave threenations for pressing the AZ5 scrambutton:1. the power was increasing2. the test was proceeding well (and

therefore a repeat would beunnecessary)

3. the automatic control rods weremoving."

Table I. (cont'd)

Date Time Power Event Comment/Moti vati on

0123:43 Soviet The "runaway period came to be much less than 20 seconds."calc. [There are indications that period was 1 second.]a Overindicates power and short-period alarms come on. The positive voidpower at coefficient prompted "deterioration of t) • situation."3800 Mwth Only the Doppier effect partially comr . =d for the voidand rising reactivity increase.

Soviets have stated that voidreactivity coefficient 2.0 x 10"*/%steam volume at normal operatingconditions was 50% higher at testconditions, but implied it was evenhigher at this time.

0123:44

0123:44(appr.)

[First power surge terminated by Doppier effect and rodinsertion. Water flow continued to decrease (due to run-down of 4 main cooling pumps) and the power continued toincrease.]

Operator heard banginy noises and saw rods were stoppingbefore they reached bottom and released servo mechanism toallow rods to fall into core.

The first power surge may havedistorted the core such that thecontrol rods coulo not be bottomed.The banging noises may have beeninitial ruptures of pressure tubes.

0123:45

0123:46

This likely led to fuel fragmentation causing large steamspike which reversed flow to close main reactor Dump checkvalves. Resulting void produces neutron pulse, and secondpower surge.

Steam drum pressure exceeds "accident level" and pressurerelief valves open. [Pressure rise rate calculated bySoviets was 8-10 atm/sec].

aItems resulting from Soviet mathematical model or that cannot be clearly attributed to measured values are enclosed in square parentheses.

Table I. (cont'd)

Date Time Power Event Comment/Motivation

0123:47 Large increase in coolant flow [as channels rupture due topressure spike].

0123:48

0124(appr.)

0154

0334

0354

0500

["Thermal explosion" (steam release) blows top off reactorand destroys reactor hall building].

Two explosions - hot fragments and sparks emitted from topof reactor building. Hot fragments caused about 30 fireson roofs, etc. [Mixture of gases containing hydrogen andcarbon monoxide capable of thermal explosion if mixed withoxygen, were created in core region].

Thirty fires started in three primary fire sites: (1)turbine room above TG #7; (2) reactor room; and (3)partially destroyed compartments adjacent to reactor room.

Fire fighting units from Pripyat and Chernobyl arrive.Focused on fighting fire in turbine room to prevent spreadto Unit 3. Hand extinguishers and "stationary firecranes" were used to fight fires in the compartments.

Most of fires on turbine room roof were out.

Fires on reactor building roof were out.

All fires out. Shutdown of Unit 3.

If initial ruptures occurred as above,this flow increase could be consistentwith the upper biological shield beinglifted by the pent-up steam in thereactor cylinder and the resultantsevering of all pressure tubes andcontrol rod channels.

[Observation from outside reactorbuiding. From this point on, there isno more information from the controlroom].

Table I. (cont'd)

Date Time Power Event Comment/Motivation

April 27 0113 Shutdown units 1 and 2.

Immediately after the accident, an attempt was made toreduce the temperature in the reactor cavity and preventcombustion of the graphite using emergency and auxiliaryfeedwater pumps. This was unsuccessful. Decision wasmade to fill the reactor cavity with heat discharging andfiltering materials.

April 28- Dropped 5000 tons of boron compounds, dolomite, sand,May 2 clay, and lead onto damaged reactor. Discharge of

radioactivity dropped to several hundred curies/hr.

Problem of reducing fuel heatup was solved by pumpingnitrogen into space under reactor. Temperatures rose,stopped, and began to drop. As insurance against"extremely improbable" failure of the lower tier ofstructures, construction of an artificial "heat dischargehorizon" was established under the core. Completed by endof June.

ANS Topical Meeting on Radiological Accidents—Perspectives aid Emergency Planning

Uranium Yellow Cake Accident—Wichita, KansasHarold R. Borchert

ABSTRACT-A tractor and semi trailer containingUranium Yellow Cake, had overturned on 1-235,Wichita, Kansas on Thursday, March 22, 1979.The truck driver and passenger were transport-ed) with unknown injuries, to the hospital byambulance. The shipment consisted of 54 drumsof Uranium Ore Concentrate Powder. Half of thedrums were damaged or had their lids off.Since it was raining at the time of theaccident, plastic was used to cover the barrelsand spilled material in an attempt to containthe yellow cake. A bulldozer was used toconstruct a series of dams in the median and theditch to contain the run-off water from thecontaminated area. Adverse and diverse weatherconditions hampered the clean up operations overthe next several days. The contaminated waterand soil were shipped back to the mine forreintroductioti into the milling process. Theequipment was decontaminated prior to beingreleased from the site. The clean up personnelwore protective clothing and respiratory protec-tion equipment, if necessary. All individualswere surveyed and decontaminated prior toexiting the area.

The Kansas Highway Patrol was notified at8:51 a.m. on Thursday, March 22, 1979 that asemi tractor and trailer had overturned on1-235 at the North Meridian overpass Wichita,Kansas. A Sedgwick County Fire DepartmentEngine along with a Wichita-Sedgwick Countyambulance were dispatched to the scene at 8:58a.m. Upor. arrival at the scene, the respondersobserved that the vehicle had radioactiveplacards affixed which indicated that radio-active materials constituted the shipment.

*At the time of the accident the authorwas a Health Physicist and Chief of theMaterials Licensing and Control Section, Bureauof Radiation Control, Division of Environment,Kansas Department of Health and Environment,Topeka, Kansas 66620.

Further investigation of the shipment indicatedthat the cargo consisted of numerous barrels.Upon reviewing the markings on the barrels theresponders concluded that the cargo was uranium.The Radiation Specialist, Kansas Department ofHealth and Environment Wichita Office wasnotified of the accident at about 9:00 a.m.He telephoned the State Office in Topeka, Kansasat 9:05 a.m. relaying the accident informationavailable to him and then proceeded to theaccident site.

The Sedgwick County Assistant Fire Chief,upon learning that the cargo was uranium, re-quested that the Hazardous Material Responseteam be notified and activated. The SedgwickCounty Civil Preparedness Office was notifiedat 9:10 a.m. The coordinator and assistantcoordinator arrived at the scene about9(20 a.m. They found that the top of theoverturned semi trailer had been ripped open anda number of barrels were scattered about thesite. The vehicle which was traveling west andsouth on 1-235 had the right wheels leave thepavement about 213 meters (m) (700 feet) fromthe site and became mired in the soft unfin-ished shoulder. The driver had attempted tobring the vehicle back onto the pavement andunder control when it overturned about 30.5 m(100 feet) east of the overpass. The overturnedsemi tractor and trailer was blocking the westbound lanes of 1-235 including the median andouter . shoulder. The contents consisting of 2.1X 10 m (55 gal) drums was spilled duringthe mishap towards the west and since theoverturned vehicle blocked the area, trafficcould not proceed past the accident site. Thisprevented the spread of the material away fromthe accident scene.

The Civil Preparedness Radiological Officerarrived about 9:25 a.m. and began performingsurveys of the area using a Civil DefenseCDV-700 GM survey meter. According to theRadiological Officer's survey, the radiaftionlevelSWere found to be about 215 X 10~ Cfound to be about 2.15 X 10 Ckg s (3 milliroentgens per hour) (mr/hr) intheqimmediate area of the spill and about 2.15 X10 C kg s (30 mr/hr) close to thematerial itself. A Kansas Highway Patrolman

39

40

also attempted to obtain radiation readings inthe area but was unsuccessful. No otherradiation levels were obtained in the area andthe previous readings could not be confirmedbecause it had started raining.

The truck driver and passenger, withunknown and undetermined injuries, were trans-ported by ambulance to the hospital. Assistancewas requested from the McConnell Air Force BaseDisaster Preparedness Office and they sent anofficer with survey equipment to the hospital toperform contamination surveys. The truckdriver, passenger, ambulance personnel andambulance were found to be free from contamina-tion. The truck driver and passenger weretreated for their respective injuries. Thepassenger did not exhibit any signs of injuryand was held over night for observation and thenreleased. The truck driver had a couple ofcracked ribs and minor cuts and a few bruises.He was hospitalized for about three days andthen released.

The Sedgwlck County Fire Department andthe Hazardous Materials Response Team estab-lished a control line about 30.5 m (100 feet)from the site. The traffic was able to bedetoured around the site on the eastbound lanesof 1-235. Since it was raining the decision wasmade to cover the material with plastic tolessen the possible dispersal from the site.

The shipping documents were recovered fromthe cab of the semi tractor and it was foundthat the contents were indeed radioactivematerial and listed as Radionuclide U-238Uranium Ore Concentrate Powder 3,700 megaBequerels (MBq) (0.10 curies) per drum. Theshipment contained 54 drums 19,859 kilograms(kg) (43,782 pounds) gross weight of which thenet contents were 18,595 kg (40,996 pounds).The Uranium "Yellow Cake" was being transportedby Salt Creek Freightways, Casper, Wyoming fromPetrotomics Company, Shirley Basin, Wyoming tothe Kerr-McGee Nuclear Corporation, SequoyahRefining Facility, Gore, Oklahoma for furtherprocessing.

The Sedgwick County Civil PreparednessCommunications Director notified all threecompanies about the accident. He also contactedChem-Trec, the national agency that providesinformation about hazardous materials.Chen-Tree confirmed that the material was notdangerous to persons in the area. The communi-cations personnel notified the Kansas Divisionof Emergency Preparedness, the Kansas AdjutantGeneral's Office and the Kansas Department ofHealth and Environment (KDHE) about theaccident.

The KDHE, Bureau of Radiation Control(BRC) Radiation Protection Specialist from theWichita Office arrived about 9:30 a.m. andassessed the area. He was appointed sitecommander by the Civil Preparedness Agencyjecause KDHE would ultimately be in charge ofthe accident area for clean up purposes. Thedecision was made to cover the area with moreplastic because it was raining. Since theaccident occurred in a construction area (bridgeand highway), a workman from the nearbyconstruction crew was requested to use his

bulldozer to construct several bar dikes ordams in the median and ditch to contain thematerial and water run-off from the contaminatedarea. More plastic was brought to the scene tocover as much material as possible to lessen therun-off or dispersal from the scene.

There was considerable discussion betweenthe Radiation Protection Specialist at the siteand KDHE Topeka Office to fully r^sess thesituation. The rain continued, the dikes anddams had been constructed and the materialwas covered with plastic, therefore it wasrecommended that the area be kept secured andstatic until a Health Physicist from KDHE TopekaOffice arrived at the site. The author wasdirected to proceed to the site, take overcommand and supervise the activities necessaryto clean up the site.

Arriving In the early afternoon theaccident scene was assessed and surveys wereconducted. The radiation levels . found wereless than 1.43 X 10 C kg s (2 mr/hr)around the area. Because of the low radiationlevel from Uranium Yellow Cake, the radiationexposure to personnel did not need to beconsidered or limited. Since Uranium YellowCake is more hazardous chemically thanradiologlcally, dealing with the accident wasmade easier. The initial assessment of theaccident site indicated that of the 54 drumstotal, 28 drums were undamaged. The remaining26 drums were damaged to some extent. Therewere 13 of these drums with their lids off andabout 907 kg (2000 pounds) of Uranium YellowCake spilled in the trailer, ditch and on thehighway. Representatives from the carrier,shipper, a consulting firm specializing inradioactive splllf, the insurance carrier and ahazardous materials company arrived over thenext several hours. After several discussionsand more surveys, a plan of action was formu-lated to obtain a barrel loader to pick up thedrums, load them on another trailer and shipthem to Gore, Oklahoma. Control lines wereestablished around the site to prevent thefurther spread of contamination and limit theaccess of personnel to the site. Security lineswere established about 30.5 m (100 feet) fromthe control lines. This was done to provide anadditional buffer area and was very valuable Inlimiting personnel access. Discussions wereheld concerning procedures to be followed and itwas decided that there needed to be someflexibility in establishing these as changeswould be necessary.

The clean up started with a new semitractor and trailer being brought to the siteand parked outside of the control line. Thetrailer was lined inside with plastic and theundamaged drums were picked up, inspected andloaded. The remaining damaged drums were alsopicked up, wrapped in plastic and loaded on thetrailer.

Two wreckers were brought to the site toupright the semi tractor and trailer. Duringthe time that the wreckers were positioningthemselves and attaching cables to the semitractor and trailer, there was a wind shiftfrom a north-easterly direction to a west

41

northwesterly direction. This placed thecommand center and support vehicles downwindof the site. Operations were halted while thecommand center and support vehicles were movedfrom the east side of the accident site to thewest side placing them upwind from the site.The traffic detoured over the eastbound lanes of1-235, around the site, was stopped for a periodof 10 minutes while the semi tractor and trailerwere uprighted. There was a cloud of yellowcake dust raised that traveled in asouth-easterly direction over the east boundlanes of the highway and the area where thecommand center and support vehicles had beenpreviously located. This cloud dissipated veryquickly and no measurable evidence of contami-nation was found in the area traversed by thecloud. The east bound lanes of 1-235 werereopened and traffic was allowed to proceedaround the site.

The clean up progressed rather slowly forthe next several days as the crew attempted torecover and pickup as much of the spilledmaterial as possible. This material was placedinto barrels for transport back to thePetrotomics Company Mine, Shirley Basin, Wyomingfor reprocessing. There were a number ofproblems associated with the clean up, mostof which were beyond anyone's control. Theweather was most uncooperative as we experiencedsnow, drizzle, sleet, rain, strong- winds andcold temperatures for several days. Duringbreaks in the weather, surveys were conducted ofthe area and contaminated areas were outlinedwith red flags. The surveys were conductedutilizing modified Civil Defense Geiger Mueller(GM) survey instruments. The modificationconsisted of replacing the side window GM tubewith a thin end window GM tube, thus increasingthe instruments' sensitivity. Soil and watersamples were collected around the perimeter ofthe marked contaminated area. These sampleswere sent to the KDHE laboratory for analysisfor confirmation of the contaminated area.

Part of the crew, after briefing, startedto decontaminate the semi tractor and trailer.The tractor was relatively easy to decontaminateby washing with water. The contaminated waterwas diverted from the roadway to a diked area inthe median. This was done for the ease ofremoval of the water. The trailer, however,proved to be a much more formidable task in thatmaterial had become lodged between the flooringand wooden sides. Several attempts were made todecontaminate the trailer by sweeping, vacuumingand washing without success. The decision wasmade to remove the interior sections of thetrailer, decontaminate them to the extentpossible and wrap them in plastic prior toshipment back to Wyoming.

Very tight security was maintained duringthe esitire operation with the insurance companycontracting for a private security force.Those individuals necessary for the clean upoperations were allowed and authorized ir. thearea. The personnel working in the area wheredust or airborne particles posed a problem wererequired to wear respirators and protectiveclothing.

When the weather broke, a very thorough andcareful survey was conducted of the area toidentify and outline the area of contamination.Red flags were used to thoroughly mark andoutline the area. Several soil and watersamples were collected and analyzed for uraniumconcentration. Upon receipt of the data, themagnitude of the contamination was more clearand was rather extensive. There were severaldiscussions held over the next few hours inorder to outline the direction of the operation.It was determined that it would be necessary topick up and dispose of a large volume of soiland water. The clean up crew decided that largeholding tanks were needed for the water, pumpsand tankers would also be needed to pick up andtransport the contaminated water. The contami-nated soil would need to be moved and handledwith a bulldozer, road grader and backhoe. Thewater could be hauled in transport tankers fordisposal and the soil would need to be packagedin drums prior to its transport for disposal.The water and soil would be transported back tothe Petrotomics Company mine for reprocessing.The water was pumped from the diked areas bymeans of a slurry pump into a .transport truck.This fruck load of about 3.79 m (1,000 gal) wasthen transferred Into the large holding tank.This process continued until all of the waterhad been picked up from the diked.areas. Largetransport tankers, about 15.14 m (4,000 gal)each were used to haul the contaminated waterback to Wyoming.

The bulldozer was used to push the contam-inated soil from the ditches and median Intoseveral piles for ease of removal. The contam-inated soil was picked up by hand in areas ofthe ditch and median that were too wet for theuse of a bulldozer. Soae of the crew put onrubber boots and walked out on boards, placedover the extremely soft: areas and with shovelsremoved the contaminated soil. The shoulders ofthe ditch and median were bladed several timeswith a grader to remove the contaminated soil.Each pass with the grader removed about 2-3 cm(1-1.5 in) and this continued until there was noevidence of contamination remaining. In someareas along the shoulder, it was necessary toremove 15-20 cm (6-8 in) of the soil in order toget all the contaminated soil.

The contractor, for this constructionarea, designed a "nuclear grade barrel filler"from a cement bucket for use in filling the 2.1X 10 a (55 gal) drums with the contaminatedsoil. The contractor had four tubular legswelded to the cement bucket so that each drumcould be positioned under it for filling. Thebackhoe was used to pick up the contaminatedsoil from the ditch and median dumping it intothe cement bucket for loading into the drums.The drums were filled one at a time by thismethod. They were then wheeled by a hand truckto an area where they were surveyed and decon-taminated if necessary. As the drums wereidentified as clean they were moved to a cleanarea ready for loading. The drums were loadedtwo at a time, by means of a barrel loader,onto a semi tractor and trailer rig fortransport back to Wyoning. An individual load

42

consisted of about 70 drums each of whichweighed 272 kg ,(600 lbs) and contained about0.085 mJ (1/3 yd ) of dirt.

Upon picking up and removing the contami-nated soil and water, the final soil samples ofthe area indicated a residual concentration of1.11 mBq-5.18 nBq/gm (0.3-1.4 pCi/g). Thiscompared very favorably to the 2.5 mBq/g (0.67pCi/g) average for naturally occurring uranium238 found in the soil in the surrounding area.

The soil and water were all shipped to themine in Wyoming for reprocessing and recoveryof the Uranium Yellow Cake. This was donebecause they were considered a resource sincethe concentration of uranium in the water rangedfrom 592 Bq/in (16 pCi/1) to a high of 2.8 MBq/m (75,853 pCi/1) and in the soil it rangedup to 20.5 k Bq/m (555 pCi/1).

Decontamination of the heavy equipmentnamely the backhoe, bulldozer and road graderwas formidable in that several successivescrubbings and washings were necessary overseveral days in order to accomplish the task.The most arduous, however, was decontaminationof the str.-1 bridge girders that were on theshoulder of the highway. Yellow cake powderhad been spilled on these, since they were veryclose to the overturned vehicle. The girderswere picked up individually, surveyed, decon-taminated and then placed on plastic in a cleanarea. They were also covered with plastic toprevent any possibility of recontamination. Thedecontamination process was very slow andcovered several days before all the equipmentwas cleaned. Only the road surface remainedwhich was almost as difficult to decontaminateas the girders. Numerous scrubbings andwashings were done on the road with very littlesuccess. Detergents, liquid soap dilutehydrochloric acid and various other solventswere tried, but were not effective. Finally ahydrolazer (hig:< pressure spray) system wasutilized and the decontamination of the roadsurface was accomplished. All of the equipmentwas decontaminated to the limits of : a) 5000disintegrations per minute (D/M)/100 cm fixedalpha, b) a maximum of 15,000 D/M/100 cm fixedalpha and c) 1,000 D/M/100 cm removable alphaprior to being released from the site.

Air sampling was conducted during theoperation and no significant air concentrationswere observed. The filter cartridges from theface respirators were analyzed of the workersperforming decontamination techniques and theconcentrations ranged from 2.3 Bq(61.6 pCi)/filter to 5.3 Bq (141.8 pCi)/filter.Urinalysis bioassay was done on 29 individualsassociated directly with the clean up opera-tions. The maximum bioassay result was 31 HgUranium (U)/liter. This level is less than 24Zof the recommended action level of 130ugU/liter for a single uptake on a biweeklysampling program. The personnel dosimetry ofthose individuals utilizing such did not recordany significant exposures above that of back-ground. There were about 24 water samples and65 soil samples collected and analyzed over thecourse of the clean up.

In summary, the clean up decontamination

operations covered a period of 12 days undersome adverse weather conditions. Both thedriver and passenger completely recovered fromtheir respective -injuries. There were about700-2.1 X 10 ni (55 gal) drums of soil andseven tanker trucks of 106.12 m (28,000 gal) ofwater were removed from the site. Severalattempts were necessary prior to the successfuldecontamination of some of the heavy equipment,girders and road surface. The cooperation ofall the parties during the entire operation wasvery good.

The area was decontaminated to the levelestablished by the state. The entire site wasofficially released from state control at6:40 p.m., April 2, 1979. The State HighwayDepartment reopened the highway to traffic thenext morning.

ANS Topical Meeting on Radiological AccMeats—Perspectives and Emergtacy Ptaaaiag

Aerial Survey Efforts in the Search for Radon-Contaminated Housesin the Reading Prong Area Near Boyertown, Pennsylvania

Raymond A. Hoover and Dennis E. Mateik

ABSTRACT At the request of the Common-wealth of Pennsylvania, the Department ofEnergy requested EG&G Energy Measurementsto fly an aerial radiological survey overa portion of the Reading Prong nearBoyertown, Pennsylvania. The survey goalwas to help locate regions where buildingscontained elevated levels of radon gas. A250 km2 area was surveyed. A number ofsices were located. These sitescorrelated fairly well with known geologicfaults in the area.

In December of 1984, the home ofStanley Watras in Boyertown, Pennsylvania,was found to contain excessive amounts ofradon gas. The Commonwealth of Pennsyl-vania's Department of EnvironmentalResources (DER) requested assistance fromthe U.S. Department of Energy (DOE) Inlocating regions with high probability ofcontaining houses with excessive amountsof radon.

In order to respond to emergencies ofthis nature, the DOE maintains a RemoteSensing Laboratory (RSL) In Las Vegas,Nevada, and an extension facility inWashington, D.C. The RSL is operated forthe DOE by EG&G Energy Measurements, Inc.(EG&G/EM), a contractor of the DOE. Oneof the aajor functions of the RSL is tomanage an aerial surveillance programcalled the Aerial Measuring Systems (AMS)(Jobst, 1979). Since its inception in1956, the AMS has continued a nationwideeffort to document baseline radiologicalconditions surrounding nuclear facilitiesof interest. These facilities Includenuclear power plants, nuclear materialsprocessing plants, and researchlaboratories employing nuclear materials.At the request of federal or stateagencies, and by direction of the DOE, theAMS is deployed for various aerial surveyoperations. In response to theCommonwealth of Pennsylvania's request,AMS was deployed to conduct an aerialsurvey of a portion of the Reading Prong

near Boyertown, Pennsylvania. The EG&Gsurvey team was tasked with determiningwhich areas in the survey area were mostlikely to contain houses with excessiveamounts of radon, and reporting thisinformation to the local DER officials asrapidly as possible.

II. RADON

Radon is an odorless, colorless,radioactive gas which is produced by theradioactive decay of radium. Radium ispart of both the uranium and thorium decaychains, two widely distributed, naturallyoccurring, radioactive materials. Radondecays by emitting an alpha particle.Long term exposure to elevated concentra-tions of radon is considered to be ahealth hazard.

The movement of radon gas from itspoint of formation in a soil or rock intoa building is controlled by many factors.Chief among these are the ability of theradon atom to escape from the interior ofthe mineral grain of the soil or rockwhere it was formed, its diffusion andtransport through permiable rock and soil(Tanner, 1964) and its ability topenetrate the walls and basement of abuilding (Fleischer and Turner, 1984).The buildup of radon in a building isfurther controlled by its ability todiffuse out of the house. Since radon istransported slowly through soils,radon-222 (Rn-222, half-life of 3.8 days)is the isotope which is mostly found inbuildings. The other two naturallyoccurring radon isotopes (Rn-219 and -220)have half-lives which are too short topermit either isotope to accumulate insufficient amounts to cause any concern.

III. SURVEY SITE DESCRIPTION

I n i t i a l l y , i t was decided that anarea of approximately 9 square kilometers(km2) (4 square miles) would be surveyed.t h i s area was flown at an altitude of 30meters (100 f t ) and at a line spacing of45 meters (150 f t ) . This survev was

43

44

roughly centered on the Stanley Watrashouse in both the north-south andeast-west directions. At the conclusionof this survey it was decided to increasethe size of the survey area to about 250ko2 (100 square oiles). At the same time,it was decided that a higher altitude anda wider line spacing would not seriouslydegrade the data quality and would alsoshorten the survey tine. The new altitudewas 60 oeters (200 ft) with a line spacingof 90 neters (300 ft). The new surveyarea was to be roughly centered on theBoyertown Reservoir and was to include allof the townships of Earl and ofColebrookdale, which Includes Boyertown.

The Reading Prong is a large geologicformation composed mostly of roetamorphisedprecambrian rocks, whose western edgestarts in the vicinity of Reading,Pennsylvania, and stretches severalhundred kilometers in a Northeasterlydirection. Parts of the Reading Prong arefound in Pennsylvania, New Jersey, and NewYork; Boyertown lies on a small lobe whichdrops down from the naln axis of theformation. In many places the ReadingProng contains rock formations chat areenriched in uranium and thorium.

IV. AERIAL MEASUREMENTS

Measurements of the total count ratesand the energy spectrum of gamma radiationwere made on each of the flight lines atone second intervals. The gamma rays weredetected by eight thallium-activatedsodium iodide, Nal(Tl), scintillationcrystals mounted in external cargo pods ona MesserBchmitt-Bolkow-Blohm (MBB) BO-105helicopter. Each Nal(Tl) crystal was 40cm x 10 cm x 10 cm In size. Helicopterground speed was about 36 meters persecond. Altitude was monitored by a radaraltimeter. (Jobst, 1979; Fritzsche, 1982)

Scintillation pulses from the eightdetectors were electronically sunned androuted to the Radiation and EnvironmentalData Acquisition and Recorder (REDAR)syatea on board the aircraft.

The system's ability to detect anisotope on the ground depends on severalfactors. These include the system'sgeometry and the source's strength anddistribution. System geometry includeseffects due to the detector efficiency,collimation, and the movement of thehelicopter. These combine to Halt theability of the system to precisely locatea point source. Source distribution willaffect the system sensitivity because thespectra collected are averaged over thedetector's field of view.

The REDAR system is composed ofseveral microprocessor-based subsystems.The control subsystem collected andformatted gamma ray spectral data from thedetector system. It also collectedaircraft positional data and system livetime Information. Records containing four

one-second data polntB for theseparameters were stored on magnetic tapeevery four seconds. The tape subsystemconsisted of a microprocessor and a dualmagnetic cartridge digital recorder.Radiological data, along with selectedoperational parameters, were presented inreal time on the CRT's of the on-boarddisplay subsystem.

The helicopter position over thesurvey site was determined by two systems:an ultrahlgh frequency ranging system(URS) and the radar altimeter. The URSconsisted of two remotely locatedtransponders and an on-board interrogator.The on-board interrogator used the transittime of the UHF pulses from the trans-ponders to obtain the distance from theaircraft to each remote unit. ThisInformation was then used to preciselydetermine the helicopter's position in thesurvey area. The radar altimeter measuredthe aircraft altitude above ground levelin the same manner. Position and altitudeinformation were processed in real time bythe steering microprocessor. These dataprovided steering indications to the pilotfor flying the predetermined flight linesat the desired altitude. Positional data,gamma ray spectral data, atmosphericpressure, and temperature were allrecorded on magnetic tape.

The magnetic tapes with the recordeddata fro* the aerial radiological surveywere processed after each flight with theRadiation and Environmental Data Analyzerand Computer (REDAC) system. Thiscomputerized data analysis system wasbuilt Into an Airstream motor home whichwas modified to carry the REDAC.

V. DATA ANALYSIS

Data from the helicopter flights wereanalyzed while in the field and theresults were reported to members of thePennsylvania DER team who were conductingthe ground-based building measurements.

Data analysis for the Initial lowaltitude survey was complicated by abroken radar altimeter. A replacementcould not be obtained for several days.Half of the low altitude survey was flownwithout the radar altimeter. At thataltitude there are many visual referencesfor the pilot to key in on and, therefore,It was felt that the pilot could maintainthe altitude with a fair degree ofaccuracy. It was also important that theaerial data be gotten to the DER personnelas rapidly as possible. Another majorcomplicating factor during the survey wasthe weather. During the course of thesurvey, it snowed three tines and rainedheavily once.

The presence of varying amounts ofsnow and rainwater on the ground makescomparison of the data on a day to daybasis difficult. Normally, an exposurerate contour map of an area would have

45

been prepared. The exposure rate valueswould have been derived from the grosscount rates due to te r res t r ia l gamma rayemi t t e r s by summing the counts In eachchannel of each spectrum between 0.040and 3.0 MeV. These resu l t s would thenhave been converted to exposure rates atone meter above ground by application of apredetermined conversion factor. Thepresence of snov or rainwater on theground will change the spectral shape andthe count rates due to attenuation. It ispossible to correct for the presence ofthe rainwater or snow (Fritzsche, 1982)but making these corrections are timeconsuming and were not done. Ins tead,contours of standard deviations above thebackground were plotted. For this purposean average count ra te and the standarddevia t ions of the count ra te for eachd a y ' s f l i gh t were found. These valueswere then used to set the contour plotlevels. At the end of the survey mission,the levels from each day's flights werenormalized to an arbitrary reference valueto give continuity to the contour l ines.

VI. RESULTS

Figure 1 shows the locations wherethe count rates are at least two (2)standard deviations above the background.The dashed lines indicate the approximatelocations of known faults in the ReadingProng (as obtained from Geologic Map ofthe Precambrian Rocks and HardystonFormation of the Boyertown Quadrangle,Berks County, Pennsylvania, Department ofInternal Affairs, Atlas 197).

Many of the high count regions can beseen to correspond with the approximatelocations of the faults as indicated bythe geologic nap and with high elevations.A number of high count regions do not seemto correspond to any Indicated fault line.This may be due to the presence of unknownfaults. The area of high counts whichcontains the Stanley Watras House is amongthese areas (Site 1 in Figure 1). Anumber of houses in this area were foundto also contain elevated concentrations ofradon gas by the DER ground searchers.The correlation of the high count regionswith elevation is probably due to uraniumbearing rocks being closer to the surface.The areas of high counts in the farwestern side of Figure 1 are located

outside of what is considered to be theReading Prong boundaries. Spectra ofthose areas which do not correspond withfault lines show only the presence ofmembers of the uranium and thorium decaychains.

VII. SUMMARY

A portion of the Reading Prong wassurveyed by air in an effort to locateareas which would have higher proba-bilities of containing radon withexcessive radon levels. A number of theseareas were located. Most of these areascorrolated with known faults in theReading Prong. The land around theStanley Watras house was identified as asite of increased probability of contain-ing excess radon levels. A number ofbuildings in this area were subsequentlyidentified by the DER ground searchers forfurther study. No information has beenreleased on the other areas identified asbeing likely to contain houses havingelevated radon levels because of anagreement not to divulge the informationwhich the Commonwealth of Pennsylvaniamade with the owners of the buildingswhich were inspected.

This work was performed by EG6G/EMfor the United States Department ofEnergy, Office of Nuclear Safety, undercontract Number DE-AC08-83NV10282.

REFERENCES

Fleischer, R. L. and Turner,L. G., 1984, "Indoor Radon Measurementsin the New York City Capital District",Health Physics, 46(5), 999.

Fritzsche, A. E., 1982, The NationalWeather Service Gamma Snow System Physicsand Calibration, U.S. Government PrintingOffice, NWS-8201.

Jobst, J. E., 1979, "The Aerial MeasuringSystems Program", Nuclear Safety, 20, 136.

Tanner, A. B., The Natural RadiationEnvironment, 1964, Ed. Adams, J. A. S. andLowder, W. M., University of ChicagoPress, Chapter 9, Page 161.

46

-----'-'. FAULT LINES

rGROSS COUNT CONTOUR ^

LEVELS

A

I B

CPS

2 - 3 SIGMA

> 3 SIGMA ^

FIGURE 1MAP OF THE BOYERTOWN PRONG SURVEY AREA- SITE 1 INCLUDES THE LOCATION OF IHhWATRAS HOUSE. AREAS WITH NO LABEL ARE BACKGROUND (BKGU). THE SHOWN FAULTLINE LOCATIONS ARE APPROXIMATE.

ANS Topical Meeting on Radiological Accidents—Perspectives and Emergency Planing

Emergency Planners, Look Back at TMI-2Robert L. Long

ABSTRACT. This paper looks back atemergency preparedness and response lessonslearned while going through the TMI-2Accident crisis. It is based on theauthor's active role as a GPU response teamparticipant and on the results of a two-dayworkshop conducted by GPU Nuclear to assessthe lessons learned, both from the accidentand the evolution of emergency preparednessprograms following the accident.Generally, the focus is on areas notcovered in depth by various regulatoryrequirements and guidelines.

Emergency preparedness at nuclearpower plants changed dramatically followingthe TMI-2 Accident. NRC and FEMAregulations and guidelines imposedextensive new requirements. And it isgenerally believed that, as an industry, weare far better prepared to handle nuclearpower plant emergencies than was GeneralPublic Utilities/Metropolitan Edison in theSpring of 1979.

However at GPU Nuclear, the new GPUcompany which now operates the TMI andOyster Creek Nuclear Generating Stations,we believe that a periodic look back at theevents of the TMI-2 accident is aneffective reminder of the bases for all thenew emergency preparedness requirements.We also find that some of the importantlessons we learned are not necessarilycaptured in guidelines and regulations.

For example, radiological supportactivities both on and off-site identifiedthe difficulties of developing andmaintaining high quality, retrievablerecords. Particularly, during the firstseveral days of the accident, in-plant

radiological conditions were changingdramatically. The survey records beingkept often failed to note the date andtime that measurements were made.Sometimes the description of the locationwas also too general to really allowother individuals to determine the sourceof the radiation hazard. Thus, wecontinue to place strong emphasis in radtech training on the need to develop andmaintain clear and accurate records.

In fact, the subject of records ofall kinds was a major concern after theaccident as we tried to determine thesequence of events. Thus, we emphasizethe need to keep accurate logs, telephoneconversation records, data sheets, checklists, etc. such that events can berecreated. Operators and technicians arereminded of the importance of placingaccurate time narks on strip charts andnoting chart drive speeds. Periodicallywe have an independent group develop thesequence of events of an emergencypreparedness drill scenario using therecords kept by the drill participants.This "records" scenario is then comparedwith the detailed drill scenario data andany problems are identified andaddressed.

During the first few days of theTMI-2 accident several workers receivedexposures in excess of 10CFR20 allowableexposure limits. This has led toextensive attention to and practice ofthe post-accident radioactive liquid andgaseous sampling techniques. Since noblegas clouds are capable of giving doserates in excess of 100 rem/hr, andprimary water samples can also have veryhigh dose rates, it is also necessary tohave available teletectors and other highrange portable Instruments (up to 1000rem/hr) to be able to properly perform

47

48

radiation surveys under accidentconditions. Appropriately these activitiesreceive close attention from NRC inspectorsand Institute of Nuclear Power Operations(IHPO) review teams.

During the THI-2 accident the abilityto handle the plant transient, which lastedfor a number of days, and respond to therequirements for emergency plannotifications and continuing demands forupdates on plant status and radiologicalconditions placed unusual demands on theplant staff. One senior manager told areview committee that the plant staff hadpracticed extensively on handling planttransients at the simulator and on makingemergency plan notifications duringin-plant drills but they had never reallytried to do both at the same time. Thus,our emergency preparedness drills at boththe simulator and the plant require thestaff to perform both functions and torecognize and use the extensive supportcapabilities now provided by the emergencyresponse teams.

As we have cone to understand the trueextent of the damage to the TMI-2 core, wehave emphasized the need for continuingawareness and understanding by plantoperators, radiological control and othertechnicians, and management to recognizeand accept the real possibility for acore-damaging accident to occur. Thiseffort Is enhanced by the use ofsymptoa-based emergency operatingprocedures and simulator training whichfocuses attention on maintaining corecooling under a wide variety of transientscenarios.

As is true for many utilities, themajority of company engineers andtechnologists who provide technical supportfor the plant, both normally and during anemergency, are located some distance fromthe plant sites, in our case In Parslppany,N.J., 100 and 150 nilea from the OysterCreek and TMI sites respectively. Duringthe first few days following Initiation ofthe TMI-2 accident, most of these supportpersonnel, plus many others froa outsidethe company, moved to the plant site andset up temporary working arrangements,e.g., In "Trailer City." As discussed in aprevious paper (Long, e£ al_ 1981), "theworking conditions, poor living conditions,the seriousness of the situation, thepressure of the exaggerated publicperception of the problems, the evolvingorganizational problems, separations froaand the need to reassure families, and theurgent nature of the work assigned allcombined to place the technical supportpersonnel under great stress." Thus, ourpresent emergency plan includes facilities

with dedicated communications links -telephones, telecopiers, and computerterminals tied Into the plant processcomputers - which permit and facilitatethe accomplishment of the technicalsupport by the staff while allowing themto remain in their home office and homeenvironment. Provisions are made for24-hour shift coverage of the supportactivities. Personnel have ready accessto all of the needed resources - drawings,technical manuals and references, computerterminals and software packages.

It is generally acknowledged thatcommunications with the news media and thepublic was very confusing and ineffectiveduring the first several days of the TMI-2accident. More recently the Chernobylaccident, while much more complicated withrespect to accessibility by the press, hadnumerous examples of misinformation,exaggeration and overreactlon to both thereal and potential dangers from theradioactive releases. A large full-timecommunications staff is now engaged atGPUN in developing, fostering, andmaintaining effective relationships withnews media personnel and the public.After several years effort, we now haveagreement for media representatives froathe utility, county, state, and NRC tooperate from a coiwon, dedicated mediacenter. This permits on-the-spotcoordination and verification ofinformation released to the media and thepublic. We also have trained technicalspokesman assigned to the media center towork with the communications personnel atInterpreting and making understandable theemergency event Information.

Another related activity has beenpersuasion of the technical personnel,particularly operations and radiationcontrols, of the need to provide timelyand accurate information to thecommunications staff In the aedla center.This is a particular challenge when theseplant personnel are focusing essentiallyall of their attention on identifying thecauses of the emergency and taking stepsto correct and mitigate any adverseconsequences. Clearly, both the TMI-2 andChernobyl accidents emphasize theImportance of providing accuratemeaningful information to the media andpublic while the event Is ongoing. Thisneed la emphasized and practiced in everyemergency drill conducted at TMI-1 andOyster Creek Nuclear Generating Stations.

As technical support personnel froathe GPU companies, as well as companiesand government organizations froa all overthe United States, began to arrive at theTMI-2 site in the first few days following

49

Initiation of the accident, major demandswere placed on security and trainingpersonnel to maintain appropriate controlswhile enabling rapid access to the site forthese hundreds of additional personnel. Itis essential to have in-place (or readilyavailable) facilities to check-inpersonnel, perform rapid security checks,validate previous radiological workertraining, provide rapid sitefamiliarization training, and Issueappropriate badges, key cards, anddosimetry. The need for an ability tomobilize outside resources was againclearly demonstrated during the Chernobylaccident.

In the summer of 1983, GPU Nuclearconducted a two-day workshop to review andidentify the lessons to be learned from theTMI-2 accident and the whole realm ofactivities evolving In the aftermath of theaccident. The results of this workshop,including 95 lessons learned statements andassociated questions to evaluate theeffectiveness of the responses to eachlesson learned statement were reported inLong, 1985. One whole section of theworkshop was focused on emergencypreparedness. This section and the othersassociated with operators, radiologicalcontrols, maintenance and technical andother support are periodically reviewed todetermine whether GPU Nuclear ismaintaining the required readiness torespond to any emergency at our nucleargenerating stations.

A recent review of these GPUN lessonslearned indicates that many of them arereemphasized by the Chernobyl accident(DOE, 1986). While the Russian report 1Bvery sketchy with regards to the extent ofemergency planning and effectiveness ofimplementation of any plans, we need tocontinue to seek information which will aidus In being prepared to handle anyemergency which may occur at any nuclearpower plant. The viability of nuclearpower as a contributor to the world'sfuture energy needs depends on our abilityto look back at TMI-2, and now Chernobyl,and learn our lessons well.

REFERENCES

Long, R. L., Criramins, T. M. and Lowe, W.W., August, 1981, "Three Mile IslandTechnical Support," Nuclear Technology,54, 155-173.

Long, R. L., 1985, "Summary Report of theGPU Nuclear TMI-2 Lessons LearnedWorkshop", Proceedings of the AHSExecutive Conference, TMI-2: A LearningExperience

Department of Energy, 1986, Soviet Reporton the Chernobyl Accident, EnglishTranslation, Prepared for IAEAConference, Vienna, Austria

Section II

Technical Aspects

Chairman: Harold Lamonds

EG & G, Inc.

ANS Topical Meeting on Radiological AccMetfs—Perspectives and Emergency Plaaaiag

Source Terms Derived from Analyses ofHypothetical Accidents, 1950-1986

William R. Stratton

ABSTRACT - This paper reviews the history ofreactor accident source term assumptions. Afterthe Three Mile Island accident, a number oftheoretical and experimental studies re-examinedpossible accident sequences and source terms.Some of these results are summarized in thispaper.

Subsequent to the developments during WorldWar II, the Atomic. Energy Commission (AEC) wasrequired to site, construct, and operate newproduction reactors. To obtain advice on howbest to solve the policy and technical problemsinvolved, a Reactor Safeguards Committee wascreated. This committee issued a Summary Report,WASH-3, (AEC, 1950) which advised the AEC on thehazards involved and on siting policies.

The assumptions in WASH-3, relative toescape of fission products (the source term),were uncomplicated; they simply assumed that mostof the fission products become airborne sub-sequent to an accident. This was stated in oneplace "as practically the total fission productcontent of the pile," or, in another section,stated as, "assume that ...50% is present in theradioactive cloud." These releases, then, alongwith various atmospheric diffusion assumptions,served to define hazards as a function ofdistance from the reactor. Later publications(e.g., Weil, 1955; Parker, 1955; McCullough,1955) in the early 1950s made comparable as-sumptions about the release of fission products.In none of these analyses was the releaseconsidered mechanistically, taking into accountthe chemistry of fission products, reactordesign, confining buildings, etc.

In 1956, the AEC and the Joint Committee onAtomic Energy recognized the inevitable large-scale application of nuclear energy for theproduction of electricity, and they undertook astudy to gain a more comprehensive understandingof the potential public hazards of nuclear powerreactors. This study became The TheoreticalPossibilities and Consequences of Major Accidentsin Large Nuclear Power Plants, WASH-740, (AEC,1957).

This study created what the Nuclear Regula-tory Commission has described as hazard statesrather than severe accident analyses. Threecases were considered, each arbitrary and with nodiscussions of sequence of events, chemistry,structural retention, etc. In Case A, the

containment is assumed to remain intact (andtight), and hazard to the public derives onlyfrom direct radiation from the building. In CaseB, the volatile fission products (noble gases,halogens, and 17. of the strontium) are assumed tobe released from containment, but all othersremain within the protective shielding. Case Cis the most severe hazard state, and the releaseof 50% of the core inventory is postulated.These cases are summarized in Table I.

The Windscale reactor accident on October 9,1957 (Morowitz, 1981; Clark, 1974; Stewart, 1958)had a significant influence on later studies andon the development of nuclear reactor regulationsby the AEC. This United Kingdom reactor was anatural uranium metal, air-cooled, graphite-moasrated, plutonium-production reactor. Thetemperature of the graphite in this reactor wassufficiently low that, as it absorbed neutron andgamma-ray energy, it tended to grow in volume andbecome distorted. This energy storage effect wasfirst identified by Dr. Eugene Wigner and, hence,is sometimes called "Wigner energy" or "Wigner'sdisease."

On October 7, 1957, procedures to releasethe "Wigner energy" were started at Windscale,but the annealing operation did not go smoothly,as had been the case in the eight previousattempts. For a complicated set of reasons, theheating of the graphite was not controlledadequately, and by October 9 (10:00 p.m.), onegraphite thermocouple indicated such a hightemperature that the pile physicist was requiredto allow air flow through the core to cool thegraphite. The situation continued to deteriorateuntil, on October 10, increasing amounts ofradioactivity were recorded in t>->. 400-ft stack,and the fuel was presumed to be burning. Whenthe fire could not be controlled by injectingcarbon dioxide, it was finally quenched onOctober 12 by flooding the reactor with water.

Because the fuel became very hot and theaccident extended over several days, the escapeof fission products was significant. Theassessment of the amount of radioactivity thatescaped to the environment is given in Table II,along with the estimated inventory in the 150channels in the core that became hot. These rtataare taken from "Long-Range Travel of the Radio-active Cloud from the Accident at Windscale",(Stewart and Crooks, 1958).

53

Table I. WASH-740 hazard statosa

Containment

Release

(Source

Tern)

Hazard

Case A

(Contained)

Tight

0.0

Radiation from

containment

Case B

(Volatile

Release)

Fai led

100% Xe,

Kr, I, Br;1* Sr

Airborne

activity

Case r.

(Release of 50%

ot Inventory)

Failed

50% Fission

Product

Inventory

Airborne

activity

aAEC. (WS7).

Table II. Release during Uindscale accident'1

Isotope

8 5Kr

89Sr

9 0Sr

9 7Zr

106Ru

1 2 9 ^

132To

131,

131mI

1 3 3Xe

" 7 C S

""ce

Inventory

(150 channels, Ci)

275,000

11,500

32,000

161,000

162,000

12,300

7.18,000

Release

(Ci)

1,600

80

2

l>60

80

684

12,000

20,000

1,754

333,600

600

80

Percent ofInventory

0.03

0.02

0.2b

7.4

12.3

4.9

0.04

aStewart and Crooks (1958).

The amount of iodine that escaped from thecore was at least twice the listed value; between20,000 and 30,000 Ci are mentioned as beingdeposited on the filters. The amount of cesiumretained by the filter also was comparable tothat which escaped to the environment. Theestimates of this deposition were 800 to 1000 Ci.Thus, between 11.3Z and 13£ must have escapedfrom the core. The chemical form of this iodineis not known, but most was either gaseous oradsorbed on very small particles. This escape ofiodine can be explained if the hot, dry, oxi-dizing environment is recognized.

In 1960, the Advisory Committee on ReactorSafeguards and the AEC staff (ACRS, 1960; Beck,1960), both commenting on procedures, assump-tions, and rules being developed to licensenuclear reactors (and other facilities), used thephrase, "source term," for the first time. Thedefinition given was "an arbitrary accident is

assumed to occur which results in the release offission products into the outermost buildings orcontainment shell. About 1007. of the totalinventory of noble gases, 502 of the halogens,and 1% of the non-volatile products are assumedto be released (from the fuel). It is thenassumed that this mixture leaks out of theoutermost barrier at a rate defined by thedesigned and confirmed leak rate." The recogni-tion that iodine may be the most importantelement against which to construct defenses isclear.

These severe assumptions, modified only to257. of the halogens, were incorporated into theAEC document, TID-14844, Calculations of DistanceFactors for Power and Test Reactor Site(DiNunno, 1962). In time, this document becameembedded in the federal regulations as a part ofthe rules defining the criteria to which nuclearreactors must be designed and constructed. Thesuccess and high degree of acceptance of TID-14844 has led to its use for the past 25 yearsand its replication worldwide.

The Reactor Safety Study, WASH-1400, (Nf.C,1975) was a major step forward in accidentevaluation. It clarified the meaning of risk andmade quantitative estimates of the probabilityand consequences of each of the factors involvedin an accident. The study showed conclusivelythat the public risk from even serious core-damage accidents is small; comparisons were madeto other risks to which the public is exposed.Even though rigorous analyses were involved,where uncertainties existed, conservativeassumptions were made. Important among thesewere the chemical forms of iodine and cesium.Evidence (thermodynamic and experimental) existedthat the preferred chemical forms were cesiumiodide and cesium hydroxide, but this evidencewas not deemed sufficient to overcome theaccepted belief that iodine and cesium wouldbehave as gases or inert aerosols at hightemperatures. This neglect of the chemistry offission products, along with assumptions aboutcontainment capability, led to high estimates ofthe release of some fission products duringsevere core-damage accidents. Accidents oflesser severity (during which engineered safe-guards survived and operated) led to much smallerreleases. Essentially all the risk is derivedfrom only the most severe core-damage accidents,even though their probabilities were very small.The full set of accidents was divided into ninerelease categories for pressurized water reactors(PWRs) and five release categories for boilingwater reactors (BWKs). These are summarized inTable III.

The initial criticisms of WASH-1400 were inthe probability estimates for the severalaccidents. More recently, the estimates ofrelease fractions for different accidents havebeen very closely examined, and while, ingeneral, large reductions are found, the magni-tude of the change is being debated.

The accident at Three Mile Island on March28, 1979, was severe in terms of physical damageto the reactor and trauma to the public, but therelease of fission products, except for noblegases, was very small (Kemeny, 1979). Theaccident was triggered at 4:00 a.m. by failure of

55

Table III. WASH-l'tOO accident release categories3

Release Prob./ Fraction of Core Inventory Released to

Cat. Reactor- the Environment

Vr.

PWR1

PWR2

PWR3

PWR4

PWRS

WR6

PWR7

PHR8

PVR9

BWR1

BWR2

BWR3

miu

BWR5

"The^The

9E-7h

8E-6

J.E-6

5E-7

7E-7

bZ-b

'.E-b

4E-5

UZ-k

l£-6

6E-6

2E-5

2E-6

1E-4

Xe-Kr

0 . 9

0 . 9

0 . 8

0 . 6

0 . 3

0.3

6E-3

2E-3

IE-6

1.0

1 .0

1 .0

O.b

5E-*

• I

0 .7

0.7

0.2

0.09

0.03

8E-*

ZE-5

1E-4

1E-7

O.U

0 . 9

0 . 1

8E-U

6E-11

Cs-Rb

O.U

0 . 7

0.2

0.0".

9E-3

8E-<«

1E-5

5E-4

6E-7

OA

O.b

0 . 1

SE-3

fcE-9

data are taken from NRC (197S)common convention i s used, i . e

Te-Sb

0 . 4

0 . 3

0 . 3

0.03

5E-3

IE-3

2E-5

1E-6

1E-9

0 . /

0 . 3

0 . 3

t>t-l

8E-12

, App.. , 3 X

Ba-Sr

0.05

0.06

0.02

5E-3

1E-3

9E-5

1E-6

1E-8

IE-11

0.05

U.I

0.01

6E-fc

8E-1U

V, p .

io-6 =

Ru

0.'.

0.02

0.03

:IE-3

6E-4

VE-5

1E-6

0

0

O.b

0.03

0.02

6E-1»

0

V-U.• 3E-6.

La

3E-3

4E-3

3E-3

UY.-U

7E-i

IE--.

2E-?

0

0

5E-3

I.E-3

3E-3

1E-I.

0

the feedwater system which in turn caused theturbine-generator to shut down, the PORV (reliefvalve on the pressurizer) to open, and thereactor to scram. The drop in system pressurecaused the high-pressure injection system tostart in 2 minutes, but the operators misinter-preted the information available to them andturned this system off at 4 minutes and 38seconds. Some of the confusion was caused byclosed valves in the auxiliary feedwater system(in violation of technical specifications).These valves were opened at 8 minutes, thusrestoring a heat sink for the primary system, butthe emergency cooling was not restored tooperation for 3 hours and 20 minut.es. The maincoolant pumps were turned off at 1-1/4 and 1-3/4hours.

By 6:00 a.m. fission products (probably Xe,Kr) were detected in the containment building, at6:54 a.m. in the auxiliary building, (causing astation alarm), and by 7:24 a.m. they weredetected offsite (causing a general alarm to bedeclared). Within the reactor vessel, thesituation worsened monotonically for about3 hours (until 7:00 a.m.) when one main coolantpump was activated briefly. Following thisinjection of water, the condition of the coreagain worsened steadily until 3 hours and20 minutes into the accident (7:20 a.m.), whenthe high-pressure injection system was restarted.We now know that much of the core disintegratedand some fuel melted or liquified a few minutesbefore this time.

Fission products escaping from the hot fuelinto the water (or steam) in the core werecarried in the primary system flow. Someremained, some were carried into auxiliarybuilding tanks by way of the let-down system, andsome escaped into containment by way of the PORV.Between 2.5 X 106 and 13 X 106 Ci of xenon-133(inventory 154 X 10°) were released from theplant, but only about 15 Ci (inventory 64 X 10°)of iodine were detected offsite. No metallicfission products were found outside the contain-ment and auxiliary building (Kemeny, 1979).

The fact thar a severe fuel-damage accidentoccurred in a commercial nuclear power station,the incredibly wide publicity given to the event,and the surprisingly low release of iodine (andno metals) relative to the noble gases, stimu-lated the technical community to create aresearch program of unprecedented magnitude. Inthe United States, this program has been fundedby the Nuclear Regulatory Commission through thenational laboratories and their contractors,through the Electric Power Research Institute(EPRI), the Industry Degraded Core RulemakingProgram (IDCOR), and through programs funded bysingle utilities (New York Power Authority(NYPA)/Risk Management Assoc.) and by architectengineering firms [Stone and Webster EngineeringGorp.(SWEC)]. Major safety-oriented experimentshave been completed at the Argonne, Sandia, andIdaho National Laboratories, and theoreticalstudies have been done by several organizations.

The theoretical studies have had theobjective of rcexanining the risk dominantaccident sequences found in WASH-1400 andassociated with PWR-1,2,3, and BWR-1,2,3. Thesesequences involve postulated failures of piping,of electric power, and selective failure of theengineered safeguards (containment, containmentspray, cooling, emergency core cooling, etc.).Because of redundancy and diversity, the prob-abilities are low, but a de minimus probabilityhas not yet been defined and accepted. Thereexamination of these sequences has beenunderway for several years.

A sampling of theoretical results from themajor U.S. study centers is given below. PWRdata are given first for several reactors fromvarious organizations, with BWR resultsfollowing. The notation insofar as is possibleis that of WASH-1400. Some explanatory notes areadded to clarify some of the inevitableopaqueness. Generally, the tables are not acomplete representation of the organization'sefforts, and contain only a representativesampling.

Table IV summarizes a recent EPRI publi-cation relative to the Surry reactor. The S2C6sequence did not lead to core uncovering.

Table V contains results that the New YorkPower Authority and Risk Management Associatescalculated for the Indian Point-Ill plant.

The IDCOR program (IDCOR; 1984, 1986) hasreexamined a number of reactors and sequences indetail. Some of their data are shown in TableVI. The drains referred to are those between thetwo volumes in the ice condenser. The IDCORprogram has been underway since 1981 and isdrawing to a close this year.

56

Table IV. Calculated source terms for Surry (PWR)a Table V. Caculated source terms; lor Indian Point (PUN)*1

rission

Product

Croup

Fraction of Core Invontory Released to

Environment Touring Accident Sequences

Xe-Kr

I-l!r

Cs-Rb

Ba-Sr

Ku

La

<.K-3

2E-3

?.E-3

UK-U

•>K-U

5E-/

5F.-7

1.0

6E-2

6K-?.

ZE-2

8E-3

;.E-b

2E-S

0.0

0.0

0.0

0.0

0.0

0.0

0.0

"Hitman et al. (1985).bFailure of secondary system steam relief valves, power

conversion system, and atixiliary feedwater system, with

failure to recover either onsite or offsite electric power

within about 1-3 hr following initiating transient which is

loss of power.cDry; low pressure injection system check valve failure.

Flooding in Safeguards Building ignored; deposition in

building included.

'small loss of coolant accident (equivalent diameter 0.5-2

in) with failure of containment spray injection system and

containment failure due >n overpressure. Core remained

covered, spray operative. Same conclusion in Silberberg et

al. (1986), p. '.-b7.

d,.

Fission

Product

Groui*

Fraction of Core Inventory Released to

Environment During Accident Sequences

AB" mi.B' TMI.B'

Kr-Xe

Csl

CsOH

Tc

3E-6

9K-6

1.3E-6

2E-S

1./E-5

6.5E-9

u

3

.2E-6

.2E-6

.3E-9

aThese studies were completed in 19B3-198fc and quoted in

ANS (1984). Reexamination with improved codes is planned.

"Intermediate to large loss of coolant accident with failure

of electric power to engineered safety features.cFailure of secondary system steam relief valves, power

conversion system, and auxiliary feedwater system, with

failure to recover either onsite or offsite electric power

within 1.-3 hr following an initiating transient which is

loss of affsito power.

Pump seal loss of coolant accident with other characteristics

as in footnote c.

VI. Calculated source terms for Zion and Sequoyah (PWRs)a

Fission

Product

Group

Fraction of Inventory Released to Environment During Accident Sequences

Zion Sequovah

TMLB'-BC TMLB'-6e Vd S?HK-6jff S2HF-fl

h

Xe-Kr

I-Br

Cs-Rb

Te-Sb

Ba-Sr

Ru

La

1.0

1.7E-3

1.7F.-3

2E-5

IE-5

1E-5

__

1.0

IE-?

1E-2

3E-4

6E-&

6E-4

__

1.0

8E-5

8E-5

8E-5

bE-5

1E-5

1.0

5E-<.

6E-4

3E-5

lE-"i

IF-5

1.0

IE-/.

IE-it

lE-i.

IE-5

1E-5

1.0

2E-5

7E-S

1E-S

1E-5

!E-b

1.0

1E-5

2F.-5

2E-5

1E-5

1E-5

0

1.0

l.ftE-2

1.6E-2

A-E3

5E-4

5E-'i

0

aData taken from IDCOR (1986).

Failure of secondary steam relief valves, power conversion system, and auxiliary teed water system, with failure to recover

either onsite or offsite electrical power within about 1-3 hr following the initiating transient, which is a loss of offsite AC

power. Containment failure due to overpressure.cSame sequence as described in footnote b with containment failure resulting from inadequate isolation of containment openings and

penetrations.dLow pressure injection system check valve failure.eSame sequence as in footnote b except containment failure due to overpressure.

Small loss-of-coolant accident (equivalent diameter of O.S-2 in) with failure of emergency core cooling recirculation system and

containment spray recirculation systems. Containment failure due to overpressure, Drains blocked.

£same sequence as describe*! in footnote f with drains opc#i.

''Same sequence as describe^ injfootnote f with containment failure resulting from inadequate isolation of containment openings and

penetration!;. Drains open. I f [

57

Table VII is a small sample of the PWR workcompleted by the Stone and Webster EngineeringCorporation. In addition, a number of parametricstudies of the effects of different physical andchemical parameters and their importance insource terms evaluation has been published in theANS Report of the Special Committee on SourceTerms (ANS, 1984).

Table VII. Calculated source terms for Surry (PWR)a

FissionProductCroup

Fraction of Core Inventory Released to theEnvironment During Accident Sequences

AB-BD

1 Ft2AB-6C

1 ft?-THLB'-P"

1 f t 2

Kr-Xe

ICsTe

b

63

.4E-2

.3E-2

.9E-2 1.9E-3

1.5E-21.3E-?6.4E-2

IE-21E-26.5E-3

aData taken from ANS (1984).

Intermediate to large loss of cooling accident with failure

of electric power to engineered safety features.

Containment failure results from inadequate isolation of

containment openings and penetrations.cSame sequence as footnote b, except containment failure at

24 hrs. due to overpressure.

Failure of secondary steam relief valves, power conversion

system, and auxiliary feedwater system, with failure to

recover either onsite or offsite electric power within

about '.-3 hr following an initiating transient whxch is a

loss of offsite AC power. Containment failure results from

inadequate isolation of containment openings and pene-

trations.eLow pressure injection system check valve failure. A

decontamination factor of 50 was assumed for piping aitd

water in the safeguards building. Decontamination by

building structures was not considered.

Table VIII. Calculated source terras for Surry (Pifl?)

Fraction of Core Inventory Keleased to theFission Environment During Accident SequencesProductGroup

S-jG3 '13 AGC TMLB l d> e T M L B ' e > f V d ' S V d >"

S2G

Xe-Kr

I 0.0

Cs

Te

Sr

Ru

La

0.0

0.0

0.0

0.0

0.0

0.57

0.57

0.47

4.7E-3

8E-7

1.6E-4

0.2

0.2

0.1

2E-2

JE-3

.E-4

4.6E-2

3.9E-2

0.11

0.3

0.3

6E-2

5E-3

2E-7

3E-4

5E-2

4E-2

9E-3

1E-3

4E-8

5E-5

0.0

0.0

0.0

0.0

0.0

0.0

aData from Denning et al. (1986).

Small loss of coolant accidents (equivalent diameters less

than about 6 in.) with failure of the containment heat

removal system. Containment is tight; no failure predicted.cInteraediate to large loss of coolant accident with failure

of containment heat removal system. Steam generator fails

due to hot leg break. Containment fails at 50 hr.dData from Silberberg et al. (1986).eFailure of secondary systea steam relief valves, auxiliary

feedwater systen, and power conversion system with failure to

recover either onsite or offsite electric power within about

1-3 hr following an initiating transit which is loss of

offsite AC power. Containment fails at 2.5 hr.fData from BHI (1981.).

^Low pressure injection system check valve failure (dry).

Low pressure injection system check valve failurs (wet)1Small loss of coolant accident (equivalent diameter of

0.5-2 in) with failure of containment spray injection system.

Containment is tight; no failure predicted.

Tables VIII and IX show a small fraction ofthe results obtained by the NRC through i t scontractors, Battelle Memorial Institute (BHI)and Sandia National Laboratory, for the Surry and

hZion reactors. The S2G,fd

S3G and S2C^ 2 3 2sequences for Surry are found to be negligible,but other Surry sequences are significant intheir evaluation and, indeed, are found to becomparable to WASH-1400 results.

The recent Risk Management Associatesexamination of the Fitzgerald BWR-MkI plant,sponsored by EPRI and NYPA, is shown in Table X.

The IDCOR examinations of the Peach BottomBWR-I and the Grand Gulf BWR-Mklll are presentedin Tables XI and XII.

The BWR-Mkll has not been studied in thedetail that other designs have been examined.However, SWEC has examined the Shoreham BWR-MkII,and some of the results are shown in Table XIII.Additional studies have been completed by BMI andNYPA, but these were not received in time to beincluded in this report.

Finally, th<? recent reexamination of thePeach Bottom BWR-MkI by the NRC, BMI, and Sandiais presented in Table XIV.

The dichotomy in source term results fromdifferent investigating organizations is apparentfrom examination of the tables. The suspicion isthat the differences derive both from computercode differences and from boundary conditionassumptions (hydrogen deflagration, steamexplosions, core-concrete interactions, and thepostulated direct heating phenomenon). Theseconjectures, however, remain to be investigated.

Within the past two and one-half years, atleast three reviews or summaries of the state ofknowledge have appeared. These are:

o The Report of the Special Committee on SourceTerms, American Nuclear Society (ANS, 1984).

o Report to the American Physical Society of theStudy Group on Radionuclide Release fromSevere Accidents at Nuclear Power Plants(APS, 1985, presented to the NuclearRegulatory Commission, February, 1985).

58

Table IX. Calculated source terns for Zion (PWR)a Table X. Calculated source terns for Fitzgerald (BWK)a

FissionProductGroup

Fraction of Core Inventory Released to theEnvironment During Accident Sequences

Fraction of Core Inventory Released to theEnvironment During Accident Sequences

S2DCF2d TMLUe

Xe-Kr

I

Cs

Te

Sr

Ru

La

Ce

Ba

3E-6

X.6E-5

1.9E-3

9E-4

2.6E-8

5E-6

1.3E-5

7.3E-4

0.22

0.23

0.32

1./E-4

1.8E-3

2.8E-2

2.6E-2

2.7E-2

3.4E-2

2.5E-3

SE-7

8E-5

6E-5

2E-3

5.7E-3

6E-3

4E-2

9E-5

4E-7

fcE-7

2E-8

2E-3

aData fron Denning et al. (1986)Small loss of coolant accident (equivalent diameter of 0.5"2 in) with failure of emergency core cooling injectionsystem and containment spray injection systea. Containmentfailure at 24 hr.

cSmall loss of coolant accident (equivalent diarater of 0.5-2 in) with failure of emergency core cooling injectionsystem, containment spray injection systea, and containmentspray recirculation system. Containment failure at 2.2 hrcaused by hydrogen bum or direct heating.Same sequence as in footnote c except containment failureoccurs at 14.9 hr due to hydrogen burn or over-pressurization.

transient event with failure of secondary steam reliefvalves, auxiliary feedwater system, and power conversionsystem with loss of emergency core cooling systaa andinoperability of sprays or coolers. Containment failure at3.2 hr caused by direct heating.

o Nuclear Reactor Accident Source Terms, Reportby a Nuclear Energy Agency Group of Experts(NEA, 1986).

The last and most recent of these coveredinformation from Canada, France, the FederalRepublic of Germany, I taly , Japan, Sweden, theUnited Kingdom, and the United States (the manyorganizations mentioned in this summary). Thisreview report identified a number of (1) maindevelopments since WASH-1400, (2) areas wheresource term information i s suff ic ient , and(3) uncertainties and areas of disagreementbetween studies. The most important points ofthese three categories are l i s ted below:

1. Main developments since WASH-1400A. Source terms were overestimated in

the past for most accidents.B. Source terms technology i s more

complex than the treatment in WASH-1400 indicates, and completegeneralization is not possible.

FissionProductGroup

TCC TCd TQUVe

Kr-Xe

Csl

CsOH

Te

Sr

Ru

La

5.6E-2

3E-4

3E-4

2K-4

1E-7

4.5E-7

4.9E-5

4.4E-")

1.1E-S

1.0E-4

9.8E-6

2.2E-6

0.0

0.0

0.0

0.0

0.0

0.0

IE-5

IE-5

1E-5

IE-.'

IE-7

IE-7

aBieniarz and Deeo (1986) and Ritzman et al . (1985).Rupture of reactor coolant boundary with equivalent diametergreater than 6 in with failure of emergency core coolinginjection. Containment failure at 501 Bin due to over-pressure. Release is through reactor building.

°Transient event with failure of reactor protection systea.Vent or assumed failure of containment occurs at 75 psi, 65min.Same sequence as footnote c, with no vent.

^Transient event with failure of normal feedwater xystea toprovide core aska-up water, plus failure of high pressurecoolant injection, raactor core isolation cooling, and lowpressure emergency core cooling systea to provide coreup water. Containment fails at 902 ain.

Areas where source term information issufficient:

A. In-vessel steam explosions of suf-ficient magnitude to fail both theRPV and the containment are veryunlikely.

B. Some containments are stronger thanpreviously thought.

C. Experimental release data for thenon-volatile fission products arelimited; the impact of data isunlikely to be important to thefinal source terms.

D. The most probable chemical compoundsfor iodine and cesium have beenidentified.

Uncertainties and areas of disagree-ment:

A. The several source term studies haddivergent views as to the impli-cations of the present source termpredictions.

B. The inadvertent omission of possiblyimportant phenomena.

C. The boundary and initial conditionsrelated to specifications of theaccident sequences and the plantgeometry.

D. Assumptions relating to applicationof computer codes.

59

Table XI. Calculated source terns for Peach Bottom (BWR)a Table XII. Calculated source terns for Grand Gulf (BWR)a

Fraction of Core Inventory Released to theEnvironment During Accident Sequences

FissionProduceG**oup

Xe-Kr

I-Br

Cs-Rb

Te-Sb

Ba-Sr

Ru

TW»

1.0

0.1

0.2

0.1

4E-4

6E-U

BCC

1.0

0.1

0.1

0.1

4E-4

1E-3

BCd

1.0

">E-2

3E-2

6E-2

1E-4

2E-4

BCe

1.0

3E-2

3E-2

4E-3

8E-5

3E-4

BCf

1.0

6E-4

6E-4

4E-4

'iE-6

1E-5

1.0

5E-2

SE-2

6E-2

3E-5

1E-4

1.0

4E-2

4E-2

6E-2

IE-5

2E-S

aData from IDCOR (1984), p. 10-6.transient event with failure to remove residual core heat.No operator action taken although operator actions canpreclude any release.

cFailure of power to engineered safety features and failureof the reactor protection systea. No operator action taken.

dSaae sequence as footnote c, except operator vents throughwetwall when drywall pressure reaches US psia.

eSaae sequence as footnote c, except operator ref i l l scondensate storage tank to provide continuous CRD flow.

*Sane sequence as footnote c, except operator both ventsthrough wetwall and re f i l l s the CST.

^Transient event with failure of noraal feedwater systea andthe low pressure emergency core cooling systea to providecore make-up water and failure to reaove residual core heat.

hSmall pipe break (equivalent diaaeter of about 2-6 in) withfailure of emergency corfl cooling injection.

FissionProductGroup

Fraction of Core Inventory Released to theEnvironment IHiring Accident Sequences

TQWD AE°

Xe-Kr

I-Br

Cs-Rb

Te-Sb

Sr-Ba

Ru-Ho

1.0

3E-4

3E-4

2E-4

1E-5

1E-5

1.0

8E-4

BE-t

8E-4

IE-5

1E-5

1.0

1E-5

1E-5

JE-5

1E-5

1E-5

1.0

7E-5

7E-S

3E-5

1E-5

lE-b

"Data from 1DCOR (1984), p. 10-11."Transient event with failure of normal feedwater systea toprovide core make-up water and failure to remove residualcore heat.

^Transient event with failure of the reactor protectionsystea.Rupture of reactor coolant boundary with an equivalentdiameter of greater than 6 in and failure of emergency corecooling injection.

^Transient event with failure of normal feedwater systeai toprovide core make-up water and failure of high pressurecoolant injection, reactor core isolation cooling, and lowpressure emergency core cooling systea to provide core cake-up water.

E. Details of aerosol modeling incontainment.

F. Assumptions relative to steamspikes, steam explosions, hydrogendeflagration, and the postulateddirect heating phenomenon.

G. Operator actions can affect themagnitude of tiie source termsignif icantly. The operator actionsthat can be accepted for regulatorypurposes have not been defined.

In final summary: In view of the enormouseffort put forth in the past half dozen years,and now approaching a climax, one must ask thequestion, how can the community of nuclearreactor safety specia l i s ts arrive at a consensuson such an important topic as the magnitude ofthe predicted source term (relat ive to publichealth and safety) , following serious core-damageaccidents? The question i s far too important toremain unresolved.

ADDENDUM (September, 1986)The accident at the Chernobyl plane in the

Soviet Union on April 26, 1986, has been the mostserious, by far, in terms of damage to the plant,magnitude of the source term, health consequences

Table XIII. Calculated source tents for Shorehaa (BWR)a

Fraction of Core Inventory Released to theEnvironment During Accident Sequences

FissionProductGroup ADb AEC •nfi

Kr-XeICsTe

<6E-2<6E-2<6E-7.

<2E-2<2E-2<2E-2

aData froa Warman (1985).

TJupture of reactor coolant boundary with an equivalentdiaaeter of greater than 6 in and failure of vaporsuppression. There is in-leakage to reactor building for13 hr.

cRupture of reactor coolant boundary with an equivalent

diameter of greater than 6 in and failure of emergency corecooling injection,

transient ewmt with failure to remove residual core heat.Retention or suppression based on an assumed decontaminationfactor of 100.

transient event with failure of the reactor protection

system. Retention or suppression based on an assumeddecontanination factor of 100.

60

Table XIV. Calculated source terms for Peach Bottom (BWR)a

Fraction of Core Inventory Released to the

Fission Environment During Accident Sequences

Product

Group

Xe-Kr

I

Gs

Te

S r

Ru

l.a

Ce

FSa

TC-lb

1.0

3E-2

3.3E-2

0.7.6

0 . M

3.5E-7

1.2E-2

2E-7.

0.39

TC-2C

1.0

1.3E-2

l.fcE-2

B.7E-?

0 . 3

2E-6

9E-3

l.SE-f

0.19

TC-3d

1 . 0

6E-4

9 . 6

6.5E-3

8.7.E-3

1.3E-7

3.2E-*!

'4.9E-i.

6.7.E-3

1B-Ie

1.0

1.2E-2

l.ftE-2

0.22

0.37

6E-7

3E-2

fc.8E-2

0.28

TB-2f

3.6E-2

'..IE-2

0.23

O.hb

7E-7

3.5E-?.

3.5E-2

0.3^

0.&6

0.44

0.26

0.2't

1.4E-6

1.3E-2

2.3E-2

0 . 2

aData from Denning et al. (1986).

''Transient event with failure of reactor protection system.

Containment fails at 85 rain.cSaroe sequence as in footnote b except containment failure

occurs at 126 min.dSame sequence as in footnote b. Containment is vented at

88 min.

^Transient event with failure of power to the reactor

protection system. Containment fails at 914 min.

Same sequence as in footnote e except containment fails at

736 rain.

^Failure of low pressure emergency core cooling system to

provide core make-up water.

to station personnel, and integrated radiationdose away from the plant site. The RBMK reactoris graphite-moderated and cooled with boilingwater at about 1000 psi. The design is such thata significant positive coefficient of reactivityexists for the removal of water (or steam) fromthe vertical fuel-coolant flow channels.

The Soviet Report on the Chernobyl Accident(State Committee, 1986) describes the prepara-tions for an experiment. Apparently, thesepreparations led to a condition of the reactorwith little or no boiling in the fuel-coolantchannels. A- reduction of flow at the same timethat the experiment was started led directly to areactivity induced autocatalytic power excursion.The Soviet estimate of the source term by May 6,1986, is given in Appendix 4, Table 4.14, page 21of the DOE translation. They estimate that 20%of the iodine, 13% of the cesium, 5-6% of thebarium, and 3-4% of everything else escaped fromthe plant. Some postulated conditions during theaccident are such that it would be reasonable toexpect some of the cooling water and fuel to havebeen ejected vertically at high velocity.

REFERENCES

Advisory Committee on Reactor Safeguard (ACRS),1960, Letter to AEC Chairman John A. McCone,L. Silverman (October 1960.)

American Nuclear Society, (ANS), 1984, Report ofthe Special Committee on Source Terms, 1984.

American Physical Society, (APS), 1985, "Reportto the APS by the Study Group on Radi...-nuclide Release from Severe Accidents atNuclear Power Plants," Reviews of ModernPhysics 57(3), Part II (July).

Atomic Energy Commission (AEC), 1950, SummaryReport of the Reactor Safeguards Committee,WASH-3 (March 31).

Atomic Energy Commission (AEC), 1957, TheTheoretical Possibilities and Con-sequences of Major Accidents in LargeNuclear Power Reactors, WASH-740(March).

Battelle Memorial Institute (BMI), 1984, Radio-nuclide Release under Specific LWR AccidentConditions, (Draft) BMI-2104 (July).

Beck, C , 1960, Transactions of the WinterMeeting of the American Nuclear Society,American Nuclear Society, LaGrange Park,Illinois.

Bieniarz and Deem, 1985, "An Analysis of a BWRMkl Utilizing Advanced BMI-2104 ComputationalTechniques to Calculate Source Terms," IAEA-SM-281/48, NYPA, Columbus, Ohio (October).

Bieniarz and Deem, 1986, Source Term Analysis ofFitzpatrick Nuclear Power Station, ElectricPower Research Institute, Technical Document(January).

Clarke, R. H., 1974, "An Analysis of the 1957Windscale Accident Using the Weerie Code,"Annals of Nuclear Science and Engineering(73).

Denning, K. S., et al., 1986, RadionuclideRelease Calculations for Selected SevereAccident Scenarios, NUREG/CR-4624, BMI-2139(July).

DiNunno, J., 1962, Calculations of DistanceFactors for Power and Test Reactor Sites,USAEC, TID-14844 (March).

Industry Degraded Core Program (IDCOR), 1984,Nuclear Power Plant Response to SevereAccidents, Technical Summary Report(November).

Industry Degraded Core Program (IDCOR), 1986Reassessment of Emergency PlanningRequirements with Present Source Terms,Technical Report 85.4 (5.3) (February).

61

Kemeny, John G., 1979, Report of the President'sCommission on the Accident at Three MileIsland (October).

McCullough, C. R., M. M. Mills, and E. Teller,1955, "The Safety of Nuclear Reactors,"Proceedings of the International Conferenceon the Peaceful Uses of Atomic Energy, 13,79.

Morowitz, H. A., 1981, "Fission Product andAerosol Behavior Following Degraded CoreAccidents," Nuclear Technology, 53(2) (May).

Nuclear Energy Agency (NEA), Nuclear ReactorAccident Source. Terms, Report by a NuclearEnergy Agency Group of Experts, OECS, Paris(March).

Nuclear Regulatory Commission (NRC), 1975,Reactor Safety Study, An Assessment ofAccident Risks in U. S. Commercial NuclearPower Plants. WASH-1A00 (October).

Nuclear Regulatory Commission (NRC) ContainmentPerformance Working Group, 1985, DraftReport.

Nuclear Regulatory Commission (NRC) ContainmentPerformance Working Group, 1985, DraftReport, NUREG-1037.

Parker, H. M., and J. W. Healy, 1955,"Environmental Effects of a Major ReactorDisaster," Proceedings of the InternationalConference on the Peaceful Uses of AtomicEnergy, 13, 1006.

Ritzman, et al.,1985, Surry Source Term andConsequence Analysis. NP-4096, ElectricPower Research Institute, (June).

Silberberg, Meyer, Mitchell, and Ryder, 1986,Reassessment of the Technical Bases forEstimating Source Terms, (Final Report),NUREG-0956 (July).

State Committee for Using Atomic Energy ofU.S.S.R, (1986), The Accident at ChernobylAES and Its Consequences, DOE TranslationNE-40, International Atomic Energy AgencyExpert Conference, August 25-29, Vienna,Austria.

Stewart, N. G., and R. S. Crooks, 1958, "Long-Range Travel of the Radioactive Cloud fromthe Accident at Windscale", Nature, 182, 627(September 6).

Warman, E. A. 1985, "Investigation of SevereAccident Source Terms", Executive Conferenceon Ramifications of the Source Term,Charleston, S. C. (March 1985).

Weil, George L., 1955, "Hazards of Nuclear PowerPlants," Science, 121, 315 (March).

ANS Topic*! Meeting on Radiological Accidents-Perspectives and Emergency Piawuag

Radiological Source Term Estimation Methods

P. C. Owczarski, M. Y. Ballinger, J. Mishima, and J. E. Ayer

ABSTRACT Pacific Northwest Laboratory is devel-oping methods for estimating the amount and sizedistribution of aerosols generated frompotential accidents in nuclear fuel cyclefacilities. These accidents include fires,explosions, spills, pressurized releases,tornadoes, and criticalities. Experiments havebeen performed investigating releases fromfires, some types of explosions, and spills.Using data from the experiments and informationfrom published literature, models and computercodes are being developed to estimate airbornesource terms from all the accidents listedabove. These methods will be part of an Acci-dent Analysis Handbook to be published in 1987.

I. INTRODUCTION

Pacific Northwest Laboratory (PNL) is inthe final stages of a 7-year project to developtools for estimating airborne source terms fromaccidental fires, explosions, spills, tornadoes,and criticalities in nuclear fuel cycle facili-ties. New experimental data, models, and compu-ter codes have been developed to help estimatesource terms. The final product of this projectwill be an Accident Analysis Handbook to be pub-lished in 1987 along with supporting documents.

Two other laboratories are participating indeveloping project tools. Martin Marietta Cor-poration at Oak Ridge National Laboratory hasdeveloped tools for assessing UFg accidents, andLos Alamos National Laboratory has developedcodes and experimental data for the transportand deposition of airborne materials in filtersand building ventilation systems.

"Pacific Northwest Laboratory is operated forthe U.S. Department of Energy by BattelleMemorial Institute.

bWork supported by the U.S. Nuclear RegulatoryCommission under Contract DE-AC06-76RL0 1830,NRC FIN B2481.

This paper briefly describes experimentalresults and modeling efforts developed in thisprogram.

II. FIRES

FNL has extensively studied accidentalfires involving radioactive material. Experi-mental data were generated from small-scalefires of combustible materials contaminated withdepleted uranium dioxide (DUO) or uranyl nitratehexahydrate (UNH). Models developed from theseand earlier release experiments were programmedinto a compartment fire code (FIRIN). The fol-lowing section briefly describes the experimentsand code.

A. DataExperiments were performed in which

pjlymethyImethacrylate (PMMA), polystyrene (PS),polychloroprene (PC), paper, kerosene/tributylphosphate (TBP), and mixtures of paper/PC andpaper/PMMA were burned. Uranium in both thepowder (DUO) and liquid (UHH) forms was used asthe contaminant. Other parameters that werevaried in the experiments Included air flow,heat flux, and oxygen concentration. PMMA andPS gave the greatest release (up to 4Z of thesource contaminant). Most of the release occur-red while the material was melting and bubbling,before flaming combustion. The smallestreleases (up to 1J) came from burning paper andappeared to be emitted throughout the burn(Halverson and Ballinger, 1984).

A mixture of 30Z TBP in kerosene wasburned in the presence of aqueous nitric acidsolutions with uranyl nitrate dissolved ineither phase initially. The weight percent ofairborne uranium ranged from 0.2 to 7.1 with thehighest releases coming from a burn of anunloaded organic layer over acid with uranium.In all configurations, the mass rate of airborneuranium seemed proportional to the mass rate ofsmoke made airborne.

The size distribution of smoke andradioactive particles made airborne in the fireexperiments was measured using a cascade impac-tor. Size distributions of smoke particles wereall within close range of each other. However,

63

64

the size distribution of radioactive particlesvaried widely depending on the fuel type.

B. ModelsDynamic models describing an acciden-

tal fire with time-dependent formation of air-borne radioactive aerosols have been programmedinto the FIRIH code (Chan et al., 1935). Thiscode is basically a compartment fire model.Heat and mass transfer are modeled to provideinput to a ventilation code (FIRAC). Compart-ment pressure, the airborne smoke and radioac-tive particle characteristics, hot-layercharacteristics, aerosol deposition on compart-ment surfaces, and aerosol flow out of the com-partment are calculated as the fire progresses.Aerosol deposition can occur by settling, Brown-ian diffusion, thermophoresis, and diffusio-phoresis. The compartment decontaminationfactor for fire aerosols can typically rangefrom 1.0 to 10. Filter resistances are calcu-lated as a function of smoke loading. Sourceterras from heated surfaces, hot overpresaurizedvessels, boiling liquids, and spills are alsocalculated in FIRIH.

III. EXPLOSIONS

Explosions have been classified into fourtypest fast and slow physical and fast and slowchemical explosions. The likelihood of fastphysical explosions in fuel cycle facilitiesappears to be remote (Halverson and Mishima,1986). Thus, source term calculation methodswere not developed for this type of explosion.The T'17-equivalent concept provides a simplesdutlor. to the energy release characteristicsfor fast chemical explosions. Mass releaseInformation for metals and liquids is availabletrcra the model of Steindler and Seefeldt (1S30).For slow chemical explosions, the methods usedfor fa3t chemical explosions can be used ifapplied carefully. A series of pressurizedrelease experiments were performed to providedata on seme types of slow physical explosions.Data from these experiments and Information inpublished literature have been used to model theother three type3 of explosions. Pressurizedrelease data and resulting explosion models aredescribed below.

A. DataMethods to estimate source terms frons

explosions were supported by experimentalreleases of pressurized powders and liquids.Pressure was applied using pressurized air, CO2or direct heat. Results of air pressurizationexperiments were reported by Sutter (1983).

During the air pressurizationexperiments, 100 g of DUO (1 pa dia) and TIO2(1.7 13 dia) powders ar.d 100 enr and 350 ca"' ofuranine and uranyl nitrate solutions werereleased at pressures ranging froa 50 to 500pslg. The largest percent of powder made air-borne was 24Z. The saxiarjia aaount of airborr.eIlould source was significantly less, about0.15%.

Experiments in which CO2 was used asthe pressurizing gas for liquid pressurizedreleases gave a larger source tsrra than com-parable air pressurized experiments. Maximumweight percent airborne for C02/uranine releaseswas 0.22. Flashing sprays in which uraninesolution was heated before release gave up to 9%airborne, the greatest of a:iy of the liquidpressurized releases.

B. ModelsModels are being developed for

pressurized releases of powders and liquids.For powders, the model calculates the drag forceon a sphere of powder as it moves upward andgrows after Initial release. The amount air-borne is assumed to be proportional to the dragforce, and the proportional constant is a func-tion of material properties and controllingparameters. Liquid pressurized releases produceaerosol on release of dissolved gases. Thequantity and size distribution of these aerosolsalso are functions of controlling parameters(pressure, viscosity, surface tension, density,and gas solubility).

For other slow physical explosions, avariety of standard fluid-flow equations aresufficient to calculate energy and mass flowfrom ruptured vessels. The hoop-stress equation(Halverson and Mishima, 1986) for rupturedpressure vessels provides a good estimate. Theisothermal expansion equation provides a con-servative estimate of the total energy releaseIn a rupture-type event.

Models for fast chemical explosionsare detailed by Halverson and Mishima (1986).Releases from slow chemical explosions can becalculated using the fast chemical explosionmodels, in some cases, or a universal correla-tion.

A universal correlation, which can beused for estimating explosion releases, wasdevised from spill, pressurized release, andexplosion data (Halverson and Mishima, 1986).This correlation is the upper boundary of theairborne fraction of material at risk as a func-tion of the energy available for aerosol for-mation per unit mass of material at risk.

The bounding equation is

log (wt% airborne) =

.2

-2.6 + fl8.8 log (g-\ - (log JJ-)[ \ 0/ V '0/

E = energy, ergMo * source raass, g

Frcn the numerous subsets of data,other best-fit correlations are provided.Figure 1 is a plot of the above equation aicr.gwith the data subsets.

IV, SPILLS

The lover boundary accidental release eventis a free-fall spill of radioactive powder orliquid In static air. Experiments performed at

65

100

0.0001 l l l . l l I I I I t l l l l I I l l l l l l l

• Uranine PR with CO2• TiO2 Spill• DUO SpillA DUO PR• TiO2 PRvUNH SpillsA Uranine Spills• Uranine PRoUNH PRO Slurry Spills

PR = Pressurized Release• • I " " ' ' ' • I m i l

10b 10' 10°E/Mo, dyne cm/gm

10 1010

Figure 1. PNL Experiment Results and Curve Fits

PNL measured the mass airborne and particle-sizedistribution of spill aerosols for varioussource sizes, source types, and spill heights.Thess experiments and modeling efforts resultingfrom them are described in this section.

A. DataFor the first set of experiments, two

powder and two liquid sources were used: T102and DUO and aqueous uranine and UNH (Sutter etal., 1981). For the source powders used (quan-tities from 25 to 1000 g and fall heights of 1 mand 3 m), the maximum source airborne was 0.12%.The maximum source airborne was an order of mag-nitude less for the liquids (with source quanti-ties ranging from 125 to 1000 cc at the samefall heights).

Subsequent experiments investigatedspills of solutions varying viscosity and sur-face tension as well as slurries. Weight per-cent airborne for these experiments was lessthan half that of previous liquid spills.

The median aerodynamic equivalentdiameters for collected airborne particlesgenerated from spills ranged from 1 to 36 pm.All of the spills produced a significant frac-tion of respirable particles measuring 10 (ini diaand less.

B. ModelsModels for fractions made airborne in

spills are being correlated with controlling

parameters. For liquids «t low viscosity, theairborne fraction correlates well with adlmensionleas combination of liquid density,surface tension, spill height, and liquidviscosity. All liquid spills correlate wellwith spill height and liquid viscosity.

Powders were dispersed while falling,producing aerosols; in contrast to liquid aero-sols formed on impact. A spill model similar tothe pressurized release model for powders wasdeveloped. The drag force on the volume of pow-der is calculated as the powder falls. Theamount airborne is assumed proportional to thedrag force. The Galileo number given below isused to calculate the proportional constant.

Ga

where 6a

gM

Galileo numberacceleration of gravitypowder masspowder bulk densityair densityair viscosity

These models will be published in a documentscheduled for completion later this year.

66

V. OTHER ACCIDENTS

Methods to estimate the source term forother types of events will be considered. Incases where realistic scenarios cannot beformulated for criticalities, current U.S.Nuclear Regulatory Commission Regulatory Guides(3.33, 3.34, and 3.35) appear adequate. Phenom-ena (e.g., earthquakes, tornadoes) that mayresult in severe loss of structural integritycan lead to subdivision of solids on crush/impact. Suspension of powders from homogeneousor heterogenous substrate by high-velocity gases(tornadoes and explosions) are also phenomena ofconcern.

VI. CONCLUSIONS

Experiments have been performed investigat-ing source term releases from potential acci-dents in nuclear fuel cycle facilities. Usingdata from these experiments and Information intha literature, methods are being developed toprovide radiological source term estimationmethods for fires, explosions, spills, torna-does, and criticalities. Before this studythere were few methods available.

Using methods in the Accident AnalysisHandbook, a user can estimate the amount air-borne and particle-size distribution of aerosolsgenerated by the above accidents. Other partsof the handbook will provide source termcalculations for UF5 accidents (Martin MariettaCorporation at Oak Ridge National Laboratory)and methods for determining the transport anddeposition of airborne materials in filters andbuilding ventilation systems (Los AlamosNational Laboratory).

REFERENCES

Chan, M. K., et al., 1985, "User's Manual forFIRIN: A Computer Code to Estimate Acci-dental Fire and Radioactive Source TermReleases in Nuclear Fuel Cycle Facilities,"NUREG/CR-3037, U.S. Nuclear Regulatory Com-mission, Washington, D.C.

Halverson, M. A., and M. Y. Ballinger, 1984,"Radioactive Airborne Releases from BurningContaminated Combustibles," In Abstracts1984 Annual Meeting of the American NuclearSoeiaty. Vol. 46, American Nuclear SocietyAbstracts, La Grange Park, Illinois.

Halverson, M. A., and J. Mishima, 1986, "InitialConcepts on Energetics and Mass ReleasesDuring Explosive Events in Fuel CycleFacilities," NUREG/CR-4593, U.S. NuclearRegulatory Commission, Washington, D.C.

Steindler, M. J., and W. B. Seefeldt, 1980, "AMethod for Estimating the Challenge to anAir Cleaning System Resulting from AirAccidental Explosive Event," In Proceedingsof the 16th DOE Nuclear Air Cleaning Con-ference, Harvard Air Cleaning Laboratory,Boston, Massachusetts.

Sutttr, S. L., 1983, "Aerosol Generated byReleases of Pressurized Powders and Solu-tions in Static Air," 'JUREG/CR-3093, U.S.Nuclear Regulatory Commission, Washington,D.C.

Sutter, S. L., J. W. Johnson, and J. Mishima,1981, "Aerosols Generated by Free FallSpills of Powders and Solutions in StaticAir," NUREG/CR-2139, U.S. NuclearRegulatory Commission, Washington, D.C.

ANS Topical Meeting on Radiological Accidents—Perspectives and Emergency Planning

Changing Perspectives on Severe Accident Source TermsR. S. Denning, P. Cybulskis, and R. DiSalvo

i . INTRODUCTION

Since the potential hazard to thepublic from nuclear power plants i s a l -most ent ire ly associated with the pos s i -b i l i t y of' the accidental release ofradionuclid.es from the boundaries of theplant (source terms), i t i s not at a l lsurprising that source term estimatespervade the regulations that govern powerplant operations. Because there has beena substantial quantity of research sincethe Three Mile Island Unit 2 accidentrelated to the processes that governsevere accident source terms, i t i snatural at this time to determine if theregulations should be changed, either torelax unnecessarily re s t r i c t ive regula-t ions or to add further margins of pro-tect ion , i f appropriate. In this regardthe U.S. Nuclear Regulatory Conmissionhas coimitted to reviewing i t sregulations to evaluate the need for

change 1In this paper we describe changes

in the understanding of processes thatinfluence source term behavior that haveoccurred since the regulations werei n i t i a l l y formulated.

It i s not the purpose o f . t h i s paperto prejudge how or i f the different regu-lat ions should be changed. The formu-lat ion of regulations involves consider-ations of policy quite outside technicalconsiderations of source term phenomen-ology.

I I . HISTORICAL OVERVIEW OF SOURCE TERMPERSPECTIVES

In the 1960s the nuclear industryin the United States moved rapidly fromthe operation of low power demonstrationplants to the construction of plants thatwere many times larger. The l icensingagency, the U. S. Atomic EnergyCommission at that time, was faced withthe challenge of establ ishing regulatorycr i t er ia supported by a very limited baseof experimental data related to the

behavior of accidents in large nuclearpower plants and of plant operatingexperience. To account for the largeuncertainties in this limited data baseconservative safety margins were includedin the regulations. Most of the regu-lat ions and regulatory guides whichinvolve source terms rely on what arereferred to as TID-14844^ release terms.These release teems, which were intendedto characterize a severe accidentinvolving some fuel damage but not pro-gressing to vesse l fa i lure , assumed100 percent release of noble gases,50 percent release of iodine (25 percentfrom the reactor coolant system) and1 percent release of a l l other f i s s ionproducts as so l ids . This simple pres-cript ion, which i s independent of a c c i -dent sequence and plant design features,was convenient for use in regulations.Furthermore, in an era in which the l i k e -lihood of severe accidents was generallybelieved to be very small, the TID-14844release terms were considered to providea conservative basis for regulation.These release terms only described therelease to the containment, not to theenvironment. Consistent with the designbasis c r i t er ia for the plant, the regu-latory framework did not recognize thepotential for containment fa i lure .

A. Changes in PerspectivesSubsequent to WASH-1400

In 1975, the Reactor Safety Study(WASH-1400) was i s sued . 3 This study ledto major new insights into the characterof reactor risk. It indicated thatsevere accidents were substantial ly morel ike ly than previously believed and thatthe consequences of a severe accidentcould vary dramatically, depending on thecharacterist ics of the accident sequenceand on the performance of plant features.

The methods of analysis used topredict severe accident source terms inthe Reactor Safety Study were quitecrude. A minimal data base existed on

67

the release of radionuclides from smallsamples of fuel (primarily at ORNL) andintegral experiments had been performedon the removal of elemental iodine andaerosols from containment atmospheres.The analyses of many complex severe acci-dent processes were performed with verysimple models which primarily attemptedto conserve mass and energy. Neverthe-less, the severe accident source terrasdeveloped for WASH-1400 were recognizedas being the best available characteri-zation for severe accidents with theavailable technology. Following WASH-1400, it was clear that some aspects ofthe regulations, such as the treatment ofsevere accidents in environmental impactstatements and requirements for offsiteemergency planning could no longer beconsidered adequate, and regulations werechanged to incorporate WASH-1400 sourceterms.

B. Changes in PerspectivesSubsequent to TMI-2

After the Three Mile Island Unit 2accident, the importance of developing abetter understanding of severe accidentprocesses became widely recognized. Whathad been considered hypothetical condi-tions prior to WASH-1400, had at least inpart been experienced. The Severe Acci-dent Research Program, that had beeninitiated by the Nuclear RegulatoryCommission after WASH-1400, was accele-rated and expanded.^ Over the past sevenyears this program, complemented by in-dustrial and international programs, hasresulted in a vastly improved under-standing of severe accident processes,the ability to model these processes, anda data base for model validation.

As the WASH-1400 methods wereexamined criticallv in this time period,there was widely-st:ated belief that theWASH-1400 source terms were very conser-vative. Some estimates were that theWASH-1400 values were high by at leastone to two orders of magnitude. As aresult of this frame of mind, it wasgenerally believed that a review of theregulations could result in significantrelaxation of some regulatory cri teria,particularly related to emergency offsiteplanning.

One of the major elements of theNRC's Severe Accident Research Programhas been the development of improvedmethods for source term analysis. Underthe direction of the Accident Source TermProgram Office, an extensive reassessmentwas made of the technical bases forsource terra analysis. A systematic,mechanistic approach to source term anal-ysis was developed. As the ini t ial re-presentation of this approach, a set ofstate-of-the-art source term codes wasassembled, tested, and exposed to exten-sive peer review. In J-.me of 1986 the

"Reassessment of the Technical Bases forEstimating Source Terms", NUREG-09565 waspublished with the recommendation that"The Source Term Code Package isrecommended as an integrated analyticaltool for NRC evaluation of source termsin regulatory applications provided thatuncertainties are considered for eachtype of application". (Underlining pro-vided by authors.)

C. Current Perspectives

In a previous paper, a comparisonwas made between results obtained withthe suite of source term codes (the pre-decessor to the Source Term Code Package)with WASH-1400 results. <> An examinationof the models which predict the principalelements of the source term: in-vesselrelease from fuel, transport in thereactor coolant system, ex-vessel releasefrom fuel, and transport in the contain-ment and secondary buildings, indicatedthat there is a tendency for the WASH-1400 models to overestimate source terms.In particular, current models predictsignificant retention of radionuclides inthe reactor coolant system for most acci-dent sequences. Retention in the reactorcoolant system was not typically givencredit in the WASH-1400 analyses. How-ever, the WASH-1400 analyses do notappear to overestimate all sequences.Perhaps more importantly, the uncertain-ties in source term estimates are quitelarge and typically encompass WASH-1400results.

One element of the source term re-assessment effort supporting NUREG-0956was the QUEST' program in which quanti-tative uncertainty estimates were devel-oped for three accident sequences. TheQUEST study indicated that the uncer-tainty in source term estimates is large.The band of uncertainty about the SourceTerm Code Package results was found to beon the order of a factor of 100 orgreater.

The first application study of theSource Term Code Package is the NRC'sRisk Reference Document, NUREG-1150,8

which will describe the risk to thepublic for six reference nuclear powerplants. In this study particular atten-tion is being given to the uncertaintiesin the description of severe accidentphenomena that influence the source term.Based on the results of the QUEST study,it is likely that the uncertainties insource term estimates will propagate intosignificant uncertainties in the riskestimates for these plants.

In recent years, a much moredetailed understanding of severe accidentsource terms has developed. The methodsare much more mechanistic. They arebetter supported by data. The moredetailed examination of severe accidentprocesses has also helped to identifytechnical issues where a good

69

understanding of source term behaviordoes not exist. There is a much bettercharacterization of what isn't knownabout source term behavior, and as aresult, the implications of theseuncertainties can be better understood.

Thus a perspective on source termsis emerging of broad uncertainty bandsrather than point values. The centers ofthe bands are usually lower than the oldWASH-1400 point estimates but the spreadin uncertainty is typically broad enoughto encompass the WASH-1400 values.Anomalously, a decade ago when the stateof understanding of source term phenomenawas very limited, the use of point es t i -mates for source terms gave the impres-sion of greater precision than currentrepresentations of source terms usinguncertainty spreads. It should be recog-nized that this perspective on sourceterms is not universally held. Thesource term estimates obtained in theIDCOR* program using the MAAP code tendto be lower than Source Term Code Packageresults and the sensitivity studies thathave been performed by IDCOR*0 (and in-dependently by EPRI)** yield narroweruncertainty bands. This further supportsthfl need to attempt to resolve the tech-nical issues that underlie differences inmethodologies.

I I I . THE SIGNIFICANCE OF DIFFERENCES INSOURCE TERMS

It is difficult to compare sourceterm estimates involving release frac-tions for seven to nine different ele-mental groups and to draw inferencesabout the implications to public health.In this section of the paper, we will usesome indices for the relative health con-sequences for the different elementalgroups to develop a measure for comparingrelease fractions. We will also comparesome results of source term analyses withcrude criteria for the onset of differenttypes of health effects.

Before continuing, it is imperativeto warn the reader about the errors in-herent in the simplifications made inthase analyses. The contributions tohealth effects from the different radio-nuclide groups are neither linear norindependent. Furthermore, the magnitudesof accident consequences, particularlyfor threshold types of effects, are verysensitive to assumptions in the analysissuch as the effectiveness of emergencyprotective actions. With regard to thissimplified treatment of source terms,perhaps the most grievous error is toignore the influence of duration of

release on the measure of consequences.The IDCOR program results^ indicatedquite long periods of release for most ofthe sequences analyzed ( i . e . , one to twodays). The major part of the release inthe Chernobyl^^ accident occurred over anine day period. The likelihood of earlyeffects that are threshold in nature canbe substantially influenced by the dur-ation of the release, since for an ex-tended release an individual will havemore time to evacuate, or the wind canshift before the individual is fully ex-posed.

Despite the limitations of simpli-fied measures of health effects, they canallow the reader to obtain a feeling forthe significance of quoted release frac-tions. In a recent study of the"Relative Importance of Individual Ele-ments to Reactor Accident ConsequencesAssuming Equal Release Fractions",^Alpert developed importance weights for25 elements with radioisotopes (60) thatare potentially significant contributorsto offsite consequences. Table 1 repro-duces these weights for two contributorsto early effects, 24 hour bone marrowdose and lung dose, and for total latentcancers. In the analyses performed withthe MACCSl code the following importantassumptions were made:

An end-of-cycle core inventoryfor a 3412 MW PWREqual release fractions of a l lelementsCategory D stability vith awind speed of 5 meters persecond

- Release occurs outside growingseason (no direct depositionon crops or pasture)A nonbuoyant release of eachelement at ground levelA uniform, one-hour release ofeach element, one hour aftershutdownRadiation protection factorsof 0.75 for cloud shine, 0.45for graundshine and 1.0 forinhalationNo emergency protectiveactions.

In the development of accidentsource terms, a number of elements withsimilar chemical behavior are typicallyanalyzed as a group for simplicity. Inthe Reactor Safety Study^ seven elementalgroups were used. In Table 2 the weight-ing factors for each of the elements inthe Reactor Safety Study groups aresummed to give a weighting factor for the

70

group. These weighting factors then pro-vide a simple means to compare the healthsignificance of accident source terra est-imates among themselves or in referenceto health effects benchmarks.

Two benchmarks for comparison arethe TM1-2 accident and the Chernobylaccident, both severe accidents but withdramatically different consequences.From the regulatory viewpoint, however,the most significant benchmarks for com-parison relate to the thresholds ofspecific health effects, in particularthe thresholds for early fatalities andearly injuries.

In the paper, "The Implications ofReduced Source Terms for Ex-Plant Conse-quence Modeling"^ Kaiser examined alarge number of consequence calculationsin order to examine the importance ofdifferent levels of source term reduc-tion. Kaiser developed two crude cr i -teria which can be used to evaluate thepotential significance of source terms.

Criterion 1. If the releasefraction of volatile fission pro-ducts is less than 0.1 and therelease of the ruthenium andstrontium groups is less than 0.05,and the release of the lanthanumgroup is less than 0.01, the con-ditional mean number of earlyfatalities will be very small orzero.

Criterion 2. If the releasefraction of volatile fission pro-ducts is less than 0.01 and therelease of the ruthenium andstrontium groups is less than0.005, and the release of thelanthanum group is less than 0.001,the conditional mean number ofearly injuries will be very smallor zero.

IV. EXAMPLE COMPARISON OF SOURCE TERMRESULTS

In t h i s s e c t i o n we w i l l examine oneof the r i s k dominant sequences from WASH-1400, TMLB'5, to e v a l u a t e the s i g n i f i -cance of improvements in source termmethodology. As p o i n t s of r e f e r e n c e , theTMI 2 a c c i d e n t , Chernobyl a c c i d e n t , Ea r lyFatality Criterion, and Early InjuryCriterion are plotted in Figure 1. Notethat the TMI 2 accident falls substant-ially below either of the early effectscriteria. The principal contribution tothe TMI 2 consequences is the release ofapproximately 7 percent of the noble

gases, not the estimated release of 14 Ciof iodine. The Chernobyl accident isfound to exceed the threshold criterionfor early fatalities. The lung dose isparticularly high because of the largerelease of the non-volatile elements inthe lanthanum group. It should berecognized, however, that although theearly fatality criterion was exceeded, nomember of the public actually received alethal dose because of the duration ofthe release and the emergency protectiveactions taken.

The WASH-1400 analysis of theTMLB'S is illustrated in the figure as apoint of reference. Case 1 is the SourceTerm Code Package analysis of the samescenario. In comparing the two results,it must be recognized that point estimateresults of the Source Term Code Packagecannot be considered best estimate re-sults, since the effect of some importanttechnical issues with high uncertainty isnot included in the analysis. Whereasthe WASH-1400 source term exceeds theearly fatalities criteria by a substan-t ia l margin, the STCP analysis for thissequence is essentially equal to theearly fatality criterion for bone marrowdose and is less than half the criterionfor lung dose. The latent effects healthmeasure for the STCP analysis is approxi-mately one-third of the WASH-1400 value.

If the uncertainties in the anal-ysis are taken into account, however, theperspective changes. One technical issuewill be considered as an example of thepotential effect of source term uncer-tainties. In this accident the predictedretention fractions within the reactorcoolant system of Csl, CsOH, and Te are76 percent, 80 percent, and 69 percent,respectively. The fraction of thesespecies that would be released as theresult of decay heating following vesselfailure is quite uncertain. In Case 2,it is assumed that 50 percent of the ori-ginally deposited volatile elements iseventually released from the reactorcoolant system and that 50 percent of thematerial released late in the sequenceafter containment failure will bereleased to the environment (based onSTCP analyses). In this case, the con-sequence measure for bone marrow dosewould exceed the early fatality cr i -terion.

For the Surry plant, tha likelihoodof early failure of the containment isbelieved to be substantially lower thanat the time of the Reactor Safety Study.One mechanism that could lead to earlyfailure that was not recognized earlieris direct heating, however. If signifi-cant dispersal of the core and direct

71

heating were to occur, the release ofradionuclides to the containment would beenhanced. This effect is shown inCase 3. Again the consequences are sub-stantially increased over the base caseand the early fatal ity criteria are ex-ceeded.

However, i f the containment were tiavoid early failure, the release offission products to the environment wouldbe reduced. Case 4 shows the effect ofdelaying containment failure until12 hours after the start of the accident.Although the consequence measures arereduced, they s t i l l exceed the criteriafor early injuries. In Case 5, i t isassumed that the containment removal pro-cesses are completely effective in atten-uating a l l radionuclides but noble gases-This i s equivalent to a very late failurecase or a filtered vent case. Althoughthe consequence measures are below theearly injury cri teria , the consequencemeasures are close to these criteria. Asmall contribution from the uncertaintyin any of the issues affecting sourceterm magnitude could result in exceedingthe early injury criteria .

V. CONCLUSIONS

Subsequent to the TMI-2 accident,knowledge of severe accident processeswhich influence source terms has expandeddramatically. The methods of analysisthat have been developed or are underdevelopment are more mechanistic and arebetter supported by experimental data.The uncertainties in source terms arelarge, however. Considering the magni-tude of these uncertainties, it is un-likely that in the near future it can bedemonstrated with a high degree of con-fidence that source terms have been re-duced to the degree that early fatalitiesor early injuries cannot occur in a severeaccident. Although it is appropriate atthis time to review source term basedregulations, it is not clear that sub-stantial relaxation of regulatory cri-teria will result. Point estimates ofsource terms should not be used withoutfull consideration of the associateduncertainties.

Furthermore, the magnitude of thesource terms is affected not only by theprocesses of radionuclide release andtransport, but by the timing and mode ofcontainment failure and by the perfor-mance of containment safety systems suchas sprays, suppression pools, icebeds,and air coolers. Thus there is nowincreased awareness of the potentialbenefits of accident management to reducethe consequences of severe accidents.

REFERENCES

1. L. Softer, "Source Term RelatedRegulatory Changes", NRR StaffPresentation to the ACRS,February 24, 1986.

2. J. J. DiNunno, et al , "Calculationof Distance Factors for Power andTest Reactor Sites", TID-14844,March 23, 1962.

3. U.S. Nuclear Regulatory Commission,"Reactor Safety Study—An Assess-ment of Accident Risks in U.S.Commercial Nuclear Power Plants",WASH-1400, October, 1975.

4. U.S. Nuclear Regulatory Commission,"Nuclear Power Plant Severe Acci-dent Research Plan", NUREG-0900,January, 1983.

5. M. Silberberg, et al , "Reassessmentof the Technical Bases for Esti-mating Source Terms", U.S. NuclearRegulatory Commission, NUREG-0956,July, 1986.

6. R. S. Denning, P. Cybulskis, andJ. A. Gieseke, "Changes in SourceTerm Perspectives 1975-1985",Supplemental Proceedings for theThermal Reactor Safety Conference,San Diego, February, 1986.

7. R. J. Lipinski, et al , "Uncertaintyin Radionucli.de Release UnderSpecific LWR Accident Conditions",SAND84-0410, May, 1985.

8. U.S. Nuclear Regulatory Commission,"Risk Reference Document", NUREG-1150 (to be published).

9. Technology for Energy Corp."Nuclear Power Plant Response toSevere Accidents", IDCOR TechnicalSummary Report, November, 1984.

10. Fauske and Assoc , Inc., "TechnicalReport 85.2-Technical Support forIssue Resolution", IDCOR ProgramReport, July, 1985.

11. E. L. Fuller and Z. T. Mendoza,"Structured Sensitivity AnalysesUsing the MAAP 2.0 ComputerProgram", in Proceedings of theInternational ANS/ENS TopicalMeeting on Thermal Reactor Safety,San Diego, February, 1986.

72

12. USSR State Committee on the Utili-zation of Atomic Energy, "The Acci-dent at the Chernobyl Nuclear PowerPlant and Its Consequences",August, 1986.

13. D. J. Alpert, D. I. Chanin, andL. T. Ritchie, "Relative Importanceof Individual Elements to ReactorAccident Consequences AssumingEqual Release Fractions", NUREG/CR-4467, March, 1986.

14. D. J . Alpert et a l , "The MELCORAccident Consequence Code System(MACCS)", in Proceedings of CECWorkshop on Methods for Assessingthe Off-Site Radiological Conse-quences of Nuclear Accidents",Luxembourg, April 15-19, 1985.

15. G. D. Kaiser, "The Implications ofReduced Source Terms for Ex-PlantConsequence Modeling", ANSExecutive Conference on theRamifications: of the Source Term,March, 1985.

CoKr?-SSrYZrKb

TcRuRh5bTeIXeC»BaI*CM?rW!*ptutmCa

E«rl? EipoiuNoraiUied

2U Hour toutMarrov Doit

0.008o.:o.r0.0?1 . 00 . 30 . 10.0]0 . 30.010.04o.a1 . 00.010.140.51.20.30.0030.031 .4

0.0030.0010 . 4

reNorMlistd

Lung

G.OI0.04ZL-U0 . 93.52.00 . 40 . 7

0.033.00.08o.:0 . 61.0o.oos0.090 . 61.68.0

o.a0.35.01 . 40.015 . 0

Nora* HiedToc.L

Utent Canceri

0.Q70.00;l£-50 . 70 . 40 . 70 . 2

0.060.C61.00.004O.OOi0 . 10 . 11E-41 .00 . 2&.082.00.080.030 . 43 . 00.031.1

Table 1. Relative Importance of Individual Elewnts13

Croup

I

2

3

u5

t,

7

X

I

c7

S

t

*»»

P

Kr

Sr

Si)

Sb .

S .

Eleacnti

9e

Pi, *>, Tt

«. Y, C*. ?u, Mp

Zcrlv EEfcctt24 Hour Bone

Hsrrov

0.11

1.0

G.lt

o.ati . ?

0.44

lung

0.045

1 . 0

0.09

0 . 7

1.5

3.a

28.0

UtentCancsr

0.0011

0.1

1.0

0.100.9

J . I

8 . 0

Ul* 1 Ut* t C4M 3 ClM 4 t*M S

J « __

Figure 1. Normalized Consequence Measures for Variations onB'^ Scenario

ANS Topical Meeting on Radiological AccidentsPerspectives and Emergency P

NRC Perspective on Severe Accident Consequence AssessmentThomas McKenna and James A. Martin

ABSTRACT - One of the major roles of the Nu-clear Regulatory Commission (NRC) during a se-vere accident (core melt) is to monitor thereactor licensee to assure that they are rec-ommending the appropriate protective actionsto offsite officials. Initial protective ac-tion recommendations should be made by theplant staff based on predetermined instrumentreadings (emergency action levels) that indi-cate the status of plant systems(s) requiredto protect the public. The determination ofthe need for additional protective action mayrequire dose assessments/projections. In thepast dose assessments have been based onlicensee estimates of releases based on thestack monitor reading. This is not adequatebecause severe accident releases cannot becharacterized by the stack monitor. Dose as-sessments should consider the "probable range"of accident conditions and not just the rela-tionship of stack monitor readings to doses.This paper will summarize the basis for theNRC perspective on consequence assessment andthe tools used to monitor licensee consequentsassessment and protective actionrecommtndations.

One of the major roles of the NRC duringa severe reactor accident (core damage) is tomonitor the reactor licensee to assure thatthe licensee is recommending the appropriateprotective actions to offsite officials.

The NRC first monitors and decides if theplant conditions warrant taking action, basedon an assessment of core and containment sta-tus. Releases resulting in observable earlyhealth effects (injuries and deaths) or highindividual risks (very high radiation doses)offsite can occur only as e result of a severecore damage coupled with eany containmentfailure (within 24-hours of release from thecore). If containment integrity is maintainedor if mitigative engineered safety features(e.g., containment sprays) are functioning,the offsite consequences may be limited to the

immediate vicinity of the site and earlyhealth effects may not be induced even forcore melt accidents. However, during a severenuclear power plant accident early containmentfailure could not be ruled out or "predictedwith confidence." Therefore, severe core dam-age accidents should be classified as generalemergencies, which would warrant initiation ofimmediate, early protective actions offsite.

Three basic conclusions have been derivedfrom the considerable research on severe acci-dents concerning the effectiveness of earlyprotective actions for preventing early healtheffects from a several reactor accident re-sulting In a major release. These are:

1. To be most effective, protective actions(shelter or evacuation) must be taken be-fore or immediately at the tine a majorrelease to the atmosphere occurs.

2. People should immediately evacuate areasnear the plant (within a 2-to-5 mile ra-dius) and remain in shelter elsewhere toreceive additional information or in-structions during the early time frame.

3. Following a major release, the dose fromground contamination in the area beyondwhich people have seen evacuated based onplant conditions (item 2) may become veryimportant in a few hours (e.g, 4 hr), re-quiring immediate radiological monitoringto locate hot spots where evacuationwould be required.

These basic concepts form the foundationfor the NRC's protective action guidance andfor the assessments conducted by the NRC dur-ing its response.

The specific NRC guidance fromNUREG-0654/FEMA REP 1, Rev. 1, Appendix 1, forgeneral emergency protective actions is at-tached as Figure 1. This figure was sent tolicensees via Information Notice 83-28 and isthe basic tool used by the NRC response organ-ization when monitoring the recommenda-tions being made by licensees during a severereactor accident.

73

74

Several points concerning Figure 1 shouldbe discussed. Initial scoping of an emergencyis to be based on a comparison of the plantinstrument readings (i. e., plant conditions)with the predetermined emergency action lev-els. This comparison will identify the plantconditions referred to in the diagram (e.g.,20% cladding failure, fission products in con-tainment, or containment status). Core damageis defined as release of 20% of the gap activ-ity. This level of core damage was chosen be-cause it is well beyond that expected for anyaccident if safety systems operate as planned.

The criteria also call for evacuation inwhat is called a keyhole, that is, in all di-rections close to the plant and a further dis-tance in the downwind direction (e.g., a2-mile radius and 5 miles downwind). This en-sures that people at greatest risk would evac-uate early in the event of a general emergencyeven if wind direction projections areincorrect.

This guidance indicates that people closeto the plant should not evacuate if imminentcatastrophic containment failure that couldrelease large fractions of the radioactive ma-terial in a puff (<1 hr) is likely. The goalis to shelter people inside their homes as thepuff release moves by. This at best would beeffective only for puff releases (short dura-tion); for a long-duration release, most ofthe release and accompanying dose can beavoided by early evacuation. Therefore, ef-fective application requires that the durationand timing of the release be known (predict-able), which may not be the case early in asevere accident. In most cases, puff releaseswould be the result of containment failuresthat were (1) unpredictable (e.g., bypass ac-cidents or explosions) or (2) the result ofoverpressurization. Over-pressurization failures would most likely befar from imminent (following detection of coredamage), allowing considerable time for pre-cautionary evacuation of areas near the plant.The actual time of failure resulting fromoverpressurization would be unpredictable.Therefore, this part of the criteria is gener-ally applicable only for those situationswhere the timing of an imminent puff releaseis certain. An example would be the con-trolled venting of the containment by thelicensee. Consequently, for most core damageaccidents (general emergencies) the populationclose to the plant should be evacuated.

After implementation of protective ac-tions near the plant (based on an assessmentof plant conditions) the determination of theneed for additional protective action may re-quire dose assessments/projections. Boundingdose calculations also may be very iseful incomparing the consequences of varicjs plantresponse options (e.g., venting the contain-ment versus allowing later containment fail-ure). In the past, dose assessments have beenbased on licensee stack monitor readings.However, during a severe accident (core dam-age), releases may not be characterized by thestack monitor. Severe accident releases can

bypass the stack monitor or the accident con-ditions could prevent adequate characteriza-tion of release. In addition, since effectiveprotective action requires promptimplementation, an attempt should be made toproject the magnitude of a release before itoccurs (rather than waiting until it goes outthe stack). Therefore, in performing dose as-sessments, the NRC estimates the possibleoffsite releases based on the probable rangeof accident conditions and not just on stackmonitor readings. These consequence estimatestake into consideration the current and pro-jected status of the core and possible releasepathways.

Some simple tools have been developedthat could at least provide a bound of possi-ble offsite consequences, based on considera-tion of plant (core and containment)conditions. Figure 2 is a very simple example.

Figure 2 uses the concept of an eventtree to display the potential consequences forpublic health due to severe accidents. Movingfrom left to right in the figure, "yes/no" an-swers to scenarios at the top result in a se-ries of branches, possibly to offsiteconsequences. For example, if only the radio-active material contained in the fuel pins(gaps) is released with late containment fail-ure, the offsite consequences would be small(branch 7 in Tig. 2). If all the answers areyes, branch 1 indicates extremely severeoffsite consequence. Figure 2 illustrates twofundamental public health questions during anemergency response at a light water reactor:

What is the status of the reactor core?What is the status of the reactorcontainment?

The answers to these two questions scope thelevel of threat to the public and the need foroffsite emergency response.

Other examples of simple tools used tobound offsite consequences based on plant con-ditions are the 15 precalcuiated offsite doseprojections for severe accident conditions andfor many meteorological conditions presentedin NUREG-1062 Dose Calculations forSevere LWR Accidents (USNRC, 1984).

NRC recognizes that dose projections mustbe used with great caution because of thegreat uncertainties associated with makingsuch assessments during severe accidents. Ta-ble 1 gives an overall estimate of the uncer-tainties associated with dose assessment forsevere reactor accidents. The NRC responsestaff has estimated that, at best, dose pro-jections made during a severe accident may bewithin a factor of 10 of the average fieldmonitoring results—at worst, dose projectionscould be off by several factors of 10. Thelargest single component of this uncertaintyis source term estimation. Unanticipated cat-astrophic containment failure is an example ofan accident for which source term and conse-quently the offsite dose could be underesti-mated by a factor of 100,000 or more.Nevertheless, performing dose projections is auseful tool for determining where to conductearly monitoring or where implementation of

75

additional protective measures should be con-sidered, but reliance should be placed onfield monitoring results as soon as possibleafter an actual release.

I would also like to reassure you thatthe NRC recognizes that its role early in theresponse to an accident is to monitor. In co-operation with local officials, licensees havedeveloped site-specific criteria for recom-mending protective actions to the public ac-cording to the previously discussed concepts.Licensees are required to report these eventsto the NRC within 1 hour (versus 15 minutesfor offsite officials). It is expected totake an additional hour after notification forthe NRC response organizations tc be activatedand to be prepared to comment on protectiveaction decisions or recommendations. Callingthe NRC to confirm a preplanned protective ac-tion would only delay protective actionimplementation.

Margulies, T. S., and Martin, J. A. Jr, 1984,"Dose Calculations for Severe LWRAccident,"NUREG-1062, U. S. Nuclear Regu-latory Commission,

76

FLOW CHART FOR GENERAL EMERGENCY OFFSITE PROTECTIVE DECISIONSTh« foHowinf actions win ba baud on prad«arminad obsarvabla mstrumsniation and plant status indicator!(EAUI contain**% «fv* amargancy Plan and thai hava t**n raviavnad by olfiita officials. Howavar. raaponsiblsoffsita off!clan muat dsclda on tha faanbility of implamantlng ma protactlva actions at tna tima of lha accidant.

MCOMMINDSHUTS*

1 M i l l KAOIUI» MIUS OOWNWINO

JUIITANTIAICOME OAMACIN •KOCKISSOX

"OJICTIO»'ACTUAL

4CCOMM«NDlOTfCnvl ACTIONS>N ACCORDANCEWITH (PA MOa

HICOMMIN0SHILT IH nmAUCAS THAT

CANT I IfVACUATID

HIOM ft ' jKIAflHIVAL.IVACUATI

OTHIHI •Mi(«Hat13-tMIUtl

•ICOMMtNOtVACUATION O>I M i l l RADIUS.

I MIICS DOWNWIND

KICOMMINO(VACUATION OfVMIH KADIUS

10 WlltS DOWNWIND• I1HAI

'111 SITUATION* MQUIKING UHQENT ACTION I Y OPFSITf OWCIALSllaaad on Control Room indicators. No Oos* r^rojacrJona Haquiradl

15-Mlnuta Oacisionmakins. Activation of Alartjnf Systam and 51S Maaaafa

'121 Actual or prejactad r«tw ol 2 0 * 8a» from cors or leaa of pnysical control of Tna plant to intrueara.

*I3I "Puff" ralaaaa I rat a much graatar ttian daaignao; laak ratal.

"(41 For all •vacuarJona. thaltar cfia rarnaindar of tha pluma tPZ and promptfy rtlocata cha popularonaffactad by any ground contamination following plum* paaaaga.

•15) Concantrata on avacuatlon ef araM naar ma plant (a.g. may ba tlms to avacuata 2-mila radioa *n*not tha 5>mila radlual.

Figure 1.Flow Chart for General Emergency Offsite Protection Actions.

77

CORE STATUS CONTAINMENT STATUS CONSEQUENCES

CORE ! CORE MELTUNCOVERY(GAPRELEASEFROM FUELPINS)

EARLYTOTALCONTAIN-MENTFAILURE(BYPASSI

EARLYMAJORCONTAIN-MENTLEAKAGE

LATE (24-hr)CONTAIN-MENTFAILURE

YES

i

NO

7. EARLY HEALTHEFFECTS LIKELY

2. EARLY HEALTHEFFECTSPOSSIBLE

3. EARLY HEALTHEFFECTSUNLIKELY

. A. NO EARLYHEALTH EFFECTS

5. EARLY HEALTHEFFECTS VERYUNLIKELY

. 6. EARLY HEALTHEFFECTS VERYUNLIKELY

7. NO EARLYHEALTH EFFECTS

, 8. NO EARLYHEALTH EFFECTS

9. NO EARLYHEALTH EFFECTS

Figure 2.Event Tree For Severe Reactor Accident Consequence Assessment

78

ESTIMATED RANGE OF UNCERTAINTYBETWEEN PROJECTED AND ACTUAL OFF-SITE DOSE

FOR A SEVERE ACCIDENT (CORE MELT)

Element

Source term (eventand sequence)

DispersionDiffusion (concentration)Transport (direction)Transport (rate)

Dosimetry

Overall (dose anddirection)

At Best

5

222°1

3

10,22°

Uncertainty

Most Likely

100-1,000

545°2

4

100-10,000,45°

Factora

Near Worst

100,000

10180°10 (low wind speed)

5

100.000,180°

aThese estimates are for an averaged dose at a location (e.g.,15-30 min), not for a specific or single monitor reading.

Table 1.Estimated range of uncertainty between early projected dose andactual off-site dose for a severe reactor accident.

ANS Topical Meetiag oa Radiological AccWeits—Perspectives ««d EaiergeMy Pfauukg

A History of Aerial Surveys in Response toRadiological Incidents and Accidents

Joel E. Jobst

ABSTRACT EG&G Energy Measurements Inc., operatesthe Remote Sensing Laboratory for the U.S.Department of Energy (DOE). The Laboratory playsa key role in the federal response to aradiological incident or accident. It assiststhe DOE in the establishment of a FederalRadiological Monitoring and Assessment Center(FRMAC). The Remote Sensing Laboratory hasplayed a major role in more than 14 incidents,including lost sources, accidental dispersions,and nuclear reactor incidents.

I. INTRODUCTION

EG&G Energy Measurements, Inc. (EG&G/EM)operates the Remote Sensing Laboratory (RSL) forthe United States Department of Energy (USDOE).The Laboratory has major facilities in Las Vegas,Nevada, Washington, D.C., and Santa Barbara,California. It maintains nine twin-engineaircraft, helicopters and fixed-wing, equipped asaerial survey platforms. Remote sensingtechnologies include: large area radiologicalmapping, ground-based radiological measurements,high altitude aerial photography, multispectralphotography, multispectral aerial scanning, andairborne gas and particulate saapling. Thelaboratory has acquired and developed a broadvariety of remote sensing equipment. Itspersonnel acquire, analyze, and report data tofederal and state agencies.

The RSL Is a research facility withstate-of-the-art equipment and modern dataanalysis techniques. It has emerged in responseto a unique, clearly-understood need which wasapparent more than 40 years ago, viz., the use ofradioactive materials demands an aerialmeasurement system capable of surveying largeareas in a short time. Aerial measurements ofsurface radioactivity were made in the UnitedStates as early as 1948, originally to determinethe feasibility of airborne prospecting for rad-ioactive ore dtposits (Davis and Reinhardt,1957).In the mid-1950s a series of events prompted theU.S. Geological Survey (USGS) and Oak RidgeNational Laboratory (ORNL) to develop thesesystems for entirely different applications.

Such events included the United Kingdom Windscalereactor accident, the release of radioactiveclouds from nuclear weapon tests in Nevada andthe mid-Pacific, and the emergence of commercialpower reactors.

In 1959 the U.S. Atomic Energy Commission(AEC) asked EG&G, Inc., to develop a dedicatedsystem for use at the Nevada Test Site. TheAerial Measuring System (AMS), as the program isnow known, became operational In November 1960;the first large-area survey was conducted in1961.

System development and operationalcapability have grown continuously under the AECand its successors, the U.S. Energy Research andDevelopment Agency (ERDA) and the USDOE. Aerialradiological surveys are now conducted for avariety of scientific objectives. And becausethe RSL provides aerial survey platforms, a broadvariety of data acquisition and analysishardware, and a cadre of dedicated reaote sensingexperts, the Laboratory has become a developmentcenter for many remote sensing technologies.

Hundreds of routine aerial surveys have beencompleted over the entire United States, someover international waters, in the raid-Pacific,and in foreign countries. This report, however,primarily sunmarlzes those Laboratory surveysconducted in response to radiological incidentsand accidents. Thirteen surveys will bediscussed.

II. LOST COBALT-60 SOURCE

In the summer of 1968 the Laboratory wascalled by the Radiological Assistance ProgramCenter in Albuquerque, New Mexico. A 330 mCicobalt-60 source was lost during routineinterstate shipment from Salt Lake City, Utah, toKansas City, Missouri, a distance of 1930 km. Atthe Kansas City freight terminal the empty leadstorage container was found on its side, the lidbroken open, the source missing. From Salt LakeCity a Laboratory fixed-wing aircraft flew theexact route of the truck at 120 m above thehighway. On the second flight day, the crewobserved a strong radioactive signal just east ofthe Missouri River, near St. Joseph, Missouri.

Subsequent passes pin-pointed the location;spectral data confirmed the presence of Co-60.

79

80

After landing in St. Joseph, the crew drove tothe site and located the source within 10minutes, using hand-held instruments. The sourcewas a 9-cm-long stainless steel capsule, lyingjust off the road, within 75 m of the nearestresidence. Using standard health physicsprocedures and an improvised radiation shield,the crew recovered the source and turned it overto the authorities.

III. ATHENA MISSILE

On 5 July 1970, an Athena missile waslaunched by the U.S. Air Force from Green River,Utah, as part of a routine test program. Asystem malfunction caused the missile toovershoot the target zone at the White SandsMissile Range in New Mexico. The missile wastracked to low altitude about 650 km south of theUnited States/Mexico border, near Torreon,Mexico. Because of the rugged terrain and mapswith little detail, ground crews failed to findthe missile, despite an intensive two-week-longsearch. Since the missile contained two 470 mCiCo-57 sources as part of its payload, theLaboratory was requested to join the search.After flying a grid pattern for 2-1/2 hours, afixed-wing aircraft located the source. Positiveidentification was obtained from gamma spectraldata. The following day the aircraft guidedground crews to the impact site, a shallow craterapproximately 5 m in diameter.

IV. HAVE S1NEH-1

In August 1971 the Laboratory cooperatedwith the U.S. Air Force and several otheragencies in a research project called HAVESINEW-1. An Air Force Athena rocket carrying sixIndependent payload packages was launched fromthe Green River Launch Site in Utah. Thepackages were expected to land, separately) atthe White Sands Missile Range. Each packagecontained either a 50 mCi or a 100 nCitantalum-182 gamma radiation source as an aid torecovery at White Sands. After the packages wereradar-tracked to the Range, the Laboratoryconducted an aerial search with a large array ofNal detectors mounted in a fixed-wing aircraft.Each of the payload packages was a simulation ofa radioisotope heat source intended for use ininstrumented satellites.

Four of the six payloads were recoveredintact, with their radiation sources. Inaddition, two of four 50 mCi sources were

* i

recovered, detached from their payloads. Thelatter apparently disintegrated upon reentry.The two remaining 50 mCI sources, as well astheir respective payloads, wer» never recovered.It has been speculated that these disintegratedprematurely, never reaching the expected impactzone. This early experiment demonstratedeffectively that the aerial detection sy3temcould be used to search for lost radiationsources.

V. MID-PACIFIC ISLANDS

From 1946 through 1958 the United Statesconducted a number of nuclear weapons relatedtests in the mid-Pacific at Enewetak and BikiniAtolls. These are part of the Northern MarshallIslands, 4000 km southwest of Honolulu. Forty-three tests were conducted at Enewetak, 23 atBikini. In 1972 the AEC conducted a detailedsurvey of the total radiological environment ofEnewetak Atoll (USAEC 1973). More than 4500samples from the marine, terrestrial, andatmospheric components of the Atoll environmentwr;re analyzed with Instruments and radiocheraicaltechniques. The Remote Sensing Laboratoryprovided photographic base maps and a detailedaerial survey of the gamma-radiation levels onall islands of the Atoll. It was found thatSr-90, Cs-137, Co-60, and Pu-239 were thepredominant radioactive isotopes still present.

From mid-September through mid-November 1978an even more extensive aerial survey wasconducted in the Northern Marshall Islands(Tipton and Maibaum, 1981). A large array of Naldetectors was mounted on a U.S. Navy helicopterwhich operated from the flight deck of the USNSWheeling, a missile tracking ship. A total ofeleven atolls and two islands were surveyed.Detailed contour maps were prepared for exposurerate and concentrations of Cs-137 and Co-60.

In addition to the aerial survey work, theLaboratory also provided in-situ gammameasurements during the Enewetak clean-up programfrom 1977 to 1979. These data were obtained withhigh-purity germanium detectors mounted ontracked vehicles (Tipton et al., 1981).

VI. B0MARC MISSCLE SITE

In November 1973 the Laboratory flew aradiation survey over a deactivated U.S. AirForce Bomarc missile site near McGulre Air ForceBase, New Jersey. A fire had occurred in one ofthe missile bunkers in 1960; fallout from theresulting cloud of smoke and debris distributedweapons grade plutonium around the site. A largearray of Nal gamma ray detectors was externallymounted on a military UH-1N helicopter. Byflying a flight pattern at 23 ra and hovering at15 m, a Laboratory crew was able to map thedispersion of Am-241, an isotopa whichaccompanies the plutonium and allows it to bemapped from an aerial platform. Surfaceconcentrations as high as 0.81 |j g/m2 weremeasured and a distribution map was provided toaid in site clean-up.

VII. COSMOS 954 SATELLITE

Perhaps the most celebrated incidentinvolving the DOE Remote Sensing Laboratory beganwhen U.S. observers noticed a change in orbit ofthe Russia satellite. Cosmos 954 (USDOE, 1978).The satellite, powered by a nuclear reactor,plunged back Into the earth's atmosphere on 24January 1978. It impacted in the NorthwestTerritories of Canada, scattering radioactive.

81

debris over an estimated ar^a of 52,000 km2. Thesearch and recovery effort, called OperationMorning Light, was a joint program of theCanadian Combined Forces and the USDOE, whichcoordinated the efforts of many U.S. FederalAgencies and private contractors. Large arraysof detu'tors, operated by radiation physicistsprimarily from the United States, were flown atlow altitudes in Canadian C-130 aircraft whichcovered thousands of kilometers each mission.Helicopter teams, equipped with hand-heldequipment, investigated hits and recovereddebris. The recovery effort was extremelydifficult because of major logistics problems inso large a search area, the remoteness of thesite, snow, and extreme cold. More than 100large pieces of radioactive debris wererecovered. But it has been concluded that thereactor core disintegrated during reentry,scattering radioactive partlculate matter over abroad area. Since it was impossible to locateand recover any more than a small percentage ofthis debris, even with the best technologypresently available, further recovery effort wasterminated after several months of intensivework.

VIII. THREE MILE ISLAND

On 28 March 1979 the Three Mile Island (TMI)nuclear reactor was shut down after suffering amajor accident which has forever changed thehistory of nuclear power (Beers et al., 1984).The RSL flew 167 helicopter missions between 28March and 25 June 1979. Airborne instrumentationwas used to identify Isotopes, to determine thedirection of the plume of radioactive gasesemanating frora the crippled reactor, and tomeasure the maximum radiation levels inside theplume. During the first few days the effluentwas dominated by Xe-133m and Xe-135. On thefirst day Kr-88 and its daughter, Rb-88, werealso observed. Later the predominant radiationwas from Xe-133. The highest exposure rate, 15mR/h, was measured at 90 m altitude, 0.4 km fromthe reactor, on 30 March 1979. On the same dayan exposure rate of 4.0 mR/h was measured 5 kmfrom the reactor.

The preponderance of radioactive gases, andthe absence of radioactive partlculate matter inthese measurements, suggested that the accidentwould produce little radioactive fallout downwindof the reactor. This was corroborated by anelaborate sampling program for soil, water, andbiota, performed by many Federal and privateagencies working cooperatively after theaccident.

Confirming evidence was obtained in anaerial survey of an 82-square-kilometer areacentered on the nuclear reactor, conducted by anRSL helicopter in October 1982 (Colton, 1983).The highest exposure rates, up to a maximum of200 pR/h, were inferred from data measureddirectly over TMI facilities. This radiation wasdue to Co-58, Co-60, and Cs-137, which wasconsistent with normal plant operations. For theremainder of the survey area, the inferredradiation exposure rates varied from 6 to 14L/R/h Ground measurements, including soil samplesand pressurized ion chamber readings, were in

agreement with the corresponding aerial data.With the exception of the activity measureddirectly over TMI facilities, no evidence wasdetected of any contamination which might haveoccurred as a result of past reactor operationsor che TMI Unit 2 accident. Ground and aerialresults from 1982 agreed with those obtained in a1976 aerial survey of a considerably larger area(Fritzsche, 1977).

IX. NEUTRON LOGGING SOURCE IN VENEZUELA

In early March 1982 an oil well loggingcompany lost a strong neutron source inVenezuela. It was an americium-beryliium sourcewith a neutron yield of 4x10 n/sec; such sourcesare routinely used for down-hole research in oildrilling operations.

It was assumed, at first, that the sourceaccidentally fell from a storage container brokenopen during transportation from the field. Thetransportation route was searched by teams invehicles and on foot, using conventionalhand-held detectors. Villages were searched. Alarge reward was offered. After six weeks a teamfrom the RSL was called. They brought propor-tional counters filled with hellum-3 to 3atmospheres absolute pressure. Each of threemodules contained eight tubes, 5.1 cm in dlaneterby 1.8 n long. They were mounted aboard aVenezjelan helicopter for the aerial search.

The helicopter flew the 175-km road betweenthe operations base and the well site, as well asa 0.8-kilometer-wide area on each side of theroad and all river crossings. All cities, towns,and villages within several kilometers of theroad were flown. The logging site and alldrilling rigs In the vicinity were alBo flown.This work was done at 91 m altitude, with 300-to460-m line spacing. Seven rivers In the vicinitywere flown at 46 m altitude. CalculationsIndicate that at a survey altitude of 91 a and anaircraft velocity of 35 m/sec, the neutron sourcewould have been readily detectable at lateraldisplacements as great as 580 m. Despite thissensitivity and despite a well—planned,carefully-executed search, the source was neverfound. It haB been speculated that the sourcewas stolen and quickly removed from the area,either for profit or to embarrass its owners.

X. COBALT-60 DISPERSAL IN MEXICO

In November 1983 one of the most seriousaccidental dispersions of radioactive materialoccurred in Cludad Juarez, Chihuahua, Mexico. APicker C-3000 Co-60 teletheraphy machine wasdisassembled for scrap steel. The sourcecapsule, containing an estimated 450 Ci of Co-60,was ruptured. The capsule contained 6000 to 7000cobalt pellets, each 1 mm x 1 mm in size andhaving an activity of 40 to 100 mCi. Individualpellets were found at the disassembly site, inthe transporting vehicles, along the roads tofoundries In Juarez and Chihuahua and in thescrap piles and processing equipment. Anestimated 218 Ci of Co-60 was melted into 7000tons of steel products such as reinforcing bar,metal table bases, and electric motor parts.

When the RSL was deployed on 27 February

82

1984, nearly 3 months after the initial dispersalof the Co-60 pellets, contaminated steel productshad been distributed In northern Me;:ico, much ofthe United States and several other countries.Using a helicopter and an array of Nal detectors.Laboratory personnel conducted an aerial surveyover El Paso, Texas; Anapra, New Mexico; and thecities of Juarez and Chihuahua in the state ofChihuahua, Mexico. The total area surveyed overthese cities was approximately 520 km2 . Bothsides of the highway linking Ciudad Juarez andthe city of Chihuahua, were also surveyed, adistance of 365 km. Between March 2 and March 24a total of 700 km2 was covered; 45 Co-60anomalies were detected, due either to pellets orto contaminated steel. All of the pelletsdiscovered with hard-held instruments or in theaerial search have been recovered. Most of thecontaminated steel producLs have been returned toMexico and buried. Unfortunately, because of thenature of the source and its wide dispersion inthe environment, complete recovery is neitherpossible nor certifiable.

XI. READING PRONG SURVEY

In December 1984 the home of Stanley Watrasin Boyertown, Pennsylvania, was found to containexcessive levels of radon gas (Hoover and Mateik,1986). Shortly thereafter the RSL began anaerial survey covering 260 km2 over the ReadingProng. This is a huge geological formationunderlying Boyertown and many other cities inPennsylvania, New Jersey, and New York. Thesurvey was requested by Pennsylvania's Departmentof Environmental Resources to help I c :te regionswhere buildings might contain elevAt i levels ofradon gas, a naturally radioactive gas producedby the decay of radium contained in the ReadingProng.

The results of this survey were presented incontour maps showing regions which wyre above thenatural background radioactivity in the area. Itwas observed that elevated activity frequentlycoincided with known geological faults. It issuspected that other highactivity areas maycorrespond to previously unknown fault zones orto regions where uranium- and radiumbearing rockare unusually close to the surface. There issome evidence that faults act as conduits,allowing radon to escape to the earth's surface.The results of the survey are being used infurther research on the radon problem.

XII. SOUTR HOUSTON

On 6 May 1985 the Laboratory began a 5-weeksurvey of many contaminated sites south ofHouston, Texas. During the 1950s and 1960s theHastings Pharmaceutical Company and severalsuccessor companies manufactured technicium andradium sources for medical use, Cs-137 sourcesfor well-logging, and other radiochemicalproducts. Because of poor quality control inmanufacturing and carelass dumping of trash andproduct residues, many different sites werecontaminated. The Texas Department of HealthResources, acting on anonymous tips, foundcesium, cobalt, and radium contaminationexceeding 30 to 35 mR/h contact readings in

completely uncontrolled areas. After a thoroughreview of the available data, Health Resourcescalled upon the USDOE for assistance. The RSLflew a prid pattern over South Houston with ahelicopter at an altitude of 46 a. A 455 kmarea was covered with tight line spacing.Industrial areas in four small cities were alsosurveyed, as well as seven waste dumps. ElevenCs-137 anomalies were found, five of which haribeen previously identified. The new sitesappeared to be the result of contaminatedlandfill used at various home construction sites.Several other anomalies were observed: 3 due toRa-226, 1 to Co-60, and 3 to Ir-192.Thirty-seven other hits were just above theminimum detectable Cs-137 activity of the aerialsystem. The Texas Department of Health Resourcesis still investigating and evaluating the resultsof the aerial survey.

XIII. RESSEMER, ALABAMA

On 2 July 1985 the Laboratory was requestedto assist the Alabama Department of Public Healthin locating three Cs-137 sources whichdisappeared when a Uniroyal Tire and RubberCompany production line was scrapped. Two of thesources were 50 raCi, one was 10 mCI; they wereused as gauging sources on a tire productionassembly line in Opelika, Alabama.

Seven potential sites were identified byAlabama Public Health: steel companies, scrapyards, and a rubber company. In two days allseven were overflown by a helicopter search teamat an altitude of 30 m. Because of their smallsize, all positioning was based on visualnavigation and radar altimeter information.

The remains of the three sources were foundat the V.S, Pipe and Foundry Co. in Bessemer,Alabama. It has been theorized that the sourceswere dumped into a hopper along with tons ofsteel scrap, and smelted to produce steel pipe.The sources appear to have been volatilizedduring the smelting operation, went up the stackas particles or vapor, which was trapped andcollected in a bag house. The debris from thebag house was dumped onto an area measuring 80 nby 18 n which indicated elevated exposure rates.Gamma spectra confirmed the presence of Cs-137.None of the other six sites showed anycontamination above background levels.

XIV. SEQUOYAH FUELS UP.ANIUM HEXAFLOURIDEACCIDENT

On 4 January 1986 a tank containing morethan 12,700 kg of uranium hexaflouride explodedat the Sequoyah Fuels Corporation facility inGore, Oklahoma. The RSL responded quickly to arequest from the U.S. Nuclear RegulatoryCommission (NRC). Ground-based measurements witha high purity germanium detector were begunwithin 36 hours of the explosion and continuedfor a week. The measurements were madeprincipally in the downwind direction from thefacility and In public areas. These results weremade available to the NRC within hours ofcollection. An aerial gamma survey was conductedin an 8 km by 8 km area surrounding the facility.Radiation contour lines were overlaid on an

83

aerial photograph of the site. The radiological,maps were available on the eighth day followingthe accident.

Laboratory personnel also obtained highaltitude and intermediate altitude photographicimagery, In both the visible and infrared regionsof the spectrum. Hundreds of documentaryphotographs were also taken of the damage andclean-up operations at the facility.

XV. ROUTINE SURVEYS

The purpose of this paper was to present asummary of the radiological accidents andincidents to which the RSL has contributed itsunique capabilities. A proper perspective,however, demands wider focus. The routine workof the Laboratory, over more than a quartercentury, has made a significant contribution tothe definition of our environment. The nature ofnatural radioactivity, the extremes of its range,and the human contribution to our background arebetter understood in the United States thananywhere else In the world. Routine Laboratorywork has provided the equipment, facilities, andprograms for developing truly unique techniquesand capabilities.

Routine work for the Laboratory hasincluded:

1. Detailed radiological maps of the Nevada TestSite. Dozens of surveys, conducted with bothfixed-wing aircraft and helicopter, haveprovided contour maps of fallout patternsproduced during atmospheric testing (Boyns,1986).

2. Nuclear Regulatory Commission (NRC) facili-ties. The NRC has jurisdiction over uraniumprocessing facilities and radiological wastesites which are regularly surveyed by theLaboratory. In addition, commercial nuclesrreactors are surveyed. A background surveyis conducted before the reactor achievescritlcality. Regularly scheduled surveys aremade over each operating reactor every fewyears. To date, 144 surveys have beenconducted for the NRC.

3. Research and production laboratories. Forthe DOE Office of Nuclear Safety the Labora-tory surveys the DOE national laboratories,nuclear production facilities, other researchfacilities and sites of special interest. Atotal of 57 radiological surveys have beenconducted at such sites, not including thosespecifically discussed above.

4. FUSRAP and UHTRAP sites. For the DOE Officeof Nuclear Engineering the Laboratory hassurveyed many sites formerly used for theproduction or disposal of nuclear materials,under the Formerly Utilized Sites RemedialAction Program (FUSRAP), and mines and millsites, under the Uranium Mill TailingsRemedial Action Program (UMTRAP). A total of53 such sites have been surveyed.

5. Miscellaneous sites. The Laboratory has alsoconducted aerial surveys of seven other

sites, for the Environmental ProtectionAgency and the U.S. Geological Survey. Ateach of these sites there were specialconcerns related to radiological safety.

6. Ground surveys. The Laboratory has deployedhigh-purity germanium detectors on ground-based survey platforms for detailed analysisof the isotopic composition and concentra-tions at contaminated sites. In addition tothe work at Enewetak, mentioned above, groundsurveys have been completed at 33 additionalsites.

In summary, the Remote Sensing Laboratory isa major technical resource of the U.S. Depart-ment of Energy. Hence, it plays a key role inthe federal response to a radiological incidentor accident. It also assists the DOE in theestablishment of a Federal Radiological Monitor-ing and Assessment Center (FRMAC) (Doyle, 1986).Many sophisticated remote sensing systems aredeployed to a FRMAC, along with an advancecommunications system to link the participatinglocal, state, and federal agencies. In antici-pation of emergency deployment, Laboratorypersonnel regularly participate in radiologicaltraining exercises; they have played key rolesin 16 major exercises since May 1975.

This work was performed by EGSG/EM for theUnited States Department of Energy, Office ofNuclear Safety, under Contract Number DE-AC08-83NV10282.

REFERENCES

Beers, R.H., Z.G. Burson, T.C. Maguire, andG.R. Shipnan, 1984, "Airborne Cloud TrackingMeasurements During the Three Mile IslandNuclear Station Accident," M i d d l e t o w n ,Pennsylvania, Date of Survey: March-June1979. Report No. EGG-10282-1009. Las Vegas,NV: EGSG Energy Measurements, Inc.

Boyns, P.K., 1986. "Aerial Systems Support forthe Nevad.i Test Site Weapons Testing".Presented at the Topical Meeting onRadiological Accidents - Perspectives andEmergency Planning, American NuclearSociety, Bethesda, MD, September 15-17,1986. To be published.

Colton, D.P., 1983, "An Aerial RadiologicalSurvey of the Three Mile Island NuclearStation and Surrounding Area, Middletown,Pennsylvania, Date of Survey: October 1932,"Report No. EGG-10282-1021. Las Vegas, NV:EG&G Energy Measurements, Inc.

Davis, F.J,, and P.W. Reinhardt, 1957,"Instrumentation in Aircraft forRadiation Measurements," Kucl. Sci. Eng.,2:713-727.

Doyle, J.F., "Role of the FRMAC Following anAccident," Presented at the Topical Meetingon Radiological Accidents - Perspectives andEmergency Planning, American Nuclear Society

84

Bethesda, MD, September 15-17, 1986.published.

To be

Fritzsche, A.E., 1977, "An Aerial RadiologicalSurvey of the Three Mile Island StationNuclear Power Plant, Goldsboro Pennsylvania,Date of Survey: August 1976," Report No.EGG-1183-1710, Las Vegas, NV: EG&G, Inc.

Hoover, R.A., and D.E. Mateik, 1986, "AerialSurvey Efforts in the Search for RadonContaminated Houses in the Reading ProngArea Near Boyertown, PA," Presented at theTopical Meeting on Radiological Accidents -Perspectives and Emergency Planning,American Nuclear Society, Bethesda, MD,September 15-17, 1986. To be published.

Tipton, W.J. and R.A. Meibaum, 1981, "An AerialRadiological and Photographic Survey of

Eleven Atolls and Two Islands Within theNorthern Marshall Islands, Dates of Surveys:July-November 1978," Report No. EGG-11831758. Las Vegas, NV: EG&G Energy Measure-ments Group.

Tipton, W.J., A.E. Fritzsche, R.J. Jaffe, andA.E. Villiare, 1981, "An In SituDetermination of Am-241 on Enewetak Atoil,Date of Survey: July 1977 to December1979," Report No. EGG-1183-1778, Las Vegas,NV: EG&G Energy Measurements Group.

USAEC, 1973, "Enewetak Radiological Survey,"Report No. NVO-140, Las Vegas, NV: U.S.Atomic Energy Commission.

USDOE, 1978, "Operation Morning Light, NorthwestTerritories, Canada," Report No. NV-198.Las Vegas, NV: U.S. Department of Energy.

ANS Topical Meetiag on Radiological Accideats—Perspectives awi Energeacy PUaaiag

Status of Aerial Survey Emergency Preparedness and GroundSupport Equipment, CaHbration, and Sensitivities

Thomas S. Dahlstrom

ABSTRACT During the course of EG&G EnergyMeasurements, Inc. history in aerial sur-veillance, the scope of response has broadenedfrom routine surveys and accident response withaerial systems, to being prepared to respond toany radiological incident with aerial, groundmobile, and hand-held instrumentation.

The aerial survey system presently consistsof four MBB BO-105 helicopters outfitted withgamma pods and specialized navigation systems(MRS or URS) that allow the operator and pilotto fly well-defined survey lines.

Minimum detectable activities (MDA) forvarious isotopes range from a few tenths of amCi to 100 mCI for point sources and from 1 to200 pCi/g for volume sources.

fixed-wing aircraft, and various ground-basedvehicles. The system displays to the operatorall required radiation and system information inreal time, via a 5-inch CRT display and multipleLED readouts. All pertinent data are recorded on3M cartridge tapes for post-mission analysis onminicomputer systems.

The system employs five Z-80 micro-processors with AM9511 arithmetic processingchips to perform the data collection, dataanalysis, data display, position end steeringcalculations, and data recording which are allunder operator control. The system allows accessto the main processor bus through both serialand parallel data ports under control of theControl Processor.

The system consists of the followingsubsystems:

I. AERIAL SYSTEMS

A. Gamma Survey EquipmentA Mfcsserschmitt-Bolkow-Blohm (MBB)

BO-105 helicopter is used for low altitude gammaradiation surveys. The aircraft carries a crewof two and a lightweight version of the Radia-tion and Environmental Data Acquisition andRecorder (REDAR) system. Two pods, each contain-ing either ten 12.7 cm diameter by 5.1 cm thick,or 10.2 cm square by 40.6 cm long sodium iodide,Nal(Tl), detectors, are mounted on the sides ofthe helicopter.

The preamplifier signal from eachdetector is calibrated with a Na-22 source.Normalized outputs of each detector is combinedin a 10-way summing amplifier for each array.The outputs of each array are matched andcombined in a 2-way sunning amplifier. Finally,the signal is adjusted in the analog-to-digitalconverter (ADC) so that calibration peaks appearin pre-selected channels of the multichannelanalyzer of the REDAR.

1. The REDAR System. REDAR is amulti-microprocessor, portable data acquisitionand real-time analysis system. It has beendesigned to operate In the severe environmentsassociated with platforms such as helicopters,

1. Two independent radiation data collectionsystems

2. A general purpose dati 1/0 system3. A tape recording/playback system4. A CRT display system5. A real-time data analysis system6. A microwave ranging system with steering

calculation and display

The multichannel analyzer collects 1024channels of gamma ray spectral data (4.0 keV/channel) once every second during the surveyoperation. The 1024 channels of data are sent tothe single channel processor and are compressedinto 256 channels with partitions. Table 1summarizes the spectral data compressionperforated by REDAR.

The spectrum is divided into the threepartitions with the appropriate energy coeffi-cient to make the width of the photopeaksapproximately the same in each partition. Theresolution of Nal(Tl) crystals varies withenergy, permitting the compression of thespectral data without compromising photopeakidentification and stripping techniques. In thefirst partition (channels 0-75), the data is notcompressed permitting stripping of low energyphotopeaks, such as the 60 keV photopeak fromAm-241.

85

86

Table 1. REDAR Spectral Data Compression

Ey(keV)

0-300

304-1620

1624-4068

4072-4088

>4088~analog

cut off

Channel Input

0-75

76-405

406-1017

1018-1022

1023

1024

Energy CoefficientE{keV/channel)

4

12

36

N/A

N/A

Unused

CompressedChannel Output

0-75

76-185

186-253

254

255

256

The spectral compression techniquereduces the amount of data storage required by afactor of four.

The 256 channels of spectral data arecontinuously recorded every second. The REDARsystem has two sets of spectral memories. Eachmemory accumulates four individual spectra. Thetwo memories are operated in a flip-flop mode,every 4 seconds, for continuous data accumula-tion. While one memory is storing data, theother is being transferred to magnetic tape.

2. Helicopter Positioning. Helicopterposition is established using a Del Norte UHFranging system (URS) and an RT-220 radaraltimeter. The URS master unit, mounted in thehelicopter, interrogates two remote transponderslocated outside of the survey area. By measuringthe round trip propagation, time between themaster and the remotes, the unit calculates thedistance to each station. These distances arerecorded on magnetic tape once each second andin later computer analysis of the data thesedistances are converted to position coordinates.

Tablt 2.

Isotope

2*'Am

'3?CS

'3«Cs

60CO

238U

Minimum Detectable Activity torSeveral Selected Radioisotopes asa Function of Source Geometries

Minimum Detectability Activity"

PointSource(mCi)

3.0

0.5

0.4

0.3

90.0

SurfaceSource

OuCi/mZ)

0.5

0.08

0.06

0.04

10.0

VolumeSource

(pCi/g)"a = 10cm

16.0

1.3

1.0

0.7

200.0

'Assuming a survey altitude of 46 meters"Conversion factor to pCi/g relate to the

average value of a 5 cm deep soil sample.

Under optimum conditions, rangeaccuracy is _+2 a line of sight to 50 km. Theposition accuracy under typical surveyconditions is 5 m.

The radar altimeter similarly measuresthe time lag for the return of a pulsed signaland converts this to aircraft altitude. Foraltitudes up to 150 m, the accuracy is +0.6 m or+2Z, whichever is greater. These data are alsorecorded on magnetic tape so that any variationsin gamma signal strength caused by altitudefluctuations can be compensated.

3. Minimum Detectable Activity. Table2 indicates the minimum detectable activity forseveral Isotopes as a function of sourcegeometry for the aerial system as employed Inthe Sandia and ITRI survey of April 1981.(1)These MDA« are typical for most surveys flownwith 12.7 cm diameter by 5.1 cm thick crystals.

4. Comparison of Cylindrical andRectangular Log Detectors. It was mentioned inthe description of equipment that the helicoptercan be outfitted with either cylindrical orrectangular detectors. In our terminology <-'erefer to these as 5x2 pods or log pods,respectively.

The original thought was to Increasethe overall area and volume of the detectorsystem to increase sensitivity at the higherenergies. This would be particularly useful inthe case of looking for lost sources like 137-Csand 60-Co or other isotopes with higher ener-gies. The usefulness when looking for contami-nation like 241-Am or other isotopes with gammarays in the low energy region may not be aspronounced.

In order to evaluate these pods, theeffective areas for two sets of two pods (countsIn the photopeak divided by uncollided mono-energetic fluence at the listed energies) weredetermined and are presented in Figure 1.

It can be seen that at 137-Cs energies(662 keV), the difference can approach a factorof two. (Other curves of effective area vs.angle and energy are displayed on the poster.)

One interesting Jgct is that theoverall response of the two©/?;terns when lookingat distributed sources such as 241-Am showalmost no difference in sensitivity. It would

EFFECTIVE AREAS20, Sx2's (2 POOS)8, LOGS (2 PODS)

(SOURCE 1 TO POD DOWNWARD FACE)

I M

PHYSICAL Altf A

L M > XHicm/

Hi • MM em>

300 900E - MY

Figure 1. Effective Areas of 5x2 and Log Pods

be more useful operationally to employ the light5x2 pods because of the weight difference (62kg) which translates into fuel and allows longermission times.

B. Neutron EquipmentThe MBB BO-iO5 helicopter can also be

outfitted with a neucron pod consisting of fourmodules of eight Helivm-3 proportional counterseach for a total of thirty-two tubes. Each tubeis 5.1 cm diameter by 1.83 n long and is filledto 2.99 Ata pressure. Each eight tube arrayconsists of 0.64 en thick polyethylene tubeswhich are welded to fora a rigid system holdingthe Heliun-3 tubes.

The four modules are enclosed in analuninua pod mounted between the skids on thehelicopter. Each module contains its ownpre-amplifier and high voltage supply andrequires only _+12 VDC available f ro« the 8EDAR.

The principle use of this system wouldbe the recovery of lost neutron sources; e.g.,well logging sources commonly used by cilcompanies.

II. IM SITU GAMMA SURVEY SYSTEM

A. System ComponentsEGSG/EM has successfully fielded an in

situ gamma analysis system employing high puritygersanium gamma detectors. This system consis-ted of the following major components:

1. A vehicle (two types were employed, a four-wheel drive vehicle and a tracked vehicle.)

2. A HPGe detector and collinator shield.3. A telescopic pneumatic mast capable of loca-

ting the detector from 0 to 7.4 meters aboveground level (AGL).

4. A 4096 channel pulse height analyzer.5. A computer, printer, and data storage device.6. A microwave ranging system (MRS).7. A 4 leW generator.

1. High Purity Germanium Detector andCollinator Shield. A HPGe detector was need tomeasure the gamma rays from the radlonuclldesdispersed in the surface and near-surface soil.The detector was capable of high energy resolu-tion, typically 1 to 2 keV Full-Width-Half-Maximum (FWHM) of detected photopeaks. Thishigh resolution enhanced the ability to identifyphotopeaks and quantify their emanatingisotopes.

A lead and cadmium collinator shieldwas used to limit the detector field-of-view.The shield forced the detector to view arestricted solid angle and circular area on theground. The physical angle of the cone was 50"fron the vertical; however, the cut-off angle atwhich gammas could not enter the crystal(detector) was approximately 60° for gammas withenergies between 0 and 300 keV.

The detector was calibrated usinglaboratory point source angular responsemeasurements folded Into a sensitivity compu-tation. Several HPGe detectors were used withthe in situ system. Each detector of the Basedesign had a similar angular response with theshield mounted in place. There were, however,some variations in the absolute efficiencybetween the detectors. Conversion factors weregenerated based on the angular response of onedetector and were applied to the other detectorswith a relative response factor folded in. Therelative response factor was used to correct fordifferences in the absolute efficiency of eachdetector. Detector efficiencies were checkedperiodically and relative response factors wereadjusted as necessary.

2. Minimum Detectable Activity (MDA).The aininum detectable soil activity for the insitu gamma system was dependent on the signalcount r'.r.e and the background count rate. Thein situ system used a photopeak (signal) energywindow and two background windows on either sideof the signal window such that the total widthof the background windows was equal to the widthof the signal window. The MDA for a bare N-typegermarium detector is shown in Figure 2.

The software in the computer of the insitu system evaluated specific photopeaks inreal time from the energy spectrum obtained fromeach measurement. The results were printed outimmediately for interpretation by the operator.These results included the isotope name, concen-tration in the soil, and exposure rate. Table 3lists the routinely monitored radionuclldes andthe associated gamma ray energies.

More detail of the in situ system aaybe found in References 3 and 4.

III. OTHER SYSTEMS

A. Multi Purpose Modular SystensEG&G/EM is developing other systems

that can be used in a variety of manners in

88

Table 3. Radionuclides and Associated GammaRay Energies Routinely Monitored bythe Sn Situ System

Source Energy (keV)

"»Rh

106Ru

'«Sb

'"Ba

'"Cs

'•17Cs

' " E u

S36U

" ' A m

1460.8

834.8

1173.21332.4

127.2197.9326

475.1697.5766.8

1046.6

621.8

427.9463.4600.6

81302.7355.9

604.6795.8801.8

661.6

39.5121.8244.7344.2411.1444778.9964

1408

591.8723.3873.2

1004.81274.5

6086.5

105.3

76.51241.8

238.6583.1911.1

186.1351.9609.3

1120.4

129.3

59.5

89

MINIMUM DETECTABLE ACTIVITYFOR A BARE 20% N-TYPE GERMANIUM @ 1 METER

PAMHfTElO:MANCHWO l.t §•!*<«•SO*L OCNBITV 1.1 | / c n lAM DCMS4TV «.M12t3 fl/cm3• O H HOIITUKE 10SALPHA 10.0 cm-1Of PTH - J« en)ACOUMI T I M PN ( K o f i *

0.04S

ENEROV (k«V)

Figure 2. Minimum Detectable Activity (MDA)

non-dedicated platforms for aerial survey orlooking for lost sources. These systems consistof discrete modules housing the electronics anddetectors.

One example of such a system makes useof an electronics module containing batteriesand associated power supplies to provide powerto the detector nodules. The detector nodulesare fiber jjlass containers consisting of thefollowing:

1. Gamma nodule containing two "log" detectors2. Neutron module containing eight Heliuo-3

tubes, 5.1 cm In diameter by 91.4 cm long inthe same polyethylene configuration as de-scribed in I.B.

All the real time data reduction andsygtea control Is done In the electronics unitutilizing a NSC-800 microprocessor.

With the Inherent portability of thesesystems they can be employed in any helicopterwith sufficient space or any vehicle (e.g.station w?.gons or vans). One can then tap 12 VDCpower from the platform or use internal batter-ies to power the system. Data output is in the

form of audio or analog to a strip chart.The flexibility of these systems lends

itself greatly to enhanced emergency response.This work was performed >J EG&G/EM for

the United States Department of Energy, Officeof Nuclear Safety, under Contract NumberDE-AC08-83NV10282.

REFERENCES

Boyns, P.K., November 1981, "An AerialRadiological Survey of the USDOE'a SandlaNational Laboratory and Inhalation ToxicologyResearch Institute.

Soberaan> R.K., Phys. Rev. 102, 1399 (1956).

Tipton, W.J. et al. November 1981. "An In SituDetermination of Am-241 on Enewetak Atoll (Dateof Survey: July 1977 to December 1979)." ReportNo. EGG-1183-1778. Las Vegas, NV: EG&C/EM.

Relaan, R.T. 1983. "In Situ Gamma AnalysisSupport for Phase I Middlesex Cleanup Project(Date of Project: July-November 1980). ReportNo. EGG- 10282-1003. Las Vegas, NV: EG&G/EM.

ANS Topical Meeting on Radiological Accidents—Perspectives and Emergency Plaaauf

Aerial Systems Support for Nevada Test Site Weapons TestingPhilip K. Boyns

ABSTRACT EG&G Energy Measurements, Inc. opera-tes two aircraft for the Department of Energy insupport of the Nevada Test Site (NTS) ac t iv i -t i e s . A King Air B-200 and a Turbo Beechaircraft are used to perform wind measurements,cloud sampling and cloud tracking operations insupport of each t e s t .

I . MISSION

Hind sounding measurements are made with aDoppler radar or a Loran C low frequency navi-gation system. The aircraft measure winds fromapproximately 1,000 feet Above Ground Level(AGL) to 12,000 feet Mean Sea Level (MSL) atmultiple locations. Meteorological data aresupplied to the Test Controller at the NTSbefore each test.

In case of a venting, the aircraft cancollect whole gas and filter saaples and providethe Test Controller with visual observation ofcloud development and isotopic identification ofthe consistency of the cloud. The radiologicand navigation equipment allow long-rangetracking of airborne radioactivity. The dataacquisition system can provide longitude,latitude, altitude, direction, and isotopiccontent of the cloud on a one second basis.These two aircraft provide both pre- andpost-event radiological safety information fornuclear testing at the Nevada Test Site. Theaircraft and measuring systems are maintained ina state of readiness to meet routine andaccident measurement requirements.

The King Air B-200 (Figure 1) and TurboBeech (Figure 2) have the same equipment and canperform either cloud sampling or trackingmissions.

II. THEORETICAL SENSITIVITY CALCULATIONSFOR TERRESTRIAL AND POINT SOURCES

A computer has been programmed in FORTRANto compute sensitivity of an airborne detector

Figure 1. King Air B-200 Aircraft

Figure 2. Turbo Beech Aircraft

system. The sensitivity of the detector systemis computed for point, surface (<*»»), exponen-tially and uniformly distributed (<*- 0) sources.Conversion factors are also computed for concen-trations of various depth soil samples versusdistributions of the isotopes in the soil. Pre-selected isotopic distributions in the soil are0.10, 0.33, 0.50, 1.0, 3.0, 5.0, 7.C and 10.0cm. A specific distribution can be entered, ifnecessary. The conversion factors are given inunits of gammas/cm2 -sec, uCi/m2 and pCi/g forsoil samples up to a specified depth arecomputed in 1.0 cm increments. Model andequations used is shown in Figures 3 and 4.

All conversion factors are calculated foran lsotropic, cosine (angle 6 ) and an averagebetween isotropic and cosine detecor responsefunctions for all source distributions. A

91

92

MODEL

h

z

PCItJf

K\

\

SOil

COS I) =

\

\

GPOrr

4

ef'v

let

h

COS H

cos »

niiai voof So."

n depth ol source >n Ihe soil lem ')

Figure 3. Model Used for Theoretical Calcula-tions for Terrestrial Source Detect-ability.

special response function may be entered for thedetector system. The response may be dividedinto 20 increments, of any width for angle from0 to 90°.

The point source conversion factors arecalculated for isotropic and cosine responsefunctions with lateral displacements of thesource up to two times the altitude. The signalFull-Width-Half-Maximum of the signal is calcu-lated in both time and distance for each lateraldisplacement for a specific aircraft velocity.(Figures 5 and 6).

Point SourceEffective Bottom Detector Surface Area

Source Location

A = elfective surface area of the detector tyslem

Figure 5. Bottom Surface Area of the DetectorSystem Equation

° Jo Jo Jo 4"[|h + z) sec »]> * *

Point SourceEffective Side Detector Area

Volumedv ** idid^dz• •= (h + j ) tan t

da = (h + z) sec' 9

2ir a da 2tr (h + 1) tin g (h + z) sec'r} at tan t>Xf°°e-°z e"2""'*- to-

Jo

Jo dZ " « * » o

H" 2 sec»/A.)

Figure 4. Equations Used to Compute SystemSensitivity

Detector Path

Source Location

Figure 6. Side Area of the Detector SystemEquation

93

Various parameters can be plotted uponrequest: flux percentage vs. angle (for finitesource conversion factors), flux per unitangle, flux per unit area versus angle and thepCi/g versus depth of soil samples.

Conversion factors are shown in Figure 7for terrestrial sources of Co-60 and Cs-137.Table 1 is the minimum detectable activitiesfor the same isotopes at 91m (300 feet) AGL,normally flown by the aircraft.

AIRBORNE SOURCES

. i.-

Jf *wn* leriK*

CwmrH

• I c»t

1

• * fK»f*

I

i

nl *••# Otf* IS ft*#e

;_'••. -I-.'"

i ' " i * ' • •

«na**m

1 . . , •....

! ' i' <

Figure 7. Conversion Factors for Terrestrialand Point Sources

Table 1. Minima Detectable Activities

• r U N I M OT WVfvt «M#fHVWIW

Uolop*

'•"Co

'Cs

Surfsc* Source*

PointSource(mCl)

126

24 3

Dlttribul*4Source

(//Ci/mi)o = 00 34

0 77

Volum*Sourc*

CD'o = 10 cm

43

11 4

' Assuming i survey altitude of 91 meters

Conversion factor to pCi/g relate* to the average

value of a 5-cm deep soil simple

III. THEORETICAL SENSITIVITY CALCULATIONSFOR AIRBORNE SOURCES

Programs were developed t o c a l c u l a t e thesensitivity of the detector system to radio-activity uniformly distributed in a cloud. Thecalculations assumed the clouds were atleast3 ha in radius for each gamma energy (Figure 8).

Minimum detectable activities and otherpertinent data are in Table 2 for airborneclouds of radioactivity for selected isotopes.

IV. EQUIPMENT ON EACH AIRCRAFT

REDAR is a multichannel analyzer thatcollects 1000 channels of ganaa spectral datafrom 40 keV to 3 MeV. The spectral data are

AIRCRAFTDETECTORS

AREA A

THE COUNT RATE (CR) AT THE DETECTOR AREA A FROMAN INFINITE CLOUO (r>3A.) OF UNIFORMLY DISTRIBUTEDACTIVITY (Co):

CR = 2AC0 r*C0S9 dOdrdp

CR = AC / / e"-co»8drdeJo JoJo Jo

f"/' 0

fCR = AC0 / e"'dr

CR = AC 0 A.

Figure 8. Model Used for Airborne Sources

Table 2. Minimum Detectable Activities forAirborne Radioisotopes

Minimum Delectable Activity (MDA) for Airborne Radioisotopei

Two 4X4X1* Inch LofltNil (T/>

Surface An* 2*00 cm'1.10 •//

" * •

lit*

' ' « ' ' 0 514

i-ci ' ^ HO

. ; «

I(MMMM

096

0 ? '

Cfe>

• ]

A

P

OVi

DM

.„

0 14

<>Yi

!

i-

[.

A * , j

•Mi j

. . . . . .

V'.'"':

•*v/»

, « ,

!j

A. '

. 0 » ,

1

"'

• " " "

v '?

B U

••- > '

-; <•

] ftcl* \ MDA

' ' " " • ' ' . " ' " •

; , , . ; , , ; • < • *

compressed into 256 channels and written on tapswith various other parameters, includingaircraft altitude above ground level, positiondata (processed by the L0RAN unit into latitudeand longitude coordinates), discrete datalabels, and outside air pressure. The system

94

also provides the capability for realtimeonboard spectral analysis. All of the data arewritten on tape once every second to provide arecord of the mission for extensive post-missionanalysis.

The gamma detector package consists of two4 x 4 x 16 in. Nal(Tl) log detectors and one 1in. diameter x 2 in. thick Nal(Tl) detector.The detector sensitivities depend on the energyof the penetrating gamma rays and the branchingratio of the parent nuclide; hence, onlyapproximate sensitivities can be given:

Airborne - 2 pCi/liter (5 jjR/h)Ground deposition - 1.0 uCi/m2 (10 /jR/h)Exposure rate ranges -

Logs: S to 100 fiR/h1 x 2 : 50 (iR/h to 6.0 mR/h

HPI tissue equivalent ion chamber data arerecorded by the REDAR system every second. Doserate is recorded from 1.5 mrad/h to 10 rad/h.

A 660 Victoreen Integrating Exposure Meterprovides a total integrated exposure to amaximum reading of 100 R.

A XETEX 30SB Digital Exposure Rate Meterprovides an auto-ranging display in units ofmR/h or R/h.

Two strip chart recorders are used to plotvarious parameters recorded by the REDAR system,usually the radar altitude and the gross countrates from the log detectors and the 1 x 2 in.crystal.

Portable survey instruments:

a. Eberline E-500Bb. Eberline R0-2Ac. Baird Atomic

High volume charcoal and IPC-1478 paper filters-Paper filters:

Charcoal filter:

Flow Rate:

four each 10.16 cm(4 in.) in diameter22.86 cm (9 in.) x26.64 cm (10.5 in.) x3.175 cm (1.25 in.)700 1/min (25 ft.'/min)

The whole gas sample system includes a 5.08(2 in.) IPC paper filter with a variable flowrate up to 700 1/min (25 ft.7min). The air fil-ter can be removed and analyzed with a 2 x 2 in.Nal crystal that is shielded with 0.5 in. oflead. The spectral data are displayed andrecorded by the REDAR system. Whole gas samplesare compressed into bottles for post missionanalysis.

LORAN C navigation system provideslongitude and latitude for real time displaysand are recorded on magnetic tape every second(+1500 feet).

Auxiliary equipment outputs are alsorecorded every second:

a. REDAR Altitude (0-2000 ft. AGL)b. Absolute Air Pressurec. Outside Air Temperature (OAT)

This work was performed by EG&G/EM for theUnited States Department of Energy, Office ofNuclear Safety, under Contract Number DE-AC08-83NV10282.

ANS Topical Meetiag oa Radiotogicfel AccMeato—Perspectives aad Emergency Plaaaiag

Mobile Robot Response to Actions Associatedwith the Release of Hazardous Materials

Harvey B. Meieran

ARSTNACT: This paper presents a rationaland composite summary of tasks and mis-sions that could be assigned to mobilerobots and other teleoperated devices inresponse to accidental releases of ra-dionct.ive and other hazardous/toxic ma-terials to the environment. This paperwill also discuss specif o missions thathave been, or could be,mobile robots operating atCh«rnobyl--4 nuclear power plant sites.Other items and issues that will also beconsidered are: availability of appli-cable mobile robot units, expendabi 11i-ty/durabiIity/decontamination, j>or tab i -I ity/maneuverab il i ty/raob i I i ty, communi-cation techn i']ues, power supply, andartificial inlet I 1 igcnce/aulonomous navi-gation interactions.

assigned tothe TMI--2 and

r . INTHODUCT1ON

Many man-made and natural disaster-ous accidents precipitate loss-of-lifeand injuries among persons residing inthe vicinity of the incident. Further-more, releases of contaminants from theincident can also create a long-term(i.e., more than a few days) harsh envi-ronment for those who survived the ini-tial effects from the disaster as wellas to members of the emergency responseteams who are trying to mitigate theconsequences of the disaster.

The monitoring, rescue, relief, andcleanup actions associated with the ac-cidental release of radioactive, chemi-cal, biological, and other toxic materi-als has been historically handled byemergency response team personnel.These incidents also include fires,security, . and civil disorder sitations.

liven though individual emergencyrespose team members can be appropriate-ly attired to protect them from theeffects of the hazardous material conta-mimin's, they remain to some degreesusceptible to exposure of these mate-rials. Furthermore, the efficiency oftheir motions and general activities inthe affected areas can becoae compro-mised and somewhat restricted. On theother hand, the transference of some ofthese manipulative and monitoring func-tions to currently available mobilerobots and other teleoperaled vehiclescan eliminate the probability of expo-sures by the relief worker to the poten-tially harmful health hazards.

Activities associated with thedevelopment and deployment of mobilerobots and/or teleoperated devices whichmay conduct repair, maintenance, decon-tamination, decommisisioning, and sur-veillance/inspection missions in nuclearpower plants and other nuclear industryfacilities are being pursued by a spec-trum of domestic and international orga-nizations. These activities are beingcomplemented with parallel efforts forrobots in other "hazardous industries",such as explosive ordnance disposal (orbomb disposal, EOD) fire-fighting, andmining.

This paper will discuss the appli-cability and technologies of currentlyavailable vehicles which can assist ii>responding to incidences associated withthe accidental release of hazardousMaterials. These vehicles have beenmanufactured/assembled in nine differentcountries, as noted by Gelhaus andMcieran (1985), and represent a spectrumof locomotion techniques, sophisticationof design and operation characteristics,and degree of autonomy and onboard/off-board intelligence. They have been

95

96

o p e r a t i n g i n v a r i o u s " h a z a r d o u s i n d i i Kt r i e ! : : " , s u c h as n u c l e a r , in i I i I n r y a n dc i v i I i a n HOP , !. ox i o ma t e r i ;i I h a n d I j ri}i ig e n e r a l s e c u r i t y , c i v i l i a n c i v i l d i s o r -d e r s , f i r e - f i g h t i n g , am) m i n i n g . T i mcommon; ! I I j t y o I' c o n f i g u r a t i o n s , ;11<|> I i < ;il i o n s , a n d m i s s i o n a s s i g n m e n t s l o r t h e s ev e h i c l e . ' ; e n a b l e s I hc-ni t o he e m p l o y e d i nh a z a r d o u s s i 1 ua I i f i s u l . l i « r t. l i a n I l i n s cf o r w h i c h t h e y w e r e o r i g i n a l l y d»'S i g n o d .F o r e x a m p l e , a r o b o t , t h a i was r les i g r i ndf o r c i v i l i a n !i()l> a c t i v i t y h a s IJr-«-ri <'inp l o y e d i n t h e n u c l e a r i m l i m l . r y . C o i rv c r s c l y , a r ^ c n n /' i K u r c d mob i I f r o b o tt h a t was i n i t i a l l y d e s i g n e d f o r I l ien u c l e a r i n d u s t r y w i I I be erap I o y e d i n 1 li f-c i v i l i a n MOD " i n d i i s t r y " a i i ' l i an be u s e dI n j n o n i t o r t h e ( o n r i ' i i l n i l i o n s o f r e I e n sed ( t o t h e a t m o s p h e r e ; t o x i c <: hem i <a ] s .

II. MISSION ASKICNMKNTS

A. General Consi'lnrnl ions.Thoi c a r c m a n y c o m m o n o h a r a c

terist ics assnriiiti'd w i t h <i i s a s t I T Sw h i c h h a v e o e e u r e d in s e v e r a l nunr e l a t e d i n d u s t r i e s m i d B i t ua( ions ; l h " s ei n d u s t r i e s a r e , for e x a m p l e , nucleiirpowt't'i c h e m i c a l p l a n t s , a c c i d e n t s o Isystem?; t r a n s p o r t ing h a z a r d o u s m a t e rifils, and v o l o a n i i erupt i o n s . T h eb o t t o m line for t h e s e a c c i d e n t ? ; is the|irosi;iicc of a l i n g e r i n g half h e n v i r o n •merit c r e a t e d by the a c c i d e n t w h i c h c o u l dbo d e t r i m e n t a l to the survivor's o f I. liea c c i d e n t as w e I I as to m e m b e r s o f thee m e r g e n c y r e s p o n s e t e a m .

Tin: c o m m o n f e a t u r e s for t h e s e accid e n t s in w h i c h t h o r n a r e run I o r I lire a-I cried i n i I i a I o r r e s i d u a l d e a t h s a n d / o ri n j u r i e s c a n he l i s t e d o h r o n o l officii 1 lya c c o r d i n g to the f o l l o w i n g ten cale|;or i o s : a ) i n i t i a l h u m a n d e a t h s ; b ) i ri iH a l h u m a n i n j u r i e s ; <:) i n i t i a l l i v o -s toe k d e a t hs ; d ) e r c a t i o n of a ha•/.ard•ous/'uirsh e n v i i o n m o n t for the s u r v i v o r sa n d m e m b e r s o f t h e e m e r g e n c y r e s p o n s e /r e s c u e / r e l i e f t e a m s ; ei a t t e m p t to o l in i n a t e t h e m a i n s o u r c e of the c a u s e ofth e i n c i d e n t ( p r o b l e m ) ; I') e v a c u a t i o n ol'the s u r v i v o r s f r o m t he s c e n e / I o c a Iity ofth e i n c i d e n t ; R ) ret;l n n i l i o n o f thes c e n e o f t h e a c c i d e n t (if possible.) andd e a d Jiver.toch b u r i a l ; h) d e l a y e d d e a l In;a n d / o r i n j u r i e s to t.!ie s u r v i v o r s ; i)d e l a y e d l i v e s t o c k d e a t h s , i n t e n t i o n n Is l a u g h t e r , a n d b u r i a l ; a n d j ) r e s i d u a ] ,lorif! t e r m dcrcont ami na t i on o f the loiale n v i r o n m e n t ( p r i m a r i l y r a d i o a c t i v e p a r -tides*, f r o m n u c l e a r p o w e r a c c i d e n t s ) .

It r a n he s e e n in T a b l e I thaithere. a r e ninny c o m m o n c h a r a c t er i a I i <;;s h a r e d by s e v e r a l wflJ k n o w n anil e x e m p1 i f i e d a c c i d e n t s w h i c h h a v e recent Ivo c c u r r e d . T h e most c o m m o n and |re<ni"rtl

c l i i i n i i l i T i i i l l e a a r e ; c r e a t i o n o 1 a n e nv i r o m n c i i I I 11 a I c a n b e d e I i i m e n t a I I D t h ere. 1: c n e M o r h o i ; '. a h e a l t h h a '/• a r<l t ; c I i ju ia a t. e t h e m a i n s o u r c e o f (lie c a u s e o f 1 li"i n c i d e n t ; .ir>d e v a o i a l i ;>n o f p e r s o n i K - Ia n d : ; u r v i V D I :; o f t h e a c c i d e n t f r o m I h e;i t'I't<f t <-d a r e a / m o s t m a n d a i i t o i v a n i l o u ev o l u n t a r y ) , a n d t h e l o n g l e r m c o n l a m i n at i o n l e f t a s a re, r; i d u e 1 r run I lie a* c i d e n 1

H. 'I'M I 2 AcciJ n I he ens

i len I , wh i r l i o e e u r r eI he con I am i nat i o na I t r i l»ii I ed t o I he r aI r • f I w i t h i n I l ie :; t i l lco iH a i nmen I l>ii i I d i nI ow i| ii a n I i t i '••: o t Imi r y, i de en v i ronnie n If i r a n t I e ve I ;; w i t liace i den t . IIj> I o fit)t i I y r-va'°uti I ed t her ear-1 o r s i 1. e 1 o r a I< u r r o n t I y l i v e o p e r ain i ng a v a r i <;t y o I" l abeen ret i red from uro l i o l f ; were i n s t a l lI o lUH'A, or more t hacc i den t o c c u r r e d .

C.

dent .

e o f I lie TMI 2 a c c i

it o n M a r c h '/.H, \'.iT.i,w a s ' a n d s I i I I i <; )d i m i l ' l i v e p a r t i l l or,el III e o f 1 III r o n e I oiV, • T h e r e l a t i v e ! v

I :11 i c | e a s e r l I o I h ed e c a y *'rl t. O i n s i i\ n i

i n d a y s a f t e r I In, 000 poop I p. vo I unt a

areas around I bl-ew days. There aret ing robots per for.s k:; ; two robots haveso. None o( I hoseed i nlo 'I'M I 2 pr i oran <\ vears a ft er I he

Oher'ioby I - " Acci drill .In the case of the Chernobyl 1

net: i dent, which occurred on April 2\>,HtHG, the initial two deaths were causedby the fire in and doslruol ion of I hereactor building. Several individuals•lied from accute radiation exposurewithin days after the H O C ident and thoprognosis for others who have beenhospitalized if, somewhat pessimistic.The environs surrounding the reactorsite were heavily contaminated andforced the mandaiitory ovnciial ion of morethan 135,000 persons from a 2iV.H) KIJ km< 1000 si| wile) area around Chernobyl.II was also necessary to destroy andbury ii cons i rlerab J e amount of con I aminat.ed livestock. The area continues Inremain contaminated find local residentsari> not expected to be able to return totheir residence for periods! of up tofour years.

The njmrni orfj of the Chernobylplant have purchased (or leased) atleant throe mohile robots from Germanyt.h~ two-tracked MF2 and four-tracked Mh'3robots and a 33 Mlonne remote control ledbulldozer. These have been operating alI ho Chernobyl site w i I Ii i n one monthafter the accident occurred.

C . l.ako N i os , Canierooris •The chemicals released on 1 he

night of August XI, l'.)HH from the bed ofthe volcanic Lake Nio.s, Cameroons,instantly kilted more than IfiOO res i

97

dents and left more than 500 injured.Survivors were later eventually evacua-ted. Authorities initially expressedsome concern over the lingering chemicalcontaminations which could cause somehealth problems to the survivors and tomembers of the rescue teams. There is astronger and more realistic concernregarding the health hazards being gen-erated by the decay of thousands of headof cattle that were instantly killed bythe initial release of chemicals fromthe volcanic lake.

D. Other Incidents/Accidents.A large fraction of transpor-

tiit ion accidents reported almost dailyto FEMA involve trucks and/or trainscarrying hazardous nnd toxic: chemicals.When there is a release of these mate-rials, which are frequently accompaniedby fires and explosions, the local resi-dents art; ordered to evacuate the area.Some recent accidents have forced theevacuation of thousands of people fromtheir residences; they have also beenforced to stay away from the residencesfor periods in excess of several days.The accident at the Bohpal, India UnionCarbide plant killed more than 2000people and injured another 200,000, manyof whom were evacuated. These threeaccidents have also been included inTable I as examples of incidents thathave similar characteristics to thoseassociated with nuclear power plants.

E. Locations for Mobile RoboticVehicle Act i vity.The locations where mobile

robots could be employed in and aroundthose accident sites are: a) indoors(possibly more than one floor or eleva-tion); b) within the physical site boun-dary for the facility having the acci-dent; and 3) in and around an exclusionarea (external to the physical siteboundaries) surrounding the affectedzone, to include structured and unstruc-tured terrains.

F. Specific Mission Assignments.It can be seen from the dis-

cussion in Section 11.A. above and fromTable 1 that there are several commoncategories for which mobile robots couldhave been used to mitigate the conce-quences of accidents. First of all, itshould be noted that the use of a mobilerobot(s) in the specific accidents lis-ted in Table 1 is a speculated considei—Htion. Robots may not necessarily beavailable to the rescue team at the timethey would be most needed. Secondly,suitable mobile robots may not haveexisted at the time an accident, as in

the case for TMI-2.The primary category which will

dictate the need for a mobile robot isd), the development of a harsh environ-ment which can become a hazard to thehealth of the rescue/relief worker. Ifthere is no hazard, then theie is littleincentive to employ a mobile robot andlittle chance that the vehicle couldconduct a mission/task that would bemore efficient or cost effective thanthat which could be conducted by a humanworker. For this reason it is not anti-cipated that the utilization of mobilerobots could offer a reasonable alterna-tive solution to the employment humanrescue workers at the location of anearthquake where hazardous environments,outside of collapsing buildings, wouldnot be expected.

Specific suggested missions asso-ciated with each relevent category arelisted in 'fable 2. Some of the specifictasks associated with the various acti-vities are amplified below: security -crowd control (during evacuations) andpatrolling deserted streets and environsfrom which the local population has beenevacuated. Core samples are beingdrilled and taken from the concretewalls of the TMI-2 reactor; thesesamples will be analyzed for fissionproduct penetration which in turn willassist planning personnel to develop anefficient strategy for decontaminatingthe reactor's facilities. The carcasesfrom dead and destroyed animals shouldbe buried in an excavated trench as soonas possible (to limit the onset ofhealth hazards among the survivors).The specific activities associated withother missions listed on Table 2 aredescribed by the title of the mission.

III. DESIGN/PEHFOHMANCB CONSIDERATIONS

A. Locomotion.

1. Mobility/portability.Specific mobility aspects that wouldhave to be considered are levels ofoperation in a building (more than onefloor), complexity of infrastructurewithin a building (stairs, location ofpipes and other obstacles, width ofpassageways and doorways, height ofwalls and ceilings in buildings), typeof structuring in an outdoor terrain(texture of surface, placement andheight of obstacles, distances to tra-vel) .

The robot should be easilytransported to the scene of the accidentin a small van or a small truck (lessthan I ton) It should also be able to

98

travel under its own power to the find Idestination. Delays in committing itsutilization may cause the res cur: workersto be unnecnssaril ly exposed to bazardons mater i al s .

As it is probable Hint, under somecircumstances, a vehicle may not havethe ability to travel to an iniKcessalilflocat Lon under its own power-, it may benecessary to manually carry that vehicleto the work station. The overall weightof a vehicle- should then be restrictedto a level that will enable i) to becarried by no more than four persons upa flight of stairs.

2. Maneuverability. Themaneuverability of the vehicle shouldconsider whelher or not the robot willbe operating in a confined indoor,multiple level indoor, sl.ruclurod outdoor terrain (parking lots, roads) orunstructured outdoor terrain (crosscountry, around destroyed facilities).

3. Locomotion techniques.Most of the current generation of mobilerobots possess either wheeled or track-type locomotion techniques. The thirdbasic technique, legged, has been rele-gated to the role of a research item atthis time. The criteria for the seleetion of the specific form of locomotionis based upon the requirement for mohility and specific placement of the robotiu an outdoor or indoor environment.

B. Cost/Benefit Analyses.Under normal and non- emergency

circumstances, the robot operator mayhave some reluctance to assign the robotto a mission where there is a finitechance that the vehicle could becomelost. Under emergency situations, how-ever, the philosophy of operationdemands that the consequences of thc-hazardous situation be compromised assoon as possible so that there is aminimum loss of life or threat upon thehealth of the community, and finally tominimize damage to the facilities. Thooperator thus has the freedom to consi-der the robot more expendable underemergency situations than under normalsituations.

One of the key issues regarding anassured and durable life for the robotis its ability to stay relatively'clean". A device will soon lose itsability to conduct, missions if it is notpossible to restore the robot to its"mint" condition. The levels of conta-mination and durability will be dependendenl upon the design features whichwill: a) minimize the number and expo-sure of as- f(iliri('!il.i"fl erevinis: b> nsi'%

suitably hardened components that wil Inot rapidly degrade under high levels ofradiation; and ci use materials of con-struction that will not del I T i oraleunder high levels of radiation nor incorrosive atmospheres; and d) the abiIit.y of the vehicle to withstand washdowndecon turn i nn I i on arlio/j.s lor either r;idiological or chemical i.on t am i nant agents.

It is desirable that there will b<-a high degree of assurance that therobot will cont inue to funct ion uni literrupliid during its missions in an emer-gency situation. That is, the vehicleshould have a high availability factor.This assurance can be gained if therobot designer utilizes prudent designand planned maintenance programs (Meir-m n , ISM). Spare parts and prudentoperating scenarios: wi I 1 be able to extend the; operating life of the vehicles.

C. Communication Teci.n i ques .There are two basic modes for

communication between the vehicleoperator and the robot : untethered andtethered. In the former case thecommunication is usually by radio (RF,UHF, microwave) or by infrared. Otherwireless methods have been consideredbut. are not. common. The tethered modeis based upon coaxial cable, fiber-opticcable, or twisted pair wires. The for-mer technique permits the robot to behighly mobile and not restricted totravelling distances limited by thelength of the cable, along as the powersupply package is located on-hoard therobot. Distances of noinunication canexceed 2 km (1.2 m i l e s ) .

The limitations in communicationfor the teterless robot can be attri-buted to HK interference and liue-of-sight obstacles.

I). Power Supply.There are two modes of power

supply: on board or off board systems.In the former case the vehicle can betetherless (i.e., no umbilical cord fromthe control station or a stationarypower supply system), other 1h«n thaiwhich may be required for communicationbetween the operator and the vehicle.In the latter case a tether will berequired from either a portable powersupply system (usually a gasolineengine/generator) or from hour, e powi-r(plug in to the nearest available recep-tacle). The untethered vehicles will ofcourse be able to have the higher levi-lsof in obi] ity and the greater drurens offreedom. The travel distance for tb*3

tethered vehicles will bo limited by I lielength of the umbilical cord and I heform of the paths for travel with obr-.ta

99

cles around which the cord could notnave.

The untethered robots are generally-powered by batteries (or gasolineengines in a few cases) and the tetheredrobots by house power (110/220/440 VAC).The limitations for use of the batterypowered vehicles will be dictated by thetime that the robot can operate betweenbattery recharging periods (which usual-ly lasts between I and 5 hours). Trick-le chargers can extend the period ofuninterrupted operation by only a relatively short time. On the other hand,the period of uninterrupted operationfor an AC-powered robot is unlimited.

E. Artificial Intel licence/Auto-nomous Navigation.No currently available off-

the-shelf mobile robot possesses anydegree of practical on-board or off-board intelligence which will enable thevehicle to independently transport it-self to a workstation (autonomous navi-gation) and then conduct its missionwithout having further human instruc-tions. These degrees of sophisticationand intelligence are being pursued byseveral current government/non-govern-ment sponsored projects.

F. Status of Current Technology.In a recently completed survey

(Gelhaus and Meieran, 1985), it wasnoted that there are 69 separate modelsof mobile robots that are used to con-duct a variety of surveillance, inspec-tion, and manipulative missions in aspectrum of hazardous environments. Itwas also noted by Meieran (1986a and1986b) that most of these vehicles thatwere still functioning could be used toconduct missions in and around nuclearfacilities during normal circumstancesand to a lessor extent they can be uti-lized as assistive devices during radio-logical emergencies.

IV. SUMMAHY AND CONCLUSIONS

The capabilities and advantages ofremote controlled (teleoperated)/roboticmobile vehicles are now coming to theattention of those individuals who arecharged with the responsibility to di-rect efforts to mitigate the conse-quences of radiological and other hazar-dous incidents. The direct employment ofthese technologies can remove the emer-gency response team member from a harshenvironment created by the emergency andthereby eliminate the potential ofhaving the individual exposed to thehazardous materials in that environment.

The optimum features that should bepossessed by a currently available, off-the-shelf mobile robot that is directedto respond to radiological and otherhazardous environment emergencies areincluded in the following items: rela-tively light weight - less than 175 kg(385 lbs) for a vehicle that must workindoors (as well as outdoors) and lessthan 700 kg (1500 lbs) for a vehiclethat is committed to an outdoor terrainhaving some obstacles; basically able totravel through standard doorways andclimb stnirs; battery or gasoline fueledpower supply; teleoperator controlled;untethered; and an optional manipulatorthat should be able to lift/transportmore than 20 kg (44 lbs). The recentadvances in minaturization and the em-ployment of microprocessors and compu-ters have enabled mobile robots tobecome more reliable and versatile intheir missions conducted in the hazai—dous environment. There are, however,several issues which may limit the abi-lity of these devices to respond with100 X assurance in these situations;these issues are being addressed to byseveral national and government-spon-sored H & D programs, as for example theElectric Power Research Institute (PaloAlto, CA) and the CESAR facility at theOak Hidge National Laboratory.

REFERENCES

Meieran, H.B., 1984, "Remote Maintenanceand Assured Availability of Remote-ly Controlled Robots Operating inHarsh Environments", Proc^ RobotsWest Conference", Robotics Interna-tional, Anaheim, CA (Nov 28-30,1984), Paper MS84-1065.

Gelhaus, F.E., and Meieran, H.B., 1985,"Currently Available Mobile Tele-operators and Their Applicabilityto Radiological Emergencies," Proc^°f the. Workshop^ on Requirements ofMobile Teleogerators for Radigiogi;£§I IBS£S§3£X ES5E235.? §QSl B??gv-ery, US Dept. of Energy/ArgonneNational Laboratory, Dallas, TX(June 24-25, 1985), 65.

Meieran, H.B., 1986a, "Robotics: A Tech-nology on the March", Nuclear Engi^n§?EinJ£ ini§rDati2s?i i 3i (380)"(March 1986) , 33?

Meieran, H. B., 1986b, "Mobile Robots inNuclear Facilities, a Status Re-port," Proji of lgt Regional Con-ference, American Nuclear Society,Pittsburgh, PA (Sept 22-23, 1986).

100

TABLE 1SUMMARY OF BKCENT INCIDENTS INVOLVING KELEASES OF HAZARDOUS MATERIALS

Nuclear Plant

Chemical

Transportali on NaturaJ

Cherno- Miamis- San LakeChronological TMI-2 byl-4 Bohpal burg Antonio N i os ,

Sequence , PA USSR India OH TX Cameroon

a Initial Human X X XDeaths

b Initial Human X X XInjuries

c Livestock X X(?) XDeaths

d Hazardous to X X X X X XRescue Workers

e Eliminate Main X X X X XSource of Inci.

f Kvacuato 50K 135K >200K 2K >1K .5KSurvivors

g Bury Dead X X(?) XLivestock

h Delayed Inju- Xries + Deaths

i Delayed De- Xstruction &Burial Live,

j Residual Long- X X XTerm Contamin.

TABLE 2SUMMARY OF SPECIFIC MISSION/TASKS TO BE CONDUCTED

BY MOBILK ROBOTS DURING HAZARDOUS INCIDENTS

Chronological Sequence

Initial Eli'mi- Evacu-Human nate ate

Bury Delayed ResidualDead Liv«- Long-

Miss von/Task

SecurityRadiation MoitoringChemical MonitoringCore Samp]ingBury Livestock

Trench ConstructDestroy LivestockSample Aquisition

SludgeAirSmear

Video SurveillanceFire/Chemical FightTunnelingReactor RepairReactor MaintenanceDecontaminate;

Scrape/ScabbleSprayCoat/Strip

DeathsInjury

XXX

X

XXXX

MainProblem

X

XXXX

XXXXXXXX

XXX

Survi-vors

X

XX

X

Live-Si ock

X

XX

X

X

StockDestroy

X

XX

X

X

X

TermContain.

•A

XX

XXXXX

XXX

ANS Topical Meetisg on Radiological AccMeato—Perspectives and Emergeacy Plaawfag

Communications Systems for Emergency Deployment ApplicationsCharles A. Gladden

ABSTRACT The Emergency Response Team(ERT) c o m m u n i c a t i o n s s y s t e m wasd e v e l o p e d by the U. S. Department ofEnergy (DOE) to p r o v i d e r a d i o andt e l e c o m m u n i c a t i o n s s e r v i c e f o rs c i e n t i f i c and management e l e m e n t sl o c a t e d i n , and a d j a c e n t t o , anemergency a r e a . The t e l e p h o n e systemc o n s i s t s of s i x n o d e s , i n t e r c o n n e c t e dv i a microwave l i n k s tha t support T-ldata l inks and simultaneous two-way l i v ev i d e o . The rad io network i s a s e l f -c o n t a i n e d VHF sys tem arranged aroundp o r t a b l e and programmable r e p e a t e r s .The system i s comprised of approximately133 DES v o i c e - p r i v a t e r a d i o s and 168c l e a r t e x t r a d i o s . C a p a b i l i t y i sa v a i l a b l e in t h e form of p o r t a b l eI n t e r n a t i o n a l M a r i t i m e S a t e l l i t e(INMARSAT) terminals that a l l o w d i r e c td i a l a c c e s s to c o a s t e a r t h s t a t i o n s inthe U. S. or o t h e r c o u n t r i e s .

C o m m u n i c a t i o n s s y s t e m s f o remergency f i e l d dep loyments have beendeveloped by s e v e r a l organizat ions suchas the N a t i o n a l F i r e Radio Cache, FBI,White House C o m m u n i c a t i o n s Agency(WHCA), and m i l i t a r y , e t c . With rareexcept ion , however, these are developedto p r o v i d e r a d i o c o v e r a g e o v e r theimmediate area of the emergency and dov e r y l i t t l e to a d d r e s s the problem ofproviding voluminous telecommunicationswithin the emergency area or l ink ing thee m e r g e n c y a r e a w i t h t h e undamagedp o r t i o n of t h e i n t e r n a t i o n a l d i a lnetwork.

During the Cosmos incident (MorningL i g h t ) in January 1978, and the ThreeM i l e I s l a n d i n c i d e n t in March 1979 , i tbecame i n c r e a s i n g l y o b v i o u s that a l lc o m m u n i c a t i o n s s u p p o r t s h o u l d betransportab le and independent of l o c a lcommunicat ions s y s t e m s . With t h i s inm i n d , t h e D e p a r t m e n t of E n e r g y ' sE m e r g e n c y R e s p o n s e Team (ERT)

communications system has been developeda r o u n d t h e p h i l o s o p h y t h a tcommunicat ions i n , or a d j a c e n t t o , theemergency area were n o n - e x i s t e n t p r i o rto the i n c i d e n t or would be damaged ors e v e r e l y o v e r l o a d e d as a r e s u l t of theinc ident . With t h i s e s t a b l i s h e d as theb a s i c p r o b l e m , s y s t e m s h a v e beend e v e l o p e d which p r o v i d e b o t h r a d i oc o v e r a g e of t h e e m e r g e n c y a r e a andt e l e p h o n e s e r v i c e which i n c l u d e then e c e s s a r y microwave l i n k s to a s s u r ei n t e r f a c e to the o u t s i d e i n t e r n a t i o n a ld i a l network. A part of t h i s microwavesystem i s a l s o used to t r a n s p o r t l i v ev i d e o as w e l l as t e l e c o m m u n i c a t i o n sbetween the s e v e r a l major DOE operat ionsnodes wi th in the emergency area.

One of the more important majore l e m e n t s i s t h e p o r t a b l e t e l e p h o n esystem that provides telecommunicationsb e t w e e n v a r i o u s s c i e n t i f i c andmanagement e l e m e n t s l o c a t e d i n , anda d j a c e n t t o , the emergency a r e a . Thiste lephone system c o n s i s t s of s i x nodes,packaged and as sembled in r u g g e d i z e dpacking c a s e s , t h a t s e r v e as o p e r a t i n gracks as w e l l as t r a n s p o r t c o n t a i n e r s .These nodes are i n t e r c o n n e c t e d v i a ERTmicrowave l i n k s tha t support T-l datal i n k s and s i m u l t a n e o u s two-way l i v ev i d e o . The m i c r o w a v e l i n k s a r ei n s t a l l e d i n " r e a l t i m e " on anya v a i l a b l e high po ints such as b u i l d i n g s ,t o w e r s , m o u n t a i n s , e t c . and a l l o w them a j o r i t y of command and s c i e n t i f i cs tructures to funct ion at safe d i s t a n c e sfrom nuclear emergencies. In the eventtrunk c i r c u i t s are u n a v a i l a b l e in theimmediate area, up to 24 c i r c u i t s can beprocured at a remote l o c a t i o n (up toapproximately 50 m i l e s ) and transportedusing e x i s t i n g m u l t i p l e x and microwavea s s e t s . Trunk c a p a c i t y of the b a s i ct e l e p h o n e sys tem i s 48 "loop s t a r t "c e n t r a l o f f i c e t r u n k s , w i t h ana d d i t i o n a l 48 trunk c i r c u i t s that can beaccommodated in a T-l format . S t a t i o nl i n e c a p a c i t y i s 261 l i n e s when a l lnodes are f u l l y deployed. The s i x major

101

102

n o d e s a r e t h e T e c h n i c a l O p e r a t i o n sC e n t e r , Command P o s t , Forward C o n t r o lP o i n t , Working P o i n t , Forward S t a g i n gArea, and Main Staging Area.

The r a d i o n e t w o r k i s a s e l f -c o n t a i n e d VHF s y s t e m a r r a n g e d aroundp o r t a b l e and programmable r e p e a t e r s .The system i s comprised of approximate ly200 DES v o i c e p r i v a t e r a d i o s and 200c l e a r t e x t r a d i o s . A l l are p o r t a b l ehand r a d i o s and h a v e an R.F. powero u t p u t r a t i n g of a p p r o x i m a t e l y s i xwatto .

VHF r a d i o n e t w o r k i n s t a l l a t i o nnormal ly beg ins with a r r i v a l of the ERTa d v a n c e p a r t y on s c e n e . The f i r s tr e p e a t e r c e l l i s i n s t a l l e d in t h ev i c i n i t y of the a d v a n c e p a r t y commandp o s t or in the v i c i n i t y of the i n c i d e n ts i t e , i f r e q u i r e d . R e p e a t e r c e l lcoverage i s p r o g r e s s i v e l y increased bymount ing equipment on a v a i l a b l e h i g hp o i n t s (mounta in t o p s , t o w e r s , e t c . )u n t i l the area of emergency has thoroughR.F. c o v e r a g e w i t h a minimum of t h r e en e t s . Thie coverage area t y p i c a l l y hasa 2 5 - 5 0 m i l e r a d i u s . S l a v e r e p e a t e r sare o f t e n used to augment the c o v e r a g ei n shadow a r e a s or in major b u i l d i n g sand c o m p l e x e s w i t h h i g h R.F .a 11 enuat i o nf a c t o r s . A l l r e p e a t e r s are r e g e n e r a t i v ewhen opera t ing in the d i g i t a l mode, thusa l l o w i n g t r a f f i c to be r e p e a t e d 2 -3t i m e s w i t h m i n i m a l b i t e r r o r r a t e ,b e f o r e a r r i v i n g at the f i n a l d e s t i -nat i o n .

L i m i t e d c a p a b i l i t y h a s b e e nd e v e l o p e d i n t h e form of p o r t a b l eI n t e r n a t i o n a l M a r i t i m e S a t e l l i t e(INMARSAT) t e r m i n a l s that a l l o w d i r e c td i a l a c c e s i v i a the appropr ia te P a c i f i cor A t l a n t i c s a t e l l i t e , to c o a s t e a r t hs t a t i o n s i n t h e V. S . o r o t h e rc o u n t r i e s . Although t h i s has been usedp r i m a r i l y in e x e r c i s e s , i t s p o t e n t i a lv a l u e in responding to eaa1 1 to medium-s c a l e a c e i d e n t s / i n c i d e n t s in i s o l a t e da r e a s i s immense. A p r i s e example oft h i s w o u l d be t o p r o v i d e e m e r g e n c yc o m m u n i c a t i o n s i n t h e e v e n t of ana c c i d e n t in an i s o l a t e d area i n v o l v i n gweapons daaage or r a d i o a c t i v e d i s p e r s a l .

Th i s e a r t h t e r m i n a l has been u t i l i z e dv e r y s u c c e s s f u l l y in the a n n u a l DOEi n s p e c t i o n and s c i e n t i f i c a n a l y s i s ofthe Amchitka p r o j e c t in A l a s k a . I t ss m a l l s i z e ( a p p r o x i m a t e l y 3 s t a n d a r dl u g g a g e c a s e s ) a l l o w s i t t o bet r a n s p o r t e d on c o m m e r c i a l a i r l i n e s asluggage . Setup time i s approximate ly 15m i n u t e s , and the use of a r e l a t i v e l ys m a l l 4 - f o o t d i s h a l l o w s the s y s t e m toremain o p e r a t i o n a l in 40-50 knot winds ,i f proper ly sandbagged.

The ERT c o m m u n i c a t i o n s s y s t e m i ss e r v i c e d , t e s t e d , and s t o r e d i n ew e s t e r n U. S. l o c a t i o n . D e p l o y m e n tt imes vary with equipaent requirements ;however, the advance party can normal lybe ready to depart with a l l equipment intwo h o u r s , w i t h a c o m p l e t e d e p l o y m e n ttaking approximate ly 8-10 hours .

Future improvements are planned ins e v e r a l a r e a s ; however, the most c r i t i -c a l i s the assembly and f i e l d i n g of them u l t i c h a n n e l s a t e l l i t e system which w i l lp r o v i d e the ERT c o m p l e t e i n d e p e n d e n c efrom any e x i s t i n g e a e r g e n c y a r e ac o m m u n i c a t i o n s . Other major i m p r o v e -m e n t s p l a n n e d a r e t o c o n v e r t t h em i c r o w a v e baseband to an a l l d i g i t a lo p e r a t i o n , conver t the VHF radio networkto a ce 1 lu l a r - e a t e 11 i t e o p e r a t i o n , andp r o c u r e 1.5 Mbps l i v e a o t i o n v i d e ocompressors .

In summation , we f e e l EmergencyResponse Team emergency c o a a u n i c a t i o n ss y s t e a a e e t s t h e o b j e c t i v e s of the DOEby p r o v i d i n g a p o r t a b l e , l i g h t w e i g h t ,v e r s a t i l e , h igh c a p a c i t y a y s t e a that canbe f i e l d e d in a r e a s o n a b l e t ime and bet r a n s p o r t e d by m i l i t a r y a i r c a r g o orcommercial passenger carry ing a i r c r a f t .V a r i o u s i n c i d e n t s and e x e r c i s e s h a v ep r o v e n the s y s t e m can be e f f e c t i v e i np r o v i d i n g emergency c o m m u n i c a t i o n s inareas where no other s y s t e a s e x i s t , havebeen des troyed or o v e r l o a d e d .

This work was performed by EG&G/EMf o r the U n i t e d S t a t e s Department ofEnergy, Of f i ce of Nuclear S a f e t y , underContract Huaber DE-AC08-83HV10282.

ANS Topical Meeting on Radiological Accidents—Perspectives and Emergency Planting

A Mobile Laboratory—Emergency SampleAnalyses at the Accident Site

R. H. Wilson

ABSTRACT The Pacific Northwest Laboratory (PNL)provides technical support to the Department ofEnergy (DOE) with rapid response to the site ofa radiological accident for initial assessmentof radiological conditions in the surroundingenvirons. A mobile laboratory manned with atrained crew 1s capable of collecting, preparing,and analyzing air, liquid, vegetation, and soilsamples from the areas surrounding the accidentsite. Rapid assessment of radiological conditionsis important to enable responsible agencies toImmediately administer appropriate response andcoordinate any necessary longer-term managementof the accident site. Sensitive Instrumentationand support equipment provide the capability torapidly assess environmental conditions foruncontrolled areas as stated in DOE Order 5480.1Chapter XI. A regimen is required to maintainproficiency and to train new crew members. Aprofessional team responding with a well-equippedmobile laboratory capable of providing quickanalytical results has a positive Impact on publicrelations. Confidence is also generated amongother agencies with an awareness of the technicalexpertise that 1s available for assistance andbackup in the event of a radiological accident.

INTRODUCTION

The Federal Radiological Monitoring andAssessment Plan (FRMAP), a part of the FederalRadiological Emergency Response Plan (FRERP)establishes a means to request and provide federalradiological assistance and a framework tocoordinate radiological monitoring and assessmentactivities of government agencies In support offederal, state, and local government activities.

The DOE coordinates all federal offsiteradiological monitoring and assessment operationsduring the Initial phase of emergency response.The DOE has the responsibility to provide personneland equipment required to coordinate and performoffsite monitoring and evaluation activities 1ncooperation with other federal components. TheDOE 1s also designated the cognizant federal agency(CFA) for both onsite and offsite radiologicalevents involving DOE-owned materials. In addition,

the DOE 1s also responsible for maintainingnational and regional coordination offices aspoints of access to federal radiological emergencyresponse.

The DOE operations office in Richiand,Washington (DOE-RL) has been designated as thepoint of access for Region 8 (Washington, Oregon,Alaska). To assist DOE-RL in fulfilling Itsobligation for response 1n the event of a Region8 radiological emergency, PNL Is charged withproviding appropriate technical supportcapabilities.

Of primary concern to both PNL and DOE-RL Isthe rapid and appropriate response to aradiological Incident. Following the Initialassessment by DOE and PNL, 1t may be concludedthat more sensitive and detailed analyses arerequired to characterize radiological conditionsand define the affected area as quickly aspossible. A mobile unit equipped with the properInstrumentation 1s required to make rapid on-the-spot analyses that identify the radiocon-taminants and their concentrations In theenvironment. With the In-depth measurementsprovided by a mobile unit, governmental agenciesand local authorities can Immediately administerappropriate follow-up actions and coordinate anyrequired longer-term management of the accidentsite.

A mission statement was developed for a mobileunit and from that mission statement the followingspecifications were set forth to provide the levelof analytical capability and support required byDOE-RL:1. space for the Instrumentation and support

equipment necessary to make a comprehensiveassessment of an accident area In a maximumof three days

2. overnight living accommodations for a maximumof four crew members

3. capability for deployment, operation, andreturn through the mountainous terrain ofremote areas (exclusive of travel enunimproved roads) where supplies may belimited

4. auxiliary power-generating capability suf-ficient to Independently operate laboratoryequipment without replenishment for a maximumof three days

5. equipment capable of detecting concentrationsof gamma-emitting radioisotopes associated

103

104

with major radiological accidents (theinstrumentation should be sensitive enoughto determine whether or not concentrationsin liquid and air samples are above 10 CFR20 Table II concentration guides and theinstruments should be capable of determiningcomparable concentration levelsforvegetationand oil in short counting times, <5 minutes)

6. equipment and supplies for the collection,preparation, counting, and analyses of aminimum of 50 environmental samples per day

7. capability to communicate with the UnifiedDose Assessment Center (UDAC) facility inRichland when distance and terrain will allowand with crew members working away from theunit

8. dosimeters and read-out instrumentationcapable of determining radiation dose inthe environment and the dose received bypersonnel worki pg i n the radi ati on envi ronment

9. mobilization capability within one hour.

MOBILE LABORATORY DESCRIPTION

The mobile laboratory is a 27-foot-long fifthwheel trailer that has been converted into alaboratory for use in the field. A large pickuptruck with four-wheel drive and a crew cab is usedto pull the mobile laboratory and transportmaterial, equipment, and people.

The basic trailer was built by Conestoga.The Snobird-model trailer interior was redesignedto accommodate counting equipment and stillmaintain a reasonable level of comfort for thecrew. Doors, counter space, and cupboards wererearranged and a heavy support beam was installedunderneath the floor above the axle where theheavy lead counting chamber was placed. The dryweight of the redesigned trailer was about 6770pounds on delivery to PNL. After the equipmentand materials were installed, the dry weight ofthe trailer was 10,000 pounds.

Some of the major items installed in thetrailer for the crew's comfort are a 6-ft3gas/electric refrigerator, 24,000-BTU ductedfurnace, 6-gal hot water heater, 40-gal watertank and pressure water pump, single sink, 40-galgray water tank, wet-bowl toilet, shower, 27-galwaste tank, air conditioning, storm windows,awning, 6.5-kW-rated generator with 10-gal gastank supply, two twin beds, and Weathertronicsrecording wind system. A radio base station wasinstalled with charging units for four portableradios.

COUNTING EQUIPMENT

The detector used in the trailer 1s a EG&GORTEC GEM Series HPGe (high-purity germaniumcoaxial detector system) and a Tracor NorthernTN4000 analysis system, keyboard and display,anda printer. A special counting chamber was designedto allow the sample container (450-mL Marinellibeaker) to be raised into counting position andyet have 4 Inches of lead shielding on essentiallyall sides. A precision track with roller bearingswas installed in the trailer floor to hold thecounting chamber and allow easy movement to itscounting position under the detector or anchor

it in the center of the trailer for the travelposition. Because of its weight (about 800 pounds)positioning and anchoring the counting chamboris very important. If the lead chambe» I.loose during travel it would probably go throughthe wall of the trailer during a turn. A pulleyarrangement lowers the detector to the countingposition or holds it securely in its travelposition.

The detector has been calibrated for fourdifferent counting geometries using 450-mLMarinelli beakers containing liquid, soil,vegetation, and 2-in-diameter air filter papers(see Table 1. Counting Efficiencies). Calibrationis accomplished using a National Bureau ofStandards (NBS) source composed of 126Sb, 1MEu,and Eu which supplies several calibrationpoints over a wide range of energies. Otherstandards such as 6BCo, 137Cs, and m A m arealso available. The equipment has been verystable and recalibration is scheduled on aquarterly basis only. A table of calibrationfactors (Table 2) has been established for thefour counting geometries across a broad spectrumof energies so the efficiency for any energy canbe determined with reasonable accuracy. Detectionlevels for gamma-emitting isotopes are well belowthe concentration limits stated in Table II of 10CFR 20, Appendix B and in DOE Order 5480.1. Theselevels are stated for Individuals; one-third ofthese values are considered acceptable for asuitable sample of the population. Thissensitivity can be achieved using one-minutecounting tines.

THE MOBILE LABORATORY TEAM

Fourteen staff members from the PNL HealthPhysics Department having a variety of expertisein dosimetry and Instrumentation make up the mobilelaboratory team. These people are on a call list,and in the event of a radiological Incident aminimum crew of four team members will be assembledwithin one hour. During an Incident memberswould be alerted to stand by If the need for themobile laboratory services had not been determinedimmediately.

Crew member training is conducted on aquarterly basis. During these training exercisesvarying aspects of operation such as thepreparation for travel, travel to some location,set-up for operations, sample preparation, samplecounting, and evaluation of results are practiced.It 1s not possible to trafn all members at onetime because of other work commitments. A trainingsession usually consists of five or six membersand all members participate at least twice a year.

The operational duties of the mobilelaboratory crew have been divided into fivefunctions: team leader, communicator/recorder,counter/analyzer operator, sample preparer, andthermoluminescent-dosimeter (TLD) reader operator.

The team leader will be appointed by the PNLEmergency Preparedness Manager or his delegate atthe time team members are notified to activate themobile laboratory. If this individual has notbeen appointed at the time of notification, thefirst member arriving to activate the laboratoryshall act as team leader to ensure the unit is

105

readied for travel and all members of the creware assigned to the tasks necessary for thelaboratory, pickup truck, and equipment operation.

The communicator/recorder is responsible formaintaining communications with the base station(usually UDAC), crew members, and other teams withportable radios working in the field. A recordof all activities will be maintained In a mobilelaboratory logbook starting with the preparationfor travel and closing with the return to thestorage location.

The counter/analyzer operator is responsiblefor preparing the equipment for operation andchecking the supplies for operation in the field.

The person assigned to sample preparationsat the accident site also participates In theinventory of equipment to prepare the laboratoryfor travel.

The TLD-reader operator will haveresponsibility to ensure that the reader andequipment are operable and adequate supplies arein the laboratory. This operator will also assistin preparation for travel, Inventory of equipment,and set-up for operations at the work site.

Because of the versatility and expertise ofcrew members, work assignments can be changed torelieve crew members or to concentrate effort 1nsome area of the laboratory operation such assample preparation. This function Is time-consuming; when a large number of sample resultsare required, most of the crew members directtheir efforts In this function.

PUBLIC RELATIONS ASPECTS

Wherever the mobile laboratory goes, whether1t is on display or Involved In an emergencyresponse exercise, 1t stimulates a great deal ofInterest among people who work with or tour It.People participating In large emergency responseexercises Involving federal, state, and localgovernments, military and civilian agencies areIntrigued with the capability, sensitivity, andamount ->'- equipment that is readily portable andavailable 1n the field.

In the local communities where radiation 1sstill a mysterious term, people are Impressed andsomewhat overwhelmed at the display of strange-looking equipment, blinking lights, and strangenoises. However, there seems to be somesatisfaction In having this peculiar equipmentand the "experts" available to help them If anaccident did occur near their community.

This response has been particularly noticeablein the state of Oregon during the exercises that

have been conducted along the Interstate-84corridor. This Interstate highway is a mainroute for the thousands of trucks that transportmostly low-level radioactive waste through easternOregon to the Hanford Site for disposal. Inrecent years the public has become much more awarethat these shipments are passing through theircommunities on a dally basis and the people areconcerned about potential accidents. PNL providestraining 1n the detection of radiation to localpolice and fire departments; PNL also presentsinformation on emergency response and demonstratesthe capabilities of the mobile laboratory. Thetraining and information relieves a lot of theanimosity toward nuclear energy and creates anawareness that an effort 1s being made to protectpeople and the environment from the effects ofradiation.

ACCIDENT EXPERIENCE

The capability of the mobile laboratory wasdemonstrated 1n May of this year during a fieldtraining exercise foil owing the Chernobyl accident.The laboratory was moved to a location near theHanford Site and samples of vegetation and soilwere collected for analysis. When the samples werecollected, prepared, and counted using theprocedural methods established for Initial responseto a radiological accident, activity In the rangeof 5 />C1 per gram of I was detected. Ameasurable difference In activity was noted amongsamples collected from the outer and toner foliageof the vegetation sampled. No " I activitycould be detected In soil samples collected Inthe same location as the vegetation samples.

At the time the Chernobyl "cloud" waspredicted to pass over Washington, rainshowersoccurred. The mobile laboratory collected andanalyzed rainwater samples and found thatconcentrations of I were similar to thoselevels reported by other laboratories.

The measurement of I in the range of 5 pCiper gram was considered very good and demonstratedthe capability to meet and greatly exceed thespecifications for detection under fieldconditions.

REFERENCES

U.S. Department of Energy (DOE). 1981. "Require-ments for Radiation Protection." In DOE Order5480.1, Chap. 11. Washington, D.C.

Standards for Protection Against Radiation,10 C.F.R. Part 20 (1983).

u . <_>a

• -^?m a

10 i-U- C>a.

NNNNMt-lHOHOOOOO

14- -O<*- V)UJC\J

I

ce ff. d%

E +2 NrtMMN UJ CM

oto

Cou

ntE

ff.%

WNNNNHHHHOOOOO

ai—«

crO H-O UJ

UJUJUJlOUJl>0(>Ot/)LLJUJL'.~UJL>JUJ UJUJLUMlUMl/IOIliiUJlUUJLlJ

c

UJ

en

u. oa

10 ao O r o N o

it-ic\IC\JCiJC\JC\J

fO - r -U. LJ

a

3oo

LTER

PA

PE

t—<

u_

UJ OJ

+

Cou

ntE

ff.%

keV

• - H • »•«

CSJ TM f C"J CTl CM I

UJ

Zt-H

ION

UJ CM1

• &«M-T3UJ OJ

3 3 3 ^ 3 ^ 3 ^ . 0 3 3 3 3 3LLJLUUJtOI-UtOCOtO LLJ UJ UJ UJ LUui la IO ^ l o i a i o ^ * * ^ *U ) « U ) N U ) n N N U ) U ) U ) u ) U )

107

EnergykeV

<100100-150151-200

201-250251-300

301-350351-400

401-450451-500

501-550551-600

601-650651-700

701-750751-800

801-850850-900

901-950951-1000

1000-11001101-1200

>1201

Veg.Factor

12.6911.5511.55

13.8615.81

17.3318.77

20.4823.10

25.7428.15

31.0733.37

36.0439.17

39.1740.95

42.9045.05

45.0550.0552.99

SoilFactor

18.7715.8116.38

17.6619.17

21.4523.10

25.0328.15

31.0733.37

36.0437.54

40.9542.90

45.0547.42

50.0552.99

56.3156.3160.06

Table 2, Calibration

AirsFactor

6.225.395.49

7.518.58

9.3810.01

10.4811.40

12.0113.06

13.6514.08

16.0918.02

18.7719.58

21.4522.52

23.7125.0326.50

Factors

Liq.Factor

16.6814.7715.02

19.5821.97

21.9723.10

23.1026.50

27.3028.15

30.0332.18

40.9542.90

45.0547.42

47.4247.42

47.4250.0552.99

ANS Topical Meeting on Radiological Accidents—Perspectives and Emergency Planning

Developing a Comprehensive and Accountable Data BaseAfter a Radiological Accident

Hollis A. Berry and Zolin G. Burson

ABSTRACT After a radiological accident occurs,it is highly desirable to promptly begin develo-ping a comprehensive and accountable environ-mental database both for immediate health andsafety needs and for long-term documentation.The need to assess and evaluate the impact ofthe accident as quickly as possible is alwaysvery urgent, the technical integrity of the datamust also be assured and maintained. Care musttherefore be taken to log, collate, and organizethe environmental data into a complete andaccountable database. The key components ofdatabase development are summarized as well asthe experience gained in organizing and handlingenvironmental data acquired during:

I. TMI (1978).2-. The St. Lucie Reactor Accident Exercise

(through the Federal RadiologicalMeasurement and Assessment Center (FRMAC),March 1984).

3. The Sequoyah Fuels Inc., UraniumHexafluoride Accident near Gore, Oklahoma(January 1986).

4. Chernobyl Reactor Accident in Russia(April 1986).

I. INTRODUCTION AND BACKGROUND

After a radiological accident of substantialmagnitude occurs, physical data obtained can bedivided into two categories: (1) data thatconcerns immediate health and safety needs, and(2) data that satisfies long-term documentationrequirements. It is then highly desirable tobegin developing a comprehensive and accountableenvironmental database as soon as possible afterthe occurrence of a radiological accident.Assessing and evaluating the impact of theaccident is always urgent. Yet the technicalintegrity of the data must also be assured andmaintained before any assessment, evaluation, orsummary is completed. To adequately assess theaccident, the data must be managed and organizedinto a complete and accountable database.

The Three Mile Island (TMI) reactoraccident revealed the fact that environmentalmonitoring and assessment tasks are extremely

important, and become increasingly complex asthe magnitude of the accident increases. Inresponding to the State of Pennsylvania'srequest for federal technical assistance theUnited States Department of Energy (DOE)provided ground and aerial monitoring teams,coordinated the monitoring efforts of all teamsperforming off-site environmental measurements,and created an assessment team of laboratoryexperts. The assessment team examined themonitoring results to ensure that the correctinstruments and samples were utilized, that theinstruments were properly calibrated, thatcommon geographic terras and units of measurewere used, and that any. ambiguous readings wereconfirmed or denied by additional monitoring.

During the response to the TMI accident,the amount of data being examined soon becameoverwhelming; it was evident that an environ-mental database needed to be developed. Someeffort was made to put the data in properperspective but because of the large number ofdata points being gathered, the lack ofqualified personnel and appropriate equipment onsite, and the lack of a definitive plan, thistask was not adequately addressed.

A definitive plan for handling reactor typeaccidents has been formalized into DOE's respon-sibilities in managing the Federal RadiologicalMonitoring and Assessment Center (FRMAC). Exer-cises and responses to accidents have furtherunderlined the importance of databasedevelopment in relation to the monitoring andassessment effort. During the St. Lucie ReactorAccident Exercise in March 1984, preparationswere made to develop a comprehensive andaccountable database but time was short and theplay and r"ata were artificial. Nevertheless, itwas evident that more detailed plans andprocedures, as well as an adequate portablecomputer system and more qualified personnel,would be needed in a real accident situation.

A real accident did occur at the SequoyahFuels Inc., Uranium Hexafluoride Plant nearGore, Oklahoma in January 1986. Several daysafter the accident EG&G/EM was asked by DOE toformulate a database for the Nuclear RegulatoryCommission (NRC) task force assigned to assess

109

110

the accident and its impact on the environment.The accident seemed small and perhaps

insignificant, but more than 2000 data pointswere incorporated into the database in arelatively short time. The need for thoroughnessand accountability was again evident.

On April 26, 1986 the Chernobyl Reactoraccident occurred in Russia. EG&G/EM assistedDOE in formulating a database of the falloutdata received. The importance and usefulness ofa database were recognized at the outset, butthe effort required to manage incoming data wasunderestimated and our involvement with thedatabase formulation was delayed until a weekafter the accident. The amount of data beingreceived at the point of EG&G/EM's involvementwas tremendous and the task ahead was stagger-ing. Approximately 10,000 points were enteredinto the database in a short period of time.Yet, this represented only a small portion ofthe data available for assessment andevaluation.

The evaluation and entry functions performedon the Chernobyl data were similar to those thatwould be performed in any FRMAC or emergencyresponse operation. We were able to design thedatabase and its implementation procedureswithout the extreme pressures inherent in anactual radiological emergency situation.

The Sequoyah Fuels and Chernobyl accidentsclearly emphasized the urgency of beginning thedatabase development and assessment processes atthe earliest possible time after a major acci-dent so that credible assessments and decisionscan be made Quickly. We learned that it isunrealistic to design, modify, or formulate fromscratch a database system while officials andmanagers are demanding to kno«r how serious theproblem is and what actions should be taken.Instant answers will always be requested of themonitoring and assessment teams in aradiological response. The authorities' abilityto provide substantive assessments of the threatto the public health and safety hinges on theability to rapidly assemble and integrate allavailable information on radiation measurements,location of surrounding populations, identifica-tion of farm animals and crops, surface-fedwater supplies, etc. Through the experiencegained at Three Mile Island, the St. Lucieexercise, the Sequoyah Fuels accident, and theChernobyl response, much has been accomplishedin identifying major elements of the task ofdeveloping a comprehensive and accountabledatabase alter a radiological accident.

II. DEVELOPMENT OF EMERGENCY RESPONSE DATABASECAPABILITIES

The complexities of the assessment tasks inan emergency response are easily overlookeduntil one examines the many variables of theproblem. Short-term and long-term effects,complex food chain and concentrating mechanisms,alpha and beta dose contributions, potentialwatershed impact on surface drinking watersources, and protective action recommendationsfor all age groups and for farm animals are justa few of the many results, issues, and question:.

the assessment team must consider when providinginformation to the decision-makers. Databasedevelopment therefore becomes an integral partof data assessment.

The experience gained from previous exer-cises and operations in formulating databasesduring an emergency response have identifiedthree major areas of development.

1. Development of plans in cooperation withpotential users and/or customers.

2. Development of computer hardware andsoftware capabilities for field and homebase support roles.

3. Development of procedural guidelines forreceiving and assembling raw data, andverifying its quality and applicability tothe response and assessment requirements.

In an emergency response situation, per-sonnel from federal agencies, contractors andstate and local organizations will be calledupon to assist in the response and alsoparticipate in the assessment and monitoringteams. Later, as the incident proceeds out ofthe emergency stages Into the long-term moni-toring and recovery phases, a final custodianwill be determined. Therefore, it is necessaryto coordinate emergency response databasedevelopment plans with all potential users.Their needs and requirements should be con-sidered in the development of this capability.

The Information collection problem in anemergency response is massive and computers havebeen used extensively to aid in compiling andorganizing data. Currently and in the past,within EG&G/EM, emergency response databaseshave been stored in small, portable tabletopcomputers. This may not be the best computerfor this task, particularly if the database Isvery large. The right computer system should becompatible with the needs and capabilities ofthe group assembling the database and doing theassessment as well as potential users andcustodians. Other basic requirements are asfollows:

1. The system must be easily transportable inruggedized containers.

2. The basic components must also be fairlyrugged to withstand the harsh environmentsthey encounter (baggage handlers, highhumidity, etc.).

3. It oust support a multiuser/multitaskenvironment.

4. It must be capable of the timely handlingof large volumes of data.

5. It must be able to utilize third partyhardware and software that are supportedand proven.

6. The software must be compatible with othersystems.

7. Power requirements must be low tomoderate.

8. The equipment must be readily available.9. It should include the software and

peripherals for automated plotting of thedata and information transfer.

A larger computer (or an assemblage of the

Ill

transportable work stations considered above)will be needed at home base. This system can beused during the early stages of the emergency toinput and store data fro:* locations not In thecritical area ..nd prior to the arrival of thedatabase and assessment teams.

Later, the entire database would reside inthis system and be used to produce reports of amore conclusive nature until a final custodianis determined.

The technical integrity of the data, at allstages of development and use, must be assuredbefore the Information can be released foraction or information purposes. Key aspects ofdatabase development can be summarized asfollows:

1. All original data sheets should be logged,collated, and cross-referenced for easyreferral.

2. The database should be complete, organi-zed, and trackable in every detail.

3. Methods and plans for dealing withqualitative vs. quantitative data need tobe developed.

4. The information must be in a form that al-lows technical integrity of the data to beconfirmed as quickly as possible. Alldata should be evaluated as to quality,validity, and priority of importance tothe response prior to its entry into thedatabase. In the early stages of theresponse, the assessment team should dealonly with data that is consistent andappropriate to the incident. All otherdata can be considered as soon as theimmediate threat to the public hasdiminished. In addition, all entries intothe database should be edited andverified.

5. Key data should be retrievable in anycategory desired. The results need to besummarized, evaluated, and placed inproper perspective quickly and easily andon a continual basis.

6. All the environmental data should be plot-ted on maps or overlays separately by datatype and medium and in appropriate detail.

7. The requirements for distribution of thedata will be very great. The need will becontinuous for detailed, organized datafor assessment purposes and for highlysummarized data for managementinformation. The database must beorganized so that these requirements canbe met easily and quickly.

8. Adequate and qualified staffing should beassigned.

As with any database management program,development evolves and expands as more tasksand requirements are identified. In developingemergency response database capabilities anexpansive, dynamic approach must be maintainedto remain flexible and responsive to needs andrequirements,. In addition, the database itselfshould be dynamic so it can be easily and quick-ly modified to meet specific requests. The abil-ity to manipulate, massage, and plot the datawill also facilitate assessment and decision

making in an emergency response situation.

III. IMPLEMENTATION CONSIDERATIONS

In implementing the development of responsecapabilities In the event of a significantemergency, a number of specifics need to beconsidered. These include detailed planning,database construction, staffing and deployment.

In planning emergency response databasecapabilities several activities can aid in thisdevelopment. These include:

1. The preparation of maps on a variety ofscales using a common coordinate system.This (1) aids outside responding organiza-tions in locating sample sites Inunfamiliar areas, (2) facilitates plottingthe data, (3) assists the assessment tea*in their evaluation, and (4) allows foreasy retrieval of information from commonareas.

2. A compilation of baseline environmentalinformation including routine monitoringnetworks and associated results, popula-tion distributions, locations of criticalfacilities, agricultural areas, potablewater supplies, etc. is helpful to theassessment team in determining the impact,if any, resulting from the accident.

3. The standardization of collection andanalysis techniques, the formulation ofdata and information collection forms, theestablishing of common and consistentunits of measure, and determining the typeof information required for a variety ofaccident scenarios.

4. The development of detailed Implementationplans and procedures and the exercise ofthese plans.

5. The gathering of required supplies includ-ing standard office equipment andsupplies, a copy machine, an adequatecomputer, etc. This equipment should beprepared and packaged for deployment.

6. Personnel assignments, training, andcrosstrainlng.

The following information should be includedin the computer based database:

1. The reference ID number assigned to eachpiece of original data received.

2. The time and date the sample ormeasurement was taken.

3. The identity of the organization thatcollected the sample and the organizationthat conducted the analysis.

4. The sample collection ID number and thelaboratory analysis ID number.

5. The location at=|hich the sample wascollected, base\j on a common coordinatesystem.

6. The medium and type of sample in additionto the type of analysis.

7. The results and units of measure.8. Any significant comments that may affect

data assessment.

A considerable amount of the information

112

provided is of a bookkeeping nature but allowsthe database and assessment team to backtrackand verify incomplete or inconsistent informa-tion either with the laboratory that conductedthe analysts or the agency that collected thesample. The database format is dependent on thetype of accident and response. Preplanned anddeveloped databases should be very generic informat and be able to be modified, as required,during the response.

Adequately staffing the database center atan accident site is essential. In addition tocomputer operators and clerks, technicallyqualified people are needed to evaluate theincoming data on a continuous basis. Theseindividuals can also be utilized in the assess-ment of the date, when time and activities allow.For a small response similar to the SequoyahFuels Inc., accident four to six people would besufficient to handle the data. To handle the

data from a major reacttr accident that wouldInvolve the participation of several agenciesand organizations as many a two dozen or morepeople would be required 1T 24-hour operation.This level of involvement w<_:-ld allow for con-tinuous data entry and evaluation- It wouldalso allow the assessment team to assure thetechnical integrity of the data.

Finally, in the matter of deployment, data-base and assessment personnel should be an inte-gral part of the first or advance party respond-ing to a major incident. This will allow smoothdata and information flow on a continual basis.In addition, the database and assessment centershould be fully staffed as soon as possible.

This work was performed by EGSG/EM for theUnited States Department of Energy, Office ofNuclear Safety, under Contract Number DE-AC08-83NV10282.

ANS Topical Meetiig on Radiological Acciderts—Perspectives and Emergency Pbudag

A New Method for Presenting Off-Site Radiological MonitoringData During Emergency Preparedness Exercises

M. P. Moeller, G. F. Martin, E. E. Hickey, and J. D. Jamison

ABSTRACT As scenarios for exercises becomemore complicated and flexible to challengeemergency response personnel, improved means ofpresenting data must be developed to meet thisneed. To provide maximum realism and free playduring an exercise, staff at the Pacific North-west Laboratory (PNL) have recently devised asimple method of presenting realistic radiolog-ical field monitoring information for a rangeof possible releases. The method utilizes onlytwo pieces of paper. The first is a mtp of thepffsite area showing the shape of the plume forthe duration of the exercis.e. The second is a^emi-log graph containing curves relating expo-sure rate and iodine concentration to downwinddistance and time. Several -techniques are usedto maximize the information 'on the graph.

-• r rINTRODUCTION * < ,

Current government regulations requirethat the operator of a commercial nuclear powerplant make provisions for the conduct of emerg-ency preparedness 'exercises. Such provisionsinclude the- development of a scenario and thesupporting data necessary to drive the exercise.For those in the role of preparing exercise sce-narios, data presentation has become an impor-tant part of this task. In the last few years,data presentation on inplant reactor parametershas been advanced through the use of simulatorswhich add a new degree of realism and free play.To date, however, the full benefit of the simu-lators has not been realized due to limitationsin the methods used to present radiological datato monitoring teams. Previous methods have notallowed operator response actions on the simu-lator to determine the timing and magnitude ofthe radiological release.

In support of a major exercise conductedin the Spring of 1986 at the Department of En-ergy's Hanford Site, staff at PNL prepared theoffsite radiological onitoringm data. A majorgoal of the exercise was to design a scenariowhich would provide data for all anticipatedparticipant response actions. Consequently,the release pathway, timing of the release and

magnitude of the source term were all to be de-pendent on the participants' response actions.To meet this objective, a new method for pre-senting the radiological data was required.

This paper describes the new method forpresenting radiological data which was devel-oped for the Hanford Exercise. In order toshow the method's effectiveness, a review ofthe limitations of conventional methods is pre-sented. The new method provides consistentoffsite monitoring data for an emergency exer-cise scenario in which operator actions changethe progression of the accident, and hence, themagnitude of the radiological release. Inaddition, the ease by which the offsite moni-toring team controller could pass the data toplayers was emphasized.

Conventional methods of data presentationdo not readily allow for variations in thesource term responsive to operator actions. Inorder to provide data covering a range of po-tential operator actions, many different setsof tabular data are needed to cover the rangeof possible radiological releases. Therefore,a tremendous volume of data sheets is requiredif the scenario is to be flexible. As a result,it is not practical to use conventional methodsto provide field team controllers with data forrelease scenarios corresponding to every poten-tial combination of operator response actions.

The volume of data sheets currently re-quired is determined by the number of dependentvariables. Dependent variables include the mag-nitude of fuel damage, the release pathway, thetime of initiation, and the duration of the re-lease. Data that players request include gen-eral area radiation exposure rates using bothan open and a closed instrument window, grossparticulate and iodine air concentrations incounts per minute above background based on thevolume of air sampled, radioactive deposition,and contamination levels on the ground incounts per minute per square meter.

METHODTnstead of computing the required data for

all possible combinations of the dependentvariables and providing the information to con-trollers in tabular form, the new method con-

113

114

denses the data into a series of curves pre-sented on a semi-log graph. Data is calculatedusing a source term of 100 percent fuel damage.Maximum centerline values over the duration ofthe release are determined for downwind dis-tances. Similarly, the off-centerline distancewhere radiological data values are equal tobackground levels is determined for selecteddownwind distances. These data define the edgeof the plume and, hence, its shape. Values fordata midway between the centerline and theplume edge are assumed to be two-thirds (66.6%)of the centerline value.

Plotting such data on a semi-log graph isuseful in that multiplication and division by aconstant on a log scale always results in anincrease or decrease on the scale of the samedistance. In other words, a new curve which isa multiplicative constant of an existing curveon a log scale will have the same shape and bedisplaced from the original on the scale. Itis this important fact that allows a cursor tobe used which adds the desired flexibility.Having plotted the curve for 100 percent fueldamage, the cursor can be used as a measuringstick to determine values corresponding to lessfailed fuel if the operators take actions thatwould reduce the level of core damage. Whilewithin the plume, values for closed-windowreadings are assumed to be some fraction lessthan open-window readings, depending upon thescenario's assumed source term radionuclidemixture. Therefore, the closed-window readingsare a constant distance from the open-windowcurve and can be determined using the cursor.

Data prepared for a scenario can be con-densed on a semi-log graph as shown in Figu.-e1. A map of the exposure area is presented inFigure 2. This map indicates the outer plumeedges, the plume centerline, and the midwaydistance between each plume edge and the cen-terline. Curves on Figure 1 depict both thegeneral area radiation exposure rate (left-handabscissa scale) and the gross participate andiodine air concentrations (right-hand abscissascale). The air concentration information isgiven in counts per minute above background percubic foot of air sampled. Therefore, if anair sampler pumped at 10 cubic feet per minutefor five minutes, then the controller wouldmultiply the value obtained on the graph byfifty. Rules of thumb and other relevant in-formation are presented on the border of thegraph.

To implement this method of data presen-tation, controllers must be informed of theevent start time (initiation of release) andrelease category (duration and pathway of re-lease and percent of fuel damage). The releasepathway can affect the source term through con-sideration of filtration systems and releaseheight. Information can easily be communicatedby the lead controller through simple, pre-determined coded messages to offsite monitoringteam controllers. When required to provideinstrument readings, a controller should usethe following procedure:

• Prior to providing any instrument readings,the field controller should receive the eventstart time and release category informationfrom the lead controller.

• Knowing the event start time, the field con-troller will fill in the timeline at the top ofthe graph. The time of plume arrival at a down-wind distance is determined by the wind speed.The timeline, when filled in, should reflectthis relationship.

• When a request for an instrument reading ismade by a player, the controller will determinethe team's current location on the map. It maybe useful for the controller to carry a secondmap without the plume shape indicated so thatif there is a question concerning the team'slocation, the controller can ask a player forassistance without revealing the plume trajec-tory.

• The controller then determines if the team'slocation is within the plume. If not, the con-troller reports background readings. If it is,then the controller determines if the plume hasarrived at this downwind location by checkingthe timeline. If the plume is more than tenminutes upwind of the team's location, the con-troller reports background readings. If theplume is less than ten minutes upwind of theteam's location and still has not arrived, thecontroller should will report a reading betweenthe background level and the maximum valueafter the plume has arrived. If the plume hasarrived, the controller will report the maximumvalue for the team's location by proceeding tothe next step. If the plume is elevated atthis location, report closed-window readings(gamma only) when either open- or closed-windowreadings have been requested. If the plume hastouched down upwind of this location and is atground level, then open- and closed-windowreadings should be given as requested.

• To determine the maximum value after theplume has arrived, the controller should Iden-tify the team's downwind lateral location rela-tive to the plume centerline, outer plume edges,and the midway distance (midline) between.

• The controller then pinpoints the team'slateral location relative to the three curveson the graph (plume centerline, midline andplume edge) at the team's downwind distance.

• Using the cursor, the controller should ad-just the point on the graph based on the mea-sured distances associated with each releasecategory.

• The controller then reports the general arearadiation exposure rate (mR/hr) from the lefthand abscissa values.

• After the team has taken and measured an airsample, the controller can report the air con-

115

centration values using the values from theright hand abscissa multiplied by the volume ofair sampled (cubic feet).

Because of the cursor and condensed natureof the data, the method is quite versatile andreadily adaptable to different conditions. Forexample, during the 1986 Hanford Site Exercise,one scenario included a highly-radioactive puffrelease followed by a continuing, less radio-active release. Data for both of these releas-es were put on one graph by utilizing two time-lines relating the passage of each release overa given location. Another example is to pro-vide for a radical plume direction change mere-ly by utilizing two plume maps indicating thetwo dominant plume shapes.

Acceptance of the method was noted duringthe training of controllers for the Hanford Ex-ercise. In less than two hours, controllersunderstood the method, were able to respondcorrectly to simulated requests for data, andwere able to interpolate between known valuespresented on the curves to get a value for any

location. Many noted correctly that the levelof accuracy that they could provide from read-ing the curves and interpolating is consistentwith that available from actual field measure-ments. Many improved their graphs by includinglines that indicated a transformation of afixed monitoring route roadway into the down-wind and lateral coordinate system of thegraph.

SUMMARYAT scenarios become more flexible and com-

plex, as through the use of simulators, moreversatility is needed to present realistic datato offsite monitoring teams. Conventionalmethods are not adequate because of the volumeof tabular data necessary to correspond toevery potential operator response action. Themethod described in this paper provides a solu-tion that allows flexibility and data ingraphic form. Utilization of this new methodduring a complex exercise indicated that con-troller performance is enhanced.

116

Plume Travel Time to the Downwind Centerline Distance (hours:mint).

Bafora tha tima shown for each specific distance, give background readings. Aftarthat time, give data from the curves balow.

to =

0.00:20:401:00 1:30 2:00 2:30 3:00 3:30 4:00 4:30 5:00 5:30 6:00i :30 :60

KE

1E6

6E6

JZ

s.

i

0.02

0.01

Downwind Centerline Distance (Miles)

Conversion Factors and Other Controller Aids:

• Wind Sp*td = 6 mpti • Btfora PlumsArriva* Giva Background Readings

• 1 mR/hr = 3000 cpm on a GM Countar

t Dotim«t«r Raadingi - Stay Tima x Doa* Rat*

a Micro R Malar Ofttcala > 3 mR/hr

lodmaCountt , S M B p | i n f l Volume (himin-ft

a lodina cpm ~

Figure 1. Semi-log graph with radiological data.

117

300 Area

Figure 2. Map of plume exposure area.

ANS Topical Meetiag oa Radiological AccMeats—Perspectives ami Eawrgeacy Ptaaaiag

Enhancement of the 1985 Browns Ferry ExerciseThrough the Use of Spiked Samples

James L. McNees i

I

ABSTRACT The use of spiked environmentalsamples has proven to be a beneficial part ofnuclear power plant exercises. Milk, soil, andair sampling cartridges can easily be spiked withrealistic concentrations of radionuclide withoutviolating Nuclear Regulatory Commissionregulations or policy and with no significant"adiation hazard to exercise participants.Analysis and reporting of data from realisticsamples significantly improves motivation,attitude, moral, and overall preparedness of theindividuals participating in that phase of theexercise.

Since 1979 the State of Alabama hasparticipated in at least one nuclear power plantexercise per year involving the demonstration ofthe capability to analyze environmental samples.For the first five years, only non-radioactivesamples of milk, water, and other items wereanalyzed in a effort to establish the fact thatthe State could adequately evaluate realradioactivity at their field laboratory location.Since everyone knew that there was notiiing in thesamples, those who had participated for severalyears developed an attitude of flippancyconcerning the exercises. Other participantsexpressed concerns about the level of radiationexposure that they would receive from collecting,transporting, and analyzing environmental samplesduring a real incident. This problem washeightened when a federal observer suggested to aState participant that we might need to put up ashield to reduce the exposure from milk samplesthat were awaiting analysis.

In order to fully accomplish exerciseobjectives of demonstrating our capability ofbeing able to analyze samples and to initiateprotective actions on the basis of the U. S. Foodand Drug Administrations (FDA) Protective ActionGuides, the State of Alabama decided to utilizespiked or radioactive environmental samples forthe 1984 nuclear power plant exercise conductedwith the Joseph M. Farley plant operated by theAlabama Power Company. Both the State'sradiological laboratory and elements from the U.S. Environmental Protection Agency's radiological

laboratory from Montgomery participated in theexercise and analyzed the spiked samples at thefield location in Dothan, Alabama.

All parties involved acclaimed the use ofspiked samples as a positive method of improvingthe attitude and involvement of laboratorypersonnel in the exercise. Following the successduring the 1984 exercise, it was decided toutilize spiked samples for the November 1985exercise at Browns Ferry. Starting from thescenario, calculations were made of the amount ofiodine-131 that .would be anticipated in milk atthe nearest dairy to the plant, in air samplecartridges taken at the plume centerline at apoint on the boundary of the five m'le evacuationsector, and in surface soil samples takenadjacent to the plant following the mockaccident. The radioiodine to be used to spikethe samples was obtained from a local hospital'sradioactive waste storage, where it had beendiscarded due to insufficient remainingradioactivity for clinical use.

The capsules, which had originally eachcontained approximately 93 microcuries of iodine-131 had decayed to 0~r2 microcuries per capsule bythe exercise date of November 13, 1985. Thusthree capsules placed in a gallon of milk wouldproduce an approximate concentration of .17microcuries per liter which fit the scenario, andwould adequately demonstrate that the State couldrapidly analyze milk against the Food and DrugAdministration's emergency protective actionguide of .15 microcuries per liter (FDA, 1982).For the air sample cartridge, two capsules placedinside a previously use cartridge would have thesame analysis as a cartridge from a samplecollected according to the State's field airsampling procedures in a concentration of .69picocuries per cubic centimeter of air. Thespiked soil sample utilized two capsules ofiodine-131 placed in cesium contaminated soilthat had been previously collected as a sample ata known location of cesium contamination. Itcontained 0.4 micorcuries of iodine and 0.08microcuries of cesium-137 in 500 milliliters ofsoil. The air and soil samples were preparedseveral days prior to the exercise, while themilk sample was prepared the nx>rning of theexercise. The spiked samples were placed intoexercise play at the time they would normally be

119

120

collected during an ar.tual accident response.All parties who were expected to handle the

spiked samples were informed in advance that someof the samples utilized in the exercise wouldcontain real radipactivity. The radioactivitywould be consistent with the scenario, but wouldbe utilized in only nominal quantities. The needto inform all participants in advance of the useof real radioactivity can not be overemphasized.

The U. S. Nuclear Regulatory Commission(i\i?C) has established exempt quantities ofradioactive materials such that any person whoreceives, possesses, uses, or transfersindividual quantities below these amounts isexempt from the requirements for a license andcertain regulatory controls (NRC, 1982). Foriodine-131 that quantity is 1.0 microcurie.Nuclear Regulatory Commission regulationsprohibit the transfer of licensable material inexempt quantities for purposes of commercialdistribution without a specific license to do sofrom the Commission (NRC, 1984). However fornon-commercial distribution, as in the case ofthese spiked samples provided to exerciseparticipants, no license to distribute exemptquantities is required.

Radiation exposure problems in exercises arebest controlled by controlling the quantities ofmaterials involved. The exposure, rate from eachof the 0.2 nicrocurie iodine capsules utilized onNovember- 13, 1985 was .07 milliroentgen per hourat one inch or 0.5 mircoroentgen per hour at onefoot. Placing three capsules in a gallon of milkproduced a milk sample containing approximatelythe same concentration that the Food and DrugAdministration stated in their October 22, 1982Federal Register statement as being permissiblefor sale to the general public following anuclear power plant emergency (FDA, 1982).Utilizing a typical hand held survey instrument,'no external radiation can be measured from agallon of milk containing this concentration ofradloiodine. Extensive decontaminationprecautions for a possible spill duringtransportation or at the field laboratory werenot deemed necessary for this concentration ofradioactivity. While all samples must containless than the exempt quantity for theradionuclide involved, lessor amounts can oftenaccomplish the same purpose. Due to the natureof activities that occur during exercises, thequantities of radionuclides utilized in spikedsamples should be kept at the minimum necessaryto demonstrate laboratory capability and still berealistic for the accident scenario being used.

At the 1984 Farley exercise, during thepreexercise briefing, the use of realradioactivity in spiked samples was discussed.Likewise during the preexercise briefing for the1985 Browns Ferry exercise, the use of realradioactivity in both the environmental samplephase and the medical drill was discuss.Representatives of the various federal agenciesincluding the Nuclear Regulatory Commission werepresent at these briefings and none of theagencies present expressed any concerns ordisapproval of the intended usages of these smallamounts of actual radioactive material.

The medical drill, that was a part of the1985 exercise, utilized technetium 99m and

natural thorium as a source of low levelcontamination. Only inanimate objects werepurposefully contaminated. A simulated tornadowas to have struck the plant and blown somecontaminated objects from the plant. One of thecontaminated objects was to have struck andinjured a farmer approximately one mile from theplant. The injured farmer and contaminatedobjects were to be found by field monitoringteams. The magnitude of the potential hazardfrom this use of real contamination was to becontroiled by the limited amount of material tobe used and by the six hour hall life oftechnetium 99m. The benefits of the respondershaving actually seen some real contaminatedobjects that would at least make their metersrespond was thougnt to outweigh the risks of theuse of real radioactivity. The problems andsubsequent concerns following this medical drillwere a result of the following:

1. The utility's "Scenario Director" wasaware of, and had agreed in advance to the use ofthe technetium 99m, however the utility personnelwho participated in the drill had not beeninformed of the planed us of real radioactivity.

2. An error in calculation had been madeconcerning the anticipated concentration of thematerial being used. This resulted in the amountbeing used being five times the intended amount.

3. Due to a miscommunication concerning thetime to begin setting up the scene of the drill,the liquid Technetium did not have time to dryprior to the drill beginning, thus it was easilyremovable.

4. Utility participants were being informedvia their radio system at the plant that theradioactive materials were being simulated andthus ignored warnings to the contrary by theState representative at the scene.

5. The utility's radio system at the jointState-Utility monitoring location was unable tobroadcast on the day of the exercise, thus theycould hear the participants being incorrectlytold that the radioactivity was simulated, butwere unable to transmit the correct information.

As a result of the abo:e named problems, sixindividuals received detectable amounts oftechnetium contamination on their clothing and/orperson. One responder, who did not utilize theprotective gloves that were available to him,received a spot of contamination on the back ofthe middle finger of his right hand of no morethan 3 microcuries of technetium 99m. Theradiation exposure received by this individual asa result this incident was less than 5 percent ofthe allowable quarterly limit. Never the less,his finger was needlessly contaminated. Thehazard of this amount of external contaminationis best illustrated in relation to the fact thatthousands of people every day receive aninjection into their blood stream of up to 10,000times as !;;uch of this isotope as a part ofroutine medical diagnostic tests. This is not toimply that the magnitude of these accidentalexposures in niyway justify the mistakes thatoccurred. It does however imply, that themagnitude of the risk inherent with the use oflimited quantities of short lived materialsshould not be sufficient to prevent State andlocal emergency responders from receiving the

121

types of training experiences necessary to enablethem to function in the conditions such asexisted adjacent to Chernobyl in the daysfollowing the accident.

The events of this medical drill caused aprompt reconsideration of regulatory policy withregards to the use of real radioactivity duringexercises. All such usages, including the use ofspiked environmental samples were reconsidered byboth the State and the Nuclear RegulatoryCommission. In interim guidance furnished to theStates in March 1986, the Commission staffapproved the use of unsealed sources ofradioactive material for spiking environmental orfood samples for laboratory analysis and the useof sealed check sources and thorium lanternmantels for use in medical drills (NRC, 1986).The Alabama Department of Public Health hasincorporated this interim guidance into ourprogram policy.

A group of firefighters would not beconsidered qualified to be sent to a fire if noneof them had ever seen, been to, or helpedextinguish a real fire. For this reason firedepartments conduct controlled practice burns totrain firefighters. Likewise the legions ofState and local off site response personnel, whoare in theory being trained to mitigate theconsequences of a Chernobyl type accident, needto have a training experience that will preparethem for dealing with real radioactivity. Thereis no better way to provide that experience thanthrough the controlled use of very small amountsof short half life radioactive r>,aterials incarefully planned training exercises.

The use of spiked sanples has proven to beof positive value to nuclear power plantexercises for both the participants and thefederal observers. There is no better way todispel many of the myths that have beenassociated with what could be anticipated fromsuch samples. Having actually encounteredsamples containing levels of radioactivity thatcould be realistically anticipated to occur, ourpersonnel are much better prepared to handle realsamples when the actual event occurs.

REFERENCES

U. S. Food and Drug Administration, 1982,"Accidental Radioactive Contamination ofHuman Food and Animal Feeds, Recommendationfor State and Local Agencies," FederalRegister, Vol. hi, No. 205.

U. S. Nuclear Regulatory Commission, 1982, "Rulesof General Applicability to DomesticLicensing", 10 CFR 30, Part 30, Section30.18(a).

U. S. Nuclear Regulatory Commission, 1984, "Rulesof General Applicability to DomesticLicensing", 10 CFR 30, Sections 30.18(c)and 30.18(d).

U. S. Nuclear Regulatory Commission, 1986,"Guidance For Use of Radioactive MaterialsDuring Reactor Emergency .Exercises", datedMarch 21, 1986, unpublished.

ANS Topical Meetiaf o» Radiological Accents—Perspectives and Eaergeacy Plawiag

Integrated Emergency Response Programat the Savannah River Plant

Richard W. Benjamin

ABSTRACT. The Savannah River Plant (SRP) hasan emergency response system to evaluate rapidlythe possible consequences of a release of toxicor radioactive pollutants to the atmosphere or tothe onsite streams. This system is based upon astrong research and development technology, andcomputer simulation. The path, extent, andconcentration of a pollutant are predicted by acomputer system linked to field instrumentation,and monitored by a mobile laboratory and samplecollection teams. Field measurements, using anontoxic tracer gas, serve as emergency responseexercises and provide an extensive data base withwhich to test atmospheric dispersion codes.

INTRODUCTION

The Savannah River Plant (SRP), a governmentowned, du Pont operated production facility inSouth Carolina, produces special nuclear mate-rials for the United States Government andoperates virtually the entire nuclear fuel cyclein support of these operations. The wide varietyof processes at SRP makes necessary a broadly-based emergency response capability for evaluat-ing and responding to unplanned radioactive andtoxic chemical releases to the atmosphere orstreams on the site. The emergency responseprogram at SRP includes support in the event of arelease and basic research in meteorology, aque-ous processes, transport monitoring technology,and computer simulation. The strength of theprogram lies in the interaction between emergencyresponse and basic and applied research.

The research and development programsinclude experimental and theoretical dispersionmeteorology, dye and thermal studies in streamsand lakes, and Che development of a unique remoteenvironmental monitoring system. The programsprovide the basis and format for an effectiveemergency response capability, including regularemergency response exercises and effective24-hour emergency response personnel coverage.

This paper serves as a programmatic overviewand introduction to five papers [Addis (1986),Hayes (1986), Hoel (1986), Schubert (1986), andSigg (1986)] to be presented in the postersession at this conference, which give consider-ably more detail on particular aspects of theintegrated emergency response program.

REVIEW OF SR? FACILITIES ANDPOTENTIAL RELEASES

SRP is the third largest production site inthe DOE complex, comprising some 315 square milesof mostly wooded, gently rolling terrain. Thesite and its surroundings are shown in Figure 1.

f

CCORCI* " ^ ^ ^ s t j ^

c'"'''" IZ1."^,™"™1""""* k'* <M fuil >nd Tarfd P»bric«Tl«n0 f « M ' I tit Ion

A4mlnkfr«(i*n Art!

T/tf Savannah River Plant and Laboratoryare operated for the U.S. Departmentof Energy by E. 1. du Pont de Nemoitrt itCompany, "the primary purpote fj toproduce radioactive isotopej for nationaldefense and for peacetime applications.

i

\lntltlMI

'•Mm,

±JS SOUTH ZR

\ x s.c! \\ V . \ • COLUHtlA }>~S~'

\ AUtUSTA*Ml pi|vCP> ^ J

UIZYFIGURE 1. The Savannah River Plantsit*. The

Inset Shows SIP's Regional Location

123

124

It is bordered on the southwest by the SavannahRiver, which provides cooling water for the fouroperating production reactors. The reactors areheavy-water moderated and cooled; they operate atnear atmospheric pressure and at a temperaturebelow the boiling point of heavy water. Twolarge man-made lakes are on site to help cool thethermal effluents of two of the reactors. Inaddition to the reactors, there are two largechemical reprocessing plants with associatedhigh-level waste tank farms, a tritium processingand handling facility, and a fuel fabricationfacility. A high-level waste solidificationfacility and a naval fuels fabrication plant areunder construction. All of these facilitiesprovide the possibility for unplanned releaseswhich would require emergency response procedures.

Radioactive effluents which might have to bemonitored in the event of an unplanned releaseare, e.g., tritium (elemental or oxide forms), ,gamma ray emitting aerosols, radioiodine, transu-ranic aerosols, and noble gases. The most likelyrelease from SRP facilities which could travel ;offsite is tritium; either from the tritiumprocessing facilities or as the result of areactor moderator spill. Over the years therehave been several tritium releases too small tobe a health hazard either on or off site. Theyhave, however, served to improve tritium confine-ment and to develop appropriate tracking andmonitoring capabilities. As a result, confine-ment of tritium has improved dramatically andmethods to track and monitor any potentialrelease have been developed and demonstrated.There is also the capability to monitor toxicchemicals, but except for limited quantities ofchlorine gas, there are no toxic chemicals onsite which could represent a significant hazard.

THE ENVIRONMENTAL TRANSPORT GROUP

The Environmental Transport Group (ETG) hasthe responsibility for the prompt characteriza-tion of atmospheric and aqueous releases from anyof the SRP facilities. The primary aim is toprovide the proper decision makers with therelease location and the predicted dosimetry andexposure information at each stage of an emer-gency. These decision-makers may be, forexample, the shift supervisor at the earlieststages of a release, DOE and du Pont managers,and the monitoring teams seeking to evaluate theenvironmental effects of the release. Rapidresponse to a release at SRP is possible becauseall components of the response system are underthe control of one group, the EnvironmentalTransport Group. The system components, person-nel and facilities, include meteorology, dosim-etry, the dedicated emergency response computers,and the atmospheric monitoring instrumentation.

ETG has a staff of fourteen professionalsand eight cechnicians. Included in this groupare five meteorologists, an oceanographer, twohealth physicists, a nuclear chemist, a computersystems manager, and a small maintenance team toservice the widespread meteorological instrumen-tation on and off site. At least two meteorolo-gists and a supervisor are on call at all timesfor emergency operations, and additional supportpersonnel are called in as required.

THE EMERGENCY RESPONSE SYSTEM

The emergency response system is a highlyautomated, real time system which consists of twomajor components, the WIND system .and the TRACmobile laboratory. The Weather INforiration andDisplay (WIND) system is used for data acquisi-tion and release consequence calculations forboth atmospheric and aquatic releaser. TheTracking Radioactive Atmospheric Contaminants(TRAC) mobile laboratory is used for plume track-ing and real-time radioactive monitoring in thefield. In addition, real-time sample collectionteams provide post-ialease contamination evalua-tion. These components are described in consid-erable detail in Addis (1986), Hayes (1986), Hoel(1986), Schubert (1986), and Sigg (1986).

The WIND computer system is used sitewide.Every operating area has a WIND system computerterminal in or near its control room. The systemis user-friendly with all of the necessary emer-gency response codes on a general menu, and thesystem is used routinely by the site healthprotection technicians. There is also a WINDsystem terminal in the Emergency OperationsCenter (EOC), and an automated alarm system basedon stack monitor release limits will be installedin the 1TOC in the near future.

All data acquisition for the WIND system isachieved through the Remote EnvironmentalMonitoring System (REMS) (Schubert, 1986), whichprovided the following data:• Real-time turbulence quality meteorological

data are collected from the eight onsitetowers and from the WJBF-TV tower just offsite

• Regional meteorological data are obtained fromthe National Weather Service over theAutomated Field Operations Services (AFOS)computer

• Monitors in SRF streams and the Savannah Riverprovide temperature and flow data for aqueoustransport

• Real-time source term information and plantboundary concentrations "re obtained from thestack and perimeter monitors, respectively

These data are used to provide forecasts of theexposure pathway from an atmospheric release as afunction of time, as well as radioactive doseestimar.es. For aqueous releases, stream andriver concentrations as a function of time can bedetermined.

RESEARCH PIOCRAMS

Several basic and applied research anddevelopment programs are in progress, and all areaimed at improving, quantifying, and extendingthe emergency response capabilities of SRP andthe national DOE production complex. Researchprograms also provide the assurance of recruitingand maintaining a highly competent staff ofscientists and engineers. These efforts aresupported in part by the Office of Health andEnvironmental Research of the Department ofEnergy.

125

Typical of these programs is the AtmosphericResearch Program. This program combines:• Theoretical studies in atmospheric

dispersion,• Experimental field studies aimed at providing

a statistically sound data base with which totest atmospheric dispersion models,

• Instrument development to provide the means tomeasure meteorological parameters, radioactiveatmospheric contaminants and tracer gases on areal-time basis

The heart of the Atmospheric Research Program isthe Mesoscale Atmospheric Transport Studies(MATS) effort. Its focus is to provide experi-mental data to test the predictions of thedispersion models. This is done through the useof an experimentally released tracer gas (SFg),and by tracking the routine small releases fromSRP's manufacturing facilities. The data basesfrom MATS and several other iiseteorologicalstudies were used as the basis for the SecondModel Evaluation Workshop in 1984 [Weber (1985)a,Weber (1985)b, and Kurseja (1985)]. Radioactiveemissions of Kr-35 tritium, and Ar-41 in additionto nonradioactive He-3 are continually used forfurther dispersion model evaluation and improve-ment. The development of highly sensitivemeasurement technology has extended these tracerexperiments from on site out to several hundredkilometers.

Studies begun with MATS are being extendedin an inter-laboratory program called the STableAtmospheric Boundary Layer Experiment (STABLE).The objective of STABLE is to develop an under-standing of turbulence and dispersion in thestable (nocturnal) boundary layer, a conditionwhich is not well-understood and which would beexpected to lead to the potentially highest off-site doses in the event of a release. STABLEwill include study of a five year data base ofturbulence measurements, field experiments, andextensive model evaluation and development.

The Aqueous Research Program of ETG isprimarily concerned with aqueous transportprocesses in the onsite streams and in theSavannah River, and in thermal processes in theonsite lakes and ponds. The major transportconcerns are liquid releases, such as tritiatedwater and hazardous chemicals, and transport ofmaterials adsorbed, such as Cs-137, on streamsediments. A one-dimensional model is used topredict travel times, maximum concentrations, andconcentration distributions as a function of timeat downstream and river locations for pollutantreleases. Stream velocities and dispersioncoefficients that were needed in the model weredetermined from dye studies done in each of theonsite streams. Models are being improved as newinformation from studies of Cs-137 and uraniumtransport in onsite streams becomes available.New instrumentation installed in onsite streamswill provide real time data on stream/river flowfor input to the model. Discharges to onsitestreams typically take at least one day to traveloffsite and do not require the urgency of anatmospheric release which can cross the SRPboundary in an hour.

CONCLUSIONS

The integrated emergency response programat SRP provides rapid response to an unplannedrelease. The staff who are active during unplan-ned releases are the same people who conduct thebasic and applied research programs. Many of therequirements for emergency response are the sameas those required for the experimental researchprograms, e.g. operable equipment, good communi-cations, good forecasting, and skilled datataking. Each experiment is conducted as alimited emergency response exercise so thateveryone gets plenty of practice. In addition,the participants are knowledgeable about both theresearch and the emergency response actions, andable to suggest improvements in both areas. Theinteraction has proven to ba very valuable.

ACKNOWLEDGMENT

The information contained in this articlewas developed during the course of work underContract No. DE-AC09-76SR00001 with Che U.S.Department of Energy.

REFERENCES

Addis, R. P., R. J. Kurzeja, and A. H. Weber,1986, "Improving Emergency Response throughField Exercises," This Conference.

Hayes, D. W., 1986, "Aquatic Emergency ResponseModel at the Savannah River Plant," ThisConference.

Hoei, D. D., 1986, "Savannah River PlantEmergency Response: Operations andExercises," This Conference.

Schubert, J. F., 1(»86, "Savannah River PlantRemote Environmental Monitoring System," ThisConference.

Sigg, R. A., 1986, "Mobile Laboratory for NearReal-time Response to Radioactive Releases,"This Conference.

Weber, A. H. and A. J. Garrett, 1985, Proceedingsof the DOE/AMS Air Pollution Model EvaluationWorkshop, Kiawah, South Carolina, October23-26, 1984, Vol. 1, Papers Presented, USDOEReport DP-1701-1.

Weber, A. H. and R. J. Kurzeja, 1985, Proceedingsof the DOE/AMS Air Pollution Model EvaluationWorkshop, Kiawah, South Carolina, October23-26, 1984, Vol. 2, Statistical Analysis ofResults, USDOE Report DP-1701-2.

Kurzeja, R. J. and A. H. Weber, 1985, Proceedingsof the DOE/AMS Air Pollution Model EvaluationWorkshop, Kiawah, South Carolina, October23-26, 1984, Vol. 3, Summary Conclusions, andRecommendations, UDSOE Report DP-1701-3.

ANS Topical Meeting OH Radiological Accidents-Perspectives and Emergency Planning

Savannah River Plant Emergency Response:Operations and Exercises

Doris D. Hoel

ABSTRACT. The Savannah River Plant is acomplete nuclear complex with reactors, fuelfabrication, and fuel reprocessing onsite. Dueto the potential for unplanned releases, anemergency response program has been developed toprovide information onirelease consequences. Theautomated system was developed initially as areal-time site-specifLc system for the predictionand analysis of atmosnheric and aquatic releases.However, it is flexible and has application forincidents at offsite facilities as well. Theemergency response system, its components, equip-ment and capabilities, and the emergency responseprogram, procedures and activities are presented.This includes the special case of joint responsewith Georgia Power Company's Vogtle ElectricGenerating Plant.

HmODUCTIOK

The Savannah River Plant (SRP) is located inSouthwestern South Carolina, 25 miles southeastof Augusta, Georgia, along the Savannah River-(Figure 1). It is a government owned, contractoroperated facility which produces nuclear mate-rials for defense purposes. The complex includesfour operating nuclear materials productionreactors, a fuel and target fabrication area, twochemical separations facilities, high-levelradioactive waste management operations, a low-level waste burial ground, and (under construc-tion) a new facility to provide fuel for the U.S.Navy's nuclear powered submarines. During SRP'sthirty-five years of operation, an emergencypreparedness program has been developed, whichincludes response to nuclear incidents, both onthe site and in the southeastern United States.

The following discussion describes emergencyresponse at SRP. It includes components of thesystem, equipment and capabilities, responsibili-ties, procedures and activities during an emer-gency, response exercises, and incidents on siteand in the southeastern United States.

DESCRIPTION

A site-specific, automated, real-timeemergency response system has been developed atthe Savannah River Laboratory (SRL) to predict,measure, and analyze the consequences of any

M Finland f.ifit0 P o w St.iisnA (a*.nn*> flkar LaboraMiv and

Attntnjttralion/

® MaMwolaaiealT

The Savannah River Plant and Laboratoryare operated tor the US. Deparlmtntof Energy by E. I. dit Pont tie NemoursCompany. The primary purpose It toproduce radioactive isotopes for nationaldefense and for peacetime applications.

FIGURE 1. Savannah River Plant

unplanned radionuclide release as a result ofoperations at SRP. There are two major componentsof the system. One is the Weather INformationand Display (WIND) (Garrett et al., 1983) systemfor data acquisition and calculation of predictedrelease consequences. The other is the TrackingRadioactive Atmospheric Contaminants (TRAC)mobile laboratory (Sigg, 1985) capable of nearreal-time monitoring and plume tracking. Inaddition, real time sample collection teams areavailable for post incident radioactive plume

127

128

tracking. Standard environmental monitoringanalytical techniques have also been adapted toprovide support for aquatic transport of a radio-active or hazardous waste release.

The WIND System consists of eight meteor-ology towers for real-time onsite data gathering,an Automated Field Operation and Services (AFOS)National Weather Service (NWS) computer foracquisition of regional meteorological data(Figure 2), two mini-computers, and a network ofterminals (Figure 3) for data display and calcu-lation of predicted consequences of either anatmospheric or aquatic release. The WIND Systemcomputers are Digital Equipment Corporation VAXH/780 and VAX 11/750 super mini-computers.Duplicate sets of emergency response codes arekept current on both systems for backup purposes,and data collecting and archiving is automati-cally switched from one computer to the other inthe event of computer malfunction. In additionto meteorological -iata, stream data from onsitestreams and the Savannah River and stack monitor-ing data from the reactors and separations areasare collected on the VAX computers through theRemote Environmental Monitoring System (REMS).These data provide stream flow and source terminformation. WIND system data sources are shownin Figure 4. The network of terminals is locatedthroughout the plantsite. Terminals are also inthe homes of appropriate personnel for responseduring off shift hours.

The TRAC (Figure 5) mobile laboratory isdesigned for plume tracking and near real-timedata collection and analysis during an atmosphericrelease. Its primary purpose is to measurelow-level air concentrations of specific radio-nuclides as they move off the plant site. Duringan emergency, the TRAC is used to aid in deter-mining dose estimates, to aid in determiningadjustments to trajectory calculations, to pro-vide release analysis, and meteorological codeverification. Current equipment includes moni-tors to detect radioactive plumes, gamma andtransuranic aerosols, radioiodines, tritium formsand noble gases, a mini-computer for data analyz-ing and archiving, and an engine/generator forelectric power in the field.

FIGURE 3. WIND System Graphics Terminalsfor Display of Data and Calcula-tions 1 Results Located ThroughoutSRP/SRL and in the Homes ofEmergency Response Personnel

FIGURE 4. WIND Emergency Response SystemData Sources

FIGURE 5. Tracking Radioactive AcmosphericContaminants (TRAC) System

FIGURE 2. The AFOS MWS Computer Used for RetrievingLocal and Regional Meteorological Data

129

RESULTS

Onsite Incident ResponseAn emergency situation on the plant site in

which SRP's Emergency Operating Center (EOC) isactivated requires immediate notification ofsupport and management personnel from SRP, SRL,Department of Energy Savannah River (DOE-SR),and Wackenhut Services, Inc. (WSI). A 24-hournotification procedure has been established inwhich required emergency response personnel arecontacted. All emergency response personnel havebeen issued pagecoms for contact when away fromtheir office daring work hours or away from homeduring offshift. All pagecoms can be accessedfrom on or off plant. In addition, three groupsof all-call pagecoms have been issued to primaryemergency response personnel. The all-callpagecom system is used by EOC communicationspersonnel to simultaneously notify groups of 29persons of an EOC activation. The all-callsystem is used in addition to phone notification.Secondary notification for support is made byphone through appropriate departments.

Response time in an emergency depends upontime of notification. The EOC is staffed withinminutes during regular shift, in 30-40 minutesduring offshift, weekends, and holidays. SRLpersonnel involved with the WIND System provideimmediate response in the Weather Center AnalysisLaboratory (WCAL), and in the EOC in about 10minutes during regular shift; TRAC is operationalin approximately 30 minutes; and sampling vansare deployed in 15-30 minutes. During offshift,SRL personnel may respond immediately from homeusing WIND System terminals. EOC/WCAL responsetime is 30-40 minutes, TRAC response is about1.5 hours and sampling van deployment is 45-60minutes. Increased response time during offshiftis the result of travel time to the plant fromoffsite.

The EOC/WCAL. personnel have the responsibility

• to provide weather information, meteorologicaldata, stream data, WIND System trajectory, anddose and deposition calculations

• to make adjustments to calculations for plumerise and mixing depth

o to interact with health protection personnel,area supervision, and technical personnelknowledgeable of the process involved in theincident

• to obtain source term and code input

• to communicate with TRAC and the sampling vansand direct monitoring team activities

» to assist healch protection with theirmonitoring efforts

'» to provide data assessment and analysis, andprovide data and interpreted results tomanagement and DOE

Figure 6 shows the information flow and interac-tions during an onaite incident. Finally, afterthe incident SRL response personnel provide athorough analysis of all significant releases.The analysis includes trajectory, concentrations,doses, and health effects for both on and offsitepopulations. The response program continues toimprove through research and regular WIND Systemand plantwide exercises.

FIGURE (i. Onsite Emergency Response:Information Flow and Interactions

Offsite Incident ResponseIn the event of a major incident involving

nuclear materials in the southeastern U.S.,DOE-SR is responsible for organizing and coordi-nating the radiological monitoring and assessmentactivities of. Federal agencies outside of thesite or incident boundary (USD0E, SR-503). Theemergency response system at SRL/SRP, althoughdeveloped as a site specific system, is a majorresource for DOE-SR's Region 3 RadiologicalAssistance Program (RAP).

Under normal circumstances, a request forSRL's assistance would come through DOE-SR'sOffice of External Affairs. Depending upon thenature of the incident and the amount of assist-ance required, response would be handled at SRPor may require travel to the incident area. Ineither case, the responsibilities for onsiteresponse apply. If travel is required, initialWIND System calculations can be made at SRL bypersonnel who would remain at SRL to support SRP,while designated response personnel are enrouteto the incident urea. A graphics terminal with abuilt-in dial-in modem and hard copy unit (aswell as a portable backup terminal with printcapability) would be transported to the incidentarea for WIND System calculations. The TRACvehicle remains in standby mode for either on oroffsite incident response. As with onsite emer-gencies, all calculations, results, analyses, andrecommendations concerning offsite impact arereported to DOE. DOE is responsible for approvaland forwarding of information to the agencyrequesting assistance, and the cognizant Federalagency (CFA). Offsite response is shown inFigure 7.

130

Support is Provided

At

WfrVO SyiumIThrougta OiaMn Mrtdctnl

Areas of Support

* Pradietiva Capabilitiaa

• Raid Monitoring

> Data Evaluation and Assattmant

FIGURE 7. Offsite Emergency Response

On March 6-8, 1984, Savannah River partici-pated in the St. Lucie Federal Nuclear EmergencyResponse Exercise in St. Lucie, Florida. SRL/SRPpersonnel utilized the WIND System, TRAC vehicle,and sampling vans to provide assistance in theareas of predictive capability, field monitoring,and data evaluation and assessment. This was thefirst major field exercise which provided aframework for a coordinated Federal response Co amajor incident.

A special case of offsite response (USDOE,SR 402.1) has also been established with theGeorgia Power Company (GPC). GPC's VogtleElectric Generating Plant is located just acrossthe Savannah River in Georgia (see Figure 1).Due to proximity, DOE-SR and GPC have establisheda memorandum of agreement to provide for planningand responding to emergencies originating ateither facility. SRP/SRL will respond to aVogtle incident in the same manner as an onsiteincident. SRP will be responsible for all moni-toring and assessment and protective action onSRP, provide monitoring in South Carolina asrequested, provide meteorological data to GPC,and advise GPC and the States of Georgia andSouth Carolina concerning the incident.Responses are reciprocol for an SRP incident.To facilitate response, direct communicationlines between the two facilities and 24-hour-per-day points of contact have been established.Both facilities participate jointly in emergencyexercises, the first of which was the Vogtlelicensing exercise in April 1986.

AcknowledgmentThe information contained in this article

was developed during the course of work underContract No. DE-AC09-76SROO0O1 with the U.S.Department of Energy.

RE7KKEMCES

1. A. J. Garrett, M. R. Buckner, andR. A. Mueller, 1983, "The Weather Informationand Display Emergency Response System."Nuclear Technology, 60, 50.

2. R. A. Sigg, 19b5, "A Mobile Laboratory ForNear Real-Time Measurements Of Very Low-LevelRadioactivity." Proceedings Of The Fifth DOEEnvironmental Protection Information MeetingHeld In Albuquerque, New Mexico, November5-8, 1984, CONF-841187, 629.

3. Federal Radiological Monitoring AndAssessment Plan. SR 503, U.S. Department OfEnergy. Radiological Assistance Region-3,Savannah River Operations Office.

4. Emergency Management Plans Vogtle ElectricGenerating Plant Response Guide. SR 402.1,U.S. Department Of Energy, Savannah RiverOperations Office.

ANS Topical Meeting on Radiological Accidents-Perspectives and Emergency Planning

A Mobile Laboratory for Near-Real-Time Responseto Radioactive Releases

R. A. Sigg

ABSTRACT. A mobile labnratory is improvingradiological emergency response and environmentalmonitoring capabilities at the Savannah RiverPlant. The laboratory can monitor low concentra-tion levels of radionuclides and can rapidlyconfirm the location and radionuclide compositionof a downwind plume. The analysis system devel-oped for the laboratory includes radionuclide-specific monitors for tritium forms (HT and HTO),gamma-ray emitting aerosols, volatile radio-iodine, transuranic aerosols, and noble gases.The monitors can detect radionuclides transportedby the atmosphere to offsite population centersat concentrations well below those consideredhazardous. An analyzer for a passive chemicaltracer gas (sulfur hexafluoride, SFg) also isonboard. The diverse instruments, their goodsensitivity and fast response, and the ability tooperate while in transit, make this a uniquelaboratory.

INTRODUCTION

The diverse nuclear facilities at theSavannah River Plant (SRP) have a very lowpotential for the release of radionuclides atlevels that could create an offsite hazard.However, in keeping with the philosophy thatprotection of the public and the environment isessential, the Savannah River site has a contin-uing research and development program to improveits radiological emergency response and environ-mental monitoring capabilities. A mobile labora-tory (Sigg, 1985) (Figure 1), to monitor lowconcentration levels of radionuclides and torapidly confirm the location and radionuclidecomposition of a downwind plume, has beendeveloped as a part of this program.

The large variety of radionuclides producedat the site prompted the development of a numberof specialized atmospheric collection and analy-sis systems for the Tracking RadioactiveAtmospheric Contaminants (TRAC) mobile labora-tory (Sigg, 1985). These include radionuclide-specific monitors for tritium forms, gamma-rayemitting aerosols, volatile radioiodine, transu-ranic aerosols, and noble gases. (HT and HTO areused generically to represent the several

Critium-containing elemental and oxide chemicaiforms of hydrogen isotopes.) The monitors candetect radionuclides, transported by the atmos-phere to offsite population centers, at concen-trations well below those considered hazardous.Their measurements can address legitimate publicconcerns following any abnormal radionucliderelease. An analyzer for a passive chemicaltracer gas (sulfur hexafluoride, SFg) also isonboard. The diverse instruments, along withtheir good sensitivity and fast time response,make this a unique laboratory. The ability tooperate while in transit is an improvement overmost mobile facilities. Some of the importantfunctions of TRAC are to• Provide prompt analyses in the field instead

of delayed analyses at site laboratories• Obtain timely concentration results in the

event of an abnormal release in order to- assure the plume has been located andprofiled before returning from the field

- provide information almost immediately tothe onsite Emergency Operating Center (EOC)

o Test and refine emergency response procedureso Provide data to evaluate rwJ improve

atmospheric transport models by

FIGURE 1. Tracking Radioactive AtaosphericContaminants (TRAC) Laboratory

131

132

- monitoring small releases from SRP's normaloperations

- measuring releases of SF6 tracer gas tosimulate emergency response operations

- gathering data during emergency responseexercises.

PLUME LOCATION

The extensive monitoring capabilities cf theSavannah River Plant can be applied most effec-tively when personnel have downwind plume loca-tion information to assist sample collectionoperations. Three plume locating methods areavailable to guide sampling operations.

TRAC initially receives forecast plumetrajectories and concentrations by radio from theonsite Weather Center Analysis Laboratory(Garrett, 1983).

A direction-sensitive gamma-ray monitor(Figure 2) determines plume locations relative tothe vehicle heading. This onboard plume monitoruses twelve large Nal(Tl) spectrometers similarin volume to those used in some aerial surveyingapplications. The detector array is shielded onthe bottom and sides to minimize interferencesfrom naturally occurring radionuclides found insoils and rocks. Successive short (~15 sec)counting intervals yield a time series histogramas the laboratory transects a plume (Figure 3).Low plume concentrations of short lived ^Ar(2 hr) from normal reactor operations have beendetected as far as 50 kilometers downwind of anSRP reactor.'

A real time analyzer for low concentrationsof SF6 tracer is being tested (Milham, 1986).SF6, intentionally released from onsite facili-ties, was detected 50 kilometers downwind (Figure4) in tests of our real-time plume locatingcapabilities. When SF6 and Ar are releasedsimultaneously from two different points on theplant, measurements downwind show the plumetrajectories are nearly parallel. Location ofeither plume downwind allows us to deduce thelocation of the other. In the event of an inad-vertent atmospheric radioactivity release, 1*1Arplumes could be useful in locating plumes fromanother facility. The possibility of simultane-ously releasing tracer gas at the effluent pointis also being explored; in this case, detectionof the tracer gas would coincide with theradioactive plume location.

S 0.6

5

I5 0.4

i i i i r i i

S 12 16

Distance Across Plume (Km)

FIGURE 3. Crossvind Transect of an Ar-41 Pluaeat the SRP Boundary

FIGURE 2. Plume Monitor Nal(T«) Detector Array

Distinct Across Plumes, Km

FIGURE 4. Successive Crosswind Transects of anSFg PIuse at 25 Km Downwind

COLLECTION AID ANALYSIS SYSTEMS

Specialized monitoring systems collect andmeasure radionuclides from large volumes of air.Concentrations less by a factor of 10~s than themaximum permissible concentration (MPC) limitsfor the general population in uncontrolled areascan be detected for some nuclides (e.g., 3H and99mjcj> Sensitivities are less for monitoringother radionuclides, but all radionuclides ofinterest can be detected at or below MPC limitsin near-real time. These are significantimprovements over previous field measurementcapabilities.

TRITIUM FORMS

A tritium forms monitor (Figure 5) collectsand analyzes tritium either as moisture (HTO) oras elemental gas (HT). Adsorption techniquesconcentrate individual forms from large volumesof air, and desorbed sample material is analyzedby liquid scintillation spectrometry. The com-plete process takes about forty minutes. The

133

sensitivity (~20 pCi/m ) permits detection of anyabnormal release, even 40 km downwind. Figure 6shows separate plumes that, were detected onsitefor HT and HTO from different facilities.

iiv Peteciion

4] i •

i <1 ")IN>-

t,on Cou

- | !

l|

nun*

isLow-Level Germanium

Detector

FIGURE 5. Real-Time Tritiint Forms Monitor

: ii*^

i•

%

i . i . - %

Djiunce Acrott Plume (km)

FIGURE 6. Crosswind Transect of HT and HTOPlumes from Two Separate SRPFacilities

GAMMA-RAY AEROSOLS

A monitor collects aerosols from air flowingthrough a filter at twenty-five cubic meters/minute. In about twenty minutes, post-collectiongermanium spectrometry (Figure 7) identifiesindividual gamma-ray emitting radionuclides atconcentrations ranging from 1/15 to better thanlCT5 of MPC (for " T c , 132Te, and m C e )without waiting for normal radon daughter inter-ferences to decay. Radioactive debris from theChernobyl accident was easily detected in thesoutheastern United States by the monitor(Figure 8).

FIGURE 7. Gamma-Ray Aerosol Monitor andChernobyl Debris Aerosol Spectrum

Sample Collection

HEPAPrelilter

Activity Detection

Germanium Gammi-Ray Spectrometry

Radionuclkle SpecificNear Realtime

Low Level

Spectrum of ChernobylRadioiodine Debris

'L-?'~1-E.~'.ui:

1.000

migy IktVI

FIGURE 8. Volatile Radioiodine Monitor anil aGaana-Ray Spectrum of ChernobylRadioiodine Debris

134

RADIOIODIHE

A charcoal canister concentrates radioiodinefrom an air stream and then gamma-ray spectrom-etry (Figure 7) evaluates the canister activity.The sensitivity for 13iI is 1/3 of MPC whensamples are collected for ten minutes and countedfor ten minutes. 131I from the Chernobyl acci-dent was also easily detected in the vicinity ofthe SRP.

TRAMSUIAHCS

A Teflon*-based filter medium collectsaerosols from an air stream (Figure 9). Theaerosols are deposited at the surface of themedia, allowing alpha spectrometry by silicondiodes to distinguish transuranics from radondaughters. Results with detection limits equalto the MPC of 239Pu are produced in an hour.

SwipltColltction

SLAlpha

Spectromitry

.-. FluoroportFilttr

oAil Flow

(1000 litcri/min)

— Ion ImplinttdSiliconSptctrometm

(20 lorrl

FIGURE 9. Transuranic Aerosol Monitor and aBackground Alpha Spectrum

NOBLE GASES

About 6 cubic meters of air are analyzed fornoble gas activities. Filtered air is compressedto about 170 atmospheres to concentrate the airin a pressure vessel. A germanium spectrometerin the pressure vessel annulus (Figure 10)analyzes noble gas activities.

APPLICATION AMD SUPPOET SYSTEMS

The above laboratory systems are supportedby a real-time computer and a Loran-C navigationsystem.

itnium DtieciorLiquid Niuo9fn Dtwar

FIGURE 10. Noble Gaa Monitor Counting Apparatus

The TRAC mobile laboratory supports onsiteenvironmental research programs through atmos-pheric transport model validation studies and bymonitoring for trace concentrations of radio-nuclides wir.h the potential for release duringnormal SRP operations. The laboratory monitoredand easily detected radioactive debris from theChernobyl reactor accident in air samplescollected in population centers within 160 km ofSRP. The TRAC mobile laboratory also partici-pated in the Federal Emergency Management Agency(FEMA)-sponaored St. Lucie Federal FieldExercise, as well as a joint exercise between SRPand the states of South Carolina and Georgia.TRAC's participation in these exercises andresearch programs have spinoff benefits thatinclude testing and improving sampling strate-gies, logistics, communications, and personneltraining (Addis, 1986).

ACCTOWLKDCMHT

The information contained in this articlewas developed during the course of work underContract No. DE-AC09-76SR0O0O1 with the U.S.Department of Energy.

REFERENCES

Addis, R. P., R. J. Kurzeja, and A. H. Weber,1986, "Improving Emergency Response ThroughField Exercises," This Conference.

Garrett, A. J., M. R. Buckner, and R. A. Mueller,1983, "The Weather Information and DisplayEmergency Response System." Nucl. Tech.60(1):50-59.

Milham, R. C , A. H. Weber, and R. A. Sigg,1986, "Mobile Measurement of an AtmosphericTracer." Accepted for Proceedings of theFifth Joint Conference on Applications ofAir Pollution Meteorology with APCA, ChapelHill, NC, November 18-21, 1986.

Sigg, R. A., 1985, "A Mobile Laboratory for NearReal-Time Measurements of Very Low-LevelRadioactivity." Proceedings of the FifthDOE Environmental Protection InformationMeeting, Albuquerque, NM, CONF-341187,November 5-8, 1984.

ANS Topical Meeting on Radiological Accidents—Perspectives and Emergency Planning

Improving Emergency Response Through Field ExercisesR. P. Addis, R. J. Kurzeja, and A. H. Weber

ABSTRACT. Exercises using a tracer gas Cosimulate the accidental release of gaseous efflu-ent to the atmosphere have been conducted at theSavannah River Plant since 1983. Although thetracer released is a passive gas, the operationsare similar to those required for unplannedreleases of radionuclides. These exercises,called Mesoscale Atmospheric Transport Studies(MATS), have produced many tangible benefits,including efficient, dependable performance ofsampling teams, refinement and documentation ofrequired skills and logistics, and better windforecasting. The exercises have provided datarequired to evaluate the performance of disper-sion models, and identified areas requiringimprovement. For example, the operationalGaussian model was found to accurately predictpuff centerline and puff width, but to over-predict the maximum concentration. The relativemerits of moving samplers versus stationarysamplers under various atmospheric conditionsare also discussed.

Array ol 2B SF,Sequential Samplers

SF, PuttSFt Release (15 Minutesfrom 60 m Slack)

MATS Observation Roads

FIGURE 1. A schematic map of the Savannah RiverPlant, the ring of road* (called theHATS arc) on which air sampling i*conducted, and a typical configurationof moving and stationary aaatplera

INTRODUCTION

Fifteen-minute stack releases of the tracergas, sulfur hexafluoride (SFg), are used tosimulate the unplannad release of radionuclidesfrom facilities at the Savannah River Plant(SRP). Atmospheric dispersion of the tracer gasis detected and measured by both fixed and movingsampling systems.

The goals of these exercises are to:• Determine atmospheric dispersion model

accuracy* Provide an operational test of the capability

to respond during emergencieso Provide information on crosswind and downwind

spread of gaseous effluentA downwind distance of 30 km was chosen for

monitoring because it corresponds to the distancebetween the release point and the major popula-tion centers of Augusta, Georgia, and Aiken,South Carolina. Thirty kilometers also corre-sponds geographically witu an approximatelycircular series of highways where vehicles candeploy samplers alongside the roads. A schematicmap in Figure 1 shows the SRP, the ring of roadsencircling the plant, and the configuration for atypical tracer experiment (MATS).

DESCRIPTION

During a typical exercise, personnel areadvised of the release time and are organizedinto three sampling teams. Two of the teamsload twenty-eight programmable, sequential airsamplers into vans. The third team operates amobile laboratory called Tracking RadioactiveAtmospheric Contaminants (TRAC, Sigg, 1985) whichcontains a continuous operating SFg samplingsystem and other atmospheric monitoring devices.When the three teams leave the site, they receivea puff trajectory forecast. The teams then driveto the intersection of the puff centerline andthe sampling arc. While the vans are travelingto the intersection point, meteorologists in theweather center study the surface pressure fore-casts and observations, balloon soundings, andmeasurements from eight instrumented meteoro-logical towers to refine the accuracy of theinitial puff impact-zone forecast. Theseinstrumented towers are located near each produc-tion facility within the plant boundaries. Thespacing of the drop points along the roadside forthe sequential samplers is based on turbulencemeasurements from tower wind instruments.Deployment instructions are then relayed from

135

136

the weather center to the vans.As the vans deploy samplers, the TRAC

vehicle with its mobile detection capabilitytravels crosswind along the sampling arc todetermine the puff's arrival time and actuallocation. The SFg concentration data are storedon computer tapes and discs in real time. SinceTRAC is in radio communication with the WeatherCenter, information can be passed to advise thefixed sampling teams to reprogram or to movesamplers to a better location. This procedurecan be extended to greater distances downwind asthe puff moves along its path. A computer pro-gram determines the 3tart time and optimizes thesampling interval for the 10 sequential motor-driven syringes used to collect whole air samples.

After the sequential samplers are retrievedfrom the roadside and transported back to thelaboratory, the SF6 concentrations are determinedby gas chromatography. The TRAC SFg concentra-tions are also transferred to the weather centercomputer system. Calibration gas standards areavailable in four concentrations to ensure theaccuracy of measurements of both the fixedsequential samplers and the mobile continuoussampler. The concentration data are plotted intime-space cross-sections and compared withmodels used during emergency response.

RESULTS AMD CONCLUSIONS

The exercises have provided the datarequired to determine the performance of thedispersion forecast models. The operationalemergency response Gaussian model, PUFF/PLUME,accurately predicts the width and centerlinelocation of the puff, but overpredicts thecenterline maximum concentration on average by5.A times. A sequential Gaussian model, 2DPUF,also used in emergency response, more closelypredicts the centerline maximum concentration -only overpredicting by 1.5 times.

The logistics of sampler deployment havebeen improved through heuristically optimizingthe spatial distribution of the samplers, thesampling intervals, and perfecting communicationsand deployment timing between the meteorologicalforecasters and field crews. The importance ofaccurate and timely forecasts of transport windspeeds and directions became evident when thefield exercises were begun. For example, atransport distance of 30 km results in a meantransport time of about 2 hours for a 15 km/hrwind. Thus, a forecast error of only 10 degreesin direction and 3 km/hr in speed for the trans-port wind results in a centerline location errorof 5 km at the samplers and a timing error ofabout 30 minutes.

Another practical benefit of the exerciseswas the testing of sampling strategies. Samplingwith a series of stationary sequential samplersis useful in steady or predictable wind regimes.Such a strategy enables the measurement of severalcross-sections of the puff as it passes overhead.Results from a typical stationary sampler deploy-ment are shown in Figure 2. However, duringvariable conditions the stationary samplingstrategy often proves to be less than ideal. Itis during such conditions that the moving samplerstrategy of the TRAC vehicle is most valuable.

The TRAC is able to cross the puff andmeasure the SFg concentrations in real time.Figure 3 shows five successive cross sections ofa typical SFg release. However, if wind speedsare high, only one transect through the puff maybe possible.

«^

FIGURE 2. Typical result* from stationarysequential air samplers showing theSFg concentration in tiae and spacea puff crossed the HATS arc

0 2 4 • •Distinca Ikm)

FIGURE 3. Typical cross sections of an SF6 puff•a •assured by a continuous SF6 analyseronboard TRAC. These are five successiveconcentration cross-sections for MATS#26, 12/11/85

ACOOWLEDCMNT

The information contained in this articlewas developed during the course of work underContract No. DE-ACO9-76SROOOO1 with the U.S.Department of Energy.

SKFERENCIS

Sigg, R. A., 1985, "A Mobile Laboratory for NearReal-Time Measurements of Very Low-LevelRadioactivity." Proceedings of the 5th DOEEnvironmental Protection Information MeetingHeld in Albuquerque, New Mexico, November5-8, 1984, CONF-841187, April 1985,pp. 629-636.

ANS Topical Meettag oa Radiological Accidorts—Perspcctires aid Eoerfeacy Planiif

Savannah River Plant Remote Environmental Monitoring System/ . F. Schubert

ABSTRACT. The SRP remote environmental monitoringsystem consists of separations facilities stackmonitors, production reactor stack monitors,twelve site perimeter monitors, river and streammonitors, a geostationary operational environ-mental satellite (GOES) data link, reactor cool-"ing lake thermal monitors, meteorological towersystem, Heather INformation and Display (WIND)system computer, and the VANTAGE data basemanagement system. The remote environmentalmonitoring system when fully implemented willprovide automatic monitoring of key stackreleases and automatic inclusion of thesesource terms in the emergency response codes.

INTRODUCTION

The Savannah River Plant (SRP) remoteenvironmental monitoring system is a new conceptin emergency response, designed to provide real-time accurate source terms to the emergencyresponse computers in the event of an unplannedrelease of radioactive or toxic chemicals to thebiosphere. Data are collected and archived forimmediate use during the emergency response, orfor later use in research and development.

The 800 square kilometer Savannah RiverPlantsite operated by E. I. du Pont de Nemoursand Company for the U.S. Department of Energyproduces nuclear materials for the U.S.Government. SRP is considered a complete nuclearcomplex. Major facilities at SRP include:* Four nuclear production reactors moderated and

cooled with heavy water* Two chemical separations plants for separating

and purifying the reactor products• A Fuel Materials Facility for producing fuel

for the U.S. Navy's nuclear powered ships• A Defense Waste Processing Facility which will

process high-level radioactive waste forincorporation in glass for underground storage

Although the facilities of the SavannahRiver Plant have a very low potential for hazard-ous releases to the environment, the Environ-mental Technology Division of the Savannah RiverLaboratory with the aid of SRP personnel havedesigned and begun installation of a real-timeremote environmental monitoring system. Thissystem places the latent monitoring technology inexisting production facilities to provide accu-rate source terms to a central computer systemfor effluents which might be released to thebiosphere.

Key stack release instruments are presentlymonitored by control room personnel and theresults are transmitted verbally by telephone tothe Emergency Operating Center (EOC). Thesedata are then manually entered into the HeatherInformation and Display (WIND) System computer(Garrett et al., 1981). The WIND System combinesthe stack release data and local meteorologicaldata and calculates the transport and dispersionof toxic and/or radioactive pollutants. The WINDSystem also calculates the predicted integrateddoses and dose rates for both onsite and offsitepopulations.

Delays can occur in obtaining the necessaryinformation about the possible dose consequence.*to the offsite populations because of the largeamount of manual data entered and the verbalcommunications between the EOC and operatingareas. To eliminate these delays and the poten-tial for human error, a program is in progressfor the installation of electronic stack monitorsand the necessary associated electronics to scaleand digitize the signals for direct transmissionto the WIND System computers.

SYSTEM DESCRIPTION

The SRP remote environmental monitoringsystem is composed of the following instrumenta-tion: Separations Facilities Stack Monitors,Production Reactor Stack Monitors (Figure 1),Twelve Site Perimeter Monitors (Figure 2), Riverand Stream Monitors (Figure 3), Reactor CoolingLake Thermal Monitors, Meteorological TowerSystem (Figure 4), WIND System Computer, and theVANTAGE Data Base Management System.

137

138

rf REACTOR BLDG.

• I SEPERATION BLDG./STACK

FIGURE 1. Map of the Savannah River PlantShowing Locations of Reactors andSeparation Buildings

PERIMETER MONITORS

FIGURE 2. Location of the Twelve PerimeterMonitors

FIGURE 3. Location of the River and StreamSatellite Data Collection Platforms

FIGURE 4. Location of the Seven MeteorologicalTowers, Weather Center AnalysesLaboratory and the WJBF-TV Tower

139

A Geostationary Operational EnvironmentalSatellite (GOES) data link is used to transmitdata from remote locations in and around theplantsite (river, streams, and perimetermonitors). Radio-Telemetry data links connectreactor cooling lake temperature monitors with acomputer controlled base station. The computeralso transmits the received data over telephonelines to the WIND System computer. The noble gasraoniLor uses fiber optic cables to link the stackmonitor to the control room. The fiber opticcable is immune to electro-magnetic interferencefrom motors and relay switches in the local area.

RESULTS

As each element of the system is completed,it is linked to the WIND System computer and theremote environmental monitoring data base.Elements completed to date are: the meteorologi-cal tower system; reactor cooling lake tempera-ture monitors; river and stream monitors, andassociated GOES satellite up-link and down-link;and a tritium forms stack monitor. The noble gasdata links will be available shortly vith theinstallation of data collection and transmissionequipment. The real-time monitoring data arecombined with the meteorological data using theWIND System VAX 11/780 computer and the VANTAGEData Base Management System. The VANTAGE DataBase Management System provides seven levels ofalarm and alert functions. It also can displayup to four variables (in any combination) on anelectronic strip chart. VANTAGE archives all thevariables in such a fashion that they are immedi-ately available to the WIND System emergencyresponse codes, or are available at a Later datefor any other purpose.

The remote environmental monitoring systemwhen fully implemented will provide automaticmonitoring of key stack releases that arepresently monitored by control room personnel.This will eliminate verbal communications anddelay in receiving the necessary informationneeded for the emergency response codes. Thissystem will provide complete real-time accuratesource term information for automatic inclusionin the emergency response codes.

ACKNOWLEDGMENT

The information contained in this articlewas developed during the course of work underContract No. DE-AC09-76SR00001 with the U.S.Department of Energy.

REFERENCES

Garrett, A. J., M. R. Buckner, and R. A. Mueller,1983, "The Weather Information and DisplayEmergency Response System." Nucl. Tech.60(l):50-69.

ANS Topical Meeting on Radfotogka! AccMeats—

Perspectives aad Emergency Pbuwteg

Aquatic Emergency Response Model at the Savannah River PlantDavid W. Hayes

ABSTRACT. The Savannah River Plant emergencyresponse plans include a stream/river emergencyresponse model to predict travel times, maximumconcentrations, and concentration distributionsas a function of time at selected dowstream/riverlocations from each of the major SRP installa-tions. The menu driven model can be operatedfrom any of the terminals that are linked to thereal-time computer monitoring 9ystem for emer-gency response.

INTRODUCTION

The Savannah River Plant (SRP) has emergencyresponse plans to be put into'effect in the eventof a significant release to the environment oftoxic or radioactive material. This paper con-cerns the emergency response to the escape ofliquid pollutants to plant streams which flow tothe Savannah River and thence to the AtlanticOcean. More specifically, the paper describes amathematical model used to predict the immediateconsequences of the liquid release.

The computer input information required forthe stream/river emergency response modelincludes the known or estimated quantity, compo-sition, and location where the pollutant entereda plant stream. The model then predicts traveltimes, maximum concentration, and concentrationdistributions as a function of time for selecteddownstream/river locations. The menu drivenmodel can be operated from any of the terminalsthat are linked to the real-time computer moni-toring system for emergency response. Theseterminals are in the control rooms of all majorSRP facilities. The selected downstream loca-tions stored in the model include the publichighway (SC Highway 125 that crosses the SRPfacility, mouths of streams at the SavannahRiver, two bridges that cross' the Savannah Riverbelow SRP, and the water intakes for theBeaufort/Jasper and City of Savannah WaterTreatment Plants.

OPEKATIOH OF THE MOMEL

The menu-driven aquatic emergency responsemodel contains all of the required stream infor-mation in its data base and requires only entryof data pertinent to the release. In the examplehere, a release from C Reactor to Four Mile Creek

will be simulated. The execution of the modelrequires entry of date, time, type, and units ofrelease (part (a) of Table 1), facility identi-fier (part (b) of Table 1), the amount ofrelease, and the duration of release (part (c) ofTable 1).

TABLE 1

Input Data for a Release(a) Type, (b) Location,(c) Aaount and Duration

itxttttxttttttnuttttttiiittttttttttttttitttxtt

(a) EMERGENCY RESPONSE FOR RELEASES TO PLANT STREAMS

tttttxxtttsstttttttttxxttttttttttttttxtttttxittt

INPUT FOR THE TIRE A W DATE OF RELEASE INOON IS M M Pfl, H1DHICHT IS M M Afl.INPUT RELEASE START (DEFAULT, »7SC Pfl Ml786)

IS RELEASE RADIOACTIVE (V or N, dafiult is VM

ENGLISH OR flETRIC UNITS (E or «, default 1* E X (1

ARE RELEASE UNITS I , KC, CI (Dafiuti Is C D :

(b) IF RELEASE LOCATION ISi

D. AREA OUTFALLHIGHWAY 27» UPPER 3 RUNSRD. C (FLOUINC STREAMS)RD. A (UPPER 3 RUNS)H WfEnF AREAC AREA OUTFALLP AREA OUTFALLK AREA OUTFALLRD. A 14 RILE CREEK)RD. A (STEEL CREEK)PAR POND DABflOUTH OF UPPER FOUR NILE(MOTH OF LOUER 3 RUNSHOUTH OF STEEL-PENBRANCHHOUTH OF « NILE CREEKBOOTH OF BEAVER DM*IWUTH OF UPPER 3 RUHSALTERNATE (MANUAL INPUT)

ENTER CHOICE HERE - CC

PLEASE TVPE

DOHSFSUAHHFFCCPPKKA4ASPDRUPIL

nsmmAL

(0tsmt*tttt**tt»titx»*»tt*Mmt*t*z»xt»ttt

SLUG RELEASE OPTION

XXXXXXXXXXXXXXXXXXIIXXXXXXXXXXXXXXXXXltXXlXXXXXt

ENTER THE NUMER OF CI OF MATERIAL RELEASED: 5

ENTER THE DURATION OF THE RELEASE IN flINUTESi 25

141

142

The output is designed to show graphicallythe impact of a release to downriver users and tohelp in the design of a sample plan to monitorthe release. The output consists of a table andseveral graphs. Table 2 summarizes the inputinformation, in this case for C Reactor at SRP,and gives the results of the model calculation(the maximum concentration and time of arrivalat each of the preselected downstream/riverlocations). The maximum concentration atC-Reactor outfall was 3.3 E-07 Ci/L and thisconcentration had been reduced to about 1.9 E-07Ci/L by the time it had arrived at SC Highway 125230 minutes later due to dispersion processes(Table 2).

TABLE 2

SHL Stream Transport Program

SRL STREAK TRANSPORT PROGRAIt 17-AUG-B6 7:56 Fn DT

DOUNSTREAn CONCENTRATIONS CALCULATED FOR A RELEASEAT C-AREA OUTFALL

RESULTS FOR A SLUG RELEASE OF 5.«« CI, LASTING 2<

INITIAL CONCEHTRATIONI •.33ME-K CI/L

Pn ONRELEASE TIKE IS 7:56THE REACTOR IS UPTHE RELEASE IS RADIOACTIVE

ROAD AISAIMWMH RWEK:US HICHUAV 3811STATE HIGHUAV 1191BEAUFORT-JASPER:SAU. UATER PLANT:

HAX CONCENTRATION(CI/L )

9.l8812E-e6E7

1656EeS6.94951E-89«.91S51E-»9e96696Ee9

ETA(HIN)

1635.842368.M4681.MS490.M5899.»»

TIHEi j u s pnHU Pfl

i«:24 AH1213? AH3126 Pn18:IS pn

DATE8/17/868/18/868/19/863/21/868/21/868/21/8G

The travel time to the Savannah River isabout 1000 minutes from C Reactor Area and theconcentration was reduced by a factor of ten dueto dispersion processes. Once flow from FourMile Creek entered the Savannah River, it wasimmediately diluted by a factor of about 25.Concentration changes were small after mixingwith the Savannah River.

Graphical illustrations are then used topresent the tabular results on area maps and toshow the time/concentration profiles atdownstream/river locations. The first graph(Figure 1) is a map of SRP that shows thelocation of the releaee, identifies the publicroad and mouth of the creek for that release, andshows maximum concentration and estimated time ofarrival. The second graph (Figure 2) is a mapshowing the same type of data as the first graphbut for downriver locations. The third graph(Figure 3) is a plot of the concentration as afunction of time for each preselected location.The third graph is used to estimate the durationof release and the time the concentration couldbe above some guideline concentration level.Because the concentration reduction is largeafter mixing with the Savannah River, a plotshowing the log of the concentration is alsogiven so that details of the concentration in theSavannah River can be seen.

SRL STREAfl TRANSPORT PROCRAH I7-AUG-8S 7:56 Pf! DTDOUNSTREAn CONCENTRATION CALCULATED FOR A RELEASE

RELEASE T I IC I AT C-AREA OUTFALLTINE DATE7S*

IE7iS*

ES/17/K

(11 ROAD Aj~~ETA~"

Tine DATE1H45 PH B/17/8G

RAX CCWC. (CI/L )».l8812E-«6

(2) SAVANNAH RIUERi

" E T V "TINE DATEIt 11 PH S/18/SS

IKX CONC. (CI/L )*.22694E-«7

FIGURE 1. Downstream Concentration Calculatedfor a Release at C-Area Outfall -Onplant

SRL STREAK TRANSPORT PROGRAH 17-AUG-86 7:56 PH DTDOUNSTREAn CONCENTRATION CALCULATED FOR A RELEASE

RELEASE TIfl£: AT C-AREA OUTFALLTINE DATE7lS* Pn S/17/IC

(1) US HICHUAV 3011~ETA "

TINE MTE1*184 M 8/19/M

MX CONC. <CI/L )• .l*S£7E-*t

(2) STATE HICHU4V 1191gj^ -

TinE DATE12137 m S/21/SC

RAX CONC. (CI/L )*.«49SlE-*9

(3 ) tEMJFORT-JASPERl' " E T » "

Tine DATE3126 PN t /21 / tS

RAX COHC. (CI/L )• .91SS1E-W

(4) SAV. WATER PUHTIET*

TIRE MTE• • U S PR I/Z1/SC

HAX COHC. <C1/L )#.SMSSE*4Jt)

/ S i p '(1

\L>6.US HICHUAV 2»1

I C" hSTATE HICHUJW 119 /

\ lEMJFORT^MPER

[i SAU. UATCM/LMfT

FICUKK 2. Downstream Concentration Calculatedfor a Kclease at C-Area Outfall -Offplant

SKL STXEM TRANSPORT PROCRMI 17-AUG-8G 7:'C Ptl DT PAGE 4TINE SERIES PLOT OF DOUNSTREAn CONCENTRATIONS

(1) ROAD A: (41 STATE HICHUAV 1191... (2) SAVANNAH RIVERl (S) lEAtJFOHT-JfiSPE*!

(3) US HICHUAV 3*11 (S) S M . UATER PLANT;

• ».* l.« 1.5 2.* 2.5 3.« 3.S 4.* 4.S

MVC SINCE KLEME (RELEASE TlnE 1 7.K HI )/17/*C>

FIGURE 3. Tiae Series Plot of DownstreamConcent rat ions

143

MODELA one-dimensional model was used to describe

pollutant transport in SRP streams and theSavannah River.

3x

3C

where C * the average cross sectionalconcentration

x " coordinate in the direction of flowu * the mean velocity of flowD « the longitudinal mixing coefficientt * time

The model is conservative (i.e., it isassumed that no material loss occurs duringtransport) and can be used to calculate theresults for any release distribution by use ofrouting. A one-dimensional model is an adequatedescription of pollutant transport except in thezone of discharge. Mixing in the discharge zoneis dominated by convective dispersion whichproduces a skewing of the pollutant concentrationdistribution and is not predicted by the aboveequations. The mixing zone is a small feature(the effective zone is equal to about 50 streamwidths) when compared to the length of thestream. These types of models have been used todescribe pollutant transport in many streams andrivers (Fischer, 1966; Thackston, Hays, andKrenkel, 1967).

A conservative model from an emergencyresponse view results in higher pollutantconcentrations and faster travel times in astream or river than a nonconservative modelwhere adjustments are made for deposition,sorption, and chemical reactions for pollutantsin a stream or river. The conservative model isconsiderably easier to implement and by use ofthe pollutant routing technique can handle almostany pollutant release distribution.

EXPERIMENTAL EVALUATION OF STREAM/RIVER TRANSPORTCOEFFICIENTS

It was necessary to determine the velocityand dispersion coefficient for each stream andthe Savannah River to estimate the travel timesand pollutant concentrations at a location.These coefficients were determined using dyetracer studies. A one-dimensional model, TETRAD,was programmed to allow the calculation of streamvelocities and dispersion coefficients frommeasured dye concentration profiles. For eachstream, dye was released at the facility and thedye concentration curve was measured as afunction of time at downstream locations.

Stream velocities and dispersion coeffici-ents were calculated from Che dye curves usingTETRAD. TETRAD routed the point-by-point dyeconcentration from an upstream to a downstreamlocation using an initial set of stream velocityand dispersion coefficients estimated from thestatistics of the dye distribution (Godfrey andFrederick, 1963; H. B. Fischer, 1968; Thackston,Hays, and Krenkel, 1967; and Buckner and Hayes,

1975). The stream velocity and dispersioncoefficients were then incremented and theprocess continued until nonlinear least squaresestimators were minimized for the coefficients.When the least squares estimators were minimizedfor the routing, the best estimate for velocityand dispersion was obtained. An example of thedye response curves in Lower Three Runs Creek andthe resulting coefficients are shown in Figure A.Slight deviations of the model from the actualcreek dye response can be seen. These coeffici-ents were then used as a data set for theemergency response model. Dispersion coeffici-ents and velocities determined for streams on theSavannah River site ranged from 1 to 25 m/sec andfrom 0.1 to 3 ktn/hr.

The model is continually updated with newdata and re-evaluated velocity and dispersioncoefficients. Real-time flow data telemeteredfrom the streams and river are being incorporatedin the model, thus assuring that the appropriatedilution factors are available for each stream atthe time of emergency response.

FIGURE 4.

SUMMARY

Lower Three Run* Strew* Dye TracerResponse Curves

A stream/river emergency model has beendeveloped for use at the Savannah River Plant topredict travel times, maximum concentrations, andconcentration distributions as a function of timeat selected downstream/river locations from eachof the major SRP installations.

ACKNOWLEDGEMENT

The information contained in this articlewas developed during the course of work underContract No. DE-AC09-76SR00001 with CheU.S. Department of Energy.

144

UraiENCES

Buckner, M. R., and Hayes, 3. W., "PollutantTransport in Natural Streams," ANS Meeting onComputational Methods in Nuclear Engineering,Charleston, South Carolina, April 15-17, 19 75.

Fischer, H. B., "Longitudinal Dispersion inLaboratory and Natural Streams," W. M. KeckLaboratory of Hydraulics and Water Resources,California Institute of Technology,Pasadena, California, Report No. KH-R-12,June 1966.

Fischer, H. B., "Dispersion Predictions inLaboratory Natural Streams," Journal ofSanitary Engineering Division, ASCE, Vol. 94,No. SA5, pp. 927-943, October 1968.

Godfrey, R. G. and Frederick, B. J. , "Dispersionin Natural Streams," United StatesGeological Survey open file report (1963).

Thackston , E. L. , Hays, J. R., and Krenkel, P. A.,"Least Squares Estimation of Mixing Coeffi-cients ," Journal of Sanitary EngineeringDivision, ASCE, Vol. 93, No. SA3, pp. 47-58,June 1967.

ANS Topical Meetteg OB RrtMogical AccMeato—Penpectires Md EaergeKy

Meeting NUREG-0737, II.B.3 Requirements—BackupAnalysis of Postaccident Samples

Todd L. Hardt, Mark J. Bradley, and Trudy E. Phillips

ABSTRACT - The Babcock & Wilcox Company has inplace a mult'fparticipant program designedspecifically to meet the requirements ofNUREG-0737, II.B.3 for offsite analysis ofbackup, post accident samples. This paperprovides a brief review of B&W's experience inanalyzing the Three Mile Island (TMI-2) NuclearStation post accident samples and thesubsequent establishment of the Post AccidentSample Analysis Program to meet NUREG-0737requirements.

Highlighted are the program's "EmergencyProtocol" and its contribution to effectiveemergency planning. The paper also reviews thetechnical requirements for analyzing andhandling post-accident samples. Specifically,the methods for analyzing low-levels ofchlorides in a highly borated sample and theproblems associated with transportingpost-accident samples will be addressed.

INTRODUCTION

NUREG-0737, "Clarification of TMI ActionPlan Requirements," I I .B.3, "Post-AccidentSampling Capability," Criteria (8) states thefollowing, " . . . the license shall providebackup sampling through grab samples, and shalldemonstrate the capability of analyzing thesamples. Established planning for analysis atoff-site facil it ies is acceptable. Equipmentprovided for backup sampling shall be capableof providing at least one sample per day for 7days following onset of the accident and atleast one sample per week until the accidentcondition no longer exists."

In early 1982, Babcock S Wilcox (BSW)Initiated the Post-Accident Sample AnalysisProgram for backup, offsite analysis of post-accident samples. The program was specificallydesigned to meet the requirements of NUREG-0737and benefited from the lessons learned duringBSW's handling and analysis of TMI-2's postaccident samples. The program consists ofthree important documents:

Emergency Protocol: A one-pageguideline identifying whom to call andthe required Information for BSWnotification of post accident sampleshipment.

Project Technical Plan: Describes thestate of readiness to be maintained atB&W's Lynchburg Research Centerlaboratories, response times, andscope of analyses to be performed.

Quality Assurance Plan: Documentscompliance with the requirementsimposed by 10CFR61, Appendix B.

This paper reviews BSW's experience withanalysis of the TMI-2 post-accident samples andthe subsequent genesis of the Post-AccidentSample Analysis Program. Discussed in detailare the program's "Emergency Protocol" andbackup analysis of post accident samples.Transportation of post accident samples is alsodiscussed.

BACKGROUND

In early April 1979, Babcock S Wilcox wascontracted to perform radiochemical analysis onsamples taken from the TMI-2 plant followingthe March 29th accident. The samples wereshipped to BSW's Lynchburg Research Center inlead-shielded, 55-gallon drums. The earlysamples were taken from the plant to theairport by truck and then air l i f ted to BSW inLynchburg, Virginia by the National Guard. Thelater samples were transported by privatecharter from Harrisburg, Pennsylvania, toLynchburg, Virginia. Once the sample arrived1n Lynchburg, I t was transported to the lab bya commercial transport company. Shipment ofreactor coolant samples continued on a weeklybasis for over 2 1/2 years.

Several valuable lessons were learnedduring analysis of the Init ial samples fromTMI-2. For example, the fact that a singlepoint of contact did not exist at TMI-2 or BSW

145

146

to respond to this kind of emergency createdconfusion. To address this and other problemsexperienced following TMI-2 and to meet therequirements of NUREG-0737, BSW's Post-AccidentSample Analysis Program was established inmid-1981 as a multiparticipant program for theBoiling Waste Reactor (BWR) Owners Group. Thepurpose of the program was to ensure that thecapability for backup analysis of post-accidentsamples was available on an emergency responsebasis. I t was designed specif ical ly to meetthe requirements of NUREG-0737, I I .B.3 foro f f - s i te analysis of post accident samples.

The analysis done on the TMI-2 samples wasthe minimum required to answer questions ofplant condition and s tab i l i t y . Gamma-rayanalysis, in particular the ident i f icat ion ofcesium and iodine ac t i v i t i es , was coupled withstrontium analysis to assess the extent ofdamaged fue l . Boron was measured to verify theshutdown margin. Water chemistry attr ibutessuch as pH, conductivity, chlorides and oxygenwere used to determine the corrosion potentialof the coolant. Analysis of the dissolvedgases aided in the estimation of core damageand coolant corrosion potential .

The current program provide.': the u t i l i t yemergency planner with an "emergency protocol,"which includes a single point of contactprotocol for program in i t i a t i on in the event ofan accident. The program provides a 24-hourturnaround confirmatory analysis capabil ity toover one-third of the operating nuclear plantsin the United States. The emergency protocolhas been exercised many times during plantd r i l l s with two u t i l i t i e s testing the completesystem of sample shipment, receipt, analysis,and reporting of the results.

EMERGENCY PROTOCOL

Following the TMI-2 accident, samples weretransported to the Babcock & Wilcox researchcenter i n Lynchburg, V i r g i n i a , f o r ana lys is .At the t ime, no mechanism existed fo r then o t i f i c a t i o n of E3W, or f o r rapid coordinat ionof resources to analyze the samples. As such,a s i g n i f i c a n t amount of e f f o r t was required togear-up f o r the i n i t i a l samples.

In developing the Emergency Protocol (EP)fo r the Post Accident Sample Analysis Program,the need to make i t simple and r e l i a b l e wereconsidered paramount"! Simple meant t ha t the EPcould be exercised quick ly w i th l i t t l e chance

of making a mistake. Rel iable meant tha t theEP would accomplish i t s purpose every t ime.

The Emergency Protocol i s a one-pagedocument. I t provides the names and telephonenumbers of the emergency contacts; theinformation required at the time of i n i t i a t i o n ;and BSW's emergency response ac t i v i t i es . Forr e l i a b i l i t y , both a primary and alternatecontact are l i s ted , and the work and home phonenumber for each is provided. The EmergencyProtocol has been successfully exercised manytimes over the past few years as part of plantemergency d r i l l s .

BACKUP ANALYSIS OF POST-ACCIDENT SAMPLES

The NRC's "Clar i f icat ion of TMI ActionPlan Requirements," NUREG-0737, specifies therequirements that must be met by licensees ofcommercial nuclear power reactors for postaccident sampling and analysis. Cri ter ion 8provides the guidelines for using an o f f - s i t elaboratory for backup sample analyses. Al l ofthe post accident analyses have beensuccessfully tested and the required accuraciesachieved, as l i s ted in Table 1 .

Post accident sample analyses areperformed on three sample types — l i q u i d , gas,and part iculate and charcoal f i l t e r s .Depending on whether the plant is a PWR or BWR,the type of sample to be handled could be anatmospheric gas or l i qu id or a pressurizedl i q u i d . The gas phase of pressurized samplesis expanded and flushed from the l iqu id bymeans of an argon sparge into an evacuatedexpansion v i a l . Gas samples are removed by asyringe for analysis by gas chromatography andgamma-ray spectroscopy. The l iqu id phase 1ssimply drained from the sample container into abott le for subsequent analyses.

To perform these analyses within a 24-hourperiod, the 38W Lynchburg Research Centerlaboratory is kept in a state of readiness ata l l times. Reagents are kept fresh and a l lanalytical Instruments are maintained in aready and calibrated state. Therefore, 1n theevent of an accident or a practice d r i l l , acomplete set of analyses on a post accidentsample (except for strontium 89 and 90) can beperformed wi th in 24-hours of receipt of thesample.

The NRC c la r i f i ca t i on l e t te r forNUREG-0737 requires that laboratory proceduresfor the analysis of post-accident reactor

Analysis

Gamma ActivitySr-89,-90Boron

Cilor ide

ConductivityK, 0, Kr

TABLE !.. Post-Accident Sample Analyses

Range Accuracy Method

50-2000 keV0-1000 uCI/cc0.1-100 ppm100-10,000 ppm>500 ppb<500 ppb1-10,000 uohms> 0.5%< 0.5%1-13

20%25%5%10%10%50 ppb10%5%20%

SpectroscopyRadiochemicali.e.Titrationi.e.i.e.Cond. BridgeGCGC

0.3 units Electrode

141

coolant samples be tested with a simulatedpost accident sample. The B4W procedures havebeen successfully tested for the analysis ofchloride and boron in the prescribed testmatrix. Both analyses used the method of ionchromatography and were shown to meet or exceedthe specifications of the guidelines.

TRANSPORTATION OF POST-ACCIDENT SAMPLES

An undiluted liquid, post accident grabsample drawn for off-site analysis willtypically be 10 to 30 mis and have an activityas high as 1 curie/ml. The sample may bepressurized or unpressurized and be in anythingfrom a glass vial to a shielded, steel bomb.The transportation of these highly radioactiveliquid and gaseous samples is difficult.

There are two methods for shippingpost-accident samples — Type A containers andType 8 containers.

o Type A - Containers are designed toretain the integrity of containment andshielding under normal transportconditions.

o Type B - Containers are designed toretain the integrity of containment andshielding under normal transportconditions and hypothetical accide.ittest conditions.

Type B packaging is more restrictive andcan thus carry higher activity samples. Theprimary regulations that govern thefabrication, testing, licensing and use of TypeA and B containers are:

o 10CFR71 - Nuclear Regulatory Commissionregulations on transportation ofradioactive material,

o 49CFR173 - Department of "Transportationregulations on transportation of hazardousmaterials.

A typical post accident sample w i l lrequire a Type B container for shipmentimmediately following an accident. However,after a short decay period (16 hrs to 7 days),these samples can be shipped as Type A. Thus,the two options for transporting apost accident sample are:

1. Type B Shipment: The post accident sampleis drawn and then shipped immediately viatruck or ra i l to the o f f -s i te laboratoryfor analysis.

2. Type A Shipment: The post accident sampleis drawn and then stored for deacy. Whenthe sample reaches Type A, levels, they aretransported by a i r to the of fs i telaboratory.

The decision to use a Type A or Type Bcask is a function of many factors. Two keyfactors are the technical adequacy of thetransport method selected and the 'up-front1

cost. The up-front cost is a function of the

cost of the container, and the lead pigs neededto hold samples while the containers are inuse. Typically, a Type A cask w i l l cost muchless than a Type B. As a minimum, two casksshould be available. This is a cost of$150,000 to $300,000 for two Type B's versus$20,000 to $40,000 for two Type A's.

For ALARA purposes, a post accident sampleshould be drawn, stored, and transported in thesame lead pig. A Type B container can take upto nine days for a round t r i p between theo f f -s i te laboratory and the plant. Thus, evenwith two Type B casks, up to seven lead pigsw i l l be needed to handle the f i r s t week'ssamples. The number of lead pigs neededthe Type A transportation scenario w i l l dependon the roundtrip f l i gh t time (one day or less)and the sample decay time. For a BWR, 10-ml,depressurized sample, the decay time to Type Ais approximately 16 hours. For a PWR, 15-ml,pressurized sample, the decay time to Type A Isapproximately seven days. Thus, use of a TypeA cask w i l l require no more, and probably fewerlead pigs for drawing and storing samples thanfor a Type B.

In general, using a Type A cask fortransport or post accident samples w i l ls igni f icant ly reduce up-front costs. Keep Inmind, following the TMI-2 accident, the . i n i t i a lsamples were flown to Babcock * Wilcox'sLynchburg Research Center by Air NationalGuard and the subsequent samples were flown inType A containers using a private carr ierservice.

EXPERIENCE

During the four years that this programhas been in effect, the entire system has beenexercised one time. During this drill,unexpected problems arose concerning the sampleshipment and the actual sample bomb used tohold the sample. A power station notified BSWthrough the emergency protocol that a post-accident practice sample was being shipped toLynchburg by air for analysis. The sample hadnot been received by noon the following day asexpected and a tracer located It at the airportIn Richmond. Oue to its weight, the samplecould not be transported on small plane, andthis, subsequently arrived by truck. Thisdrill pointed out the need for each utility todefine the most efficient way to transportthe sample to its backup laboratory. A followup sample was shipped a month later by truckwith no delays.

An additional problem surfaced when thesample bomb was removed from the shippingcask. We found that the sample container wastoo large to fit onto the B4W sampling rig.Several hours were spent replacing fittings onthe container and constructing the tubing tobe used so that the smallest possible volumewould be Introduced to the expansion side ofthe system. This pointed out the need foreither a standardized sampling bomb or eachutility to inform Its backup lab of thecorrect dimensions of the sample container.

Despite the problems encountered, we wereable to complete all analyses and report

148

results to the ut i l i ty within the prescribed24 hours. The single point of contactemergency protocol worked very smoothly duringthis dr i l l as i t had in other, less-extensivedr i l l s .

SUMMARY

The Sabcock & Wilcox Post-Accident SampleAnalysis Program successfully fu l f i l l s therequirements of NUREG-O737, II.B.3 foroff-site analysis of post-accident samples.Key to that success 1s the emergency protocolfor program initiation and the abil ity of thelaboratory to analyze post-accident samples tothe required precision and accuracies.

When Initiating similar programs formeeting regulatory requirements, Importantfactors to consider are:

1. The emergency protocol needs to be bothsimple and reliable.

2. Use vendors and multiparticipant programsto meet regulatory requirements. The idealsituation is to be able to have the vendoraddressing regulatory questions on theprogram rather than each participant on anindividual basis.

ANS Topical Meetiag on Radiological AccMeats—Perspectives aid Eawrgeacy Pltaaisig

An Evaluation of Emergency Systems for the Monitoring,Sampling, and Analysis of Reactor Coolant, Containment

Atmosphere, and Airborne EffluentsAndrew P. Hull, Wayne H. Knox, and John R. White

ABSTRACT

A postimpleroentation review has been made in NRCRegion I of the postaccident sampling systems(PASS), the gaseous effluent monitors, and pro-visions for sampling effluent participates andradioiodines, which were required by the NRCsubsequent to the TMI-2 accident (NUREG-0737).Prefabricated PASSs were predominant. Problemsincluded inadequate purge times, separation ofdissolved gases, excessive dilution, and Cheaccuracy of analytical techniques in thepresence of interferences. Microprocessorscontrolled high-range gas monitors with integralprovisions for sampling particulates and radio-iodines in high concentrations were widelyused. Calibration information was generallyinsufficient for the unambiguous conversion ofmonitor reading to release rates of a varyingpostaccident mixture of radiogases. The refer-enced sampling guidance (ANSI-N 13.1-1969) wasinappropriate for the long sampling linescustomarily used. Generic research is needed toestablish the behavior of particulates andradioiodines in these lines.

I. INTRODUCTIONThe accident >it Unit-2 of the Three Mile

Island Nuclear Power Station (TMI) on March 28,1979 disclosed numerous deficiencies in theinstalled system for the collection and analysisof primary coolant and containment atmospheresamples under post-accident conditions. Thissystem was typical of those then installed atU.S. Nuclear Power Stations. The accident alsodisclosed limitations in the capability of itsgaseous monitors and the adequacy of samplingsystems to deal with the concentrations ofairborne effluents that might be anticipatedunder post-accident conditions. They were alsotypical of those being employed at the time ofthe accident.

Subsequently, the U.S. Nuclear RegulatoryCommission (NRC) drew up a number of short-termrecommendations which were based on the lessonslearned from the accident and which were de-scribed in the report NUREG-0578.1 They in-cluded measures for the improvement of post-accident sampling capability and for theexpanded range of radiation monitors. These

recommendations were developed into specifictasks in report NUREG-0660 and fir.ilized forimplementation in a clarification, NUREG-0737 . Its specific requirement for PoetAccident Sampling Capability are set forth inItem II.B.3.

Those for High-Range Noble Gas EffluentMonitors are set forth in Item II.F.I, Attach-ment 1, and those for the sampling and analysisor Measurement of High-Range Radioiodine andParticipate Effluents in Gaseous Streams are setforth in Item II.F.I, Attachment 2.

An implementation deadline of January 1,1982 was specified. It was also indicated inNUREG-0737 that these systems would be subjectto a post-implementation review. Responsibilityfor their post-implementation review was as-signed by the NKC Office of Inspection andEnforcement to the NRC's regional offices. Inmid-1983, Region I contracted with the Safetyand Environmental Protection Division ofBrookhaven National Laboratory for technicalassistance in their performance. Each hasrequired the identification and documentation ofthe licensee's commitments, clarifications,schedules and orders. A subsequent on-siteinspection has Included the physical verifi-cation and validation of the installation andoperability of equipment, as well as the verifi-cation of the adequacy of the licensee's pro-cedures and of the qualification and training oflicensee's personnel.

Starting in late 1983, on-site reviews havebeen completed at the rate of about one permonth for the twenty operating licensee sites inRegion I, which currently contain a total oftwenty-five operating reactors. They arelocated in four of the New England states: NewYork, New Jersey, Pennsylvania and Maryland.

II. APPROACHBefore the on-site reviews commenced, the

individual elements that should be included Inthe overall review effort were considered in aManagement Oversight and Readiness Tree (MORT).Following this, a specific set of instructionsand/or questions related to each review compo-nent was prepared. These included such sub-categories as design, monitoring system,procedures, structures, hardware and supportservices, readout an^ recording, personnel andtraining.

149

150

III. FINDINGS

A. POST ACCIDENT SAMPLING SYSTEM (II.B.3)As indicated in NUREG-0576, the purpose

behind the requirement for the improved post-accident sampling capability was the promptprovision of information for the assessment andcontrol of the course of an accident. In par-ticular, the II.B.3 required chemical andradiological analyses are intended to provideInformation for the assessment of core damageand coolant characteristics. The requiredanalyses of containment atmosphere are Intendedto establish the presence and concentration ofhydrogen, as well as to provide information forcore damage assessment.

The principles of core damage assessmentare based on the grouping of fission productsaccording to their volatility. Thus, thefraction of each group that would be releasedwould depend on the temperature and fuel frag-mentation during an accident. An extendeddiscussion of this subject was presented in theRogevin Report , from which the grouping shownIn Table I is excerpted. An Inventory of themajor fission products in a reference 3651 Mw(t)BWR which has been in operation for three years,arranged according to release groups, is shownin Table II.

In very broad terms, an accident in whichonly noble gases were released would be indica-tive of fuel cladd?"ff failure, one in which thevolatile nuclides of I and Cs were present wouldbe indicative of high fuel temperatures and onein which the non-volatiles were also present.In a more complex situation, the core damagewould be determined by a set of simultaneousequations which take into account the ratios ofthe release groups.

Table I

Categories of Fission ProductReleases in Order ofDecreasing Volatility

I Noble Gases (Kr, Xe)II Halogens (I, Br)

III Alkali metals (Cs, Rb)IV Tellurium (Te)V Alkaline earths (Sr, Ba)VI Noble metals (Ru, Rh, Pd, Mo, Tc)VII Rare earths and actinides

VIII Refractory oxides of Zr and Nb

The licensees of the twenty operating powerreactors In NRC Region I have Installed a varie-ty of systems to meet the requirements of NUREG-0737, Item II.B.3. As shown in a summary InTable III, they range from relatively simplelicensee designed systems which are intendedsolely to obtain samples of reactor coolant andof the containment atmosphere for subsequentlaboratory analysis, to elaborate vendor orarchitect engineer designed systems which areIntended to perform all of the required analysesin line, with the laboratory only as a back-up.

Table 11

Inventory of MajorReference Plant Operated

Group-Rogevln Report

Noble gases 1

Halogens 11

Alkali Metals 111

Tellurlua Croup IV

Al/allne Earth9 V

Noble Metals VI

Rare Earths VII

Refractories VIII

Fissionat 3651

Isotope*

Kr-85.Kr-85Kr-87Kr-68Xe-133JCe-135

1-131I-I321-1331-1341-135

Cs-134Cs-137Cs-138

Te-132

Sr-91Sr-92»a-140

Ho-99Ru-103

Y-92La-140Ce-141Ce-144

Zr-95Zr-97

Products In aMWt for Three Years

half-Life

4.48h10.72y76.3a2.84h5.25d1. llh

6.04d2.3h

2O.B)i

52.6.6.61h

2.06y30.17y32.2.

78.2h

9.5h2.71h

12.lid

66.02h39.4d

3.54h40.2h32.5d

284.3d

64. Od16.91,

Inventory**106 Cl

24.6I.I

47.166.8

202.026.1

96.0140201221189

19.612.1

178.0

138

115123173

183155

124184161129

161166

* Only the representative Isotopes which have re la t ive ly large inventory an<considered to be easy to Mature are listed1 here.

None of the reviewed PASS Systems wereadjudged perfect in every respect. However, ofthe seventeen which could be fully tested at thetime of the review, all met the basicrequirements of Item II.B. 3. Of the three thatcould not, one was inoperative, one had anImproperly installed valve which made itimpossible to obtain a sample of the containmentatmosphere and one could not conduct a test ofthe containment atmosphere sampling system dueto its Technical Specifications, which did notpermit the opening of the valves which maintaincontainment isolation during operation.

The representativeness of the PASS samplesreactor coolant and the licensee's radiologicalanalytical capability were tested by making acomparison of the results of their analysis withthose from the plant's normal sample sink. Theaccuracy of the licensee's chemical analyticalcapability was tested by the use of standards ofknown content. With one exception, the systemswhich were designed to perform all or most ofthe required analyses in-line demonstrated thegreatest operational readiness and ability toprovide prompt data. This was especially evi-dent for the one system that was used to analyzeroutine samples. The exception required a longstartup, so that the required sampling andanalysis with it could not be completed withinthe stipulated three hours.

151

Table III

Summary of Installed Post-AccidentSampling Systems

Sample Collectioncapabil i ty)

(no on-line analysis

No. Remarks57

Sample collection (limited on-line analysiscapabili ty)

General DynamicsQuadrexSentry

Stone & Webster

No. Remarks

3 In-line-pH, Cond.1 In-llne-pH, B, Cl1 In-line-pH, Cond,

DO.Dis H1 In-line-pH, Cond,

IX), Dis H, Cl1 In-line-pH, B, Cl

Full in-line analysis capability (includingisotopic)

SentryCombustion Eng.

Ho.

11

Remarks

A schematic of the coolant sampling portionof a relatively simple PASS Is shown in Figure 1and that for containment atmosphere sampling Isshown in Figure 2. The principal features ofthe associated control panel are shown in Figure3. The panel also contains a mimic diagram ofthe system, with pilot lights to indicate thatthe Intended steps (i.e. value close up opening)have occurred. It is obvious that even thisrelatively simple system is in fact quite com-plex, so the detailed and lengthy procedures arerequired to guide the PASS operator through thesequence of steps necessary to obtain thedesired samples.

The principal deficiencies that wereidentified during the review of the PASS systemsare summarized in Table IV. It should be notedthat In many instances the findings of Inadequa-cy of surveillance was made on the basis of thelack of a suitable schedule and/or the excessivetime Interval required to get a system fullyback on line after a fault had been identifiedby the licensee. Although most purge tiroesseemed adequate, In many instances the licenseehad not conclusively established this by makingcalculations of the volume of the line(s) to bepurged. Midway In the review, the Office ofInspection and Enforcement indicated that thecontainment atmosphere sample was primarily forfission gas measurements, so that sample linelosses should not considered a significantfactor unless the licensee Intended to usemeasurements of airborne radiolodines In thecontainment atmosphere in the assessment of core

LIQUID POISON

Principal Deficiencies Identified in lUvlew ofPoet-Accld«nt Stapling Systems

Deficiency

Inadequate surveillance and Maintenance

Asiuranee from atudy that purg« tlmea adequate

Assurance of representativeneae of radloiodlneain containment air sample

Adequacy of documentation that shielding In eampleroom «ad/or of sample during transport sufficientto enAbl* operation within GLVC-19 criteria.

Procedures did not call for proper pressure and/ortesjperature corrections

Procedures Inadequate or In need of revision to conformto actual operation of PASS

Cxceafllve dilution of aaaple prior to screening for actuallevel of activity

Liquid in stripped gas

Flow not afsured in the absence of a flow sjeter

Insufficient or no backup for ona or sore in-line analyses

inadequate cest of a l l features of ayetam by llcanae* priorto on~slce review

Inadequate training or insufficient number of trainedperiionnel to assure abil ity to operate system during post-sccident conditions*

Inadequate assurance that aample could be obtained whenreactor depressurlzed \no pump In PASS}*

Needle bant during attempt to perforate status of samplecollection vial (UE design system)

Improper interpretation of flow producsd by cr l t lcs l orifice

Unsuitable caek/shiald vial for saanle transport

Volums delivered by ball valve (for dilution) not establishedby sctual msasurement (GE design system)

Chemlcel analysis procedure not adequately teated for possiblelnterfarencee

Figure 1PASS LIQUID SAMPLING SIMPLIFIED PSIO

-IXl—V-I9-I9 V-I9-20

RECIRCLOOP

V-24-29 HV-24-30

0 /# fllBLOG

V-37-9

S/0 COOLING

V-I7-54 H0/W RiBLOG.

fV-40-8

V-40-37CORE JV-40-12SPRAV

- t t

TO

TOOLS. U i R ! "-CH2—>*-

V-40-2

iAMPLI

V-40-3S

, .V-40-36 FROMDEUINWATERTANK

X rV-40-39

TOTORUS

V-40-24

TORBEDT

PASS*

fv-40-38 FROM

FROMV-40-75 RASS

V-40-87

damage. This in the view of the authors ' V-40-26

152

Figure 2CONTAINMENT ATMOSPHERE PASS SAMPLING OIAGRAM

SIMPLIFIED P S 10

PASS ROOM

Figure 3CE MSS COaTIOL HNEL

JP I - H I PI-701

LIQUID PRESSURE PI-H2 SAMPLE G&l PffESSUHE0I53OLVEOMSME55UM

f\ —1*4 TI-IIO Tl - 724

P0«» STITUS LUWTS

HC-TI9-1

DISJOLVED

ANOLIOUIO

HC-C0

O

10

•

1 Ai 3

45

downplays the value of this measurement for theestablishment of the potential source term ifthe containment should leak or fail outright.

In most cases, the shielding provided forPASS systems and sample transport appearedadequate, but many licensees had not conducted aformal study to establish that the GDC-19criteria (5 rem whole body, 5 rem extremitydose) could be met.

The balance of the listed deficiencies andthe measures necessary to address them should beself-explanatory.

B. HIGH-RANGE NOBLE GAS MONITORSA summary of the installed high-range noble

gas monitors, according to their location (on-line or off-line), type of detector, and vendoris shown in Table V. It is evident that theRegion I licensee have chosen a variety ofapproaches to comply with the requirements ofItem II.F.1-1. The typical Boiling WaterReactor (BWR) contained either one monitoredrelease point under accident conditions, theunit vent, or a second monitored release pointfor the standby gas treatment system. ThePressurized Water Reactors (PWR) were morevariable, with from one monitored unit vent anda main steam line monitor to three monitoredvents and a steam relief monitor.

Two licensees installed on—line monitors,using ion chambers which were located in orimmediately adjacent to stacks or ducts, whileseventeen installed off-line monitors. Of thelatter, six installed "gas only" high-rangemonitors as additions to their pre-existing low-range monitors. A schematic of such a monitor

153

TAB1.K V

S1IMHAKY Of 1NSTAM.KD HID- ANIJ HH.'ll-HANUr'. N0B1.K CAB HONITOH.S

Wo. Bange Detector Vendor Model

On-Une

2 Hld/Hlgh Ion Chamber ( I ) GA M1-2A( I ) VU lo recn M 7

Operat lnKMode

Data ftaclt^round

Off-Line

Una Only

1 Mrf/Hlfth I" I Jl«t 1 c

1

NH(; UA-27U

H i d <W V l c t o r c e n

High ]on Chamber VJctoreeu M7

3 Hld/Hlgh Ion Chamber Vlctorecn 647

1 Mld/Hlgh Ion Chamber Router-Stoke* C4-2!)10-10l

Integrated Gag Honltorfi and P a r t l c u l a t e - l o d l n c Simpler*

5

3

Z

1

1

MidHlKh

HidHigh

HidHigh

Mld/lllgh

MidHI ill

Cd-TeCd-To

m

CH<«

Co-LI

Kberllnu

SAI

Kbcrllnc

WKCH

svim-

WJM-HH

HACKH5

AXH-]

High A1ura

Contlniiouu

Continiiriito

HIKI) DIAL'D

HI«h Alarm

Contlniioiia

lllp.h Alnrn

Contlniiotiu

High Al/iiw

Mo

No

No

No

Ye

Ye

Ve

Yo

In

No

No

No

which utilizes an ion chamber, ia shown inFigure 4. Twelve licensees installed com-mercially available Integrated monitors withmodules for both monitoring and sampling. Aview of a typical one (the Kaman KDGM-HR) isshown In Figure 5.

These installations have also incorporateda variety of approaches to the problem ofachieving the required full-range sensitivity.Typically, three overlapping-range detectorshave been provided, as shown in Figure 6 (forthe Genral Atomics WRGM). In order to achievethe upper limit of 10 uCi/cm ( Xe equiva-lent), most of these monitors are designed sothat their high-range detectors view a limitedvolume of gas, as compared to that viewed bytheir mid- or low-range detectors. An example,for the enhanced high-range detector of theSaman HRH, is shown in Figure 7.

Although Item II.F.1-1 was not specific onthe calibration of noble gas monitors up to therequired upper range, the NRC has provided someguidance. It recognized the problem of theavailability of suitable noble gases, i . e .

Xe In sufficient concentrations and of theirutilization by licensees if they were available.Therefore, the Staff recommended that a one-time"type" calibration in the laboratory over theintended range be performed and that the trans-fer procedure of ANSI NJ23-1978 be utilized inconjunction with solid sources at appropriateenergies for on-site calibrations.

Figure

ADDED EQUIPMENT

High-range effluent process radiation

monitors

FiRure 5. Kaman HRH high-range noble gas monitor

and Bampler

154

As suggested by Table VI, most of the ven-dors appear to have performed onlv a "one point"primary calibration, utilizing 1JJXe and S 5Kr.They have then performed a number of transfercalibrations with solid sources with a range ofactivities and energies, to establish the energyresponse and/or range capability of a givendetector.

A summary of the sampling arrangementswhich have been provided to achieve compliancewith Item II.F.1-2 and which have been reviewedto date is shown in Table VII. Again, a varietyof approaches is evident. Some licensees (in-cluding the five who have utilized "gas only"monitors to comply with Item II.F.1-1) installedindependent sampling facilities. One licenseewrote emergency sampling procedures with incor-porated pre-existing unshielded collector forroutine sampling. Five added additionalshielded particulate and iodine sample positionswhich were connected to an existing low-rangesample line, while one added a pre-fabricatedmultiple sample-position module.

Eleven licensees have installed integratedmonitor/samplers which contain micro-processormodules that provide for the automatic or remotecollection of a sample at one of three indivi-dual sample positions, as is also shown inFigure 8. Another licensee located its inte-grated unit in what would become a high-radia-tion field during post-accident conditions, soelected to create another more remote samplingstation. These integrated monitor samplerstypically provide for a much reduced flow of afew hundred cm /min, as compared to the 1-2 cfmflow that is typically provided for low- andmid-range sampling. The intent is to therebylimit the total amount of activity that would becollected at concentrations which approach theuy,.er design criterion of 100 uCi/cm for thestipulated 30-minute sampling period.

IV. LESSONS LEARNED

with distance of Eberline's high-range detectorto each of these nuclides. When correctedrespectively for absorption and bremstrahlung,the true energy response of this detector isabout midway between the two curves, so usingone point from either could lead to a factor oftwo error.

Beyond this, these uninterpreted cali-bration data were in some Instances also em-ployed to calculate release rates (in uCi/sec),without regard to the variable energy responsecharacteristic of the detector. This charac-teristic may be close to linear with energy, asshown in Figure 9, for the Kaman KDGM-HR, or maybe quite non-linear as shown in Figure 10, forthe General Atomics WRGM. Beyond the inherentresponse of the detector itself,, its energyresponse may also be dependent on the geometryin which it is installed and the type andthickness of the intervening duct or pipe wallswhich may absorb radiations before they reachthe detector.

All of the reviewed licensees haveinstalled monitors which in principle net theupper range criterion of 10 uCi/cm . However,only two had calibrated the installed high-rangemonitors on-site with radiogases in concentra-tions approaching 105 uCi/cm . The vendorcalibration information supplied by Kaman, asshown in Figure 11, suggested that a test withactual radiogases approaching tbese concentra-tions had been performed with Xe. However,on the basis of field testing which employed

Kr it was found by another investigator thatthis monitor could not meet the specified upperrange. It is our understanding that sincethese tests, the Kaman high-range detector hasbeen modified so that it can do so. A similarfall-off which appeared to be due to a largedead-time at high concentrations was reported bya consultant to a Region I licensee in a field

calibration of the high-range detector (SA-9) ofthe Eberline SPING.li

A. HIGH-RANGE NOBLE GAS MONITORS

Oversimplifications in the conversion ofthe direct indications of the installed gasmonitor, typically in cpm or mR/hr, to effluentconcentrations and/or rates of release wereamong the principal shortcomings encountered inthe reviews.

The guidance in NUREG-0737, II.F.1-1 states"Design range values may be expressed in Xeequivalent values for monitors employing gammaradiation detectors" (as most do). This concept

has not been generally understood or employed byvendors or by the reviewed licensees. In someinstances, they have employed uninterpretedactuai calibration data for Xe or 83Kr toestablish detector response, without a recog-nition of their limitations. The former emitslow energy photons, with a mean energy of 0.045MeV per disintegration. Thus, they may besignificantly absorbed in the housing or wallsof a detector. In contrast, Kr is principallya beta emitter, with accompanying bremstrahlunggamma radiations and a 0.51 MeV photon with ayield of only 0.4%. This Is apparent fromFigure 8, which illustrates the direct response

CONCENTRATIOHS FOK VEMDOr CALI1KATTOW6Of I I P.1-1 HIGH KAHCE HOMIQJS

Eber l ine

Mid-Hinge SPINGNGD-1 (SA-13)

Hlgh-IUnge SFINGAXM-USA-l*)

SA-15, SA-9

General Atonici

Mid/High Range-WRCM

Hlgh-Ringe-UKH

0.13

0.26

1.75

0.65

5x10*

0.47

1.47

9.98

U.I*

1.5xlO5*

Baaed on calibration data euppllcd by vendor, asinferred for NBS Reference Dec**

155

I 2X 10

t. I

1 11

=i—M4+-

T

I

y/

j

=1

Figure 6. Ranges of General Atomics wide-range gas monitor.Figure 9. KDGM-HR enhanced detector in KSG-HRH saapler,

enhanced high-range position energy dependenLecharacteristic.

Figure 7. KHK-HRH-Enhanced Mgh-r.ngc geometry. Flgur* 10. General Atomlca wide-range ga^ monitor RIK72high-range detector energy reiponae curve.

Figure 8. Reeponae of Eberllne SA-9 Mgtj-range detector to 85Kr and l33Xe.

Figure 11. HDGH-HR Enhanced detector In KS(7-HRH enhancedhigh-range position efficiency to Xenon-l-'S.

156

VII

lndcpundunt Ut i l i ty

Ho.

5

Vendor

1

1

Intern

4

3

2

3

1

Range

-

Devlgn

-

-

i ted UnltB

Hld/Hlgh

Mld/Hlgh

Hld/Hlgh

Al l

Mld/HICh

De«lRn

Vendor

-

NKC In

Raman

CA

Eberlt

Kjioan

SA1

liber 11

SAHH.INC ABM ANALYSIS at HLANT KffLUKHTS. H . f . 1 - 2

Sample

Yea1

HAP-5

HRH

WKCM

KGM-IIRH

RACKMS

A X H - l

i

1

3

I

3

1

1

YeB

No

Yea

Yes

Local/remote control

Fixed

Hawk*

(la, each Instance)

TlMd •••pie

'fined BSIiple

Automatic

Fixed (CM Monitor)

tlaied Ba>|

Nore I

Note: One licensee has Installed this system, but doe» not utlllce Its Ce-1.1 detection feature.

Sorae licensees have recognized the variableenergy response of high-range monitors by theprovision of corrections in their software formaking off-site dose assessments. However, thisdoes not provide guidance tor % reactor operatoror supervisor who may have to make manualcalculations of effluent release rates beforeskilled post-accident dose assessors are likelyto be available.

As indicated in Table VI, three licenseesselected the Eberline SP1NG-4 as a high-rangemonitor for effluent noble gasses. During thereviews, it was ascertained that the micro-processor of this monitor is not radiationhardened, thus making it doubtful that it wouldoperate reliably in high-radiation fields.However, in one case the monitor was supple-mented by the Eberline SA-10 and SA-9 mid- orhigh-range detectors, for which the sensitivecomponents are remotely located. When theSPING-4 component of this unit senses highradiation fields, it is isolated from the samplestream, thus increasing its reliability offunction throughout an accident sequence.

in several instances, licensees withinstalled micro-processor controlled high-rangegas monitors were found to have a limited numberof plant personnel with sufficient training tobe able to retrieve data beyond that routinelydisplayed. The review also revealed thatseveral of these monitors had experiencedfrequent and/or extended down time of theirautomatic features, due to the failure of theirflow sensors which appear to be sensitive toentrained dust particles and which thereforecall for frequent preventative maintenance.

Except for those with installed integratedunits which function automatically, provisionand/or procedures had not been incorporated bymany licensees for the isolation and/or purgingof their low-level gas monitors, should theirrange be exceeded. Thus their recovery andavailability would be doubtful following anaccident as effluent concentrations declined towithin the low-range region.

C. SAMPLING AND ANALYSIS OF PLANT EFFLUENTSThe principal deficiency encountered in the

review of arrangements for the sampling ofradioiodines and particulates was the inabilityof licensees to document that their samplingsystems could collect representative samples.This is particularly so for those with longsampling lines, in which considerable depositionlosses of elemental radioiodines could occureven when installed in accordance with thedesign guidance of ANSI N13.1-1969.

The transmission of elemental iodinesthrough long sampling lines has been measuredunder controlled conditions in the laboratory byUnrein et al. Their studies suggest that itdepends upon the relative rates of depositionand resuspension from th^ir walls. Transmissionfactors greater than 50% were found for 1"sampling lines at flow rates of 2-3 cfm, forinjection periods of several hours. However,these studies did not indicate how long it tookto reach equilibrium between deposition andresuspension after an initial injection. Only asmall fraction (<1%) of the injected elementaliodine was transmitted through the 1/4" samplingline with a 0.06 cfm flow rate as utilized inthe General Atomics WRGM, which is shownschematically in Figure 12,.

The NRC's proposed guidance suggests thatthe closest approximation to representativenessmay be achieved at equilibrium, when depositionand reentrainment or re-suspension are equal.This could be expected to occur most rapidly ina continuously operated system, rather than onein which flow is initiated only upon the occur-rence of higher-range concentrations. The Kamanand the Eberline AXM-1 monitors approximate thisIn that, upon an Indication of abnormal gasconcentrations, they isokinetically obtain asmall local side-stream flow (of a few hundredcm /min) from the low-range monitoring/samplingline, in which a much greater flow (1-2 cfm) ismaintained.

157

From the reviews, it was apparent that mostarchitect/engineers and ltcensees have beenaware of the need for the heat tracing ofsampling lines wh^n they are exposed to "out-door" conditions • However, it was also apparentthat many of them have not recognized a similarneed for the heat tracing of long Indoorhorizontal sampling lines in which condensationcould occur, especially under the high moistureloads of some accident sequences. In a fewreviews condensation was found in the samplingmedium of sampling positions.

Although II.F.1-2 calls for continuoussampling, the procedures of five licenseescalled only for the analysis of a grab sample tobe col/c-'ted post~accid"nt over a short periodof time (to limit the amount collected \.o thecapability of their laboratory Ge-Li analysissystems), with no Indication in their proceduresof how they would evaluate the preceding sampleto establish the total amount released from theonset of accident conditions.

In several instances, which included thethree SPING-4s, the three SA1 KAGEMS and onelicensee devised installation, the filterassembly for the collection of partlculates andiodines was either unshielded or inadequatelyshielded. None had conducted an analysis toassure that with such an arrangement, thesamples could be collected, retained and trans-ported within the GDC-19 dose limits (5 remwhole body and 75 rera to the extremities). Itshould be noted that by two successive 1/200dilutions, the RAGEMS should collect only rela-tively low activity samples under all accidentconditions.

All of the licensees had Ag-Zeolitecollection media available for sampling underaccident conditions. Almost all of the Instal-lations provided for lsokinetic sampling atnormal stack flow rates, but only a few couldmaintain it if large deviations from these flowswere to occur under accident conditions. Ofthose that could not, none had developed cor-rection factors, as called for in Item II.F.1-2.

Only a few licensees had developed adequateprocedures for the analysis of "hot" samples, inwhich the collected activity might considerablyexceed the upper limit which could be analyzedby their GeLi counting and analysiB systems.Although several had established procedures forcounting samples with greater than normal ac-tivity In a geometry distant from the detector,only a few would be able to cope with samplesapproaching the 85-170 Ci of radioiodines whichwould J>e collected at a concentration of 100uCl/cm at normal flow rates of 1-2 cfm for thestipulated 30-minute sampling period.

VI. COMMENTS ANU RECOMMENDATIONS

Except for the GE designed PASS, which wasbasically the same except for variations insample line arrangements and the associatedvalves jt individual facilities, a wide varietyin PASS systems were encountered in the reviewswhich have been conducted over the past twoyears. Many required frequent and considerableattention to keep them fully operational and all

required frequent retraining to maintain oper-ator proficiency with their controls and theirdetailed operating procedures.

The example of the one In-line system thatIs also used for routine sampling suggests thatthe readiness and availability of the othersystems could be enhanced if they were too werealso periodically used for routine sampling, inbetween the Infrequent occasions they are util-ized for exercises or mandatory retraining.

A wide variety of approaches to themonitoring of noble gases and the sampling ofparticulars and radioiodines in highconcentrations have also been encountered in thereviews.

If the monitoring requirements were solelythose for the noble gases, ion chambers wouldseem the most straightforward detectors, in viewof their simplicity, wide range capability, andlinear energy response characteristics.However, they are relatively insensitive andtherefore require a large volume of containedgas which is difficult to shield from extraneousradiations. An example of one such installationis shown in Figure 13. The 0.1" -thick steelwall in which the detector was housed would have:a large absorption for low energy photons, suchas those from Xe, compared to a much smallerabsorption of the higher energy photons fromshorter-lived noble gases.

The integrated monitoring/sampling deviceswhich incorporate microprocessor data handlingand control accomplish the full range require-ments of Item Il.F-1.1 by routing the flow tomore then one detector, each of which is de-signed to be sensitive to portions of the fullrange requirement. This permits the isolation ofthe low-range detector during periods of highconcentrations. It also facilitates the routingof flow to a selected shielded filter assemblyat the same time. Their ability to store and toprovide a history or release rates over timemakes them attractive for both routine andaccident monitoring. Additionally, the use of 3monitor for every-day purposes adds to theirreliability for accident monitoring. If not soutilized, they require regular surveillance andmaintenance to assure their availability.

Much of the confusion over the use of theXe equivalent concept In the calibration of

high-range noble gas monitors could bc> elminatedby the adoption of the "Ci-Mev" concept asdescribed by Mourad. A simplified version ofthe same concept, which utilizes the averagenoble gas energy as a function of time post-shutdown as shown in Figure 14 in dose calculations,was described by Lahti at the 1985 AnnualMeeting of the Health Physics Society.

To minimize the ambient post-accident radi-ation fields, most of the post-accident monitorsand/or samplers are located at considerabledistances from the points of effluent release,thus necessitating long sampling lines {typi-cally 1" x 100-250'). fhis creates a dilemmabetween the desirability of maintaining a highflow rate In the sample line so as to minimizedeposition losses and the desirability of mini-mizing the amount of collected radioactivity on

158

ISOKINETICNOZZLES

rrr-rrrrr

HIGH RANGE tO.Ot SCFM)

NORMAL RANGE (2 SCFM)

RETURN LINE

SAMPLECONDITIONER ,

13)WIDE-RANGE

GAS DETECTORS

15)READOUTS ELECTRONICS

PURGEA(R IN

INTERCONNECTING PIWIG ISQUO UNEI AND WINING IDQnEO LINE!IV UtfN

BLOCK DIAGRAM, WIDE-RANGE GAS MONITOR

Figure 12. Block diagrsn, wide-range gas monitor. F i g u r e 13< H l g h _ r a n g e n o b l e g a a wmLtMm

Avtng* nobl* ga> energy In M«V it*

Ltg«nd

A: 2-hour rtlciaaB: 4-hour relcaicC: t-hour reltaat

Tim* ift«r«huldown of nh iu (houra)

Figure 14. Average Noble Gas Energies of Total Releases.

159

Che sanpler. It is solved in some monitors, bythe provision of a second stage of isokineticsanpllng with a probe situated within the high-flow line close to the sampling head, but with amuch small flow (a few hundred cm /min) throughthe high-concentration sampler. This seemsdesirable on the grounds of both convenience inhandling and analysis and of ALARAconsiderations.

ACKNOWLEDGEMENTS

The participation in the gathering andinterpretation of data in r.hese reviews by JohnBaum, Alan Kuehner, Edward Lessard, RobertMiltenberger, Stephen Musolino, Norman Rohrig,and Christopher Weilandics of the staff of theSafety and Environmental Protection Division ofBrookhaven National Laboratory and of MarieMiller of the staff of the Emergency Prepared-ness and Radiological Protection Branch of theUS NRC, Region I is gratefully acknowledged.

Appreciation is also appropriate for thetypists, Marie Cooney and Cheryll Christie whoprepared the text of many revisions of thereports and of this manuscript.

REFERENCES

1. US NRC, "TMI-2 Lessons Learned Task ForceStatus Report and Short-Term Recom-mendations", NUREG-0578 (July 19/9)

2. US NRC, "Action Plans for ImplementingRecommendations of the President's Commissionand other Studies of the TMI-2 Accident",NUREG-0660 (May 1980)

3. US NRC, "Clarification of TMI Action PlanRequirements", NUREG-0737 (November 1980)

4. M. Rogerin et al "Three Mile Island: A Reportto the Commissioners and to the Public" USNRC (1980)

5. C.C. Lin, "Procedures for the Determinationof the Extent of Core Damage Under AccidentConditions", NEDO-22215 (August 1982)

6. A.P. Hull, "A Preliminary Dose Assessment ofthe Chernobyl Accident", BNL-38550 (Sept1980)

7. US NRC, "Reactor Safety Study", Table VI,3-1, Appendix G, WASH-1400 (1975)

8. W.G. Pasedag et al, "Technical Bases forEstimating Fission Product Behavior DuringLWR Accidents", Table 4, 5, NOREG-0772 (1981)

9. L.L. Borzon and N.A. Lures, "Best EstimateLOCA Radiation Signature", NUREG-1237 (1980)

10. L.A. Cunningham, "Counterpart Meeting on PostAccident Sampling <PAS,S) and RadiologicalEffluent Technical Specifications - April 24-26, 1984" Internal NRC Memo (5/16/84)

i ' i ••'• ! - .,

11. D.G. Eisenhut, "Proposed Guidance for Cali-bration and Surveillance Requirements forEquipment Provided to Meet Item II.F.IAttachments 1, 2 and 3, NUREG-0737", Memo-randum to Regional Administrators (Aug. 16,1982)

12. D. McClure, "Accident Range Monitor Response"HPS Newsletter, 18:6 pg. 7 (1985)

13. J.W. Holloway, "Calibration of EberlineSPING-4 High Range Noble Gas Monitors atSusquehanna Steaa Electric Station", RCT-8307(Feb., 1983)

14. P.J. Unrein et al, "Transmission of Radio-iodines Through Sampling Lines", Proceedingsof the 18th DOE Nuclear Airborne WasteManagement and Air Cleaning Conference, CONF-840806, Vol. 1, p. 98-114 (1985)

15. R. Mourad "The Use of Ci-ttev in Evaluatingthe Effective Doae Equivalent from Noble-GasReleases to the Atmosphere" Presented at the5th Annual Meeting of the Canadian RadiationProtection Association (CRPA), April 30- May3, 1984. Author's address: Atomic Energy ofCanada, Sheridan Park Research Community,Missisauga, Ontario L5K 1B2.

•••t

ANS Topical Meeting cm. Radiological Accidents—Perspectives! awl Emergency Piavuig

A State-of-the-Art Approach to Emergency Preparedness—Remote Monitoring of Nuclear Power Plants

James A. Blackburn and Michael C. Parker

ABSTRACT Immediately following the nuclearaccident at Three Mile Island, the State ofIllinois began to design a state-of-the-artemergency preparedness system for the thirteennuclear power reactors within its borders.This system incorporates an on-line reactorparameter data communication link, an on-lineautomated isotopic gaseous effluent monitoringsystem, and gross gamma monitors installedaround each site. Liquid effluent monitorswill soon be installed also.

The sensitivities and capabilities of thisremote monitoring system have been clearlydemonstrated both during abnormal events atthe reactor sites and during emergencypreparedness exercises. These experiencesreadily illustrate the system's abilityrapidly to provide comprehensive, technicaldata to the Department's staff should an acci-dent occur at an Illinois reactor site.

I. INTRODUCTION

Immediately following the nuclear acci-dent at Three Mile Island in 1979 f the Stateof Illinois began to design a state-of-the-artemergency preparedness system for the 13nuclear power reactors being operated orconstructed within its borders. In 1980, theIllinois Department of Nuclear Safety (IDNS)was formed, initially by Executive Order ofGovernor James R. Thompson, then by confirminglegislation. An immediate challenge was toprovide a mechanism whereby governmentalagencies would receive more timely andaccurate information regarding both the radio-active composition and the magnitude of anyaccidental release of radioactivity to theenvironment.

The initial goals of the Remote Monitor-ing System (RMS), in the event of accidents,were: to analyze as accurately as possiblethe discrete radioactive components beingreleased from the reactor site; to assess themagnitude of their radiological impact on the

populace; and to transmit the results of theanalyses to the Departmental decisionmakers asrapidly as possible. It wa3 quickly realizedthat predicting radiation doses to members ofthe public was almost impossible without addi-tional knowledge of reactor conditions. Since1984, the RMS has been significantly enhancedand expanded by incorporating on-line informa-tion regarding the status of essential safetysystems at the plant. Crucial factors intheRMS design were to provide reliability inthe acquisition and transmission of data andto minimize the amount of time required ofutility staff for collection and transmissionto IDNS.

The current objectives of the RMS arethreefold: early warning of nuclear reactorevents having a potential off-site impact;fast risk analysis of reactor systems; andrapid identification, quantification, andverification of a radioactive release to theenvironment. Each of these objectives playsan essential role in assuring the ability torecommend prompt off-site protective actions.

II. SYSTEM DESIGN

The Illinois Department of NuclearSafety's Remote Monitoring System incorporatesthree major components: gross gamma detectorsradially positioned around each nuclear powerstation; on-line automated, isotopic gaseouseffluent monitors; which sample from majorengineered release points; and an on-linereactor parameter data communication link toeach facility's on-site computer. Inaddition, on-line liquid effluent monitors,which will be located at each plant's liquiddischarge points, are scheduled forinstallation at two sites within the nextyear. All RMS components are connected,through dedicated data communication links, tothe IDNS Radiological Emergency AssessmentCenter (REAC) located in Springfield,Illinois. There, a technical staff comprisedof nuclear engineers, health physicists, andother nuclear safety specialists reviews thedata and performs analyses of plantconditions. Thi3 REAC staff is divided into

161

162

two analytical groups: one concerned with thestatus of reactor safety systems; the otherwith environmental assessment.

Confirmation of a gaseous release to theenvironment is accomplished by a network of upto sixteen pressurized ion chambers installedradially around each plant at a distance ofapproximately two miles. These monitors havea dynamic range of 2.6 x 10"'° to 2.6 x 10"^ Xunits per hour (10 to 10 Roentgen3 perhour). The 3iting of these instrumentsinvolved a balancing of many factors, includ-ing maximizing the probability for plumedetection, while minimizing plume transit timeto the detector and the potential for error,either due to observing building shine fromthe containment or sky shine from an elevatedpluae. Using typical Illinois atmosphericconditions, the off-site system will detect aplume with a centerline reading of twomilliRoentgens per hour approximately 90percent of the time. This criterion fordetection requires one monitor to reportvalues double the levels associated withnatural background. Non-detection is limitedto extremely stable conditions with winddirections bisecting the distance betweenadjacent detectors. Since these atmosphericconditions can be easily described, theDepartment is planning to develop software torecognize 3uch conditions rapidly and displayan appropriate warning for the analyticalstaff.

Quantification of a release is accom-plished through radiological analyses of datasupplied by the IDNS gaseous effluentmonitor. This monitoring system ia compli-cated, incorporating high purity germaniumdetectors and gamma spectroscope to identifyan<\ quantify components of the isotopic sourceterm, i.e., those radioactive contaminantsreleased into the environment, whether partic-ulates, volatile iodines, or noble gases.This Instrumentation has a dynamic range of3.7 x 10~9 to 3.7 x 102 Bq/cc (10~13 to 10"^uCi/cc)

ong,

radiological data gathered by this gaseouseffluent monitoring system is continuallyreceived and updated by a combination of PDP11/31 and 11/70 computers. Software has beendeveloped to compute atmospheric dispersionand postulate environmental exposures from arelease, based upon current meteorologicalconditions and effluent radioactivity levels.

Early warning and risk analysis of reac-tor events are accomplished utilizing the IDNSData Link (DDL). DDL receives approximately1200 to 1500 key reactor and engineered 3afetysystem parameters every two to four minutesdirectly from each reactor's on-site com-puter. Early warning of abnormal reactorconditions will soon be provided through soft-ware which will monitor each plant's engi-neered safety system configuration, identifythe presence of key abnormal events, andindicate proper operabillty of engineeredsafety systems. Additionally, DNS is re-

for the particuljate and iodine sta-tions, and of 3.7 x 10~5 x 3.7 x 10° Bq/cc(10"" to 10^ uCi/cc) for noble gases. The

searching the use of expect system technologyto develop a decision-aide to diagnose reactorsystem 3tatus during abnormal events and toapprise the REAC team of possible sequencesleading to a release of radioactivity.

The analysis of the liquid effluentstreams may use only gross gamma monitoringrather than complete isotopic analysis. Sucha reduction in complexity and cost is beingconsidered due to the longer lead time beforethe onset of exposure to the general public bythe water pathway and the relative ease withwhich such exposure can be reduced or averted.

III. CURRENT STATUS

At the present time, the gamma monitorsare installed and operating around allreactor sites, providing baseline data forthose facilities still under construction. Agaseous effluent monitoring system has beenoperational at LaSalle since 1982. Additionalsystems are currently under construction forinstallation later this year at Zion andDresden. The contract has been signed forfabrication and installation of systems atQuad Cities, Byron, Braidwood, and Clintonduring the next four years. The DNS data linkis installed and operational for all reactorsexcept Braidwood. Braidwood will beincorporated into the DDL network prior toinitial loading of fuel. As previously men-tioned, the liquid effluent monitoring systemsare currently undergoing initial review a3 apilot project for the Zion and Dresden facili-ties. Upon satisfactory completion of thatevaluation, additional systems will beprocured and installed at all Illinois reactorfacilities.

IV. DDL SOFTWARE DEVELOPMENT

Although the DDL analysis .software isonly in the first of three developmentalstages, it has already become a aajor facet ofthe Departmental approach to Emergency Prepar-edness and Response. The initial phase in DDLsoftware development is to establish a database with access to both historical andcurrent data. At the present time, a four-daydata base is available for instant access.The basic capability is to display data for upto seven signals chosen at random from amongthe 1200-1500 points available from eachreactor. These data may be either sequentialtwo-minute measurements or samples chosen atspecified iiilei «ai=>. The user is also allowedto track real-time trends of the chosensignals. Under this mode, the oldest data,approximately forty minutes old, automaticallyscrolls off the terminal screen as the currentinformation is received. An alternate formatallows real-tine display of the current statusof up to twenty different signals, althoughthe trending capability Is sacrificed. Thislatter format is truly user selectable, havingthe flexibility to group points associatedwith a givan subject, a given reactor, orselected at random from the system at large.

163

The second phase of software developmentfor the DDL system will be to design an "earlywarning" alarm into the system. This featurewill consist of software programs which willconstantly monitor various data packages forindications of abnormal conditions, such anloss cf off-site power. Although not neces-sarily an indication of accident conditions,such degradation is important to assess theability of the facility to respond adequatelyin the event of an accident. The Departmentalresponse to such an alarm will be to alert thereactor analysis staff to ensure its awarenessof the situation.

The third and final phase of DDL softwaredevelopment is presently envisioned as anexpert system to aid the reactor analyst inhis ability to determine the current status ofreactor safety systems, project in a fa3ter-than-real-time environment the possible conse-quences of the abnormal events and advise onprotective measures off-site. It is not theintent or the desire of the Department todirect utility remedial actions based on thisdata. DNS will use this information only tothe extent that likely consequences of anevent can be analyzed in support of off-siteprotective actions.

V. TECHKICAL DIFFICULTIES

The development and operation of such anextensive remote monitoring system has had itsshare of technical difficulties and chal-lenges. The off-site monitors, for example,require the use of a 300 volt battery which isboth relatively expensive and short-lived. Inaddition, impending failure is characterizedby erroneous indications of increasing radia-tion levels, sometimes as high as ten to fif-teen times normal levels. Additional problemsencountered with the off-site detectorsinclude loss of electrical power, failure ofthe telecommunications link to Springfield,vandalism, and susceptibility of the elec-tronic eompononts to power surges caused bylightning.

Despite these difficulties, over the pastyear (July 1985 through June 1986), the off-site component of the system, comprised of 96detectors with their modems, approximately2130 kilometers (1325 miles) of telecommunica-tion lines, two computer systems with atten-dant hardware peripherals and required soft-ware, achieved a system-wide operability fac-tor (availability of data for display inSpringfield) of 93.1 percent.

To increase the reliability of the off-site monitors even further, the Department isdeveloping a 300 volt power supply to replacethe dry cell battery, researching the abilityto use solar collectors for back-up power, andevaluating radio communications to a centralcollection point rather than leased lines to atelecommunications bridge. In addition, theDepartment has already installed electricalsurge protectors on the power supplies toreduce susceptibility to lightning.

The complexity of the isotopic gaseous

effluent monitoring system creates complicatedproblems. A major cause of system failure isfailure of off-the-shelf components and a lackof environmental support systems, rather thanproblems with the custom-designed components.For example, the system has failed severaltimes following failure of the air condi-tioning system for the Technical SupportCenter (TSC), the location for the system'son-site process computer. Although a TSC isrequired by KflC regulations, apparently theoperability of its support systems is not.Sometimes, substantial effort is required toobtain repair of this essential support systemby the utility.

A recent failure of the effluent monitor-ing system was the result of loss of electric-ity to the TSC. Although supplied with back-up generating capability, the circuit breakerfeeding the system had been manually throwndue to a "noisy transformer" which wasannoying a worker in the area.

To eliminate these types of operationalproblems, the Department is requiring theconstruction of a dedicated building on-site. This building will house the entiregaseous effluent monitoring system, includingits process computer and dedicated supportequipment such as back-up power supplies,heating and air conditioning, and instrumentair.

Not all of the problems with the effluentmonitoring system are the result of variablein-plant environmental conditions. A substan-tial amount of difficulty was finally tracedto microphonics in the coaxial cables connect-ing the gross count rate meter to the multi-channel analyzer. Since this gross indicatoris designed to signal the system to rapidlyincreasing radiation levels within the sample,the presence of these spurious signals causedthe system to decrease the analysis tines andconsequently the system's sensitivity.Although the system never specifically faileddue to this problem, the ability to detectradioactive contaminants being released intothe environment was substantially impaired.

In spite of these difficulties, the gase-ous effluent monitoring system has perfomedwell, particularly for a prototypicalsystem. Over the past year (July 1985 throughJune 1986) the system's three stations, par-ticulates, iodines, and noble gases, wereoperational 86.7, 83.3, and 89.8 percent,respectively.

In contrast to the approach used for theoff-site and the gaseous effluent monitoringsystems, the DDL incorporates no State-ownedsensors or instrumentation, relying exclusive-ly on utility installed and maintained instru-mentation. The Departmental components to thesystem are limited to a pair of statisticalmultiplexors, dedicated telecommunicationlinks, and the REAC computer system. To mini-mize the cost and time required for systeminstallation, this component is currentlylimited to a subset of the signals presentlyresiding on the utility's computers at thevarious reactor sites. Such an approach mini-

164

mizes the direct responsibility for properoperation and reliability, although at asacrifice of control and assurance that thesystem will be functional when needed. Theutility13 computers, for example, may beremoved from service for preventivemaintenance with little, if any, advancenotice to the Department or Departmentalinput. In addition, utility personnel have,on occasion, unilaterally altered the contentsof the transmitted data stream, adverselyimpacting the Department's data analysiscapability.

VI. OPERATIONAL RESULTS

The primary intent of the RMS is to pro-vide the capability of independently assessingthe possible off-site impact of an accident ata nuclear power reactor. Such a capability isa major support in recommending specificprotective actions for the general public.Since monitoring effluent streams duringnormal operations is of secondary importance,only a limited amount of data analysis hasbeen performed to date. Gaseous effluentsfrom the reactors have remained below theminimum detectable levels of the off-sitemonitoring system. Environmental radiationlevels nominally average between seven and tenmicroRoentgens per hour due to natural back-ground. During times of precipitation, how-ever, the ambient levels are substantiallyelevated, often reaching as high as fifteen totwenty microRoentgens per hour.

The data generated by the gaseous efflu-ent monitoring system has routinely indicatedthe presence of radioactivity in each of thethree stations, when the reactor is a^power. Typical fission products includingisotopes of Cobalt, Manganese, Sodium, Iodine,Krypton and Xenon, for example, were identi-fied by the system during May, 1986. Researchis being done to match the ongoing resultsobtained from the gaseous effluent monitoringsystem and the daily grab samples gathered bythe utility, as a portion of its Appendix Irequirements. Although not yet available forpublication, these comparisons appear to cor-respond remarkably well, considering the vari-ation in methodology used for obtaining andanalyzing samples.

Late in January, 1966, a small leak de-veloped in one or more of a reactor's fuelrods. The DNS Gaseous Effluent MonitoringSystem readily identified both increasedlevels of radioactivity and additional nu-clides within the effluent stream. Th« noblegas station, for example, reported effluentconcentrations averaging 2.7 x 10"1 Bq/co (7.4x 10 uCi/cc) for the 64 analyses performedbetween January 23 and February 1. This com-pares .with an average of 3.4 x10"^ Bq/cc (9.4x 10 uCi/cc) for the 96 analyses obtainedbetween January 1 and January 15. By combin-ing these data with DDL parameters, such asoff-gas radiation levels and the readings fromthe utility's effluent radiation monitors, theDepartment was able to verify the presence of

this leaking fuel.The DDL was directly utilized by the

Department during the feedwater transientevent at the LaSalle station on June 1,1986. The system did not document the waterlevel as being less than the low water levelscram setpoint, due to the transient's shortduration when compared to the two-minute sam-pling frequency. However, the sy3tem clearlyshowed the initial downward trend of the waterlevel and other affected parameters, bothduring the initial event and the subsequentmanual shutdown of the reactor. During thisevent, the utility telephoned the BEAC athourly intervals, indicating the power levelsand the reactor status. By monitoring DDL,technical staff within the REAC were ableindependently to verify the status of individ-ual control rod banks, reactor power levels,reactor water levels, etc.

VII. EMERGENCY USE OF THE

Although the RMS monitors the activitiesof the Illinois reactor facilities duringnormal operations, its purpose is to providedata during emergency conditions at a nuclearpower reactor. During a major accident, theREAC staff would initially analyze the variousDDL parameters for trends and to diagnose theproblem. The reactor analysis staff wouldconcentrate on the status of the reactor'sengineered safety systems in an effort topredict key trends and their consequences,while the environmental analysis staff wouldfocus on containment radiation levels, arearadiation monitors, and effluent concentra-tions. In anticipation of a possible radio-active release into the environment, the envi-ronmental analysis staff would also calculatedose reduction factors due to atmosphericdispersion, using the meteorological param-eters available from the DDL. If the situa-tion warrants, calculations would be performedto determine the protective actions whichwould give the least dose to the generalpublic living in the immediate environssurrounding the facility.

If a release occurs, its presence wouldbe verified by both the DDL and the gaseouseffluent monitoring system, and the postulatedimpact on the general public calculated.Using the isotopic source term provided by thegaseous effluent monitoring system, the envi-ronmental analyst would also be able to pre-dict the presence of released radionuclideswhich would contribute to ground contamina-tion. Verification of the wind direction andthe anticipated off-site radiation levelswould be available from both the environmentalmonitors and the Department's field teamsdispatched to the area. Following the termin-ation of the release, the isotopic sourceterm, in conjunction with contamination read-ings obtained from the mobile field teams,would be used to determine and document theintegrated dose to the off-site population.

Although the iiemote Monitoring Systemdoes not solve all problems associated with

165

assessment of reactor accidents and their In providing accurate and tlnely InfornationImpact upon the general public, the Illinois to decision makers during a radiologicalDepartment of Nuclear Safety believes that accident at a nuclear power reactor,this approach is a technological advancement

Section HI

Computer Applications

Chairman: R. J. CatlinElectric Power Research Institute

ANS Topical Meetaf oa RadioloficaJ AccMeatf—Perspectires aid Eawrgeacy

Computer-Assisted Emergency Preparedness and ResponseJames W. Morentz

ABSTRACT. No tool a v a i l a b l e to emergencyplanners and response personnel offers moreactual and potential management capabil ityimprovement that the microcomputer. The lowcost , fast access speed, data format f l e x i b i l -i t y , data manipulation v e r s a t i l i t y , and colorgraphic display clarity make the microcomputer— when applied appropriately — an extraordin-ary resource .for the emergency manager to drawupon. A computer program called the EmergencyInformation System* i s in use around severalnuclear powerplants. The system helps describeand clarify situations at a nuclear faci l i ty forknowledgeable observers. It i s used for commun-ication with surrounding jur i sd ic t ions , and i seven linked into s tate emergency operationscenters where the broadest emergency decisionsare made. The EIS uses computer—generated mapsto display evacuation routes and merges thegraphic displays with several databases of valueto emergency planners: special populations,resources, traffic control, alert l i s t s , plans,and others. The microcomputer, operating aloneor in a network of up to 20 computers, providesreal-time decision support to emergency responseoperations staff.

While a nuclear powerplant stays in onelocation, an emergency sets in motion a swirl ofevents that traverse geography, demography, andpolit ical jurisdictions. The spatial movementof an emergency i s , we believe, the single mostcomplex aspect of i t s effective management.

No tool available to emergency planners andresponse personnel o f f er s more actual andpotential management capability improvement thanthe microcomputer. The low cost , fast accessspeed, data format f lexibi l i ty , data manipula-tion v e r s a t i l i t y and color graphic displayclarity make the microcomputer — when appliedappropriately — an extraordinary resource forthe emergency manager to draw upon.

This paper describes actual experience andpotential applications of a computer softwarepackage for nuclear emergency planning. Thecomputer Program, called the Emergency Informa-

tion System*, has been developed by the author'sfirm for use on-s i te at a nuclear f a c i l i t y , incommunication with surrounding jurisdictions,and even linked into state emergency operationscenters where the broadest emergency decisionsare made.

I. NUCLEAR FACILITY EIS

At a nuclear faci l i ty , the Emergency Infor-mation System becomes an important tool of theemergency planner. The combination of database*which can be related geographically give the on-s i te planner superior capabilities in emergencyresponse analysis.

Let us take three examples, warning sirens,special populations and public-technical infor-mation on accidents, a l l of which are responsi-b i l i t i e s shared by the local government and thefacility.

The special capability of the EIS is tolink geographical data with text data. In thecase of warning sirens, the following scenarioillustrates the power of this tool.

Siren sound propagation distances areentered on the geo-relational database of theEIS. When sirens are ordered to sound, thetyping of the word "siren" brings to the screena map of the evacuation zone and circles Corother shapes) showing tht s i ren soundpropagation. If a siren is found to fail (via acall-back system or report from the field), theuser selects the failed siren froa a l i s t .Pressing the [Mjap key returns the evacuationzone to the screen and the failed siren's entirepropagation zone is blinking. The user thenmoves the screen cursor into the blinking arearepresenting the failed siren sound and pressesa single key. Appearing almost instantly is theplan for replacing the warning siren with mobilesirens, loud speakers, or door-to-door contact.Exact street names, responsible agencies.contact persons and phone numbers, and thenumber of people to be warned all appear. Inseconds, the emergency manager can confer withthe responsible authorities and set in motionalternative warning methods.

Now. for another example, let us look atspecial population needs. The EIS is equipped

169

170

with a geo-relational special emergency needsdatabase. Contained on the database is infor-mation on the name of the facil i ty, address,phone, contact name, number of people at thedependent care faci l i ty, notes on evacuationplans, the agency responsible for carrying outthe evacuation, the lead time necessary for anevacuation to take place, the shelter locationto which people w i l l be moved, and otherimportant information.

Of premier utility to the emergency manageris the fact that all of this information is alsoreferenced by geographic location. Thus, withthe press of a few keys, the user has displayedon the screen's map the location of each specialcare faci l i ty. Pressing another key. a textl ist ing is displayed showing the name of al lfac i l i t i e s , the number of people at each, andthe total number of people to be evacuated fromthe care facilities.

With the press of two keys, the user canidentify how many busses, for example, withwheelchair l i f t s are necessary to evacuate aspecific nursing home. Then, with the press ofa single key, the user can return to the map andhave that specific nursing home appear, blinkingat its proper location on the map.

As a final example of on-site applicationsof the Emergency Information System, we refer toproviding public-technical information about theThree Mile Island boiling water reactor. Theversatile graphic capabilities of the EIS haveenabled us to "zoom" right inside the plant todisplay a graphic depiction of the steam cycleof the plant. An initial graphic encompasses theentire steam cycle. Moving to any of the threemain parts of the cycle and pressing a singlekey "zooms" the user to a detailed graphic ofthat portion of the steam cycle, as shown below.

Then, using the geo-relational databasecapability of the EIS, the user can request textinformation about any of the component parts ofthe reactor.

For example, in i t s development of theThree Mile Island EIS, the Pennsylvania Emer-gency Management Agency focused on public-technical information; that is , generally avail-able public information of a technical naturethat could be used to inform knowledgeable mediaand aid decision-making. The focus of the textdeveloped about Three Mile Island was on theimpact of a reported failure of the componentparts of the reactor.

Thus, if a report comes in that a particu-lar pump has fa i l ed , public informationpersonnel at the plant, or emergency personnelresponsible for contacting governments, have attheir fingertips a sophisticated graphic andtext display to use to provide appropriatepublic—technical information. The speedy datadisplay process is as follows:

First, the user "zooms" in on Three MileIsland and the steam cycle. Second, the usertypes the word "pump." Instantly, circlesappear around all the pumps designated as vitalto the functioning of the facility. Third, theuser can move the cursor directly to the pump ofchoice if i ts location is known, or switch to asummary text screen and select the right pumpfrom a l i s t of pump names. If the l i s t approachis selected, the user can go directly todetailed text descriptions or return to thesteaa cycle graphic where the selected pump willbe blinking to note its location. Fourth, fromthe graphic or the summary text display, aseries of full text displays are retrieved.These describe the function of the specific pumpselected, the implications of i t s failure, and

171

the potential for expanded emergency conditions.The full text description may be a single screenor many screens, each of which are displayedless than a second after the user presses asingle key.

The value of this graphic and textcapability for briefing the public news media orgovernment executives with explicit technicalinformation accompanied by excellent graphicsshould not be underestimated. If one thing hasbeen made clear by past incidents, i t is thateffective communication is essential. The EISas an executive briefing tool is unexcelled.Six different types of graphic pointers areavailable on the system (cross-hair, checkmark,arrow, hand with pointing finger, hand held instop position, and user-sized outline box). Theuser can select the most appropriate pointinggraphic at the press of a key, changing i tirstantly to match the presentation's need.Thus clarity of presentation in graphic and textbecomes one of the improved products of the useof the Emergency Information System.

I I . OFF-SITE AUTOMATED PREPAREDNESS AND RE-SPONSE

The Emergency Information System iscurrently used in large and small municipalitiesand counties in this country and abroad. It hasbeen purchased by nuclear faci l i t ies and othercorporations for the use of local governments.As a result, the EIS can serve as an importantinformation link between powerplante and theoff-site emergency responders in local andcounty government. The EIS has been used inradiological emergency planning dr i l ls andexercises. It is upon that exercise experiencethat this and the following section are based.

Off-site emergency preparedness is aidedtremendously by the Emergency Information Systemin planning, evacuation route monitoring, andresource management.

The EIS Automated Emergency Plan allowsspeedy retrieval of standard operating procedurechecklists for a l l types of emergencies. Forradiological emergencies, the Automated Planallows the user to quickly arrive at specificl i s t s of assignments for all personnel. Thesecan be used in advance of an emergency forprinting out plans for everyone to review andfor training new personnel. During an emer-gency, top management can use the plans as achecklist of performance to monitor whetherspecific benchmarks have been achieved. Follow-ing an emergency, plan evaluation and improve-ment are aided by the computer. Debriefingmeetings can be held about plan u t i l i ty , inwhich a user calls up each part of the plan insequence. As recommendations are made, they canbe typed directly onto the computer for all inattendance to verify. Almost instantly, then, anew or revised plan section is ready fortraining, exercising, and use.

Evacuation route monitoring is readilyaccomplished with the EIS. At the local level,three elements of evacuation are crucial: routeplanning, alternate route identification, andtraffic control point staffing.

Using the EIS, a local governnent candesign an evacuation route system that takesinto account a l l prior facility planning. Thegovernment can then enter routes as part of theEIS (as shown on county-level map on the follow-ing page), providing street names, number ofpeople to be evacuated, and other importantinformation. The evacuation routes can be codedby sector and can each identify a lead time orclearance time and make note of the destinationof evacuees. When the user requests evacuationroutes to be displayed, they are drawn on thescreen as red lines, and a text l ist ing of allroutes can be generated.

Alternate routes can also be entered ontothe system so that if a main route is blockedthe user can quickly display alternate routes(as dotted or dashed lines of red, blue, oryellow) and obtain a l ist of those route names.If a main route is closed, i t and i t s alternateroutes can be set to blink on the map as analert to al l personnel to the deviation fromplans.

Resource management is another off-siteapplication of the EIS to radiological emergencyplanning and response. The EIS can be used tolocate dosimeters, monitor their deploymentaround the Emergency Planning Zone (EPZ), andrecord their recall after an exercise or emer-gency. In the evacuation stage, the EIS willidentify each Traffic Control Point, i ts staff-ing numbers, responsible agency, and lead timefor staffing. (See EPZ-level evacuation map onnext page.)

With these brief examples i t should beclear that the advantages of microcomputer de-cision support tools for radiological emergencyplanning and response are great. The advantagesare even greater when the computer is viewed aspart of a communications network linkingpowerplants, local governments, and the stateemergency office. In the next section we willdiscuss the aspects of the EIS that bring thestate into the network.

I I I . STATE EMERGENCY OFFICE COMPUTER APPLICA-TIONS

The s ta te emergency operations center (EOC)has a key mandated ro le in rad io logica l emer-gencies. The EIS supports t h a t ro le not onlythrough better information f i l t e r e d up throughloca l governments by means of the automatedEvent Log, but also by giving the s t a t e EOC afirm grasp on the SPATIAL DIMENSIONS of theemergency.

Far removed as i t normally i s , the s t a t eEOC nevertheless i s usually responsible fortremendously important decis ions regarding aradio logica l incident . In order to make theproper decisions, the s ta te EOC must be able toprovide an accurate p ic ture to the Governor ofthe options and consequences of decisions.

There i s no better tool for achieving thisend than computer graphics. And, we suggest,the geo-relational data management system thati s the backbone of the Emergency InformationSystem brings to emergency decision-making al lthe best a t t r ibutes of this new tool.

172

County-Level Map of Evacuation Routes

EPZ-Level Evacuation Map

173

At the state EOC, the EIS can portray anaccurate picture of the geographic aspects ofthe emergency through a series of "zoom" mapsthat fully depict the EPZ from 50 or more milesout right down to the crucial intersections.Each map is separately digitized, showing al lthe information that an EMERGENCY MANAGEMENTDECISION-MAKER needs. Thess are, above a l lelse, decision-makers' maps, not geographers'aaps.

Most importantly, the EIS does not .stopwith maps. It builds into the decision-supportsystem the geographic representation of data.Thus, when considering an evacuation, the state-level decision-makers can "play" with "What if?"options and graphically see their outcome.

"What if we decide to evacuate in onehour?", the user might ask. The EIS would showall reception centers that can be opened in onehour. Or in two hours, if desired.

What if a plume spread northwest and we hadto move all people immediately out of i t s pathto the east and southwest? What evacuationroutes are available in this area (and the userdraws a rectangle on the screen)? Shownimmediately on the screen are the evacuationroutes. A text l is t ing of the routes is alsoavailable at the press of a single key.

Now, what about Access Control Points? Ifwe evacuate that zone, where will our ACPs beplaced? Again, ACPs appear as dots on the map.Those selected for staffing in this specificevacuation can be set to blink by the user.

Now, the Governor might ask, what if wedon't do a northwest sector evacuation but do anen t i r e 10-mile zone evacuation? Show meeverything I've just seen for the whole EPZ.And in a few seconds the interrogation of allthe constituent geo-relational databases of theEmergency Information System is complete and theGovernor has the options.

IV. SUMMARY

Microcomputers have opened t o l o c a l andstate emergency managers capab i l i t i e s fordecision analysis that were possessed only bythe military in the past. As our technologicalsociety has become more complex, at least onetechnological advance is now available to helpgovernment and industry leaders sort out thecomplexity and make the best possible decisionswhen i t comes to public safety.

ANS Topical Meetiag oa Radiological AccMeats—Perspectives awl Eatergeacy Plaaaiag

FEMA'S Integrated Emergency ManagementInformation System (IEMIS)

Robert T. Jaske and Wayne Meitzler

ABSTRACT - FEMA is implementing a computer-ized system for use in optimizing planning,and for supporting exercises oE these plans.Called the Integrated Emergency ManagementInformation System (IEMIS), it consists of abase geographic information system upon whichanalytical models are superimposed in orderto load data and report results analytically.At present, it supports FEMA's work inoff-site preparedness around nuclear powerstations, but is being developed to deal witha full range of natural and technologicalaccident hazards for which emergency evacu-ation or population movement is required.

I. INTRODUCTION

At the present time, IEMIS is opera-tional at FEMA Headquarters, the 10 FEKARegions, the National Emergency TrainingCenter, twelve States and a number of localjurisdictions.

The IEMIS computer equipment is dividedinto three major components: the operationalfield workstations, the FSMA mini-computerand data base, and networked user nodes.

Presently, a field workstation supportseach Regional Office. In the future, eachRegion would became a node with its own mini-computer acting as a local hub for State andlocal use. The field workstation consists ofa medium resolution color graphic terminal tomake system requests and display results, aletter quality printer to print alphanumericreports, a color hacdcopy device to providehardcopy of the infonnation on the colordisplay, and communications equipment inter-connecting the field terminal and the FEMAinini-ccroputer.

For the users located near the FEMAmini-computer, high resolution workstationsare available to display results, and filmrecorders to make hardcopy of the displayed

information. The high resolution terminalsare also supplemented by large screen projec-tion, 2x3 meters.

Both the field workstations and the highresolution workstations connect to the IEMIScomputer which performs the modeling calcula-tion and data storage. The system is builtaround commercially available hardware. Thecomputer is a Digital Equipment CorporationVAX 11/750 with four megabytes of theinternal memory and two gigabytes of diskstorage, magnetic tape units, and a lineprinter. At present, over 50 field work-stations are operational in the ISMISenvironment. Included are IBM compatiblePCs which can be adapted to serve a3 IEMISterminals. Any VT 100 compatible terminalmay connect to IEMIS, but graphic functionsrequire special features.

As currently operating, TEMIS includesthe following elements:

a. Computer utilities, word processing,(VMS/TCMS/Datatrieve/INGRES);

b. Emergency Information CoordinatingCenter elements-message functions, notifica-tion, weather abstracting, team briefing;

c. The National Map and resources database as a geographic information system;

d. Program data bases, such asRadiological Emergency Planning Program;

e. Exercise Evaluation and SimulationFacility; and

f. Colorgraphic chart and graphpreparation.

Other ccmpoi ,nts will be added as modelsan program data bases are integrated into thesystem. Under development at this time are:hurricane evacuation planning and exercising,transportation accident evaluation, manage-ment of accidents involving toxic andhazardous chemicals, dam failure, conflagra-tions and earthquake.

FEMA is working with State and localgovernments in the development of data basesand in developing the application of IEMIS tothese program areas. In order to stimulate

175

176

interest, FEMA makes developmental copies ofIEMIS in executable fonnat to state andlocal governments, associatel universitiesand contractors.

II. NATIONAL MAP AND GEOGRAPHIC INFORMATIONSYSTEM

Presently, IEMIS contains a digitalcartographic data base derived from the1:2,000,000 scale, sectional maps of theNational Atlas of the United States. Thiscolor graphic version is a unique adaptationof the original work by the United StatesGeological Survey, and is arranged WithinIEMIS for unlimited pan and zoom for theentire United States. All displays may bemade at one of six standard map projections,and the data structure will handle globalinputs including International Dat? Linetransactions.

As implemented in IEMIS, the mappingsystem includes political boundaries, trans-portation networks (roads and railroads),hydrographic data features (streams and waterbodies) elevations, and names of geographicand populated places. These data wereentered into the computer in a format thatrecords topologic relations with relatedfeatures on the source map. This formatallows graphic applications, such as drawingstreams and roads for autonatic map plotting,as well as analytical applications, such aspopulation and area calculations, andchecking data for consistency and accuracy.The format also permits display of maps acany scale and introduction of data from mapsof any scale.

An important feature of the attributecoding is that individual features withineach category are ranked from the mostsignificant to the least significant. Thisscheme allows the user to select a minimalamount of data and selectively increase thisamount to the level of detail needed tosupport the theme and scale of the desiredmap. The data can also be grouped togetherin various ways to produce logical sets ofinformation.

As implemented, the system permits thecutting out of precise sections, performingunlimite.1 editing? additions, color changing,and reinsertion of edited sections. It is,thus designed to be continuously maintainedwithout service interruption. Each user maycreate unique maps filed in a personal filefor use in support of individual projects orin specialized instruction, or alternativelyreturn edited map segments to the master datafile which may exist in several locations.These functions can be performed off-line onPCs for economy in developing local databases.

The user can obtain the location of anypoint in latitude/longitude, linear distancesbetween any points, and the perineter of anypolygon. The user can also add or deletetext, color or shade any polygon, and findtile population or resources data in any

streams for evacuation (nodeIs. The systemalso includes the geographic place nameslisted in the National Atlas. FEMA is alsoplanning to upgrade the National Map systemsubstantially over the next several yearsinto a generalized geographic infornationsystem.

The USGS 1:100,000 scale maps presentlybeing prepared for the 1990 census (by 1988),will be integrated into the system as thequadrangles are digitized. Digital eleva-tions and resources data in topologicallystructured layers may now be enteredinto the data base.

As part of the ongoing map applications,FEMA maintains a Memorandum of Understandingwith the USGS whereby the parties agree tothe 1:2,000,000 scale map to 1:100,000 scale.As (NMSS) segments are produced by USGS theywill blend into the FEMA system. The twoagencies will share a common, computerizedstatus management sub-program maintained onT.FMTS and available to all users.

III. PRINCIPAL IEMIS COMPONENTS

While the long range goal of IEMIS is tosupport the entire range of emergencies forwhich planning, exercising and response isthought feasible, as a practical matter,developments progress in an incrementalfashion. At this time, the following programareas are addressai, with a conment as to thestatus of development.

Table 1

Radiological EmergencyPreparedness

Hurricane EvacuationPlanning

Hazardous Materials

Crisis RelocationPlanning

Dam Break

Status Ccmnent

Supports allregulatoryobjectives.

Model in advancedstate, preliminaryplanning possible.

Diffused gasmodeling and eva-cuation support.

Preliminary worknow possible.More data neededin system.

Base maps andelevations encoded,coding in progress.

As a means of describing basic ISMIS capa-bility, a short discussion of the majormodels involved in the programs abovefollows. The model suite is available toall the program areas either singly or inmultiple combination. For example, theMESORAD model will interact with the I-DYNEVevacuation model and allow studies of

177

concentrations of highly diffused radio-isotopes which could occur as a result ofpower accidents. Since as an interim step,it reports concentrations of duffused (notsubject to gravity effects) gases, it canalso estimate concentrations of highly toxic,diffused chemicals over the simple terrain(no deep valleys or shore breeze effects)where two dimensional puff models areapplicable.

MESORAD canputes the dose arising f ranthe release of radioisotopes listed in thejoint NRC/FEMA Preparedness Guide NUREG-0654/FEMA-REP-1 Rev. ltU. MESORAD consistsbasically of the diffusion code MESOlt2] anda simplified radiation dose code described byScheL-pelz et al, 1985 t3l. Outputs from thediffusion code are considered geometricallyand dose calculated using a semi-infinitecloud assumption over a grid of 1000 cellswhose size can be varied from a minimum of0.4 mi 2 (1.04 km2) to a maximum of about10 mi 2 (26.8 km 2). This rather coarse repre-sentation provides a relatively macroscopicestimate. This almost matches the approxi-mate size of neighborhoods in evacuationcentroids of the evacuation model and allowsintegration of the dose and transportationmodels by applications of a shelter factor toeach population centroid. This model formatis computationally efficient and allows quickestimation of dose by successive interactionin support of plans reviews, exercises anddrills. In its existing form it is notintended for response activities.

The techniques utilized in MESORAD con-tribute to advancing the state-of-the-sciencefor real-time dose assessments with thedevelopment of a new method to treat thefinite cloud approximation. The method,entitled the Discrete-Point Approximation,is used in the estimation of external doseswhen the dimensions of the cloud are smallcompared to the mean-free-paths of gammaradiation. This method, used in place of apoint-kernel integration, provides a twodecade savings in computational ti-te at thecost of typically a one percent (but up to10 percent under sane conditions) differencein estimate. This represents an acceptabletolerance in light of the decades uncertaintyin source term and, to a lesser degree,atmospheric dispersion.

A revised version of MESORAD is nowoperational on the system. This revisedversion may, after a suitable period oftraining and indoctrination, become a firstorder reference model for some responseactivities.

IV. EVACUATION - I-DYNEV, INTERACTIVEDYNAMIC EVACUATION

I-DYNEV is the principal model frameworkwhich ties a number of sub-models into anintegrated subsystem. The main source of thesubsystem is the U.S. Federal HighwayAdmininstration model named TRAFU) described

by Liebermanl^J before the January 1982meeting of the Transportation Research Board.TRAFLO was developed to operate at threelevels: Level I treating individualvehicles, Level II treating a macroscopicgrouping of vehicles in clusters, and LevelIII a gross representation in terms oftraffic parameters. I-DYNEV uses the levelIT process, treating groups of vehicles bymsuns of movement specific histograms.

TRAFLO is actually a system of modelscomposed of:

a) A macroscopic urban network modelcalled NETFLO;

b) A macroscopic freeway model calledFREFLO; and

c) A traffic assignment model.The traffic assignment model used ori-

ginally in TRAFLO is the equilibrium modelTRAFFIC developed at the "Universite deMontreal", which has been extensivelyvalidated. It is planned, however, to alsointerface other assignment models withTRAFLO.

The physical traffic environment, whichmust be specified as input data by the userin order to exercise f.^ I-DYNEV system,consists of the following features:

a) Topology of the roadway system;b) Geometries of each roadway compo-

nent;c) Channelization of traffic on each

roadway;d) Motorist behavior which, in aggre-

gate, determines the operational performanceof vehicles in the system;

e) Circulation pattern^of traffic onthe roadway system;

f) Specification of the traffic controldevices and their operational characteris-tics;

g) Traffic volunes entering and leavingthe roadway; and

h) Traffic composition.To provide an efficient framework for

defining these specifications, the physicalenvironment is represented as a networkcomprised of uni-directional links and nodes.The links of the network generally representurban streets or freeway sections. The nodesof the network generally represent urbanintersections or points where a geometricproperty changes (e.g., a lane drop, changein grade or a major mid-block trafficgenerator) 15].

Over a period of several years, I-DYNEVwill be supplanted by a more generic formwhich will be capable of simulating evacua-tions ranging from point source (typicallyhazardous materials sources) to regionalevents (typically hurricanes).

V. REGIONAL EVACUATION (REGEVAC)

I-DYNEV is being expanded to deal withregional problems which require a higher

178

degree of origin/destination, control andaccountability. Mien complete, the regionalevacuation system will allow the user tospecify which destination nodes are asso-ciated with each origin node. This is beingdone to allow specific assignment of popula-tion groups by neighborhood or district tospecific relocation centers. Wiile thiswas done in the trip distribution phase oEI-DYNEV, this distribution was tentative andwas to some extent overruled by the tripassignment and ultiinate loading pattern forthe simulation. A more specific directionand control strategy is thought necessary inlarge cases where the impact aone of thehazard is variable (e.g. hurricanes) andrequires neighborhood assignment to shelterareas when the onset rate of the developingincident exceeds the dispersal rate of evac-uees to individual destinations. Data fromhurricane evacuations show significantincreases in persons using public shelterswhen the hurricane suddenly changes courseand threatens areas previously thoughtrelatively secure.

REGEVAC will contain 3 levels of model-ing (2 for I-DYNBV) of which two!15} arecurrently available for preliminary work.It allows a preliminary development of tripdistribution based on the available highwaynetwork and control tactics h*iich provides afirst approximation. Then based on the knownpopulation dynamics, a trip assignment canbe executed which readies the input factorsfor the final or simulation model whichdetermines the actual traffic performance.Because of the sizra o£ the problem, thesimulation model will operate on a segmentednetwork, but except for data entry andmodification purposes is transparent to theuser.

Features of REGEVAC not possible withthe existing I-DYNEV include the ability tocreate networks directly in the iteritivegraphic mode, to modify existing links inpreviously defined cases, and to signifi-cantly relate the denial of links to a systemas progressive flooding from an advancingflood wave. These features will signifi-cantly reduce manual coding burden, decreaseerrors previously created by incompletelycoded network changes and allow flood levelinput from other models such a'* the NationalWeather SLOSH models, the Corps of EngineersHEC-1 and others. In addition, throughtesting with pilot projects involving realtime flow measurements of both water andtraffic, FEMA plans to work with localjurisdictions in the applications of REGEVACas a means of building public confidence.

REGEVAC is designed to operate inconjunction with a computerized geographicinformation system in which resourceselanents with locational significance can beidentified, incorporated into traffic flowplanning, and ultimately used to guiderecovery operations. Such attributes asdemography, highway information and sheltersand public facilities can be attached to

visible map vectors and incorporated directlyinto modes which simulate planning options.

The mapping data of choice at this timeis a combination of 1:2,000,000 NationalAtlas Data and the USGS/Census 1:100,000scale data for streets, roads and streans nowbeing digitized for the 1990 Census. Inaddition, FEMA is supplementing these datawith facilities' data from its extensivefiles for ultimate use for a number of majordisaster problems involving transportation,energy and fuel distribution, populationprotection and national resources management.

VI. Hazardous Materials

Currently, FEMA is deploying advanceelements of IEMIS related to hazardous andtoxic materials accidents. The U.S. CoastGuard Hazardous Assessment Conputer System(HACS) is operational as a study sub-modelsuite, and a heat radiation envelop model isbeing deployed as a training aid. As amulti-purpose element applicable to manyhazards, a siren alerting, sound propagationmodel is being integrated into the colorgraphic, menu driven format. Releases ofvapors and aerosols following assumptionsgoverning diffused gases can be modeled withthe concentration sub-model of MESORAD.Several excellent heavy gas models are beingevaluated by the Department of Energy, andone will be selected for inclusion in IEMIS.

As the data within the national IEMISGIS is developed, the universal capabilityto support accidents froti fixed and mobilesources will develop. This capability isalso applicable to the modeling of radio-active exposure from accidents to mobilefacilities.

VII. NEAR TERM DEVErjDPMENTS

Because of the technical achievementspossible with the system, IEMIS is beingexpanded to serve as the basis for a generaldistributed data processing systan combiningthe elements of a geographic informationsystem and a suite of action oriented simula-tion models which can be selectively broughtto bear on disaster management problems.

VIII. FUTURE; DEVELOPMENT

While the future is by no means assured,IEMIS provides the basis for a truly nationalsystem using a fully distributed data base,ccr ton data exchange standards, and fullaccess by all components o£ civil government.

° In conjunction with a pilot projectin the Tulsa, Oklahoma area, FEMA willdevelop software which will operate anationally accepted code for simulating danfailure and incoporate it into the system.Using a combination of data picks from thesimulation, produce an evacuation plan forthe manageTient of the threatened area. Inaddition, FEMA will support development ofmeans to integrate flood insurance maps.

179

° FEMA is also arranging to integrate asample of a metropolitan data base created bya gas utility company with the operationalsimulation capability described in thispaper. In this formal demonstration, thesoftware will be used to show how utilitylines and services can be integrated withtransportation models and hazardous impactmodels in order to create a metropolitanplanning instrument. In addition, becauseof its high quality, the data base will bebased to illustrate other uses for digitaldata of this type such as vehicle management,real estate development and cadastral appli-cations. Subway planning and operationcould also be integrated.

Ey the completion of the 1990 Census inthe United States, a national geographicinformation system will be operational tosupport a multitude of administrative andcaronercial activities. Through data exchangestandardization, all parties with interestsin public resources should be able to par-ticipate.

On the equipment side, we see thedevelopment of computing equipment continuingunabated through 1990 such that the ordinarypersonal computer will be a very powerfulmachine, probably a 32 bit address, virtualcore with at least 10xl0*> bytes of randomaccess memory. Such machines will be exten-sively networked using common file protocolsand will exchange information at unprece-dented rates over both land lines and radiofrequencies.

Thus, the emergency management communitywill become a full partner to the burgeoninguse of electronic media and will be able toavoid the necessity of building immense databases for statistically infrequent events byusing everyday geographical and cadastraldata bases in a special way.

FEMA fully intends to maintain theleadership posture evident in IBHIS, andwithin the achieveable time limits of tech-nological adaptation, assist in creation ofwhat could become an international system.

STLU rcr/EUKi ggg/ige2D0M INamourRESTORE PLOTRUNSTEPEXITTEXT DEH.BY DISABLEDICOti DISPLAY DISflBUDCONTINUEPICK FDR DOSE REPORT

6FEB.86- 12:30CUMULATIVE UNDEPLETED

PLUrc CONCENTRATIONLOG QLPNTITY-H0UR/ri**3

OUTER ocrrrauR «-i?.EftCH IKKEI-gNT* 3.

n u m . PARTICLE rem

SCALE: 1 DIY • £.?E t i l .

fiDHtNISTRfiTH£ lOUMOflRYPOLITICAL BOUHMIESRODS 1 TWOSUATE* BODIES

1B-HOE S"Z GRID _EUftCUffTiiQN LINK-NODE NET

28-PUG-1986 1E:03

Figure 1 - Section of Map for Fort Pierce, Florida area with hypotheticalrelease from St. Lucie Power Station.

180

IX. REFERENCES

U.S. Nuclear Regulatory Commission andFederal Emergency Management Agency,(Cr i te r ia for Preparation and Evaluationof Radiological Emergency Response Plansin Support of Nuclear Fewer Plants . )NUREG-0654/FEMA-REP-l, Itev. 1,October 1980. (Available from FEMA)

Ramsdell, J . V. , Athey, G. P . , Glantz, C. S."MESOI Version 2.0: An In terac t iveMesoscale Lagrangian Puff DispersionModel with Deposition and Decay",NURB3/CR-3344/PNL-4753, Pac. N.W. Lab.Richland, WA, Sept. 1984. (Availablefran Nuclear Regulatory Commission)

Scherpelz, R . I . e t a l , "MESORAD DoseAssessment Model", NUREG/CR-4000,PNL-5219, Pacif ic Northwest Laboratory,Richland, WA, August 1985. (Availablefrcm the Nuclear Regulatory Cownission)

Lieberman, E. , "Macroscopic simulation forUrban Traffic Management", Vol. 1,Executive Summary, FHWA-RD-80,Washington, DC, June, 1980. (Availablefran FHWA, DOT, Wash, DC)

Lieberman, E . , "Formulations of the DYNEVand I-DTOEV Traffic Simulation Modelsused in EESF", TR 154, KLD Assoc.,Huntington Sta t ion, N.Y., March 1984.(Available from FEMA)

Description of an Integrated Trip Assignmentand Distr ibut ion Model (TRAD) for theI-DYNEV System, KLD: TR-187, KLD Assoc.,Huntington Sta t ion, N.Y., Apri l , 1986.(Available from FEMA)

ANS Topical Meetiag oa Radiological AccMmts—Perspective* ami EaKrgeacy Ptaaaiag

The MAPSS Nuclear Emergency Management SystemHoward A. Price, Jr., Larry D. Sadler, and Robert W. Johnson, Jr.

ABSTRACT: With the MAPSS NEMS in place,a distinctly higher level of confidenceis imparted to the emergency personnelmanaging a nuclear plant accident. Thegraphics are unique and are the samethroughout all levels of response.Standard Operating Procedures areavailable instantly to all personnel,and system security providespartitioning of critical levels ofcommand as well as documentation of allactions taken through the emergency.Maintenance is reduced to a very simplelevel and the entire system operates atmeaningful life saving speed.

INTRODUCTION

The MAPSS Corporation is a smallhigh-tech company with experience in thedesign and implementation of computer-based systems. MAPSS personnel havedesigned and implemented systems forseveral governmental agencies, includingthe Department of Defense and theDepartment of Energy. They range fromsmall business systems, running onpersonal computers, to nuclearmaterials tracking systems involvinglarge scale mainframes. MAPSS personnelhave considerable experience in thedesign, manufacture, and application ofspecialized nuclear instrumentation.MAPSS has developed a system, calledMAPSS=MAGIC which is flexible enough tobe the core of the MAPSS NuclearEmergency Management System (NEMS). TheMAPSS=MAGIC system is the culminationof more than 7 years of design anddevelopment. MAPSS personnel haveworked in conjunction with the nuclearindustry for over 20 years. Some ofMAPSS consultants have made significantcontributions to the nuclear industry.

DEFINITION OF THE PROBLEM

The MAPSS Corporation has learnedthat, in spite of the intensive effortwhich has been made since the Three MileIsland (TMI) accident in Pennsylvania,and the subsequent establishment of theFederal Emergency Management Agency(FEMA), no satisfactory command, con-trol, and communication informationsystem has been established at the stateor federal level.

The experiences at Three MileIsland (TMI) and more recently atChernobyl, USSR, dramatically illus-trate the need for effective controland communications at every operatingnuclear installation. TMI andChernobyl have demonstrated what can andwill happen if an adequate nuclearemergency management system is not inplace. If TMI and the USSR had had aresponsive operating nuci?«»r emergencymanagement system, much of the panic andconfusion that resulted would have beenavoided. Vet today, more than sevenyears after TMx, no computer basedstate-of-the-art system is in use and notwo manual systems now in place arealike.

Existing varning systems, not onlyin the United States but throughout theworld, are tr the great majority apiecemeal and awkward combination oftelecopiers, telephones and manual pro-cedures. Even with the stand alone PCbased systems now on the market, consid-erable time is still wasted in passingcomplex details via the telephone withthe recipient copying data on paper byhand. The telecopying of materialbetween operations centers often suffersfrom network jamming because oftelecopier overload. In addition,telecopiers are unreliable because offrequent equipment failures at criticaltimes; telecopier performance hasdropped lower than 50% in someexercises. In short, the existing

181

182

communication systems have been stressedbeyond their capabilities in exercisesand most likely will not work during anactual emergency. It is obvious that animproved method of handling thecritical information is necessary foradequate and timely decision making.

In order to establish a commonground on which to address what an emer-gency management system should do, thefollowing list of the minimum require-ments have been identified. Require-ments may be different for particularapplications, but an emergency manage-ment system should:

1. monitor situations and direct theactivities of emergency personnel,resources, and equipment;

2. noti fy the affected areas in theevent of an accident;

3- inform and djjr_ect the media in aresponsive manner;

4. prepare for, coordinate, and conductevacuation as events dictate;

5. conduct, training of emergencypersonnel; and

6- £59SJ^lSi§L%.?. and standardize operatingprocedures for emergency situations.

The MAPSS Corporation has developeda comprehensive Nuclear Emergency Mana-gement System (NEH5) which addressesthese requirements from a "systems"

point of view rather than the fragmentedapproach of the existing systems. TheMAPS5 NEMS is designed to be usedprimarily by personnel with little or nocomputer background. Information trans-mission, which now can take minutes,hours or longer, can be accomplished inseconds.

The MAPSS NEHS permits theestablishment of a command, control andcoinniijiii rations network which will allowthe timely flow of information amongparticipating parties. These may be anyone? of several levels of government:from federal to state to localgovernment. The system will communicatewith any of the previously mentionedagencies in any combination or sequence.

The I1APS5 NEMS is a microcomputer -based system designed to aid in thecommand, control and coordination ofemergency response teams in the event ofan accident at a nuclear installation.Since the system is built around amicrocomputer, it is portable in that ifan Emergency Operations Center (EOC) hasto be relocated, the system may berelocated with the personnel. Thesystem would require a telephone lineand 110V AC to become operational. TheMAPSS NEMS uses a distributed processingapproach which incorporates:

o a high-speed local area network;

o remote networking;

o voice and data transmissioncapabiliti es;

II .'i Ml: ' .YMI M

MOM',I

Kl YHflAWI!

C O I O K

I ' H I N I I H / I ' l <i ! I I |J

MAI'Mi MSMII MM KAU

IliM l'(:I Mill MOl i

Cf'l OH CNAI ' I I ICM Oll l l l I

in r mi

I I K I I I / I K

111 MO I ICD HI It III

II O X

( K C I - i )

> I AHI.I SC.KI I NI I P . I ' I AY ( I M l )

y Ml I) If. I H I I 1(1 A l l

IV roi OK

M o N11 n I;

I'lUK.I V.IIK

IJM A( i 7 I I I I I I A I I I I )

V O I C If l ' I I K )

183

o inulti-tasking real-time oper-ating system;

o high resolution color graphicsdisplays;

o up to 34OM bytes of hard disk;

o color printer/plotters;

o graphic digitizers;

o video output drivers; and

o IBM PC corapatiblity.

The software which drives the NEMSis a combination of a menu and functionkey driven system. The MAPSS NEMS takesfull advantage of the real-time multi-tasking operating system in that mostof the communications are done inbackground. The software is written inFORTRAN 77 and assembly language. Thesystem responds to input from externalcomputer links. The operator mayinitiate various functions, eitherthrough menus or the use of functionkeys. However, as defined by systemrequirements, certain functions areavailable only at predetermined timesand are accessable only by authorizedusers. An example might be that theoperator can run the plume model onlywhen the; situation has reached the SiteArea Emergency level. The entire systemis password protected: Passwords areused for two purposes, first, torestrict access to the system toauthorized users only, and second, torestrict access to critical informationor software routines. This featureinsures that only authorized personnelhave the capability to calculateevacuation recommendations, run plumemodels, ov review the event log. Theentire software system is implemented ina modular fashion; which permits theaddition of new features and/orenhancements to the system withouthaving to recompile the entire softwaresystem. For instance, the plume modelto be used may not be decided on untilthe system design phase.

SYSTEM OVERVIEWThe managing utility has respon-

sibility for informing the appropriateEOCs of an emergency situation and anysubsequent escalation or closeout. TheMAPSS NEMS monitors the managing util-ities' Emergency Command Center (ECC)via a computerized link and will auto-matically notify the EOCs of anyemergency situation. On command fromthe operator at the EOC, the StandardOperating Procedures (SOP) for theparticular situation will be initiated.When the EOC has been informed of achange in conditions, it will inform the

appropriate state agencies. The EOC hasthe additional responsibilty of keepingthe news media aware of the situationand coordinating news releases for thepublic. The MAPSS NEMS will permitauthorized users at the utility, state,and local level to review and composeguidance messages and media releases.All news releases can be viewedsimltaneously at the appropriate sites;and once agreed on by all parties, therelease is printed and distributed tothe media. If the situation requires,the MAPSS NEMS can feed video directlyto the news media via standard video(RGB) outputs. The image on thesystem's monitor would be the same Oi>ethat would be released to the media.This allows important messages,evacuation routes and sectors, etc. tobe viewed by the public. Informationfrom the various radiation monitoringteams can be entered into the system andnew contamination patterns can be calcu-lated. This process is accomplished byusing a contouring technique which willcombine both real data from the fieldand the projected data from a plumemodel. In addition to monitoring theECC's computer system and supplying datato the outlying operations centers, theMAPSS NEMS automatically maintains andupdates an accurate record (Event Log)of all incoming and outgoing messages,as well as the acknowledgements tothose messages. This Event Log ismaintained at all times, not just whenan emergency is in progress.

At the city/county level during anemergency the operators have, the abilityto display the location of the shelters,the status of those shelters, theevacuation routes, the pickup points,paired schools, and roadblock status.If during the operation of the system,a new incoming message is received, theoperator will be notified by an audibletone and a flashing one-line message inthe upper left hand corner of thescreen. This feature insures that allmessages are received and are availablefor recall. The city/county can alsoput its own data in their system that ispertenient locally such as city/countySOPs, phone list of key personnel,radio frequencies and call signs,equipment status, other city/countyofficals phone numbers, or.. whateverelse is necessary for the city/county tomanage a emergency. The city/county issupplied with data directly from themanaging agency and coordinates itsactivities with the EOC.

If the communication link is lostbetween EOCs, each EOC has the abilityto run in a stand alone fashion. Itwill update its Event Log with theexact time communications were lost andwhat actions were taken and by whomduring the loss. When communications

184

have been reestablished, the Event Logon the controlling master computer willbe updated to reflect the actions takenat the EOC which had been "lost".

EXAMPLES OF THE SYSTEMAll of the examples used are from

the State of Tennessee. AlthoughTennessee's reporting network may besomewhat different, the philosphy anddata flow of the systems are verysimilar. The following examplesillustrate the graphics capabilities ofthe system and demonstrate itsflexibility.

NORMAL MODEThe following is an example of what

the screen would display during NORMALcondition. Notice that the date/timegroup will be updated as will thecondition of the entire notificationnetwork. The system is displaying thename of the duty officer at theEmergency Operations Center (EOC). Italso displays the location and status ofall the outlying centers. In addition,the system displays the currentcondition of each of the nuclear powerplants and the operating status of eachindividual unit. If the color of thesite or the plant is green, the

condition is normal; if the color isblue, the site or plant is offline; ifthe color is violet, the site or theplant is not responding to the com-puter's polling. The system willdisplay the status map of the affectedarea on a large screen display (LSD) andit will be updated at regular intervalsreflecting situation changes.

UNUSUAL EVENT MODEThis is an example of the screen

when an Unsual Event occurs. The statusof the plant has changed to yellow, theplant site as shown on the state map isnow yellow, and the unit status at theplant is also yellow. At this time thesystem will start to update at a muohfaster rate. The system normallyupdates once every ten minutes. Once anUnusual Event has occurred, the updaterate is increased to once every fiveminutes. since update time varies witheach system, the times given here aremerely examples. As the situationcontinues to degrade, the update timewill be reduced as needed until thesystem becomes fully dedicated to aparticular event (full dedication wouldnot take place until a General Emergencywas declared). The notification networkcontinues to check each outlying site

TENNESSEE STATUS MAP

DATE: 23JAN86TIME: 13:21 CST

STATE DUTY OFFICER: ANDY EDLYNUCLEAR POWER PLANT STATUS

SEQUOYAH PLANT CONDITION: NORMAL

UNIT 1 100%UNIT 2 OFFLINE MAINTENANCE

WATTS BAR PLANT CONDITION: NORMALUNIT 1 100%UNIT 2 100%

STATE NOTIFICATION NETWORK:o n r A O HAM EOC (SON) McMIN EOC (WBN)S E O C B R A D EOC (SON) MEIGS EOC (WBN)

p r r r FCC/RMCC (SQN) St^lSfrCECC JIC(SQN) ra^gjoc

185

TENNESSEE STATUS MAP

DATE: 23JAN86TIME: 13:30CST

STATE DUTY OFFICER: ANDY EDLYNUCLEAR POWER PLAN! STATUS

SEOUOYAH PLANT CONDITION: NORMAL

UNIT 1 100%UNIT 2 OFFLINE MAINTENANCE

WATTS BAR PLANT CONDITION:UNIT 1 100%UNIT" ? 100%

STATE NOTIFICATION NETWORK:, . , - „ _ HAM EOC (SON) McMIN EOC (WBN)SEOC B R A D EOC (SON) MEIGS EOC (WBN)

n c n n FCC/RMCC (SQN) RWAbOCCECC MC(sa«) FCC^MOC

UNUSUAL EVENTPRESS <G0>

and report the status of each even ifthe site is not involved in thisparticular event.

The next picture is the StandardOperating Procedure that would come upon the operator's display when he/shepresses the <GO> key on the keyboard.Some of the sites have responded and acouple have not. Those sites which havenot responded would be displayed inviolet. At this time the operator wouldverbally contact the agencies or personsnot responding to the alert. As therespondents react, their status willreturn to the normal color code which iswhite. At the same time, the event logis updated each time someone is notifiedand it will also record who acknowledgedthe message. However, if the operatormust verbally contact any party, theoperator must log manually the time ofnotification. To notify a selection,the operator simply places the screen'scursor on the selection and presses the<G0> key. In this example Rad Healthand the Highway Patrol have not beennotified. It is obvious that the systemcontinues to monitor the entirenotification network (SNN).

EVACUATION RECOMMENDATIONSThis picture shows evacuation

recommendations based on the predictedplume from the plume program. The MAPSSNEMS will calculate which evacuationsectors within the 10 mile EPZ will becovered by the projected plume and makeevacuation recommendations based onthose projections. The plume program tobe used would be the one selected by theutility and integrated into the MAPSSNEKS. The system will give evacuationrecommendations based on the timeintervals selected during the plumecalculation process. In this examplethe plume has been calculated for 0 to 4hours, 4 to 8 hours, and greater than 8hours. The system has identified thosesectors which would be affected by theplume and has listed them to the rightof the picture and color coded them tocorrespond wi th the color coding of theplume. It should be noted that MAPSSdoes not write plume models; MAPSS willintegrate the plume model that theuti]ity selects.

CONCLUSIONEven though the MAPSS NEMS is

designed primarily for nuclear emergencymanagement applications, the system canbe utilized to manage other enprgenicessuch as natural diasters, hazardousmaterials spills, and other catastrophic

186

EVACUAIION KLCOMMLNDAi IONSEVACUATE THE FOLLOWING SECTORS:

0 A HOURS

A I , m, n?, m, o?, 1.1

1 - 8 HOURS

A1, B1, B2, D1. D?, E1

8+ HOURS

i A1. L31.B2. [33. [34, [3S', D1.D2.E1

events that would require real-timeaccurate and responsive emergencymanagement.

Although the MAPSS NEMS is designedprimarily as a nuclear emergencymanagement system, it is versatileenough to be utilized in day to dayoperations. The word processor, spreadsheet, and graphics, which are standerdin our system, can be used for reports,graphs, equipment lists, maintenance andtraining schedules.

MAPSS has developed the most com-prehensive and responsive system avail-able. In order to keep pace with theincreasing demands and complexity ofemergency management, MAPSS continuesto enhance the system. Some of thefuture enhancements may include:

o interface with FEMA's IntegratedEmergency Management InformationSystem (IEMIS); and

o use FM radios and other techno-logies as a backup to transferdata between sites in the eventof phone line failure;

o a version of the NEMS whichwill run on the IBM PC andcompatibles.

The MAPSS Corporation is committedto a systems approach to the emergencymanagement challenge. MAPSS will cont-inue to be at the forefront of emergencymanagement systems by offering acomprehensive cost effective turnkeysystem which includes:

o delivery and installation of thehardware and software;

o training the operators andproviding the technicalassistance;

o expanding the system asrequirements dictate;

o upgrading the system asenhancements come online.

ANS Topical Meeting on Radiological Accidents-Perspectives and Emergency Planning

A Comparison of Computerized Dose Projection Models and TheirImpact on Protective Action Decision Making

Susan M. Reilly

ABSTRACT. The increased use of computers atnuclear power plants provides more accurate andtimely assessment of conditions and consequencesduring emergency situations. Dose projectionhas been one of the prime beneficiaries of thisnew technology, and the resulting calculationshave a direct impact on protective actiondecision-making. In this paper, a quantitativecomparison of three dose projection modelsillustrates that projected doses can vary fromfactors of 2-3 up to factors of 100 or more forsimilar input conditions. The impact onprotective action recommendations is great,resulting in recommendations that vary from "noaction" to "evacuation."

I. OBJECTIVES

This project ha3 two objectives: first, toquantitatively compare three different doseprojection models; and second, to compare theprotective action recommendations that might bemade based on each set of dose projectionresults. The models which have been selectedfor this comparison are: (1) the U.S. NuclearRegulatory Commission's IRDAM model; (2) aatraight-line Gaussian (Class "A") modelcurrently in use at a U.S. utility; and (3) atime-dependent (Class "B") model, also availableat a U.S. utility.

The IRDAM model runs on a personalcomputer, while both utility models run on VAX11/780 mainframes. Each of these models aredescribed in the following section of thisreport.

These particular models were chosen torepresent a cro3s-section of those doseprojection methods currently available In theUnited States. IRDAM provides a particularlyapplicable point of reference since it is themodel used by the NRC site dose assessment teamwhen they respond to emergencies at operatingplants. Since the NRC would be involved indiscussions of protective actionrecommendations, it is useful to know how theSRC dose projections might differ from those of

the utility; and therefore, how protectiveaction decision-making could be affected.

II. DESCRIPTION OF MODELS

A. Model 1: IRDAM

The Interactive Rapid Dose Assessment Model(IRDAM) i3 the model currently used by the U.S.NRC site dose assessment team during emergenciesat nuclear power plants. Version 5.0 (August,1985) was used in this analysis.

IRDAM provides a straight-line Gaussianprojection based on user-supplied input data.The IRDAM program is used for all U.S.utilities, and is designed to reference a givenfile for 3ite-3pecific data upon the userentering the site name. (Such site-3peciflcdata might be: type of plant (BWH/PWR), numberof units on site, etc.) Other input, such ascurrent meteorological and release information,is provided by the user. The output of theIRDAM program is in tabular form, with bothwhole body and thyroid dose rate3 provided atseveral pre-determlned or user-3upplieddistances.

B. Model 2: Utility "A"

Model 2; is a straight-line Gaussian (Class"A") model which produces transport anddiffusion estimates in the plume exposureEmerganoy Plarinlr.g Zone (EPZ). The model isdesigned to run on real-time data: the computeris hard-wired directly into the plant'smeteorological tower and effluent monitoringsystem. It is possible, however, to overridethe automatic mode and manually input data.

Model 2 is structured around an accident"menu" wnich permits the user to provide as muchor as little information as is available. Themore specific the inputs, the more accurate theoutput, but the computer has a set of defaultvalues which are used when no user inputs areprovided (e.g., accident type, time aftershutdown, isotopic mix). Dose projectionresults are provided either in tabular or incolor graphic form. Graphics Include a site mapand a multicolor plume plot.

187

188

C. Model 3s Utility "B"

Model 3 is a time-dependent (Class "B")model which can represent actual spatial andtemporal variations of plume distribution andoan provide estimates of relative depositionwithin the ingestion EPZ. No "eenterline" dosevalues are calculated; values are integratedwithin individual segments of the plume.

Aside from the dose projection calculationsthemselves, Model 3 and Model 2 are verysimilar. Like Model 2, Model 3 also runs onreal-time data, and is menu-driven by theuser. • Data are available in tabular or graphicform.

Both Model 2 and Model 3 have severalcharacteristics which are common to each otherbut which differentiate them from IRDAM (seeTable 1). Key differences are:

o Isotopic mix is a function of accidenttype rather than being based on astandard iodine: noble gas ratio,

o Decay time between reactor shutdown andstart of release is incorporated on anisotope-by-isotope basis,

o Downwind deposition and depletion ofthe plume is considered.

III. VARIABLES IN THE AM1LISIS

Each of the three models was run for thefollowing sets of input conditions:

Hind speeds:Stability Classes:Filtered Release:Unfiltered Release:Time Between Shutdown

and Start ofRelease:

Wind Direction:Release Duration:Gamma Dose Type:

4, 10 miles per hourA through GLoss of Coolant AccidentSteam Line Break

0 and 2 hoursConstent during release2 hoursSemi-infinite

IV. RESULTS OF DOSE PROJECTION COHPABISOMS

Table 3 summarizes the ratios of projecteddose rates from IRDAM to those of Model A andModel B for the input conditions indicated onthe table. Dose projections vary from factorsof 1-3 up to factors of several orders ofmagnitude. Sources of discrepancy can bedivided into three categories:

1) Dispersion

Figures 1 and 2 illustrate that X/Qvariations between models account for up to afactor of 3, with IRDAM being consistentlyconservative.

2) Source Terms

Models "A" and "B" use a decay-correctedsource term of 15 specific isotopes (13 noblegas and 2 iodine). Dose rates are calculatedfor each isotope, then summed to obtain the

total dose rate. For the IRDAM runs, nobls gasto iodine fractions were provided that wereidentical to those being used by Models A and B(Table 2); however, since apejiflc isotopes werenot indicated, IRDAM results depend on theaverage noble gas and iodine dose factors beingassumed in the IRDAM program.

TABLE 1. MODEL SUMMARI MATRIX

Characteristic

Straight Line Gaussian ('A1 model) X XTime Dependent (Class 'B1 model) XWhole Body Gamma Dose Calcs X X XAdult Thyroid Dose Calcs (1) (1)Child Thyroid Dose Calcs X (1) (1)Elevated Releases (2)Ground Level Releases X X XR. G. 1.109 Dose Factors X X XTime Between Reactor Shutdown

and Release Considered (3) (3) (3)Gros3 ItNG Ratio Incorporated XI:NG Ratio Determined by Isotope X XDownwind Decay of Plume X X XDeposition Considered XLake Effect Capability XFinite Gamma Dose XSemi-Infinite Gamma Dose X X

I = IRDAM, A = Utility "A" model, B = Utility"B" model.

KO_[ESt

('X' indicates model has that charaoteristo)

(1) User chooses adult or child.(2) IRDAM calculates ground level releases as

elevated, with a 10-meter release height.(3) All three methods consider time after

shutdown, but in different ways.

TABLE 2. PERCENT IODINE IN EFFLUENT

AccidentType

% I of Total I + NG0 2 8(hours after shutdown)

Filtered(Loss of Coolant)

Unfiltered(Steam Line Break)

2.1 H.5 5.8

80.2 96.6 98.0

189

LOSS OF COOLANT ACCIDENT

StabilityIRDAM:

Hours After

WBCT

WBCT

WBCT

0

1.20.78

0.460.58

0.360.79

AShutdown

2

1.21.1

0.950.85

0.361.1

IRDAM:Hours After

0

2.00.78

2.00.95

2.41.6

BShutdown

2

1.31.1

1.21.4

2.03.5

STEAM LIME BBEAI

StabilityHours

WBCT

WBCT

WBCT

IRDAMAfter

0

0.160.97

0.0630.70

0.0480.94

: AShutdown

2

0.0281.6

0.0311.2

0.00911.6

IRDAM:Hours After

0

0.0410.95

0.0291.2

0.0281.9

BShutdown

2

0.00661.6

0.00612.0

0.00505.2

TABLE 3. Summary of Site Boundary (0.7 miles) Dose Projection Comparisons.Values are the ratios of IHDAM results to those of Models A and B, asindicated. (WB = Whole body, CT = Child Thyroid).

190

5.OT

< b.(T '

3.CT

2.(r •

i.o-

STABILITY

• B

t

•rtDownwind Distance (nlles)

Figure 1. Ratio of X/Q values between IRDAM andModel A for various stability classes. Ratiosshowed less than 31 variation between 4 sph and10 mph.

5.Q-

I

2.0

SXUItlTT

• B

• *• F

•

10Downwind Distance (ailes)

Figure 2. Ratio of X/Q values between IRDAM amiModel B for various stability classes. Ratiosshowed less than 3Z variation between 4 »ph and 10•ph.

191

3) Other Factors

a. The major discrepancy betweenmodels is seen for the case of whole body doseresulting from a steam line break. Thisdiscrepancy results from the fact that Models Aand 3 include the iodine isotopes in theircalculation of whole body dose, while IRDAM doesnot. In the case of a filtered release, thisomission is not significant; however, in arelease that consists of greater than 80%iodine, the whole body dose resulting fromiodine is substantial. It is important torealize, however, that in a plume that is 80Siodine, Protective Action Recommendations (PARs)will be driven by thyroid do3e, not by wholebody dose.

b. Finite vs 3emi-infinite dosemodels, deposition, and depletion of the plumeall contribute to differences between results ofIRDAM and the other models. For th-5 casesexamined in this paper, these factors contributeto Ies3 than a factor of two difference inresults (as evidenced by comparing lose ratiosat the site boundary to those at ten miles).

T. IMPACT OH PROTECTIVE ACTIOB HECOttlEIIDATKMIS

The basic guidance for protective actiondecision-making is found in EPA-520/1-75-901"Manual of Protective Action Guidea andProtective Actions for Nuclear Incidents."Protective actions are based on total projecteddose (i.e., do3e rate multiplied by the expectedduration of the release), and when this doseexceeds certain triggerpolnt;j, actions toprotect the public may be required. TheseProtective Action Guides (PAG's) and theircorresponding actions are as follows:

Total Do3e (Rem)

Whole Body Child Thyroiddose u 1 dose t- 5

1 h. dose L. 5 51 dose .255 h. dose 25 4 dose

Action Considered

noneshelterevacuate

(Kotes non-metric units are used herein as boththe dose models and the PAG'3 themselves areexpressed in these units. For reference, 1 Rem= 0.01 aievert, 1 Cl = 3.7E10 becquerel, and 1mile = 1605 meters.)

Figures 3 and H illustrate the effects thatdifferent model results have on protectiveaction decision-making. In Figure 3, resultswere compared for the case where there la goodagreement (less than 20} difference) betweenmodels: IRDAM va. Model A for the case of aLOCA, 0 hours after shutdown, B stability, 4mile per hour wind, 1000 Ci/sec release rate, 2hour duration. In the case of the whole bodydose, both models exceed the PAG limit forshelter it the site boundary, but require noaction at other downwind distances. One wouldexpect a similar protective action

recommendation (PAR) to be made in both cases;e.g., shelter to 2 mlle3, no action beyondthat. In the case of the thyroid doseprojection, however, Model A exceeds the PAGlimit for shelter at 5 miles while IRDiM doesnot. This could result in a PAH based on IRDAMto evacjate to 2 miles, shelter to 5 miles, noaction beyond that; while Model A indicatesevacuation to 2 miles, shelter from 2 miles to10 miles, and no action beyond 10 miles.

Now examine the impact on PARs when modelsproduce signficantly different projecteddoses. Figure <J illustrates the integratedwhole body and thyroid doses for the ca3e ofIRDAM vs. Models A and B, steam line break, 2hour3 after shutdown, F stability, 4 mph wind,660 Ci/sec release rate for whole bodycalculation, 1.6 Cl/3ec for thyroid calculation,2 hour release duration. In the case of thewhole body dose, IRDAM predicts that no actioni3 required - while Models A and B indicate thatthe evacuation PAG is exceeded beyond 5 milesand the shelter PAG is exceeded beyond 10miles. In the case of the thyroid dose, IRDAMpredicts that the evacuation PAG 13 exceededbeyond the site boundary (0.7 miles), while theshelter PAG is exceeded beyond 2 miles. Incontrast, Model B results show that theevacuation PAG is never exceeded, and that theshelter PAG is exceeded between the siteboundary distance and 2 miles, while no actionis required beyond that distance. Model Aresults are similar to IHDAM's

T.

This paper illustrates the dramatic effectthat the use of different dose projection modelscan have on protective action decision-making.Even though models may appear similar on thesurface, subleties such aa time-dependent dosefactors contribute to widely-varying doseprojections - and .jonsequently to widely-varyingprotective action recommendations. To avoidconfusion during emergencies, dose assessmentpersonnel at both the utility and governmentalagencies should have a thorough understanding ofthe capabilities and limitations of theirparticular models.

192

0.7 2 5Downwind Distsnce (miles)

figure 3. Compariuon of Projected Do*t* Relative to PAG Llait*.Case 1: LOCA, 0 hour* after shutdown, B (.'ability. Ca*e 1 resultsrepresent agreement of projected doses to within 20Z, yet there isa difference in which PAG triggerpoints are net at each distance.

Downwind Distance (miles)

Figure 4. Comparison of Projected Doses Relative to PAG Limits.Case 2: steam line break, 2 hours after shutdown, F stability.The asrked differences in projected doses between these model*result in widely varying protective action recommendations.

0

A N S Topica l Mee t ing o a Radiological AccMeate—Perspectives nai EawrgeKy P l i

An Importance Ranking of Various Aspects of Off-SiteRadiological Emergency Preparedness

John W. Hockert and Thomas F. Carter

ABSTRACT

Under contract to the Edison Electric Insti-tute, IBAL developed a method to assess therelative importance of various aspects ofoffsite radiological emergency preparedness.The basic approach involved structuring the 35objectives that the Federal Emergency Manage-ment Agency expects offsite emergency plannersto demonstrate during nuclear power plant emer-gency preparedness exercises into a hierarchybased upon the emergency response capabilitiesthey support. The analytical hierarchy process(ASP) was employed to derive the quantitativerelative importance of each of the 35 objec-tives based upon its contribution to the over-all capability of offsite agencies to assist inprotecting public health and safety in theevent of an emergency at a nuclear power plant.The judgments of a cross-section of state andlocal emergency planners, federal regulators,and intervenors were solicited to rank the 35objectives.

I. INTRODUCTION

At present, thn primary mechanism employedby federal evaluators to determine the adequacyof offsite emergency preparedness is observa-tion of the full-scale exercise conducted atleast once every two years at each nuclearpower reactor. Although the Federal EuiergencyManagement Agency (FEMA) has continued to re-fine the methodology used during these evalua-tions since taking over this responsibility in1979, it is generally recognized that addition-al improvements can be made in the evaluationmethodology. In particular, a useful adjunctto the current evaluation methodology would bea systematic method to recognize and incorpo-rate into the exercise evaluation process theconsiderable differences in the real importance

among the NUREG-0654 planning standards andevaluation criteria. Without such a method,there is the possibility that utility andstate and local government attention andresources will be directed toward relativelyinsignificant exercise deficiencies at theexpense of real emergency preparednessproblems. In addition, this method would alsoserve to foster a consistent understandingamong the FEMA regions and evaluators aswell as state and local governmentagencies and utility decision makers as toexactly what constitutes adequate emergencypreparedness. Furthermore, such a method couldalso be used to best allocate governmental andutility radiological emergency preparednessresources and minimize compliance costs torate-payers.

This paper describes a means to achievethe first part of this objective: that in,to systematically assess the relativeimportance of the various aspects of offsiteradiological emergency preparedness. Themethodology presented builds upon the modularexercise evaluation concept developed byFEMA to create a hierarchy describing re-quired offsite emergency preparedness capabili-ties. Based upon fiis hierarchy, provenanalytical methods for tizantification of expertjudgment were employed to develop a structuredinterview format to be used with experts in thefield of radiological emergency preparedness.Based upon their qualitative judgements,expressed as responses to a carefully selectedset of questions, the methodolocnj provided ameans to calculate a. quantitati-- measureof the overall importance of each nf the 35standard objectives that FEMA expects offsiteemergency planners to periodicallydemonstrate during exercises. The methodologyalso provided quantitative measures ofthe consistency of the qualitativejudgments leading to the relative impor-tance measures.

193

194

II. METHODOLOGY

The evaluation of the level of offsiteagency preparedness to respond to nuclear powerplant emergencies is a difficult and complica-ted undertaking. Proper emergency response maywell depend upon the actions of from dozens to,conceivably, hundreds of individuals. Theactions of these individuals throughout theemergency are interrelated in a complex mannerthat depends upon both the specific nature ofthe emergency and the consequences of previousemeigency response actions. In evaluatingsuch a complex system, it is necessary tofoster and to maintain an awareness of theimportance of the proper functioning of eachcomponent of emergency preparedness to theoverall capability of offsite agencies toprotect the public health and safety. Other-wise, there is a distinct possibility thatevaluators will concentrate their attentionon activities that they are expert in or thatthey find interesting, independent of theimportance of such activities.

In developing their modular approach forevaluating radiological emergency preparednessexercises (FEMA, 1983), FEMA has made greatstrides in structuring the evaluation process.This approach identifies a set of 35standardized objectives to be demonstratedduring radiological emergency preparednessexercises. These objectives have been drawnfrom those elements of NUREG-0654/FEMA-SEP-l,Rev. 1, Criteria for Preparation and Evaluationof Radiological Emergency Response Plans andPreparedness in Support of Nuclear PowerPlants, that are both observable duringexercises and pertinent to state or localgovernments. The next logical step in thisstructuring process is the identification ofthe relative importance of each of these 35objectives.

The identification of the relative import-ance of each of the 35 objectives is also madedifficult by the complexity of their interrela-tionship with one another. This process ismade even more difficult because the requisiteoffsite emergency response capabilities differconsiderably for different types of nuclearplant emergencies. Therefore, the capabilitiesrequired to achieve certain objectives may bequite important in responding to extremely un-likely accidents and yet be unimportant for thegreat majority of nuclear plantemergencies. In making a determination ofoverall relative importance, suchcapabilities must be compared with othercapabilities of moderate importance forvirtually all nuclear plant emergencies.Thus, the method employed to determine therelative importance of the 35 objectives mustfacilitate such complex comparisons.Furthermore, in order to be useful, the methodmust be reasonably simple to use andunderstandable. These considerations stronglysuggest the use of a methodology that makesmaximum use of the considered judgment ofindividuals experienced and knowledgeable inradiological emergency preparedness.

An executive body of literature exists de-scribing proven methods to consistently accountfor and effectively use expert judgment inimportance ranking. (Freeling, 1982, Saaty,1977, 1978 a,b,) Thomas L. Saaty, formerly ofthe University of Pennsylvania's Wharton Schooland now with Pennsylvania State University, haspublished several journal articles that addressa structured method for developing quantitativemeasures of the relative importance of variousobjectives supporting the same goal. The firststage of this mezhod is to identify the goalsthat each objective supports. This leads tothe construction of a hierarchy.

The use of such a hierarchy has severaladvantages. First, the hierarchy provides ameaningful integration of related objectives.This not only focuses the evaluator's attentionon those aspects of the related objectives thatcontribute most directly to their goal, but italso relieves the evaluator of the difficultproblem of directly comparing objectives thatsupport disparate goals. This increases boththe effectiveness and the efficiency of the pro-cess. Second, the use of a hierarchy struc-tures the evaluation so that issues related tosystem pu. ••• formance occur at lower levels, whilepolicy issues dominate at the higher levels.Thus, information about, the relative importanceof higher level goals can be effectively ob-tained from policy makers, while informationabout the relative importance of individualobjectives in achieving those goals can beobtained from individuals more familiar withsystem details. Third, the use of hierarchiesincreases the reliability and flexibility ofthe process. The achievement of the overallgoal is divided among the levels of the hier-archy in such a manner that each achieves aportion of the goal and the totality meets theoverall goal. Therefore, changes to the mannerin which one portion of the goal is met (i.e.,changes to one part of the hierarchy) do notafiect the analysis for those parts of the hier-archy supporting other portions of the overallgoal.

The hierarchy constructed to support theevaluation of the relative importance of FEMA'sset of 35 standardized objectives to be demon-strated during radiological emergency prepared-ness exercises was derived by identifying thethree capabilities necessary to protect publichealth and safety:

1. The capability to obtain informa-tion necessary to determineactions to be taken to protectpublic health and safety;

2. The capability to maintain thecommand and control necessary tosupport effective decision-makingand incident management; and

3. The capability to implement appro-priate protective actions whenthe decision is made to do so.

These three capabilities were then analyzedinto subordinate capabilities until, even-tually, all 35 of the objectives were linked

195

with the capabilities that they support. Thedetailed hierarchy and a listing of all 35objectives are included in the project report(IEAL, 1985).

The next stage of the method involved theestimation of the kro^ieageaole individual'sunderlying estimate of the importance of eachgoal or objective through use of ratio judg-ments, which are a form of pairwise compari-sons. This was accomplished by having theindividual answer a series of questions regard-ing the relative importance of each pair ofobjectives. The questions are derived directlyfrom the hierarchy, and the possible answersare taken from a set of preselected responsesthat have evolved from Saaty's research. Theset of questions used to rank the importance ofthe 35 objectives and the possible answers tothese questions are presented in the projectreport (IEAL, 1985). The method for quanti-fying these answers and using them to derive anindividual quantitative importance weightingfor each of the 35 objectives is also discussedin the referenced report (I1IAL, 1985).

Expert judgments of the relative importanceof objectives supporting each higher level goaland the goals themselves were solicited by send-ing detailed questionnaires to FEHA and NuclearRegulatory Coimission (HRC) headquarters andregional staff. Regional Assistance Committee(RAC) members, other federal, state, and localemergency planners, and members of publicinterest groups. Thirty-five responses warereceived from individuals representing a reason-able cross-section of these organizations.Rankings were derived based upon each indivi-dual 's responses, and an overall composite rank-ing was developed, based upon each individual'sresponses within his or her areas of expertise.In addition, the responses were analyzed for in-ternal consistency and to determine the degreeto which the responses were correlated with therespondent's organizational affiliation.

Ill. RESULTS

The individual respondent's detailed rank-ings of the 35 exercise objectives were foundto vary. This is not surprising in light ofthe complexity of emergency preparedness andthe extent to which the relative importance ofspecific capabilities during a given emergencydepends upon the detailed accident scenario andsite conditions. The observed variation didnot appear to be significantly related to therespondent's organizational affiliation, geo-graphical location, or area of expertise. Thisvariation indicates that the specific relativeimportatjce weights provided by the AHP shouldbe regarded as soixwhat uncertain.

However, in reviewing the overall response,a pattern did emerge. While the detailed rank-ings varied, there was general consensus on agroup of seven exercise objectives thHt wareconsidered very important by a large majorityof respondents and a group of ten exercise ob-jectives that were considered relatively

unimportant by a large majority of respondents.With one or two exceptions, the uncertaintiesin the individual rankings were r.uch that objec-tives could be ambiguously assigned to one ofthese two groups or to a third group of objec-tives of moderate importance.

The seven objectives that were consideredvery important by the large majority of respon-dents were, in approximate order of decreasingimportance:

Ability to project dosage to the pub-lic via plume exposure (Objective10);Ability to communicate with allappropriate locations, organiza-tions, and field personnel (Objec-tive 5);Ability to make decisions and coordi-nate activities (Objective 3);Ability to mobilize staff and acti-vate facilities promptly (Objective1);Ability to project dosage to the pub-lic via ingest ion exposure (Objec-tive 11);Ability to monitor and control emer-gency worker exposure (Objective20); andAbility to evacuate onsite personnel(Objective 23)

The ten objectives considered relatively unim-portant by the large majority of respondentswere, in approximate order of decreasing impor-tance:

Ability to administer potassium io-dide to the general public (Objec-tive 22* )Ability to brief the media in aclear, accurate, and timely manner(Objective 24);Ability to establish and operaterumor control (Objective 26);Adequacy of facilities for mass careof evacuees (Objective 28);Ability to deal with impediments toevacuation, such as inclement wea-ther (Objective 16);

Ability to coordinate informationreleases (Objective 25);Adequacy of procedures for registra-tion and radiological monitoring ofevacuees (Objective 27);Ability to control access to evacu-ated areas (Objective 17);Ability to estimate total populationexposure (Objective 34); andAbility to determine and implement

FEMA's original Objective 22, the abilityto administer potassium iodide, wasseparated, for this analysis, into theability to administer potassium iodide toemergency workers and the ability toadminister it to the general public.

196

recovery and reentry measures (Objec-tive 35)

The remaining objectives were considered tobe of moderate importance by the majority of re-spondents. The variation in the relative impor-tance accocded the objectives in this group bythe various respondents was greater than thevariation seen in either of the other twogroups.

IV. CONCLUSION

The results of this analysis have potentiallyfar reaching implications for the conduct andevaluation of offsite radiological emergencyresponse exercises. For example, it is reason-able to expect that those objectives consideredvery important would be exercised most frequent-ly, evaluated most thoroughly, and, if not de-monstrated satisfactorily, would be the basisof a Category A deficiency or negative finding.Likewise, the relatively unimportant objectiveswould be exercised infrequently, accordedlittle evaluation effort, and, if not demon-strated satisfactorily, would rarely result inanything more severe than a citation as a cor-rectable weakness. The moderately importantobjectives would, on average, receive moderateemphasis or be emphasized within a reduced emer-gency planning zone, or in specific cases, beclassed as relatively important or unimportantdepending upon the detailed local conditions orexercise scenario considerations.

REFERENCES

FEHA. 1983. "Procedural policy on radiologicalemergency preparedness plan reviews, exer-cise observations and evaluations, andinterim findings." Memorandum from D.McLaughlin to Regional Directors.

IEAL. 1985. "An Importance Ranking of VariousAspects of Offsite Radiological EmergencyPreparedness." IBAL-R/85-98.

Freeling, A. 1982. Unpublished paper.

Saaty, T.L. 1977. "A scaling method for prior-ities in hierarchical structures." Journalof Mathematical Psychology. Vol.25:234-281. Academic Press.

Saaty, T.L. 1978a. "Exploring the interfacebetween hierarchies, multiple objectives,and fuzzy sets." Fuzzy Set.i and Systems.Vol. 1:57-68. North Holland PublishingCompany.

Saaty, T.L. 1978b. "Modeling unstructured deci-sion problems -- the theory of analyticalhierarchies." Mathematics and Computers inSimulation. Vol. XX-.147-158. NorthHolland Publishing Company.

ANS Topical Meeting oa Radiological AcctteitsPerspectives ami Emergency f

A Goal-Oriented Functional Tree Structure forNuclear Power Plant Emergency Preparedness

Richard V. Calabrese and Marvin L. Roush

ABSTRACT Guidelines for development and imple-nentation of emergency response plans do notprovide the planner/implementer with an adequateoverview of the functions to be achieved or ameasure of their relative importance. To pro-vide a framework in which this importance can berecognized, understood and quantified, a logicalgoal oriented tree structure has been developedwhich integrates and gives a clear visual repre-sentation of the functions required to meet theemergency preparedness objective. The treeconsiders a spectrum of both high and low prob-ability events which may require mitigation bothonsite and offsite.

The ultimate objective: to Minimize the111 Effects of a Nuclear Power Plant Incident issatisfied by functions concerned with preventionand mitigation of human injury and propertydamage. A complete and detailed structure whichspecifies the subfunctions and success pathswhich satisfy these functions has been devel-oped. Institutional activities such as plan-ning, training, procedure development, nor.itor-ing and decision making do not enter the treedirectly. Instead, the logic structure definesthe extent to which these activities must beconsidered and the information systems and deci-sion models required for successful implementa-tion of the plan.

The top structure of the tree is presentedand a few branches are considered in detail.The Impact of Institutional activities, informa-tion systems, etc. is discussed. Tree quantifi-cation is considered.

I. INTRODUCTIONNumerous documents provide guidance toward

development and implementation of nuclear powerplant emergency response plans. Most are con-cerned with the adequacy of personnel, suppor-ting facilities and institutional activitiesperformed in support of the plan. These activi-ties include contingency planning, task organi-zation, procedure development and specification,monitoring and dose projection, and practicedrills to aid in plan maintenance and improve-

ment. Unfortunately, such documents do notprovide the planner/implementer with a goodbasis for understanding the functions to beachieved and a clear definition of their overallimportance toward attaining the ultimate objec-tive. To provide a framework within which this"importance" can be recognized, understood andquantified, we have developed a logical treestructure which integrates and gives a clearvisual representation of the functions requiredto meet the emergency preparedness objective.Rather than focus solely upon low probability,high consequence events with radiological impacton the public, we consider the spectrum ofevents which may require mitigation both onsiteand offsite.

The goal oriented tree presented herein isan outgrowth of the logic structures defined aspart of the Top-Down Integrated Approach toSafe, Reliable, Economic Nuclear Power (e.g.,Combustion Engineering, 1982; Technology forEnergy Corp., 1982). The approach and resultsrepresent refinements of a detailed functionalclassification performed for the Dept. of Energy(DOE) by these authors (University ResearchFoundation, 1983) as part of the DOE Informa-tion Systems Project. The structure providesan easily understood overview of all requiredfunctions and their interrelationships, and alogical framework for quantitative evaluation.It id anticipated that the model can be a usefultraining tool to allow an individual to viewspecific parts of a plan in proper context. Itmight also be useful to provide a contextualoverview of regulations to allow a proper per-spective of the need for and the contribution ofa single regulation to the overall objective, aswell as the impact of the regulation uponachievement of other functions.

We have attempted to structure this goaltree according to the following simple guide-lines: For each box, one can look "down" in thetree to see how the functional objective isaccomplished. One can look "up" in the tree tosee why the functional objective is required.Near the top of the tree, each function isspecified in a very general way. Subfunctionsor components of a higher level goal are thendeveloped at lower levels. The lowest levels

197

198

represent one or more success paths by whichspecific objectives can be accomplished. Efforthas been made to address issues in a generic waythat is applicable to any plant and to accountfor all possible contingencies, except sabotage.Therefore, at lower levels not all subfunc-tions/success paths will be relevant at aspecific plant.

II. OVERVIEW OF TREE STRUCTUREThe top structure, shown in Figure 1,

states the ultimate goal of emergency prepared-ness; that is, to Minimize the 111 Effects of aNuclear Power Plant Incident that could poten-tially threaten the health, safety and economicwell being of both the public and plant person-nel. The objectives to be accomplished areprevention and mitigation of human injury andproperty damage. The functions which addressthese objectives can be classified as eitherpreventative (pre-injury) or mitigative (post-injury). Mitigative functions are includedsince any realistic criteria for success mustaccommodate sone casualties and/or propertydamage, if only due to the initiating event.

HINIHIZE ILLEFFECTS OF NUCLEARPOWER PLANT INCIDENT

0_L

PREVENTION ANDHITIOATION OFHUMAN INJURY

A

JLPREVENTION ANDMITIGATION OFPROPERTY DAHAOE

Figure 1 - Top Tree Structure

Portions of the tree that are presentedherein are each assigned a number. This numberappears inside a triangle located to the left ofthe top box of that branch. If the lowest boxesof that branch contain a triangle beneath thea,then the number inside that triangle specifiesthe tree number on which it is continued.

The functions that satisfy the ultimategoal are given in Trees 1 and 2. These are:

a) Prevention of Human Injury. Plant personneland the public must be protected fromconventional Injury and from exposure to radia-tion and toxic chemicals.

b) Prevention of Psychological Stress. Thepsychological impact on the public from an inci-dent of any degree is to be minimized.

c) Human Casualty Control (Mitigation). Peopleinjured as a result of conventional accidentsand/or exposure to radiation or toxic chemicalsmust be identified, treated and otherwise caredfor both onsite and offsite.

d) Prevention of Property Damage. Damage toplant and public property is to be preventedduring an incident.

e) Property Damage Control (Mitigation). Plantand public property suffering conventional dam-age and/or contamination by radioactivity ortoxic (or corrosive) chemicals must be identi-fied, secured, and possibly restored to its pre-incident condition.

Except for Prevention of PsychologicalStress, each function is divided into its compo-nent parts. These are conventional, radiologi-cal and chemical. The conventional and radiolog-ical subfunctions are further divided intotheir onsite and offsite components. Onlyonsite measures are considered for the case of achemical spill. The utility is not responsiblefor the impact of offsite spills on the public.Furthermore, it is unlikely that an onsite spillwill impact the public. However, said impactcould be included if warranted. The need toconsider chemical spills becomes apparent whenone considers that release of toxic vapors fromthe failure of an onsite ammonium hydroxidestorage tank or chlorine cylinder can have ahigher probability of impacting control rooapersonnel than does a radiological release tothe environment. Licensing issues for the pro-posed Midland Nuclear Power Plant included de-tailed consideration of the impact of toxic andcorrosive chemicals released from nearby chemi-cal plants.

Several conditionals appear in Trees 1 and2. They serve to identify the conditions forwhich each subfunction should be considered.Onsite conventional injuries and property damagemust be considered for all incidents. The sameis true for psychological stress. Offsite func-tions must only be considered when the poten-tial exists for a radiological release to theenvironment. The tree is structured to accommo-date radiological incidents confined to plantstructures and to consider the impact of offsitechemical spills on plant personnel and property.

The functions defined above are accom-plished by a complex series of human actions/in-teractions. Numerous activities must be per-formed by designated plant personnel and civil-ians, often under less than ideal conditions.Many require cooperation among plant officialsand civilian authorities. Their success isdependent upon proper planning, accurate dissem-ination of information and an ability to makerational decisions in a timely manner. However,such considerations do not directly enter intothe functional classification. Planning, whichIncludes the Identification, organization anddesignation of tasks and the development ofImplementation procedures to Insure that theyare accomplished in a timely fashion, is aninstitutional activity. Each functional activi-ty is best accomplished when a well designed andtested plan exists. Information can only bedisseminated in a timely fashion when a welldesigned communications network is in place.Again, development of such a network and proce-dures for its use are institutional activities.The functional classification serves to identify

199

PREVENTION M OMITIGATION OFHUMAN INJURY

orrtiTERADIOLOGICALCASUALTYCONTROL

£vPIICVENTION AM)H1T1EAT1ON OF

DMIAOC

CONVENTIONALBAMAKCONTROL

ONCITECtCIIICAL

OAJ1AOE CONTROL

RAOIOLOSICALDAnwtCONTROL

ONSITECONVENTIONAL

OAHAOCPREVENTION

OMITECONVENTIONAL

D«1A«ECONTROL

OFFSITECONVENTIONAL

DAHAKCONTROL

ONtHERAOIOLOSICAL

DAiugcCONTROL

OFFtlTERADIOLOSICAL

DAHAKCONTROL

the requisite institutional activities. Eachinstitutional activity could be classifiedaccording to a tree structure which parallelsthe functional tree, once Its goals have beenproperly defined.

The acquisition, analysis and interpreta-tion of data to insure that accurate information

Is disseminated and that proper and timely deci-sions are made are activities which aid in theimplementation of the plan. Like institutionalactivities, decision models and informationsystems can also be classified according ,.o aparallel tree structure. Furthermore the func-tional classification serves to identify infor-

200

nation needs and the decisions which must bemade. Separate consideration of institutionalactivities and Infornation Syst im/DecisionModels also serves to facilitate quantificationof the goal oriented tree.

While a detailed functional classificationhas been carried out, it is too extensive to bepresented here. Instead, the major elements ofthe tree will be presented followed by a discus-sion of selected individual success paths. Aspreviously noted, the tree super structure pre-sented here represents both a refinement of anda supplement to the tree developed for the DOEInformation Systems Project (University ResearchFoundation, 1983). The work was undertaken toprovide a more logical and easily understoodrepresentation by emphasizing common featuresamong the various subfunctlons and successpaths. The lower levels of the tree remain thesame and are given in complete detail in theaforementioned report. In this sense, thestructures presented here serve as an introduc-tion to ' and a useful aid in interpreting themore detailed functional classification.

In presenting the logic structures, theimpact of institutional activities, data acqui-sition and decision making will be considered.Emphasis will be given to the function OffsiteRadiological Dose Prevention which appears underPrevention of Human Injury. This is satisfiedby subfunctions for source mitigation, evacua-tion, sheltering, etc.

III. CONVENTIONAL/CHEMICAL INJURY PREVENTION

A. Conventional Injury Prevention.This function is concerned with protec-

ting emergency response workers and the publicfrom conventional injury during the course oftheir actions.

tion.quired

1. Onsite Conventional Injury Preven-Tree 3 specifies the subfunctlons re-to meet this goal. Non-essential per-

ONSITECONVENTIONAL INJURY

PREVENTION

6

sonnel must be protected or escorted to the siteboundary. The required actions are similar tothose taken to protect the public. Two types ofessential personnel are identified, simply be-cause the tasks they must perform are quitedifferent. At the next tree level (not shown),specific tasks are Identified. These includeresponding to equipment failures, fires, seismicevents, etc., or providing transportation/equip-ment/supplies to other response teams. For mostof these, conventional injuries are prevented bya common type of success path illustrated InTree 4. Individuals must be capable of perfor-ming the task; that is, they must understand allaspects of the task. In order to identify ade-quate personnel, the magnitude of the task andpersonnel limitations muBt be understood. Teammembers must be aware of potential dangers andthe requisite protective equipment (task specif-ic) must be available. This tree is applied inturn to each task with lower tree levels speci-fying the details of these requirements.

2. Offsite Conventional Injury Preven-tion. This goal includes protecting the publicand emergency response teams assisting them(Tree 5). Protection of essential personnel issatisfied by subfunctions similar to those forthe protection of those assisting onsite non-essential personnel. The public is protected bysatisfying thxee general subfunctions. Accessto dangerous areas must be restricted. Thisaubfunction is activated only if property damageis caused by initiating events such as earth-quakes, tornadoes, etc., or by public panic.Damaged property must be identified and securedand the potential danger must be publicized.Rioting and looting must be controlled by avoid-ing panic and providing proper deterrents. Thisis Insured by adequate response capability,competent teams and adequate security. Trans-portation hazards must be minimized on all ap-propriate types of carriers. A common successpath is shown in Tree 6. Traffic congestion andpanic tendencies must be minimized. Vehiclesmust be loaded and operated safely. The needsof special population groups such as the ill andelderly must be accommodated.

Lower levels of the Offsite ConventionalInjury Protection tree identify specific items/population groups that must be considered and

MOTECTIONOF ESSENTIAL

PERSONNEL

MOTECTION OFNON-ESSENUM.

PEMOMCL

PROTECTION OFPERSONNELINVOLVEP IN

MTIMTIVE ACTIONS

PROTECTION OFPERSONNEL ASSKTIN6

NON-ESSENTIALPERSONNEL

PROTECTIONOF ESSENTIALPERSONNEL

PERSONNEL CAPABLEOF PERPSRMNS

REQUISITE TASKS

ADEOUATEPERSONNELAVAILABLE

POTENTIALDAIISER*

PROTECTIVEEQUIPMENTAVAILABLE

201

detail the specific tasks and facilities (in-

cluding communications) that are required. This

la common to the success paths for all major

functions.

B. Onsite Chemical Dose prevention.

Tree 7 reveals tnat the top tree struc-

ture for this function is similar to that for

Onsite Conventional Injury Prevention. Since

workers must be protecced from overexposure the

general subfunctions are quite similar to those

for Onsite Kadiological Dose Prevention to be

discussed subsequently. Of course, specific

tasks and equipment requirements will differ.

RIOTING ANDLOOTING ARECONTROLLED

ADEQUATERESPONSE CAPABILITY

AVAILABLE

COMPETENT TEAMSCOORDINATE

PUBLIC ACTION

SECURITYTEAMS PATROL

AREA

[ToROAD TRANSPORTATIONHAZARDSMINIMIZEDI

MARINE TRANSPORTATIONHAZARDSMINIMIZED

AIR TRANSPORTATIONHAZARDSMINIMIZED

RAIL TRANSPORTATIONHAZARDSMINIMIZED

rTRAFFICCONOESTIONMINIMIZED

TRANSPORTATIONHAZARDSMINIMIZED

1PANIC

TENDENCIES, MINIMIZED

0

r~

1VEHICLES

ARE SAFELYLOADCD/UNLOAUepo

PUBLIC AT LAPSE 1IS SAFELY

LOADED/UNLOADED

1

VEHICLESOPERATEDSAFELY

SPECIAL POPULATIONGROUPS ARC SAFELYLOWED/UNLOADED

ONSITECHEMICAL DOSEPREVENTION

Q.PROTECTIONOF ESSENTIAL

PLANTPERSONNEL

IL.-"PROTECTION OTNON-ESSENTIALPLANT PERSONNELAND VISITORS

s.PROTECTION OFPERSONNEL IN

CONTROL ROOM AMD OTHERSAFETY RELATED AREAS

IL0

PROTECTION OFPERSONNEL

RESPONDING TOCHEMICAL INCIDENT

J-INITIAL

PROTECTIVEMEASURESTAKEN

_LSUBSEOUENTPROTECTIVEMEASURESTAKEN

202

IV. RADIOLOGICAL DOSE PREVENTION

The onsite component of this function isdefined to include monitoring teams and thoseassisting the public.

A. Onsite Radiological Dose Prevention.Like the other onsite "prevention"

functions, the radiological function involvesprotection of both essential and non-essentialpersonnel (Tree 8). Essential personnel are di-vided into those Involved in emergency responseoutside plant buildings and those who remaininside the plant. Those outside plant buildingsinclude monitoring teams and those assistingnon-essential personnel. Their safety is in-sured by adequate knowledge of environmentalhazards, and use of protective clothing/equip-ment and/or shift rotation to avoid overex-posure. Personnel within the plant must beprotected from radiation sources within theplant and from radioisotopes released to theatmosphere. The former corresponds to the pro-tection of workers entering high radiationareas. This subfunction has not been developedsince it is more the concern of health physi-cists than emergency planners. The latter sub-function is satisfied by success paths whichinsure the habitablllty of the control room,emergency response center and other safety re-lated areas.

Initial and subsequent measures are takento insure the safety of non-essential personnel.Initial measures include identifying, notifying,accounting for and sheltering these individuals.Subsequent measures include sheltering orevacuation and possibly the use of blockingagents.

B. Offsite Radiological Dose Prevention.To protect the public, two subfunctions

are required as shown in Tree 9. Both acute andchronic exposure pathways must be controlled.Acute pathways are those to which the public is

exposed during and immediately after a releaseof radioactivity to the environment. Theseinclude inhalation, cloud shine, and short-termground shine. Chronic exposure pathways arelonger term pathways of concern during the pe-riod of days (or possibly years) after the inci-dent. These include longer term ground shine,inhalation of resuspended material, and inges-tion of contaminated foodstocks.

1. Control Acute Exposure Pathways.Methods currently accepted for controlling pub-lic exposure to radioactivity are evacuation andsheltering. However, the use of blocking agentssuch as potassium iodide (KI) is included forcompleteness. Each of these methods may be usedalone or in conjunction with one or both of theothers. For instance, it may not be practicalor desirable to evacuate the entire population.During a severe incident it is possible thatthe people residing in areas to which the ra-dioactive plume is carried by the winds will beevacuated and those in surrounding areas will besheltered. Blocking agents may be distributedto evacuees and those sheltered in areas adja-cent to those under evacuation. Evacuees may bedirected to engage in mobile sheltering tech-niques such as breathing through wet towelswhile en route to their cars and keeping theircar windows and air intake vents closed to avoidthe influx of contaminated air. Furthermore, Itmay be recommended that only particularly sus-ceptible members of the population, such aspregnant women, be evacuated and all others besheltered or given blocking agents. It may onlybe possible to evacuate the mobile populationrequiring sheltering for immobile persons, suchas bedridden individuals and the like. It maybe desirable to evacuate the general public butnot institutionalized individuals.

The decision as to which of the aforemen-tioned methods should be employed, and where,when, and to whom is a complex one requiringplant, monitoring and weather data. As pre-

ONSITERADIOLOGICAL

DOSEPREVENTION

_LPROTECTION OFESSENTIALPLANTPERSONNEL

_LPROTECTION OFNON-ESSENTIALPLANT PERSONNELAND VISITORS

PROTECTION OFPERSONNEL INVOLVED INEMERGENCY RESPONSE

OUTSIDE PLANT BUILDINGS

PROTECTION OFPLANT PERSONNELWITHIN THE PLANT

INITIAL PROTECTIVEMEASURESARE TAKEN

SUBSEOUENTPROTECTIVE MEASURES

ARE TAKEN

PROTECTION OF PLANTPERSONNEL FROM RADIATION

RELEASED WITHINPLANT BUILDINGS

J_PROTECTION OF PLANT

PERSONNEL WITHIN PLANTBUILDINGS FROM RADIATION

RELEASED TO THE ENVIRONMENT

203

OFFSITE

PUBLIC ISEVACUATED IN ASAFE AND TIMELY

MANNER

PUBLIC ISSHELTERED IN ASAFE AND TIMELY

MANNER

BLOCKING AGENTS CK1>ARE ADMINISTERED INA SAFE AND TIMELY

MANNER

RADIOACTIVEEMISSIONS ARECONTROLLED ORTERMINATED

viously mentioned, the decision process with itsrequisite information is not included as part ofthe functional classification presented here.The output of the decision process isrepresented by conditionals placed on the sub-functions "Evacuation," "Sheltering" and "Block-ing Agents" displayed as Rl, R2 and R3, re-spectively, in Tree 9.

In principle, the area surrounding theplant can be divided into several sectors, thesize and number depending upon population den-sity, road network, meteorology, terrain fea-tures, etc. The decision to evacuate, shelter,use KI or to use a combination of these can bemade for each sector and population group andfor various time periods. The conditionalstherefore represent a matrix of decisions. Thematrix coefficients prioritize and define therequisite subfunctions. For instance, a coeffi-cient of 0 can be used to define the associatedsubfunction as unnecessary for the sect ">r/popu-lation group/time period it represents, a 1 candefine it as a primary success path, a 2 as analternate success path, etc. Depending upon thesituation, two or more of the conditionals mayhave the same coefficient for a common element.For example, sheltering with ingestion of KI maybe the primary success path. Again, the coeffi-cients of the matrices are the output of thedecision process.

We have defined an Information System/Deci-sion Model whose output in the form of condi-tional matrices specifies the particular successpaths that should be utilized. Such modelsenter the goal tree at many locations and a

sample model will be presented later. The needfor particularized decisions becomes apparent inareas of high population density during incle-ment weather, when the road network cannot ac-commodate a full scale evacuation or when thepresence of water bodies and terrain featurescould force some evacuees to travel through theradioactive plume.

The role of evacuation, sheltering, use ofblocking agents and the output of the decisionmodel can be modified to account for an addi-tional success path under certain circumstances.A controlled release (note the conditional dis-played in Tree 9) is defined as one that can bereduced in emission rate or terminated for aperiod of time, or permanently. Consider asituation where the containment building con-tains substantial airborne radioactivity and thepressure is rising but has not yet reached anunsafe level. Plant authorities may wish tovent the containment to prevent its rupture. Ifthe wind direction is shifting away from popula-tion centers, the decision may be made to wait,if possible, until the wind shift occurs. Ifthe wind is steady, the decision may be made toevacuate in the direction of the wind but tohold off completely or vent slowly until theevacuation is well underway or until the publiccan be safely sheltered.

The other extreme of a controlled releaseis the situation that existed in the TMI-2 con-tainment after the plant was under control. Thecontainment pressure level was safe but ventingwas necessary. Radioactivity was released at awell-controlled rate during optimal meteorolog-

204

leal conditions tc maintain acceptably small

levels of radioactivity in the environment.

Evacuation, sheltering, or use of blocking

agents were noc required.

Tree 10 provides the elements common co

successful evacuation and sheltering. Lower

levels of the tree (noc shown) identify and

specify the specific tasks/facilities/equipment

needed to achieve each goal. Tree 1L is a

similar structure for administering blocking

agencs. Tree 12 applies to a controlled re-

lease. By Release Mechanism Insured we mean

that the venting aperature will function proper-

ly. It is evident that lower levels of this

tree are intimately linked to knowledge of plant

status and other emergency response neasures.

2. Control Chronic Exposure Pathways.

This function is accomplished by Interdiction as

noted on Tree 9. Again, an Information System/

PUBLIC ISEVACUATED/SHELTERED

IN A SAFE ANDTIMELY MANNER

1PUBLIC

NOTIFIED IN ATIMELY MANNER

rYPUBLIC RESPONDSIN A SAFE ANDTIMELY MANNER

1ADEQUATE

FACILITIES/EOUIPHENTAVAILABLE

d d0FFICIALS/T6AMSNOTIFIED OFRECOMMENDEDACTIONS

RECOMMENDEDACTIONS

COMMUNICATEDTO PUBLIC

1i .

PUBLIC AWARE OFPURPOSE OF ACTIONAND IMPLEMENTATION

PROCEDURES

1

ADEQUATESUPPORTSERVICESAVAILABLE

ADEQUATESUPPORT

FACILITIESAVAILABLE

01

ADEQUATEFOOD ANDSHELTER

AVAILABLE

BLOCKINB AOENTS CKI)ARE ADMINISTERED INA SAFE MO TIMELY

MANNER

PUBLICNOTIFIED IN ATIMELY MANNER

Kl SUPPLY ISAVAILABLE FORDISTRIBUTION

OFFICIALS/TEAMSNOTIFIED OF

RECOMMENDED ACTIONS

RECOMMENDEDACTIONS COMMUNICATED

TO PUBLIC

ADEQUATELOCAL SUPPLYAVAILABLE

SUPPLY MADEAVAILABLE AT

TIME OF INCIDENT

KI DISTRIBUTEDWITHIN AFFECTED

AREA

dKIDOSE

RECEIVED

Q._L

PUBLIC AHARE O f HPURPOSE OF ACTIONAND IMPLEMENTATION

PROCEDURE

KI PROPERLYIMESTED OR'ADMINISTERED

ADEQUATE DISTRIBUTIONFACILITIES/EOUIPMENT

AVAILABLE

AOCQUATE SUPPORTSERVICESAVAILABLE

CONTAINMENTBOUNDARY CAN BE

INSURED FOR REQUISITETIME PERIOD

RADIOACTIVEEMISSIONS ARECONTROLLED ORTERMINATED

RELEASEDMECHANISMINSURED

RATE AKI TINEPERIOD OF RELEASE ARECOORDINATED WITH OTHERMITISATINS ACTIONS

205

Decision Model impacts the tree. This modeluses weather and monitoring data and analyticalpredictions to locate contaminated areas and todetermine the proper course of action. Lowerlevels of this tree are specified in Trees 13and 14. Again, the success paths (not shown)specify the specific means to accomplish eachgoal.

V. OTHER FUNCTIONS

A. Prevention of Psychological Stress.It is not our intention to imply that

the utility is responsible for the psychologicalwell being of the public. Rather, we note thatthe probability of success is enhanced whenpanic is avoided and the public demonstrates acooperative spirit. It is evident from Tree 15that this is accomplished through education, bycompetent response efforts and by accurate andtimely dissemination of information. In otherwords, this function is satisfied by a success-ful plan.

B. Human Casualty Control.Tree 1 reveals that consideration of

conventional, chemical and radiological issuesfor both onsite and offsite areas involves fiveseparate functions to accomplish this goal. Allhave a similar top tree structure. This isshown in Tree 16. At lower levels of the tree(not shown), detailed success paths for eachhave been developed. Except for degree of in-jury/overexposure, onsite and offsite considera-tions are similar. The success oaths give the

means by which casualties are identified anddiagnosed, first aid is administered and subse-quent treatment is obtained. Transportationconsiderations include use of road, air andpossibly marine transportation, depending uponthe site. Specific treatment facilities areIdentified. Alternate success paths includefacilities committed in advance and at the timeof the incident.

r. Property Damage Prevention.The functions/success paths which in-

sure that property damage is prevented have manyelements in common with those which insure theprevention of human injury. While certain tasksare specific to damage prevention others will beImplemented by the same teams responsible forhuman safety. These latter tasks are includedin the property tret for completeness and toemphasize to implementers the dual nature oftheir responsibilities. The top tree structurefor conventional property damage prevention isgiven In Tree 17. It applies (in turn) bothonsite and offsite. That for onsite chemicalproperty damage prevention is given in Tree 18.Structures for radiological damage preventionare not given. These have not been developed indetail since the requisite subfunctions areintimately linked to termination of the inci-dent. The offsite branch of this tree doescontain provisions for sheltering of farm ani-mals and the like.

D. Property Damage ControlThis function is concerned with the

INTERDICTINSESTIONPATHWAYS

sAIRBORNEROUTES

INTERDICTED

WATERBORNEROUTES

INTERDICTED

1CONTAMINATED

PLANTIN8ESTI0NINTERDICTED

1CONTAMINATED 1

MEATJN6ESTI0NINTERDICTED

1CONTAMINATEDDAIRY PRODUCT

INSEST1ONINTERDICTED

1CONTAMINATED

WATERINGESTIONINTERDICTED

CONTAMINATEDSEAFOODINGEST1ONINTERDICTED

INTERDICTINHALATION ANDSHINE PATHWAYS

RESTRICT PUBLICCONTACT WITH

CONTAMINATED AREAS

CONTROLCONTAMINATION

SPREAD

REAL HAZARDSMINIMIZED

ANDCONTROLLED

PREVENTION OFPSYCHOLOGICAL

STRESS

UNFOUMXD FEARSOF PLANT HAZARDSMINIMIZED BEFORE

INCIDENT

ILPERCEIVED HAZARDSHADE NO GREATERTHAN ACTUALHAZARDS

206

mitigation of property damage caused by theincident, the initiating event and emergencyresponse actions. During the actual emergency,the subfunctions which satisfy conventionalproperty damage control (Tree 19) are concernedwith controlling fires and securing areas toprevent further damage. The subfunction dealingwith other initiators is taken to include damageto structures caused as a result of emergencyresponse actions. A separate subfunction dealswith damaged equipment (e.g., motor vehicles)that must be moved to prevent interference withother emergency actions. Subfunctions whichdeal with the restoration of conventionallydamaged property have not been developed sincethis is usually not the responsibility of emer-gency plan Implementers.

The radiological (onsite and offsite) andonsite chemical property damage control func-tions are concerned with decontamination and/or

disposal of contaminated materials and property.Their common top structure is shown in Tree 20.The need for an Information System/DecisionModel to locate contaminated areas and determinethe proper course of action is noted. Of coursethe success paths for radiological and chemicaldecontamination are quite different.

VI. DISCUSSION

A. Task Implementation/Institutional Activ-ities.The lowest levels of the goal tree

detail the specific tasks that must be implemen-ted and the mechanisms for their accomplishment.Figure 2 illustrates the elements common toimplementation of many tasks. It is at thislevel that institutional activities have theirgreatest impact on the tree. Table 1 gives theinstitutional activities that insure successful

CONVENTIONAL/RADlOLOeiCAL/CHEMICALHUMN CASUALTY CONTROL

INJURY/OVEREXPOSUREDIAMOND IN ATIMELY MANNER

CASUALTYIDENTIFIED j

CONVENTIONALPROPERTY DAMAGE

PREVENTION

EVACUATIONRELATED DAMA8EIS PREVENTED

LOOTING ANDRI0TIN6 AREPREVENTED

FIRST AIDADMINISTEREDCAS NEEDED)

6EXTENT OF

INJURIES/DOSEDETERMINED

FACILITYAVAILABLE

FIRST AIDPROPERLYAPPLIED

VICTIMMOVED TOSAFE AREA

DAMAGE TOESSENTIAL EOUIPHENT

IS PREVENTED

VICTIM TRANSPORTED TOMEDICAL/TREATMENT

FACILITY

TRANSPORTATIONAVAILABLE

VICTIMDELIVEREDTO VEHICLE

ONSITECHEMICAL

PROPERTY DAMASEPREVENTION

CHEMICALSTILL* ARECONTAINED

HITltATEEFFECTS OF

TOXIC/CORROSIVEFUMES

CONVENTIONALPROPERTY DAMASE

CONTROL

FIREEXISTS

FIRESARE CONTROLLED

AND DAMA9ED AREASARE SECURED

DAMMED EOUIPHENTHAMPEftMB RESPONSEACTIONS IS REMOVED

207

_L

RADIOLOGICAL/CHEMICALPROPERTY DAMAGE

CONTROL

METHOD ISDEVISED FORCLEANUP OF

CONTAMINATED AREA

AFFECTED AREASARE

DECONTAMINATED

/DECONTAHINATIO

CONTAHINATEDMATERIALS ARC

DISPOSED OFH W K S L Y

•CLEANLINESS' OFDECONTAMINATED

AREA I S CERTIFIED

RESPONSIBLEAUTHORITYrMTIFIED

TEAMNOTIFIED

TEAMAVAILABLE

TEAMRESPONDS

TEAMEXECUTES

TASKS

RESPONSIBLE AUTHORITY IDENTIFIEDCOMUNICATIONS AVAILABLE

TEAM IDENTIFIEDTEAM LOCATION KNOWNCOMMUNICATIONS AVAILABLE

NO HIGHER PRIORITY EXISTS

TRANSPORTATION EQUIPMENT AVAILABLETRANSPORTATION ROUTES AVAILABLEAREA ACCESSIBLE

TEAM ADEQUATELY TRAINEDREQUISITE EQUIPMENT AVAILABLE

Figure 2 - Elements Common to TaskImplementation

task performance.

B. Information System/Decision Model.Figure 3 illustrates a typical Informa-

tion System/Decision Model to aid in the controlof acute exposure pathways (Section IV.B.I andTree 9). The decision model considers the avail-ability of specific success paths at a particu-lar time. For instance, the evacuation successpath would not be available to al l if certainroads were not passable. The adequacy of such amodel is crucial to the success of the entireemergency plan. The model itself can be classi-fied according to a functional/goal tree struc-ture. The need for institutional activities insupport of the Information System/Decision Modelis apparent. Other information systems whichenter the tree have many elements in common withthat of Figure 3.

C. Tree Quantification.The very nature of emergency response

defies quantification since there is no clearly

Table 1

Institutional Activities for Task Performance

1. Task Implementation RequirementsA. Task DesignationB. Implementation ProceduresC. Auxiliary Procedures

(Communications, Notification, Assem-bly)

D. Training(Programs and Exercises)

2. Equipment RequirementsA. Task Specific Equipment/Facilities

1. Specification and Procurement2. Storage, Installation or Construc-

tion3. Maintenance

B. Peripheral Equipment(Communications and Transportation)See 2.A.I to 2.A.3

3. Disseminate Information to UsersA. Identification of Responsible Authorities

and How to Contact ThemB. Notification and Response ProceduresC. Equipment/Facility Availability and Lo-

cation

TO F1MCT1OW. THEE

Figure 3 - Information System/DecisionModel for Control of AcuteExposure Pathways

208

defined measure of overall success. However, itis possible to define quantitative measures ofsuccess at lower levels of the tree to testalternate success paths. For instance, a mea-sure of success for the subfunction ControlAcute Exposure Pathways (Tree 9) can be thetotal or maximum dose to the population for agiven accident scenario. Consequence modelssuch as CRAC and CRACIT can thus be employed totest alternate sheltering, evacuation and con-trolled release strategies. Proposed modifica-tions in the Information System/Decision Modeland the success paths can thereby be assessed.For instance, different evacuation paths, speedand distance scenarios can be considered. Theimpact of decision time can be related to war-ning and delay time.

D. Summary.A goal oriented tree structure has been

developed which provides a clear visualrepresentation of the functions that must beperformed to insure successful emergency pre-paredness. In addition to employing it as atraining tool, the structure has several otherpotential uses. For instance, the particularinformation needs, Implementation procedures andregulations which apply to each success path canbe noted on the tree to provide a contextualoverview of their importance. The adequacy ofpresent capabilities can thereby be assessed.The tree structure can be customized to meetspecific needs. Success paths that are notrelevant at a particular site can be deleted.Civilian authorities and planners can deleteonsite functions and those only responsible foronsite measures can remove the offsite func-tions.

VII. ACKNOWLEDGEMENTS

The lower branches of the tree were devel-oped at the University of Maryland by a taskgroup under the direction of the first author.The team members were George Harrison, DonMarksberry, Christine Reilly and Bob Starrett.Special thanks to Niall Hunt and Malcolm Patter-son of Baltimore Gas and Electric Co. and JoeBraun of Combustion Engineering for many usefuldiscussions concerning the tree structure.

VIII. DISCLAIMER

Any opinions expressed or Implied hereinare solely those of the authors. Several con-troversial functions and success paths appear onthe goal tree for the purpose of completeness.The presence in no way implies that the use issupported or condoned by any other persons oragencies associated with this work.

IX. REFERENCES

Combustion Engineering, "An Integrated Approachto Economical, Reliable, Safe Nuclear PowerProduction," CE Report No. ALO-1011, April1982.

Technology for Energy Corporation, "The Assess-ment of the Standard Review Plan Using theIntegrated Approach," TEC Report No. R-81-038, September 1982.

University Research Foundation, University ofMaryland, "DOE Information Systems Project,Task G Report - Emergency Preparedness," May1983.

ANS Topical Meeting on Radiological Accidents—Perspecti1""1 ••••••* ''•av tfency Planning

Results of the Early NRC Analyses of the RadiologicalMonitoring Data from the Chernobyl Accident

Rosemary Hogan and Thomas McKenna

ABSTRACT: This paper describes the work ofone group of the NRC task force for theChernobyl accident during the first twoweeks following the accident. This groupfocused on organizing the available environ-mental measurement data into a data basewhich could be used to gain insights intothe consequence throughout Europe and thescope of accident.

The Nuclear Regulatory Commission'sIncident Response Branch, maintains anOperations Center in Bethesda to be usedduring a nuclear emergency involving anyof its licensees, especially the nuclearpower reactors. When news of the accidentat the Soviet Union's Chernobyl NuclearPower Plant reached the NPX, it was immedi-ately realized that information, explana-tions and assessments would be requestedfrom the NRC by many sources includingCongress, the media and other Federalagencies. It was apparent that the NRCmust be prepared to respond to these querieswithin a short time period. Since theOperations Center is designed to be usedfor this kind of situation, it was anobvious decision to activate the Centerand perform the necessary monitoring andassessment activities using our emergencyprocedures which were already in place.This paper contains the assessment ofthe environmental data from the accidentwhich was prepared for an immediate responseto requests for information. As expected,the staff was asked to brief the ACRS, theCommissioners and Congressional subcommittees.Within about ten days after activating theemergency organization, the task group wasable to provide an analysis of the firsteight days data.

This report contains that early analysis.After their report was prepared, that partof the task group was dissolved and no furtherrefinements of the data or computer analyseshave been performed since then.

On April 26, 1986, the fourth unit ofthe Soviet Union's Chernobyl Nuclear PowerPlant experienced a severe accident that re-

leased a large quantity of radioactivity tothe environment. The first evidence of thisaccident appeared in Sweden during routineradiological monitoring on April 28. Theelevated readings were first attributed toa Swedish source. However, further investi-gation indicated that the source was some-where in the Soviet Union.

With the world alerted to these ele-vated radioactivity levels, Soviet officialsacknowledged the accident and confirmed therelease of radioactive material from thereactor. Increased radiological monitoringby many countries around the world producedlarge amounts of environmental data thatcould be used for protective action consi-derations by potentially affected countries.These results also could be used to betterunderstand the accident because little in-formation on the accident scenario and sourceterm was available for the first several daysafter the accident.

A task force was formed at the U. S.Nuclear Regulatory Commission (NRC) in sup-port of the Environmental Protection Agency(EPA) to perform accident assessments andproject consequences. To accomplish thesetasks, information was requested from avai-lable sources throughout the world. Mostof the information received was radiologicalenvironmental monitoring data.

Radiological environmental monitoringdata were sent to the NRC from European coun-tries beginning on April 29, 1986. The analy-ses were performed by various state radiationagencies and nuclear facilities. Early datawere received from the Scandinavian countriesfollowed by data from West Germany, Poland,Romania, Hungary, Yugoslavia, Austria,Belgium, and the Netherlands. Later, datafrom Switzerland, England, Spain, Italy,Korea, Japan, Canada, Israel, and the UnitedStates were obtained. Data continuallyreceived from the Scandinavian countriesand Eastern Europe during the days followingthe accident provided a picture of radiationlevels over time. Although data are stillbeing provided, this report contains onlyreports received by NRC between April 28 andMay 9, 1986.

209

210

Samples of air, drinking water, milk,rain water, human food, and a-Kimal feed wereanalyzed by the various countries for radio-activity concentration. Ground contaminationvalues also were provided. Although grossmeasurements and complete radioisotopicanalyses were provided for this report, thedata used were primarily limited to theiodine-131, cesium-134, and cesium-137 re-sults.

The data were first reviewed .to identifymeasurements that were candidates for consi-deration in the analysis. The team reviewedapproximately 5000 measurements. Those mea-surements that were of interest and appearedto be complete were selected for further analy-sis. The data were entered on computer formsand then formatted into a data base. Foreach reported result, information was includedon the source document, city, country, lon-gitude, latitude, date and time of sampling,and the isotope. Any results that did notinclude this essential information were re-jected from the data base. To simplify thedata base, all human food and animal feedresults were entered as forage. The resultswere entered in the reporting units as theywe"e received on the original source document.

Once entered, the data were verifiedagainst the raw data. Additional reviewswere performed to ensure that the longitudeand latitude were consistent for each cityor region and that the reporting units foreach sample matrix were appropriate. Con-version factors were used to convert allthe data to standard units of picocuriesper liter (pCi/1), kilogram (kg), cubicmeter (m 3), or square meter (m2). The datawere then sorted from lowest to highest valueby sample matrix. The highest and lowestmeasurements were checked for "reasonable-ness" based on location and date. About1000 field measurements remained in the database after all the edits were performed.

Any conclusions derived from this analysismust consider the quality of the data. Withvery few exceptions, the data were receivedwithout information on the sensitivity ofthe measurements or the propagated errorof the analyses. Without this information,all data were necessarily treated equally.In addition, some data were received inEnglish directly from the organizationperforming the analyses. These data werereadily understood and there was some confi-dence in the validity of the results. Therewas less confidence in data that may havebeen recorded by personnel who were unfami-liar with radiological nomenclature and theparameters nscessary for interpretation. Forexample, some reports listed results of radio-activity without indicating whether the resultswere gross beta, gross gamma, or a specificradionuclide. Some data failed to indicatethe city within the country or the Englishtranslation of the city name, thereby makingit difficult to locate the city name on a map.Some data were not fully translated. As aresult of the unknown quality of the data,the analysis w?s performed with all data

treated equally and with the understandingthat the conclusions should not be consideredquantitative but rather would indicate rela-tive trends during the time of interest.

The results were searched and the max-imum daily concentration for ground deposi-tion (pCi/m2), forage (pCi/kg), milk (pCi/1),rain (pCi/1), water (pCi/1), and air (pCi/m3)were selected for each 2° square of latitudeand longitude. These values were comparedwith the baseline concentrations given 1nTable 1. The highe'st ratio of baseline valueto measured result was calculated.

The baseline value for air is the calcu-lated concentration in air that would resultin exceeding 1.5 rem [U. S. Food and DrugAdministration (FDA) preventive action Protec-tive Action Guide (PAG) basis] to the childthyroid in 5 days based on the assumptionsin NRC Regulatory Guide 1.109.

The FDA has issued recommendations* forState and local agencies In the event of acci-dental radioactive contamination of humanfood and animal feed. These recommendationsor guidelines are projected doses at whichresponsible officials should take protectiveactions. There are two levels of guidelinesincluded in the recommendations. The emer-gency PAG should trigger isolation of foodto prevent its introduction to the public,even 1f the impact of such actions is signi-ficant. These drastic measures are justifiedbecause of the high projected health ha-ards.The preventive PAG should trigger actions onthe part of the State or local officials onlyif there would be minimal impact.

The preventive PAGs are much lower pro-jected doses that do not represent significanthealth hazard. These more conservative PAGswere used as the baseline values in theanalysis of the environmental data from theChernobyl accident. The preventive PAG Is1.5 rent projected dose commitment to thethyroid or 0.5 rem projected dose commitmentto the whole body, bone marrow, or any otherorgan. The FDA has derived specific radio-nuclide concentrations for area deposition,forage, and milk that are equivalent to theprojected dose commitment PAGs. The radio-nuclide concentrations or "response levels"were calculated used the total intake of thePAG (1.5 rem), the average dally consumptionof specified foods, and the days of intakeof the contaminated food. Response levels,measured in units of radionuclide concentra-tions, provide a convenient value for Stateand local officials to make protectiveaction decisions.

The baseline values used for ground, milk,and forage are the FDA response ?vsls forpreventive PAGs. The rain and water baselinevalues used are the same as the FDA PAG for

* Federal Register Notice 47 FR 47073,October 22, 1982, "FDA Accident Conta-mination of Human Food and Animal Feeds;Recommendations for State and LocalAgencies."

211

milk. Obviously, this is not valid for mostareas because humans do not consume rain waterdirectly.

To show when and where the baseline valuesin Table 1 (preventive PAGs) had been exceeded,the highest daily measurement in each countrywas determined. Three ranges were shown onmaps (1) those that exceeded the baselines,(2) those within 50% of the baselines, and(3) those below the baseline.

The measurements used to characterize acountry were validated by assuring that therewere other measurements that would place acountry in the same category or other measure-ments that were within a large fraction of thehighest value. For countries with only onemeasurement, this measurement was comparedwith those from nearby countries for reason-ableness.

The single measurements were used tocharacterize an entire country, because ofwhen displayed on a global scale the limitednumber of actual measurement locations didnot provide an effective briefing tool.

Examples of the resulting maps for April29, May 1 and May 4 are shown in figures 1,2, and 3.

If there were no data from a country fora particular day, the map indicates no mark-ings. By following the maps by day, a generalpattern where preventive actions would be con-sidered is evident. It is interesting to notethat in the early days after the accident, theair and rain PAGs were approached or exceeded.With the passage of time and distance from the

accidents, the PAGs were exceeded in forageand ground. Milk results appeared as the high-est measurement in Sweden on May 3, inYugoslavia on May 5, and Switzerland on May 6.

The NRC staff also examined ground conta-mination for long-lived isotopes of cesium-134(2.1 years) and cesium-137 (30.2 years). This(night provide insight into the extent of pos-sible long-term prob'oms with agriculture.The data base contained only 28 locations withreported levels of ground contamination forcesium-134 or cesium-137. Only one locationreported levels above the preventive PAGfor ground contamination; all other locationsreported levels less than one tenth of thePAGs.

Although data were not available fromall countries, for all days and for all samplemedia, the available radiological measurementsprovided a general pattern of contaminationthat would affect the ingestion pathway. Themaps show the extend and general movement pat-terns of the contamination at a level of detailconsistent with the data. However, due tolimitation of the data discussed in the paper,the staff always took care to assure eachaudience was aware of the limitations of thedata.

Despite the incomplete data and the un-known quality of the data, sufficient infor-mation was available to characterize the leveland scope of radiological contamination fromthe accident at the Chernobyl Nuclear PowerPlant beyond the Soviet Union.

Table 1 Baseline values

Matrix Units

Ground deposition

Forage

Milk

Water

Rain water

Air*

Unit ofMeasure

pCi/m2

pCi/kg

pCi/1

pCi/1

pCi/1

pCi/m3

1-131

1.3X105

5.X104

1.5X10*

1.5X10*

1.5X10*

7X10J

Cs-134

2X1C6

8.X105

1.5X105

1.5X106

1.5X105

-

Cs-137

3X1O6

1.3X105

2.4X105

2.4X105

2.4X105

-

"All values except air are based on FDA preventive PAG response levels.

212

CHERNOBYLMAXIMUM REPORTED CONTAMINATION LEVELS

VS. LOWER LIMIT PAG

APRIL 29. 1986

GERMANY / POLAND

IllillllSWITZERLAND

FRANCE

LEGEND:

ABOVE PAG

[ WITHIN 50% OF PAG

Illlll l l l i LESS THAN 50"« OF PAG

MONGOLIA

S J JAPAN

CHINA

Figure 1

213

CHERNOBYLMAXIMUM REPORTED CONTAMINATiON LEVELS

VS. LOWER LIMIT PAG

MAY 1, 1386

LEGEND:

ABOVE PAG

WiTHIN SO".' OF PAG

IHIIIIIIil LESS THAN 50".. OF PAG

MALTA

JAPAN

Figure 2

214

CHERNOBYLMAXIMUM REPORTED CONTAMINATION LEVELS

VS. LOWER LIMIT PAG

MAY 4, 1986

ABOVE PAG

WITHIN 60% OF PAG

LESS THAN 50% OF PAG

S J JAPAN

Figure 3

ANS Topic*! Meeting on Radiological Accidents-Perspectives and Emergency Planning

ARAC: A Centralized Computer-Assisted EmergencyPlanning, Response, and Assessment System for

Atmospheric Releases of Toxic MaterialM. H. Dickerson and J. B. Knox

ABSTRACT The Atmospheric Release Advisory Ca-pability (ARAC) is an emergency planning, response,and assessment service, developed by the U. S. Depart-ments of Energy and Defense, and focused, thus far, onatmospheric releases of nuclear material. For the past14 years ARAC has responded to over 150 accidents,potential accidents, and major exercises. The most no-table accident responses are the COSMOS 954 reentry,the Three Mile Island (TMI-2) accident and subsequentpurge of 85Kr from the containment vessel, the recentUF6 accident at the Kerr-McGee Plant, Gore, Okla-homa, and the Chernobyl nuclear reactor accident inthe Soviet Union. Based on experience in the area ofemergency response, developed during the past 14 years,this paper describes the cost effectiveness and other ad-vantages of a centralized emergency planning, response,and assessment service for atmospheric releases of nu-clear material.

I. INTRODUCTIONThe Atmospheric Release Advisory Capability

(ARAC), (Dickerson et al., 1985) has developed overthe past 14 years from merely a concept in 1972 to itspresent role as a federal emergency planning, response,and assessment resource. From the beginning, ARACwas designed to be a centralized resource of a highlytrained and specialized staff devoted to all aspects ofemergency response, and to reduce duplication of capa-bilities, software development, and maintenance. Thisconcept was not intended to replace local functions orresponsibilities; in fact, it was designed to complimentand enhance local emergency response capabilities ofindividual nuclear facilities.

During the development and implementation ofARAC, the Department of Energy (DOE) and Depart-ment of Defense (DOD) have been major supporters

and users of the service. Presently there are approxi-mately 50 DOE and DOD facilities connected directly toARAC, each having on-line databases for terrain, geog-raphy, and meteorological measurement locations perti-nent to their own facility. In addition to these facilities,ARAC supports the DOE response to any nuclear eventcapable of releasing radionuclides into the atmosphere,the Nuclear Regularatory Commission (NRC) responseto nuclear power plant accidents, the Federal AviationAdministration (FAA) response to incidents in whichaircraft might intercept radioactive material, and theEnvironmental Protection Agency (EPA) response! toincidents in which radioactive material has left or might,leave facility boundaries.

Advantages of a centralized emergency planning, re-sponse, and assessment system are that it:

• Avoids duplication of resources, and provides astate-of-the-art, proven response capability;

• Provides experienced staff devoted to emergencypreparedness, response, and assessment;

• Is cost effective when applied to a large num-ber of nuclear facilities and integrated into thefederal emergency preparedness programs;

• Provides a standard (or criterion) for emergencyresponse assessments while maintaining flexi-bility to meet site-specific and agency require-ments;

• Focuses research and development on timelyimprovement and evaluation of emergency re-sponse resources; and

• Applies integrated research and development re-sources to specialized emergency response re-quirements in real-time, e.g., Cosmos 954, TMI,Gore, (Oklahoma), aSd Chernobyl events.

On the other hand, the disadvantages of a centralizedsystem are that it:

• Is cost effective only when applied to a broadbase of nuclear facility and federal agency re-quirements;

215

216

• Can be viewed by local authorities as a "threat"to their capabilities and responsibilities; and

• Can be viewed by local authorities as a mecha-nism for reducing or eliminating their responsi-bilities.

During the development and implementation phasesof ARAC, a balance between the advantages and disad-vantages of a centralized system has emerged. As statedearlier, ARAC now serves as a national emergency re-sponse resource for several federal agencies. It has de-veloped an extensive background in emergency responseby responding to over 150 accidents, potential accidents,and major exercises. The most notable ARAC responsesare:

• Savannah River Plant (SRP) Tritium Release,1974Train accident involving UFg, 1976Chinese 200 kt and 4 mt atmospheric tests, 1978COSMOS 954 reentry, 1978TMI Nuclear Power Plant accident, 1979Titan II accident, 1980"SRP H2S leak and transfer, 1981*Ginna Nuclear Power Plant accident, 1982Gore, Oklahoma, UF6 accident, 1986"Chernobyl USSR Nuclear Power Plant accident,1986

The remainder of this paper will discuss the roleresearch and development has played in responding toaccidents, both in real-time and in the model evaluationarea—two significant attributes of a research and devel-opment group co-joceted with an operational emergencyresponse center. f

II. RESEARCH) CONTRIBUTIONS TO REAL-TIME EMERGENCY RESPONSES

The various roles ARAC played during and afterthe TMI accident response provide an excellent basisfor describing how the system is used, and the value ofclosely associated research and development. The fivebasic roles ARAC filled during and after the accidentare that it:

• Provided guidance on deployment of radiologicalmeasurement systems;

• Helped interpret surface and airborne radiolog-ical measurements,

• Estimated the 133Xe source term;• Provided guidance to the FAA for air traffic

safety in and out of the Harrisburg airport; and• Estimated total population dose for the Presi-

dent's Commission on TMI.A few of these roles, such as advising the mea-

surement teams and the FAA, were relatively straight-forward. Estimating the source term and modifyingthe MATHEW/ADPIC (Sherman, 1978; Lange, 1978a)models to estimate "man-rem" for the President's Com-mission required model modifications and interpreta-tions of data by the research staff. To estimate thesource term required a knowledge of both the response

* Toxic Chemical Releases

function of the instruments used in the aircraft to mea-sure the 133X, and of the characteristics of the ADPICtransport and diffusion estimates, which use particles tosimulate radioactivity. This coupling of information ledto estimates of the source term that were made avail-able during the first four days of the accident. Later acomparison of the model calculated source term with asource term estimate provided by the President's Com-mission showed agreement within a factor of two tothree.

During the ARAC response to the Soviet Cosmos954 satellite reentry into Canada, the ARAC researchteam was able to modify a nuclear weapons falloutmodel (KDFOC2) (Serduke, 1978) so that it could beused to simulate the depositional "footprint" of radioac-tive particles generated with varying densities at alti-tudes between 20 and 60 km. These "footprint" simu-lations, together with the ground measurements, wereused to define and to limit the search areas to manage-able sizes. For this event, ARAC also provided guidanceon the appropriate time and positioning for launchinga balloon-borne measurement system into the strato-sphere to observe the amount of fine particulate mate-rial (i.e., the material that remained for months in thestratospheric circulation regime). This guidance servedto eliminate costly and prematurely arranged balloonflights when the regularly scheduled future flights wouldprovide the required concentration measurements.

For the UFg release at the Gore, Oklahoma facil-ity of Kerr McGee, the ARAC research team workedwith LLNL chemists, Oak Ridge National Laboratory,and the Atmospheric Turbulence Diffusion Division(NOAA) scientists to define the chemical and physi-cal characteristics of the source term and the dispersionprocesses. These input data were used to define andparameterize the amount of HF and UQ2F2 released inthe process, the cloud rise due to exothermic reactions,and the particle size attributed to the UO2F2. Withoutthis research support, the ARAC response to this acci-dent would not have been as timely and useful as it wasto the NRC on-site assessment team.

The Chernobyl accident, because of its magnitudeand limited information, has provided the largest chal-lenge to date for the ARAC research and developmentteam. The team was "called on" to (1) expand theMATHEW/ADPIC grid from a horizonal size of 200 x200 km to 1920 x 1920 km, (2) estimate the verticalextent, time history, and magnitude of the source, and(3) retrieve a global particle-in-cell model (PATRIC),(Lange, 1978b) which was originally developed for esti-mating transport and diffusion in the stratosphere, from6 years of storage and modify the model to simulate theupper level (tropospheric) release created by the initialexplosion and fire at Chernobyl.

The first part of the MATHEW/ADPIC calculationcovered a 200 x 200 km region centered on the Cher-nobyl reactor site (Figure la); it became apparent thatthis calculation was insufficient to answer the questionsarising from the spread of radioactivity across the So-viet boundaries into the rest of Europe. Thus, the first

217

Figur| 1. The calculational grids used for the MATHEW/ADPIC models: a)the calculated infant thyroid dose due to 1:"I inhalation for a six-hour period on29 April sjiown on a 60 km expanded subregion of the 200 km grid; b) the final1920 x l{ 20 km grid (48 km cells) with the nested subgrids (3 km, 6 km, 12 km,and 24 km cells) outlined pud the initial grid (200 x 200 km) shaded.

MATHEW/ADPIC modeling effort was terminated, aspart of our ARAC response, and a second, larger-scalesimulation was initiated, covering the largest area pos-sible within the limitations of the model and availablecomputer resources.

The MATHEW/ADPIC grid chosen was 1920 kmon a side, and extended 2100 m vertically. The horizon-tal grid mesh was 48 km and the vertical grid spacingwas 150 m. This particular grid size was chosen becauseit represented the largest grid that could be used for theMATHEW/ADPIC models without a major revision ofthe computer codes. In addition, this grid allowed forcoverage of a reasonably sized area of interest for theinitial dose and deposition calculations. In Figure lb, anested SP ^ling grid is shown around the source (Cher-nobyl) for horizontal cell sizes of 3 km, 6 km, 12 km,and 24 km, respectively. These nested grids were usedto sample the particles that produced surface air con-

centration and ground deposition estimates near the re-actor site.

The source term for the reactor accident was di-vided into two parts, a lower and an tipper cloud. Thelower cloud was assumed to be produced over a periodof six days as a result of heat from the burning reac-tor. The upper cloud was assumed to be produced byone or more of the following: explosions followed by ahot fire for several hours, convective activity near theChernobyl reactor site associated with thunderstorms,or lifting over a warm front located between Chernobyland the Baltic Sea. One major part of the ARAC effortduring the first two weeks following the accident was as-sociated with the determination of a lower level sourceterm and the associated consequences.

Employing both the grid shown in Figure lb and theinitial source estimate derived from the environmental

218

22001— 2100^

20

(A) (B) (C)

Figure 2. Depictions of the vertical distribution of source material (explosion andfire cloud): a) average distribution as used in the ADPIC calculations; b) the nightand c) day, respectively, vertical distributions calculated with a non-hydrostatic cloudmodel.

measurements of 137Cs and 13II in Scandinavia and Eu-rope, the MATHEW/ADPIC model was used to refineinitial estimates of the iow level source term for the firstsix days of the accident. Providing a reasonable descrip-tion of this source term required a vertical distributionof the radioactivity, as well as a time dependence of therelease.

Figure 2a shows the vertical distribution of mate-rial used to define the lower level source term in theMATHEW/ADPIC calculations. This estimate wasbased on prior experience simulating heated plumeswithin air masses—air masses which contained verti-cal temperature distributions sl-nilar to those shown byvertical atmospheric soundings near Chernobyl. Eightyper cent (80%) of the released material was assumed toreside between 1000 m and 1500 m, with the remain-ing 20% located between the surface and 1000 m. Themaximum concentration was at 1000 m.

Toward the end of the two-week period immediatelyfollowing the accident, the vertical distribution of thelow level source term was quantitatively estimated by atwo-dimensional, high resolution, non-hydrostatic cloudmodel. This model was originally developed to simulatethunderstorms but, more recently, has been applied tosimulate plumes of smoke induced by large-scale urbanfires. The estimated source strength for the model cal-culations was 62 megawatts, about the resident heat en-ergy expected from the shut-down of a 3200 mw thermalreactor. Figures 2b and 2c show the vertical distributionof material, calculated by this model, for both night-time and daytime atmospheric soundings taken nearthe Chernobyl reactor site near the time of the acci-dent. During the nighttime (Figure 2b), the model esti-mated that 80% of the material was contained between600 m and 1000 m. The remaining 20% of the materialwas located between 100 m and 600 m with the maxi-mum concentration located at 500 m. For the daytime(Figure 2c), 80% of the material was determined to bebetween 800 m and 2100 m, with the remaining 20%between 100 m and 800 m. The maximum value for

this case was at 1300 m. Differences between dose anddeposition estimates, based on the distribution of ma-terial in the original source term estimate (Figure 2a)and the non-hydrostatic model estimate (Figures 2b,2c), would not be large, particularly at distances of sev-eral hundred kilometers and beyond. For this reason,the dose and deposition estimates were not recalculatedusing the quantitatively-modeled vertical distributionsof material (Figure 2).

From the PATRIC hemispheric scale model cal-culations, it rapidly became apparent that some ra-dioactive material was convected or lofted (or both)to much higher altitudes than that assumed for theMATHEW/ADPIC calculations (described above). Aseries cf calculations with material placed at 2500, 4200,and 5500 m failed to transport contamination to Japanand North America even close to the recorded arrivaltimes—if at all. While limitations of the model re-duce some of the precision desired, it presently ap-pears that at least some material had to be injectedto altitudes above 5500 m in order to account for air-borne air concentration and surface rainwater and milkmeasurements of the radioactivity in Japan and theUSA. Original estimates of dose and deposition fromI31I and 137Cs, made during the first 3 weeks of theaccident, for eastern and western Europe using theMATHEW/ADPIC model and global estimates of dis-persion with the PATRIC model, are reported by Dick-erson and Sullivan, 1986. Further refinements of theseestimates have recently been reported by Gudiksen andLange, 1986.

III. MODEL EVALUATION STUDIES

One of the most significant continuing research anddevelopment efforts has been in th» area of model eval-uation and improvement. Over the past several years,better diffusion parameterizations utilizing space vary-ing surface roughness heights, and the use of multiplevertical wind .profiles and nested grids for better con-

219

METEOROLOGY TRACEREXPERIMENT TERRAIN STABILITY WINDSPEED MEASUREMENTS

(m/s) (km)

INEL 1971SRP 1974TMI 1980ASCOT 1980ASCOT 1981EPRI 1981SRP MATS 1983MONTALTO 1983

ROLLINGROLLINGROLLINGCOMPLEXCOMPLEX

FLATROLLINGCOASTAL

CF-CF-CF-EF-EF AD-BC-B

2-61-41 40-40-41-51-81 6

7-803 3040-60

1-81 101-50~ 2 01-6

Table 1. Summary of MATHEW/ADPIC Model Evaluation Studies

centration estimates near the source point have con-tributed to improving the MATHEW/ADPIC models.Many of these improvements were designed and imple-mented as a result of model evaluation studies whichwere done to define the expected accuracy of the mod-els under various terrain and meteorological situations.Table 1 lists the evaluation studies conducted with theMATHEW/ADPIC models during the past 12 years.Contained in these studies are 26 individual experimentsconducted in 6 different geographical areas. They rep-resent approximately 3000 tracer measurements span-ning a wide variety of diffusion categories. (Dickersonand Lange, 1986) The experiments shown in Table 1utilized a multitude of tracers, including routine emis-sions of 41Ar from the SRP nuclear reactors, the con-trolled venting of 85Kr from the TMI containment, I31Ireleases at the Idaho National Engineering Laboratory(INEL), sulfur hexafluoride releases from the SRP, theMonttslto, and Kinraid power plant sites, and perfluo-rocarbon and heavy methane releases that were part ofthe ASCOT experiments. The releases occurred fromthe 62 m stacks at the SRP and TMI, and from the187 m stack at the Kincaid power plant. The remain-ing releases generally occurred near the surface, exceptfor one heavy methane tracer that was released at 60 mduring the 1980 ASCOT experiments, and one perfluo-rocarbon tracer released in a cooling tower plume dur-ing the 1981 ASCOT experiments. The duration of thereleases varied from 15 minutes to several hours. Exten-sive surface sampling networks were employed in eachseries of experiments. Maximum distances were 80 kmfor the 1971 INEL studies, 50-60 km for the EPRI andTMI studies, 30 km for the MATS experiments, 10 kmfor the ASCOT experiments, and approximi tely 6 kmfor the studies at Montalto, Italy. The experimentswere supported by a variety of surface and upper airmeteorological observations; data was provided by mea-surements ranging from normal meteorological coverageprovided by the National Weather Service (NWS), to alocal site tower and an extra upper air sounding duringthe TMI purge of 85Kr. Data were supplied by a widespectrum of measurement systems, including acousticsounders, tethersondes, rawinsondes, optical anemome-ters, and towers that were an integral part of the AS-COT experiments.

It is difficult to devise a statistical process that ad-equately describes a model's performance when com-pared to tracer field data, particularly when the fielddata span a broad spectrum of release and samplingtimes, sampling distances, terrain, and meteorology.For example, the standard correlation coefficient isused sometimes; however, one point at the high endof the scale can influence the entire data set. Earlyon, a rigid technique, but one considered a stan-dard, was chosen for comparisons of tracer measure-ments to the MATHEW/APDPIC model calculations.A factor R is computed for each pair of measure-ments (Cm) and model calculations (Cc) which repre-sents the whole - number ratio between the two. Foreach experiment the percent of comparisons within afactor R are plotted as a function of R. The defi-nition of R is R = (Cm + B)/(CC + B), except if(R < 1,) then R = (Cc + B)/(Cm + B), and B is back-ground.

Figure 3, based on the factor R, depicts a summaryof the performance of the MATHEW/APIC models todate. The best simulation of the experimental datais given by the upper curve, which is associated withrolling terrain and near-surface tracer releases. Themost difficult simulation is associated with complexterrain and elevated releases. Other situations provideresults that are intermediate to these curves. Hence, thebest results indicate that the calculated concentrationsare within a factor of 2 for 50% of the measured concen-trations and within a factor of 5 for 75% of the compar-isons. This performance degrades to 20% and 35% forfactors for 2 and 5, respectively, for the comparisons as-sociated with elevated releases in complex terrain. Thisdegradation of results in complex terrain is due to a va-riety of factors, such as the limited representativenessof measurements in complex terrain, the limited spatialresolution afforded by the models, and the turbulenceparameterizations used to derive the eddy diffusivities.

IV. FUTURE RESEARCH OPPORTUNITIES

The most significant improvement in emergency re-sponse can be attained by the development and imple-mentation of an operational mesoscale (out to 200 km)time-dependent forecast model. Technology is availabletoday to develop a model that can be applied to a range

220

100%

I

Ia.

- Rolling terrainnear-surface

" release

Complex terrainelevated release

10" 101

Factor R

102

Figure 3. Percent of computed air concentrations withina factor R of measured values. The figure provides a mea-sure of the spectrum of model evaluation results that spanfrom near-surface tracer releases in rolling terrain to elevatedreleases in complex terrain.

of forecast problems and consequently would be uso-ful (but not all inclusive) for emergency response pur-poses. This is definitely a high pay-off, low risk areaof research and development. Chernobyl has shown thevalue of having the capability of simulating rainout, inthe transport and diffusion models. The state-of-the-science in this area has advanced to the point whererainout is technically feasible to implement, althoughadditional research would be required to improve thesimulations and make them more realistic. A logisti-cal problem remains which involves obtaining spaciallyvarying rainrate data, and providing a mechanism forincluding these data directly into the transport-diffusionmodels.

ARAC does provide a foundation for addressingtoxic chemical response as a centralized system; how-ever, many more technical unknowns are involved indealing with toxic chemicals as opposed to radioactivematerial. Also the frequency of accidents and the num-ber of chemicals are considerably greater and the healtheffects are known to a lesser degree than for nuclear ma-terial. In general, toxic chemical releases can be dividedinto four classes based on their physical and chemical re-activity and their density with respect to density of theambient air. These 4 classes are: non-reactive chemi-cal/ambient air density, non-reactive chemical/heavier-than-air density, reactive chemical/ambient air density,and reactive chemical/heavier-than-air density. Given atoxic chemical release where the chemical is non-reactiveand whose density is approximately that of ambient air,the MATHEW/ADPIC models would be expected toperform as well as they do for nuclear material. If thereleased material is non-reactive and heavier-than-air,models are available to estimate consequences; however,considerable effort would be required to place them inan operational environment. (Gudiksen et al., 1986)

Chemically toxic releases that are both chemicallyand physically reactive at the source point and dur-ing the dispersion processes, would require a large re-search effort before the environmental consequences canbe modeled with confidence. A joint research and de-velopment and implementation effort is required beforeARAC or any other centralized or local emergency re-sponse system can be expected to address a range ofaccidental releases of toxic chemicals with any degree ofconfidence.

V. ACKNOWLEDGEMENT

Work performed under the auspices of the U.S. De-partment of Energy by the Lawrence Livermore Na-tional Laboratory under Contract W-7405-Eng-48.

VI. REFERENCES

Dickerson, M. H., et al., 1985, "ARAC Status Report:1985," UCRL-53641, Lawrence Livermore Na-tional Laboratory, Livermore, CA.

Dickerson, M. H., and T. J. Sullivan, 1986, "ARAC Re-sposne to the Chernobyl Reactor Accident," UC1D-20834, Lawrence Livermore National Laboratory,Livermore, CA.

Dickerson, M, H., and R. Lange, 1986, "An Exampleof Emergency Response Model Evaluation Stud-ies Using the MATHEW/ADPIC Models," UCRL-94388, Lawrence Livermorc National Laboratory,Livermore, CA.

Gudiksen, P. H. and R. Lange, 1986, "Atmospheric Dis-persion Modeling of Radioactivity Releases Fromthe Chernobyl Event," UCRL-95363, LawrenceLivermore National Laboratory, Livermore, CA.

221

Gudiksen, P. H., et al., 1986, "Emergency ResponsePlanning for Potential Accidental Liquid ChlorineReleases," UCRL-53685, Lawrence Livermore Na-tional Laboratory, Livermore, CA.

Lange, R., 1978a, "ADPIC—A Three-DimensionalParticle-In-Cell Model for the Dispersal of Atmo-spheric Pollutants and Its Comparison to RegionalTracer Studies," J._Appl. Meteor^ 17, 320-329.

Lange, R., 1978b, "A Three-Dimensional Particle-In-Cell Sequential Puff Code for Modeling the Trans-port and Diffusion of Atmospheric Pollutants,"UCID-17701, Lawrence Livermore National Labo-ratory, Livermore, CA.

Sherman, C. A., 1978,"A Mass-Consistent Model forWind Fields Over Complex Terrain," .7. AppJ.Meteor., 17, 312-319.

ANS Topical Meetiag on Radiological AccMemti—Perspectives aid Ewrgeacy Plaaaiaf

Digital Imagery Manipulation and Transmission for EmergencyDeployment Applications

C. A. Gladden and D. S. Phillipson

ABSTRACT The D i g i t a l I m a g e r yTransmission System (DITS) was designedas a replacement for the older slow-scans y s t e m . DITS i s an i n t e r a c t i v emicrocomputer graphics system capable ofd i g i t i z i n g maps or p h o t o g r a p h s .D i g i t i z e d images are s tored on a l o c a larea network f i l e server and can then bedistributed to other units on the systemor be transmit ted over a d i a l - g r a d ephone l i n e to a remote s i t e where otherDITS termina l s e x i s t . A DITS terminalc o n s i s t s o f an i n e x p e n s i v emicrocomputer, an o f f - t h e - s h e l f RGBmonitor and an e l ec tron ic drawing devicefor enter ing o v e r l a y data. Hardcopyp r i n t s can be o b t a i n e d in m i n u t e s .

Dur ing t h e mid 1 9 7 0 ' s , t h eDepartment of Energy (DOE) recognizedthe need for an emergency response teamto provide quick response to a v a r i e t yof n u c l e a r - r e l a t e d emergencies . As ar e s u l t , an emergency response team wasformed, and the t a s k undertaken toassemble the necessary e x p e r t i s e andhardware to support such a mandate. TheA e r i a l Measuring System (AMS) forr a d i o a c t i v e mater ia l was a l ready inexistence and provided the core for thenew emergency response team.

Since the response would be to av a r i e t y of acc ident t y p e s , in of tenunknown l o c a t i o n s , a method was badlyneeded for moving v i s u a l informationsuch as maps and photographs over theinternational d ia l telephone network. Avideo compander (commonly known as slowscan) sys tem was s e l e c t e d for t h i sa p p l i c a t i o n , and whi le i t provided aw e a l t h of i n f o r m a t i o n , i t a l s o hads e v e r a l shortcomings. These problemsb a s i c a l l y i n v o l v e d the i n t e r a c t i v ec a p a b i l i t y of a v a i l a b l e s low scanequipment. While the video information

could be d i g i t i z e d , i t c o n s i s t e d ofo n l y one image, and no method wasreadi ly a v a i l a b l e that would a l low theadd i t i on of m u l t i p l e l e v e l o v e r l a yinformat ion.

With the advent of the inexpensivemicrocomputer, i t became p r a c t i c a l toconvert the maps to a d i g i t a l data base ,g e n e r a t e t h e o v e r l a y i n f o r m a t i o ne l e c t r o n i c a l l y , and to s i m u l t a n e o u s l yimplement the balance of the on-scenecommander's ' t a t u s information systemthat had been in planning for s e v e r a lyears .

The D i g i t a l Imagery TransmissionSystem (DITS), as i t was f i n a l l y named,i s composed of an assortment of standardand s p e c i a l i z e d sof tware , a s e r i e s ofinexpens ive microcomputer*, moni tors ,cameras, and a f i l e s e r v e r . A t y p i c a lDITS t e r m i n a l c o n s i s t s o f t h emicrocomputer, a standard scan RGB videom o n i t o r , and an e l e c t r o n i c drawingd e v i c e for e n t e r i n g the o v e r l a y ormodification information. The system i spackaged in the standard deploymentconta iners for easy movement into af i e l d environment.

Once the area of the emergency hasbeen de f ined , maps are procured andd i g i t i z e d u s i n g the v i d e o camera.Multiple images are d ig i t i zed and storedin a contiguous data base on c commonsystem f i l e server. This map data basei s then a v a i l a b l e to any terminal on thesystem for v iewing and/or input t ingo v e r l a y information such as r a d i a t i o nc o n t a m i n a t i o n l e v e l s and l o c a t i o n s( a c t u a l and p r o j e c t e d ) , p o p u l a t i o nd i s t r i b u t i o n , m e a s u r e m e n t t eaml o c a t i o n s , e v a c u a t i o n r o u t e s , e t c .O v e r l a y s may be v i e w e d in f u l lresolut ion or l /16th resolut ion to showmore than one at once.

A key s y s t e m f e a t u r e i s t h erequirement for a l l o v e r l a y data orm a n i p u l a t i o n s to be c e r t i f i e d at as ing l e s c i e n t i f i c control point prior to

223

?24

d i s t r i b u t i o n to any l o c a t i o n other thant h e o r i g i n a t o r . Once the o v e r l a y datah a s b e e n a d d e d t o t h e i m a g e andv e r i f i e d , i t i s a v a i l a b l e for e l e c t r o n i cd i s t r i b u t i o n to l o c a l and remote u s e r sor f o r d u p l i c a t i o n as hard c o p y orv i e w g r a p h s . In o r d e r t o m i n i m i z et r a n s m i s s i o n t i m e s , map d a t a b a s einformat ion i s g e n e r a l l y on ly moved onceb e t w e e n the s t a t i o n s and common f i l es e r v e r . A f t e r w a r d s , o n l y o v e r l a y s orm o d i f i c a t i o n s are t ransmi t ted and theseare compres sed to f u r t h e r enhance thet r a n s f e r t i m e s . The t r a n s m i s s i o n t imeof a complex o v e r l a y i s o n l y m i n u t e s( t y p i c a l l y l e s s than 3) .

In a d d i t i o n to such f e a t u r e s as"roam" and t w o - s t a g e "zoom", a w i d ev a r i e t y of predef ined geometric p a t t e r n sare a v a i l a b l e to the u s e r as an a i d inc r e a t i n g o v e r l a y in format ion . Freehandd r a w i n g i s a l s o a l l o w e d , and g r i dp a t t e r n s w i t h d e f i n a b l e d i m e n s i o n s ord i s t a n c e s a r e a v a i l a b l e t o p r o v i d eass i s t a n c e .

T h i s s y s t e m h a s b e e n t e s t e ds u c c e s s f u l l y in s e v e r a l CONUS and OCONUSl o c a t i o n s , u t i l i z i n g the I n t e r n a t i o n a lMaritime S a t e l l i t e (INMARSAT) system andt h e i n t e r n a t i o n a l d i a l t e l e p h o n en e t w o r k . G e n e r a l l y , the da ta may bet r a n s f e r r e d u s i n g p r o p e r i n t e r f a c ee q u i p m e n t a n d a n y v o i c e g r a d ecommunications c i r c u i t . Data i n t e g r i t yo v e r l o n g d i s t a n c e s i s i n s u r e d by f u l l100 percent error c o r r e c t i n g modems andt r a n s m i s s i o n so f tware .

Current ly a DITS image c o n s i s t s of16 " s h o t s " (4 x 4) of the v i d e o camerawhich are assembled by the computer i n t oone l a r g e image ( 2 0 4 8 p i x e l s wide by1920 p i x e l s h i g h ) . Any s e c t i o n , 512p i x e l s wide by 480 p i x e l s h i g h , can beviewed in f u l l r e s o l u t i o n , or the wholeimage can be viewed in 1/16 r e s o l u t i o n .

The IBM c o m p a t i b l e computer usedf o r a DITS t e r m i n a l i s the AT&T 6 3 0 0 .I t f e a t u r e s low c o s t , h i g h r e l i a b i l i t yand f a s t o p e r a t i o n . The current drawingd e v i c e i s a t h r e e b u t t o n "mouse" w i t hp l a n s to add a l i g h t pen for a d d i t i o n a li n p u t c a p a b i l i t y . C u r r e n t h a r d c o p yo u t p u t s are 8" by 10" and 35mm p o l a r o i dp i c t u r e s and v i e w g r a p h s . A l s o , a. c o l o rp r i n t e r may be c o n n e c t e d f o r p r o d u c i n gp i c t u r e s with 125 c o l o r c h o i c e s . Ease ofo p e r a t i o n i s a c c o m p l i s h e d w i t h many" h e l p s c r e e n s " t h a t c o n s i s t of t e x tmixed w i t h g r a p h i c i l l u s t r a t i o n s t ob e t t e r demonstrate system o p e r a t i o n . Thes o f t w a r e i s menu d r i v e n and t h e s y s t e mprompts the u s e r f o r a l l i n f o r m a t i o nneeded to accompl i sh the task at hand.

In s u m m a t i o n , t h e DITS s y s t e mp r o v i d e s t h e a e r i a l m e a s u r i n g s y s t e mo r g a n i z a t i o n with an i n t e r a c t i v e d i g i t a lg r a p h i c s s y s t e m t h a t e a s i l y a l l o w s t h ep r i n c i p a l s to be kept in formed of suchi t e m s a s c o n t a m i n a t i o n l e v e l s andl o c a t i o n s , o p e r a t i o n s p r o g r e s s , ors t a t u s . The s y s t e m i s l i g h t w e i g h t ,i n e x p e n s i v e , and b a s e d upon r e a d i l ya v a i l a b l e m i c r o c o m p u t e r t e c h n o l o g y .Transfer of informat ion over v o i c e - g r a d ec o m m u n i c a t i o n s c i r c u i t s a l l o w s remotet e r m i n a l s t o be l o c a t e d , w h e r e v e rr e q u i r e d , on a g l o b a l b a s i s . S i n c e t h einformat ion i s t r a n s m i t t e d in a d i g i t a lf o r m a t , i t i s e a s i l y e n c r y p t e d u s i n gg e n e r a l purpose data encrypt ion d e v i c e s .The s y s t e m c u r r e n t l y c o n s i s t s o f 7t e r m i n a l s and a common f i l e s e r v e r .P r o v i s i o n s have been made t o i n c r e a s et h a t n u m b e r t o a p p r o x i m a t e l y 20t e r m i n a l s for future a p p l i c a t i o n s .

This work was performed by EG&G/EMf o r t h e U n i t e d S t a t e s Department ofEnergy, Of f i ce of Nuclear S a f e t y , underContract Number DE-AC08-83HV10282 .

ANS Topical Meeting oa Radiological AccMeate—Perspectives tad Emertency PUuwfag

Image/Data Storage, Manipulation, and Recall Using Video/Computer Technology for Emergency Applications

James M. Thorpe

ABSTRACT Employing a blend of broadcast videoand state-of-the-art computer technology theManagement Emergency Response Information System(MERIS) is designed to control , manipulate, anddistr ibute the graphic and visual informationnecessary for decision-making 1n an emergencyresponse situation or exercise. Instant storageand recall of an extensive l ibrary of frames ofvideo imagery allow emergency planners the timeand freedom to examine necessary informationquickly and e f f i c i en t l y .

I . INTRODUCTION AND BACKGROUND

For a number of years, EGSG EnergyMeasurements, Inc. (EGSG/EM) has been providingthe Emergency Response and Aerial MeasuringSystem (AMS) programs with a field-deployabiephoto and video capabil i ty in support of tech-n ica l , management, and public af fairs functions.Imagery collected from both actual operations andmajor exercises has been a signif icant tool fortraining team members, f i e ld managers, and otheragency personnel. In response to a continuingneed to provide better information handling anddisplay methods, and fo r Improved reca l l anddisplay of a l l types of graphic data products,development of the Management Emergency ResponseInformation System (MERIS) was begun. MERISpresently consists of a number of components ofexisting video and computer equipment and state-of-the-art image handling systems.

In MERIS, images are Instantly stored andreca l led , e i ther one frame at a time or insequence. This c a p a b i l i t y , coupled wi th thenewest computer-controlled videodisc technology,w i l l make the EGSG/EM l ibrary of aerial photo-graphs, videotapes, and other imagery for a l lmajor nuclear sites available for instant recallfor cr is is management purposes or other urgentinquir ies.

Up to 10 Input stations can be used to feedinformation into the MERIS base station for imagehandling. The images are combined and t i t l e d ,have graphics added, then are time-coded and

stored in the MERIS l i b r a r y u n t i l needed fo rcr is is management. Four large monitors are usedto display the information to the cr is is manage-ment teams. The choice of information on themonitors is at the control of the Crisis Manager;data can be recalled within seconds of i t s input.

In emergency operations, MERIS can be usedto b r i e f important v i s i t o r s by reca l l i ng anddisplaying Information stored in the MERISl ibrary . Images can be displayed in any sequencethe manager requests. Hardcopy color prints ortransparencies and black and white prints canalso be made of any image stored 1n the MERISl ib rary .

MERIS concepts have already proved of inter-est to a number of emergency planners and othersjvho must store and recall large quantities ofSgnaphic data. MERIS 1s an operational f i e l d -portable system that could: be 'u t i l i zed forAccident Response Group (ARG) and FederalRadiological Monitoring and Assessment Center .(FRMAC) emergency operations and exercises. Someof the unique MERIS capabilit ies are currently,being u t i l i z e d to solve spec i f ic problems tn 'Emergency Operations Centers (EOCs) and cr is iscontrol centers.

I I . DESCRIPTION OF MERIS COMPONENTS

DIGITAL FRAME STORE - Stores frames of videoimagery within 2 seconds. Those frames can berecalled at random or in sequential order.

VIDEO SWITCHER - Sometimes called a specialeffects generator. Electronically keys or over-lays one video source (or Image) over another.Also allows for colorization of white overlays aswell as fades and dissolves among a l l videosources.

MULTI-LEVEL KEYER (Graham-Patten) - Works Inconjunction with the video switcher to overlay orkey up to six different sources or Images at onetime over a base image.

PAINT BOX - Electronic drawing computer usedas a source that is keyed or overlayed by videoswitcher and multi-level keyer over a base image.Used to produce backgrounds and animation also.

CHARACTER GENERATOR - A t i t l i n g and graphicbackground-generating computer keyed over the

225

226

227

base image via the video switcher.COPY STAND - A f l oo r stand unit with l igh ts

and a video camera on which graphic material canbe converted into a video source for manipulationand overlay.

DIGITAL FRAME STORE LIBRARY COMPUTER CONTROL- Controls the f i l e s of the frame store 's d i g i t a lmemory. Contains a cross-referencing recal lsystem to a l low quick access to any frame instorage.

3/4 UMATIC-VCR - Conventional news broadcastformat video recorder used to record or play backan event videotape in to the system. An i nd i v i d -ual frame can be stopped and stored at any t ime.

VIDEODISC PLAYER WITH ATXT COMPUTER - Allowsstorage and cont ro l le r recal l of an extensivel i b ra ry of photo and video frames (up to 54,000per disc s ide) . Organizes and houses a data basethat allows indiv idual frames to be categorizedand labeled.

RTS INTERCOM - Allows communication betweenMERIS team members and emergency response manage-ment personnel.

TIME BASE CORRECTOR/FRAME SYNCHRONIZER (4) -Distr ibutes composite images to each of the fourindividual channels.

VIDEO MONITOR (output) - Conventional NTSC(National Television System Committee).

I I I . HOW DOES MERIS PRODUCE AN IMAGE?

Since i t is based on broadcast video tech-nology MERIS can easi ly create a compositeoverlay of several pieces of graphic informationwi th in a very short t ime. For example, when anaerial photograph of a s i t e is placed under thevideo camera on the copy stand, that image can bestored in the d i g i t a l frame store with the pushof a button.

When r e c a l l e d , t ha t ae r i a l image can beoverlaid by any number of sources. A charactergenerator can overlay le t te rs for t i t l i n g , etc.The pa in t box can be ove r l a i d (v ia the videoswitcher) for de ta i l ing and graphics, plust i t l i n g . The Graham-Patten keyer, along with thevideo switcher, allows up to six d i f ferentsources (character generator, paint box, copycamera, photo, e tc . ) to be keyed or overlaid onthe or ig inal aer ial photo. The f ina l compositeimage w i l l usually have the time keyed over ( inthe lower r ight corner) as i t i s entered onto thes t i l l store d isks. This system can store up to400 frames of information on i t s b u i l t - i n harddisk and 200 each on a number of removable diskcart r idges.

In addi t ion to storage and recal l of i n d i -vidual frames of imagery, MERIS provides c r i s i smanagers with other valuable visual informationalassets.

Conventional o f f - a i r broadcasts (news, pressconferences, e t c . ) can be d i s t r i b u t e d on anychannel when necessary. These broadcasts canalso be recorded fo r la te r evaluation and analy-s i s . Video images recorded by emergency responseauthorized crews could be s imi lar ly played back.

Publ ic a f fa i rs br ief ings could also be con-ducted with MERIS imagery to give the s ign i f icantvisual deta i l necessary for media release.

Post-exercise or response evaluation couldalso be greatly enhanced with MERIS.

IV. FUTURE DEVELOPMENTS

Since broadcast video and computer t e c h -nology are constantly changing, and many of thosechanges w i l l have a d i rect effect on MERIS capa-b i l i t i e s , constant evaluation of new technologyin these areas is necessary. MERIS w i l l re f lec tnew ways of stor ing and recal l ing images and dataas they are developed.

Several areas are current ly being researchedfor potential MERIS appl icat ions. Fiber optictechnology offers numerous d i s t r i bu t i on optionswithout the signal degradation of conventionalcable when transmission over a long distance isneeded. This w i l l enable the MERIS remoteconsole to display images from the main system atv i r t u a l l y any point necessary.

This technology w i l l a lso a l low a videocamera t o be placed at a designated s i t e f o rmonitoring while i t s image is transmitted to themain terminal .

Technology advances in the storage andrecal l a b i l i t y of videodiscs w i l l also improvegreat ly in the coming years. This combined withthe in teract ive t ra in ing assets of videodisctechnology w i l l help expand MERIS and i t s simu-lat ion and cr is is scenario training capabil i t ies.

As MERIS is used in emergency response s i t u -ations and exercises, design engineers w i l l beable to f ine-tune i t s response capab i l i t ies tomeet changing demands.

This work was supported by the U. S.Department of Energy under Contract No.DE-AC08-83NV10282. NOTE: By acceptance of t h i sa r t i c l e , the publisher and/or recipientacknowledges the U. S. Government's r i g h t t oretain a nonexclusive, royal ty- f ree l icense toany copyright covering th i s paper.

ANS Topical Meetimg OH Radiological AccMemK—Perapectires and Emwrgemcy Plaaamag

Using MENU-TACT for Estimating Radionudide Releases Duringa Reactor Emergency

Andrea L. Sjoreen

ABSTRACT MENU-TACT i i t computer codedesigned to estimate potential atmosphericreleases of radionuclides froai tbe reactorcoaplex during a reactor emergency. It i s partof a suite of Models, tbe Intermediate DoseAssessment System, which is used at tbe U.S.Nnclear Regulatory Commission's OperationsCenter. MEND-TACT incorporates only thoseprocesses tbat can make major changes in tbemagnitude of a reactor release. It providesquick reaults of bounding calculations. MEND-TACT was written to be simple to use. Data areentered via fil1-in-the-blank displays orselection of menn options. Plant-specific andgeneric default values for data are providedwhere appropriate. The code models transport ofnuclides from containment through a series ofvolume* to the environment. The analyst controlsthe pathways from the volumes to the environment,timing of events, and depletion parameters, alongwith other required data.

I . INTRODUCTION

During a severe reactor accident the U.S.Nuclear Regulatory Commission (NRC) must f i rs tdecide if plant conditions warrant takingprotective actions. This in i t ia l decision isbased on consideration of the consequences of awide range of core damage accidents and i s notbased on dose assessements. Determination of theneed for additional protective actions includesconsideration of dose assessments. To producedose assessments, NRC must f irst estimate theprobable range of magnitudes of tbe release fromthe reactor complex. The basic tool used in theNRC's Operations Center to estimate these reactorsource terms i s MEND-TACT. (TACT is an acronymfor Transport of ACTivity.) MENU-TACT was createdby modifying TACT-III (til lough, et a l . . 1983).The Accident Evaluation Branch, NRC, developedTACT-III so that doses and releases from avariety of accidents could be estimated. MENU-TACT and TACT-III model transport of nuclides

from containment through a series of volumes(which are also called nodes) to the environment.The number of node*, the node volumes, and thepathways through the nodes and to the environmentare all entered by the analyst. MENU-TACT andTACT-III use matrix Inversion routines that solvesimultaneous equations describing transport,radioactive decay, and removal processes in thevolumes. Radionuclide daughter build-up i s notconsidered in MENU-TACT and should not beimportant fox the time period of concern whenresponding to an emergency ( e . g . , a few hours ordays). The mathematical framework of TACT-III i sdescribed inKil lough, et a l . (1983).Modifications made to TACT-III to produce MENU-TACT serve to simplify tbe input required and tosimplify the processes nodeled.

MENU-TACT i s part of the Intermediate DoseAssessment System (IDAS). IDAS also contains adata baae management system and MESORAD (Sherpelzet a l . , 1986), an air transport and doaeprojection model. The source terms computed byMENU-TACT can be atored in the IDAS data base.They are read by MESORAD according to the namegiven to that case in MENU-TACT.

This document explains the use of MENU-TACTduring an accident. The processes modeled aredescribed. Examples of the input scheme* usedare preaented and *ampl* result* are shown.Future enhanceasents to MENU-TACT are described.

I I . PROCESSES MODELED IN MEND-TACT

The processes included in MENU-TACT and thetreatment of reactor volumes are diagrammed inFig. 1. This figure shows two volumes,containment and an auxil l iary building.Transport processes include leakage fromcontainment, transport through the building,recirculation through f i l t e r s , and escape to theenvironment. Removal processes are radioactivedecay, f i l t er ing , and user-specified removalprocesses. In Fig. 1 the user-specified removalprooess i s spray removal. Filtering and removalapply only to non-noble gases. A single constanti s entered for each proceas for all radionuclidesaffected. While thia i s not a very sccuratetreatment of removal and f i l t er ing , i t i sappropriate for bounding calculations.

229

230

FRACTION OF CONTAINMENTOR

VOLUME OF COOLANT

FORCED FLOWS

RECIRCULATIONFILTERS

FILTERS

NODE 2

AUXILIARYBUILDING

SPRAYREMOVAL

ANDDECAY

LEAKAGETO THE

ENVIRONMENT

Figure 1. The processes modeled in MEND-TACT and their relationships tothe voluaes defined.

The tiaing of processes is controlled by thetiaes entered by the analyst as shown in Fig. 2.Radioactive decay operates fron reactor shutdownto end of release. (Note that decay continuesdaring atmospheric transport »t aodeled inHESORAD.) Daughter build-up is not included inMEND-TACT calculations, because it was determinedto be nninportant for bounding calculations. Theprocess start tine begins in-p]ant processes. Noactivity i s released to the environment beforethe start of release tiae. Note that shutdownand process start tiae aay be the saae. butrelease start aust be later, if only a fewseconds, than process start. The tiae of the endof the release aost be later than all othertiaes .

III . MODEL TESTING

MENU-TACT has been tested both for internalconsistency and against TACT-III for correctness.Etch type of input data was varied froa i t sainiBBE to i t s aaxiaun value, with a]] otherinpnt data kept at their default values. Theresults were inspected to insure that theexpected changes occured. These tests insuredthat MENU-TACT was aodeling i t s processescorrectly. More that 30 pairs of MEND-TACT /TACT-III runs were aade to insure that both codesproduced the saae resnlts. Four types of tests

were run. (1) A baseline case was run withdefault data. (2) One and two time-step caseswere run, with shutdown tiae both equal and notequal to process start t iae. (3) Two voluaecases were run with coabinations of large andsaall volnaes with large and saall flow rates.(Note that the aatrix solution routines used inTACT-III i s soaewhat aore robust than those inMEND-TACT in handling cases that were nuaericallyextreae.) (4) Large and saall f i l terefficiencies were coabined with large and saallreaoval coefficients. The TACT-III and MEND-TACTresults agreed to three significant figures forall tests .

IV. DSING MENU-TACT

A coaplete u s e r ' s aanual for MEND-TACT hasbeen published (McKenna e t a l . , in p r e s s ) . Thefol lowing descr ibes b r i e f l y the use of MEND-TACTin the Operations Center. The aajor s teps inperforming a MEND-TACT a n a l y s i s are:

1 . Preparing the re l ease pathway fora

2 . Logging onto the systea and branching in toMENU-TACT

3 . Selecting plant site (and case)

231

SHUTDOWN PROCESSSTART

START OFRELEASE

END OFRELEASE

ENVIRONMENTAL RELEASE

IN-PLANT TRANSFER AND FILTERING

ak.

DECAY OF FISSION PRODUCTS

Fi|ore 2. The relationships between the tiaes entered in WENU-TACT.

4. Specifying basic tiae dependant information:

a. case description or t i t l e

b. power level

c. node naaes and volumes

d. shutdown tiae

5. Specifying, core dsaage state, either:

a. coolant release

b. gap release

c. (rain boundary release

d. cote aclt release

6. Specifying pathway data:

a. event tiainf

b. reaoval coefficient for each node

c. portion of the release in i t i a l l yinjected into each node.

d. transfer rates between nodes andbetween nodes and the environment andf i l t er efficiencies assnaed fortransfers.

All data are entered on *rou screens (Fig. 3) oron fill~in-the-blank screens (Fig. 4 ) . A defaultvtlne i s provided for each data i tea . Ifappropriate, the defanlt valne i s plant-specific,that i s , the valne is read froa the reactor s i t edata bate. If the default valne i s appropriate,the analyst can jast type the NEWLINE key forthat data. Each type of data has boundingvalues. If the analyst enters an inappropriatevalne. am inforaative aeasage i t printed and thevalue aost be corrected before continuing. Atthe end of each data screen, the analyst i t askedto check the entries. If any are incorrect, theanalyst can return to the top of the screen toreenter the data. If a aistake i s noted on thescreen before i t i s finished, the analyst aayback up through the entriea with the ESCAPE key.Once all data are entered, the aenn la Fig. 3appears and the analyst aay run the case oraodlfy the data farther.

The data that are plant-specific areprovided as the default for that i tea . Once aplant aita (and unit, if appropriate) i sselected, the fact that i t is a pressarixed waterreactor (PffK) or boil ing water reactor (BWK) iaknown. FVK's are given two default voluaea (ornodes), priaary and containment. BWR't ere giventhree default voluaes, priaary, contaiaaent, andsecondary containaent. The analyst can changethe nuaber of voluaes, their s ize , and theirnaaes. The n t e d reactor power (Nwt) isprovided. The plant design leak rate andpredicted failure pressure are ahown on the

232

NRC - IDAS TACT MAIN MENUIDAS TACT - MENU VERSION

CREATE A NEW CASEMODIFY AN EXISTING CASERETURN TO IDAS MAIN MENU

USE ARROW KEYS FOR SELECTION AND (CR) WHEN SELECTION tS THE CORRECT ONE

Fifnre 3, An example of • MEND-TACT aean tcreen.

BASIC TIME INDEPENDENT DATA FOR PWR ARKANSAS 1BASIC CASE DESCRIPTION/CASE NAME:testNUMBER OF NODES (NODES <=4): 2NAME OF NODES VOLUME IN FT**3NODE (1 ) = Pr i tnary . 1 000E+07NODE(2) = Contain. .1V0OE+07

REACTOR POWER (MEGAWATTS THERMAL) : 256B.REACTOR SHUTDOWN TIME - 24 HOUR CLOCK SITE TIME(SDTIME) (<= TIME OF FJRST CASE : 86/08/04 14:00 (YY/MM/DD HH:MM)

ARK 1:REFERENCE VOLUMES <ft*«3)Primary system 9.7E+03Containment 1.9E+06

ARE YOU DONE WITH THIS SCREEN ? Y

Figure 4. An exaaple of a MENU-TACT data entry screen.

233

Isotope

1-1311-1321-133I-X341-135Kr-85Er-8SmKr-87Ki-88Xe-131mIe-133mXe-133Ie-13 5Xe-138Cs-134Cs-136Cs-137Te-129mTe-131«Te-132Sb-127Sb-129Sr-89Sr-90Sr-91Ho-99Ru-103Ru-106Ba-140Y-91La-140Ce-144Npj23 9

Decay1/hr

3.6E-33.0E-13.3E-27.9E-11.1E-l7.4E-61.5E-15.4E-12.4E-12.4E-31.3E-25.5E-37.6E-22.9E+03.8E-52.2E-32.6E-68.6E-42.3E-28.9E-37.5E-37.4E-25.7E-42.8E-67.3E-21.1E-27.3E-47.8E-52.3E-34.9E-41.7E-21.0E-41.2E-2

SourceCi/Mwt

2.8E+44.1E+45.9E+46.JE+45.5E+41.9E+27.9E+31.5E+42.2E+43.7E+22.1E+35.7E+41.2E+41.2E+42.3E+39.2E+22.OE+36.4E+24.1E+34.1E+43.5E+39.0E+33.4E+41.7E+33.SE+45.2 E+44.2E+41.1E+45.1E+43.7E+45.4E+43.0 E+45.5E+5

CoolantConcentrati on

PWRCi/gn

3E-O71E-074E-075E-082E-072E-071E-07«E-082E-07IE-072E-072E-054E-074E-083E-081E-082E-081E-093E-093E-08

00

4E-101E-117E-108E-085E-111E-112E-106E-112E-103E-H1E-09

nWRCi/jB

5P.-093E-082E-O85E-082E-08

000000000

3E-112E-117E-114E-111E-101E-11

00

IE-106E-124E-092E-092E-113E-124E-104E-13

03E-127E-09

<;«Pfraction

2E-22 E-22E-22E-22 E-23 E-23 E-23 E-23 E-23 E-23E-23 E-23 E-23 E-25 E-25F.-25 E-21E-41E-41E-41E-41E-4

00000000000

Grai nBoundaryfraction

5E-15E-15E-15E-15E-1SE-15E-15E-15E-5E-5E-5E-5E-5E-]5E-15E-]5E-]1E-11E-11E-1

1

2 E-22 E-21E-31E-31E-31E-21E-41E-41E-2

0000

CoreMel t

fraction

1E+01E+01E>O1E+01E+01E+01E+01E+01E+01E+01E+01E+01E+01E+01E+01E+01E+03 E-l3 E- l3E-12 E-22E-27E-27 E-27 E-21E-17E-37E-32E-11E-41E-4I E - *1E-4

Table 1. Nuclide-specif ic data ntcd In MEND-TACT.

bottom of the screen on which the rate* oftransfer to the environment are entered.

The amount of the reactor inventoryavailable for release i s determined by ooredamage state and either plant power level and thenumber of Caries of each nnclide per Mwt or thenumber of Caries of each nnclide per ft ofcoolant. The nuclidet vhich aay be included in anassessatnt are shown in Table 1. The fractionsof the inventory released for each nuclide areavailable for a gap release , grain-boundary(TMI-Iike) re lease , or core Bel t . The inventory•ay be scaled up or down. One factor applies toal l noble gases and another applies to a l l othernacl ides . Nnclide speci f ic scaling i s notavailable for the above three release type*.Coolant concentrations are provided for bothPWK's and BWK's. For a re]ease.of coolant theanalyst enters the nuaber of f t of coolantreleased frost the reactor. As the accidentprogresses and better estimates of the coolantconcentrations are avai lable , the analyst has theoption of revising the default coolant

concentrations ( C i / f t ) for each nnclideindividual ly .

The transfer frpai the core are specified asa total ntmber of ft or a rate of f t / s i n forcoolant releases and as fractions of inventory orfract ions /a in of inventory for a l l other accidenttypes.

V. MENU-TACT RESULTS

MENU-TACT f i r s t prints a simaary of the datafor this case (Fig. 5 ) . It always displays thei n i t i a l inventories in the voluaes, theconcentrations in the nodes at the beginning ofthe release to the environment (F ig .6 ) , and thetotal releases (CD to the environment (Fig. 7)for each nnclide considered for the time periodof the release . More detailed results areavai lable , i f requested. If the results areneeded for father assessment in NESOKAD, a f i l ei s written to the IDAS data base.

Site

234

is:ARKANSAS 1 Record is:TIME INDEPENDENT INPUT

ARK1 026

RUN DA IE 8/ 5/ 86TITLF.: testNUMBER OF NODES:POWER <MWT):TIME OF SHUTDOWN:RELEASE FRACTION:

PARTICULATES:NOBLE GASES:

SOURCE OPTION: .PLANT OPTION:

22.568E+0386/08/04 14:00

l.OOOE+00l.OOOE+00

4-CORE1-PWR

Node names:Pr imary Conta i n.

TIME DEPENDENT INPUTPROCESS START TIME END OF RELEASE

86/0P/04 14:00 86/08/04 14:10VOLUMES

NODES VOLUME <FT*»3>Pr imary 1.000E+06C o n t a i n . 1.900E+06REMOVAL COEFFICIENTS (1/HR)

NODE PARTICULATES NOBLE GASESPr imary 3.50 O.OOOE-01Conta in . O.OOOE-01 O.OOOE-01

Press Return fo r Transfer Rates. =

Figure S. Data summary.

ACTIVITY (CURIES) IN EACH NODE AT 86/08/04 14:00

ISOTOPE1-1311-1321-1331-1341-135KR-85KR-85MKR-87KR-8BXE-131MXE-133MXE-133XE-135XE-138CS-134CS-136CS-137TE-129MTE-131MTE-132SB-127SB-129SR-89SR-90SR-91MO-99RU-103RU-106BA-140Y-91LA-140CE-144NP-239

(BEGINNING OF PROCESS)ENVIR. Primary

O.OOOE 01 5.752E+07 1.O.OOOE-01 8.423E+07 2.O.OOOE-01 1.212E+08 3.O.OOOE-01 1.335E+08 3.O.OOOE 01 1.130E+08 2.O.OOOE-01 3.903E+05 9.O.OOOE-01 1.623E+07 4.O.OOOE-01 3.082E+07 7.O.OOOE 01 4.520E+07 1 .O.OOOE-01 7.601E+05 1 .O.OOOE-01 4.314E+06 1 .O.OOOE-01 1.171E+08 2.O.OOOE-01 2.465E+07 6.O.OOOE-01 2.465E+07 6.O.OOOE-01 4.725E+06 1 .O.OOOE-01 1.B90E+06 4 .O.OOOE-01 4.109E+06 1 .O.OOOE-01 3.944E+05 9.O.OOOE-01 2.527E+06 6.O.OOOE-01 2.527E+07 b.O.OOOE-01 1.438E+05 3.O.OOOE-01 3.698E+05O.OOOE-01 4.B89E+06 1,O.OOOE-01 2.445E+05 6.O.OOOE-01 5.177E+06 1,O.OOOE-01 1.068E+07 2,O.OOOE-01 fe.040E+05 1,O.OOOE-01 1.582E+05 3,O.OOOE-01 2.095E+07 5,O.OOOE-01 7.601E+03 1,O.OOOE-01 1.109E+04 2,O.OOOE-01 6.163E+03 1,O.OOOE-01 1.130E+05 2,

Conta i n .438E+0/106E+07030E+07338E+07B25E+0775BE+04057E+06704E+06130E+07900E+05079E+06928E+07163E+0-6163E+06181E+06725E+05027E+06861E+04317E+O5317E+06595E+04245E+04.222E+06112E+04294E+06671E+06510E+05955E+04239E+06900E+03773E+03541E+03825E+04

Press Return to continue

Figore 6. Initial concentrations computed for the data in Fig. 5.

235

ACTIVITYISOTOPE

1-1311-1321-1331-1341-135KR-85KR-8SMKR-87KR-88XE-131MXE-133MXE-133XE-135XE-138CS-134CS-136CS-137TE-129MTE-131MTE-132SB-127SB-129SR-89SR-90SR-91MO-99RU-103RU-106BA-140Y-91LA-140CE-144NP-239

M i ght youY :

RELEASED TO THE ENVIRONMENT FROM 86/0B/04 14:00 TO 86/08/04RELEASEDCURIES1.192E+011.705E+012.506E+012.600E+012.323E+019.039E+003.710E+026.B22E+021.026E+031.760E+019.980E+012.711E+035.673E+024.512E+029.797E-013.91BE-018.S19E-018.178E-025.230E-015.236E+002.980E-027.622E-021.014E+005.069C-021.067E+002.213E+001.252E-013.280E-024.344E+001.576E-032.297E-031.278E-032.340E-02

want to use these results in MESO RAD? =

Figure 7. Release* to the environment computed froa the data in Fig. 5.

VI. PLANNED ENHANCEMENTS

Several enhancements are planned for MENU-TACT. Current ly , • simpler model i t beingcons tructed which requires only the WASH-1400(USNRC, 1975) source t e n s , reactor power, andevent t i a i n g . In the fu ture , another aodel w i l lbe added that w i l l r e l a t e plant c o n d i t i o n s tosource t e n s . Also, the f i l l - in-the blankscreens will be replaced by graphic data entryscreens which will show the confignation of theselected voluaes and will allow data entry on thevolumes and pathways drawn on the screen.

VII. REFERENCES

til lough, G. G., et al., 1983, A Guide to JheTACT III Coaonter Code. NDREG/CR-3287, ORNL/TM-8763, Oak Ridge National Laboratory.

McKenna, T. J., et al., in press. Source TernEstimation Dsin« MEND-TACT. NUREG/CR-4722, ORNL-6314, Oak Ridge National Laboratory.

Scherpelz, R. I., et al., 1986, Tb_s MESORAD DoseAssessment Model Voj. 1: Technical Basis.NUREG/CR-4000, PNL-5219, Pacific NorthwestLaboratories,

U. S. Nnclear Regulatory Coaaission, 197S,Reactor Safety Study - An Assessment of AccidentEllkl is D«S. Coaaercial Nuclear Power Plants.WASH-1400, NCREG-75/014.

Research sponsored by the Office of Inspectionand Enforcement, U. S. Nuclear RegulatoryCoaaission, ffashinton, DC 20555, underInteragency Agreement 40 546 75 with MartinMarietta Energy Systems, Inc. contract DE-AC05-84OR2140O with the D.S. Department of Energy.

ANS Topical Meeting on Radiological Accidents—Perspectives and Emergency Planning

The MESORAD Dose Assessment Modela

R. I. Scherpelz, J. V. Rams del I, G. F. Athey, and T. J. Bander

ABSTRACT MESORAD is a dose assessment codedeveloped for use as an emergency response toolby the U.S. Nuclear Regulatory Commission andthe U.S. Department of Energy. The codecombines atmospheric dispersion and dosecalculations to make dose estimates forreleases containing as many as 50 radio-nuclides. It evaluates those dose pathwaysthat are likely to be important during andimmediately following a release. The pathwaysincluded are the internal dose commitmentthat is due to inhalation of radionuclidesand the external dose that is caused byexposure to radionuclides in the air anddeposited on the ground. External doses arecomputed using either the semi-infinite cloudapproximation or the finite puff model. Thefinite puff model was developed to reducecomputational time and is an approximation tothe point-kernel integration technique foundin other dose models. Initialization ofMESORAD requires about 5 minutes, and simula-tion of the first 3 hours of a release canbe completed in about an additional 3 minuteson a super-minicomputer.

I. INTRODUCTION

MESORAD is a computer code that performsdose assessments for emergency responseapplications. The code was developed at thePwific Northwest Laboratory for use at theU.S. Nuclear Regulatory Commission's OperationsCenter in Washington, D.C. and at the U.S.Department of Energy's Unified Dose AssessmentCenter at Hanford, Washington. The code isused to assess the consequences prior to,during, and immediately following potentialor actual releases of radioactive material.The dose pathways represented in the code arethe external doses that are due to airborneand deposited radionuclides and the internal

(a)Work supported by the U.S. Department ofEnergy and the U.S. Nuclear RegulatoryCommission under Contract DE-AC06-76RLO1330, NRC FIN P2001.

dose commitments that are due to inhalationof radionuclides.

II. ATMOSPHERIC DISPERSION MODELS

In MESORAD, atmospheric dispersion modelsare used to estimate radionuclide concentra-tions at nodes on two receptor grids. Theprimary grid Is a Cartesian grid in which thespacing between receptors is typically severalkilometers. The second grid is a polar gridwith receptor nodes at 10-degree intervals thatare 0.8, 1.6, and 3.2 kilometers from therelease point. This polar grid provides abetter resolution of concentration and dosepattarns near the release point.

The primary dispersion model in MESORADis an extension of the Lagrangian puff modelin the MESOI computer code (Ramsdell et al.,1983). The model is used to estimate theconcentrations at the nodes on the Cartesiangrid and may also contribute to the concentra-tion estimates at nodes on the polar grid iftha wind direction reverses. A straight-lineGaussian model is used for the primary concen-tration estimates at polar-grid receptors.Surface contamination resulting from dry andwet deposition is estimated for all receptorlocations on both grids.

III. DOSE MODELS

MESORAD assesses the radiologicalconsequences of a release using three dosepathways: the internal dose commitment thatresults from the inhalation of radionuclides,che external dose that results from exposureto airborne radionuclides, and the externaldose that results from exposure to radio-nuclides deposited on the ground. Doseestimates are computed for all receptors andare available for the most recent 15-minuteperiod as well as the entire period since thebeginning of the release.

The dose estimates can be based on as manyas 50 radionuclides. Decay schemes have beenincluded in the code,, where appropriate, sothat production of progeny nuclides by parentdecay is represented. Radioactive decay andproduction are calculated during "holdup" in

237

238

the facility (the elapsed time between thespecification of the release inventory and thetime of release), during transit from therelease point to the receptor, and duringresidence on the ground (for deposited radio-nuclides).

A. Internal Dose PathwayFor the evaluation of the internal

dose pathway, 50-year inhalation dose commit-ments are calculated to three organs: wholebody, lung, and thyroid. The calculation ofinhalation dose at each receptor point usesthe ground-level air concentration multipliedby a breathing rate and a dose factor. Thecomputational procedure is described byScherpelz et al. (1986). The inhalation doseto thyroid is determined for the child-agegroup, which is generally considered to be mostsensitive to thyroid doses. The doses to theother two organs are calculated for the adult-age group.

B. External Dose Resulting from Exposureto Airborne RadionuclidesTwo models are used in MESORAD to

treat the external dose pathway of exposureto airborne radionuclides, a semi-infinitecloud model and a finite puff model. Thesemi-infinite cloud model uses a simplecalculation, multiplying a dose factor by theair concentrations. The code executes quickly,but provides poor estimates of the dose incases where the radionuclides are elevatedabove the ground, or in the cases where thespread of the radionuclides is small comparedto the range of emitted gammas. For a betterdose estimate in these cases, a finite puffcalculation using a discrete-point approxima-tion is utilized. This discrete-point approxi-mation requires more computation time than thesemi-infinite cloud calculation; the accuracyof the calculation deteriorates as the spreadof the radionuclides grows very large. Thus,the two models are complementary; the finitepuff model offers an accurate dose evaluationwhen the spread is small, and the faster-executing semi-infinite cloud model can givea good estimate when the spread of the radio-nuclides increases.

The finite puff model used in MESORADfor the calculation of external doses is asimpli-ficatlon of the point-kernel integrationtechnique. In the point-kernel technique, thesource volume (the puff) is divided into anumber of differential volumes, and the doseat the receptor point that results from eachdifferential volume is evaluated. The totaldose equals the sum of the contributions ofall the differential volumes. The quality ofthe computation depends on choosing the propernumber of differential volumes. This selectionis done iteratively, with increasingly finerdivisions of the source volume. This techniquealso uses large amounts of computer time andis impractical for use in emergency responseapplication.

The discrete-point approximationused in MESORAD is a simplification of thf

point-kernel integration technique in whichthe puffs that are used to approximate theradionuclide plume are immediately dividedinto a number of differential volumes. Theiterative nature of a point-kernel integrationis thus avoided, and the calculation proceedsmuch faster. The puffs are divided intodifferential volume elements using cylindricalcoordinates. Each differential volume ofthe puff is treated as though all the radio-nuclides were located at a point in the centerof the small voluma. The dose to a receptorpoint is then calculated as the sum of thedoses from all source points.

The discrete-point approximation wastested by comparing calculations made usinga preliminary code called DISCRTPT againsthand calculations of simple configurations andagainst calculations made usir.g PERCS (Reeceet al., 1984) and ISOSHLD (Engel et al., 1966),point-kernel codes that are used in the radia-tion shielding community. In these tests, thepuffs were assumed to be cylinders containingair with radionuclides uniformly distributedin them, because the shielding codes neededto assume a uniform distribution in the sourceregion. The comparisons showed that theDISCRTPT code and other methods were in generalagreement within +20Z when the puffs weredivided into 576 differential volume elements(8 angular divisions, 8 radial divisions, and9 height divisions). Differences in dose esti-mates from different codes in the range of 20Zare not unreasonable, especially when the codesuse different expressions for buildup anddifferent attenuation coefficients.

The performance of the discrete-pointapproximation was evaluated for use in MESORADby varying the number of volume elements inthe puff in the DISCRTPT code and comparingthe dose estimates with the dose estimatescomputed from 576 puff elements. The resultsof this process were used to develop the rulesthat determine the number of volume elementsinto which puffs are divided. These rules,which represent a balance between the computa-tional time constraints of emergency responseand accuracy in dose estimates, are based onthe magnitude of the characteristic horizontaldimension of the puff (diffusion coefficient)and the distance from the center of the puffto the receptor.

C. External Dose Resulting from Exposureto Contaminated GroundThe current dose rate and cumulative

dose resulting from exposure to contaminatedground are computed at 15-minute intervals.These computations include the contributionsof all nuclides deposited on the ground fromthe start of the simulation to the currenttime. The dose rate calculation essentiallyuses the product of the nuclide concentrationson the ground (corrected for radioactive decayand buildin while on the ground) and a dose-rate factor. The cumulative dose is a runningsummation of dose rate multiplied by durationof the dose rate (usually 15 minutes). Thecumulative dose, therefore, indicates the dose

239

that an unsheltered person standing at thereceptor location for the entire duration ofthe scenario would receive.

IV. COMPUTER CODE EXECUTION AND PERFORMANCE

The performance of MESORAD was evaluatedon a VAX-11/780 computer with Version 4.2 ofthe VMS operating system. The evaluation runswere made during normal computer operationperiods using the system clock to determinethe time required for various simulations.Both real time and central processing unit(CPU) usage were recorded to the closest 0.01second.

All simulations were based on a commonscenario, a simulation of a release followinga reactor accident. The scenario assumed thatthe data files required for MESORAD wereavailable when execution of the code began.This assumption corresponds to the situationwhere meteorological data are automaticallyentered Into a data file and radionuclide fileshave been prepared for various postulated acci-dents. The data bases used in this evaluationinclude meteorological data for 9 stations andrelease data for 27 radionuclides. The periodssimulated extended from 6 to 72 hours with amaximum release duration of 24 hours.

The real time required for interactiveinitialization of a research version of thecode ranged from about 3.5 to almost 7 minutes.During this period, the operator specified thesize of the model domain, selected from amongthe model options, established the sourcecharacteristics, and entered names for the datafiles. The length of the initialization perioddepends to some extent upon the model optionsselected. For example, the average initializa-tion for runs that included dose computationstook about 5 minutes, while the average ini-tialization time for runs that only madedispersion and deposition computations tookabout 4 minutes. The CPU time for initializa-tion was about 14 seconds for the runs withdose computations and about 9 seconds for theruns without dose computations. By carefulselection of sets of default options, it shouldbe possible to develop site-specific opera-tional versions of MESORAD in which theinitialization time is significantly reduced.

The time required for model computationsfollowing initialization varied significantlyfrom run to run, depending primarily on numberof other users on the computer. However, theCPU time required to reach a given point inthe simulation showed much less variation, andthis variation is easily ascribed to differ-ences in the model options selected.

Assuming a release at about midnight inlow winds and stable conditions, the fullMESORAD code took 156 seconds of CPU time(following initialization) to complete thefirst 3 hours of simulation, and 661 secondsto complete 6 hours. Repeating the simulationwithout the polar grid, the CPU times werereduced to 120 and 563 seconds, respectively.The CPU times were not reduced significantly

(2-second difference at the end of 6 hours)when the same simulation was run a third time,eliminating the polar grid and the trackingpuffs after they left the dose computationgrid. For comparison, the MESOI dispersioncode (Ramsdell et al., 1983), run within theMESORAD shell, took about 2 seconds to complete3 hours of simulation and about 6 seconds tocomplete 6 hours of simulation.

During these simulations, the CPU timeper hour of simulation increased initially asthe number of puffs on the computational gridIncreased. However, the time did not reachan equilibrium value as might have beenexpected, instead it entered a diurnal cyclethat appears to be related to wind speed andstability. In nighttime low wind speed, andstable atmospheric conditions, the CPU timeper simulation hour reached a maximum ofabout 180 seconds. When the wind speed pickedup and the stability decreased during theday, the CPU time required for each hourdropped to a minimum of about 70 seconds.This variation is not reflected in the totalCPU time required to complete a 3-hour simula-tion because the effects of the increasingnumber of puffs is dominant. However, it isapparent in the time required to complete6-hour simulations. Completion of 6-hoursimulations starting at about midnight, 6:00a.m., and noon took 563, 380, and 311 CPU-seconds, respectively.

Grid spacing and release height alsoappear to affect computational time, but theeffects of these sources of variation have notbeen explored in detail. It is anticipatedthat the general effect of decreasing gridspacing is to Increase computation time becausemore computations are required per puff.Similarly, an increase In release height isexpected to be accompanied by a decrease incomputation time, because wind speeds generallyincrease with height, reducing the time thatreleased material is within MESORAD*s dosecomputation domain.

Recently, the performance of the MESORADcode has been evaluated using more detailedanalysis routines. These routines identifiedthe computational time used by various portionsof the code. The programming has been opti-mized in several of these areas resulting ina 30% decrease in MESORAD execution time duringthe simulations that were timed. Further codeoptimization is expected.

V. SUMMARY

MESORAD Is a dose assessment code devel-oped for use by the U.S. Nuclear RegulatoryCommission and U.S. Department of Energy inemergency response applications. The code'spurpose is to estimate doses during and immedi-ately following an accidental release of radlo-nuclides. Therefore, the dose pathways treatedby the code are internal dose commitmentbecause of inhalation of radionuclides andexternal dose because of exposure to airborneradionuclides and radionuclides deposited on

240

the ground. Doses accumulated during the mostrecent 15-minute period and since the beginningof the release are available for use.

Initialization of the code can be com-pleted In about 5 minutes. Simulation of theinitial 3 hours of a release can be completedin an additional 3 minutes, and simulation ofthe first 6 hours of a release can be completedin 5 to 10 minutes. Further optimization ofthe code can reduce these times.

REFERENCES

Engel, R. I,., et al., 1966, ISOSHLD - AComputer Code for General Purpose IsotopeShielding Analysis, BNWL-236, Pacific NorthwestLaboratory, Richland, Washington.

Ramsdell, J. V., et al., 1983, MESOI Version2.0; An Interactive Mesoscale Lagrangian PuffDispersion Model with Deposition and Decay.NUREG/CR-3344, U.S. Nuclear Regulatory Commis-sion, Washington, D.C.

Reece, W. D., et al., 1984, Personnel Exposurefrom Right Cylindrical Sources (PERCS).NUREG/CR-3573, U.S. Nuclear Regulatory Commis-sion, Washington, D.C.

Scherpelz, R. I., et al., 1986, The MESORADDose Assessment Model. NUREG/CR-4000 Vol.1, U.S. Nuclear Regulatory Commission,Washington, D.C.

ANS Topical Meeting on Radiological Accidents—Perspectives and Emergency Planning

Accident Analysis Codes That Predict the Transport of RadiologicalAerosol Through a Fuel Cycle Facility as a Result of an Accident

Kevin C. Greenaugh

ABSTRACT. Computer codes are presently beingdeveloped which predict the gas-dynamics,material transport, and heat-transfer inducedfrom fire, tornado and explosion accidents innuclear fuel cycle facilities. These codesconsist of models which describe the majorphenomena which occur to 0.5 to 5.0 micronparticles as they are released in a room, andtransported by way of the ventilation systemthroughout the nuclear fuel cycle facility.

I. ACCIDENT ANALYSIS CODES

Assessment of the environmentalconsequences of an accident in a fuel cyclefacility ultimately Involves calculating theatmospheric dispersion of radioactive materials,and estimating the radiation dose to thesurrounding population. A radiation dose calcu-lation of this nature requires consideration ofthe following parameters: (1) wind speed;(2) diffusion stability; (3) dose conversion;(4) type of release; and (5) source amount atthe facility boundary. Studies have beenperformed to develop an understanding of thephenomena associated with each parameter.

Considerable uncertainty lies in estimatingthe nuclear facility source terra. Codes havebeen developed which are comprised of mathe-matical models that describe the phenomenaoccurring during a reactor accident (e.g, HARM,TRAP-MELT), but conditions following a fuelcycle facility accident are typically different.The reactor models provide a theoretical estima-tion of a source release at the reactor contain-ment boundary of radioactive material as aresult of a postulated accident.

Los Alamos National Laboratory (LANL) andPacific Northwest Laboratory (PNL) under thedirection and support of the Nuclear RegulatoryCommission (NRC) have been developing a familyof accident codes which can be used to analyzetornado, explosion, and fire accidents innuclear fuel cycle facilities. The developedcodes are called TORAC, EXPAC and FIRAC.

TORAC models the gas-dynamics and materialflows (radioactive partlculate) in a facility

and their perturbations due to a tornado outsideof the facility. Negative pressures areassociated with a tornado. TORAC is used tomodel backflows throughout the facility due tonegative pressures at the facility boundary. Apressure time function at the facility boundaryis used to simulate a tornado.

liXPAC is the computer code used to studyexplosion transients. The source or amount ofenergy and particulate released as a result ofan explosion are determined by a chamber(compartment) model within EXPAC called NORDAL.Explosions can also be simulated by usingpressure, temperature, or energy time functionsalong with mass (gas) and material injectiontime functions. EXPAC predicts the gas-dynamicsand material flows resulting from an explosionwithin a facility.

FIRAC is the computer code used to studyfire transients within a fuel cycle facility.FIRAC predicts material transport (both radio-active particulate and combustion products) andthe gas dynamics throughout the facility. Itcan also predict the heat-transfer within thefacility's duct work.

In addition to time functions, fires canalso be modeled by an extensive compartmentmodeling subroutine called FIRIN. An explana-tion of FIRIN and the explosion compartmentmodel NORDAL and their features will bediscussed in the next section of this paper.

A general description of the gas-dynamics,heat-transfer, and material transport modelsincorporated in the network aspect of each ofthe codes will be diseased in a later section.

It. COMPARTMENT MODELS

Compartment models were developed todescribe the amount of energy and particulatereleased in a room due to fires and explosions.These compartment models were incorporated intoFIRAC and EXPAC, respectively, and are used todetermine fire and explosion source terms.

A. Fire Compartment ModelPacific Northwest Laboratory developed

the fire compartment model FIRIN (Owczarski,1985). FIRIN was developed to estimate thesource release of smoke and radioactive

241

242

particles from potential room fires in nuclearfuel cycle facilities. The technical bases ofFIRIN consist of a nonradioactive fire sourceterm model, compartment effects modeling, andradioactive source term models.

1. Fire source term models. Sourceterms required to characterize a fire includemass loss rates, heat release rates, and combus-tion product generation rates. The mass lossrate (burn rate) In FIRIN is the net heat fluxreceived by the material in the fire (aftertaking into account the flame heat flux and theradiation loss) divided by the heat required toproduce a unit mass of :orabustible vapors. Theheat release rate in FIRIN is the product of theheat of combustion and the mass loss rate (acorrection for incomplete burning is performed).Lastly, the rate of producing combustionproducts is obtained in FIRIN by multiplying thefractional yield of each product upon combustionby the mass loss rate.

Two burn modes are also considered inFIRIN. They are combustion and smoldering.Combustion is assumed to occur when the oxygenconcentration in the room exceeds 15% and highventilation exists. Smoldering is assumed tooccur when the room has less than 11% oxygen andis under-ventilated.

2. Compartment effects. The severityof a fire inside an enclosure can be affected bycompartment constraints and barriers.Compartment barriers tend to contain and retainmost of the gases, mass, and thermal energyreleased during the combustion process. Modelsare incorporated in FIRIN which describe howeach of these phenomena Interact within thecompartment.

Three major types of heat transfermechanisms are considered in FIRIN - radiative,convective, and conductive. Radiative heattransfer is modeled as occurring from thecombustion flame, and the smoke layer to thewalls, ceiling, and equipment. In modeling thefire within the compartment, a two-layer regime,is assumed: a hot layer (smoke) on top and acold layer which contains contaminant free air.

Convective heat is assumed to be carriedoff by fire gases into the ceiling hot layer.Also, whenever a solid body is exposed to themoving hot layer having a temperature differentfrom that of the body, heat is transferredbetween the fluid and solids according toNewton's law of cooling.

Compartment floors and ceilings aretemporary heat sinks for energy generated fromburning materials. A simple one-dlraenslonalform of Fourier's heat conduction equation isused to model the thermal conduction rate withincompartment solid media. The conductionequation is written as:

32T

where:

_ . Ill' k a 3t

k " thermal conductivitya » thermal diffuslvlty (k/pcp)q, = internal energy source term

(1)

The equation is solved by a finite-difference technique with convection boundaryconditions. The Fourier heat conductionequation is also used to describe wall heatconduction with the following boundarycondition:

fc ax (2)

where Q is the heat flux from the flamedivided by the wall surface area times At forthe wall in the cold layer. The heat fluxboundary condition to the wall in the hot layermust consider radiative and convective heatrates to the wall.

An overall heat balance in the firecompartment is performed to calculate the hotlayer temperature. A simplified form of theinternal energy and enthalpy equation Is formu-lated to obtain the hot layer temperature fromthe principle of conservation of energy, and theassumption that potential and kinetic-energyterms are negligible.

Models for species inventories (carbondioxide, carbon monoxide, water, HCL, nitrogen,oxygen, methane) of both hot and cold layers areIncorporated in FIRIN. Assumptions are made inthe mass balance calculations that no mixingtakes place between layers, and only air isfound in the cold layer. It Is further assumedthat mixing within each layer Is uniform. Theideal gas law is employed to obtain total massand molar information.

The three major mechanisms of fluid motionconsidered In FIRIN are flows caused by fireplume entrainment, ventilation, and additionalflow paths. Each of these mechanisms cantransport radioactive particles out of thecompartment. Mechanisms which contribute to theradioactive particle depletion modeled in FIRINare sedimentation, Brownian diffusion, dlffuslo-phoresis, and thermophoresls.

Radioactive source term models are builtinto the FIRIN code in the form of subroutinesallowing estimation of the mass rate and sizedistribution of radioactive particles becomingairborne in the event of a fire. These modelsare based on experimental data obtained fromburning contaminated combustibles such as rags,gloves, and plastic bags in gloveboxes orsolvents, extraction fluids, and cleaning fluidsused in fuel cycle operations.

FIRIN Is designed to provide mass andenergy input (for the fire room) to FIRAC, whichmodels mass, energy, and heat transfer throughthe ventilation system of a nuclear facility.

B. Explosion Chamber ModelLos Alamos National Laboratory

developed an explosion model called NORDAL thatwas incorporated into the EXPAC computer code.The NORDAL subroutine is capable of providingthe total energy released from TNT, red oil,hydrogen and acetylene explosions in the form oftime functions. Parameters needed to modelexplosions from these materials are the totalmaterial mass, the material type, and thecharacteristic length of the room where theexplosion occurs. The characteristic length is

243

assumed Co be the cube root of the room volume.Tlie characteristic length along with the speedof sound is used to determine the characteristictime*

Pressure pulses occur with most explosions.After 4-8 pressure pulses, the room experiencesan over-pressurization. The characteristic timeis defined as the time factor after which a roomover—pressurizes and a peak energy releaseoccurs. The total energy released by theexplosion caused by the chemical exothermicreaction between oxygen and the explodingmaterial is obtained by integrating the areaunder the energy time function curve. This timefunction is used by EXPAC to describe theexplosion. The gas-dynamic and materialtransport affects of the explosion on adjacentrooms is then determined by the network modelingaspect of the EXPAC code.

III. MODELING NUCLEAR FACILITIES

The source terras determined by c~apartmentmodels or time functions are used to describethe conditions In the near-field (close to thetransient). Numerous mechanisms occur whichresult in a difference between conditions in thenear-field and conditions in the far-t^eld (atthe facility boundary). For example, particu-late deposition will cause a difference in theamount of radioactive participate released closeto the source and the amount that reaches thefacility's boundary.

For each of the codes, the ventilationsystem is assumed to be the primary path forparticulate, heat, and gas transport to thenuclear facility's atmospheric boundary.Therefore, mathematical models and a networkmodeling approach is used to describe materialand gas flows through a complex network ofrooms, gloveboxes, duct work, filtrationsystems, and other components typically found ina nuclear fuel cycle facility ventilationsystem.

A. Gas-Dynamics and Network ModelingCharacteristically, nuclear facility

ventilation systems are essentially one-dimensional with many branching and loopingducts. Therefore, a lumped parameter of eachsystem element has been used as the methodologyto simulate a ventilation system. This methodis sometimes referred to as network modeling andconsists of dividing a ventilation system into anumber of system elements called branches,joined at certain points called nodes. Spatialdistributions are not considered in this lumped-parameter approach. Ventilation components thatexhibit resistance, such as dampers, filters, orblowers, are located within system branches.Mass flow rates across branches are calculatedfrom the following momentum balance equation:

dt A <• 2- F + pAAZg

- A [(p2u2) V2 - (plUl) V j

where: p = gas densityP2 = pressure downstream

(3)

P. = pressure upstreamF = frictiong = gravitational acceleration

Vj = area out^2 = area In

Here the terra on the left represents the totalmomentum. The first two terms on the rightrepresent the pressure force and the frictionforce, respectively. The third terra on theright represents the gravitational force, andthe last terra represents the rate of momentuminflux and efflux by virtue of bulk motion. Thelast terra in equation (3) is assumed to be zerodenoting equal momentum flows into and out ofthe branch.

Frictlonal forces or resistances aremodeled through the application of a power lawrelationship between force and volumetric flowrates. For example, the resistance through aduct is modeled as being proportional to thevolumetric flow rate squared. The resistancecoefficient can either be an input parameterthat remains constant during a given simulation,or a calculated value based on the relationshipbetween force and flow rate.

The connection points (nodes) of a systemelement are at upstream and downstream ends ofbranches. Components that have larger volumes,such as rooms, gloveboxes, and plenums, arelocated at nodal points. Therefore, a node maypossess some volume or capacitance where fluidstorage or compressibility may be accounted.The appropriate mass flow rates which are deter-mined upstream and downstream of each node areused along with node energy and mass balances todetermine node densities and pressures. Theenergy and mass equations used in the nodecalculation are the following:

Energy Balance

^ C > f V P i T i > ijo^-P.ul) + f-{pfhf-Plhi)

1 A+ §gc c

" S2P2U2 l / 2 g " fc

i n i e t

o u t l e t

(4)

where: Q = heat addedo) = workIf » downstream branch gas temp.T^ = upstream branch gas temp.Sj =• upstream branch flow area&2 = downstream branch flow area

Accumulation of kinetic and potential energy areassumed to be negligible for node energybalances.

Mass Balance

At (P."

where: downstream branch densityupstream branch density

(5)

244

UJ, =• upscream branch gas velocityuf • downstream branch gas velocityM =• injected mass

Once the node pressure and density isdetermined, the equation of state is used todetermine the node temperature.

Consequently, the unknowns that appear inthe equations describing system behavior are themass flow rates for each branch and the tempera-ture and pressure at each node. These unknownsappear in the LANL codes as a potentially largeset of simultaneous equations. Solutions tothese equations are obtained by a Newton-Ralphson procedure, and decomposition of thelarge set of equations into independent sets ofsmaller numbers of equations. The calculatednode and branch gas-dynamics are used in many ofthe material transport models presented in thefollowing sections.

Decoupling of the equations is accomplishedthrough omission, in the linearize energy andmass balance for a given room, of the termswhich contain coefficients of variation withrespect to the density and pressure of air inadjoining rooms.

B. Material TransportThe material transport models

incorporated in the network aspect of FIRACdescribe the types of phenomena that affectaccident induced material transport of SpecialNuclear Material (SNM) as it is transportedthrough a ventilation system. These phenomenainclude convection, deposition and entrainment.Each of these models assume that the SNM is inthe form of solid powders with a size range of0.5 to 5.5 microns. Condensation is notconsidered in any of the models, which is unlikethe material models incorporated in FIRIN.Therefore, radioactive partlculate from anextraction column in a reprocessing facility orpartlculate from other liquid operations arehandled in FIRAC as non-condensing media.

Other phenomena such as thermophoresis,turbulent inertia deposition, and coagulationare prevalent for the size particles considered.These phenomena are being studied, and modelsare being developed to upgrade the presentmaterial transport aspect of the codes.

1. Convection. Convection is thetransport of partlculate or heat by the circu-lation or movement of a fluid or gas. When apostulated accident occurs in a fuel cyclefacility, snail particles in the micron rangemay be released. The particles can be trans-ported by the movement of a gas (air) throughthe ventilation system of the nuclear facility.The transported particulate and the host gas areassumed to form a multiphase, multlcomponentsystem. It is assumed in the material transportmodel of FIRAC that particle sizes are small,that the small particles have a small relaxationtime compared to its residence time, and thatthe particle mass fraction Is small relative togas mass in the same volume. Thus, the particu-late would have nearly the same velocity as thegas. Under these conditions the gas andpartlculate are in dynamic equilibrium. The gas

velocity obtained from the gas-dynamics calcula-tion along with the continuity equation

— / p dv = - / p up • ds + Mpv P s P

(6)

where: p = particulate density based onp mixture volume (equals gas

density)

particulate velocity (equals gasvelocity)

Mp

can be used to determine the material densityconcentration.

2. Gravitational settling. A smallparticle falling under the action of gravitywill accelerate until the drag force justbalances the gravitational force. It will thencontinue to fall at a constant velocity known asthe settling velocity. The settling velocitycan be obtained by equating the drag force tothe gravitational force, then solving forvelocity. The drag force is obtained by solvingthe Navier-Stokes equation which describes fluidmotion, using simplifying assumptions. Thesolution is referred to as Stoke's Law (Hinds,1982).

FD - 3wnud (7)

where: r\ - viscosityu « particle velocityd - particle diameter

Equating Stoke's law to the gravitational forceproduces the settling velocity equation:

P d 2

P 818n (8)

The settling velocity Is used in FIRAC as a sinkterm to account for the amount of material thatIs deposited due to gravity throughout theventilation system as the material Is convected.

3. Entrainment. Entrainnent is thedisplacement of stationary particles when aero-dynamic forces are equal to or exceed thethreshold velocity (air speed) required toovercome particulate restraining forces.

Iverson (Iverson, 1976) conductedexperiments to determine the threshold frictionspeed (U*t) for a number of material sizes anddensities for a thick bed (0.5 inches orgreater) of material. The parameters U*t andparticle diameter (d ) were plotted, and equa-tions were fitted to the data. The resultingequation along with the equation which relatesthe friction velocity to the velocity at theboundary edge can be used to determine thefriction and threshold velocities. Travis(Travis, 1975) suggested that the followingequation for the particle flux (qv) can be usedto determine the ma&s of particles per unit areaper unit time that go into suspension due toentrainment

245

qv " ql

where:

(9)

mass percentage of suspendableparticles

C , C. « empirical constants (2 x 10and 10" , respectively)

q. = horizontal mass flux along aplane (2.61 ' '-' • -<L

-10

The amount of material suspended is used as asource term in FIRAC, and is added to the amountof material being convected.

C. Duct Heat-TransferThe purpose of the duct heat-transfer

model is to determine how a combustion gas in asystem cools down or heats up as it passesthrough a duct. KIRAC is the only code of thethree which has heat-transfer features. Ductwork is the only ventilation component for whichheat-transfer is modeled. The model predictsthe temperature of a gas leaving a duct when theinlet gas temperature is known.

The heat-transfer model is comprised offive modes of heat transfer. Forced convectionbetween the combustion gas and the inside ductwalls is calculated by the following equation

out inm C

(10)

where: T l n = temp, of gas coming in

T o u C - temp, of gas going out

Q^ = energy transferred from gas toduct wall due to convection(Qcl) and radiation (Qri)

m • mass flow rate of gas coming in

C =• specific heat of gas coming in

The convection component of Q* is a function ofthe gas temperature and the duct wall tempera-ture. The radiation component of Q^ is afunction of intensity factors associated withthe duct wall, gas temperatures, and gas compo-sition. The wall temperature is calculated bysolving the heat conduction equation. Naturalconvection heat transfer from the outside ductwalls to the surroundings, and radiation heattransfer from the outside duct walls to theatmosphere are also modeled.

IV. CODE OUTPUTS

Code outputs include calculatedtemperatures, pressures, and material concentra-tions at nodal points (nodes) as a function oftime. Other code outputs include predictedflows, material flow rates, pressure differen-tials, and material accumulations on filters asa function of time. Code outputs are written toa scratch file during the transient calculationfor editing or plotting by the output processor.

Figures 1-2 illustrate sample temperatureand pressure code outputs for a room where anexplosion is simulated using NOKDAL. Figure 3depicts sample volumetric flow rates throughbranches or duct work connected to the roomwhere the explosion was simulated.

- •

r.

z?.£

AJ

e-

o-

O'

p

f s" ..•-. . T O * •

ID 0.1 GJ 0J 0« 0.5 0.1 0.7 0J 01 10HUE (»1

Figure 1. Plot of temperature versus time forroom where an explosion occurs.

CO 01 D.2 OJ 04 a.i 01 0.7 QM 0* IQ

Figure 2. Plot of pressure versus time for aroom where an explosion occurs.

246

Figure 3. Plot of volumetric flow versus timefor ducts connected to a room where atransient is occuring.

V. CODE SPECIFICATIONS

The LANL developed codes (TORAC, EXPAC, andFIRAC) were programmed to run on a CDC 7600computer. Each of the codes were also modifiedto run on a VAX 11/780 system. A personalcomputer (PC) version of FIRAC has been created.The PC version of FIRAC does not have FIRINincorporated within it. For FIRIN to be incor-porated an overlay procedure must be performed.The present PC version of FIRAC must be compiledin segments due to the size limitation of mostPC Fortran compilers. Each segment must belinked to produce an executible FIRAC file.

The size of each code varies due to theirdifferent features. FIRAC is the largest code.It has about 9000 lines of code (includingcommon and comments) and 85 subroutines. Itslarge size is primarily due to the extensivecompartment model within it (FIRIN), whichconsists of 2247 lines of code. EXPAC also hasabout 9000 lines of code and 44 subroutines.The explosion chamber model NORDAL contributesgreatly to its size. TORAC Is the smallest ofthe three codes. It consists of about 5000lines of code and only 26 subroutines. TORAClacks the capability of modeling multiplespecies and uses an incompressible flowalgorithm to model gas dynamics. Each of thecodes are comprised of numerous subroutineswhich are Indicative of their modular structurewhich allows relatively easy modification.

Manuals that describe the use of each codeare published as Los Alamos Reports (Gregory,1985). Each manual consists of a chapter whichexplains in general terms each of the codefeatures, how to develop input files, and sampleproblems.

VI. CONCLUSIONS AND FUTURE WORK

Codes that predict the gas dynamics, heat-transfer, and material transport induced fromfire, tornado, and explosion transients innuclear fuel cycle facilities are being

developed. These codes would reduce the amountof uncertainty involved in estimating the amountof material that reaches a facility boundary asa result of an accident. The source termobtained from the codes along with plume modelsand other models which describe environmentalconditions outside of the facility can estimatethe impact on public health and safety as aresult of an accident.

Because of the detailed modeling ofventilation systems incorporated in each code,the codes can be used as a tool to aide in thedesigning of ventilation systems.

The material transport models presentlyincorporated in the codes are primarily for thetransport of non-condensing, spherical particles0.5 to 5.5 microns in diameter (typical size ofSNM powders). The codes are presently inappro-priate to model fission product releases fromspent fuel storage facilities, because spentfuel gases and partlculate are smaller than thesize particles modeled in the codes. Diffusionmodels must be incorporated into the codes todetermine an accurate prediction of smallerparticle deposition. The material transportmodels are also Inappropriate for modelingaccidents involving wet solutions similar tothose that exist at solvent extraction opera-tions. Though the FIRIN subroutine has thecapability of modeling diffusiophoresis, whichis due to condensation and Stefan flow, thenetwork modeling approach of the codes Is not initself capable of modeling condensation.

Deposition due to thermophoresls andturbulent inertia Is presently being incor-porated Into the codes.

An experimental facility at New MexicoState University has been built to verify exper-imentally the heat-transfer, material flows, andgas-dynamic models Incorporated in the codes.The experimental apparatus consists of duct workconnected to a cylindrical and rectangularcompartment. A fan provides air flows similarto those encountered in a ventilation system.The air Is heated by passing it through afurnace, which simulates hot air from an explo-sion or fire. The air passes through a numberof dampers and filters in the duct work.Pressure transducers and thermocouples are alsopositioned through the duct to measure pressuredrops and heat, respectively.

REFERENCES

Gregory, W., et al., 1985, "TORAC Users Manual",LA Report 10435-M.

Hinds, W.C., 1982, Aerosol Technology Chap. 3,p. 40, John Wiley and Sons, New York.

Iverson, J.D., 1976, "Windblown Dust on Earth,Mars, and Venus", J. Atm. Sci., 33,2425-2429.

Owzarski, P.C., et al., 1985, FIRIN User'sManual, PNL-4532.

Travis, J.R., et al., 1975, "A Model forPredicting the Redistribution of PartlculateContaminants from Soil Surfaces", LA Report6035-MS.

ANS Topical Meeting on Radiological AccMeats—Perspectives ind Emergency Plaaaiag

REALM: An Expert System for Classifying Emergencies

Robert A. Touchton, Alan Dale Gunter, and David G. Cain

ABSTRACT

In la te 1983, EPRI began a program toencourage and stimulate the development ofArtificial Intelligence applications for thenuclear industry. As a part of that effort,EPRI has contracted Technology Applications,Inc. (TAI) to develop an a r t i f i c i a l lyintelligent Emergency Classification SystemPrototype. This paper presents a descriptionof and a status report on this recentlycompleted prototype.

I . EMERGENCY CLASSIFICATION DOMAIN OVERVIEW

Because there ex i s t s the p o s s i b i l i t y of anaccident at a nuclear facility which couldcause adverse health effects in the areassurrounding the plant, a great deal ofplanning and preparation takes place to reduceor eliminate public health consequences of areactor accident, should one occur. One ofthese planning elements involves theclassification of emergencies into one of fourstandardized classes by use of EmergencyAction Levels (EALs). The four classes ofemergencies are:

Notification of Unusual EventAlertSite Area EmergencyGeneral Emergency

EALs represent the observables, thresholdsand logic for mapping plant and s i t econditions into the appropriate emergencyclass . A great deal of information isavailable in the power plant (both sensors andmanual observables) to help the staff identifyand classify an emergency situation. However,this wealth of data must be reduced in anexpeditious manner during a tine when thestaff is already busy responding to an off-normal condition. This would seem to be aproblem well-suited to computerization, andyet, at most plants, manual look-up tables areused in the emergency classification process.

Attempts to computerize this process arehindered by the fact that the data containssignificant uncertainties. There may bemissing, conflicting, or ambiguous sensorreadings or there may be false alarms.Therefore, judgement is required for properinterpretation of the information. The use ofconservative assumptions within a computerprogram (a technique often used to solve thisproblem in many conventional safety-relatedprograms) does not offer a solution in thecurrent problems because over-reacting is justas detrimental as under-reacting.

The use of an Expert System approach canovercome these problems by incorporatingDomain Knowledge (heuristics) into thecomputer program such that i t can begin tomake judgements regarding the data. Such asolution not only computerizes the dataacquisition and processing, but also canappropriately compensate for the uncertaintiescontained in the data.

I I . REALM OVERVIEW

The Reactor Emergency Action Level Monitor(REALM) i s d e s i g n e d t o p r o v i d e e x p e r tassistance in the determination of theappropriate emergency status for a nuclearpower plant. REALM is an EPRI-funded expertsystem developed by Technology Applications,Inc. to operate in a real-time processingenvironment. REALM embodies a hybridarchitecture and thus utilizes both rules andobject-oriented programming techniques. Therule-base consists of two general classes:deterministic rules, which codify the logicembodied in the current EAL tables, andheuristic rules, s t i l l under development,which will resolve ambiguities and dataconflicts, identify false alarms, and drawinferences in light of missing data. Theheuristic rules will go beyond the current EALstructure to address the more problematicscenarios and en t a i l a more symbolicrepresentation of the plant information. Theheuristic rule processor (inference engine)also has the ability to accept and propagatebelief information such that conclusions caninclude a certainty factor.

247

248

REALM has been designed to ultimatelyinterface with the plant in some manner inorder to collect sensor data. The most likelyinterface is with the Safety Parameter DisplaySystem (SPDS) computer, since i t is alreadycharged with the collection of safety andemergency-related data. However, during thisprototype development stage, REALM utilizessimulated sensor data which is stored on diskand read in as if i t were actual instrumentreadings. REAli-1 must also interface with itsuser(s ) . Approximately 1/4 to 1/3 of thenecessary information must be manually enteredinto the system. However, the manually derivedinputs relevant to any given emergency issmall and, thus, the user is prompted only forthose relevant inputs.

REALM has four operating modes of whichtwo are on-line (Actual and Trial) and two areoff-line (Scenario Development and Training).In Actual Mode, REALM interprets the input toidentify and classify emergencies in real-time. Thus, REALM will display to the user itscurrent findings on the appropriate emergencyclass and the reasoning which led to thosefindings. The Trial Mode allows the user tomake assertions in order to assess theirimpact {e.g. see i f terminating flow in acertain system would cause a higher EmergencyClass to be invoked).

The Scenario Development Mode serves as astep-wise training and exercise f i l e builder.This mode operates off-line and in non-real-time, thus allowing a user from the EmergencyPlanning or Training Staff to assert a set ofplant conditions, see the impact on the EALs,look ahead to find events which would causethe next higher level of emergency, andfinally store on disk that set of conditionsif desired. The Training Mode uses a TrainingFile (developed using the Scenario DevelopmentMode) to sinulate the plant's behavior duringan emergency in order to train operators inemergency classification and users of REALM inits operation. The Training Mode operates inreal-time (as represented on the simulatedaccident scenario) but i s o f f - l ine andincludes the scoring and archiving of thetrainee's response.

Within each mode, the user may request av u l n e r a b i l i t y analysis or the system'sr a t i o n a l e for t h e c u r r e n t p l a n tclassification. The vulnerability analysisevaluates the current knowledge context,reviewing a l l of the rules relevant to thenext higher emergency class and determiningwhat n (for n = 1, 2, 3, 4) events would causet h a t h igher c l a s s to be d e c l a r e d .Conceptually, this involves backward-chaininginference to ident i fy which rules arepartially satisfied and to l i s t the missingantecedents. The processor then translates thel i s t and displays i t in the VulnerabilityWindow. The rationale processor will collectall of the facts (and certainty factors, ifapplicable), process the l i s t , and display theinformation in the Rationale Window.

REALM resides on a Xerox 1108-121 LispMachine within Inte l l icorp's KnowledgeEngineering Environment (KEE), with extensions

( i . e . methods, functions, procedures, and/orother u t i l i t i e s ) wri t ten d i r e c t l y inInterlisp-D. The power of KEE's hybridarchitecture for knowledge representation(frames, active values/demons, or rules) alongwith the powerful man/machine interfaces(graphics, windows, KEE editor, and mouse)have been u t i l i z e d t o p r o d u c e aproductive/effect ive/eff ic ient emergencymanagement decision aid.

I I I . REALM PROBLEM-SOLVING APPRCftCH

The process of emergency c l a s s i f i c a t i o nbeing developed in REALM invo lves a d ia logueamong s e v e r a l "mini -experts ," each chargedwith investigating a specific aspect of theproblem. First, REALM must read-in andevaluate the observables (sensored andmanual). This step includes flagging ofmissing and out-of-expected range readings,the setting and monitoring of clocks, andtrending of the data. The Conditions Expertreviews the data looking for specificconditions of interest (such as steamgenerator levels and pressures). Next, theRadiation Expert must assess radiation levelsand location. His job is to determine abnormalamounts or locations of radiation and tomonitor release rates and off -s i te doseproject ions . Then, the Critical SafetyFunctions Expert must evaluate the status ofthe various Critical Safety Functions and,likewise, the Fission Product Barrier Expertmust evaluate the integrity of each of thethree Fission Product Barriers. Next, theThreats Expert performs f irst-orderdiagnostics to look for known accidentsequences ("first-order" means that thediagnosis only goes deep enough to classifythe emergency but probably not deep enough tocorrect i t ) . If a specific accident i sidentified, the Resources Expert checks forthe operability or availabil ity of theresources intended to counter that accident.Finally, the EAL Expert, using the input fromthe other experts, classifies the emergencylevel by evaluating parameters against theirthresholds, by balancing threats againstcountermeasures, and by evaluating thedegradation (or potential degradation) of theCritical Safety Functions and Fission ProductBarriers.

IV. REALM STATUS

The REALM "alpha" prototype was completedin early September and i s currently undergoingextens ive evaluation by the TAI Project Team,a s w e l l a s p e r s o n n e l from EPRI andConsolidated Edison (the collaboratingutility). The primary focus of current effortsis on system performance (i .e. REALM responsetime). During the development of REALM, noeffort was expended on maintaining systemperformance. An early prototype (completed inFebruary, 1986) could process a "snapshot" ofplant data in about 30 seconds, a duration

249

considerably faster than could be accomplishedby a human expert charged with solving theproblem. However, as features and scope havebeen added, performance has gradually degradedto about 10 minutes. Therefore, our currentmission is to get REALM'S response time backto the 1/2 minute time-frame. Other currentwork revolves around comple t ion andincorporation of "advanced topics" such as

model-based conflict resolution and certaintyfactor propagation.

Finally, we have begun to actively solicitadditional industry evaluations of REAU-' andthus extend an invitation to u t i l i t i e s tovisi t TAI (in Jacksonville, FL) or EPRI (inPalo Alto, CA) to participate in our formalevaluation program.

Section IV

Economic Issues

Chairman: David AldrichScience Applications International Corporation

ANS Topical Meeting on Radiological AccMeats—Perspectives and Energeacy Ptawiag

Perspectives on the Economic Risks of LWR Accidents'*

Lynn T. Ritchie and Richard P. Burke

ABSTRACT. Models which can be used for theanalysis of the economic risks from eventswhich may occur during LWR operation have beendeveloped. The models include capabilities toestimate both onsite and offsite costs of LWRevents ranging from routine plant forcedoutages to severe core-melt accidents resultingin large releases of radioactive material tothe environment. The economic consequencemodels have been applied in studies of theeconomic risks from the operation of US LWRplants. The results of the analyses providesome important perspectives regarding theeconomic risks of LWR accidents. The analysesindicate that economic risks, in contrast topublic health risks, are dominated by theonsite costs of relatively high-frequencyforced outage events. Even for severe (e.g.,core-melt) accidents, expected offsite costsare less than expected onsite costs for atypical US plant.

I. INTRODUCTION

Several studies have examined the economicrisks from unanticipated events which occur,or could occur, during US LWR operation (Burkeet al., 1984; Starr and Whipple, 1981; USNRC,1975.) Models have been developed in one ofthese studies to estimate the economic conse-quences of LWR forced outages and accidents(Burke et al., 1984). These economic conse-quence estimates can be used together withestimates of event frequencies to calculate the"expected" losses from LWR operational inci-dents of various severities. Both "onsite"costs, which either occur at onsite locationsor most directly affect the plant licensee, and

a This work was supported by the U. S.Nuclear Regulatory Commission and performedat Sandia National Laboratories which isoperated for the U. S. Department of Energyunder Contract Number DE-AC04-76DP00789.

"offsite" costs, which most directly affectthe public surrounding the plant, are includedin the predictions of the economic risks fromLWR operation.

This paper provides a brief overview of theanalytical models which have been developed toestimate the economic risks of LWR outages andaccidents. Assumptions regarding post-accidentoffsite emergency protective measures, whichare closely tied to estimates of economicconsequences, are also reviewed. The resultsof an example analysis performed using theeconomic consequence models are discussed.Some broad perspectives derived from studiesof economic consequences of reactor accidentswhich have b«en performed are reviewed, withcomments in light of the (limited) publiclyavailable information regarding the Chernobylevent.

11. ACCIDENT ECONOMIC IMPACTS, MODELS

A. Onsite CostsEconomic consequence models have been

developed to include the following onsite costswhich either occur at onsite locations or mostdirectly affect the plant licensee:

- power production cost increases(replacement power costs) due to LWRoutage time,

- physical plant capital losses causedby severe accidents,

- plant decontamination costs,- plant repair costs,- early decommissioning costs aftersevere accidents,

- plant worker health impact costs.

The power production cost increases caused bythe need for utilizing generating facilitieswith higher fuel cycle costs during LWR outagesare the dominant loss contributors for unan-ticipated forced outage events which do notresult in core damage. These costs can beestimated using models developed at ArgonneNational Laboratory which account for regionaland/or plant-specific variations in the mix ofgenerating facilities used to provide

253

254

replacement power and the costs of replacementfuels (Buehring and Peerenboom,1982; VanKuikenet al., 1984). These costs range from approx-imately <200/MWe-day up to *1000/MWe-daydepending on the plant location in the Us,seasonal variations, and the costs of availablefossil fuels. Plant repair cost estimates forforced outage events which do not result incore damage are based on operating experiencein the US. The costs of plant repair havegenerally been small relative to replacementpower costs for events which do riot result incore damage.

For more severe accidents involving coredamage, power production cost increases,physical plant capital losses, and plantdecontamination costs are the most importantonsite cost contributors. Plant decontamina-tion costs are modeled based on experience withthe Three Mile Island Unit #2 cleanup programand engineering studies of post-accident decon-tamination (Murphy and Holter,1982). Thesecosts are estimated to range between llB and$2B depending on the physical progression andseverity of the accident. Costs associatedwith decommissioning before the end of plannedplant lifetime and the costs of possible plantworker health effects do not contribute signi-ficantly to the expected onsite losses fromeither routine forced outage events or severeaccidents. In addition to the onsite costcomponents included in the consequence models,electric utility business costs, nuclear powerindustry losses, and onsite litigation costshave been addressed in the development ofeconomic consequence models. These costs couldbe very important to specific organizationsafter severe accidents, but are not includedin estimates of societal losses.

B. Offsite CostsEconomic consequence models have also

been developed to estimate the offsite costsof severe accidents which might result in arelease f radioactive material to the environ-ment. The latest version of economic conse-quence models developed at Sandia NationalLaboratories are contained in the MACCS (MELCORAccident Consequence Code System) code (Alpertet al.,1985; Chanin et al., 1986). The MACCScode models the impact of atmospheric disper-sion and deposition, dosimetry, and publicprotective actions on offsite health andeconomic consequences. Health consequences andoffsite costs are calculated probabilisticallybased on the estimated frequencies of specificradioactive releases (source terms) and themeteorological conditions at the time of arelease. Economic consequence estimates havealso been available in earlier reactor accidentconsequence models including CRAC2, UFOMOD andMARC (USNRC, 1975; Bayer et al.,1982; Clark andKelly, 1981).

The MACCS models include the followingoffsite costs of public protective measures andhealth impacts for severe LWR accidents whichmight result In a significant release ofradioactive material to the environment:

- evacuation costs,- temporary population relocationcosts,

- agricultural product disposal costs,- land and property decontaminationcosts,

- land interdiction costs,- permanent population relocationcosts,

- health impact and medical care costs.

The costs of evacuation and temporary popula-tion relocation include outlays for food,housing, and transportation to move individualseither prior to a release of radioactivity orfrom areas contaminated immediately after arelease occurs. The loss of productivity ofindividuals moved from contaminated areas isalso included in the relocation costs.Agricultural product disposal costs are esti-mated based on the market value of contaminatedgoods which are not suitable for consumption.Land and property decontamination costs arebased on cleanup efforts aimed at achievingspecific exposure reduction factors in urbanand rural areas. In addition, the cost ofpopulation relocation during decontaminationis also included in the offsite cost models.The costs of land interdiction are estimatedusing present value discounting and theestimated tangible wealth contained withinareas which cannot be decontaminated to accept-able levels for habitation. The costs ofpermanent population relocation from inter-dicted areas, including possible periods ofproductivity losses, are also calculated in themodels. Finally, the purely economic costs ofoffsite health effects, including lost produc-tivity and medical costs, are included in themodels.

III. EXAMPLE;

The onsite and offsite economic consequencemodels have been combined with historical out-age frequency data and estimates of severeaccident frequencies to estimate the economicrisks from plant operation over the remaininglifetime of a typical US LWR facility. Table1 shows the expected losses from both routineforced outage events and severe (core-melt)accidents for the remaining lifetime of theexample plant (assumed to be approximately 30years). The present values of expected plantlosses for the remaining lifetime are expressedin 1982 US dollars and are shown for discountrates of 0%, 4%, and 10%. Table 1 shows thatthe expected losses from severe accidents aresmall ($l-$6 million dollars) relative to theexpected losses from routine outage events($84-$270 million dollars) for the remaininglifetime of this plant. This results from therelatively high frequency of forced outageevents and the substantial power productioncost increases for LWR forced outages. Table1 also shows that even for core-melt accidents,expected onsite losses are substantially largerthan expected offsite losses.

An example of the breakdown of offsite cost

Table 1. Present Value of Routine Outage and Severe AccidentEconomic Risks for Remaining Life of Example Plant

Routine Forced Outage Events Core Melt Accidents(No Core Damage) (=6xlO~5/reactor-year)*

Discount Rate

Ot

4t

104

(=ao/reactor-year)

I2.7X108

$1.6xlO8

$8.4X1O7

Offsite

$4.4xlO5

$2.5xlO5

$1.3xlO5

Onsite

$5.5xlO6

$3.3xlO6

$1.7xlO6

TpJt.aJ

5.9X1O6

$3.6xlO6

il.8xlO6

Estimated risks for core-melt accidents based on RSS PWR core-meltaccident frequencies and source terms with conseguence calculationsperformed with new economic conseguence models.

256

components for both a large release of radio-active material (the PWR-2 release category inthe RSS) and .a smaller release (the PWR-5release category in the RSS is shown in Table2. All costs are expressed in 1982 dollarsand are mean estimates conditional upon therelease. The table shows that costs associatedwith land and property decontamination andinterdiction are the most important contribu-tors to offsite costs for a large release ofradioactive material, and that the costsassociated with evacuation, agriculturalproduct disposal, and health effects and healthcare costs become more important for smallerreleases. However, the comparison of totaloffsite and onsite costs in Table 2 shows thatoffsite costs are negligible relative to onsitecosts for the PWR-5 source term. Even for thePWR-2 source term, onsite costs are larger thanoffsite costs for the example site.

Table 2 also shows other attributes of thepost-accident recovery program which are calcu-lated in the economic consequence models. Theestimates of population exposures avoided byprotective measures are useful for performingcost-benefit analyses in developing protectiveaction criteria. The attributes of therequired decontamination program, includingexposures to workers and labor requirements,are useful for determining if resource limi-tations would be a problem in postaccidentrecovery operations. For example, Table 2shows that a large number of decontaminationworkers would be required to complete therecovery program in a short period of timefollowing a very large release (PWR-2) ofradioactive material.

In general, analyses perfoi-med with theoffsite economic consequence models indicatethat the costs of population evacuation,temporary relocation, and possible agriculturalproduct disposal are important for severeaccidents which result in only in very smallreleases of radioactive material to theenvironment. The costs of decontamination orinterdiction of offsite land and propertybecome important for severe accidents whichresult in larger releases of radioactivematerial, and dominate the offsite costs of lowprobability accidents with very large releasesof radioactive material. The economic conse-quence models in MACCS are currently beingemployed in studies of LWR accident conse-quences that are based on updated source terminformation. The studies have shown that theoffsite economic consequences of an accidentare strongly dependent upon source term defi-nition and the criteria chosen for the imple-mentation of population protective measuresincluding land area decontamination, interdic-tion, and population relocation. Changes insource term definition and the criteria chosenfor the implementation of population protectivemeasures can impact offsite cost estimates byseveral orders of magnitude. The site demo-graphic characteristics also directly influenceeconomic consequence estimates, but lessdramatically.

IV. PERSPECTIVES AND DISCUSSION

The analyses performed with the economicconsequence models described in this paperprovide the following general perspectives andconclusions regarding the economic risks fromUS LWR operation:

1. In contrast to public health risks, theeconomic risks from US LWR operationare dominated by relatively high fre-quency, small consequence forced outageevents. Most of the costs of theseevents results from reduced plantavailability and capacity factors andthe need for use of higher marginalcost fuel sources for the generationof electricity.

2. The economic risks from US LWR opera-tion are dominated by onsite lossesresulting from replacement power costsfor short-term outages. Severe acci-dent economic risks are also dominatedby onsite losses including replacementpower costs, plant capital losses, andplant decontamination costs. Only verylow probability core-melt accidentswith large releases of radioactivematerial could result in offsite costsas large as onsite costs.

3. The onsite economic risks from severeaccidents may be significantlyincreased for plants at multiple unitsites. This is due to the possibilitythat a single event may result indamage and loss of power productioncapability at more than one plant.

4. Offsite costs for events which resultin negligible or small releases ofradioactive material to the environmentare predicted to be small compared toonsite costs. The costs of possiblepopulation evacuation, temporaryrelocation, and agricultural productdisposal are the most important offsitecost contributors for these events.

5. The costs of decontamination or inter-diction of offsite land and propertybecome more important for low probabil-ity, severe accidents that can resultin larger releases of radioactivematerial.

6. Offsite economic consequence estimatesare also strongly dependent upon thedefiniticn of accident source terms,including the specified radionuclidecontent of a release.

7. Estimates of offsite accident costs arestrongly dependent upon assumptionsregarding population protectivemeasures and the criteria employed fordecision making regarding the implemen-tation of protective actions at offsitelocations.

Table 2 . Onsite and Offsite Accident Costs Conditional Upon RSS CategoryPWR-2 and PWR-5 Releases. Example Plant and Site

Mean Costs (1982 U. S. Dollars)

Offsite Cost Component

Evacuation

Emergency Phase Relocation

Intermediate Phase Relocation

Agricultural Product Disposal

Population Relocation during Decontamination

Land and Property Decontanination

Land and Property Interdiction

Interdicted Population Relocation

Offsite Health Effects and Health Care

PWR-2

$4

$2

$7

$9

$7

$6

$1

$2

$1

$1

$3

Release

.4X106

.2X107

.6X107

.lxlO7

.lxlO7

.4X1O8

.6X1O8

.7X107

.7X1O8

.3X10"

.3X1O9

PWR-2 Release

1.

>s 3.

2.

1.

4.

5X105

1X105

6X1O3

1X104

6X10*

Person-Sv

Person-Sv

Person-Sv

PWR-5 Release

$4.4xlO6

$5.9xlO5

$1.7xlO6

$2.3xlO6

$2.8xlO5

$8.9X1O6

$9.9X10*

$2.4xlO2

$6.8xlO6

$2.5xlO7

$3.3x10'

PWR-5 Release

5.9X1O3 Person-Sv

1.3x10 Person-Sv

1.7x10 Person-Sv

Person-Years 1.7x10 Person-Years

Persons 6.9x10 Persons

Total Offsite Cost

Total Onsite Cost

Other Attributes froi Economic Models

Total Population Dose Incurred. 0-100 Years

Total Population Dose Avoided by Protective Measures

Total Dose to Decontamination Workers

Labor Required for Decontamination Program

Number of Decontamination Workers Required

for Completion of Program in 90 Days

258

8. The analytical models and quantitativepredictions of offsite economic lossesfrom severe reactor accidents may varysignificantly for different nations dueto fundamental differences in nationaland/or regional economies. However,the philosophy employed in the develop-ment of the MACCS economic consequencemodels, specifically the calculationof economic consequences based onprojected implementation of populationprotective measures, has broadapplicability.

9. The economic consequence modelsdescribed in this paper have beendeveloped specifically for LWR plantaccidents, and no calculations havebeen performed for other plant types.However, the limited publicly availableinformation regarding the Chernobylevent seems consistent with the assump-tions and philosophy incorporated intothe economic consequence models,particularly in the modeling ofoffsite population protective measures.

REFERENCES

Alpert, D. J., et al., "The HELCOR AccidentConsequence Code System (MACCS)," SAHD85-0884C, Sandia National Laboratories, April1985.

Bayer, A., et al., "The German Risk Study:Accident Consequence Model and Results ofthe Study," Nuclear Technology. Volume59, October 1982.

Buehring, W. A., and Peerenboom, J. P., "Loss' of Eenefits Resulting from Nuclear Plant

Outages," Argonne, IL, Argonne NationalLaboratory, NWREG/CR-3045 (ANL/AA-28),March 1982.

Burke, R. P., Aldrich, D. C , and Rasmussen,N. C , "Economic Risks of Nuclear PowerReactor Accidents, Sandia National Labora-tories, Albuquerque, NH, NUREG/CR-3673(SAND84-0178), April 1984.

Chanin, 0. I., at «1., "MKLCOR Accident Conse-quence Code System (MACCS) User's Guide,"NURKC/CR-4467, Sandia National Labora-tories, preliminary version, February 1986.

Clark, R. H. and Kelly, G. I., "MARC - The HtPBMethodology for Assessing RadiologicalConsequences of Accidental Releases ofActivity," Chilton, England, NationalRadiological Protection Board, NRPB-R137,1981.

Murphy, E. S. and Holter, G. M., "Technology,Safety, and Costs of DecommissioningReference Light-Water Reactors PollowingPostulated Accidents," Richland, HA,Pacific Northwest Laboratories,lTUREG/CR-2601, September 1982.

Starr, C. and Whipple, C., "Coping with NuclearPower Risks: The Electric Utility Incen-tives," Electric Power Research InstitutePaper, September 10, 1981.

U. S. Nuclear Regulatory Commission, "TheReactor Safety Study, Main Report,"Washington, DC, WASH-1400 (NUREG-075/14),October 1975.

VanKuiken, J. C , Buehring, W. A., andGuziel, X. A., "Replacement Energy Costsfor Nuclear Electricity-Generating Unitsin the United States," Argonne, IL,Argonne National Laboratory, NUREG/CR-4012(ANL-AA-30), October 1984.

ANS Topical Meetog <m RaitotofkuU AccMerti—Perspectives u d Eaergemcy Pl»—i«f

Restoration of a Radiologically Contaminated Site:SAGEBRUSH IV

Jack J. Tawil

ABSTRACT This paper describes the developmentof a site restoration plan for the SAGEBRUSH IVexercise held In northeastern Washington state1n mid-August, 1986. DECON, a computer programdeveloped by Pacific Northwest Laboratory, wasused to produce Information around which thesite restoration plan was developed. The featuresof DECON that are demonstrated In this paperIndicate Its potential usefulness as a planningtool for site restoration. Strategies that areanalyzed with DECON Include: 1) prohibitingspecific operations on selected surfaces; 2)requiring that specific methods be used on selectedsurfaces; 3) evaluating the trade-off betweencleanup standards and decontamination costs; 4)evaluating the sensitivity of the results tovarious assumptions; and 5) varying cleanup stan-dards according to expected human exposure tothe surface.

I. INTRODUCTION

In August 1986, several federal, state andlocal agencies participated In a field exerciseat a remote location In northeastern Washingtonstate, near Kettle Falls. Called SAGEBRUSH IV,the exercise centered around a simulated ra1d-a1rcollision between a C-141 and an F-16 aircraft.Nuclear weapons aboard the C--141 were releasedand damaged, causing dispersion of radioactiveproducts Into the environment. A site map 1sshown 1n Figure 1. Tills paper describes thesite restoration analysis conducted during theexercise, using DECON, a computer program deve-loped at Pacific Northwest Laboratory. In additionto the analyses performed at the exercise, thepaper also Is used to demonstrate other availablefeatures of DECON.

II. A DESCRIPTION OF DECON

DECON Is an analytical tool that can be usedto develop a site restoration plan for extensiveland areas that have been contaminated with radi-ological products. It works on the principle ofminimizing the social costs of the accident byselecting a decontamination strategy. Althoughnot all of the off-site social costs of an accident

are explicitly considered by DECON, a large major-ity of them are taken Into account. Informationreported by DECON Includes 1) the least costlydecontamination method that will effectively re-store each contaminated surface, 2) the cost ofthe method, 3) Its effectiveness, 4) the rate atwhich It can be applied, and 5) the manpower andequipment required to Implement It.

DECON makes use of two data bases. The first,the reference data base, contains data on decon-tamination operations and methods. An operationIs defined as a single technique for decontaminat-ing a surface. Operations currently Implementedby DECON and their symbols are presented In Tablt1. A decontamination method Is defined as oneor more sequential operations. For example, themethod VFR consists of the sequential operations:vacuum (V), foam (F), and remove and replace(R). The reference data base 1s documented 1nTawil et al. (1985).

TABLE 1. Decontamination Operationsand Symbols

Plow PVacuum Blast QStrippabie Coating RDefoliate SLeach-EDTA TFoam t3" Asphalt UHI Pressure Water VSteam Clean vWash * Scrub WResurface XLeach-Fed xClose Now YClear; Harvest ZPlane, Scarify; Radical

Thin Paved LayerVery HI Pres. WaterRemove t ReplaceSandblastSealer; FixativeFixative, AerialHydrobiastVacuumOouble VacuumLo Pressure WaterScrape 4"-6"Double ScrapeDeep PlowRemove StructurePrune

The second data base Is comprised of site-specific Information, Including the type ofproperty that Is contaminated, the value of theproperty, and the severity of contamination.The first step In preparing the site data baseIs to divide the accident site Into a grid. Thegrid elements may be of any size or shape. Twocriteria are applied In defining the grid elements.First, we assume that all exterior, horizontalsurfaces within a grid element are equally contam-inated. Second, we assume that within a gridelement physically similar property has the same

259

260

>29.6 MBq/rn- - 1.85 -29.6 MBq/m2

0.37-1.85 MBq/m2

112 - 374kBq/m2

<37.4 - 111 kBq/m2

7.4 - 37.4 kBq/m2

Colville National Forest Control Area

West BankPopulation 300 I

\__5- /

Fig. 1. SAGEBRUSH IV Exercise Site with Isopleths and Grid Elements

economic value. There Is virtually no limit tothe number of grid elements that can be processedby DECON. While a finer grid could be expectedto give more accurate results, It would alsorequire the user to provide a larger quantity ofsite-specific Information.

DECON operates on the principle that Identicalmethods can be used to decontaminate like surfacesthat have been equally contaminated. The appealof this approach Is that the entire analysis canbe based on surfaces. Some land uses, such asstreets, wooded areas and vacant land, can eachbe thought of as consisting of Just one type ofsurface. Other land use types—notably residen-tial, commercial and Industrial—are best thoughtof as consisting of a wide variety of surfaces.These types must be decomposed Into their constitu-ent surfaces If they are to be made amenable tothe "surface" approach being suggested here. Thesurface types currently Implemented by DECON arelisted In Table 2, along with exposure factors,which are discussed later.

TABLE 2. Surface Types Currently Implementedby DECON, and Exposure Factors

Agricultural FieldsOrchardsVacant LandWooded LandStreet/Roads, AsphaltStreets/Roads, ConcreteExterior Walls, WoodExterior Walls, BrickFloors, LinoleumFloors, WoodFloors, CarpetedFloors, ConcreteInterior Walls, PaintedInterior Walls, ConcreteRoofsLawnsOther Asphalt SurfacesOther Concrete Surfaces

1.04.010.010.06.06.01.51.50.50.50.51.50.51.51.01.31.01.0

261

DECON makes several adjustments to propertyvalues as a result of the accident. One adjustmentIs to reduce the value of property because of resi-dual contamination levels after site restorationhas been completed. The size of the discountshould depend on perceived health risks. The usercan specify a different discount factor for eachland use. A second adjustment Is for deterior-ation of property that may occur between the timeof contamination and site restoration, a periodduring which the property Is likely to remainIdle. A different deterioration factor can bespecified for each land use. Finally, the costsassociated with loss of use of the property areevaluated. In computing the social costs of theaccident, all costs are discounted to the end ofthe year In which they are assumed to occur.

III. COLLECTION OF DATA FOR THE SAGEBRUSHEXERCISE

An aerial survey conducted on the first dayof SAGEBRUSH IV by EGSG revealed the 37 kBq/m2 (1.0/*C1/m2) Isopleth over the accident site; thisIsopleth gave a footprint of the contaminatedarea and facilitated the planning of a detailedaerial survey for the second day. AdditionalIsopleths were then developed from the secondday's aerial survey. Because of the distortingeffect of scattered shine from an aerial survey,field monitoring teams were sent out to conducta ground survey. A few ground measurements weretaken on the first day, several on the secondday, and the filling 1n was completed on thethird day. These ground readings were enteredon the site map, enabling the Isopleths to becorrected for the scattered shine. The Isoplethsused In the site restoration analysis are shown1n Figure 1. Also shown In this figure are thegrid elements, which are bounded by the Isopleths.The number of grid elements turned out to be 15,varying In size from 22,300 m2 to 13,000,000 m2.

Radiological products 1n the environmentwere assumed to consist of 90 percent 239Pu and10 percent 241Am by activity. Measured activitylevels ranged from less than 7.4 kBq/m2 to over29.6 MBq/m2. The 7.4 kBq/m2 figure 1s of specialInterest: 1t 1s the EPA screening limit. AnIndividual exposed to this limit will not receivea dose higher than the Individual dose limit.

A variety of land uses was represented withinthe contaminated area, as shown In Table 3. ThisInformation was obtained from USGS maps and fromsources In the field.

TABLE 3. Land Uses In the Contaminated Area

ResidentialCommercialIndustrialStreets/Roads

Wooded AreasAgriculturalOrchardsVacant Land

In addition to ground concentration levelsand land use Information, DECON also uses Infor-mation on property values. Typical values forproperty represented by the different lanii useswere obtained from the Ferry County Assessorduring the exercise.

IV. SITE RESTORATION ANALYSIS OF THESAGEBRUSH IV ACCIDENT SITE

Results relating to restoration of the SAGE-BRUSH IV site using DECON are described In thissection. First, DECON was run for the entirecontaminated area within the 7.4 kBq/m2 Isopleth.For this "base case" we had to make severalassumptions. We assumed that a Class A disposalsite would be created within the contaminatedarea for radiological waste generated during thecleanup process. An average hauling distance of15 miles to this site was assumed.

The decontamination costs contained In thereference data base seriously underestimate theactual costs of labor and equipment that would beused 1n the cleanup. First of all, the data basecosts are 1n 1982 dollars. Second, labor andequipment costs are based on operations In anonhostile environment. Third, an allowance ofone hour per eight-hour shift Is allowed forradiation control measures, and this may not beadequate. And, finally, protective clothing andother radiation control measures may reduce workerand equipment productivity. For all of thesereasons, labor costs were arbitrarily Increasedby 70 percent and equipment costs by 40 percent.

A cleanup criterion of 7.4 kBq/m2 was appliedto all contaminated surfaces. As noted earlier,this Is EPA's screening limit. The results fro*this base case are reported 1n Section A. Othercases were then analyzed and the results comparedwith those from the base case. First, we wantedto determine how decontamination costs variedwith the cleanup criterion. This analysis IstcrdMcted In Section B. In Section C, we dropthe assumption that the radwaste will be disposedof on sight. Instead, we assume disposal at theHanford Reservation, about 250 miles away. Asubarea analysis Is conducted 1n Section D forgrid elements one through nine. These gridelements are within the Colville National Forest,a wooded area which Is especially costly todecontaminate. First, we consider decontaminatingthis area to a level of 37 kBq/m2. Another optionconsidered Is simply to apply a fixative andrestrict public access to the entire area. InSection E we compare base case results with ascenario In which operations that use water onexterior surfaces are banned. This analysis 1sconducted on grid element 15, a lightly contami-nated area with about four percent of the landarea devoted to residential use. Section F sug-gests an alternative restoration strategy. Adecontamination analysis Is conducted 1n whichdifferent cleanup criteria are applied to differentsurfaces; namely, those surfaces from which Indivi-duals typically receive the greatt.it exposure aredecontaminated to lower residual levels of activitythan surfaces providing relatively low levels ofexposure. Conclusions from this paper are pre-sented In Section 5.0.

A. SAGEBRUSH IV: Base Case

Major results for the base case are sun-marl zed 1n Table 4. The total cost of decontami-nating the 18 Million m1 of surface area withinthe 7.4 kBq/m2 Isopleth 1s $65.6 million, for an

262

TABLE 4. SAGEBRUSH IV Decontamination Results: Base Case(dollars, areas and volumes are In thousands)

MINIMUM CLEAN UP LEVEL ISTOTAL SURVEYING AND MONITORING COSTS ARE $VOLUME OF RADIOLOGICAL WASTE ISTOTAL DECONTAMINATION COSTS ARE $TOTAL SURFACE AREA DECONTAMINATED ISAVERAGE DECONTAMINATION C0STS/M**2 ARE $SURFACE AREA REQUIRING NO DECONTAMINATION IS...PRE-ACCIDENT PROPERTY VALUE IS $POST-DECONTAMINATION PROPERTY VALUE IS $NET PRESENT VALUE OF PROPERTY IS $TOTAL REDUCTION IN PROPERTY VALUE IS $TOTAL POTENTIAL SAVINGS FROM PROPERTY BUY-OUT

1) AT PRE-ACCIDENT PROPERTY VALUES $2) AT NET PRESENT VALUE OF PROPERTY $

SIZE OF RESIDENT POPULATION ISSIZE OF THIS AREA IS

7.4 kBq/m .3519.2483. CUBIC METERS.

65633.18161. SQUARE METERS.

3.6170. SQUARE METERS.

8361.7739.41.

8319.

57325.65543.

520. PERSONS.18137. SQUARE METERS.

average cost of $3.61 per m2. About 70,000 m 2 ofsurface—primarily, the Interiors of buildings1n lightly contaminated areas—required nodecontamination, as they were already within the7.4 kBq/m2 limit. Nearly 2.5 million m^ ofradiological wastes were generated 1n the cleanupoperations and disposed on-site. Surveying andmonitoring costs were estimated at $3.5 million.

The total market value of all real propertywithin the contaminated area 1s estimated tohave a pre-acddent value of only $8.4 million.This 1s an economically troubled area based ontimber, a depressed Industry. In addition, muchof the land is rocky and has little economicvalue—the county assessor valued much of 1t atas little as a dollar per acre—and even agri-cultural lands are marginal. It Is worth notingthat the cleanup costs far exceed the economicvalue of the property. If the federal governmentwere to provide compensation to all property owners(Including Itself, 1n the case of federal lands),social costs of up to $65 million could be avoidedby not restoring the property. (The compensationcost of $8.4 million 1s a transfer, not a socialcost.)

Total manpower and equipment requirementsfor the site restoration, Including surveyingand monitoring operations, are shown 1n Table 5.

Residual contamination levels 1n the basecase are estimated to result In property valuelosses of $622,000. This Is about 7.5 percentof the total pre-accident value of the property.

B. Costs vs. Cleanup Levels: A Trade-OffAnalysis

As expected, we found decontaminationcosts to Increase as stricter cleanup levels wereimposed. For this analysis, we re-ran the basecase, but with cleanup levels of 18.5, 27.75, 37.0,74.0 and 185.0 kBq/m2. The resulting relationshipbetween cleanup level and decontamination costs1s shown in Figure 2. Decontamination costs areseen to drop sharply for cleanup criteria of lessthan 37 kBq/m2, while above this level the decline1n costs Is considerably more modest. However,in order to determine the optimal cleanup level

TABLE 5. Manpower and Equipment Requirements,Base Case (1000s of hours)

DRIVER, HEAVY TRUCK 315.98OPERATOR, MED. EQUIPMENT 124.41OPERATOR, FARM EQUIPMENT 5.38BUILDING LABORER 131.61COMMON LABORER 1.60CLEANING WORKER .33FARM LABORER 7.97PAINTER .12FOREMAN 35.24PILOT 1.43FLIGHT CREWMAN 1.43AIR GROUND CREWMAN 2.85SPRAY OPERATOR .06FRONT END LOADER 60.725000 GAL. SPRAY TRUCK W/PUMP4B0OM .10BULLDOZER 2.31AIRPLANE 1.43WEED SPRAYER 1.94ORCHARD BLAST SPRAYER .97TRACTOR W/PLOW 5.38PICKUP TRUCK 1.59CHIPPING MACHINE 30.04HYDRAULIC EXCAVATOR 5.95VACUUM, HAND .33PAINT SPRAY EQUIPMENT .18MOBILE STREET FLUSHER .0110-TON ROLLER 25.11GRADER 30.31DUMP TRUCK 315.88

on economic grounds, one must know the dollarvalue of avoiding the additional adverse healtheffects that result from moving to a strictercleanup level. The dollar value of avoidinghealth effects, Including long-term fatalities,Is a highly controversial issue and lies beyondthe scope of this paper.

C. Hauling Radwaste Off-Site

The base case assumes that a Class Adisposal site for radwaste could be made availablewithin the accident site. In the event that theradwaste is to be disposed off-site, hauling

263

1

inO<>

7060SO403020100

40 80 120 160 200

Cleanup Level (kBq)

Fig. 2. Cleanup Costs vs. Cleanup Level

distances and costs could Increase substantially.For example. If the radwaste Is transported tothe Hanford Reservation, the one-way haulingdistance will Increase from 15 miles to about 250miles, depending upon the specific route taken.This longer hauling distance will Increase cleanupcosts tc about $250 million, as compared with $66million In the base case. In addition, disposalcosts at Hanford would,be greater than the esti-mated cost of $4.86/m (1982 $) for an on-siteClass A disposal pit (Tawil, 1985).

0. Alternatives to Decontaminating ForestLands: A Subarea Analysis

Analysis of decontamination costs Inthe base case shows that a major share of thesecosts are attributable to decontaminating woodedareas. The density of mature trees In a forestmakes 1t difficult to utilize large but efficientequipment In the decontamination process. Thereliance on less capital Intensive methods causescosts to Increase sharply. While wooded areascomprise about 42 percent of the contaminatedarea, they account for over 96 percent of thedecontamination costs lit the base case. Gridelements one through nine consist primarily ofwooded areas Inside the Colville National Forest.If, Instead of restoring this subarea to a cleanuplevel of 7.4 kBq/m2, we apply a fixative andrestrict public access, the restoration costscan be slashed from $26.7 million to $1.8 million.If we combine this strategy with cleaning upgrid elements 10 through 15 to a level of 37kBq/m2, we can reduce decontamination costs forthe entire site to just $2.4 million. While thisstrategy obviously reduces costs sharply. Itsdesirability depends on the willingness of thosewho must make this decision to accept restrictedaccess to the subarea containing grid elementsone through nine, and accepting a cleanup level —and the associated health effects—of 37 kBq/m1n the remaining areas. It should be mentionedthat while a stricter cleanup level will reducethe expected health effects to occupants of therestored area, expected health effects to radiationworkers will Increase.

E. Restricting the Use of Water In aPopulated Area

water on exterior surfaces. Contaminated waterhas the potential of creating major problems. Itcan penetrate the root systems of plants, cropsand trees and can contaminate water treatmentfacilities. The benefits from using water—acheap and effective way to reduce dosage throughthe external and Inhalation pathways—must there-fore be carefully weighed against the costs. Theresults of running OECON on grid element 15 witha ban on operations using water (I.e., operationsW, H, Q, U, L, and E—see Table 1) are presented1n Table 6. This table also shows the level ofdetail that DECON reports with regard to decon-tamination operations at the grid element level.The columns 1n Table 6 show the type of surface,Its area In 1000 m2 , the contamination levelprior to decontamination In kBq/m2, the methodused to decontaminate the surface (see Table 1),a symbol Indicating whether a restriction orrequirement Is 1n effect for that surface, theresidual contamination level In kBq/m2, the unitcost of applying the Indicated decontaminationmethod 1n $/m2, the total cost (1000 $) of applyingthe method over the entire surface area, and therate (1n m2/h) at which the method can be applied.The results show that In the absence of the re-strictions, water would have been applied toasphalt and concrete streets, parking and otherpaved areas, and to roofs and lawns. While(double) vacuuming (v) 1s a relatively Inexpensivesubstitute for water, the foam (F) used on roofsat $2.65 per m2 and the resodding (R) of lawnsat $9.51 per m2are relatively costly alternatives.

F. Cleanup Standards Based on ExpectedExposure to Activity

DECON allows the analyst to Impose differentcleanup criteria to different surfaces by selectingexposure factors. The potential usefulness ofthis feature lies In the fact that human exposuresto different surfaces vary considerably. HousingInteriors, for example, would usually offer highexposures while highways and wooded areas wouldtend to offer low exposures. The exposure factorsare Interpreted as follows. An exposure factorof 1.0 causes the cleanup standard to be Identicalto the user-specified value, 7.4 kBq/m2 In thebase case. An exposure factor of 2.0 causes theuser-specified value to be doubled, while anexposure factor of 0.5 causes the user-specifiedvalue to be halved. To Illustrate this feature,OECON was run with the exposure factors shown InTable 2.

Decontamination costs fall to Just $9.4million, and only 8.2 million m2 of surface areaneed to be decontaminated (at an average cost of$1.15/m2.) Arguably, this decontamination strategycould result In even less overall population dosethan produced In the base case. Most Interiorsurfaces are decontaminated to levels of just3.7 kBq/m2, and only vacant land and wooded areasare decontaminated to levels above 37 kBq/m2(both are decontaminated to 74.0 kBq/m2). Yetdecontamination costs are less than 15 percentof the base case costs.

In this section we consider the consequencesof prohibiting decontamination methods that use

264

Table 6. Detailed Surface Results with Restrictions, Grid Element 15(dollars, areas and volumes 1n thousands)

SURFACE AREA GRND CONC METH RES CONC COST/M**2 TOT COST RATE

AGRICULTURAL FIELDS 5334 22.20 AORCHARDS 390 22.20 TOVACANT LAND 1431 22.20 AWOODED LAND 5464 22.20 TNEXTERIOR WOOD WALLS 13 2.22 ..EXTER'R BRICK WALLS 2 2.22 ..LINOLEUM FLOORS 4 11.10 TWOOD FLOORS 4 11.10 VCARPETED FLOORS 8 11.10 VCONCRETE FLOORS 3 11.10 VINT'R WOOD/PL WALLS 25 1.11 ..INT'R CNCRETE WALLS 5 1.11 ..ASPHALT STRTS/PRKNG 130 22.20 v /CNCRETE STRTS/PRKNG 130 22.20 v /ROOFS 11 22.20 F /LAWNS 113 22.20 R /OTHR PAVED ASPHALT 1 22.20 v /OTHR PAVED CNCRETE 5 22.20 v /

NOTES:/ • RESTRICTED OPERATION(S) IS IN EFFECT

//// ' UNABLE TO DECONTAMINATE SURFACE

2.22 .0055 29.34 8500.6.29 1.1402 445.03 280..37 .0389 55.67 1770.

6.29 6.3501 34698.05 266.

2.221.114.442.96

.4070

.4260

.4260

.4260

1.451.583.461.26

40.69.69.69.

7.407.401.48.37

7.407.40

29

.0133

.0133

.6480

.5121

.0266

.0266

1.731.73

29.161070.89

.03

.14

863286328140

43164316

METHOD IS REQUIREDDECONTAMINATION NOT REQUIRED

V. CONCLUSIONS

A site restoration analysis conducted onthe SAGEBRUSH IV exercise site suggests a varietyof alternative strategies. Decontamination ofthe entire area to the EPA screening limit of 7.4kBq/m2 Is a relatively expensive option. A trade-off analysis shows that decontamination costs falloff rather sharply with less restrictive cleanupcriteria until a cleanup level of 37 kBq/m2 1sreached, after which the decline Is relativelymodest. Because wooded areas are difficult todecontaminate, restricting access to the contami-nated areas of the Colville National Forest ratherthan decontaminating them can substantially cutthe cost of the cleanup. A requirement thatsignificantly Increases the restoration costs isto haul the radwaste generated In the cleanup toa remote site, rather than creating an on-s1tedisposal area. Restrictions against using wateron exterior surfaces also Increase decontamina-tion costs significantly, especially 1n lightly

contaminated areas. Finally, Imposing differentcleanup criteria on different surfaces, dependingupon their potential for exposure, appears to bea promising way to reduce decontamination costswithout Increasing total population dose.

REFERENCES

Tawil, J.J., et al., 1985. Off-S1te Consequencesof Radiological Accidents; Methods, Costs, andSchedules for Decontamination. NUREG/CR-3413,FNL-4790 (Washington, DC: U.S. Nuclear RegulatoryCommission).

Tawil, J.J., 1985, Cost/Risk Components for Analy-ses to Support Development of Long-Term Reentryand Recovery Radiation Protection Criteria, (DraftReport), Pacific Northwest Laboratory, Richland,WA 99352.

ACKNOWLEDGEMENTS Work Supported under ContractDE-AC06-76RL0 1830 with the U.S. Department ofEnergy.

ANS Topical Meeting on Radiological Accidents—Perspectives and Emergency Plaaaiag

The NRC's Rulemaking to Require Materials Licensees To BeFinancially Responsible for Cleanup of Accidental Releases

Mary Jo Seeman

ABSTRACT On June 7, 1985, the U.S. NuclearRegulatory Commission (NRC) published anadvance notice of proposed rulemaking (ANPRM)in the Federal Register to address fundingfor cleanup of accidents and unexpecteddecontamination by certain materialslicensees.

The NRC asked for public comment to halp themdetermine whether to amend its regulations torequire certain materials and fuel cyclelicensees to demonstrate that they possessadequate financial means to pay for cleanupof accidental releases of radioactivematerials. If licensees lack adequatefinancial resources and funds are notavailable for prompt cleanup, the consequencescould be potentially significant for thepublic, the licencee and the federalgovernment.

The purpose of this paper is to explain thepurpose and scope of the Commission'sproposed regulatory action, as well asdescribing several accidents that made theCommission consider this action.Additionally, the paper will address otherregulatory precedents. Finally, the paperwill conclude by generally characterizing thepublic comments and itans of concern raisedby commenters.

Good afternoon, ladies and gentlemen.Today, I would like to briefly outline thescope and purpose of a proposed regulatoryprogram under review by the Commission. Ifimplemented, such a program would establishfinancial assurance requirements for cleanupof accidental releases by certain categoriesof non-reactor licensees. After describingthe scope of this effort, I plan to brieflyidentify issues raised by the public in thisarea, and present the staff's schedule foraddressing these items of concern.

On June- 7, 1985, the U.S. NuclearRegulatory Commission published an advance

notice of proposed rulemaking (ANPRM) in theFederal Register to address funding forcleanup of accidents and unexpectedcontamination by certain materialslicensees.

In the ANPRM. NRC asked for publiccomment to help them determine whether toamend its regulations to require certain*-materials and fuel cycle licensees todemonstrate that they possess adequatefinancial means to pay for cleanup ofaccidental releases of radioactivematerials, i.e., an authorized release ofradioactive materials due to human error,employee sabotage, system failure, an act ofGod, or defective components.

The proposed regulatory program wouldamend current NRC regulations (10 CFR 30,40, 61, 70, and 72) requiring certain fuelcycle and materials licensees to demonstratethat they possess adequate financial meansto pay for cleanup of accidental releasesof radioactive materials onsite and offsite.Licensees under consideration for this ANPRMinclude radiopharmaceutical manufacturers,pool irradiators, industrial radiographers,users of gauging devices, gas chrpmatography,well-logging, nuclear medicine diagnosis andradiation therapy. Other NRC licensedoperations under consideration include fuelcycle activities such as uranium milling,UF-6 production, and fuel processing andfabrication.

Regulated waste management activitiesinclude commercial low-level waste disposal,independent spent fuel storage facilities,and persons disposing or storing their ownwaste under special license conditions. Ifimplemented, it is anticipated that thefinancial responsibility program would exemptthe following U.S. Department of Energyfacilities: High-level waste repositories(licensed under 10 CFR Part 60), independentspent fuel storage facilities, and monitoredretrievable storage facilities (both licensedunder 10 CFR Part 72). The staff took this

265

266

approach in the ANPrtM because in the event ofan accident, DOE has access to public fundsto pay for cleanup. In the ANPRM, theCommission solicited public comments on theadvisability of having NRC require financialresponsibility for prompt cleanup ofradioactive materials both on-site andoff-site after accidental or unexpectedcontamination by fuel cycle and othermaterials licensees. Any environmentalrestoration required following an accidentalrelease into or upon the land or water wouldbe covered.

The financial assurance program fornon-reactor licensees under consideration bythe Commission is intended to address cleanupfor accidental releases of radioactivematerials. The proposed program would notaddress authorized and predictable activitiesnormally associated with decommissioning orclosure of a non-reactor licensee'soperations. The latter is being addressed ina separate, ongoing Commission rulemaking ondecommissioning.

The financial assurance program beingconsidered by the Commission in this ANPRM isalso separate and distinct from thecompensation program mandated by theCommission regulations issued pursuant to thePrice-Anderson Act, which does not providefunds for cleanup per se. At the time theANPRM was issued in 1985, that, programapplied only to nuclear reactors on amandatory basis and to plutonium processorsand fuel fabricators on a discretionarybasis.

As many of you are probably aware, theCommission has authority under section 170 ofthe Atomic Energy Act to apply Price-Andersonindemnification to any category of licensee,whenever the Commission deems it advisable inthe exercise of its licensing authority.However, the Commission has not chosen toextend Price-Anderson indemnification tomaterials licensees with the single exceptionof those engaged in the use of plutonium in aplutonium processing and fuel fabricationplant. These activities and uses ofmaterials would not be included in theproposed program. Rather, the programproposed in this ANPRM is not intended toprovide compensation to persons for personalinjury or property damage and is, therefore,not a public liability program.

The Commission is considering such aproposed regulatory program in part, becauseit appears that CERCLA, or the ComprehensiveEnvironmental Compensation and Liability Actof 1980 (P.L. 96-510) would not necessarilyprovide funds for releases involving NRClicensees. In a Federal Register Noticeissued on September 8, 1983, the U.S.Environmental Protection Agency (EPA) madethe following policy statement:

EPA has also chosen not to list releasesof source, byproduct, and special nuclear

material from any facility with a currentlicense issued by the Nuclear RegulatoryCommission (NRC), on the grounds that the NRChas full authority to require cleanup ofreleases from such facilities. The EPAprovided written comments to the Commissionregarding the ANPRM, and in their writtencomments stated, "EPA strongly supports theconcept proposed in this rulemaking."

I now want to briefly provide somebackground information regarding the scopeof the proposed effort. In the ANRPM, theCommission staff noted that although therewas little information available on thefinancial condition of NRC fuel cycle andmaterials licensees, they believed that mostof these licensees already had some financialresources or insurance coverage for on-siteand off-site cleanup as a prudent businesspractice. However, if a licensee did nothave adequate financial resources and anaccidental or unexpected contamination didoccur, the staff felt that there could beboth short and long term adverse publichealth and safety consequences from theradioactive contamination, as well as loss ofuse of the contaminated property.

for the purposes of initial discussionin the ANPRM, the Commissiori noted that itwas considering a $2,000,000 baseline as therequired maximum amount of financialresponsibility for materials and fuel cyclefacilities. This figure was chosen becauseit is in the range of known cleanup costs forNRC licensees and of other dollar amounts ofState and Federal financial responsibilityrequirements for cleanup of accidentalreleases. The amount would be changed toreflect changes in inflation and technology.Types of possible acceptable financialassurances that the Commission staff isconsidering include liability and propertyinsurance from the conventional and nuclearinsurance pools, cash or negotiablesecurities held by a third party, a financialtest, a surety or performance bond, and acorporate guarantee from a parent company.

The issue of financial responsibilityfor cleanup of accidental releases ofradioactive materials caused by non-reactorlicensees has also been an issue w^th theStates. Some States have developed financialassurance programs covering their non-reactorlicensees to address this issue.Additionally, the Conference of RadiationControl Program Directors recognized the needfor Federal standards in this area. TheCommission sent a draft of the ANPRM toofficials in all fifty States asking fortheir comments. Staff also discussed theANPRM with State radiation control programofficials earlier. State comments generallysupported the development of the rulemaking,and specific comments were incorporated intothe ANPRM.

The need to consider such aproposed regulatory program was based on

267

concerns regarding accidental releases ofradioactive materials. Both NRC andAgreement State licensees have had accidentalor unexpected releases of radioactivematerials that have been costly to cleanup.State and Federal estimates for cleanup costshave been estimated at up to $2,000,000 fora single event. Some recent examples that theNRC staff is aware of include the following:

In 1983, a cesium-137 sealed source wasaccidentally ruptured. Workers inadvertentlyspread the contamination into residences andpublic buildings. The cost for cleanup ofof this contamination was estimated to be atleast $500,000.

In 1982, an americium-241 sealed sourcein use in a well-logging operation wasinadvertently ruptured, resulting incontamination of both on-site drillingequipment and off-site homes and commercialresidences. Cleanup costs were estimated tobe up to at least $1,000,000.

During 1979 and 1980, a tritiummanufacturer's operations in Tucson, Arizona,resulted in releases both on and off-site.State officials estimated that the Statespent approximately $2,000,000 in labor andcapital costs for removal and cleanup of thetritium.

The NRC staff have had a difficult timeestablishing the scope of the problem ofcostly cleanups required as a result ofaccidental releases of radioactive materials.Relatively little data on cleanup costs foraccidental releases of licensees is availablein'NRC records, mainly because this type ofinformation is not required to be submittedto the agency.

However, a recent review by the staff ofNRC unusual events reports for radioactivereleases by materials and and fuel cyclelicensees indicates that from 1980 to 1983,accidental or unexpected releases fromlicensees' operations believed to involvesignificant cleanup costs (more than a fewthousand dollars) involved fewer than onepercent annually of the total fuel cycle andmaterials licensees authorized to possess anduse byproduct, source and special materials.

Besides lacking a comprehensive data basefor licensee cleanup costs for accidentalreleases, the Commission also has noavailable records to determine if licenseeslack (or previously lacked) adequate funds toprovide for prompt cleanup of accidentalreleases. Accordingly, in the ANPRM, theCommission solicited input from the publicand the industry on the scope and magnitudeof the problem.

However, even given this lack of astrong data base, the NRC staff does notbelieve it is prudent for the Commission to

wait until an event occurs which requiresexpensive cleanup to consider the developmentof such regulations. Other agencies such asthe U.S. Environmental Protection Agency,have gone forward with financialresponsibility requirements in the absence ofa large documented accident data base.

Additionally, several NRC-funded studieshave presented cost estimates for cleanup ofNRC licensed fuel cycle facilities related toemergency planning issues and thePrice-Anderson Act. See a 1979 paper authoredby J.P. McBride, and entitled "EconomicConsequences of Accidental Releases from FuelFabrication and Radioisotope ProcessingPlants", prepared by Oak Ridge NationalLaboratories for the U.S. NRC(NUREG/CR-0222). Another example can befound in a 1983 working paper by H.K. Elderentitled "Technology, Safety and Costs ofDecommissioning Reference Nuclear Fuel Cycleand Non-Fuel Cycle Facilities FollowingPostulated Accidents", Pacific NorthwestLaboratories, (NUREG/CR-3293).

However, since these studies wereprepared for different purposes andassumptions, it is difficult to compare theresults or to use their conclusions as thebasis for estimating cleanup costs followingan accidental release caused by a non-reactorlicensee's operations. Accordingly, in theANPRM, the staff noted that they proposed touse the limited but actual, past cleanup costexperience (discussed previously) as thebasis for setting the amount of financialresponsibility coverage for non-reactorlicensees.

In the ANPRM, the NRC staff stated thatthey would consider at a later date, the issueof financial responsibility for the smallnumber of licensees who have the potential tobe involved in the significantly more costlycleanups postulated in the NRC-fundedstudies.

The Commission's consideration of such afinancial assurance program is in line withthe regulatory precedents enacted by avariety of State and Federal agencies chargedwith protecting the public health, safety,and the environment. For example, federalagencies have enacted requirements pursuantto the Motor Carrier Act, the ResourceConservation and Recovery Act, the FederalWater Pollution Control Act, the SurfaceMining Control and Reclamation Act, the OuterContinental Shelf Lands Act, and the DeepWater Port Act.

The stipulated dollar requirementsvary from $10,000 to over $5,000,000 forthese different programs. As an example,the 1984 minimum levels of financialresponsibility for motor carrrierstransporting hazardous substances rangesfrom between $1,000,000 to $5,000,000.

268

A number of commenters responding to theANPRM suggested that the NRC staff shouldtake a closer look or even follow the U.S.Environmental Protection Agency's (EPA)regulatory approach with regards to requiringfinancial responsibility. To briefly bringyou up to speed, the U.S. EPA regulations forliability coverage required owners andoperators of treatment, storage, and disposalfacilties to demonstrate liability coveragefor sudden and accidental occurrences in theamount of $1 million per occurrence and$2 million annual aggregate, all exclusiveall legal defense costs.

Owners and operators of surfaceimpoundments, landfills, and land treatmentfacilities were required to demonstrateliability coverage for nonsudden accidentaloccurrences in the amount of $3 million peroccurrence and $6 million annual aggregate,exclusive of legal defense costs.

These commenters pointed out that after EPAhad promulgated these requirements forliability coverage, a variety of partiesinformed EPA that the current state of theinsurance market prevented them fromobtainging insurance. Accordingly, EPA wasfaced with considering whether to revise oreven eliminate their financial responsibilityrequirements.

Accordingly, NRC staff looked at how EPAaddressed the problem of a contractinginsurance market after they had promulgatedinsurance requirements. NRC staff examinedthe July 1985 testimony of Gene Lucero, whoat the time had the title of Director ofEPA1 s Office of Waste Programs Enforcement.In his testimony, Lucero noted that manyapproaches had been proposed to address thepollution liability insurance shortage. Henoted that EPA believed that no one approach,but rather a .combination of severalapproaches was needed to solve the insuranceshortage. These included a return tocareful underwriting practices through use ofenvironmental audits, premiums that reflectrisks, and contracts that address policylimits in unambiguous approaches.

Additionally, NRC staff also examined theJuly 21, 1985 EPA Notice of a ProposedRulemaking, which asked for comments toenable EPA to determine if revisions to theirregulations were necessary in light of thecurrent state of the insurance market. EPAnoted that they were considering taking oneor a combination of the following fiveregulatory actions in response to theproblem:

0 Maintain existing liability requirements;

0 Clarify the required scope of coverageand/or lower the required levels ofcoverage;

0 Authorize other financial responsibilitymechanisms;

0 Authorize waivers; and

0 Suspend or withdraw the liabilitycoverage requirements

At this time, EPA has apparently chosen thethird alternative. In July of 1986, theyissued an interim final rule stating thatthey had decided to authorize owners andoperators to use a corporate guarantee asanother mechanism to comply with the liablitycoverage requirements.

As I noted earlier, the NRC staff has,and will continue to monitor and assess thestatus of the EPA liability program in theirconsideration of a NRC financialresponsibility program for cleanup ofaccidental releases of nonreactor licensees.

Finally, I would like to brieflysummarize the statistics regarding whocommented and what their concerns were.The Commission used the ANPRM to solicitinput from all interested parties regardingthe scope and nature of such a proposedregulatory program. Besides mailing theANPRM to all non- reactor parties holding aspecific license, the NRC also targeted adirect mail to NMSS licensees, public andspecial interest groups, as well as tradeassociations representing the business,finance, and insurance communities. The ANPRMgenerated a great deal of interest,particularly among NRC materials licensees,and State and Federal regulatory agencies.NRC received 159 written responses fromlicensees, insurance and surety tradeorganizations, federal state and localgovernments, and members of special interstgroups and the public.

The ANPRM generated a great deal ofinterest, particularly among NRC materialslicensees, and State and Federal regulatoryagencies. NRC received 159 written stateand local governments, and members ofspecial interest groups and the public.

Approximately 80% of the commenters wereeither NRC or Agreement State licensees. Themajority of these commenters identifyingthemselves as licensees expressed concernregarding the cost impacts of such a program.Others felt that such requirements wereunnecessary because they were not aware ofany real problem regarding the inability oflicensees to pay for accidental releases orabandonments. Others suggested that the NRCshould consider exempting operations similarto theirs from the coverage of thisrulemaking.

Around sixteen percent of commenterswere identified as governmental organizations,

269

such as municipal, state, or federalagencies. Many supported the need for thisrulemaking. Some felt that federal financialresponsibility standards were necessary toprotect the public health and safety and theenvironment. Others indicated that such aprogram was appropriate becauseresponsibility for paying for cleanup ofaccidents and abandonments should rest withthe licensees, and not with the taxpayers.

The remaining four percent of commenterswere identified as members of the public,environmental groups, and trade associationsrepresenting insurance and surety companies.The trade associations pointed out thatsurety or insurance coverage for the programintended by NRC was either not available ortoo costly. One surety trade associationsuggested that the NRC staff allowflexibility in developing standards so thataffected licensees would have the ability todemonstrate financial assurance through anumber of different instruments. In general,environmental groups and the public supportedthe rulemaking, although one felt that NRCwould not go far enough in setting up such aprogram.

In their review and analysis of thepublic written comments, the staff identifieda number of major items that were raised by

0 What type of history has there beenfor accidental releases?What history has there been ofaccidents involving nuclear materialswhere the licensee has been unablefor cleanup?

0 How does the Commission intend todetermine coverage for such aprogram?

0 What are the cost impacts tolicensees and the public for such aproposed program? What are the costimpacts of not implementing such aproposed program?

0 What amount of coverage isappropriate? Which licensees shouldbe covered under such a program? Whoshould be exempted?

0 Is liability insurance available forthis type of risk? What willpremiums cost?

0 What types of financial assuranceshould be allowed?

The staff has reviewed and analyzed thesequestions and is continuing to assess howthese issues will be resolved. Jt would bepremature at this time to provide conclusionswith regard as to how the staff and theCommission intends to address these issues.However, I can say that the staff recogniiesthat the issues regarding who should becovered and cost impacts of such a programwere complex and required a more in-depthanalysis.

Accordingly, the staff has obtainedtechnical assistance in these areas. Thefirst project is intended to help staffbetter assess the extent to which accidentfrequencies can be used to determine amountsof coverage for different licenseecategories. Appropriate exemptions fordifferent licensee categories will also beexamined. Secondly, the staff also hasobtained technical assistance to help themarrive at an evaluation of the benefits andcosts of such a proposed financial assuranceprogram to all parties. The staff hopes touse the results from both studies to help themdetermine how to address these two concerns.

Where do we go from here? The approvedschedule for this rulemaking is found in theAgency's Regulatory Agenda, which ispublished periodically in the FederalRegister. That schedule calls for the staffto submit a proposed rulemaking package forconsideration to the Commission in early1987. If the Commission decides to proceed,then the proposed regulations will bepublished in the Federal Register forreview and comment later that spring.

In closing, I think that the issue offinancial assurance for cleanup of accidentalrelease caused by non-reactor licensee is animportant priority. I appreciate theopportunity today to bring this work to yourattention, and I will try to answer anyquestions you may have.

ANS Topical Meeting on Radiological Accidents—Perspectires and Emergency Planning

Applicability of Comprehensive Environmental Response,Compensation, and Liability Act of 1980 (CERCLA)

to Releases of Radioactive SubstancesSteven R. Milled

ABSTRACT The Comprehensive Environmental Re-sponse, Compensation, and Liability Act of 1980(CERCLA), commonly called "Superfund," provideda $1.6 billion fund (financed by a tax on petro-chemical feedstocks and crude oil and by generalrevenues) for the cleanup of "releases" of "haz-ardous substances," including "source," "specialnuclear" or "byproduct material," and other ra-dioactive substances, from mostly inactive fa-cilities. The U.S. Environmental Protection A-gency (EPA) is authorized to require private"responsible parties" to clean up releases ofhazardous substances, or EPA, at its option, mayundertake the cleanup with monies from the Fundand recover the monies through civil actionsbrought against responsible parties. CERCLA im-poses criminal penalties for noncompliance withits reporting requirements. This paper will o-verview the key provisions of CERCLA which applyto the cleanup of radioactive materials.

The Comprehensive Environmental Response,Compensation, and Liability Act of 1980 (CER-CLA), Pub. L. No. 96-510, 42 U.S.C. 59601 et_seq.. commonly called "Superfund," provided a$1.6 billion fund (financed by a tax on petro-chemical feedstocks and crude oil and by gener-al revenues) for the cleanup of "releases" of"hazardous substances," including "source,""special nuclear," or "byproduct material," andother radioactive substances from mostly inac-tive facilities. "Federally permitted releases"from active facilities are generally regulatedunder the Resource Conservation and RecoveryAct, (RCRA), 42 U.S.C. 46901 et seq. (1976) andother environmental laws.

The term "release" is extremely broad (i.e.any leaking, spilling, emitting, etc.) and cov-ers the U.S. Department of Energy (DOE) siteswhere hazardous substances are in contact with

the environment (i.e. air, water, soil) evenwithin federal or private industry propertyboundaries.

The term "release" excludes releases of"source," "special nuclear," and "byproduct ma-terial" resulting from a "nuclear incident,"subject to the financial protection require-ments established by the Nuclear Regulatory Com-mission under section 170 of the Atomic EnergyAct of 1954, as amended, 42 U.S.C. 82011, 82210.Also excluded, are releases from uranium milltailings sites being cleaned up by DOE under Ti-tile I of the Uranium Mill Tailings RadiationControl Act of 1978, Pub. L. No. 95-604, 42U.S.C. 87901 et seq.

CERCLA imposes criminal penalties for non-compliance with its reporting requirements. Un-der section 103a of CERCLA, 42 U.S.C. >9603a,any "person" (including federal agencies andprivate companies) "in charge of" an onshore oroffshore facility must report releases of hazar-dous substances in excess of "reportable quanti-ties" to the National Response Center as soon asthe "person" has knowledge of the release, andto publish information concerning such releasesin local newspapers. The failure to promptlyreport such releases can subject responsibleparties, including employees, to criminal penal-ties. Exempted from the reporting requirementsare "federally permitted releases" (i.e. re-leases permitted under other federal environmen-tal laws), which include releases of "source,""special nuclear," or "byproduct material" incompliance with legally enforceable licenses,permits, and orders, issued pursuant to the A-tomic Energy Act of 1954, as amended, 42 U.S.C.$2011 et seq.

Until the U.S. Environmental ProtectionAgency (EPA) promulgates reportable quantitiesfor radioactive materials, the minimum report-able quantity is one pound.

Continuous releases must also be reportedat least annually.

aThe views expressed in this paper are those of the author and do not necessarily represent thoseof the U.S. Department of Energy.

271

272

There are no other affirmative legal re-quirements, for federal agencies or private par-ties, except as specified by EPA.

Owners, operators, generators and trans-porters of hazardous substances are jointly,severally, and strictly liable for the cleanupof all releases.

EPA can either require private parties toclean up releases of hazardous substances or canproceed to undertake the cleanup with moniesfrom the Fund. Any Fund monies so expended byEPA can be recovered through civil actionsbrought against the "responsible parties." Mon-ies from the Fund are available for both emer-gency "removal" actions and long-term "remedial"actions on private sites listed on CERCLA's"National Priorities List."

CERCLA applies to federal agencies proce-durally and substantively, as it does to privateparties EXCEPT: (1) Monies from the Fund aregenerally not available for the cleanup of fed-eral facilities; and (2) EPA cannot sue otherfederal agencies.

CERCLA assigned all response authorities tothe President. By the terms of Exec. Order No.12,316, 46 Fed. Reg. 42,237 (1981), reprinted in42 U.S.C. J9615NT..EPA was generally delegatedall CERCLA response authorities. The U.S. De-partment of Defense was, however, specificallydelegated all CERCLA response authorities atits own sites. Exec. Order No. 12,316 permitsEPA to redclegate its authorities to other fed-eral agencies.

EPA's CERCLA authorities with respect tofederal facilities include: the ability to un-dertake assessments and feasibility studies andconduct emergency response actions, where an im-minent and substantial endangerment to publichealth and safety exists. EPA can issue admin-istrative orders against federal agencies fornoncompliance. Such orders cannot, however, beenforced in court. With respect to nuclear re-leases, however, CERCLA's implementing regula-tion, "The National Contingency Plan," 40 C.F.R.300 et seq. provides that DOE give assistance inresponding to these releases, regardless ofwhether they be on federal or private lands.

Cleanup standards are established by EPAfor federal and private facilities on a case by-case basis. The National Contingency Plan re-quires the application of "applicable" or "rel-evant and appropriate" federal standards and forconsideration to be given to the use of stateand other standards.

There is no state enforcement role underCERCLA and no waiver of sovereign immunitythat would make federal agencies responsible forcomplying with state laws and regulations. How-ever, EPA can enter into cooperative agreementswith affected states for the undertaking ofcleanup actions on private sites. States cansue both federal agencies and private parties fordamages to natural resources and for recovery ofcosts incurred in any cleanup actions the stateshave undertaken for which the federal agenciesand private parties are responsible.

CERCLA is currently up for a five-yearCongressional reauthorization because the au-

thority of the IRS to collect che tax which fi-nances the Fund expired on September 30, 1985.A $7.5 billion reauthorizatlon bill has passedthe Senate (H.R. 2005, as passed by the Senate,99th Cong., 2d Sess., 131 Cong. Rec. S12184(1985). A $10.2 billion reauthorization bill haspassed the House (H.R. 2005, as passed by theHouse, 99th Cong., 2d Sess., 131 Cong. Rec.H11619 et seq. (1985). House and Senate Confer-ees are currently meeting to resolve the differ-ences between the bills. It is expected that aSuperfund reauthorization bill will pass Con-gress by Fall 1986. As of July 1986, the Houseand Senate Conferees have agreed to a $9 billionlevel of funding but have not yet agreed on thefunding sources.

The Administration has indicated that itwould support the reauthorizatlon of Superfundat a level of $5.3 billion. Statements of Ad-ministration Policy have indicated that anybill that passes Congress with a broad based"value added" tax similar to one that passed theSenate or crude oil and petrochemical taxes atthe levels contained in the House passed billmay be vetoed by the President.

In March of 1986, the House and SenateConferees agreed on federal facility provisionswhich imposed specific cleanup requirements andEPA oversight on federal facilities. Underthese provisions, releases of "source," "spe-cial nuclear," or "byproduct material" that arebeing cleaned up under other laws would not needto be docketed by EPA under CERCLA (or presuma-bly otherwise subjected to the federal facil-ities requirements) provided EPA concurs thatthe cleanup undertaken by the federal agency isconsistent with CERCLA.

Other provisions being considered by Houseand Senate conferees affecting radioactive ma-terials Include provisions concerning indoorradon, naturally occurring radon, the temporarystorage of radium wastes currently located atcertain New Jersey sites, a research and devel-opment project on innovative and alternate tech-nologies to characterize radioactive mixed wasteat the Hanford site in Washington, and a "grand-father provision" which is intended to exemptthe radioactivelv contaminated Weldon Springsite in St. Charles County, Missouri, currentlyowned by DOE, from the federal facility require-ments of the new law. The final language onthese provisions will be contained In a confer-ence report which was not available when thispaper was written.

ANS Topical Meeting on Radiological Accidents—Perspectires and Emergency Planning

Price-Anderson—Where We've Been, Where We're GoingIra Dinitz

ABSTRACT The Price-Anderson Act,which became law on September 2, 1957,as part of the Atomic Energy Act of1954, proviJes a system to pay fundsfor claims by members of the publicfor personal injury and propertydamage resulting from a nuclearaccident. The Act as it now operatesentails a two-part insurance systemfor large utility licensees. Thefirst consists of $160 million inprimary nuclear liability insurancepurchased by utilities operatinglarge nuclear power plants. Underthe second part, these utilitiescouild be assessed up to $5 millionper reactor per accident for damagesexceeding $160 million. With 101large power reactors now licensed,the primary and secondary insurancepresently totals $665 million. Thepresent Price-Anderson Act expires onAugust 1, 1987. There are presentlytwo bills H.R. 3653 and S. 1225 beingactively considered by the Congressfor modification and extension ofPrice-Anderson. Hearings have beenheld and the two bills have beenmarked up and reported out by theSenate anc' House oversight committees.

The past three years have been part-icularly active ones for those of usinvolved in the renewal of the Price-Anderson Act. Over the past eight -teen months, various CongressionalCommittees with jurisdiction overnuclear matters have been consideringa number of bills to modify and ex-tend the Price-Anderson Act whichexpires on August 1, 1987. Thisshared jurisdiction over atomicenergy activities has led to differ-ent versions of the same bills beingreported out by oversight committees.I will go into detail on this a bitlater in the paper.

The Price-Anderson Act was orig-inally enacted in 1957 as a ten-yearlaw, and has been renewed twice withsimilar ten-year expiration dates.The Act had two main objectives: toensure that the public would be com-pensated if an accident occurred at anuclear facility, and to set a limiton the liability of private industryin order to remove a major deterrentto private participation in the deve-lopment of nuc'iear energy.

Price-Anderson provides e systemto pay funds for claims by members ofthe public for personal injury andproperty damage resulting from anuclear accident. The Act requiresutility holders of licenses of largecommercial nuclear power plants toprovide proof to the MRC that theyhave the maximum amount of privatenuclear liability insurance--general-1 ly referred to as financial protect-ion -- that is available. Thatfinancial protection, currently $665million, consists of a primary layerof nuclear liability insurance of$160 million and a secondary retro-spective premium insurance layer.This secondary layer works in thefollowing way. In the event of anuclear accident causing damagesexceeding $160 million, the licenseesof each commercial nuclear powerplant would be assessed a proratedshare of damages in excess of theprimary insurance layer up to $5million per reactor per incident.With 101 large commercial reactorsunder this system today, the sec-ondary layer totals up to $505million.

The Price-Anderson Act author-izes the Commission to enter intoindemnity agreements with reactorlicensees. These agreements specifythe amount of financial protection,if any, required of licensees andthe obligation of the federal Gover-nment to provide funds when a nuclear

273

274

accident exhausts private liabilityinsurance or when no private liabil-ity insurance is required. Thisgovernment obligation to provide fundsis called "government indemnity."

The Price-Anderson Act places aceiling on the total amount of publicliability in an accident. Thisceiling or "limit of liability," forlarge commercial nuclear power plantlicensees is currently tied to themaximum amount of insurance availablethrough private sources. For manyyears, the limit of liability was$560 million. In the 1975 amendmentsto the Price-Anderr.on Act, Congressprovided that the limitation on liab-ility would grow, once the totalprotection of the primary andsecondary layers of insurance reached$560 million. In November 1982, the$560 million level was reached and thegovernment's indemnity was essentiallyeliminated for large reactors. Thepresent limit of $665 million willcontinue to increase in increments of$5 million for each new commercialreactor licensed to operate. In the1975 amendments Congress also explic-itly provided that "in the event of anuclear incident involving damages inexcess of the amount of aggregateliability, the Congress will thorough-ly review the particular incident andwill take whatever action is deemednecessary and appropriate to protectthe public from the consequences of adisaster of such magnitude."

Both the private nuclear liabil-ity insurance polices and the indem-nity agreement that the Commissionenters into with licensees are"omnibus" in nature. That is, inrecognition of the requirement of theAct to provide coverage for the lice-nsee and other "persons indemnified,"the policies cover not only theutility licensees but also any otherperson liable for the accident. Thescope of Price-Anderson coverageincludes any accident in the courseof transportation of nuclear fuel tothe reactor site, in the storage ofnuclear fuel at the site, in theoperation of the reactor, includingdischarge of radioactive effluents,in the storage of nuclear fuel andnuclear waste at the reactor site, andin the transportation of nuclearfuel and nuclear waste from the reac-tor.

The insurance industry formedtwo insurance pools to provide nuclearliability insurance capacity to the ;utility industry at the time of pass-'age of the Price-Anderson Act. Onepool, American Nuclear Insurers iscomposed of investor-owned stockinsurance companies. The other pool,

Mutual Atomic Energy Reinsurance Poolis made up of policy-owned mutualinsurance companies. About half ofeach pool's total liability capacitycomes from foreign sources. Membercompanies comprising the pools decideindependently the amount of capacitythey wish to commit to nuclear risks.The Facility Form oolicy is for theowners and operators of nuc lear facil-ities and when provided as financialprotection is a formal part of thePrice-Anderson system. The poolsalso participate in the secondarypart of financial protection requiredby Price-Anderson by issuing policiesthat set forth the terms, conditions,end obligations of the parties tocover the secondary part of insuranceprotection. The pools are authorizedto charge the utilities and then payout the collected premium funds onbehalf of the utility which had theaccident.

While Congress placed particularemphasis in the 1957 enactment ofPrice-Anderson on the basic princi-ples of protection of the public andencouragement of the industry, otherprinciples were considered. Congressalso recognized that the substitutionof government indemnity for privateinsurance should be a short-termexpedient—that is, one limited atleast Initially to ten years. Thesubstitution of private insurance forgovernment indemnity was statutorilyrecognized in the 1965 extension ofPrice-Anderson by stipulating thatgovernment indemnity would be reducedto the degree that financial protec-tion was provided above $60 million.The process of shifting the burden ofnuclear liability risk1; from theGovernment to the nuclear industryculminated in the 1975 amendments tothe Act which established a secondaryretrospective premium insurance layerto provide funds for excess insuranceafter an accident exhausted theprimary insurance layer. As mention-ed, as the secondary layer continuedto increase, government indemnitywas gradually phased out and inNovember 1982, was eliminated.

Prior to 1977, the Joint Comm-ittee on Atomic Energy had exclusivejurisdiction over atomic energyissues. Upon the dissolution of theJCAE, however, a number of differentHouse and Senate committees weregranted jurisdictional authorityover atomic energy. Regarding themost recent Price-Anderson extension,this shared authority has meanthearings on bills to modify Price-Anderson before four subcommitteeswith bill markups by five subcomm-ittees and five full committees.This situation contrasts quite dram-

275

atically with the 1975 extension ofPrice-Anderson Act. At that time,only one bill was introduced tomodify and extend Price-Anderson withthat bill marked up and reported outonly by the Joint Committee.

Nine bills have been introducedin the 99th Congress to modify andextend Price-Anderson. Only two ofthe bills, H.R. 3653, introduced byCongressman Udall, and S. 1225 intro-duced by Senators McClure and Simpsonhave been reported out. H.R. 3653was reported out by the House Inter-ior and Insular Affairs, Energy andCommerce and Science and TechnologyCommittees. S. 1225 was reported outby the Senate Energy and NaturalResources and Environment and PublicWorks Committees.

Not surprisingly there are signi-ficant differences between H.R. 3553and S. 1225. There are also differ-ences, however, in the differentversions of each of these two billsas reported out by the respectiveoversight committees. I would nowlike to discuss some of the majorprovisions of H.R. 3653 as well aspoint out some of the differences inthe two versions of the bill. I willthen follow with a similar discussionof S. 1225.

First, as reported out by theHouse Interior Committee, H.R. 3653increases the funds available to paypublic liability claims arising out ofnuclear accident assuming the opera-tion of 101 large commercial nuclearreactors from $665 million to approx-imately $6.5 billion. This would meanthat the deferred retrospective prem-ium assessed against large commercialpower reactors would increase fromS5 million per reactor per accident to$63 million per reactor per accidentwith not more than $10 million perreactor collected in any one year.Second, large commercial power reac-tors would maintain not less than$200 million of primary financialprotection. Third, in order to paypublic liability claims arising outof a nuclear accident on a timelybasis, the NRC would be able to borr-ow funds from the Treasury or requestCongressional appropriation of addi-tional funds to pay the entire aggr-egate maximum retrospective premiumson behalf of'iti lity licensees.These licensees would then reimbursethe Commission over a period of years.Fourth, a futuro Congress could enacta revenue measure to recover fromreactor licensees, funds paid out incompensating victims of a nuclearaccident above the liability limit.Fifth, a study commission would becreated to examine and report toCongress on alternative measures ofcompensating victims of a nuclear

accident above the liability limit.Sixth, primary insurance and second-ary deferred retrospective premiumswould not be used to pay the costsof investigating, settling and de-fending claims for damages arisingout of a nuclear accident. Seventh,the costs of a precautionary evacu-ation would be compensable if liabil-ity exists under state tort law.Eighth, the 20-year statue of limita-tions would be eliminated and a three-year from discovery rule would besubstituted. This rule would recog-nize any claim filed within threeyears of discovery of damages.Ninth, DOE contractor liabilityprotection would be equivalent tothe financial protection made avail-able for NRC reactor licensees.

There are two major differencesin H.R. 3653 as reported out by theInterior and Insular Affairs, Energyand Commerce and Science and Technol-ogy Committees. First, as I mentio-ed, in the version of H.R. 3653reported out by the Interior andInsular Affairs Committee, accidentsinvolving high level radioactivewaste would not be subject to alimitation of liability. This prov-ision was deleted by the Science andTechnology Committee in its markupso that high level radioactive wasteactivities would be subject to the samelimitations of liability as otherDOE activities. The E n e m y andCommerce Committee further complicat-ed this issue by exempting all DOEcontractor activities from a limita-tion of liability. Second, theScience and Technology Committee re-moved the NRC from the three-memberpanel that would be convened to de-termine the reasonableness of pre-cautionary evacuations for incidentsinvolving Department of Energycontractors.

Turning now to the Senate sS. 1225 as reported out by both theSenate Energy and Natural Resourcesand Environment and Public WorksCommittees provides among otherthings for Department of Energy in-demnification of its contractors tothe same level as that required ofNRC licensees; confirms that theDepartment of Energy must indemnifycontractors involved with its nuclearwaste programs; provides for federalcompensation of claims arising fromincidents involving nuclear materialthat has been illegally obtainedfrom unknown sources; and allows forthe reimbursement of the cost ofprecautionary evacuations resultingfrom incidents involving DOE contra-ctors.

Two of the provisions in theversion of S. 1225 as reported outby the Energy and Natural Resources

276

Committee differ from the versionreported out by the Environment andPublic Works Committee. F'lrst, inthe Energy Committee version, theNuclear Regulatory Commission isdirected to increase by rulemakingthe present $5 million retrospectivepremium charged each reactor follow-ing a nuclear incident to a premiumof not less than $15 million normore than $20 million. The Commiss-ion would be required to annuallyadjust this premium based on theeffect of inflation.Under the version of S. 1225 reportedout by the Environment and PublicWorks Committee, however, the maximumdeferred premium that could be assess-ed against a nuclear reactor per inci-dent would be $60 million with no morethan $12 million collected in any oneyear. Second, S. 12 25 as reported outby the Energy Committee deletes the20-year statue of limitations and

adopts a 3-year from discovery rule.In the Environment Curnmittee versionof S.1225, a fixed statute of limita-tions provision is retained but incr-eased to 30-years.

It is still too eariy to predictwhether H.R. 3653 and S. 1225 will bebrought to the floor of both chambersduring the last weeks of the 99thCongress. If this does not occur,it would still be possible for thebills to be considered in a lame ducksession if such a session were to beconvened. If Congress does not enactPrice-Anderson extension legislationthis year, however, new bills toextend and modify Price-Anderson wouldhave to be introduced in the nextsession of Congress. If a new roundof hearings and markups were held, itis difficult to predict whether leg-islation to extend Price-Andersoncould be enacted by the August 1, 1987expiration date of the present Act.

Section V

Institutional Issues

Chairman: Robert WilkersonFederal Emergency Management Agency

ANS Topical Meeting OB Radiological Accidents—Perspectives and Eidergency Plamiig

Protective Action Guides: Rationale, Interpretation, and Status

Joe E. Logsdon

ABSTRACT: The Environmental Protection Agency(EPA) is developing Protective Action Guides(PAGs) Eor responding to radiological emergen-cies. The existing interim guidance for plumeexposure pathways is 0.03. to 0.05 Sv (1 to 5rems) whole body dose equivalent and 0.05 to 0.25Sv (5 to 25 rems) committed dose equivalent tothe thyroid from a 2 to 4 day exposure. InterimPAGs for ingestion exposure pathways are 0.015 to0.15 Sv (1.5 to 15 reins) committed dose equiva-lent to the thyroid and 0.005 to 0.05 Sv (0.5 to5 rems) committed dose equivalent to the wholebody or any other organ from a one-year exposure.Draft Relocation PAGs have been proposed as arange of 0.01 to 0.05 Sv (1 to 5 rems) committedeffective dose equivalent from the first yearexposure to deposited radioactive material fromall exposure pathways except ingestion of foodand water.

The guidance (i.e., PAGs for plume, inges-tion, and relocation; limits for emergencyworkers and other persons entering restrictedzones; and guidance for development of recoverycriteria) is reviewed with regard to status,values, rationale, interpretation and implemen-tation.

I. BACKGROUND

The Environmental Protection Agency (EPA) isresponsible for developing radiation protectionguidance for emergency responses to nuclearaccidents. This guidance is published in interimform in a document entitled "Manual of ProtectiveAction Guides and Protective Actions for NuclearIncidents" (PAG Manual) (EPA, 1975). After allPAGs are developed and some experience is gainedin their application, we plan to revise them asnecessary and publish them in final form.

In the ea.ly 1960's the Federal RadiationCouncil fFRC) developed PAGs for radioactivity infood due to fall- out from nuclear weapons tests(FRC, 1964 and 1965). These were not appropriatefor the early phase of a nuclear accident, whenthe concern is direct exposure to an airborneplurae of radio-active materials, and in 1975 EPAdeveloped recommended PAGs for airborne plumes.

Because of the urgent need to provide guidanceto the States, EPA elected to issue these PAGs onan interim basis, with the intent of accumulatingexperience in their application before final prom-ulgation. A similar development process hascontinued for subsequent PAGs that apply to theintermediate phase. The Food and Drug Adminis-tration (FDA), in coopsration with EPA, developedrecommendations on radioactivity in food andanimal feed and published them in October 1982.The interim PAGs for plumes and recommendationsfor radioactivity in food and animal feeds havesince been incorporated into State radiologicalemergency response plans (RERPs).

As a result of experience gained over aperiod of years from developing and testing RERPs,conducting training programs, and the accident atThree Mile Island, the need for some revisions andadditions became apparent. It also became apparentthat additional PAGs are required for exposure todeposited and resuspended radioactive materialsand that guidance is needed for developingcriteria for cleanup during recovery.

Prior to the accident at Chernobyl, revisedplume PAGs, and ingestion PAGs incorporating theFDA's recommendations for food with EPA's guidancefor drinking water, had been developed andreviewed by Federal and State agencies and thenuclear industry. Draft guidance had also beendeveloped for exposure to deposited and resus-pended radioactive materials and was in theprocess of being developed on how to establishlong-term radiation protection criteria forrecovery. As a result of current experience beinggained from Chernobyl and participation in thedevelopment of guidance for the InternationalAtomic Energy Agency, some further needs have beenidentified. These include simplification of thedecision process for protective actions and im-proved, more explicit response levels for radio-activity in food.

II. STATUS

A. Plume PAGsPlume PAGs were issued in 1975 for whole body

external exposure and thyroid inhalation exposure.A proposed revision to include PAGs for otherorgans for inhalation pathways and the technicalrationale for selection of the plume PAG values

279

280

was sent in 1985 for comment by State and Federalagencies and the nuclear industry.

The most important changes based on thosecomments are:

1. Revised units are being proposed from"dose equivalent" to "committed effective doseequivalent" based on International Commission onRadiation Protection (ICRP, 1979) dosimetry. Therevised PAGs would be a range of 1 to 5 rei s com-mitted effective dose equivalent subject to aspecial limitation for thyroid dose. This changewould not require revisions to existing emergencyresponse plans for nuclear power plants, as ex-plained later.

2. The urgency of implementing earlyprotective actions based on plant conditionsinstead of dose calculations at the time of theaccident ;.s given greater emphasis.

3. Simplified dose calculation proce-dures using nomograms are removed from theguidance, and users are encouraged to employcomputer technology in their emergency plans fordose calculations. The simplified procedures areretained as a backup in Appendix D to the PAGmanual and for training programs on accidentassessment.

4. The cost analyses supporting PAGvalue selection are revised to eliminate planningcosts as a consideration.

5. BEIR-3 (NAS, 1980) analyses are usedas the primary basis for evaluation of risk oflong-term health effects.

6. The 75-rem whole body dose limit forlifesaving activities was eliminated, and exposurefor such activities at projected doses above 25rems are recommended only for volunteers aware ofthe risk involved.

Transraittal of the plume PAGs to the FederalRadiological Preparedness Coordinating Committee(FRPCC) for their concurrence is being delayedpending incorporation of some guidance on derivedresponse levels under development by EPA incooperation with the International Atomic EnergyAgency.

B. Ingestion PAGsRecommendations regarding radioactivity in

food and animal feed were developed by FDA incooperation with EPA and published in 1982 (FDA,1982). EPA adopted these recommendations (1.5 to15 rems to the thyroid and 0.5 to 5 rems to thewhole body or other organs) aa interim PAGs andadded drinking water as a category of food. ThesePAGs, along with implementation guidance, weretransmitted for review by State and Federalagencies and the nuclear industry in 1985. Theonly major changes required as a result of thecomments relate to simplification of the doseprojection process and the inclusion of examplecalculations.

As a result of the Relocation TabletopExercise in December 1985, it became apparent that,while ingestion PAGs are adequate for applicationto situations in which supply of food is an emer-

gency need, an emergency for food and drinkingwater would not exist unless the implementation ofprotective actions would create a food or watershortage. Stated another way; it would not be rea-sonable to expect persons to consume food or waterthat is contaminated to the level of the PAGs ifother noncontaminated food or water is availableat a cost that is justifiable on the basis of riskavoided. On the other hand, there should be somelevel of dose above which cost should not be aconsideration. This dose is logically the upperrange of the PAG (the "emergency PAG"). EPA iscurrently evaluating the need to developadditional guidance for non-emergency situationsfollowing accidental releases. This will requireadditional peer review before transmittal to theFRPCC for their concurrence. Derived dose con-version factors for all major radionuclides andfood types are also being developed.

C. Relocation PAGsDraft relocation PAGs have been transmitted

for review and comment by State and Federalagencies and the nuclear industry. They arerecommended as a range of 1 to 5 rems committedeffective dose equivalent to individuals in thegeneral public from exposure to deposited radio-active materials via all exposure pathways exceptingestion of food and water. The exposure and/orintake period to be assumed for dose projectionpurposes is one year and the guidance is applicablefor up to one year after the accident. If specialradiation protection criteria for recovery are notdeveloped within one year, the existing RadiationProtection Guides (FRC, 1961) for individuals inthe population will apply beginning with the secondyear. Persons temporarily reentering the restrictedzone established on the basis of the relocationPAGs would have their exposure controlled accordingto established occupational exposure limits.

After resolution of comments from those pre-sently reviewing the draft guidance, the relocationPAGs will be sent to the FRPCC for their concur-rence prior to their transmittal for use as interimguidance.

D. Recovery Guidance

The selection of radiation protectioncriteria for recovery must consider the existingenvironmental, social, and economic circumstances.Since the public would already be protected by

way of relocation, recovery becomes primarily aneconomic issue. We do not plan to establishgenerally applicable numerical recovery criteria.Instead we are developing procedures that may beused after an accident to conduct a cost/riskanalysis of the affected area. This analysiswould provide a primary basis for selection ofradiation protection criteria for recovery.

Draft technical reports dealing with cost/risk analyses and alternative exposure pathwaysare under development. These will form the basisfor procedures that can be used in the developmentof radiation protection criteria for recovery.

III. RATIONALE

The following three principles represent theprimary considerations in establishing values forthe PAGs:

281

1. Acute radiation effects (thoseeffects that would be observaBle within a shortperiod of time, approximately 60 to 90 days,following exposure and that have a dose thresholdbelow which such effects would not occur) are tobe prevented.

2. Delayed effects (primarily cancer andlife span shortening) must be minimal and limitedto levels that EPA judges to be adequately protec-tive of the public health, and

3. The guidance must be practicable interms of cost and risk incurred in implementing it.

We have concluded that PAGs in the range of1 to 5 reins meet these three principles for evac-uation to avoid short term (two to four days)exposure to the plume and deposited materials andalso to avoid longer term (up to one year) directexposure to deposited materials. Values for pro-tective actions for ingestion pathways and forsheltering should generally be lower.

No lower limit has been specified for shortterm (a few hours) sheltering to avoid plumeexposure since the cost and risk are small andgenerally indeterminate. A lower limit for shel-tering should be established by local planners orresponders based on considerations other than dose(e.g., communication needs, social concern, andpolitical boundaries).

The lower limit of 1.5 rem to the thyroid and0.5 rem to whole body or other organs from emer-gency ingestion of food and water are justifiedprimarily on the basis of comparative risk accept-able to the public. Lower values can be justifiedon the basis of cost/risk analyses for some com-binations of radionuctides and foods, but perhapsnot for others. The possibility of establishingingestion exposure criteria based on cost/riskanalyses for response under conditions not con-stituting an emergency is being evaluated, asmentioned previously.

IV. INTERPRETATION

PAGs are stated as a range of values to allowfor informed judgment, taking local constraintsand conditions into account. They are tools to beused in planning and decision aids for use inactual response situations.

Although there are some minor differences inthe interpretation of the ranges for the threecategories of PAGs, generally speaking, the lowerend of the range is the projected dose it whichplans should be made to implement simple orlow-impact protective actions (sheltering, use ofstored cattle feed, and simple decontaminationtechniques) and to consider implementation ofhigh-impact actions (evacuation, food restriction,and relocation). The higher impact actions shouldbe implemented at the lower end of the range unlessconstraints (e.g., weather, costs, and competingemergencies) make their implementation impracti-cable. Low-impact actions should be consideredand implemented, where practicable, at levelsbelow the lower range of the PAGs. No level isestablished below which low-impact protectiveactions should not be implemented. This decisionis left to local authorities based on existing

conditions and the generally accepted publichealth practice of limiting radiation exposure toas low as reasonably achievable levels.

The upper end of the range is the projecteddose above which decisions should usually be madeto implement high impact protective actions with-out consideration of cost or difficulty. Protec-tive actions should not be implemented in anycase, however, where the health risk associatedwith the protective action will equal or exceedthe health risk from the radiation that would beavoided.

V. IMPLEMENTATION

The technical portion of PAG implementationinvolves a determination of the projected dose asa function of location in the environment, theselection of the actual PAG values to be used, andthe identification of areas where the selectedPAGs will be exceeded. (Realistically, dose pro-jections in the early phase of an accident based onradiological information are not likely to beavailable at the time of protective action deci-sions. Such decisions are expected to be made inmost cases on the basis of dose estimates relatedto existing or forecasted plant and/or environ-mental conditions.)

Dose conversion factors and example doseprojection methods, as well as guidance forselecting appropriate PAG values, are provided inthe implementation portions of the PAG manual.Users are encouraged to adopt existing computertechniques or to develop their own to supplementthe simplified dose projection methods presented.

Since "committed effective dose equivalent" isbeing adopted as the unit for plume exposure path-ways and for direct exposure to deposited radio-active materials, the thyroid, which exhibits ahigh ratio of curable to fatal cancers, requiresspecial consideration. This is because effectivedose considers only the risk of fatal cancers. Toprovide a simple method of accounting for the highrisk of curable thyroid cancers, dose conversionfactors for inhalation are derived from the "annuallimits for intake" (ALIs) from ICRP-30 (ICRP,1969). Because this ALI is bounded by the non-stochastic limit of 50 rems, this results in adose to the thyroid that is one third that calcu-lated on the basis of effective dose. Althoughthe PAGs for pathways other than ingestion of foodand water are expressed in terms of effective dose,the implementation guidance results in projecteddoses that are higher than the true effective dosein cases where thyroid inhalation dose is involved.

Most States already have RERPs based on wholebody and thyroid dose equivalent PAGs, and therewas concern that the change to effective doseequivalent would force major changes in existingplans. However, if one uses the dose projectiontechniques provided with the previously issueddose-equivalent plume PAGs and bases decisions onthose projected doses and PAGs, calculations showthat the effective dose PAGs will be satisfied formost postulated reactor accidents. In those casesin which the effective dose PAGs would not besatisfied, calculations indicated that thr effec-tive dose PAG would be exceeded only sligntly (10percent). Therefore, it is not considered neces-sary to convert RERPs for nuclear reactor accidents

282

to effective dose in cases when the previouslyissued FAGs and calculational techniques are used.

During the intermediate phase of the accident,monitoring data should be available particularlyfor decisions to return evacuees or to relocateindividuals based on relocation FAGs. The draftnow undergoing peer review recommends relocationPAGs in the range of 1 to 5 reins committed effec-tive dose equivalent from the first year exposureto and intake of deposited radioactive materials.It also provides guidance for selecting a valuefrom within the range. In general, the lower endof the PAG range will be bounded by the accepta-bility of costs for continued relocation relativeto the risk avoided. A range of costs for riskavoidance that EPA has used as one factor in itsdevelopment of environmental standards is recom-mended as a basis. The range of costs presentlyused is $400,000 to $7,000,000 per statisticallife saved for implementation of an environmentalstandard. Based on these criteria, a lower boundon the PAG range of less than 1 rem is notexpected to be cost-effective.

The upper range relocation PAG is limited bythe lower of (a) 5 rems in the first year or, (b)by a dose that could be reduced during the firstyear so that residents would receive no more thanthe RPG for individuals (FRC, 1961) during thesecond year. Based on lessons learned at theRelocation Tabletop Exercise, it was concludedthat a gradual return to contaminated areas shouldbe allowed. Return for occupancy at doses greaterthan the lower range of the relocation PAG shouldbe permitted only after completion of experimentsshowing probable effectiveness of dose reductionefforts during the first year.

Our intent is to allow for flexibility in theimplementation of PAGs. However, above the upperPAG range there is significant risk to the exposedpopulation; and protective actions should normallybe considered mandatory, recognizing that when anaccident actually occurs, unforseen conditions or

constraints may require overriding professionaljudgment to protect the public.

VI. REFERENCESEnvironmental Protection Agency, 1975 revised

1980, Manual of Protective Action Guides andProtective Actions for Nuclear Incidents.Environmental Protection Agency, Office ofRadiation Programs. Washington, D. C. 20460.

Federal Radiation Council, 1961, RadiationProtection Guidance for Federal Agencies.Federal Register, September 26, 1961.

Federal Radiation Council, 1964, RadiationProtection Guidance for Federal Agencies.Federal Register, Volume 29, pp. 12056-7,August 22, 1964.

Federal Radiation Council, 1965, RadiationProtection Guidance for Federal Agencies.Federal Register, Volume 30, pp. 6953-5 (May22, 1965).

Food and Drug Administration, 1982, AccidentalRadioactive Contamination of Human Food andAnimal Feeds; Recommendations for State andLocal Agencies. Federal Register, October 22,1982.

International Commission on Radiation Protection,1979. Limits for Intakes of Radionuclides byWorkers, ICRP Publication No. 30, PergaraonPress, Oxford.

National Academy of Sciences, 1980, "The Effectson Populations of Exposure to Low Levels ofIonizing Radiation: 1980." Reports of theCommittee on the Biological Effects ofIonizing Radiations. National Academy Press,Washington, D. C.

ANS Topical Meeting on Radiological Accidents—Perspectives and Emergency PUnning

The Federal Radiological Emergency Response Plan

Vernon Adler

Abstract - This paper describes the developmentof the Federal Radiological Emergency ResponsePlan and its application.

I. GENESIS OF THE FEDERAL PLAN (FRERP)

Work began on the Federal RadiologicalEmergency Response Plan (FRERP) in 1980 with aneffort to bring together Federal resources into aconsolidated plan for response to a peacetimenuclear emergency. Plan development became apriority subject following the emergency in March1979 at Three Mile Island.

A major recommendation of the KemenyCommission was that there should be centrali-zation of emergency planning and response in asingle agency at the Federal level with closecoordination between it and State and localagencies, and that there should be a singleagency that has the responsibility both forassuring that adequate planning takes place andfor taking charge of the response to the emer-gency.

Additionally, P.L. 96-295 (June 30, 1980)directed the President to prepare and publish aNational Contingency Plan. Three months later(September 29, 1980), E.O. 12241 delegated thisauthority to the Federal Emergency ManagementAgency (FEMA).

FEMA published the National RadiologicalEmergency Preparedness/Response Plan forCommercial Nuclear Power Plants (Master Plan) inDecember 1980, which applied to nuclear powerstation radiological emergencies.

II. DEVELOPMENT OF THE FEDERAL PLAN

The interagency Federal RadiologicalPreparedness Coordinating Committee (FRPCC)undertook development of a more detailed plan,following Master Plan publication. This FRPCCeffort, led by FEMA, involved 12 Federal depart-ments and agencies.- The scope of the Committee'seffort was expanded to include all major peace-time radiological emergencies; in April 1983,draft planning guidance was published by FEMA foruse by other Federal agencies.

The draft Federal Plan (FRERP) was publishedfor comment in January 1984 and served as thebasis for the first Federal Field Exercise

(FFE-1) in March 1984 at St. Lucie, Florida. Theoverall concept of operations for the FederalPlan was tested first in the tabletop exercise(HIEX-82).

The St. Lucie Exercise (FFE-1), establishedthe viability of the draft Federal Plan.Exercise of the Federal Plan was conducted aspart of a full field Radiological EmergencyPreparedness (REP) exercise which involved theutility, State and local government.Approximately one thousand people participated inthe FFE-1 Florida exercise; FEMA's cost ofplanning for and conducting the exercise exceeded$0.5 million. FFE-2 is to be the second in aseries of planned triennial field exercises ofthe Federal response to a major radiologicalemergency. It will simulate an accident at theZion Nuclear Power Plant and will involvecoordination of Federal efforts, the State ofIllinois and Wisconsin and the counties of Lakeand Kenosha.

The Federal Plan became fully operationalfollowing concurrence of all twelve participatingFederal agencies and its publication in theFederal Register on November 8, 1985.

III. SCOPE OF THE FEDERAL PLAN

The Plan covers a wide range of peacetimenuclear emergencies requiring Federal response insupport of State and local government requestsfor Federal assistance. It is primarily anoffsite plan, applicable within the United Statesand its territories to the Federal response inthe event of a nuclear reactor, nuclear weapon,radiological transportation, or other significantaccident involving radiological material.

IV. PLANNING ASSUMPTIONS FOR THE FEDERAL PLAN

State and local governments have primaryresponsibility for protection of public healthand safety. Consequently the Federal governmentwill respond when requested by a state or whenrequired to fulfill statutory responsibilities.

The Federal Plan creates no new Federalauthorities. It provides for the coordinatedimplementation of response by Federal agenciesoperating under each agency's existing author-ities.

283

284

V. CONCEPT OF OPERATIONS

The FRERP provides for sharing the "leadagency" responsibilities for responding to amajor radiological emergency, based on variousstatutory responsibilities, authorities, andcapabilities. The Cognizant Federal Agency (CFA)is the agency that owns, authorizes, regulates oris other wise responsible for the facility onradiological activity causing the emergency, andthat has authority to take action on site. Forexample, the CFA, which is the Nuclear RegulatoryCommission in the case of a commercial nuclearpower station, operates from the utility'sEmergency Operations Facility (EOF), as a generalrule. FEMA coordinates among Federal agenciesand between the Federal government and State/local governments; its operational center is theFederal Response Center (FRC) through whichFederal resources are coordinated in response toState requests for assistance.

FEMA, as principal coordinator for theFederal family, is attentive to numerous inter-faces, including coordination among the Federalagencies; coordination between Federal andState/local governments; coordination of offsitewith onsite efforts; and coordination betweenWashington, D.C., and onscene (between the ."?EMAEmergency Support and Response Teams).

The Department of Energy applies itstechnical expertise and resources for offsiteradiological monitoring and assessment; DOEoperates from the Federal Radiological Monitoringand Assessment Center (FRMAC), which is led by aDOE Offsite Technical Director (OSTD).

VI. FEDERAL AGENCY RESPONSE FUNCTIONS UNDER THEFEDERAL PLAN

The Cognizant Federal Agency (CFA) is eitherthe Nuclear Regulatory Commission, Department ofDefense, or Department of Energy. It is alsoprobable, in circumstances where tha organizationor entity responsible for the source of radio-activity is unknown or not of U.S. origin, suchas in the case of nuclear satellite reentry, thatFEMA would coordinate the Federal response usingthe technical assistance of U.S. Federal agencieslike DOE, NRC or EPA.

DOE may also serve as the CFA in weapons andcertain transportation accidents. The NuclearRegulatory Commission is the CFA in a nuclearpower station accident. The Department ofDefense can be the CFA in nuclear weaponsaccidents and otherwise provides support activ-ities. '

The other Federal agencies which arecurrently responders under the Federal Planinclude:

a. Health & Human Services - provideshealth care;

b. Department of Commerce (National WeatherService)-meteorological information;

c. Housing and Urban Development - housingsupport;

d. Department of Interior - public lands,ground waters;

e. Department of Transportation - coordi-nation and support of transportation;

f. Environmental Protection Agency - long-term radiological monitoring;

g. National Communications System-communications support if required.

VII. EXERCISES OF THE FEDERAL PLAN AND LESSONSLEARNED

The dozen Federal agencies, involved in theFederal radiological emergency response exerciseprogram, are cooperating in a joint effort toconduct major exercises like the Federal FieldExercise of 1984, on a triennial schedule withFederal tabletop exercises held in the yearfollowing the field exercises; consequently thesecond Federal Field Exercise (FFE-2) will beconducted in 1987 with a tabletop effort plannedfor the following year.

The Federal Field Exercise (FFE-1) in March1984, and the Relocation Tabletop Exercise inDecember 1985, provided "lessons-learned" inseveral areas. The mock accident in FFE-1followed a scenario of much greater severity thanwas experienced at Three Mile Island. Eighthours into the FFE-1 emergency, the State ofFlorida requested Federal Assistance. MajorFederal assistance to the State and localgovernment was provided through the FederalResponse Center. The FFE-1 included testing of40 interfaces, grouped into seven broad areas forevaluation including notification, activation anddeployment, onsite/offsite activity coordination,public information coordination, White House andCongressional coordination. The FRERP FieldExercise Evaluation Report of June 1984 documentslessons learned as the result of the firstFederal Field Exercise (FFE-1) of the FederalPlan.

Reentry and recovery were not examined inFFE-1, and so a relocation tabletop 'exercise, theRTE, was developed and conducted in December 1985at FEMA's Emergency Management Institute inEmmitsburg, Maryland. Lessons learned aredocumented by the March 12, 1986, RTE After-Action Report. For example, it became clear fromthis exercise that the concept of a .JointInformation Center (JIC) needs further definitionof goals and guidelines, and that substantialbenefits of information exchange were obtainedthrough liaison officers to response centers;this should be encouraged.

An American Nuclear Insurers (ANI) repre-sentative in the Federal Response Center providesthe Senior FEMA Official rapid access to impor-tant data concerning insurance industry responseinitiatives. It became clear through the RTEthat, at present, Federal agencies may or may notchoose to implement their respective statutoryauthorities through the onscene organization,that is, through the SFO or CFA. They may chooseinstead to implement these authorities throughtheir own headquarters or regional offices.

This and other important policy questionswhich came into sharper focus in the exercise arebeing examined through the FEMA-led FederalResponse Subcommittee of the Federal RadiologicalPreparedness Coordinating Committee. Answers tothese questions and development of appropriateprocedures among Federal agencies in support ofthe Federal Plan are an ongoing effort.

285

The second Federal Field Exercise (FFE-2)will be held June 23-25, 1987, at the Zion Plant.The 10-mile Emergency Planning Zone for Zionincludes both the State of Illinois andWisconsin. This exercise will be part of a fullREP exercise by these States and the CommonwealthEdison utility. The FFE-2 will inclu-le acomplete accident scenario, beginning on site andextending to an including key post-accidentrecovery considerations.

ANS Topical Meetiog on Radiological Accidents-Perspectives and Emergency Planning

Role of the Federal Radiological Monitoring and Assessment Center(FRMAC) Following a Radiological Accident

John F. Doyle III

ABSTRACT The Federal Radiological EmergencyResponse Plan (FRERP) calls for the Departmentof Energy to establish a Federal RadiologicalMonitoring and Assessment Center (FRMAC) immedi-ately following a major radiological accident tocoordinate all federal off-site monitoringefforts in support of the State and the Cogni-zant Federal Agency (CFA) for the facility ormaterial involved in the accident. Some acci-dents are potentially very complex and mayrequire hundreds of radiation specialists toensure immediate protection of the public andworkers in the area, and to identify prioritiesfor the Environmental Protection Agency (EPA)long-term efforts once the immediate protectiveactions have been carried out. The FRMAC pro-vides a working environment with today's hightechnology tools (i.e., communication, comput-ers, management procedures, etc.) to assure thatthe State and CFA decision makers have the bestpossible information in a timely manner on whichto act. The FRMAC planners also recognize anunderlying responsibility to continuously docu-ment such operations in order to provide theState, the CFA, and the EPA the technical infor-mation they will require for long term assess-ments. In addition, it is fully recognized thatinformation collected and actions taken by theFRMAC will be subjected to the same scrutiny asother parts of the accident and the overallresponse.

I. BACKGROUND

At the time of the Three Mile Island acci-dent in March 1979, the U.S. experience in re-sponding to major nuclear accidents was limitedto a few earlier accidents involving nuclearweapons and the joint Canadian - U.S. responseto the reentry of the 40KW nuclear reactor onthe COSMOS 954 satellite over the NorthwestTerritory of Canada in January of 1978. Since1960, the Radiological Assistance Program (RAP),consisting of a dozen or so agencies in the

U.S., had been assisting the State and localgovernments with routine responses to trans-portation accidents involving trucks, trains,and aircraft, as well as problems in industryand hospitals with lost sources, spills, or wetpackages. The typical Radiological AssistanceTeams, consisting of five to ten personneltrained in Health Physics and equipped withappropriate monitoring equipment and protectiveclothing, are well prepared to handle thesetypes of problems. The Canadian experience,however, taught us that a nuclear accidentinvolving dispersal of fission product contami-nation of potentially high levels over largegeographic areas presents major response manage-ment problems. Further, it was thoroughly dem-onstrated in the COSMOS 954 response that aerialradiation measurement technology would be ex-tremely valuable for rapidly defining theaffected areas.

Therefore, when the Three Mile Island acci-dent occurred, DOE immediately responded toalmost simultaneous requests for assistance fromboth the State and NRC with both RAP teams andthe Aerial Measuring System capability. ByThursday, the second day of the accident, DOEon-site management offered to host an ad hocdiscussion of off-site measurement results thusfar, for all interested parties including State,NRC, EPA, other federal agencies, and localpublic safety organizations. Following thismeeting, DOE decided to bring in additional re-sources to establish and support a federal tech-nical coordinating center for off-site measure-ment and analysis activities. By Saturday thisCommand and Control Center was fully equipped,and ten RAP teams plus the EPA and other agen-cies were coordinating a significant off-sitemonitoring and assessment activity in conjunc-tion with the State Bureau of Radiation Health.Both the State and the NRC were the major usersof the results. Following the Three Mile Islandaccident, the approach and procedures performedon an ad hoc basis were formalized into the'pre-sent Federal Radiological Emergency ResponsePlan (FRERP). In addition to coordination ofthe off-site monitoring and assessment effort,

287

288

the present FRERP plan assures that all federalsupport to the state and local authorities isfully coordinated under the direction of theFederal Emergency Management Agency (FEMA).

II. THE FRMAC ROLE

The Federal Radiological Monitoring andAssessment Center has a crucial role to play inresponse to a major radiological accident. Thelesson learned from the Three Mile Island re-sponse was not only that one must rapidly assem-ble accurate off-site measurements, but thatthese results must be clearly and quickly placedin the hands of the local decision makers andmust represent all off-site data being collect-ed. Unfortunately the Chernobyl accident, withits extensive spread of fission product contami-nation, clearly underlines the need for the roledescribed above. The complexity of the measure-ment problem cannot be underestimated. Alpha,beta, and gamma effects must all be accountedfor over potentially large geographic areas, andthe whole process may be further complicated bya continuing release of radiation as well aschanging environmental factors. The calibrationand assessment functions share the same compli-cations as the data collection. The bottomline, however, is that in spite of the very com-plex technical processes which must be carriedout, information critical to the protectiveaction processes must be placed in the hands ofthe local decision makers and the CFA at theearliest possible time. These results must alsobe available in a format that the local authori-ties and the CFA can immediately use to informnontechnical authorities and the public in aclear and unambiguous way. Ultimately the newsmedia will be the primary means for communicat-ing directly with the public concerning thestatus of the situation and recommended actionsfor public protection. There is a very realtechnical concern for having an adequate amountof time to make accurate measurements, cross-check the results, prepare assessments, and com-plete the protective action judgments beforemaking public statements. The major challengefor the FRMAC staff, working hand-in-hand withthe State and CFA, is to create a scientificwork environment for the off-site measurementeffort to compress the information turn aroundtime to the shortest possible time. Emergencieswhich are technically complex require extensivepreplanning and a major investment in technicalresources in order to meet the problem head on.

III. THE FRMAC RESOURCES

A. Management

Senior level managers with extensiveoperating experience in field situations areneeded to quickly size up the situation, deter-mine the needed major resources, and implementdeployment plans to begin the response. Inter-faces with State and local authorities and theCFA must be immediately established to determinethose resources they feel are most urgentlyrequired.

B. Monitoring Personnel and SpecialEquipment

For a large-scale release of radia-tion which may contain particulates, a largenumber of trained monitoring personnel and theAerial Measuring System resources will be placedon full alert and deployed if the release hasalready occurred or is imminent. RAT teams fromthe closest DOE laboratories and contractors andthe EPA teams will be among the first federalpersonnel to reach the scene. The aerial radia-tion measurements can provide a very rapidassessment of a large area, locate plumes, andprovide isotopic identification. The aerialdata can assist the monitoring manager in dis-patching ground monitoring teams that are bestequipped to make in-situ measurements, set upair sampling stations, etc.

C. Assessment/Data Base Personnel andEquipment

At the earliest possible time aftermonitoring is initiated, an assessment team ofhealth physics specialists must establish acommon map grid for collecting and reportingdata, specify the preferred units for reportingresults, assure the integrity of all monitoringinstruments (i.e., set up a small calibrationrange), and prepare maps with results clearlyshown. The assessment team must consist ofState, CFA, DOE, and EPA personnel. Before anydata leaves the FRMAC, each organization mustindicate that they have knowledge of the dataand the FRMAC Off-site Technical Director mustformally release the data to be transmitted orhand carried to the State and CFA senior offi-cials. Protective action recommendations arenot provided by the FRMAC since they are theprerogative of the State and CFA. If asked byeither the State or CFA for technical assis-tance, the FRMAC will respond.

D. Other Technical Support

1. Command Post Support -

The DOE support team can fullyequip a command post to support an operationinvolving 200 to 400 personnel for around-the-clock operations. Included are logging anddosimetry badging for all FRMAC staff, word pro-cessing equipment, copy machines, office sup-plies, a technical reference library, signs/labels for a1! FRMAC functions, and mundane butessential items such as coffee pots.

2. Communications

A comprehensive communicationsystem capable of providing telephone, VHFradio, HF radio, and satellite links can beairlifted to even a remote site to assure com-mand and control of the operation. In addition,the communication system plays an essential rolein the rapid dissemination of FRMAC results tothe State and CFA emergency centers.

289

3. Photo/Video

A scientific and documentaryphoto/video capability is prepared to supportrequirements for aerial photo maps, technicaldocumentation of recovery and assessment opera-tions, and copying of documents for volume dis-semination. Teams are trained to work in fullanti-contamination clothing if necessary.Graphic arts personnel also support this effortto prepare maps or appropriate graphics for usein media briefings at the Joint InformationCenter.

4. Electro-Mechanical

Large turbine powered generatorsto fully support a stand-alone FRMAC site,mechanical systems to repair and support FRMAC,and other major support items may be deployed ifrequired.

IV. FRMAC SITE SELECTION

The selection of a FRMAC site depends on anumber of factors, but the most significant arethe details of the accident. Although therewill be some pressure to prelocate a potentialFRMAC site, the specifics of the accident mayrule out such a site. All of the resources de-scribed above are packaged for military airliftand would be trucked from the nearest airport tothe FRMAC site. For a worse-case situation,over 200,000 pounds of equipment which wouldfill seven or eight 40-foot flatbed trucks maybe deployed to the site. The equipment and per-sonnel can be completely self-supporting exceptfor housing and food. Even this support couldbe provided by the military if required. TheFRMAC site must not be in the path of anyreleased radiation, and the distance from thesite can vary from five to twenty miles depend-ing on the severity of the accident. The FRMACadvance party team is prepared to make thisselection at the time a response is required.The FRMAC advance party could be expected toarrive on-site within six to eight hours, and afull FRMAC can be in operation within 24 hours.Some of the important criteria for a FRMAC siteare:

o Space providing 10,000 square feet in asingle or separate rooms, (i.e., school,National Guard Armory, tents, trailers,etc.)

o High ground for communications repeaterso Road access for large flatbeds and RVs

equipped for analysiso Nearby housing for 200 to 400 personnelo Available fuel for aircraft and

generatorso Many other items dictated by the

accident

If the accident is the result of a naturaldisaster such as a flood, earthquake, tornado,etc. then many roads, utilities, etc. could beunusable.

V. FRMAC SUMMARY

A major radiological accident creates apotentially complex set of problems for off-siteand on-site response teams. For the off-siteeffort alone, most State and RAT team resourceswill be exhausted after the first 24 to 48 hoursof continuous operations. The FRMAC conceptpermits some thirteen federal agencies toprovide a carefully coordinated and centrallymanaged support as a backstop for the State andCFA for the off-site radiation monitoringactivity. Through the FRMAC, the State and CFAmay obtain comprehensive technical assistancedrawn from the extensive resources of the DOEnational laboratories and contractors. Underthe FRMAC concept, each agency 1s also permittedto carry out their individual statutory require-ments. The concept has been fully tested in afull-field exercise held in Florida at theSt. Lucie reactor in the spring of 1984. Asecond full-field exercise is planned for thesummer of 1987. The resources described abovecontinually support the DOE in several majoremergency programs and, in these other roles,carry out a dozen or more exercises per year inaddition to actual responses to more limitedemergencies.

ANS Topical Meeting on Radiological Accidents-Perspectives and Emergency Planning

Exercising the Federal Radiological Emergency Response Plan"Kathy S. Gant, Martha V. Adler, and William F. Wolff

ABSTRACT Multi-agency exercises were an impor-tant part of the development of the FederalRadiological Emergency Response Plan. This paperconcentrates on two of these exercises, theFederal Field Exercise in March 1984 and theRelocation Tabletop Exercise in December 1985.The Federal Field Exercise demonstrated theviability and usefulness of the draft plan;lessons learned from the exercise were incorpor-ated into the published plan. Tht RelocationTabletop Exercise examined the federal responsein the post-emergency phase. This exercisehighlighted the change over time in the roles ofsome agencies and suggested response proceduresthat should be developed or revised.

I. INTRODUCTION

The Federal Radiological Emergency ResponsePlan (FRERP) was developed to define the roles ofvarious federal agencies when assisting a stateduring a radiological incident and to ensure thecoordination of the federal assistance. TheFRERP was first published in the Federal Registeras an interim, but operational, plan on September12, 1984 (FEMA, 1984). It was republished in aslightly revised form on November 8, 1985 {FEMA,1985), aftar concurrence by all the participatingagencies. The plan has been tested in exercises,so that needed changes could be identified. Thispaper will concentrate on two of those exercises,the Federal Field Exercise in March 1984 and theRelocation Tabletop Exercise in December 1985.

A basic understanding of the FRERP requiressome brief identification of the major federalparticipants and their roles. One is theCognizant Federal Agency (CFA), the agency thatowns, licenses, or regulates the material orfacility involved in the accident. The CFA hasthe primary responsibility for the events at thesite of the accident and makes federal recommen-dations to the state regarding protective

aBased on work performed at the Oak RidgeNational Laboratory, operated for the U.S.Department of Energy under Contract No. DE-ACO5-840R21400 with Martin Marietta Energy Systems,Inc.

actions or reentry In some cases, there will beno CFA, or a state regulatory agency will haveonsite responsibility. For an accident involvinga commercial power reactor, the Nuclear Regula-tory Commission (NRC) will be the CFA.

The Federal Emergency Management Agency(FEMA) coordinates all nonradiological assistanceto the state. FEMA plans call for establishing aFederal Response Center (FRC) near or colocatedwith the state response center, so that thefederal agencies and their state counterparts canwork together. FEMA also assists the CFA inpresenting federal recommendations and, alongwith the CFA, has responsibilities in thecoordination of public information and Congres-sional and White House liaison.

The radiological monitoring and assessmentassistance is initially coordinated by theDepartment of Energy (DOE) under the portion ofthe FRERP called the Federal RadiologicalMonitoring and Assessment Plan. DOE may set up aFederal Radiological Monitoring and AssessmentCenter (FRMAC) from which the state and partici-pating federal agencies can coordinate their datacollection and analysis. The radiologicalassistance portion of the plan may be implementedseparately to provide federal radiologicalmonitoring and assessment assistance in lesssevere incidents. After the emergency phase ofthe response ends, the Environmental ProtectionAgency will assume the coordination of any neededfederal monitoring assistance.

The other federal agencies will carry outany statutory responsibilities, as well ascontributing their resources to provide radiolog-ical assistance or other assistance to the statethrough the FRC or FRMAC.

Exercises were an important ingredient inthe development of the federal plan. In October1982, at the time when the interagency Subcommit-tee on Federal Response was still working on theplanning guidance for the FRERP, member agenciesof the Subcommittee held a command post exerciseto clarify the notification and communicationrequirements which would have to be articulatedin the FRERP. The differing ways that agenciesinterpreted the planning guidance became apparentin this exercise, but it offered the opportunityto deal with the differences systematically byclarifying them as the plan development pro-ceeded.

291

292

II. THE FEDERAL FIELD EXERCISE

The first major exercise of the FRERP washeld March 6-8, 1984, at the St. Lucie NuclearPower Plant on the eastern coast of Florida.Approximately 1000 people from Florida Power andLight Company, the state of Florida, Martin andSt. lucie counties, volunteer groups, andrepresentatives from 11 federal agencies andtheir contractors participated in the exercise.The goal of the exercise was to test the FRERP,which had been published in the Federal Registerin final draft form, in terms of the interfacesbetween different agencies, the usefulness of theassistance to the state, and the adequacy andcompatibility of the agency plans.

Two preliminary drills preceded the actualexercise. A limited number of participants fromali the participating agencies took part in atabletop exercise on December 1-2, 1983, at theutility's Emergency Operating Facility. Thisdrill tested some of the notification procedures,information flow, and agency interfaces using asimple scenario. The second preliminary drill,called the "Dry Run", was held January 24-25,1984, using the emergency facilities that wouldbe used in the exercise. Communications andfacilities were evaluated with a second simplescenario; the control and evaluation proceduresfor the actual exercise were also tested.

The first day of the three-day March fieldexercise was the plant's annual complianceexercise. A weli-deve'oped scenario described alarge release of radioactive material which wentoutside the plant boundaries. The state notifiedthe appropriate federal agencies and requestedfederal assistance; the responding agenciessimulated their arrival and began setting upoperations so that full federal play began by thesecond day. The second day tested the federalagencies' abilities to communicate and worktogether to meet the needs of the state and localgovernments. On the third day, the scenarioshifted to the fifth day after the accident andthe emergency phase of the response began toscale down.

All ftdtral equipment and personnel wereprcpositioned for the exercise to reduce costs. 'The FRC was established in trailers adjacent tothe state's Field Emergency Operating Center.DOE established the FRMAC in a hanger at theStuart, Florida, airport, and the state moved itsradiological monitoring activities to thatlocation. The onsite response was directed fromthe utility's Emergency Operating Facility; aJoint Information Center (JIC) was located in thesame building.

The Federal Field Exercise demonstrated thatthe FRERP was an operable plan that could beimplemented. Some of the necessary interactionsamong the federal agencies and between thefederal agencies and the state were tested forthe first time. The exercise play showed theimportance of preplanning and developing goodworking relationships with counterparts in otheragencies before an emergency occurs. Minorproblems, such as the need for all federalagencies involved in monitoring to be present inthe FRMAC and clarification of the data flow,

were then addressed in the plan before it waspublished.

III. RELOCATION TABLETOP EXERCISE

As with most emergency response plans, theFRERP is most specific with regard to theemergency phase. While the plan was designed toextend beyond the initial response period, themechanisms for communication and coordinationdescribed in the plan had never been testedduring an extended response. Tha RelocationTabletop Exercise provided tne first detailedexercise experience with the FRERP in the periodbeginning several days after an accident.

The Relocation Tabletop Exercise was con-ducted on December 9-11, 1985, at the NationalEmergency Training Center in Emmitsburg,Maryland, in conjunction with the Commonwealth ofPennsylvania and Duquesne Light Company.Duquesne Light operates the Beaver Valley PowerStation, the site of the simulated accident.Representatives of the Pennsylvania EmergencyManagement Agency, the Pennsylvania Bureau ofRadiation Protection, Beaver County EmergencyManagement Agency, Duquesne Light, the AmericanRed Cross, the Institute for Nuclear PowerOperations, and American Nuclear Insurers joinedthe federal players from 13 agencies for theexercise.

The Relocation Tabletop Exercise differedfrom the St. Lucie exercise in many ways. It wasconducted as a tabletop not a field exercise anddealt with the return-relocation phase followinga large power reactor accident, a sensitive timefor which the FRERP guidance is less specific.There were no media, no visitors, and fewobservers. Although Duquesne Light Companyactively participated in the exercise, it was notpart of any required emergency activity. Oneworkshop, to discuss the types of issues thatwould arise during the exercise, preceded theactual event.

The exercise scenario had postulated asubstantial release of radioactive materials froma fuel handling accident at the Beaver ValleyPower Station in Sh1pp1ngport, Pennsylvania,leaving radioactive materials deposited over partof the surrounding area. The exercise assumedthat over 100,000 people would have been evacu-ated (according to the Pennsylvania plan) andthat federal assistance would have been re-quested. When play began, postulated to be thefourth day after the simulated release, the plantwas stable, and no further releases of radio-active material were anticipated. It was assumedthat there was increasing economic and politicalpressure to allow the evacuated people back intothe area.

Instead of playing under one long scenario,the general scenario was divided into a series ofnine miniscenarios, each of which focused on asingle problem. The players worked from tableswhich represented operations centers or centersfor federal assistance that would have beenestablished under the FRERP. The staffs atdifferent tables consulted with each other asnecessary. At the end of each miniscenario,there were general discussion of the issue and areport from each facility.

293

The "no-fault" exercise was intended toidentify issues and problems which neededconsideration or procedures which might need tobe developed for this postaccident phase. Animportant exercise goal was to document anychanges that might take place in the federalroles and determine whether these changes wereadequately covered in the federal plan. TheEnvironmental Protection Agency (EPA) used theexercise to see whether the concepts described inthe Relocation Protective Action Guidelines(PAGs) being developed (EPA, 1985a; EPA, 1985b)for making relocation decisions could be applied.

While most of the evaluators found no needfor major revisions to the FRERP itself (FEMA,1986), several situations arose which indicated aneed for more study and possibly for the develop-ment of additional procedures for use in thelater stages of a response.

For example, after a serious accident at anuclear power plant, and when the plant is stableand onsite recovery is underway, the NuclearRegulatory Commisssion will want to focus itsefforts onsite. Some of the responsibilities itholds under the FRERP as the CFA may be lessappropriate or less in its areas of interest andexpertise. For example, the NRC does not haveexpertise related to environmental questions andtherefore does not feel comfortable coordinatingthe technical aspects of this phase. Theexercise was unclear as to what effect thischange in focus would have on the CFA's responsi-bilities.

The FRC, the coordinating facility for thenonradiological federal assistance, did n"t haveas much play as expected during the exercise.This may have been a result of the exercisestructure, where problems were handle i one at atime, resulting in little strain on the stateresources. The lack of full representation bythe appropriate state agencies may have alsocontributed to the reduced FRC participation.The question of how the FRC role may change asthe response moves into the recovery phase wastherefore not fully explored.

The exercise showed that the technicalradiological assistance provided by the federalgovernment was needed by the Commonwealth ofPennsylvania Leyond the emergency phase of theresponse. The activities of the FRMAC in thisregard were praised by most participants.

EPA's draft Relocation PAGs (EPA, 1985a;EPA, 1985b) contributed to making decisions abouthow people would be allowed to return to theirhomes. The Commonwealth of Pennsylvania,however, chose to select a lower projected one-year dose than the lower bound of the guidelinesas the criterion for immediate return. Moremeasurements and study would take place beforeother areas where projected doses from theradiation levels were approaching the projecteddoses of the PAGs would be opened for return.EPA considered the Commonwealth's approach andincorporated it into the draft Relocation PAGs(EPA, 1986a; EPA, 1986b; EPA, 1935c) that werereleased for public comment.

One problem ic<jntified in the exercise ishow federal agencies, such as EPA, the Departmentof Agriculture, or the Food and Drug Administra-tion, should go about fulfilling any statutory

responsibilities when they are also participatingin the FRC or the FRMAC. Procedures need to bedeveloped ',o that agency recommendations to thestate can be made as required without appearingto conflict with or to bypass the communicationpaths outlined for the CFA, FRMAC, or FRC in thefederal plan.

Play at the Joint Information Center (JIC)was added late in the exercise planning. Nopublic information representative was involved inthe planning. Consequently, the JIC activitieswere not closely integrated into the exercise,limiting the usefulness of any observations. TheCommonwealth's plan calls for public informationactivities to be located in several differentplaces. Although colocation of the JIC staff isdesirable, ways of coordinating the release ofinformation, when all those with the authority toapprove the release cannot be colocated, shouldbe explored.

Detailed advanced planning is probably notappropriate for the postemergency response phase.Each accident will be based on a unique set ofinitial response plans as well as presentdifferent problems, necessitating tailoredsolutions. The principles of the i-RERP can stillguide the extended federal response, but it isuseful to think ahead about how the responsemight function and what kinds of problems couldarise. The Relocation Tabletop Exercise wassuccessful in beginning this process. During thecoming months, the interagency planning groupwill be discussing some of the problems identi-fied in the exercise.

IV. CONCLUSION

The FRERP was never intended to be a staticdocument, additional changes will probably Mmade as the participating agencies get moreexperience in responding under the Plan. In thisregard, it is important for the federal agenciesto be involved in state emergency preparednessexercises whenever possible and for the states toinclude provisions for federal assistance intheir plans.

The FRERP is a generic plan, applicable toall sorts of radiological emergencies. The majorexercises of the FRERP in the past have involvedscenarios dealing with power reactor accidents.Another major field exercise of the FRERP is tobe held in 1987, in conjunction with a powerreactor exercise. Exercises involving othertypes of accidents, different agencies as theCFA, and more complicated situations where theremay not be a designated CFA will help individualagencies see how their plans must be adapted totheir changing roles in the response. Also, suchexercises will help to produce a workable,generic federal response plan.

REFERENCES

Environmental Protection Ac'-^j, idbS\, "Reloca-tion PAGs," Oraf* *o. 1338J (Sept?mber 27).

Environmental Protection Agency, 1985b, "Imple-mentation Guides for Relocation PAGs," DraftNo. 4194C (October 4).

294

Environmental Protection Agency, 1986a, "Reloca-tion PAGs," Draft No. 1338J (June 27).

Environmental Protection Agency, 1986b, "Imple-mentation Guides for Relocation PAGs," DraftNo. 4194C (June 3).

Environmental Protection Agency, 1986c, "Ratio-nale for Relocation PAGs," Draft No. 1336J(June 91.

Federal Emergency Management Agency, 1984, "TheFederal Radiological Emergency Response Plan(FRERP), Publication as an Interim Plan,"Federal Register 49(178), 35896.

Federal Emergency Management Agency, 1985, "TheFederal Radiological Emergency Response Plan(FRERP), Concurrency by All Twelve FederalAgencies and Publication as an OperationalPlan," Federal Register 50(217), 46542.

Federal Emergency Management Agency, 1986,"Relocation Tabletop Exercise (RTE) After-Ac-.ion Report" (March 12).

ANS Topical Mating on Radiological Accidents—Perspectives and Emergency Planning

Emergency Response to Transportation Accidents

Cheryl Sakenas

ABSTRACT - The Nuclear Regulatory Commission(NRC) has issued a policy statement regardingits role in transportation accidents involvingthe release of radiation. Each State has theprimary responsibility for protecting thehealth and safety of its citizens. Stateshave chartered certain agencies with theresponsibility of responding to radiologicalemergencies.

When informed of a transportation acci-dent, the NRC will notify the designated Stateagency to ensure that they have been informedof the incident and will offer NRC assistance.The NRC also will notify the U. S. Departmentof Energy (DOE), the U. S. Department ofTransportation (DOT) and any other affectedagency to ensure a coordinated Federalresponse.

Under an existing NRC/0OT Memorandum ofUnderstanding, NRC is responsible for investi-gating any accidents involving packagesregulated by the NRC. NRC staff may bedispatched to an accident scene involvingpackages not regulated by NRC whenever signif-icant aipounts of radioactive material were ormight be released and NRC activities willgenerally be limited to information collectionunless assistance is requested. DOE would beavailable to assist State and local respondersin monitoring and decontamination of affectedareas. In an accident in which Federalassistance was being provided to the State,portions of the Federal Radiological EmergencyResponse Plan would be activated to coordinatethe provision of that assistance. An exampleof recent experience in responding to atransportation event will be discussed.

On March 23, 1984 the Nuclear RegulatoryCommission issued a policy statement regardingthe agency's role in the response to transpor-tation accidents involving radioactivemateri\\. This was published in theFedei , Register on March 29, 1984 at 49 FR12335. The purpose of the policy statement

is to state clearly the extent of the NRC'sparticipation and involvement in responding toa transportation accident or incident. NRC'srole and responsibilities for responding toany potentially threatening incident involvingNRC-1icensed activities are delineated in theNRC Incident Response Plan, NUREG-0728,Rev. 1, April 1983.

The response to transportation accidentsis less predictable than the emergency re-sponse to radiological accidents at licensedsites because of the uncertainties surrounding(1) the location where the accident occurs,(2) the diversity of authority of those whowill be responding, and (3) the likely limitedradiation knowledge of the first-on-sceneresponders (who are usually local officials).Each State has the primary responsibility forprotecting the health and safety of itscitizens from public hazards.

The Commission views the States asappropriately having the lead in the overalldirection of response to transportationaccidents. In the United States, we experi-ence approximately 10-20 transpr-tatinnaccidents per year involving radioactvematerials. These are routinely handled by aState-designated agency.

Most carriers of shipments ofmaterials licensed by NRC are exempt from NRCregulations while in transit and come underthe regulatory authority of the U. S. Depart-ment of Transportation (DOT). The DOT oper-ates a National Response Center which servesto relay information to State and Federalagencies concerning transportation incidentsinvolving hazardous materials. DOT regula-tions require a carrier, to promptly notifythe National Response Center after an incidentoccurs in which, fire, breakage, spillage, orsuspected radioactive contamination occurs.Each notification of a transportation incidentof any kind is relayed by the National Re-sponse Center to the Regional Office of theEnvironmental Protection Agency (EPA) forincidents on land or to the U. S. Coast GuardCaptain of the Port for incidents in navigablewaters. When a reported incident is known toinvolve radioactive material, notification

295

296

also is made to the Regional CoordinatingOffice for Radiological Assistance of the U.S. Department of Energy (DOE) and to the NRCOperations Center. The NRC also may becomeaware of a transportation incident throughother channels, such as the shipper, thecarrier, or the police or highway patrol.

When informed of a transportation acci-dent, the NRC will inform the agency designat-ed by the State as soon as practicable toensure that the State agency has been informedof the incident. The NRC will offer technicalassistance in the form of information, advice,and evaluations. The NRC also will ensurethat DOE, DOT, and other affected agencies,including the Federal Emergency ManagementAgency, are mad" aware of the incident.

Under an existing NRC/DOT Memorandum ofUnderstanding, NRC is responsible for investi-gating all accidents, incidents, and instancesof actual or suspected leakage involvingpackages of radioactive material regulated bythe NRC. Accordingly, NRC inspectors willnormally be dispatched to the accident scenewhenever a report is received that suchpackages (i.e., Type B containers) wereinvolved in a significant accident that couldhave released or threatens to release radioac-tive materials. To maintain awareness so thatmeaningful technical assistance may be provid-ed to the State, NRC staff also may be dis-patched to an accident scene involvingpackages not regulated by NRC (i.e., Type Acontainers) whenever a significant amount ofradioactive material might be released. TheNRC will provide information on packagingcharacteristics in response to any queryregarding NRC-approved packages.

Any NRC personnel at the scene of atransportation accident will notify theState/local government on-sctf"? coordinator ofhis/her presence and make clear that, unlessNRC assistance is requested by the on-scenecoordinator, NRC activities will be primarilylimited to information collection and assess-ment. NRC personnel will pass along recommen-dations to emergency response personnel onradiological issues only if NRC assistance isrequested by the on-scene coordinator or ifNRC personnel at the scene believe additionalactions are needed to protect emergencyresponse personnel or members of the public.NRC's role, however, is not to evaluate State

and local government emergency responseactions. NRC will respond to requests forinformation on NRC activities in connectionwith the event. Requests for specific infor-mation on an accident normally will be re-ferred to the appropriate State agency, or tothe DOE if the situation relates to DOEactivities. This policy relates solely toradiological concerns. Responding to anyattempt to steal or sabotage a shipment ofnuclear material is a responsibility of theFederal Bureau of Investigation (FBI) asdelineated in the MRC/FBI Memorandum ofUnderstanding dated April 27, 1979, andpublished in the Federal Register onDecember 20, 1979 (44 FR 75535).

A Federal Radiological Emergency ResponsePlan (FRERP) has been developed for use inpeacetime radiological emergencies and definesthe authorities and responsibilities of eachFederal agency involved in a significantemergency response. In most case* Federalresponse will not be needed. Sta e and localgovernments may request Federal assistance asneeded. In most cases this would involvefield monitoring and cleanup support from DOEor EPA and technical assistance from NRC. Ifcontamination of food-stuffs is involved,assistance may be requested from the Depart-ments of Health and Human Services andAgriculture.

Since the FRERP is concerned primarilywith Federal support to State and localgovernments beyond the immediate site of theemergency, State and local governments willdefine an area "onsite" at the time of theaccident and manage all actions within thatarea. If the accident involves materialsshipped by DOD or DOE, these agencies willdefine and control the onsite area. There aresome exceptions to this, depending on the typeof material, i.e., spent fuel, and whether theinvolved State was an Agreement State, inwhich case the State may define and controlthe onsite area.

An example of how the NRC responds to atransportation event involved the spill ofuranium oxide (yellowcake) following a colli-sion between the tractor trailer carrying thematerial and a train on August 27, 1985 inWells County, North Dakota. Of the 53 drumscontaining the material, 30 ruptured, contami-nating a 5000-square foot area, including a2-mile stretch of highway which was isolatedby the responders.

The first responders, local fire depart-ment, State and local law enforcement, andIccal rescue team were backed up by the StateHazardous Materials Emergency Response Teamwhich included State Health Department andRadiation Control Program personnel. The dayafter the accident the State requested NRCassistance in the cleanup and recovery ef-forts. The Region IV office responded bysending three health physics inspectors to theaccident scene to assist the State by provid-ing recommendations on decontamination andrecovery of the area. The mining company alsodispatched a team to take care of the cleanup.Oak Ridge National Laboratory performedbioassay analysis on contaminated responseteam members. The site was returned to normalin 15 days.

ANS Topical Meetiag oa Radiological Accideats—Perspectires aad Eawrgeacy

New Source Terms and the Implications for Emergency PlanningRequirements at Nuclear Power Plants in the United States

Geoffrey D. Kaiser and Michael C. Cheok

ABSTRACT This paper begins with a brief re-view of current approaches to source termdriven changes to NRC emergency planning re-quirements and addresses significant differ-ences between them. Approaches by IDCOR andEPRI, industry submlttals to NRG and alterna-tive risk-based evaluations have been con-sidered. Important issues are discussed, svchas the role of Protective Action Guides indetermining the radius of the emergency plan-ning zone (EPZ). The significance of currenttrends towards the prediction of longer warn-ing times and longer durations of release innew source terms is assessed. These trendsmay help to relax the current notificationtime requirements. Finally, the implicationsof apparent support in the regulations for athreshold in warning tiae beyond which ad hocprotective measures are adequate is discussed.

1. INTRODUCTION

In 10CFR50.47, the Nuclear RegulatoryCommission (NRC) requires that no operatinglicense will be issued for a nuclear powerstation unless the NRC finds that there isreasonable assurance that adequate protectivemeasures can be and will be taken in the eventof a radiological emergency. Thus, utilitiesare required to develop an emergency plan foreach nuclear power plant. Generally, the zonewithin which emergency plans must be developedfor prompt protective measures against expo-sure to the pluae (known as the plume exposureemergency planning zone (EPZ)) consists of anarea with a radius of about 10 miles.

The reasons why 10 miles was chosen forthe radius of the EPZ are given in NUREG -0396 (USNRC, 1978) and are summarized as fol-lows in NUREG - 0654 (USNRC, 1980).

1. Projected doses from the traditionaldesign basis accidents would not ex-ceed Protective Action Guide (PAG)a

levels outside the zone.

2. Projected doses from most core-meltaccident sequences would not exceedProtective Action Guide levels out-side the zone.

3- For the worst core-melt sequences,immediate life threatening doseswould generally not occur outside thezone.

4, Detailed planning within the EPZwould provide a substantial base forthe expansion of response efforts inthe event -that this proved necessary.

As is explained in NUREG-0396, the secondand third criteria were applied within a risk-based framework that relied upon the Proba-bilistic Risk Assessment that was done inWASH-1400 (USNRC, 1975), including the fre-quencies of accident sequences and th* corres-ponding characteristics of the fission productsource terms. The WASH-1400 fission productsource terms were input into the CRAC (Cal-culation of Reactor Accident Consequences)

aThe PAGs were defined by the U.S. Environ-mental Protection Agency (US EPA 1980) and areexpressed as a range: 1-5 rera for whole bodyexposure and 5-25 rem for thyroid exposure ofthe general public. The lower value should beused If there are no major local difficultiesor risks that would arise In carrying out pro-tective actions at that level. In no caseshould the higher bound be exceeded in deter-mining the need for protective action. Refer-ence (USEPA, 1975) states that these recom-mendations are for use by Federal and Stateagencies in their emergency planning activi-ties. It also states that the PAGs are spe-cifically directed towards protective actionsby civil authorities in the event of inad-vertent releases of radioactive material fromfixed nuclear facilities. Thus, the PAGs areenvisaged for use in both emergency planningand response. j^

297

298

code, which was the consequence modeling codedeveloped for WASH-1400, and "Dose vs Dis-tance" curves were calculated as shown on Fig-ure 1. These curves give the probability ofexceeding a given whole body dose as a func-tion of distance, conditional on the occur-rence of core-melt.

OiII*UCt IMUISJ

Briefly, criterion 2 may be roughly translatedinto the quantitative equivalent that theprobability of exceeding the PAGs at 10 milesis no more than about 0.3 for "most core-meltsequences." The 200 rera curve on Figure 1 isdominated by the "worst core-melr. sequence"and shows a pronounced knee at about 10miles. Since 200 rem is roughly the thresholdfor early fatalities (see the PRA ProceduresGuide (USNRC, 1983)), the 200 rem curve inFigure 1 is the approximate quantitativeexpression of the third criterion.

In summary, the original reasons for thechoice of a radius of 10 miles for the EPZcontained the following important elements.

1. Risk-based calculations of the con-sequences of core-melt accidents,including consideration of both the"worst case" accident and the conse-quences of all core-melt accidents.

2. Source terms with characteristicsthat were determined using the '*«. methods of WASH-1400.

3. The EPA's Protective Action Guides.

4. A criterion that addressed the con-sequences of design basis accidents.

5. A criterion that addressed the needfor flexibility of response.

II. NEW SOURCE TERMS AND THE EPZ

Recent work on fission product sourceterms has shown that, in several cases, themagnitude of the source term is considerablysmaller than was thought to be the case at thetime that WASH-1400 was written (APS, 1985),although this statement cannot be shown to betrue for all source terms in all reactors andthe NRC has stated that generalizations areinappropriate (USNRC, 1986). However, manyhave found recent progress to be so premisingthat they have already addressed the questionof what the implications of the new sourceterm results are for emergency planning.

The Baltimore Gas and Electric Companyhas submitted a request to the NRC for an ex-emption from the requirement that the EPZshould have a radius of 10 miles (Montgomery,1986). Using a Calvert Cliffs-specific sourceterm for Event V, the interfacing systemsLOCA, which is the worst case source term forthis particular PWR, BG&E essentially usedcriterion 3 of the four from NUREG-0654 toshow that a radius of no more than 2 miles isneeded ror the EPZ. They reinforce this con-clusion by showing that the Calvert Cliffs DBAdoes not lead to doses exceeding the PAGs at,this distance. However, they do not address'the question of whether most core melt acci-dents lead to the PAGs being exceeded at 2miles, or whether planning inside 2 milesallows sufficient flexibility to expand beyondthat distance if need be.

The Electric Power Research Instituteand its contractor, NUS Corporation, (EPRI,1986) have undertaken an exercise to see whatcan be concluded if the new Industry DegradedCore Rulenaking Program (IDCOR, 1984) sourceterms are used within the framework prescribedby NUREG-0396 and NUREG-0654. Using criterion2 (most core-melt sequences and the PAGs) andcriterion 3 (worse case core-melt sequencesand life-threatening doses) EPRI has shownthat the radius of the EPZ for the four IDCORreference plants (Zlon, Sequoyah, Peach Bottomand Grand Gulf) need be no more than 3 miles.The criterion relating "most" core meltsequenrjs to the PAGs at the EPZ boundary isby fcr the most restrictive and this is sig-nificant, as will be discussed later in thispaptsr. EPRI did not discuss the influence ofDBAs on the radius of the EPZ. This is rea-sonable, because the current DBA source termis recognized to be conservative and is beingrevised using the NRC's modern sourc? terncode package (USKRC, 1986a; 1986b). Finally,EPRI did not address the flexibility questionexcept to say that, with the IDCOR sourceterms, the probability that there will be aneed to expand prompt emergency responsebeyond 10 miles, or even as far as 10 miles,is very small.

IDCOR itself has undertaken to devise anew framework for the definition of the radiusof the EPZ (IDCOR, 1986). Briefly, IDCOR de-fines "Limit tines" on a plot of frequency vs

299

mean individual whole body dose. These limitlines correspond to an individual risk cfearly fatality of lO"? per year in the regionof the line in «hich the whole body dose ex-ceeds 200 rent and an individual risk of latentcancer fatality of 10~^ per year in the regionof the line in which the whole body dose isless than 200 rera. Using the IDCOR sourceterras and also the NRC source terms calculatedwith r.he NRC's Source Term Code Package, IDCORshows that the EPZ need have a radius of only2 miles. IDCOR's work is therefore risk-basedand relies on new source terms. However, thePAGs play no role in defining the radius ofthe EPZ and the flexibility criterion is notaddressed. The DBA also plays no role.

The owners of the Seabrook plant havealso carried out extensive work on EPZ reduc-tion and have been engaged in discussion withthe NRC. At the time of writing, no publishedwork was available. However, the authorsunderstand that the work involved extensivenew analysis of the source terms for risk dom-inant accident sequences and that the approachadhered fairly closely to that of NUREG-0396and NUREG-0654.

Finally, other authors have consideredthe question of the relationship between thesize of the EPZ and the characteristics of thesource terms. There is not enough space toreview these in detail here. However, Kaiser(1986) has looked at the relationship betweenthe radius within which evacuation is neces-sary to avoid life-threatening or injury-threatening radiation doses and the sourceterm magnitude and has concluded that, if thelargest average volatile fission product re-lease fraction is less than about ten percent,a 2 mile EPZ would be sufficient to preventsuch doses. Sofer (1985) has discussed theconcept of the "graded approach" in which theradius of the EPZ is not reduced but in whichplanning for prompt evacuation is confined tothe first two miles. He shows that the gradedapproach can be justified even without newsource terms.

III. ROLE OF THE PAGS

In the approaches adopted above, theauthors all retain a risk-based approach andthey all use new source terms. Not everyoneuses a DBA-based criterion, but there seems tobe a general consensus that this is not aserious problem because the DBAs are to be re-vised and will look like severe core damagesequences in the category of "most core-meltaccident sequences" anyway. The flexibilitycriterion is universally Ignored or playeddown, and may deserve more attention than Ithas received hitherto. However, the mostsignificant difference between the approachesis the attitude towards the use of the PAGs.

Essentially, there are two approaches tothe use of the PAGs in defining the radius ofthe EPZ. One Is to abandon the second of theNUREG-0654 criteria and not to consider at allwhether "most" core-melt accidents exceed thePAGs at the boundary of the EPZ. In the

authors' view, this approach is highly desir-able. For example, the 5 rem PAP correspondsto a probability of about one In a thousandthat the affected individual would developcancer over the remainder of his/her life-time. From Figure 1, the probability of oc-currence of the 5 rem PAG at the 10 mileboundary of the EPZ is about 0.3. The WASH-1400 core-melt frequency is about 5 x 10"^ peryear. A further factor of 0.1 can be appliedto allow for the probability that the windwill blow towards a particular individual.Multiplying all of these factors together,emergency planning is being driven by anindividual risk of not more than one in abillion per year. This is six to seven ordersof magnitude below the normally occurringindividual risk of latent cancer fatality.

However, while technical arguments cancertainly show that the risk associated withthe PAGs is small, there is no certainty atthe present time that any scheme that is ulti-mately introduced to justify a reduction ofthe radius of the EPZ - if, indeed, any suchscheme is introduced at all in the forseeablefuture - will be able to avoid the incorpora-tion of the PAGs. Doubtless, Federal Agenciesother than the NRC will be consulted, and theopinions of a host of State and local authori-ties will have to be taken into account, andthere is no predicting what consensus mayemerge from these discussions.

If the PAGs cannot be excluded in thecontext of the impact of "most" core meltaccidents, what then? The immediate diffi-culty is that, if the second criterion ofNUREG-0654 is applied to a source term con-sisting entirely of noble gases, there maystill be a probability of greater than 30 per-cent of exceeding the 1 rem PAG at 10 miles.In other words, even if all source terms areso small that the noble gases dominate theoffsite doses, the second criterion of NUREG-0654 may still be very restrictive and maymake it difficult to show that the radiusof the EPZ can be reduced to as little as2 miles. This point Is discussd furtherbelow.

The next sections of this paper will de-velop a chain of reasoning that will show thatthere is a way of removing this restrictiveresult without diminishing the importance thatthe PAGs play in the underlying rationale fordetermining the radius of the EPZ. The foun-dation stone of the argument is the concept,which was not developed in NUREG-0396 orNUREG-0654, that there is a warning time be-yond which ad hoc protective measures are suf-ficient uo protect the health of the public.As is shown below, there Is support for thisconcept in the regulations.

IV. REGULATIONS AND AD HOC EVACUATION

In the discussions of the final rule onemergency planning and preparedness in theFederal Register of July 13, 1982, the NRC hasa section devoted to the risks associated withlow power operation, "here It states: 'Tor

300

issuance of operating licenses authorizingonly fuel loading and low power operation (upto 5% of rated power), no NRC or Federal Emer-gency Management Agency (FEMA) review andderennlnetlons concerning the state of ade-quacy of sito emergency preparedness shall benecessary" because: "The risks of operating apower reactor at low power are significantlylower than the risks of operating at fullpower" and "Even for a worst-case, low likeli-hood sequence which could eventually result inthe release of fission products accumulated atlow power into the containment, the additionaltim-2 available (at least ten hours (authors'italics)) would allow adequate precautionaryactions to be taken to protect the public nearthe site."

Thus, this chain of reasoning lends sup-port to the concept that there is some thres-hold in time beyond which ad hoc emergencyresponse is adequate. In NUREG-0396, however,thu doses which are the basis for Figure 1

were calculated by assuming that people con-tinue their normal activities for 24 hoursafter the passage of the plume, no matter howlong the warning time and or the duration ofthe release. It would be logical to calculateFigure 1 on the basis of doses accumulatedonly up to the time beyond which ad hoc re-sponse is deemed to be adequate, since thepurpose of Figure 1 is to give guidance onwhat kind of planned emergency response mightbe required.

This concept of a threshold in time be-yond which ad hoc response is adequate hasrelevance to the results of new source termresearch, see Table 1, which compares thewarning times and durations of release derivedfor the IDCOR reference plants with those fromWASH-1400. Briefly, most of the IDCOR warningtimes exceed those of WASH-1400, and all ofthe IDCOR durations of release exceed those ofWASH-1400.

Peach Bottom

TWTC(V)a

TC(NV)a

S ^TOVW

Grand Gulf

TjQUVAET23QWT23C

Sequoyah

TMLB'S2HF(DO)

b

SiHF(D0,IC)c

Zlon

TMLB1

TMLB'(IC)C

V

IDC

Warr.< ngTime

101242110

4657101.5

268.5230.517

301.523

OR"Duration

ofRelease

8050503030

10108050

46654

1374

WASH-1400

WarningTime

2222

22-2

12221

121

Durationof

Release

3333**

33_3

0.51.51.530.5

0.530.5

aFailure to shut down with venting in the wetwell airspace (V) or withoutventing (NV)

^DO - Drains open; DC - drains closedCIC - Ibipaired containmentdSee IDCOR Summary Report (IDCOR, 1984) for definition of symbols T, W,etc. The sequences listed above are those used to calculate the conse-quence analysis results in the above reference. The failure is by grossrupture of the containment unless otherwise stated.

TABLE 1SOMUKY <F NAMING Tims AND DOTATIONS <F KLEASE (Hr)

301

Figure 2 was derived with a source termthat consisted principally of the noblegases. The warning time was 1 hour and theduration of release was also 1 hour, and thedoses on which Figure 2 was based were accumu-lated up to 24 hours after cloud passage as inNUREG-0396. As can be seen, there Is a chanceof C.8 that the 1 rem PAG will be exceeded at10 miles „ This highly conservative case - arelease consisting mainly of noble gases wouldbe expected to have a longer warning time andduration - Illustrates the point that wasnoted above, that criterion 2 of tTOREG-0654may still be restrictive even with smallsource terms.

Figures 3A and 3B show how Figure 2changes if warning times and durations ofrelease are longer (3A) and if doses are onlyaccumulated up to a threshold time of either

10"a.<UJ>

XHIa

zMauiuiuxUJu.o

a:a.

9 - I H I c m_ I - Tp-TlK, Td-10. I<: 7 - Tr-TB-S. Td»10, Ti

I - Tr-Ttf»T«-fO. Tl-Z'• I - Tr-T«-Tt»-10, Ti-1<

tttr.Tr-tlM bitMM wctm Mut-

doim m« rt l t iM Into its <hrlTd-tfuratlsn of ri l tM* IhrlTH-MTfilni t lM t« tvKuttt b«for«

rglH» Into itMMtws thf)T«>ttM dclly bifori •VKUitlen <hi)

i0« 10H

DISTANCE (MILES)

10*'

DISTANCE (MILES)

t ra airrwci cum nm NOKI M I WUMM- eat—mnvt am

10"ccUJ>

(3zMaUJUJuXUJu.oIDO

1 - I t M CMt, Tp-T«"Tir»Ti»:'1 - IM« CM* MCMt Tr-TM*243 - • • • • E M MEHt Ttf-104 - Ti-(. TP>T>44. Td-10

Wtr— TPtlH M H M r tK tV ttHJt

mn mt P I I N H Inti its kv)Td^uritlKi if P«1MM twlTM^MPfilnf t lM t l IVKIMtl HftPf

N I H H :nti ataamMMTi hr]Tt\m M l n Mf»r« ivacuitlw tic)

10-' 10*° 10" 10*2

DISTANCE (MILES)F 1 M S» DM! It OUTAMCII - VTKT> a' HMNIM TIHCt

no OUUTICW v r

00K VI OJITUNZ - OTfCTl O> CVtCUATZDN TIMI

24 hours or 10 hour9 (3B). By a logical chainof reasoning, which is not reproduced in de-tail here, it can be shown that (i) if mostcore-melt sequences for a particular reactorare small (i.e. noble gases dominate the off-site doses); (ii) the warning times and/or thedurations of release are long; and (ill) thereis a threshold In time (say 10 hours) beyondwhich ad hoc emergency response is adequate,then the requirement that "most" oore-meltsequences should not exceed the PAGs at theboundary of the EPZ is no longer the most re-strictive. Thus even though the importance ofthe PAGs in the underlying philosophy wouldnot be diminished, the PAGs would no longerdrive the radius of the GPZ as seems to be thecase at present.

V. NOTIFICATION TIME REQUIREMENTS

In the discussion above, it appears to bepossible to make some constructive suggestionsrelating to the impact of the PAGs on theradius of the EPZ by focusing not only on themagnitude of the new source terms, but alBO onthe new timing characteristics. As shownbelow, this focus on timing may also have apositive impact on the notification timerequirements.

The notification time requirement stemsoriginally from NUREG-0396 and NUREG-0654 andis clarified in FEMA 43 (FEMA, 1983). Thealert and notification system should becapable of notifying the public within theplume exposure EPZ within 15 minutes of adecision having been taken to implement aprotective action.

The combination of longer warning times,longer durations of release and generallyreduced release magnitudes have significantimplications regarding the process of makinga decision about the most suitable protectiveactions and the subsequent need to notifythe public within 15 minutes. The discus-sions leading to a decision will, for most

302

sequences, be made without the pressure thatarises from the expectation that a matter ofminutes can make an important difference tothe health and safety of the public. For thesame reason, it is not likely to be crucial tobe able to notify the public within 15 min-utes. Thus, the IDCOR results from the Tech-nical Summary Report illustrate a trend in thetimings associated with the results of newsource term research which, if confirmed, sug-gest that, even for severe core melt acci-dents, the current 15 minute notificationrequirement may not be needed in order toimplement effective emergency actions.

VI. REFERENCES

American Physical Society, 1985, "Report toThe American Physical Society of the StudyGroup on Radionuclide Release from SevereAccidents at Nuclear Power Plants," Rev.Mod. Phys., Jj7_, No. 3, Part II.

Electric Power Research Institute, 1986,"Emergency Planning: The Effect of NewSource Term Data," NSAC-100.

IDCOR, 1984, "Nuclear Power Plant Response toSevere Accidents," Technical SummaryReport.

IDCOR, 1986, "Technical Report 85.4(5.3):Reassessment of Emergency Planning Re-quirements with Present Source Terms."

Kaiser, G. D., 1986, "The Implications ofReduced Source Terras for Ex-Plant Conse-quence Modeling and Emergency Planning,"Nuclear Safety, 27-3.

Montgomery, B., 1986, "The Case for a ReducedEPZ" presented at the Atomic IndustrialForum's Topical Conference on Reducing theCost of Nuclear Power, New Orleans,Louisiana, May 18-21, 1986.

Soffer, L. et. al., 1985, "An Examination ofGraded Response Strategy in EmergencyPlanning and Preparedness," Paper 128 inProceeding of the ANS/ENS International

Topical Meeting on Probabilistic SafetyMethods and Applications, San Francisco,CA, Feb 24 - Mar 1, 1985.

U.S. Environmental Protection Agency, 1975,"Emergency Response Protective ActionGuides - Airborne Releases form FixedNuclear Facilities," Unnumbered Report.

U.S. Environmental Protection Agency, 1980,"Manual of Protective Action Guides andProtective Actions for Nuclear Incidents,"EPA-520/1-75-001.

U.S. Federal Emergency Management Agency,1983, "Standard Guide for the Evaluationof Alert and Notification Systems forNuclear Power Plants," FEMA-43.

U.S. Nuclear Regulatory Commission, 1975,"Reactor Safety Study - An Assessment ofAccident Risks in U.S. Commercial NuclearPower Plants," WASH-1400 (NUREG-75/014).

U.S. Nuclear Regulatory Commission, 1978,"Planning Basis for the Pevelopraent ofState and Local Government RadiologicalEmergency Response Plans in Support ofLight Water Nuclear Power Plants," NUREG-0396/EPA 520/1-78-016.

U.S. Nuclear Regulatory Commission, 1980,"Criteria for Preparation and Evaluationof Radiological Emergency Response Plansand Preparedness in Support of NuclearPower Plants," NUREG-0654.

U.S. Nuclear Regulatory Commission, 1983, "PRAProcedures Guide," NUREG/CR-2300.

U.S. Nuclear Regulatory Commission, 1986a,"Implementation Plan for the Severe Acci-dent Policy Statement and the RegulatoryUse of Hew Source Term Information,"SECY-86-76.

U.S. Nuclear Regulatory Commission, 1986b,"Reassessment of the Technical Bases forEstimating Source Terms," NUREG-0956.

ANS Topical Meeting on Radiological Accidents-Perspectives and Emergency Plaining

Coordination Between NRC and FEMA: EmergencyPreparedness Issues of Current Interest

Edward M. Podolak, Jr.

ABSTRACT The framework for cooperation betweenthe Nuclear Regulatory Commission (NRC) and theFederal Emergency Management Agency (FEMA) isformally defined in an April 18, 1985 Memorandumof Understanding. Basically, FEMA coordinatesall Federal planning for the offsite impact ofradiological emergencies and takes the lead forresolving offsite emergency preparedness issuesand NRC does the same for onsite emergency pre-paredness. This paper describes the efforts ofboth agencies to coordinate offsite and onsiteemergency preparedness issues through theNRC/FEMA Steering Committee on Emergency Pre-paredness. The current status of severalsite-specific and generic emergency preparednessissues, such as offsite medical services, night-time emergency notification, fuel cycle facilityemergency planning, consideration of earth-quakes in emergency planning and ingestion path-way guidance, will be described.

I. INTRODUCTION

In April 1985 FEMA and NRC entered into anew Memorandum of Understanding (MOU) relatingto radiological emergency planning ard prepared-ness. The new MOU preserved the NRC/FEMA Steer-ing Committee on Emergency Preparedness as thefocal point for coordination of emergency plan-ning, preparedness, and response between the twoagencies.

The Steering Committee consists of an equalnumber of members to represent each agency (usu-ally 4 members from each agency) with one voteper agency. There is a provision to escalateissues that cannot be resolved by the SteeringCommittee to NRC and FEMA management, althoughthis has not been necessary to date. Eachagency designates a co-chair and they appointtheir respective members. The presentco-chairs are Ed Jordan, NRC and Richard Krimm,FEMA. The Steering Committee maintains a recordof each meeting. A meeting cannot be held with-out at least the co-chairs or two assigned mem-bers for each agency.

The NRC members have the lead responsibili-ty for licensee planning and preparedness andthe FEMA members have the lead responsibilityfor offsite planning and preparedness. TheSteering Committee assures coordination of plansand preparedness evaluations and revises accep-tance criteria for licensee, State, and 'ocalradiological emergency planning and preparednessas necessary. This entails extensive coordina-tion of site-specific issues. In addition, theSteering Committee is the principal forum forcoordination and resolution of generic emergencyplanning and preparedness issues.

As a practical matter, the Steering Commit-tee does not take votes. The Steering Committeeoperates by consensus. The reason this works isthat each agency has unique technical expertisein its assigned area of responsibility, i.e.,NRC onsite and FEMA offsite. The following fiveexamples of generic issues coordinated by theSteering Committee all concern offsite matters.Where these examples use the terms "staff" or"NRC staff" for convenience, you should read"NRC staff and FEMA."

II. OFFSiTE MEDICAL SERVICES

In 1980 the Commission promulgated a regu-lation, 10 CFR 50.47(b)(12), to require that"arrangements are made for medical ser»ices forcontaminated injured individuals." In i?82 theAtomic Safety and Licensing Board (ASLB) for SanOnofre found that the term "contaminated in-jured" encompassed "offsite exposed" persons.This was overturned by the Atomic Safety andLicensing Appeal Panel (ASLAP), which found that"contaminated injured" meant an onsite or emer-gency worker who was traumatically injured andalso contaminated. At that time the NRC Staffsupported the ASLAP position stating that theterm "contaminated injured" did not encompass"offsite exposed" persons and that ad JTOCarrangements facilitated by a list of medicalfacilities was sufficient for offsite exposedpersons. In a 1983 policy statement(CLI-83-10), the Commission stated that planningstandard (b)(12) extended to members of thegeneral public who may be exposed to dangerous

303

304

levels of radiation and that a list of medicalfacilities capable of providing appropriatetreatment was sufficient to meet the standard.The U.S. Court of Appeals in 1985, after deci-sion reviewing CLI-83-10, GUARD v. NRC, heldthat the Commission did not reasonably interpretits own planning standard when it said that alist of existing treatment facilities was suffi-cient in terms of constituting "arrangements"within the meaning of the regulation. The Com-mission asked for the staff's views on whetherthe Commission should: (1) reinterpret thephase "contaminated injured individuals" toeliminate exposed individuals from the scope ofplanning standard (b)(12) or (2) determine whatadditional arrangements are necessary for ex-posed individuals or (3) amend planning standard(b)(12) to make a list of medical facilities thesole planning requirement for exposedindividuals.

The staff favored option 2 and proposedsome additional arrangements for individualsexposed offsite. The Commission approved thestaff recommendation in June 1986, and in July1986 the Commission adopted a policy statementthat planning standard (b)(12) requirespre-accident arrangements for medicalservices—beyond a list of treatmentfacilities--for individuals who might be severe-ly exposed to dangerous levels of offsite radia-tion following an accident at a nuclear powerplant. In that statement the Commission gavethe staff some general guidance that satisfacto-ry arrangements should include: (1) a list oflocal or regional medical facilities or otherunique characteristics; (2) a good faith reason-able effort by licensees or local or State gov-ernments to facilitate or obtain writtenagreements with the listed medical facilitiesand transportation providers; (3) provision formaking available necessary training for emergen-cy response personnel to identify, transport,and provide emergency first aid to severelyexposed individuals; and (4) a good faith rea-sonable effort by licensees or State or localgovernments to see that appropriate drills andexercises are conducted to include simulatedseverely exposed individuals. The NRC and FEMAare preparing guidance to licensees and Stateand local governments on how to implement thenew Commission policy.

III. NIGHTTIME EMERGENCY NOTIFICATION

The Shearon Harris ASLB wrote the Commis-sion in November 1985 regarding the Board'sconcerns about the possible generic implicationsof certain portions of the evidence in theShearon Harris record. They were concernedabout the adequacy of nighttime emergency noti-fication of residents in the plume exposureemergency planning zones (EPZs) surroundingnuclear power plants. The siren systems accept-ed under FEMA criteria were based on daytimeconditions and the Board felt that substantialni'n>be*"s o f upZ residents wou'd not be aroused

from sleep should notification be necessarybetween midnight and 6:00 a.m., particularly ifbedroom windows were closed. Specifically, theBoard believed there was evidence that in EPZswhere sirens were the primary means of notifica-tion, such notification would not be "essential-ly complete" within 15 minutes under sometypical (summer) nighttime conditions. At theCommission's request, the staff and FEMA re-sponded to the Board's concerns in December 1985and again in February 1986. The responses stat-ed that there are no generic safety problemsinvolving nighttime emergency notification andthat the FEMA study conducted for the ShearonHarris case confirms NRC/FEMA their judgmentthat siren systems designed and evaluated inaccordance with NUREG-0654/FEMA-REP-l, Rev. 1,and FEMA-43 meet the NRC requirements for day-time and nighttime alerting. Specifically, FEMAfound the the Shearon Harris siren system can beexpected to arouse and alert approximately 90%of the EPZ residents during the hours from 2a.m. to 6 a.m., the "worst" summer nighttimehours. This conclusion was based on three majorconsiderations: (1) the effect of nighttimeactivities, so that some percentage of peopleare awake, (2) the effect of intrafamily net-working, and (3) the effect of social,interfamily networking.

The ASLB issued its decision in the ShearonHarris case in April 1986 and found that thealerting system, where sirens are to be supple-mented by a tone-alert radio in the bedroom ofeach residence in the first 5 miles of the10-mile £PZ, meets the applicable requirementsunder summer nighttime conditions. In a May 16,1986 letter to the Commission, the Board statedthat it would not accept a 90% alerting level(Shearon Harris) as meeting the Commission'srequirement of "essentially 100%" alerting inthe first 5 miles of an EPZ (NUREG-0654).' Rath-er, the Board held that "essentially 100%" meanta notification system capable of alerting great-er than 95% of the EPZ residents in the first 5miles and that this required the addition oftone-aJert radios in the Sbearon Harris case.In addition to raising this "greater than 95%alerting within 5 miles" as a generic issue, theBoard raised winter nighttime conditions, withits greater percentage of windows closed, as amore limiting case than the summer nighttimecase addressed at Shearon Harris. To illustrateits point the Board gave data from a 1982 staffanalysis titled "Evaluation of the Prompt Alert-ing Systems at Four Nuclear Power Stations,"NUREG/CR-2655, PNL-4426, which projected alert-ing rates under nighttime conditions. Amongother data cited, was the study prediction that53% of the residents at Indian Point would bealerted during a wi )ter night with snow cover.

Two members of the Board, who had heard theIndian Point Special Proceeding, wrote the Com-mission in June 1986 that, had they known aboutthis alerting rate estimate, they would haverecommended that the Commission require a backupsystem for prompt alerting, such as tone-alert

305

radios. The utility responded to the Commissionin July 1986 stating among other things, thatNUREG-2655 was based an 88-siren system as op-posed to the present 151-siren system. The twoBoard members acknowledged in July 1986 thatthis assuaged their major concern about theadequacy of the siren system at Indian Point,but that they remained concerned about the ade-quacy of siren systems on snowy winter nights.These two Board members also were concernedbecause the staff had never notified them aboutNUREG/CR-2655.

In August 1986 the Chairman requested thestaff to provide by the end of September 1986 apoint-by-point response to the generic safetyconcerns in the May 1986 letter from the ShearonHarris Licensing Board as well as recommenda-tions on whether Commission rules or guidancedocuments need revision or clarification. Fur-ther, the Chairman requested that the staffprovide a response to the June and July 1986letters from members of the Indian Point Licens-ing Board which raised a related concern regard-ing the adequacy of alert systems during stormywinter nights. The staff and FEMA are preparinga response to the Chairman's request.

IV. FUEL CYCLE FACILITY EMERGENCY PLANNING

The staff issued orders in 1981 to certainfuel cycle and other materials licensees tosubmit radiological contingency plans. Also inthat year, the Commission published an AdvanceNotice of Proposed tulemaking (46 FR 29712)proposing to codify the radiological emergencyrequirements set forth in the orders. The staffreviewed the 18 letters responding to the ad-vance notice and in 1986 submitted to the Com-mission proposed amendments that would formallyrequire emergency plans for certain fuel cycleand other radioactive material licensees. Theseproposed regulations would require about 30licensees to have emergency plans and wouldrequire about 30 other licensees to either (1)submit a plan, f.2) submit an evaluation showingthat a significant release is not plausible, or(3) amend their licenses to reduce their posses-sion limits. The staff estimates that few ofthese latter 30 licensees would submit plans;about half would submit an evaluation and abouthalf would reduce their possession limits The30 licensees that would need an emergency planare the same licensees that submitted Radiolog-ical Contingency Plans under the orders issuedin 1981.

NUREG-1198, "Release of UF, From a Ruptus-edModel 48Y Cylinder at Sequoyah Fuels CorporationFacility; Lessons Learned Report" was publishedin June 1986. This report contained many recom-mendations that are currently being evaluated bythe staff; these recommendations could amplify,expand, or clarify some of the emergency pre-paredness requirements listed in the staff pro-posal to the Commission. Hence, at theCommission's request, the staff submitted inJuly 1986 proposed changes to that would solicit

public comment on certain recommendations inNUREG-1198.

V. CONSIDERATION OF EARTHQUAKESIN EMERGENCY PLANNING

Following the San Onofre and Diablo Canyonproceedings, the Commission ruled in December1981 that its regulations do not require consid-eration of the potential complicating effects ofearthquakes on emergency planning. The Commis-sion asked the staff to initiate a rulemaking todetermine on a generic basis whether the regula-tions should be changed to require such consid-eration. In December 1984 the Commissionpublished a proposed rule that ". . .Emergencyresponse plans. . .need not consider the impacton emergency planning of earthquakes which causeor occur proximate in time with an accidentalrelease of radioactive material from a facili-ty " In August 1985 the staff forwarded tothe Commission for approval (SECY-85-283) aproposed final amendment to the regulations thatwould have required "limited consideration" inemergency planning of the complicating effectsof "severe, low-frequency natural phenomena."In April 1986, in San Luis Obipspo Mothers forPeace v. NRC, rehearing en bane, 789F.2d 26(1986) the U.S. Court of Appeals for theDistrict of ColumDia affirmed the Commission isinterpretation of its emergency planning rules;i.e., that the rules do not require consid-eration of the potential complicating effectsof earthquakes on emergency planning. Althoughthe Commission disapproved the staff proposedfinal amendment in June 1986, the matter is notresolved at this time.

VI. INGESTION PATHWAY GUIDANCE

Draft FEMA Guidance Memorandum (GM) IN-1,"The Ingestion Pathway," provides planning con-siderations that include areas of review andacceptance criteria for the protection of thehuman food chain, including animal feeds andwater, which may become contaminated following aradioactive release from a commercial nuclearpower plant. The guidance is addressed to Stateand local government emergency planners within a50-mile radius of nuclear power plants, i.e.,the ingestion exposure emergency planning zone.The implementation would be one year from thedate of publication of the final (GM) IN-1 aspart of the annual plan update. Principal fea-tures of the guidance are (1) public educationand information (2) protective response on thebasis of the Food and Drug Administration re-sponse level for dealing with accidental radio-active contamination of human food and animalfeed (3) and exercises and drills. Because theguidance is still in draft, it is not appropri-ate to provide more detail. FEMA expects toissue IN-1 in final form later this year.

306

VII. CONCLUSION

The NRC/FEMA Steering Committee is an ef-fective mechanism for coordinating emergencypreparedness issues between the two agencies.This paper has described some complex genericissues that have been coordinated between thetwo agencies. In addition, there are manysite-specific issues that are addressed by theCommittee.

ANS Topical Meeting on Radiological Accidents-Perspectives and Emergency Planing

Interactions Between Multiple Organizations Responding to aReactor Accident in Switzerland

Martin Baggenstos and Hans-Peter Isaak

ABSTRACT. The poster deals with the problems thatemerge from the distribution of responsibilitiesin the case of a nuclear accident. It is de-monstrated how clear and simple procedures andgood co-ordination between the on-site andoff-site emergency organizations are necessary forthe smooth handling of such a situation. A goodcommunications network is particularly essential.The special problem of co-ordination betveendifferent organizations performing measurements ofthe contaminated air, ground and food is alsodiscussed. From the experience gained during theChernobyl accident, it has been established thatonly those organizations, which are correctlytrained and have a perspective over the completeaccident situation, can perfora valuable work. Atthe beginning of the accident for example, therewas no organization responsible for taking samplescovering the whole of Switzerland. A first quickassessment of the radiological situation couldtherefore only be made on the basis ofmeasurements performed in the direct vicinity ofthe laboratories or from single samples which hadto be transported over large distances.

I. DISTRIBUTION OF RESPONSIBILITIES BETVEENPLANT LICENSEE AND AUTHORITIES

THE

A. General Principles of Responsibility

In the case of an accident in a nuclearpower plant, the emergency organization of theplant and several off-site authorities are in-volved. In determining the responsibilities forthe protection of the population, attention mustbe payed to the following principles: (1) Who hasthe best means for making certain decisions?

(2) How can the decision reach the population inthe fastest way possible, over a minimum number ofauthorities? The application of these principlescan lead to a case, where in an accidentsituation, certain official authorities arebypassed. Experience gained from exercises hasshown that this is generally no problem, if suchan arrangement is clearly settled in advance.

B. Distribution of Responsibilities

During the course of an accident, thereare many technical and radiological decisions tobe made. There is a connection between thetechnical and radiological measures to be taken inthe plant and the radiological consequences in theenvironment. Fig.l shows how the on-site andoff-site responsibilities are distributed anongthe on-site staff, the off-site emergency Manage-ment and the Nuclear Safety Inspectorate, whichacts as a focal point for the other organizations.The responsibilities can be summarized as follows:

Responsibility On-site. The general responsi-bility for the technical and radiological deci-sions within the plant lie with the emergencystaff of the plant.

Responsibility Off-site. The decisions forprotective measures in the environment are made bythe off-site emergency management (and not by theplant licensee), because the responsibility forthe protection of the population rests in thehands of the off-site authorities.

Nuclear Safety Inspectorate. The Safety In-spectorate is the only authority, which has expertknowledge of both the technical and radiologicalbehavior of the plant, as well as knowledge of therequirements for the protection of the population.The inspectorate is therefore able to supervisethe on-site actions taken by the plant staff andto give expert advice to the off-site managementof the accident.

307

308

Problems Responsibility

ON-SITE

OFF-SITE

On-site Staff

Nuclear SafetyInspectorate

Off-site EmergencyManagement

Flg.1 Principte Distribution of On-site and Off-site Responsibility

1. Detection of the Accident. Manyincidents occur in a nuclear pover plant, that donot necessarily require protective actions in . theenvironment. It is hovever possible, that suchaccidents may occur which sake protective measuresindispensable. For several reasons, the distinc-tion between the two cases is not always easy.(1) A severe nuclear accident nay develop from asimple failure. (2) The judgement of what isrelevant for the protection of the populationdepends somewhat on the point of view of theon-site and off-site organizations. (3) In certaincases the population may take independentprotective actions based on their ovn assessmentof the situation (e.g. in the case of an evidentfire in the plant).

The aain problem is how to set the triggerlevel for alerting the different emergency organi-zations. Different solutions are conceivable.(1) The off-site emergency management is alertedby the plant staff very early in the count of anaccident. The advantage of this solution is thatthe authorities have lots of time to build uptheir organization and are ready should they beneeded. The disadvantage is that in the case of anover-cautious assessment of the on-site situation,false protective actions could be taken. Thedamage may be larger than the benefit of theaction. Furthermore the off-site emergency manage-ment does not have such detailled technicalknowledge of the plant as the plant staff. (2) Theoff-site emergency management is not informedbefore serious radiological consequences for the

population become probable. The advantage of thissolution is that the authorities know thatprotective measures must be implemented at once.The disadvantage is that the authorities may haveto react without much time available. Thefollowing solution has been chosen in Switzerland:

Detection of tha Accident. The detection ofan accident and the correct assessment of theon-site radiological situation lies within theresponsibilities of the plant staff.

Notification of the Safety Authorities. TheNuclear Safety Inspectorate will always be noti-fied in the following cases: if the reactor hasbeen .shut down (scram), if the situation isoptically perceivable by the population (coolingtower no longer operates or fire in the plant) orif the accident has technical or radiological im-plications within the plant.

Notification of the Off-site EmergencyManagement. The off-site emergency management willonly be informed, vhen the accident threatens todevelop in such a way that protective measures forthe population become inevitable.

The criteria for alerting the off-siteauthorities are chosen in such a way that theplant operator in the main control room recognizes(with the help of a clearly visible and audiblesignal) that an incident with damage to the fuelcladding has occured. In order to prevent falsealarms, the criteria for alerting the off-siteauthorities are two-fold, namely high dose rate inthe containment and initiation of emergency corecooling. The combination of these two signals is

309

transmitted automatically to the Bain control roon

and is held fast optically and acoustically.

High Oosa Rate Inthe Containment(>1OR/h)

Warning of theOff-site Authorities

> 2

>1

Emergency CoreCooling hasCommenced

2. Assessment of the Radiological Dan-ger to the Population. In order to assess the po-tential radiological danger to the population,specific data concerning the source term (timescale and amount of radioactivity released) andmeteorology must be available. In principle theresponsibility for the correct assessment of theoff-site radiological situation could lie vitheither the on-site or the off-site organizations.(1) The emergency staff of the plant could assessthe off-site situation and communicate this to theoff-site authorities. The plant staff would needinformation on the release, the meteorology of theenvironment and knowledge of dispersion calcula-tions in the atmosphere. The advantage of thissolution is that the plant staff would be in thebest position to judge the source term and themeteorological situation at the plant site. Thedisadvantage is that the plant staff would have nofirst-hand information on the meteorological situ-ation for the whole of Switzerland. (2) The off-site authorities could assess the off-sitesituation. In this case the advantage is that theoff-site authorities would have expert knowledgeof dispersion calculations and would be in abetter position to judge the need for protective .measures. The disadvantage is that release andmeteorological data would have to be communicatedfrom the plant and from other meteorologicalstations to the off-site authorities. In an acci-dent situation this would function only if a goodcommunications network can be guaranteed. InSwitzerland the following solution has beenchosen:

The off-site National Emergency OperationsCenter is located at the Swiss Meteorological

Institute in Zurich. This institute has a completeknowledge of the meteorological situation inSwitzerland. Furthermore specialists are alvaysavailable to correctly interpret the meteorologi-cal data. The meteorological data from the plantsite and other meteorological stations areautomatically transmitted to the EmergencyOperations Center on a routine basis. Concerningthe release data (time scale and behaviour), thesewould have to be collected from the plant asneeded with the help of a telephone or telefaxtransmitter/receiver. No automatic transmission ofthe release data is forseen at the moment.

Meteorological DataPlant Sit*

Meteorological DataSwitzerland

V /Off-tit* Emergency

Management

Release Data

3. Recommendation and Realization ofProtective Measures. In principle, protectivemeasures for the population may be realized on thebasis of either a prognostic dose, which gives anestimation before or during the release ofradioactive materials, or an expected dose, whichis calculated after the release has occured. If anacute dose is expected within a short time, thenprotective measures should be established from theprognosive dose. On the other hand long-termprotective measures should be determined from theexpected dose, based on actual measurements of theradiological situation. Nowadays many sophistica-ted computer programs exist, which can calculatethe prognostic dose on-line. The question iswhether or not such an on-line computer program ora simple Gauss-Model for hand calculation is bestsuited for emergency applications. The on-lineprogram can process the actual weather data andmany other parameters and calculate different dosepathways. Vith a simple calculation one isrestricted to a few typical weather situations andparameters.

The decision which type of program is to beused, also depends on the user, namely theorganization which Is responsible for therecommendation or realization of protectivemeasures. The responsibilities for recommendationand realization of protective measures for the

On-line Calculation

310

Simple HandCalculation

Method Used inSwitzerland

••_'.: Calculation

•Protective;-v . .'.Measures

Flg.2 Examples of Possible Prognostic Dispersion Calculations

population are distributed within differentfederal, cantonal and community levels. Radiationprotection experts are generally only available onthe federal level. However, the cantonal andcommunity levels must understand, vhat the federallevel is recouending.

ResponsMMies for the Recommendationand Realization of Protective Measures

Federal Level (Recommendation)

I

Cantonal Level (Realization)

I

Community Level (Realization)

Dose calculations performed vith a sophisti-cated on-line program can correct for many specialeffects such as plume Meander. This type ofcalculation certainly has many advantages. Theproblem is that the authorities responsible forthe execution of protective measures nay betempted to overvalue the accuracy of the calcula-tions and then implement protective Measures only

within the calculated zone. Dose calculationsperformed using a simple Gauss-Model do not allowfor correction of effects such as plume meander.On the other hand the character of the estimationis more visible and the executive authorities arenot tempted to ascribe to much importance to theaccuracy of the calculation. The solution whichhas been chosen in Switzerland is shownschematically in Fig.2, along with the tvo basicmethods of dose calculation. In Switzerland thebasic calculation is performed with a sinpleGauss-model. Protective measures are, however,implemented within a much larger area covering anangle of at least 120 degrees.

4. Lessons Learned from the ChernobylAccident. Although in the case of the Chernobylaccident, Switzerland only experienced the inges-tion phase of the accident, several important les-sons could be learned concerning the distributionof responsible emergency organizations. These are:Emergency Reference Levels. The definition ofEmergency Reference Levels should be undertaken bythe sane authority, which in the case of anaccident must implement the reference levels. Allauthorities which propose or execute protectivemeasures should be involved in the planning phase.Recoomendations. Official recommendations must beannounced to the population by recognizedauthorities, not by so-called experts.Information Exchange. Information exchange must beco-ordinated in advance between the authoritiesthat propose and those that execute protectivemeasures. This includes information of the public.A good communications network is essential.

311

II. Co-ordination of Radiological Surveillance in

the Environment ai'ter a Nuclear Accident

A. General Principles of Co-ordinationIn the case of &n accident in a nuclear

pover plant it is customary, that the normalroutine radiation surveillance personnel issupported by additional civil and military surveyteams. For a good co-ordination between the manysurvey teams performing measurements! thefollowing principles should be adhered to.(1) Survey teams measuring the same quantityshould be equipped with the same measuringinstruments. (2) Survey teams should use the samesupporting documents for sampling, documentationand transmission of data. (3) The measurementsperformed should be reproducible in the sense thatthe places of measurement are easily found andaccessible to motor vehicles (places should bedocumented vith a photograph). (4) If possible,civil resources should be limited to certain morecostly and time consuming detallied measurements.(5) Military resources should be used for a rapidsurvey of the fallout situation. (6) The para-meters to be measured should in all cases beadapted to the degree of actual danger (forinstance the amount of strontium in milk is notimportant in the first week after an accident).

B. Utilization of Measuring Organizationsin the Different Phases of an Accident1. Priority of Measurements.In the case of an emission of radio-

active particles one can differentiate twodistinct phates of the accident, namely an acutephase and a medium/long-term phase. Fig.3 shows

how in the first phase of the accident theradiological danger to the population is great fora relatively short period of time. The secondphase represents a lesser radiological danger butover a much longer period of time.

Fio-3 Potential Radiological Danger to the Population

Table 1 shows the different organisationswhich perform measurements during the variousphases of the accident. Basically, permanent civilsystems, which are always operational, areemployed in the cloud phase of the accident.Measurement of the ground contamination in themedium-term ground phase of the accident areperformed by organisations, which must first bemobilized. In the long-term ingestion phase of theaccident, all available laboratories are used.

Time Cloud Phase

EXTERNAL RADIATIONFROM CLOUD

INHALATION OF CLOUD

Ground Phase

EXTERNAL RADIATION FROMGROUND

INGESTION OFCONTAMINATED FOOD

Measuring Organization

Fixed measuring systems,helicopter 1), survey teams 1)

survey teams, laboratory 1)

Fixed measuring systems,helicopter 2), survey teams 2)

Laboratories 1) and 2)

1) Permanent civil systens which are always operational2) C1v11 or military systems that are not routinely operational, but must first be mobilized to

Install their laboratories.

Table 1 : Potential Radiological Danger and Measuring Organization

312

2. Utilization of Different MeasuringOrganizations. There are many different organisa-tions involved in an accident situation, each ofvhich has a veil designated responsibility.

Survey Teams of the Nuclear Fover PlantAffected. The on-site and off-site responsibili-ties are clearly distinguished in Svitzerland. Thesurvey teams of the plant are exclusively respon-sible for tha safety of the plant personnel andfor the measurement of the radiological situationin the plant. The plant survey teams are rapidlyavailable in an accident situation, but are neededmost urgently vithin the plant. The plant istherefore not responsible for performingmeasurements outside of the plant site.

Civil Measuring Systems vhich are PermanentlyOperational. Svitzerland is covered by a permanentnetwork of doae rate measuring units (density isabout 1 probe per 1000 km2). This network quicklygives a first assessment of the external radio-logical situation. Moreover there are severalresearch institutions, vhich are in possession ofpermanent laboratories for measurements of radio-activity. Some of these institutions possess motorvehicles equipped with measuring instruments. Inthe case of an accident, these laboratoriesspontaneously take samples of the radioactivity inthe air, on the ground and in food. Helicoptersequipped vith measuring systems can also berapidly engaged. The helicopters are flown bymilitary personnel. A civil survey team performsthe measurements and takes the samples. All thesesystems are organized in such a way that they areready for action within about 1 to 2 hours and canmeasure and transmit data independently.

Civil and Military Measuring Systems whichMust First be Installed. After the radioactivecloud has passed, additional means are needed inorder to measure the radiological falloutsituation vithin a reasonable time of about 1 to 2days. A large measuring capacity is required inorder to be able to produce a detallied chart. Themeasurements must be performed for the vhole ofSvitzerland, because any background measurement inan area far away from the site of the accident isalso important vhen determining protectivemeasures. Vith the requirement that onemeasurement should be performed per 100 kmZ, thismeans for the military survey teams andhelicopters that about 400 points must bemeasured. For larger distances, the density ofmeasured points can be reduced somewhat if theveather situation can be judged reliably. For acomplete determination of the situation concerningcontaminated food, over 30 laboratories coveringthe vhole of Svitzerland may be mobilized. Theselaboratories must be able to measure a fev hundredsamples each day, and transmit the data to theEmergency Operations Center.

3. Lessons Learntd from Chernobyl. TheChernobyl accident has shown that certain problemsresult in the co.-ordination and data transferbetween the different measuring organizations.Vithin the scope of the Chernobyl accident thefolloving measuring systems vere in operation inSvitzerland: fixed measuring systems, civil groundsurvey teams, survey helicopters, and civil andmilitary laboratories. The Chernobyl accidentshoved, that only a veil organized emergencyorganization can in the case of an accident makecorrect decisions and implement them. Allauthorities concerned must be involved in theplanning of the emergency organization. This Isespecially true for the local communities, vhichmust implement the decisions taken. Specificallytha folloving problems vere encountered:

Fixed Measuring Systems. Of the 50 plannedprobes, only 11 vere in operation at the time ofthe accident. A rough estimation of the externalradiological situation vas therefore missing inthe beginning.

Civil Survey Teams- The mobile ground surveyteams have a fixed charge to measure vithin thevicinity of the nuclear pover plants. Because thetaking of samples vas not organized at thebeginning, the mobile teams also had to takesamples from far avay places and bring them to thelaboratories. These mobile units vere notefficiently used.

Helicopter. Helicopters vere the mostefficient means of taking samples during the firstfev days. The co-ordination betveen civil andmilitary personnel presented no problems. Theresults vere immediately available.

Civil and Military Laboratories. Thepermanently available laboratories performedquickly and efficiently. After a fev days theyhovever became overcharged, because of lack ofco-ordination betveen the organizations takingsamples. A uncontrolled yield of samples occured.The subsequently mobilized laboratories veremobilized too late. A fev days vere needed tomanage the large amount of data coming in.

Acknowledgment. The authors vould like tothank Laszlo Babocsay for his help in preparingthe figures and for his critical reading of themanuscript.

ANS Topical Meeting on Radiotogkal Accidents—Perspectives and Emergency Plunlng

Utility Perspective on Emergency PreparednessGeorge J. Giangi

ABSTRACT

This paper discusses soir* of the major lessonslearned from the TMI-2 accident with respect toemergency response and preparedness. Adescription of the GPU Nuclear Corporationemergency preparedness program is also providedto illustrate the significant differences in theprogram before and after the accident, Whilemajor lessons have been adopted by bothgovernment and industry there continues to beemergency preparedness related regulations andconcerns that promote the misperception thatnuclear power plants are unsafe. In addition,annual exercises send a message to the publicand news media that most, if not all, nuclearplant emergencies will result in evacuation orsheltering of the surrounding population.

Prior to the March 28, 1979 accident at theThree Mile Island station, the typical emergencypreparedness program at utilitias consisted of asimplified emergency plan with severalimplementing procedures. The responsibility formaintaining the program rested with oneindividual with resources that were shared fromother departments. Although drills andexercises were conducted on an annual basis, thefrequency, level of participation and scenarioscope were generally far less than currentprograms.

The NRC regulations and guidance governingemergency preparedness at that time were minimalin comparison to the volumes currently ineffect. (1) In 1970, 1OCFR5O.34 was the firstEmergency Planning rule. In 1975, RegulatoryGuide 1.101, which was approximately 4 pages,provided guidance for onsite Emergency Plans andoffsite emergency planning (for the LPZ areaonly). 1980 marked the start of significant

(1) Atomic Industrial Forum "Background Info"June, 1986

Emergency Planning rules and guidelines. Therules required onsite and offsite EmergencyPlans (for the two EPZ's) with requiredexercises as a condition for maintaining theplant license.

During the TMI-2 accident several areas wereof significant concern. The communicationssystem in place during the accident did notadequately allow for proper information flowamong the utility and Government emergencyfacilities. Over a matter of days, thetelephone company installed additional phonelines to enhance the existing system. Outsidelines were not always available due to thestress placed on the telephone system frompublic use.

The ability tc properly assess theradiological impact on the plant and theenvironment during the accident was also ofsignificant concern. Certain areas of the planthad very high (approximately 100 Rem/hour)radiation levels due to noble gases that made itdifficult to monitor with hand held radiationdetection instruments. The dose projectionprocess consisted of manual calculations usingisopleths which was cumbersome and timeconsuming.

The declaration of the emergency andnotification to offsite agencies were delayedpredominantly due to the lack of understandingby the operators as to why the plant wasbehaving the way it was. The current EmergencyPlan, consistent with regulatory guidancerequires the operator to promptly declare theemergency and initiate notifications whenevercertain values "emergency action levels" arereached or exceeded . This action initiates theimplementation of the Emergency Plansimultaneous with plant mitigation. Inaddition, these Emergency Action levels haveboon included into the Emergency OperatingProcedures to assure this process.

313

314

1 he planning and response process, as wollas the responsibility for maintaining theemergency preparedness:; program, havesignificantly changed from the time of theaccident. The TMI emergency Preparedness programtoday has seven full time staff members. Theirdisciplines include operations, maintenance andradiological controls. Over a dozen drills Areconducted annually which include unannouncedshift drills to test the plant crew onbackshifts, full activation quarterly drills andan annual exorcise. Scenarios are completelydeveloped in-nouse and employ the use ofcomputers and simulators for more accurate andtimely data production. Drills and exercisesare often based on real industry events and areconducted using audio and visual aids, moulagekits, radiation sources and actual meteorology.Provisions cire in place for ensuring theviability of the program on a day-to-day basis.Examples of this include periodic inventory ofemergency facilities, monthly testing ofemergency communications systems and updatingletters of agreement.

Attachment 1 provides specific informationpertaining to the G'PUNC Emergency PreparednessDepartment including budget and annual exercisecosts.

Additional examples of significantimprovements which have been made to theemergency preparedness program since theaccident include:

— Major plant modifications to improvereliability and assessment capability

— Enhanced operator training

Expansion of in-plant and effluentradiation monitors as well as portableradiation detection instruments

Dedicated Emergency Response Facilities

Dedicated Emergency CommunicationsSystem with redundant features

A sophisticated computerized doseprojection model that is terrainspecific and tracks the plume in areasonably predictable manner

A well defined emergency classificationsystem that is integrated with plantemergency procedures.

Installation of a Prompt PublicNotification System

Improved Public and News MediaEducation Program

Closer coordination and interface withFederal, State and Local Governmentofficials

Utility supported radiological trainingfor offsite emergency workers.

while a significant number of lossonslearned have boon addressed, it is stillnecessary to review the consequences of theTMI-2 accident in order to put it inperspective. In some areas this has not beendone. Perhaps ona of the most important examplesis simply that the amount of Iodine that isreleased to the environment during a core nu>llscenario in a L'igh.t Water Reactor issignificantly loss than previously postulated.In fact, approximately 20 curies of Iodineescaped Lo the environment while it was thoughtthat 20 million curies would escape. Thisinformation, along with recently publishedsource term information, suggests to some thatrelaxation of Emergency Planning rules is inorder. For example, this may be in the form ofa reduced Emergency Planning Zone.

The nuclear industry readily accepted andcorrected one of the lossons learned from theTMI-2 accident. Specifically, more education onnuclear power was needed for the public and newsmedia. This has been implemented via annualemergency public information brochure, annualnews media training, aggressive nuclear plantspeakers programs and plant tours. However, therequired annual exorcises which receivewidespread media attention continue to promotethe misperception that all nuclear plantemergencies will lead to a protective action ofevacuation or sheltering due to the harmfullevels of radiation. FEMA and NRC should allowthe states and utilities to conduct morerealistic exercises which test theimplementation of the plans without catastrophicplant failures and lethal doses. In addition,FEMA Guidance Memoranda continue to be developedwhich impose new requirements upon the state andlocal governments and ultimately on theutilities, (eg; increasing requirements for theingestion pathway 50 mile EPZ).

Except for the nuclear industry no otherindustry is required to have an outdoor warningsystem. Although recent evidence suggests thatthe probability for the occurrence of a nuclearplant accident posing harm to residents isextremely low compared to other natural ormanmade hazards it is disappointing to note thatrecent events challenge the adequacy of sirensystems which meet or exceed NRC and FEMArequirements to alert residents that may beindoors under night time conditions.

In summary. State and Local Governments andUtilities are in a better position to respond toa nuclear plant accident today than was the caseon March 28, 1979. This is a dynamic processwhich continually requires updating. If nuclearpower is to remain a viable energy source,regulations, guidance and licensing proceedingsshould be re—evaluated and modified to reflectthe latest source term studies and annualexercises should be allowed to be more realistic.

315

f,PH NUCLEAR EMERGENCY PREPAREDNESS DEPARTMENT

V. P. NUCLEAR ASSURANCE

CORPORATE MANAGEROf EMERGENCY PREPAREDNESS

SECRETARY

TMI EMERGENCYPREPAREDNESS

SFfTION

OYSTER CREEKEMERGENCY

GPU NUCLEAR EMERGENCY PREPAREDNESS SECTION STAFF

SITE EMERGENCYPREPAREDNESS MANAGER

OPERATIONSSR. EMERGENCY

RADIOLOGICALCONTROLS SR.pMFP. PI AMMFff

JR. EMERGENCYPLANNER

PROCEDURECOORDINATOR SECRETARY

316

TYPICAL CRFDENTTALS OF F_P STAFF MEMBERS

SITE EMERGENCY BS DEGREE (OR ADVANCED DEGREE)PREPAREDNESS MANAGER WITH 10 OR MORE YEARS OF

EXPERIENCE IN APPLIED HEALTHPHYSICS/EMERGENCY PLANNING ANDMANAGEMENT

OPERATIONS SR. PREVIOUS SRO CERTIFICATION WITHEMERGENCY PLANNER 5 - 10 YEARS OF EXPERIENCE

MAY HAVE AN AS OR BS DEGREE

RADIOLOGICAL CONTROLS BS DEGREE IN SCIENCE OR ENGINEERINGSR. EMERGENCY PLANNER WITH 5 - 10 YEARS OF EXPERIENCE IN

APPLIED HEALTH PHYSICS/EMERGENCYPLANNING

JUNIOR EMERGENCY AS OR BS DEGREE WITH 0 - 3PLANNER YEARS OF EXPERIENCE IN A

TECHNICAL RELATED FIELD

PROCEDURE COORDINATOR AS OR BS DEGREE WITH 0 - 3 YEARSOF EXPERIENCE

317

SITE EMERGENCY PREPAREDNESS SECTION BUDGET

ITEM (COST IN THOUSANDS OF $)

STAFF PAYROLL (INCLUDING OVERHEAD) 275

DEPARTMENT SUPPLIES 10

EMERGENCY EQUIPMENT AND FACILITY 100UPGRADE

MAINTENANCE AND REPAIR OF OFFSITE 75SIREN SYSTEM (APPROXIMATELY 80 SIRENS)

OFFSITE TRAINING AND RERP MAINTENANCE 125

OTHER CONTRACTOR SUPPORT (E.G.. 50SPECIAL PROJECTS, SIMULATOR TIME)

TOTAL 635K

318

ANNUAL EXERCISF COSTS

- EMERGENCY PREPAREDNESS OEPT:

STAFF 5 X 3 MONTHS••'1.25 MAN YEARS

- EMERGENCY RESPONSE PERSONNEL:

100 SITE PERSONNEL

25 CORPORATE STAFF

125 TOTAL PERSONNEL

PARTICIPATION IN:

PRACTICE DRILLS

TABLE TOP EXERCISES > 5 DAYS

ANNUAL EXERCISE

125 PERSONNEL X 5 DAYS (250 DAYS/MAN YR) »-»>2.5 MAN YEARS

UTILITY 3.75 MAN YEARS X S50K/MAN YR »-*iI190!<_PERSONNEL (INCLUDES OVERHEAD)

TOTAL S190K

ANS Topical Meeting on Radiological Accidents—Perspectives and Emergency Plsaaing

Improving Emergency Management Through SharedInformation Processing—Considerations in

Emergency Operations Center Design

Richard E. DeBusk and J. Andrew Walker

ABSTRACT An Emergency Operations Center (EOC)is a shared information processing facility.Although seemingly obvious, many EOCs aredesigned and operated based on other criteria.The results> measured in terms of responseeffectiveness, are difficult to determine. Areview of ' some recent disasters reveals apattern of poor performance for the EOCsInvolved. These conclusions are tentativebecause so little research has been done on thedesigni operation, or evaluation of emergencyoperations centers.

The EOC is not an onsite response commandpost but a facility removed from the responsewhere tactical and strategic decisions are madebased on information from the response site andelsewhere. The EOC is therefore the .centralfocus of emergency information processing andhigher-level decision making. Examiningexisting EOCs, several common functions emerge.These functions can be described in terms ofshared information processing. However, manyfactors impact the design and operation ofEOCs. Politics, budgets, and personal ambitionare only a few such factors.

Examining EOC design and operation in termsof shared information processing operationalizedin the seven principal functions within the EOCprovides a framework for establishing principlesof EOC design and operation. In the response toemergencies such as Bhopal or Chernobyl theetakes are high. Applying new techniques andtechnologies of management systems can improvethe probability of success. This research is abeginning step—to understand how EOCs function,to define the system. Predictive orprescriptive analysis must wait until sufficientempirical data is available to complete adescriptive model for EOC operations.

INTRODUCTION

Bhopal, Challenger, Chernobyl: majoremergencies are alarmingly frequent. Typically,these emergencies require complex,intergovernmental, multiorganlzationalresponses. In this environment, effectiveinformation sharing among the respondingorganizations is a basic requirement forsuccess. Chernobyl ±a the most recent example

of a major emergency. The response to Chernobylhas been considered inadequate by mostauthorities and poor information sharing hasbeen identified as a major factor contributingto the poor response. Poor information sharingseems to have occurred between the reactoroperations staff and the central Sovietgovernment and between the Soviet government andits neighbors.

Additionally, better information sharingcould have improved the response to Bhopal, theBeirut terrorist attack, and the Mexico Cityearthquake. William Petak (PAR, 1985:12)challenges emergency managers to developproactive approaches. Increased preparedness,increased cooperation between cognizantorganizations, and increased integration ofhazard prevention (engineered safety systems)and hazard response (emergency managementsystems) requires improved information sharing.

Emergency management is a broad,multidisciplin&ry topic and one long neglected.Research in many areas of emergency managementcan, we believe, help emergency managers meetthe significant challenge posed by majoremergencies.

We have chosen to examine one function inemergency management, (centralized command andcontrol in a facility normally called anEmergency Operations Center), in the context ofa concept applicable to all functions inemergency management (shared informationprocessing). To do this, research is organizedin three parts. Part one reviews the conceptsof and requirements for shared informationprocessing. Part two establishescharacteristics of EOCs and relates thesecharacteristics to Information sharing. Partthree, the conclusion, suggests improvements inEOC design built on information sharing conceptsand recommends other emergency managementimprovements.

I. INFORMATIONMANAGEMENT

SHARING AND EMERGENCY

What is information? What is sharing? Whyis sharing information important? We useBlumenthal's definition of data andinformation: "A datum is an uninterpreted raw

319

320

statement of fact. Information is datarecorded, classified, organized, related, orinterpreted within context to convey meaning."(Blumenthal, 1969: 27). Thus a fact such as,"The radiation dose Inside the reactorcontainment of a nuclear facility is 2,500rem/hr.," is data. When related to standardsand interpreted, the information emerges thatthe radiation dose is too high for humanexposure.

Sharing is both borrowing and lending dataand information. Effective results- from sharingonly come about when we don't hoard or protectthat which our neighbor needs. However, toomuch information is just as bad. Filtering maybe necessary to extract the most relevantinformation and not overload emergencymanagers. Further, sharing may becounterproductive if we don't return theborrowed item in original condition; otherwiseit may not be usable again.

A. Why Is Shared Information Processing InEmergency Management Important?

The manager's need for information is theforcing function in this age of information.Increasing levels of technologicalsophistication make more information availableto mor« organizations and people. Organizationsspecialize in order to cope with theever-increasing complexities of technology.Specialization leads to interdependence and theincreasing need to share information.

We are also realizing now, more clearly thanbefore, that our resources are finite.Management Systems Laboratories' hypothesis Isthat organizations previously acted to reduceinformation sharing by operating with slackresources and operating in self-containedtasks. Organizational interdependencies and abroader base of those participating in decisionmaking combined with resource constraints arenow forcing organizations to redesign theiroperations based on shared informationprocessing.

The manager's need for information is alsothe forcing function In emergency management.Industry has produced new hazards to add tonatural disasters. The public has access tomore information about these hazards and demandsthat the government provide protection. Ifgovernment wishes to maintain its credibilityand if public administrators and electedofficials wish to maintain their offices, theymust respond to this challenge.

Previously, when the lack of global,instantaneous communications made the world alarger place, informal policies In emergencymanagement were possible. Information aboutcatastrophic emergencies hundreds or thousandsof miles away didn't seem threatening. Today,the world has become very small. The mediaintensively covers a wide range of potentialhazards and the public demands protection fromrisks that seem very near. Comprehensive,Integrated capabilities are required to respondto the demand for the better management ofemergencies. Interdependence, broadeneddecision making, and constrained resources are,

therefore, forcing information sharing inemergency management.

B. Emergency Response And Emergency ManagementAre Different.

Emergency response is a technical, hands-onactivity. The operation involves physical itemssuch as bulldozers, aircraft, radiologicalmonitoring equipment, and medical equipment.The operation of emergency response cannot beseen in isolation. Successful emergencyresponse depends on successful emergencymanagement.

Emergency management includes the activitiesthat support and direct the response. Theoperation of emergency management involves data,information, information processing, anddecision mechanisms. The importance ofemergency management to emergency response wasillustrated in Mexico City. The data providingthe number of collapsed buildings and the numberof trapped people waB not received, sorted,manipulated, and portrayed to Mexican governmentofficials until forty-eight hours after theearthquake. By that time most of the trappedpeople could not be helped. Resources froawithin Mexico could not be redirected andresources from outside Mexico could not berequested until the government knew the scope ofthe problem. Better information sharing amongthe agencies of the Mexican government may havesaved thousands of lives.

C. Effective Decisions Require Good Information:.

Complex technological or natural disastersrequire a cooperative, Integrated response frommany response groups at all levels ofgovernment. Data oust be quickly and accuratelygathered, stored, manipulated, portrayed, andcommunicated as information to many groups.Information must be shared.

The "place" where data becomes informationIs a critical component in the effectivemanagement of emergencies. This place is theEmergency Operations Center. Raw data is notusually forwarded lip the reporting chain. Evenafter the EOC extracts information froaemergency incident data, the Information Is bothfunneled and filtered before being forwarded.

Funneling involves suciwrlzlng, collecting,Interpreting, and disseminating dissimilar butrelated data. This data manipulation requiresgreat technical expertise and high ethicalstandards (to prevent hoarding data to make onegroup, usually one's own, look good) to avoidmisinterpretation and distortion. Filtering isthe process of reflecting ' rends or patterns ininformation. Funneling atA, filtering extractsthe most relevant Information.

This discussion SHows that the quality ofthe Information available to most governmentofficials and to the public is dependent on theoperation of the EOC as an informationprocessing and information sharing facility.The EOC links the complex management chaintogether. Effective information processing andsharing in the EOC is a necessary condition foreffective emergency management.

321

II. INFORMATION SHARING IN THE EMERGENCYOPERATIONS CENTER

Many individual characteristics areimportant in the analysis of emergencyoperations centers. We found it useful tocombine many of the individual characteristicsinto broad categories. This method permitscomparison of EOC characteristics. However,prior to our discussion of EOC characteristicswe believe it helpful to provide a definition ofwhat an EOC is and what its functions are.

A. What Is An Emergency Operations Center?

An EOC is a facility, which may be fixed ormobile, that serves as the primary point ofcommand, control, and communications inemergency response. EOCs, and the personnel inthem, are normally removed from the physicalsite of the emergency and so do not directlyrespond to the incident. Generally, EOCs arenot on-scene command posts used for directoperational control of emergency responseelements. Usually, that direct operationalcontrol is delegated to security, healthprotection, operational, or fire departmentunits which follow established emergencyprocedures or direction from the EOC.

Because EOCs are most often remote from theincident site, and the actual emergency responseelements, emergency managers depend oncommunications systems to supply informationabout the emergency. EOCs consist of equipment,personnel, and methods of operation to processinformation and make: and direct theimplementation of decisions that terminateemergencies.

An activated EOC utilizes information tocoordinate and manage an emergency response at ahigher-than-operational level. This may includeordering specific actions by response elements,notification of appropriate governmentjurisdictions and the media, and requests forassistance. In summary, an EOC processesinformation, and based on the superior knowledgethat the information provides, determinesresponse and supporting actions to terminate anemergency.

To illustrate the complex paths emergencyinformation and decision making follow, we haveincluded Figure 1, Successful response requireswell-defined and tested paths for emergencyinformation flow and decision making. Thisfigure represents typical emergency flow forU.S. Department of Energy (DOE) facilities.Notice that the decision path, which includesboth the actions for response elements andrequests for assistance, is simpler than theinformation (notification) path. TheInformation path may involve many peripheralorganizations (e.g., the media) that couldsignificantly Impact the responding organizationover the long tern.

The EOC is often the major source ofinformation to the public and the media. Theimportance of this communication cannot beoverestimated. History often judges the successof an emergency response more on the perceptions

created by the media than the reality buried inofficial records. Therefore, although morecomplex and often time consuming duringemergency response, the information(notification) requirements cannot beoverlooked. The primary purpose of any EOC isto manage information and decisions. Thequantity of Information is great and bothinformation funneling and filtering are requiredto ensure high quality, relevant information anddecisions result.

DOE-HQ ,, „ Other Federal. - Authorities

X AMedia

LocalAuthorities

DOE OperationsOffice

n ,Facility EOC '

nResponse Elements

StateEOCs

Decision Path-

Information Path.

Figure 1. As shown in this DOE fexample,successful response requires well-defined andtested paths for emergency information flow anddecision making.

In the information processing role, whichincludes the orderly flow of relevantinformation and decision making, EOCs are theprimary interface point between on-sceneresponse and the outside world. The EOC servesas a two-way clearinghouse for information: forinformation from the incident scene and forsupporting information from outside the responseorganization. This concept is criticallyimportant for information sharing in EOCs. Astechnology permits faster and faster access andprocessing of information, there is a tendencyto be inundated by information. Post-incidentevaluations of many emergencies typically haverevealed that the management systems containedmuch significant Information that was not usedin the response. This occurs because of a lackof ability to focus on what is relevant(Information Technology for EaergencyManagement, 1984: 152-153).

We find that every EOC, as a two-wayclearinghouse, must include a carefully designedand implemented functional process that canhandle a great deal of information and separateit by relevance to the response.

First, by funneling (or channeling)information is directed to the appropriate place

322

for filtering. Second, and most important, isthe filtering of information. Information fromthe incident (internal) as well as informationfrom organizations outside the response(external) must be filtered. Only in this wayis relevant information obtained for accurateand appropriate decision making. Our concept isrepresented in Figure 2, The funnaling andfiltering of information is necessary to supportdecision making based on the most relevantinformation. It is becoming more and moreimportant for EOCs to facilitate the sharing ofinformation among those in the facility andbetween the EOC and the great number of outsideorganizations with a stake in a coordinated andeffective emergency response. Simply put, thisrequires EOCs defined as information sharingfacilities.

Information From Incident

ResponseDecisions

U 11J / /

Internal Filtering Function

Processor

(decision-making]

Eternal Filtering Function

ExternalSupport

: / M t n \/ Information Outside Responding EOC

Figure 2. The 'funne^ing and filtering ofinformation is necessar'y to support decisionmaking based on the most relevant information.

B. Characteristics of EQCs

We have combined many individualcharacteristics into, several broad categoriesrelevant to information sharing. These samecategories will permit application and designcriteria to be established.

1. The Emergency OrganizationThe eaergency organization executes

emergency management responsibilities.Differences and simil; -Ities between facilitiesand organizations define unique domains ofemergency management responsibility.Similarities in potential emergencies, policyguidance, and overall mission make facilitiesand their orga ations comparable. Differencesin geographic proximity to the facilitiesmanaged, the organizational history and culture,and organization structure and procedures usedprovides a substantial and meaningful contrastbetween EOC emergency organizations.

When activated, the EOC emergency stafffinds it difficult to make the transition fromtheir normal management responsibilities totheir emergency roles. Managers have collateralresponsibilities and this must be taken intoconsideration in emergency planning andpreparedness. Managers must be trained andexercised in their emergency roles and in theprocess of working with unfamiliarorganizations. The role of training cannot beoverstated. Managers cannot devote much time toemergency preparedness and their complexemergency management roles require orientationand familiarization. Emergency proceduresrequire drills in order to be used proficientlyin the high-stress environment of a response.Creative and innovative training is a must for asmoothly functioning EOC emergency staff.

Shared information processing can beenhanced or hindered by the physical layout ofthe EOC. Typically, a technical, specializedsupport cadre is available to receive, sort,manipulate, and portray emergency information.The management cadre uses this information tomake critical emergency decisions.

The process of receiving, sorting,manipulating, and portraying emergencyinformation is noisy. This area must beseparated from the area where the managementcadre works. In this way, internal (to the EOC)information is filtered by the support cadre foruse by the management cadre. Well-rehearsedprocedures, trust, confidence, and teamwork frompeople not used to working as a team arerequired for an emergency organization to workat all. The degree of success depends on thededication of both the people involved and theirparent organization, and on the organization ofemergency planning and preparedness.

2. Internal NotificationsAn Internal notification system gets the

right people to the EOC quickly. Following thereport of the emergency and the decision toactive the EOC, the communications staff in theEOC contacts members of the eaergencyorganization. People cannot share informationuntil they can be Hatched up with theinformation. A delay in assembling the EOCemergency organization can have catastrophicconsequences.

3. Internal Information FlowAccurate, reliable, and timely information

about an emergency is necessary for a successfulresponse. The internal information flow ensuresinformation reliability.

EOC information flow usually parallels thestructure and physical positioning of theemergency organization. From the incident tothe EOC, and once in the EOC, from functionalgroup to functional group, the path informationfollows determines its efficacy. If informationpasses outside its functional path, e.g.,security information processed by healthphysicists, the information becomes contaminatedwith loss of confidence and biased perceptions.Without strict control over information flowfrom source to decision point, the decision

323

maker loses confidence in the reliability ofemergency information'

4. Emergency Information DisplayAlthough part of internal information flow andmanagement, the display of emergency informationin the EOC is so critical it is consideredseparately. The display of emergencyinformation must support the next decision. Forthis reason, status rather than eventinformation is usually portrayed in EOCs. Eventinformation displays show multJ-incidentinformation or the chronological history ofevent status. Status information is sorted anddisplayed by function. Status informationlessens the tendency of the emergency manager toget caught up in the story that is portrayed onan event display.

Emergency information may be displayed by avariety of means of varying degrees oftechnological sophistication. The technologicalsophistication has little to do with theeffectiveness of the display. Of criticalimportance is the placement of the display: canthe emergency managers see the information theyneed? Additionally, the accuracy and timelinessof the information displayed is critical. Ifthe emergency manager does not have faith in thedisplay process, the manager will waste valuabletime confirming information.

5. External Information flowExternal information flow supports both the

information path and the decision pathillustrated in Figure 1. The informationdisplay techniques are also critically importanthere. Information should be portrayed inexactly the same format in a hierarchicallyseparate but organizationally similar EOC.Thus, the information displays in the facility,operations office, and headquarters EOCs ofDOE, as one example, should be identical. (Theyare notJ)

Well-developed and rehearsed procedures arenecessary to ensure that the media, and throughthe media, the public are provided incidentstatus quickly and reliably. This is difficultto achieve in reality because of the need toprotect classified information and because mosttop managers will not delegate the approvalprocess for press releases. The stakes are toohigh.

6. Emergency Communications SystemsControl over the flow of communications

internally and externally ensures the flow ofdata and information is fast and reliable.Communications systems support all otherfunctions, but particularly the internal andexternal information flows. Specialists in theEOC normally control the communicationssystems. These specialists receive and sortdata (funnel and filter). Data may be sorted byfunction or priority. In order to ensure thatonly high-quality, relevant information ends upon the displays the management cadre needs, thespecialists must properly process theinformation as it enters the EOC.

The communications systems are critical to

the information processing function of the EOC.Indeed, information processing cannot occurwithout good communications. These systemscould include telephones (there are a greatvariety of these, some quite exotic andexpensive), radios (usually HF, VHF, and UHF),teletypes (which can operate from radios ortelephone lines), runners (which transferinformation withia the EOC or for shortdistances outside the EOC), and finally othersystems which could include microwave orsatellite or other more exotic types.

These communications systems represent asignificsnt investment in any emergency responsesystem. Effective planning and preparedness tobest use them is essential.

7. Maintaining the Historical RecordEmergency response provides valuable

feedback that supports training. Maintaining ahistorical record of the response allowsemergency response personnel to review theiractions, decisions, and assessments and improvethem. Legal and reporting requirements are alsomet.

III. CONCLUSION

Complex organizations, especially public ones,are driven by many forces. A beginning step inan analysis of any such complex publicorganization is to define the organization'soperation, that is, what is being managed. Theemergency operations center is a facility thatmanages emergencies. Such facilities do notexecute emergency response actions but direct,control, oversee, and coordinate the actions ofone or many response groups. Additionally,actions outside of the response may Impact theEOC. Reporting requirements dictated by law ordepartmental procedure, press relations, and theconcerns of the public must be addressed.Almost any eaergency above the trivial willbecome complex and require the activation ofspecial organizations and facilities to respondand manage the response to such emergencies.The EOC is the central focus of emergencymanagement activities above the operationallevel.

Examining the response to recentemergencies, such as Three Mile Island, Bhopal,Mexico City, or Chernobyl, it is obvious thatimprovements in all functions of emergencyresponse and emergency management aredesperately needed. If we are to supportemergency management above the operationallevel, we must begin with the EOC.

The EOC is a central command, control, andcommunication facility in emergency management.We have described the actions of the EOC interms of seven critical functions: theemergency organization, internal notifications,internal information flow, information display,external information flow, communicationssystems, and maintaining the historical record.How these functions are executed largelydetermines the success or failure of emergencymanagement actions. We believe that bydescribing these functions in terms of shared

324

information processing, the probability ofsuccess in emergency management can be improved.

Sharing information begins at theoperational level. The operational responsepersonnel must collect and forward data aboutthe incident to the EOC. This takes tine. Whatimpetus do these people have to do a good job atthis data collection and communication? Becauseoperational responders are dependent on the EOCfor support they are willing to shareinformation with the EOC. Interdependence leadsto information sharing. The operationalpersonnel must have confidence that the EOC willprovide the necessary support or they will noteffectively share information, will not take thetime to collect and communicate incident data tothe EOC, and will not execute the directions ofthe EOC.

When the emergency begins, the emergencyorganization is activated in the EOC.Information from the incident begins to pour in,as do requests for information from higherlevels in the organization, other organizations,the press, the public, and employees. The EOCmust make decisions about response actions,resource allocation, press releases, and a hostof other critical matters. We have observedthat these decisions are made. All of thefunctions of the EOC identified above areperformed. We have also observed that some EOCsdo better than others.

At this point it is only possible to saythat empirical evidence suggests that EOCsdesigned and operated as shared informationprocessing facilities do a better job ofmanaging emergencies. As one example, some EOCsseparate their managers from those personnel whoreceive and process the incoming data andconvert that data into useful information.

Separating personnel in the EOC on the basisof their information processing function reducesconfusion and focuses attention on individualresponsibilities. Further, technicalspecialists usually process information anddetermine what is Important and what is not(funneling and filtering). These specialistscategorize information and add a bias. Managersalso have difficulty not looking over theshoulder of these specialists. Being aware ofthe garbage-in/garbage-out syndrome, managersare concerned about the ability of thesespecialists to give them quality information.

In some cases, budgets limit the size of theEOC and everyone is crowded into one or severalrooms. These organizations often save thousandsof dollars, but if a major emergency occurs theorganization will pay for this economy. Anotherexample of an economy-driven criterion Is theoften-noted lack of concern over the ability ofthe EOC to maintain a historical record. Thehistorical record is the best method availableto review actions and suggest improvements. Theuse at flight data recorders in determining thecause of aircraft accidents and preventingfuture accidents is evidence of the need for asimilar capability in EOCs.

EOC design and operation based on sharedInformation processing will Bignificantlyincrease the probability of successful emergency

management. An emergency preparedness programincluding adequate emergency managementdocuments, training, drills and exercises, andevaluations and audits i« necessary fororganizations to refine the skills of EOCpersonnel and mold these personnel into a team.A team of EOC professionals will be moreeffective if their operating procedures arebased on processing information, becauseprocessing Information is the operation of EOCs.

EOCs collect data, process that data intoinformation, and then use that information tomake decisions and direct response actions.EOCs will be more successful if theirorganization, activation processes, Internalinformation flow, information display, externalinformation flow, communications, and recordkeeping operate for maximum efficiency andeffectiveness in processing and usinginformation. The higher the level ofpreparedness in these functions, the moreeffective the EOC will be. In future papers, adescriptive model of EOC operation should bedeveloped. Eventually, predictive andprescriptive models will allow us to Bay withmore confidence what will Improve EOCcapability, provide a measurement and evaluationof that capability, and Improve EOC operations.

BIBLIOGRAPHY

BlymenthaJ, S.C., Management and InformationSystems—A Framework for Planning andDevelopment, Englewood Cliffs,"L-autice-Hall, Inc.. 1969.

N.J.:

Howard, Melissa M. and Philip Kronenberg."Decision Support of the Strategic Manager:Concepts and Opportunities in a New Direction,"Presented at the 47th American Society forPublic Administration Conference, Anaheim, 1986.

Kurstedt, Harold A., R. Martin Jones, and LouisI. Middleman. "A DSS Model Applied to GovernmentOrganizations Functioning as InformationProcessors," Proceeding of the Sixth AnnualMeeting, American Society for EngineeringManagement, September 1965, pp. 61-68.

Mathes, J. C. "Three Mile Island: TheManagement- Communication Role." EngineeringManagement International 1986, pp. 261-268.

Perrow, Charles. Normal Accidents: Living WithHigh Risk Technologies. New York: BasicBooks. 1984.

Information Technology for EmergencyManagement, Report Prepared by theCongressional Research Service, Library ofCongress, U.S. Washington D.C.: U.S. GovernmentPrinting Office, 1984.

ANS Topical Meeting on Radiological Accidents-Perspectives and Emergency Planning

The NRC Operations Center's FunctionEric W. Weiss

ABSTRACT The Nuclear Regulatory Commission hasmaintained a 24-hour-a-day, 365-days-a-year,manned Operations Center since- the emergencyincident at the Three Mile Island Nuclear PowerPlant in 1979. The Center functions as theNRC's point of direct communication throughdedicated telephone lines for reports ofsignificant events at licensed nuclear powerplants and certain fuel cycle facilities. TheCenter has become a key element in the agency'semergency preparedness.

The effectiveness of the NRC OperationsCenter1 depends in large measure on complete andaccurate reports from the licensees. Theinformation provided is used to: identifygeneric safety issues and precursor events thatmay compromise plant safety; develop licenseeperformance trends that are used to adjust NRCregulatory emphasis; and, evaluate and providefor the appropriate NRC response to events in areal time mode.

I. INTRODUCTION

The Nuclear Regulatory Commission hasmaintained a 24-hour-a-day, 365-days-a-year,manned Operations Center since the emergencyincident at the Three Mile Island Nuclear PowerPlant in 1979. The Center, which is located inBethesda, Maryland, functions as the NRC's pointof direct communication through dedicatedtelephone lines for reports of significantevents at licensed nuclear power plants andcertain fuel cycle facilities. The Center alsohas a commercial telephone number which is(301)951-0550. At night and on weekends andholidays when the NRC regional offices areclosed, the main telephone numbers for the NRCregional offices play a recorded message to callthe Operations Center in the event of anemergency. The NRC Operations Center receives awide variety of calls and has become a keyelement in the agency's emergency preparedness.

The staff at the Operations Center evaluatethe telephone notifications and, depending onthe safety significance of the particular event,

notify other appropriate NRC personnel and otherFederal agencies. Examples of Federal agenciesthat are sometimes notified include the FederalEmergency Management Agency (FEMA), theDepartment of Energy (DOE) and the Department ofTransportation (DOT). In all cases the NRCregional office responsible for the facilityreporting the event is notified. Response to anevent may vary from simply recording thecircumstances of the event for later evaluationto immediately activating response teams withinheadquarters and the appropriate NRC region.Events are monitored by the NRC while they arein progress from the standpoint of actions toprotect the health and safety of the public.

During the early hours of an emergency, theOperations Center becomes the focal point foraction by the NRC. The NRC monitors thelicensee's actions during an emergency to assureappropriate protective action is being takenwith respect to offsite recommendations. Whenrequested, the NRC supports the licensee withtechnical analysis and coordinates logisticssupport. The NRC supports offsite authorities,including confirming the licensee'srecommendations. The NRC keeps other Federalagencies and entities informed of the status ofthe incident. The NRC keeps the media informedof the NRC's knowledge of the status of theincident, including coordination with otherpublic affairs groups.

Not every event telephoned into theOperations Center is an emerqency; somenon-emergency events have important genericimplications for other operat.inq nuclear powerplants that could lead to an emergency, if le*tunchecked. Consequently, each event called intothe Operations Center or reported by a regionaloffice is evaluated to determine any genericimplications for similar facilities. Eventreports ere screened during the first hours ofthe first working day following the receipt of anotification. Events that may be significantfrom a generic standpoint then receiveadditional in-depth evaluation. For eventsfound to have significant generic implications,the NRC issues an Information Notice or aBulletin to the appropriate licensees andconstruction permit holders.

325

326

II. OPERATIONS CENTER PERSONNEL AND EQUIPMENT

The Operations Center is continuouslymanned by a Headouarters Operations Officer whois an engineer or scientist specifically trainedfor that job. The NRC considers events analysisan important part of the function of theHeadquarters Operations Officer. By immediatelyanalyzing the events that are reported, aserious event can be recognized at its earlieststages and events with generic implications canbe treated appropriately.

Many of our Headquarters OperationsOfficers have advanced degrees and they allreceive in-depth training on reactor design andoperations at the NRC Technical Training Centerin Chattanooga, Tennessee. Typically the courseof instruction for a Headouar'ters OperationsOfficer lasts about I?, weeks in Chattanooga plusa couple more weeks in Bethesda on the equipmentand procedures for the Operations Center.

The Headquarters Operations Officer hasspecial equipment to aid him in his function.The Operations Center has two dedicatedcomputers that serve to record, disseminate, andanalyze events. The telephone system for theOperations Center is more complicated than thatused by most telephone operators. There are twodedicated sets of telephone lines to eachlicensed nuclear power and fuel facility. Theone most often mentioned is the EmergencyNotification System (ENS) that appears as a redtelephone in the control room of each nuclearpower pl'.ni and rings in the Operations Centeronce it is picked up. The other dedicatedtelephone system is the Health Physics Network(HPN). The HPN is used during an emergency toestablish conference calls between healthphysicists involved in the event. Besides thetwo dedicated telephone systems the HeadquartersOperations Officer has access to the commercialtelephone system and the Federal Telecommunica-tions System (FTS).

Thr Operations Center also has a dedicatedtelephone line to FEHA. The dedicated telephonelines are a real advantage during an emergencywhen ordinary phone systems tend to becomeoverloaded.

The Headquarters Operations Officer is ableto control all of the various phone systemsavailable by using a electronic PBX (privatebranch exchange). This system offers theHeadquarters Operations Officer six telephonebridges to establish conference calls during anevent. During an emergency, there are usuallyat least four telephone bridges established. Oneis used for the ENS, one for the HPN, and thenthere are two counterpart links establishedbetween headquarters and regional personnel sothat headquarters and regional personnel candiscuss an event without interrupting the flowof information from the site.

The Headquarters Operations Officer who isusual'y alone in the Operations Center, isjoined by several teams during an emergency.Without attempting to describe everyone who hasa role in the Operations Center, three teams areworth mentioning to give an overall picture ofthe function of the Operations Center during anemergency. The Reactor Safety Team follows the

course of the event and attempts to predictfuture plant responses, e.g., Is the breakgetting bigger; what alternatives are there forreaching cold shutdown? The Protective MeasuresTeam follows the course of the event from aradiological point of view, e.g.. Will there beany health consequences; is evacuationwarranted? The Executive Team directs theagency response to the emergency.

III. PROBLEMS AND TWO EXAMPLES

The chief problems with reports made to theNRC Operations Center have been lack ofcompleteness in the reports and inadequatequalifications of some of those attempting tomake reports to the Operations Center.

Two recent events, one at the Davis-Bessenuclear power plant near Toledo, Ohio and theother at the Point Beach nuclear power plantnear Two Rivers, Wisconsin, highlight the needfor more complete reports and improved licenseeresponse to emergency classification.

A. Davis-Besse

At 1:35 a.m. on June 9, 1985, theDavis-Besse plant experienced a complete loss ofmain and auxiliary feedwater for nearly 12minutes. The emergency plan identified the lossof feedwater event as a Site Area Emergency,which is the second most serious of the fouremergency classes defined by the NRC's regula-tions. However, it appears that allknowledgeable personnel in the control room wereoccupied with stabilizing the plant and, thus,were not able to classify the event as a SiteArea Emergency and activate the emergency plan.Had the plant not been brought to a stablecondition quickly and had plant safety furtherdegraded, it is possible that the efforts of allknowledgeable personnel in the control roomwould have been required for recovery efforts,further delaying initiation of appropriateonsite and offsite emergency response.

At 2:11 a.m., which is C6 minutes after theloss of all feedwater, the shift technicaladvisor (STA) called the NRC Operations Centerfrom the control room using the EmergencyNotification System to report the event pursuantto the Commission's regulations, which allow amaximum of 1-hour from the time that anemergency is declared until it is reported tothe NRC Operations Center. At the beginning ofthe event, the STA had been in his quarters inthe administration building, which is outsidethe protected area about a half mile from theplant. Although the'STA mentioned the trip ofthe main and auxiliary feedwater pumps, the STAdid not describe the length of time that theplant was totally without feedwater or thedifficulty the plant had in restoring auxiliaryfeedwater. No Emergency Class was declared, norwas the fact conveyed to the NRC that plantconditions which warranted the declaration of aSite Area Emergency had existed for nearly 12minutes.

At 2:26 a.m., the STA informed the NRC thatan Unusual Event, the lowest of the 4 Emergency

327

Classes, had been declared at 2:25 a.m. The STAalso informed the NRC that although theemergency plan identified the total loss offeedwater event as a Site Area Emergency, theplant was no longer in this emergency actionlevel at this time. At 2:29 a.m., the licenseeinformed the county that an Unusual Event hadbeen declared. The licensee depended on aprocedure that required the county to notify thestate representative. However, because thecounty could not reach the local state represent-ative, the State of Ohio was not notified of theUnusual Event declaration until after the eventhad been terminated, more than 6 hours after itsdeclaration.

At Davis-Besse, the emergency plan isinitially implemented by the shift supervisor,who also has primary responsibility for ensuringthat the plant is maintained in a safe condi-tion. Because of competing priorities of (1)directing attention to necessary recoveryactions to obtain a safe and stable plant and (2)reviewing the emergency plan and initiating itsactions, there was a substantial delay indeclaring an Emergency Class and implementing theemergency plan. If the June 9 event hadprogressed in severity, valuable time needed toinitiate appropriate onsite and offsite responseto the emergency would have been lost.

Corrective actions being undertaken by thelicensee as a result of this event include anumber operational and procedural changes thatinclude but are not limited to the following:Changing the STA shift schedule from a 24-hourduty day to rotating 12-hour shifts. Having theSTA spend the entire shift within the protectedarea, and having the STA office located within 1to 2 minutes of the control room. Training theSTA as an Interim Emergency Duty Officer toadvise the shift supervisor in event classification and protective action. Having the licenseemake emergency notifications directly to theState of Ohio.

G. Point Beach

On July 25, 1985, at 7:25 a.m. (easterntime), Point Beach Unit 1 experienced an eventinvolving loss of offsite power. Point BeachUnit 2 continued to operate normally during thisevent. Because of the incomplete understandingof the event by those making the notification tothe NRC Operations Center, the NRC OperationsCenter was not made aware of the details of theevent. At 7:37 a.m., a security guard calledthe NRC Operations Center to notify the NRC thatPoint Beach Unit 1 had declared an Unusual

Event. The explanation for the Unusual Eventwas that the plant had a turbine runback. Thisdid not make sense to the HeadquartersOperations Officer because a turbine runback isneither an emergency nor even a reportablenon-emergency event in itself. When the NRCHeadquarters Operations Officer asked questions,the security guard was unable to provideadditional information because of his limitedtechnical knowledge cf the plant and because thecall was made from a location outside thecontrol room where the security guard could notobtain additional information from the operatorsinvolved.

The Headquarters Operations Officer called-the control room, and as a result of askingquestions learned that a station transformer hadbeen lost. However, not until 2\ hours later,when the plant notified the NRC HeadquartersOperations Officer that the Unusual Event wasterminated, did the Headquarters OperationsOfficer learn that there had actually been aloss of offsite power. A loss of offsite powerat a nuclear power plant is a serious eventwarranting the declaration of an Unusual Event.

On October 15, 1985, the NRC issuedInformation Notice (IN) 85-80 to describe thesetwo instances when an emergency condition wasnot classified and declared in a timely manner.Although events such as these are the exceptionrather than the rule, the NRC has counseledlicensees that it is the licensee'sresponsibility to ensure that adequatepersonnel, knowledgeable about plant conditionsand emergency plan implementing procedures, areavailable on shift to assist the shiftsupervisor to classify an emergency and activatethe emergency plan, including make appropriatenotification, without interfering with plantoperation.

IV. CONCLUSION

The effectiveness of the NRC OperationsCenter depends in large measure on complete andaccurate reports from the licensees. Theinformation provided is used to: identifygeneric safety issues and precursor events thatmay compromise plant safety; develop licenseeperformance trends that are used to adjust NRCregulatory emphasis; and, evaluate and providefor the appropriate NRC response to events in areal time mode.

If you are planning a trip to Bethesda,Maryland, we would be delighted to makearrangements for a tour of the OperationsCenter.

Section VI

Medical Issues

Chairman: Clarence LushbaughOak Ridge Associated Universities

ANS Topical Meeting on Radiological Accidents-Perspectives and Emergency Planning

Dose-Rate Models for Human Survival After Exposureto Ionizing Radiation"

Troyce D. Jones, M. D. Morris, and R. W. Young

ABSTRACT This paper reviews new estimates ofthe LD50 in man by Mole and by Rotblat, thebiological processes contributing to henatologicdeath, the collection of animal experimentsdealing with hematologic death, and the use ofregression analysis to make new estimates ofhuman mortality based on all relevant animalstudies. Regression analysis of animal mortal-ity data has shown that mortality is dependentstrongly on dose rate, species, body weight, andtine interval over which the exposure isdelivered. The model has predicted human LDJOSof 194, 250, 310, and 360 rad to marrow when theexposure time is a minute, an hour, a day, and aweek, respectively.

I. BACKGROUND

A. Mechanisms of DeathWhen mammals are exposed to high doses

of ionizing radiations, blood lymphocytes andstem cells of the active bone marrow are killed.Mammals may die from infection or hemorrhagewhen these cell populations drop below certaincritical levels. The time to death depends uponthe number of cells killed; the species andstrain of the mammal; and the cage/hospital caregiven, including therapeutic support, barriernursing, nutrition, etc. This mode of death iscommonly referred to as hematologic syndrome (ordeath from hemapoietic depression).

Hematologic death and modes of death resul-ting from gastrointestinal (GI) and central ner-vous (CN) system damage are described in Langham(1967) and many other sources (e.g., Baum et»1., 1984). However, this review will discuss,in some detail, the biological and physical con-ditions contributing to hematological depres-sion, the relevant animal studies from radiationbiology, and human radiation accident and thera-peutic experiences. This background section

will establish the justification for newanalytical tools to be used in modeling the LD50for man.

Bergonie and Tribondeau (1906) proposedthat the level of radiosensitivity of an organor tissue is related to (1) the degree of dif-ferentiation of its cells morphologically andphysiologically, (2) the mitotic activity, and(3) the length of time that the cells remain inan active stage of proliferation, which includesthe number of divisions between the youngestprecursor cell and the mature functional (ordifferentiated) cell. This proliferation ofcells, which has commenced terminal differentia-tion, is commonly referred to as amplificationand is especially important in maintaining suf-ficient numbers of lymphocytes (to fight infec-tion) and platelets (to prevent post-irradiationhemorrhage). Because stem cells of the marrowtie more radiosensitive than the rapidly proli-ferating crypt cells of the GI system or thehighly differentiated cells of the CN system,survival of the organism is dependent upon thebone marrow although the time to death of alethally irradiated animal may be determined bydamage to the GI or CN systems (Langham, 1967;Baum et al., 1984).

According to Alper (1979), the killing ofanimals by radiation is determined by the deathof "target cells" in "target tissues," and thisconcept is now a "basic part of the framework ofradiological thinking." The United NationsScientific Committee on Effects of Atomic Radia-tion (UNSCEAR, 1982) has further broadened thisconcept by proposing the basic premise that thenonstochastic response of a tissue depends uponthe level of cell killing. ("Nonstochastic" isused to describe radiation-induced injurieswhere both the frequency of occurrence and theseverity of the injury are porportional to theradiation dose.) Thus, for hematopoietic deaths,there is no controversy about the sequela ofeffects that precede death. Following whole-body exposure to lethal doses of radiation

•Research sponsored by the Defense Nuclear Agency under Task Code U99QMXMH and Work Unit Code 00018 ofIACRO DNA 85-903 with the U.S. Department of Energy, under contract DE-ACO5-84OR214OO with MartinMarietta Energy Systems, Inc.

331

332

(below about 1000 rad given over a short timeinterval), death in small mammals occurs within30 days and in larger mammals within 60 days.

When the radiation is delivered over alonger time, the magnitude of the lett;l dose isincreased, and the time to death become; longerand less sharply defined. These changes canderive from two processes. First (if the radia-tion dose is from photons), subcellular or enzy-matic repair of sublethal lesions in nuclear DNAcan occur before a second photon (or photon-induced lesion) can further damage the site andkill the cell. Second, when the dose rate issufficiently low (i.e., the dose is given over atine interval greater than 30 minutes), compen-satory cellular proliferation begins in anattempt to restore tissue homeostasis. Thus,for low-dose-rate exposures to either high orlow linear energy transfer (LET) radiations, thesurvival time can be quite long. The processhas been described by Bond, et al. (1965):

Proliferating cell systems like thosefound in hemopoiesis can, in opera-tion, be likened to a retail dealervending several items. If the factorymanufacturing a particular item isdamaged so that production is reducedor stopped, the retail dealer is notaffected until stocks in the chain ofsupply linns are exhausted. Eachindividual item (cell) in the storeeven then remains as good or as "func-tional" as ever, and the segment ofthe vendor's business (organ)represented by that item is not affec-ted until the number of individualitems falls below a critical level oris exhausted. The entire business(mammal) is not seriously hurt or des-troyed unless that particular itemrepresented a major part of (was vitalto) his operation. Survival will thenbe possible only if the factory can beput back into operation reasonablyquickly (rapid regeneration), or if asubstitute source or product can beused temporarily (symptomatic or sub-stitution therapy) until restorationof the factory eventually takes place.

When humans have been accidentally ortherapeutically exposed or when test animalshave been irradiated, mortality it commonly an"all or none" event with respect to proportionkilled in a population of individuals. That is,there is some high sublethal dose where no sub-jects die and some dose about twofold higherwhere very few, if any, individuals survive.The narrow transition zone (wherein some indivi-duals die) is defined by a dose range where theupper dose is only 2 to 3 times that of thelower dose. Thus, because the mortality func-tion is extremely steep, the dose that is lethalto 50% of the exposed individuals can in effectcharacterize the entire response function formost practical considerations with respect tonuclear safety, civil defense, and radiotherapy.

In spite of the large numbers of documentedhuman exposures (Lushbaugh, 1969), there areinadequate human data to serve as a basis for

promulgating an LD50 for man or to study how thehuman LD50 changes with different biological andphysical conditions.

Historically, the most commonly acceptedLD50 value for man has been that of the NationalCouncil on Radiation Protection and Measurements(NCRP) (NCRP, 1974). For high dose-rate expo-sures such as those that might occur followingthe explosion of a nuclear weapon, the NCRP haspromulgated a value of 450 R (in air), which hasbeen taken by the NCRP to be about 315 rad {tomarrow). This NCRP LD50 value "is the median ofa number of educated guesses made by a group ofU.S. experts in 1949," and there is stillinsufficient human data to substantiate orchange the 1949 value.

Lushbaugh et .... (1967) treated 93 ter-minally ill patients with a dose rate between0.75 and 1.6 R/min. They also gave therapeuticaid to seven victims of the Y-12 radiationaccident. From this combined population. Lush—baugh et al. derived an LD50 of 425 R (in air)or 281 rad (to marrow). Of course, all of the93 terminally ill patients died, but 18 diedwithin a time interval that suggested that theradiation treatment may have predominatedslightly over the progression of death naturallyassociated with the disease. In contrast tostudies of all-accidental human exposures, thisstudy was supported by accurate do sinetry (Becket al., 1971), and the numbers of patients weresufficient to permit a good statistical analysisof mortality.

In Lushbaugh's study, the patients wereterminally ill, a state frequently speculated toresult in increased radiosensitivity. On theother hand, the patients were given state-of-the-art hospital care, which would be expectedto decrease radiosensitivity (NRC, 1975) . Thus,although this study ia without doubt the onlyaccurate source of data that measures the LD50of man, there is much uncertainty as to how toextend the results to different dose rates or to"non-sick" humans exposed under accident condi-tions.

Mole (1984) chose to consider the Y-12,Vinca, and Ewing Sarcoma patients of Rider andHasselback (1967). Mole bases his LD50 of 600 R(in air) and 450 rad (to marrow) on survival of28 individuals from a population of 29 exposed.Mole's analysis is strengthened somewhat becausehe used animal data to describe the shape of themortality curve in the dose-normalizing techni-que of Jones (1981) .

However, doses to accident victims havebeen determined from calculations and experimen-tal "moifk up" and thus are quite unreliable onan individual basis. Mole made a series ofassumptions that combine to build a very highLD50. These assumptions include Mole's beliefthat barrier nur. ing, antibiotics, platelettransfusions, and marrow grafts did not enhancesurvival. There are, however, several sourcesof experimental data to suggest that such pro-cedures are likely to enhance survival by a sig-nificant amount, but these studies were notacceptable to Mole (NRC, 1975; Evans et al.,1985).

As a basis for evaluating the LD50 of man,Rotblat studied the atomic bomb experience in

333

Hiroshima. He presented his analysis at ameeting of the Institute of Medicine at the U.S.National Academy of Sciences (NAS, 1985). Rot-blat found an LD50 of 220 rad tissue terma inair (-250 R) and 154 rad (to marrow). Althoughthe numbers of deaths \nd exposed individualsare adequate for a gocd statistical analysis(i.e., 201 deaths in 765 exposed), Rotblat'sanalysis is quite controversial because of:

1. extremely inaccurate analytical methods (asonly one of several possible examples,Rotblat goes to great effort to compare theshapes of survival curves in a semiquanti-tative manner by transforming the scales onboth the ordinate and the abscissa of onlyone curve; he' then remarks that the simi-larity between the transformed curve forhumans and the untransformed curve for mice"is striking") ,

2. an assumption that all deaths after thefirst day were due to radiation and thatcombined injury from thermal burns andblast did not increase mortality, and

3. the fact that the survival curve isfivefold flatter than any ever observed inany mortality study (Jones, 19bl; Hole,1984) so that a few humans would die atdoses as low as 50-75 rad but some couldsurvive at doses twofold greater than theU>50> Both of these effects have no basisof support in the vast literature on radia-tion therapy and radiation biology (Jones,1981; Mole, 1984).

Many others have promulgated human LD50values over time, but the four studies reviewedhere illustrate the problem adequately. One isfaced wit:h the choice of extrapolating from sickhumans to "normal" humans or analyzing data onnormal humans where the doses and the number ofdeaths per number exposed are unreliable. Evenwith the best analytical methods, a reliableLDjo value must depend upon accurate do timetryand sufficient numbers of exposed individualsand deaths. Some experienced investigatorsexpress a great reluctance to use sick humans asanalogs of normal humans. However, it is ourview that mechanisms of death may not be changedgreatly in tome populations of sick humans andthat mortality models based on carefully donetherapeutic populations provide a technicallyaccurate estimate comparable to analyticalmodels based on many species of test animals.

Thus, from the human data collected todate, it i* not possible to define, with accep-table confidence, an LD50 value (or a mortalityresponse function) or to anticipate how humanmortality varies with dose rate and numerousother physical and biological variables.

Hence, this paper will draw on (1) the vastamount of animal aortality data published in theliterature, (2) well-in own principles from radi-ation biology that will help to fit each ofthese many animal studies into the proper per-spective, and (3) simple unifying dose-responsemodels (Jones, 1981, 19S4) that permit all dataon all specie* and experimental and biological

factors to be analyzed simultaneously instead ofthe conventional approach of sequentiallyanalyzing a few studies, which then usually can-not be merged into a coherent model that can beevaluated for man.

B. Physical and Biological Conditions That

Affect DeathMyelopoiesis is the processes whereby a

marrow stem cell divides in order to maintainthe homeostatic population of stem cells and tosupply differentiated cells to blood, bone, andthymus/lymph. Myelopoiesis is strongly depen-dent upon species and may vary to a lesserdegree within individuals or strains. Myelo-poiesis occurs rapidly in small mammals and moreslowly in large mammals. The rate within anindividual also varies according to homeostaticequilibrium or the need for compensatory cellproliferation to repair tissue injury. Figure 1is an illustration of the species variation ofmyelopoisis (Bond et al., 1965).

From Fig. 1 it is seen that new cells inrat can be observed in the blood within about 3days following exposure, whereas in man the timeincreases to about 8 days. Also, cell turnoveris more rapid in the smaller species. Bloodcell counts reach nadirs at shorter times in thesmall species so that large species survivelonger after low lethal doses. But, because ofrapid cell removal kinetics, the smaller speciescan survive higher sublethal exposures.

Thrombopoiesis results in production ofmegakaryocytes that secrete platelets, which areessential to prevent hemorrhage. As seen inFig. 2, fully differentiated megakaryocytes areproduced in about 7 days and have a meanlife span of about 10 days in peripheral blood(Szirmai, 1965).

During thrombopoiesis. megakaryoblasts canincrease in number, and this amplification isalso common in processes of erythropoietis andgranulopoiesis (Szirmai, 1965). These partiallyand fully diffezentiated blood cells are moreradio.resistant than the stem cells so that, atdoses that are just adequately lethal, time todeath may be extended through amplification ofsurviving cell populations. However, lympho-cytes (which amplify their numbers greatlywithin peripheral blood) are quite radiosensi-tive. As the magnitude of the lethal dose isincreased, the survival time is shortenedgreatly; death results when lymphocytes are kil-led instead of when stem cells are unable tomatch the homeoststic demand for new cells.Burros have been found to be quite radiosensi-tive and die within just a few days followingsuperlethal doses of radiation. Death of lym-phocytes may help explain why burros are extraradiosensitive and may die in 5-10 days whereasmost hematologic mortality in other largeanimals is seen at times greater than 10 days.Of course, 5-10-day survival times may be obser-ved in all other species if the treatment doseis sufficiently large.

RatUosensitivity is directly related tocell cycle kinetics. The typical cell cycle isillustrated in Fig. 3. When the need for newcells is low (or zero), cells cease prolifera-tive activity and are commonly referred to as

334

MYELOPOIESIS

ORNL-DWG 86-1896

REFERENCE: BOND et al., 1965

Fig . 1 . Variation of myelopoies i s by s p e c i e s . (Reference: Bond e t t l . ,1965.)

OANIDWG M 1H1

» PLATELETS * ?

MAKftOW• LOOP ~

IAHHIER

REFERENCE: SJlrmil HWS)

Fig. 2. Prooeis of thronbopoietit in nan.(Reference: Szirnai, 196S.)

being in a retting or Go itate. At seen inFig. 3, a cell with the noraal aaount of nuclearDNA ii *>id to be in the Oi ttata (i.e., aitotioinactivity gap with diploid DNA). Followingthi* 12-hour period new DNA ia eyntheaixed. andthe cell drop* into an inactive atate withtetraploid DNA preceding binary fission into twonew cell*.

AK illustrated in Fig. 4 (Case I), aanyradiogenic lesions in non-dividing DNA have ahigh probability of repair because the enzjraescan read the coapltaentary strand of DNA andrepair the damage site before the DNA helixseparates and a new helix of DNA is synthesized.In each new cell, a DNA helix contains one newand one old strand of DNA. As seen in Case II,radiogenic lesions immediately preceding DNAsynthesis have a very low probability of repair,so that it is likely that one of the daughtercells is killed or functionally altered. Formost practical considerations, DNA repair ceaseswhen the nuclear DNA replicates. However, abouthalf of the daaage can be repaired within 1 hourpreceding replication. Thus, in 4 hours betweensynthesis and aitoais the residual damage couldbe reduced to about (1/2)4, o r 6 % a Cells can bekilled by damage to other organelles auch asmitochondria or membranes, but cells can be ten-fold aore resistant during interphase periodsthan during aetaphase periods.

It ic obvious that a great aany physicaland biological factors can affect intracellularlesions, cell death, and, thus, death of theanimal. Important physical factors commonlyinclude: type (or LET) of radiation, doselevel, dose rate, fractionation of dose withtime, do»» distribution within the marrow, etc.

Important biological factors that affectintracellular lesions, cell death, and death of

335

ORNL-OWG 86-1898

CELL CYCLE FOR HEMAPOISIS

CELLULAR

DNA

TETRAPLOID

DIPLOID (M

REFERENCE: Lajtha (1957)

I12 24 36

TIME (h)

48

M - MITOSIS

S - SYNTHESIS

GD » RESTINGPHASE

G1 - GAP WITH NORMAL DNA

G2 - GAP WITH DOUBLE DNA

(NOT SHOWN)

Fig. 3. Cell cycle for hematopoiesis. (Reference: L«jth», 1957.)

the animal include: DNA content per cell; tineperiods of various phases of the cell cycleunder different in vivo conditions; sensitivityof the individual, strain or species to infec-tion and/or hemorrhage (physical conditions canhost the opportunity for infection); capacity ofthe animal for compensatory repopulation of pla-telets and granuolocytes; etc.

II. METHOD

Lethality data were collected from manydifferent animal studies. Experiments selectedto evaluate the model were restricted topenetrating photons and any combination of bodysize and irradiation geonetry that resulted in anniform dose profile to all parts of the activenarrow. A total of 224 different Mortality stu-dies were included in the data base. Dataincluded: 13 different species; body weightsfrom 25 g to 375 kg; sheep, goats, swine, andcalves with body weights near man (i.e., 70 kg);radiation sources of 6 0Co, 1 8 2T», 9 5Zr, atomicbombs, and X rays from several voltage poten-tials and many different moderating materials;and exposures from bilateral, multiple sources,unilateral, rotational, free-moving animals, 4n,and quadrilateral geometries, as long as thedose was uniform over the active bone marrow.

Typically, physical factors can be quanti-fied or ordered on a numerical scale. Butbiological factors vary with species, strain,age, etc. , and with the composite force from the

collection of other biological and physical fac-tors. Most biological factors can be treated as"classification" type variables in a regressionanalysis but usually cannot be quantified.However, the variance in the data set resultingfrom classification variables can be quantified.Regression analysis methods were used to quan-tify how much variance is left in the data setafter physical factors are considered. Then, reevaluated how much of this variance can be dueto individual "classification" variables (suchas species) and, finally, how much of thespecies effect is left unaccounted for afterbody weight is considered.

Within each species, a model that linearlyrelated the log of LD50 to the log of dose ratefitted the data fairly well. Across species,the slopes of these lines were relatively con-sistent, but their intercepts differed signifi-cantly. As evidence of this, a nodel fitted toall data with a single intercept and slopeaccounted for less than 1% of the variation inthe data (R2 < 0.01). A model that included aseparate intercept for each species but only acommon slope for the log of dose rate accountedfor 84% of the variation; a model that allowedseparate intercepts and slopes for each speciesimproved this only slightly, to 86%. Further-more, when intercepts were fitted for eachspecies, there was a clear inverse relationshipbetween species body weight and the value of theintercept (heavier species had relatively higherresponse values at a given dose rate). In fact.

336

ORNLDWG 86-1897

SIMPLIFIED SCHEMATIC OF INTRACELLULOR LESIONS

CASE I: Go OR G, PHASE G2 PHASE

(HIGH PROBABILITYOF REPAIR)

(LOW PROBABILITYOF REPAIR)

CASE I I : S PHASE

Fig. 4. Simplified schematic of intracel-lul ir lesions.

a model that contained a single intercept, atern for the log of dose rate, and a tern forstandard body weight accounted for 64% of thevariation in the data.

In consideration of the above, a two-stagemodel was adopted.

Stage 1: For species i, a model of form

[logio(IJD50)Ji ~ <»i + P*logio(dose rate)+ Y'l°glO(body weight)i + e

was used, p and y are regression coefficientsassumed to be valid across species, aj. i s anintercept specific to species i , and erepresents experimental or unexplained error inthe reported values of logio(LDso). assumed tobe distributed with mean zero and variance a 2 .

Stage 2: Across species, <zi was assumed tobe distributed with mean a and variance 6 2 .This random intercept can be thought of as aspecies effect, after correction for speciesbody weight. In order to use the model topredict results for a species not included inthis data set ( e . g . , man), i t was assumed thatthe applicable value of a± would be a new (unob-served) value from this same distribution.

Using a computational technique describedby Laird and Ware (1982), maximum likelihoodestimates were calculated for the parameters ofthe model as follows:

o = 2.743(5 = -0.070T = -0.161S2 = 0.009662 = 0.0194

So, a point estimate of the LD50 for an unspeci-fied or new species i s

estimated LD50 = iol«+P- l o810(d°se rate)+ ?#logio(body weight)]

estimated LD50 = lO^2-7 4 3 " O.701ogio(dose rate- 0.1611ogio(body weight)].

For man, a species having a 70-kg body weight,this reduces to

estimated LD50 = 281 (dose ra te )" 0 - 0 7 0 .

A common slope of -0.070 was used for al lspecies, and the intercepts ( i . e . , a i ' s ) are asfollows: mouse (2.946), hamster (2.925), rat(2.875), guinea pig (2.508), rabbit (2.967),primate (2.782), dog (2.493), goat (2.490),sheep (2.393), swine (2.500), Ban (2.743). burro(2.410), and catt le (2.205). Thus, mouse, ham-ster, rat, primate, dog, swine, goat, burro, andcatt le seem to demonstrate a consistent mono-tonic relationship with body weight, but sheepand guinea pig are more radiosensitive than mostother species and rabbit is more radioresistant,on a relative basis.

These formulae are compared with the exper-imental data, sorted according to species, inFig. 5. It should be remembered that, althoughthe experimental data in Fig. 5 reflect manyother biological and physical variables, onlyone equation was used for al l species. However,the model seams remarkably accurate for eachspecies. This consistency is unique (Baverstocket a l . , 1985) and offers almost unlimited poten-t ia l to model human response from the extensivedata basis available on test animals.

The midlethal dose for man plotted againstdose rate i s given in Fig. 6. Marrow dose wasconverted to tissue kerma in air according toJones (1977). The equation for man [viz,281/(dose ra te ) 0 - 0 7 ] was solved for midlethaldose for continuous dose rates given in 1minute, 1 hour, 1 day, and 1 week. Results arein Table 1.

III. DISCUSSION

Baverstock et a l . (1985) analyzed animaldata but did not use a dose-rate dependentmodel. Instead, they selected exposure times ofone hour or l e s s . They found a lack of homo-geneity within species. However, according toour analysis, the LD50 given in one minute i sabout 190 rad to marrow, and the LD50 given inone hour is about 250 rad. These estimates arefor a 70-kg body weight; the spread could beconsiderable larger for smaller species with

337

MOUSE = 0.025 KG1MO

1000 -

in-" ,«-!10"' io"' ioT id io*DOSE RATE (RAD/MIN)

RAT = 0.225 KGO 1900

s

I

ORNL DWG 85-16100

io"' io" id1 IO1 id"DOSE RATE (RAD/MIN)

G.PIG = 0.5 KGQ 1500

a35 10"' 10"' 10* 10" 10*

DOSE RATE (RAD/MIN)

RABBIT = 3 KG

ior" io" id" ltf id"DOSE RATE (RAD/MIN)

PRIMATE = 5 KG

10' 10"' 10* 10" 10*DOSE RATE (RAD/MIN)

DOG = 10 KGa isoo

2

10"" 10"' 10* 10* 10"DOSE RATE (RAD/MIN)

SHEEP = 60 KG1500

1000

900

10"* 10" 10* lrf itfDOSE RATE (RAD/MIN)

SWINE = 60 KG1600 r

10- MS" 10* lrf 10*DOSE RATE (RAD/MIN)

BURRO = 155 KG1900

900

10" 10" 10* irf 10*DOSE RATE (RAD/MIN)

F i g . 5 . Comparison of the s t a t i s t i c a l dose -ra te Modal with experimentaldata fox nine s p e c i e * .

f a s t e r m i t o t i c r a t e s . The aodel of Saverstocke t a l . did not d i scr iminate between dose r a t e swhen exposure times were short compared withc e l l u l a r turnover t i a e s ( i . e . , d o s e - r a t e s above10 8 / n i n were considered e q u a l ) . However, forlow LEI r a d i a t i o n (which was being modeled) ,enzrmatic repa ir must a l s o be considered. Suchrepa ir has been found to be s i g n i f i c a n t in t imesmuch shorter than c e l l turnover times (Terzaghiand L i t t l e , 1 9 7 5 ) . Thus, the Baverstock e t a l .model does not seem w e l l s u i t e d to analyze expo-sure i n t e r v a l s ranging from one minute to onehour. l a a d d i t i o n , we have made e s t imates ofdose to marrow upon which our model i seva luated . Although the Baverstock e t a l . aodelwas evaluated twice—once i s t erns of exposureand again i n a i d l i n e t i s s u e dose—no attempt wasmade to es t imate marrow dose . Because of thesesignificant differences in the two aodels andbecause of our much larger data base, i t seemsthat our model has found a coherent pattern ininterspecies LD50. Experiments using low LET

exposures (to al l species) that resulted in uni-form marrow dose a l l seem compatible with a sim-ple interpolation model based on dose rate andbody weight.

IV. CONCLUSIONS

There is no unique or practical LD50 fornan because mortality varies strongly with doserate and with several physical and biologicalfactors. The NCRP LD50 in air may be about 25%too high, and the NCRP marrow LD50 may be about60% too high—excellent agreement consideringwhat was known in 1949 when the NCKP value waspromulgated. Mole's LD50 in air (Mole, 1984)seems about 70% too high, and his value for mar-row seems about 130% too high. However, Mole'smortality curve has the correct shape because hederived i t from the animal dtta in the method ofJones (1981) ( i . e . , treatment doses were normal-ized to a multiple of the LDjo for that particu-lar experiment).

338

ORNL DWG 85-16806

7.0

6.0

5.0

4.0

3.0

an

_

^ DOS

1 1 1 1 t 1 Hi i 1

DOSE-RATE LD50(RAD/MIH) (RAD)

1 2S010 240

100 200

I I i ul i i i i i m l , I I TiTMi-

wCOoQ

a

10* 10"' 10° 10" ICDOSE RATE (RAD/MIN)

Fig. 6. Nidlethal dose for M»a tt a func-tion of dose rate according to the statistical•odel derived from animal data.

Rotblat's LDso (NAS, 19:55) seens about 20*too low, bat the slope of his nortality curve isfivefold too flat (Jones, 1981; Mole, 1984).

Lushbaugh's estimate of 281 rad to •arrowfor dose rates between 0.75 and 1.6 R/min seeasacceptable. The mathematical model based on theanimal data (ires 281 rad for 1 R/min and a 70-kg body weight (Luihbaugh et al., 1967). IfLushbangh had not included the seven Y-12 vic-tims, who were exposed to high dose rates, withhis 93 patients treated at a low dose rate, hisLD50 estimate would probably be a bit higherthan the published value of 281 rad.

The United Nations Scientific Committee hasundertaken a recent analysis of horn an mortality.Although that analysis has not been finalized,the analysis described in this paper and ourprevious experience support several importantissues discussed in the UNSCEAR 1982 report.Those issues include:

* differences in effects doe to differentphoton energies are considered to be negli-gible (p. 572),

* sublethai damage can normally be repairedin a few hours (p. 573),

* Tor a variety of different types of treat-ment, different dose rate and LET, thereduction in the proportion of survivingcells resulting in 50% death of the micewas the same for all treatments" (p. 573),

* there is little or no enzymatic repairabove 100 rad/min (p. 575) so the LDjoshould be constant at dose rates above 10^or 103 rad/min.

Conclusions presented in this manuscriptare expected to be firm, but numerical resultspresented at this time are subject to smallchanges when a final reporting of this study ismade in 1987. Because of the success of thisexploratory statistical model, a comprehensiveeffort is now under way to collect data on allindividual dose treatment groups that contribu-ted to the 224 different LD50 values analyzed inthis study. Also, other studies are being addedto the data base. This more comprehensive database will be analyzed for mortality response asa function of treatment dose expressed as multi-ples of the LD50 value appropriate for a parti-cular study. Thus, a universal mortality func-tion will be derived (Jones, 1981), and 95% con-fidence limits will be evaluated.

When that effort has been completed, thebody weight of man and different dose rates ofinterest can be used to calculate tables ofdose-response values.

V. ACKNOWLEDGMENTS

The authors wish to thank Jack Greene ofGreenewood Enterprises, Clarence Lushbaugh ofOak Ridge Associated Universities, Sheldon Levinof Espanola, New Mexico, and Curtis Travis ofOak Ridge National Laboratory for their expertreview of this paper.

REFERENCES

Table 1. Midiethai dose for man exposed tocontinuous dose rates given in 1 minute,

1 hour, 1 day, and 1 week

Exposuretime

1 Minute

l'Hour

1 Day

1 Week

Dose rate(rad/min)

194

4 .2

0.22

0.035

Photon

Marrow(rad)

194

250

310

360

midlethal dose

Tissue kermain a ir (rad)

350

450

560

650

Apler, 1979, Cellular Radiobioloiv. pp. 1-5,Cambridge Univ. Press, London.

Baum, S. J., et al., 1984. "Nuclear WeaponEffects Research at PSR-1983, Vol. 10,Symptomatology of Acute Radiation Effectsin Humans After Exposure to Doses of 75 to4500 Rads (cGy) Free-in-Air DNA. " TechnicalReport TR-85-50. 31 Aug 84.

Baverstock, K. F., et al., 1985. "Man's Sensi-tivity to Bone Narrow Failure FollowingWhole Body Exposure to Low LET IonizingRadiation: Inferences to Be Drawn fromAnimal Experiments, " nt. J. Radiat. Biol.,47(4.) , 397.

339

Beck, W. L., et «1., 1971, "Dosimetry forR»diobiologic*l Studies in the Human Hema-topoietic System, " p. 974 in Proceedings ofthe National Symposium on Natural and Man-made Radiation in Space. March ,l-£, 1971.NASA TMX-2440, National Aeronautics andSpace Administration.

Bergonit, J., and L. Xribondeau, 1906,"Interpretation de Quelques Resultats de laRadiotherapie et Essai de Fixation d'uneTechnique Rationnelle, " Conpt. Rend. Ac ad.Sci.. 143, 983.

Bond, V. P., et al., 1965, Mannalian RadiationLethality. A Disturbance in Cellular Kinet-ics. Academic Press, New York.

Evans, J. S., et al., 1985, "Health EffectsModel for Nuclear Power Plant Accident Con-sequence Analysis, Part I: Introduction,Integration, and Summary, Part II: Scienti-fic Basis for Health Effects Models,"SAND-7185 (NUREG/CR-4214). Sandia NationalLaboratories.

Jones, T. D. , 1977, "CHORD Operators for Cell-Survival Models and Insult Assessment tothe Active Bone Marrow," Radiat. Res., 71,269.

Jones, T. D. , 1981, "Hematologic Syndrome in ManModelled from Mammalian Lethality," HealthPhys. . 41, 83.

Jones, T. D. , 1984, "A Unifying Comcept for Car-cinogenic Concept for Carcinogenic RiskAssessments: Comparison with Radiation-Induced Leukemia in Mice and Men," HealthPhvs. 47(4), 533.

Laird, N. M., and J. H. Ware, 1982, "Random-Effects Models for Longitudinal Data,"Biometrics. 38, 963.

Lajtha, L. G., 1957, "Bone Marrow Cell Metabol-ism, " Phvsiol. Rev., 37. 50.

Langham, W. H. (Ed.), 1967, "RadiobiologicalFactors in Manned Space Flight," Report ofthe Space Radiation Study Panel of the LifeSciences Committee, National Academy ofSciences, National Research Council,Washington, D.C.

Lushbaugh, C. C., 1969, "Reflections on SomeRecent Progress in Human Radiobiology, " p.277 in Advances .in Radiation Biol. Vol. 3,Academic Press, New York.

Lushbaugh, C. C., et al., 1967, "Clinical Stu-dies of Radiation Effects in Man, A Prelim-inary Report of a Retrospective Search forDose-Relationships in the Prodromal Syn-drome, " Radiat. jges. Suppl. . 7. 398.

Mole, R. H., 1984, "The LDJO for Uniform Low LETIrradiaton of Man," Br. £. Radiol.• 57,677.

NAS, 1985, Symposium on Medical Implications ofNuclear War, sponsored by the Institute ofMedicine, NAS, Sept. 20-22, 1985.

NCRP, 1974, National Council on Radiation Pro-tection and Measurements, "RadiologicalFactors Affecting Decision-Making in aNuclear Attack," NCRP Report 42, Bethesda,Md.

NRC, 1975, Nuclear Regulatory Commission, "Reac-tor Safety Study - An Assessment ofAccident Risks in U.S. Commercial NuclearPower Plants, " WASH-1400 (NUREG 75/0.14).

Rider, W. D., and R. Hasselback. 1967, "TheSymptomatic and Haematological DisturbanceFollowing Total Radiation of 300-rad GammaIrradiation, " in guidelines to RadioloaicalHealth Lecture. U.S. Department of Health,Education, and Welfare, Kockville, Md.

Szirmai, E. , 1965, Nuclear Hematoloav. AcademicPress, New York.

Tenaghi. M., and Little, J. B., 1975. "Repairof Potentially Lethal Radiation Damage IsAssociated with Enhancement of MalignantTransformation." Nature. 253, 548.

UNSCEAR, 1982, United Nations Scientific Commit-tee on Effects of Atomic Radiation, UnitedNations Publication Sales No. E.82.IX.806300P United Nations, New York. 10017.

ANS Topical Meeting on Radiological Accidents-Perspectives and Emergency Planning

Radiation Carcinogenesis Following Low-Doseor Low-Dose-Rate Exposures

R. L. Ullrich

ABSTRACT. A variety of dose responses havebeen observed for cancer induction followinglow linear energy transfer (LET) radiation. Ingeneral, however, the response Is curvilinear,with a rapidly rising component in the inter-mediate dose range followed by a plateau ordecline in incidence at high doses. The res-ponse is more linear at low doses, whereas theresponse at intermediate doses is approximatedby a dose-squared relationship. Models forthis response are based on the biophysicaltheory of cellular effects. However, manytypes of effects contribute to the tumorigenicprocesses, and host factors play a major role.At low dose rates the carcinogenic effect isgenerally reduced, which is caused by a ditnuni-tion of the dose-squared component and resultsin a linear response. Effects of fractionationcan vary with total dose, fraction size, andfraction interval. High LET radiation is moretumorigenic. The dose-response relationshipsare more nearly linear and are less dose-ratedependent. The relative biological effective-ness (RBE) varies with dose, dose rate, frac-tionation, and target tissue.

The primary risk of exposure to low dosesof radiation is cancer induction. While mostexposures to radiation occur at low doses orlow dose rates, most information on the carci-nogenic effects of radiation for human popula-tions is for high dose and high dose rate ex-posures. Statistically reliable information atlow doses is very difficult to obtain becauseof the sample sizes required to detect increas-ed cancer risks which are small compared withthe total cancer risk in the population,, Sinceit is not possible to measure directly effectsat low doses and dose rates, estimates of riskare based upon extrapolation by use of mathema-tical models of the dose response relationship(Upton, 1977). The two models most often ap-plied are either a linear model or a linearquadratic model. Although the question of doseresponse models is fundamental, available humandata are insufficient to allow a choice betweenthese models and estimates of risk can varydepending upon the model used.

The linear quadratic model has as its basisbiophysical concepts of radiation effects withincells (Kellerer and Rossi, 1978; Kellerer andRossi, 1972). Biophysical theory proposes thateffects at the cellular level are a result ofthe interaction of two sublesions within thenucleus. These sublesions can be produced inde-pendently by two ioniEation tracks or, whensufficient energy is deposited, by a singleionization track. For low-LET radiation in theintermediate dose range the sublesions are gene-rally produced by two independent tracks inclose proximity in time and space, thus the' re-sponse increases with the square of the dose.When the dose is low, the probability of two in-dependent tracks passing in close enough proxi-mity to produce interacting sublesions is re- ;duced. At these low doses the dose-squaredresponse is absent and the response is linear.This linear response is a result of the produc-tion of both sublesions by a single ionizationtrack. For high-i.ET radiation the two sub-lesions are theoretically produced by a singletrack over a wide dose range because of thedense nature of the ionization events. As aresult of this, effects on cells show a linearrelationship with dose. While this model appearsto be adequate for many cellular effects, itsapplicability to tumor formation is less clear.It should be appreciated that this model isbased on the induction of initial cellular eventsbut may not be adequate for the complex, ap-parently multistage process of carcinogenesis.

Animal studies have been useful in examin-ing the question of low dose and dose rateeffects. In studies with low LET radiation avariety of dose responses have been observedranging from those with a threshold to those inwhich the response appeared linear (Fry, inpress; Ullrich and Storer, 1979a; Ullrich andStorer, 1979b; Ullrich and Storer, 1979c; Ull-rich, 1983). In spite of this range of re-sponses, in most instances the dose response hasbeen found to be curvilinear. These curvilineardose responses can generally be described by alinear-quadratic regression equation in whichthere is a rapidly rising (dose squared) compo-nent in the intermediate dose range and a moreshallow (linear) component in the low dose range(see Figure 1).

341

OQ)

LJJO

UJQ

O

40-

0

Lung Adenocarcinoma

High Dose Rate Y= 11.9 + 0.041 X + 0.00043 XLow Dose Rate Y=12.5 + 0.043X

a/ /3 = 1 Gy

0

""""""" — ——-°to

0.4 0.8 1.2DOSE (Gy)

1.6 2.0

Figure 1. Incidence of lung tumors following high (•) or low (0) dose rate -y-ray irradiation.

343

While current models of dose response em-phasize initial cellular events, many types ofradiation effects can contribute to the processof tumor development. Some effects may play arole at the level of initiating events, trans-forming a normal cell into a cell capable offorming a tumor. Other effects, some operatingat the cellular level and others at tissue,organ, or systemic levels, influence whether thecarcinogenic potential of transformed cells isrealized by modulating their expression. Infact, it may be that host factors play a majorrole in determining the expression of initiatedcells. This suggestion is supported by obser-vations in experimental animals. The incidenceand spectrum of tumors in different strains ofmice vary widely in nontreated and irradiatedanimals. Susceptibility to radiation within astrain varies with both age and sex and can bealtered by environmental or endocrine manipula-tion.

In experimental studies, when low-LET radi-ation exposure occurs at a low dose rate, thecarcinogenic effect of the radiation has beenfound to be reduced in virtually every instance(Ullrich, 1980; Ullrich and Storer, 1979b;NCRP, 1980; Ullrich, 1983). The primary effectof lowering the dose rate has been found to bea reduction of the effect in the intermediatedose range. This reduction diminishes the quad-ratic component of the dose response as would bepredicted by assuming dose squared terms comefrom interacting independent events and makesthe response shallow and more nearly linear overa wide dose range. In those instances in whichit has been examined, the slope of the linearresponse following low dose rate exposures andthe slope of the linear component from thelinear quadratic dose response following highdose rate exposure have been found to be thesame (see Figure 1).

When the radiation dose is fractionated theeffects on tumor development can vary dependingupon the total dose and the fraction size. Ifthe dose/fraction is a dose which is on thepredominantly linear portion of the dose re-sponse curve the response is similar to thatfollowing low dose rate exposures. If the dose/fraction is in a region where the quadratic com-ponent predominates the response may be similarto, or in some Instances greater, than that fora similar single dose. This is illustrated inthe following table:

Table 1

Incidence of lung tumors after a 2 Gydose delivered as acute fractionated

or protracted exposuresExposure Regimen Observed Incidence

High Dose Rate

Low Dose Rate

Low Dose Fractions

High Dose Fractions

38.6%

21.3%

32.92

High LET radiation (e.g., neutrons) is morecarcinogenic on a dose for dose basis than lowLET radiation (Upton et al., 1970; Ullrich etal., 1976; Ullrich et al., 1977). Recent datafor tumor induction indicates a linear doseresponse following high LET radiation over the0 to .25 Gy dose range (sometimes 0-.5 Gy). Athigher doses the dose response tends to bendover. Because of tiie quadratic nature of thedose response in the intermediate dose range forlow LET radiation and the linear response forhigh LET, the relative biological effectiveness(RBE) for high LET radiation increases withdecreasing dose to a point where the low LETresponse also becomes linear (Fry, in press).At this dose level the RBE becomes constant.The value for this constant RBE for tumor induc-tion varies with tissue. The carcinogenic ef-fects of high LET radiation have generally beenfound to be less dose rate dependent than forlow LET radiation. Dose rate independence isexpected for cellular effects assuming thateffects are produced by single tracks. Re-cently, however, data has suggested an enhancedtumorigfcnic effect following low dose exposuresin certain tissues even at low doses (Ullrich,1984). The general applicability of this ob-servation is not known at the present time.

Fry, R. J. M., "Experimental radiation carcino-genesis: what have we learned," Radiat.Res, (in press).

Kellerer, A. M., and H. H. Rossi, 1978, "Ageneralized formulation of dual radiationaction," Radiat. Res., 75, 471.

Kellerer, A. M., and H. H. Rossi, 1972, "Thetheory of dual radiation action," Cur. Top.Radiat. Res. Q., 8, 85.

NCRP, 1980, "Influence of dose and its distri-bution in time on dose-response relation-ships for low-LET radiations," NCRP ReportNo. 64, National Council on RadiationProtection and Measurements, Washington,D.C.

Ullrich, R. L., 1980, "Carcinogenesis in miceafter low dose and low dose rate irradia-tion," Radiation Biology in Cancer Research,R. E. Meyn and H. R. Withers, eds., p. 309,Raven Press, New York.

Ullrich, R. L., and J. B. Storer, 1979a, "In-fluence of gamma ray irradiation on thedevelopment of neoplastic disease in mice.I. Reticular tissue tumors," Radiat. Res.,80, 303.

Ullrich, R. L., and J. B. Storer, 1979b, "In-fluence of gamma ray irradiation on thedevelopment of neoplastic disease in mice.II. Solid tumors," Radiat. Res., 80, 317.

Ullrich, R. L., and J. B. Storer, 1979c, "In-fluence of gamma ray irradiation on thedevelopment of neoplastic disease in mice.III. Dose rate effects," Radiat. Res.,80, 325.

344

Ullrich, R. L., et al., 1976, "The influence ofdose and dose rate on the incidence ofneoplastic disease in RFM mice after neu-tron irradiation," Radiat. Res., 68, 115.

Ullrich, R. L., et al., 1977, "Neutron carcino-genesis: dose and dose rate effects inBALB/c mice," Radiat. Res., 72, 487.

Ullrich, R. L., 1983, "Tumor induction inBALB/c female mice after fission neutronor y irradiation," Radiat. Res., 93, 506.

Ullrich, R. L., 19S4, "Tumor induction inBALB/c mice after fractionated or pro-tracted exposures to fission-spectrumneutrons," Radiat. Res., 97, 587.

Upton, A. C , 1977, "Radiobiological effects oflow doses. Implications for radiologicalprotection," Radiat. Res., 71, 51.

Upton, A. C , et al., 1970, "Late effects offast neutrons and gamma rays in mice asinfluenced by the dose rate of irradia-tion: induction of neoplasia," Radiat.Res., 4i, 467.

ANS Topical Meeting on Radiological Accidents-Perspectives and Emergency Planning

A National Emergency Medical Assistance Program forCommercial Nuclear Power Plants

Roger E. Linnemann, M.D, and Mary Ellen Berger

ABSTRACT - Radiation Management.Consultant's Emergency Medical AssistanceProgram (KMAP) for nuclear facilitiesprovides a twenty-four hour emergencymedical and health physics responsecapability, training of site and off-si topersonnel, and three levels of care forradiation accident victims: first aid andrescue at an accident site, hospital •emergency assessment and treatment, anddefinitive evaluation and treatment at aspecialized medical center. These aspectsof emergency preparedness and fifteen yearsof experience in dealing with medicalpersonnel and patients with real orsuspected radiation injury will be reviewed.

Radiation injuries are classifiedaccording to those resulting from excessiveexposure, those resulting fromcontamination and those resulting fromcombined exposures with or without traumaor serious illness. The contaminated andinjured or seriously ill patient presentsthe only unique administrative andoperational problem for emergencytreatment. The proper management of thesepatients involves the use of procedures notnormally utilized in emergencydepartments. Because of hhis, specialplanning, training and practice arenecessary so that emergency care can bodelivered without spreading contamination.Persons sustaining external exposureinjuries can be evaluated and, ifnecessary, treated in the normal setting ofan emergency room. However, the definitiveevaluation and treatment of individuals whohave received serious external exposurescan tax the most advanced medical center.

Because of the infrequency of accidentsand the complexity and cost to provide acomplete medical program for the evaluationand treatment of radiation injuries,Radiation Management Consultants (KMC)established an Kmergency Medical AssistanceProgram (EMAP) in 1970 to serve thecommercial nuclear industry. This programis based on a regional approach to the

management of radiation injury, with threelevels of care: first aid and rescue althe nuclear facility, emergency treatmentat a nearby hospital, and definitiveevaluation and treatment at a largo medicalcenter. Unlike many medical emergencies,the clinical manifestations of radiationinjury unfold over time - usually days toweeks. Consequently, patient care can bocarefully managed and controlled at each ofthese 3 levels without impairment urdetriment, to the treatment and recovery ofthe patient. The nature of radiationinjury is such that at each level of carepriority can he given to more1 ife-threatening trauma or illness that mayaccompany the radiation injury.

KMAP is a nationwide programcoordinated and directed from RMC's officesin Philadelphia and Chicago. Currently itcovers 21 nuclear power plant sites ownedand operated by ?.O utilities in 13 statesacross the country. RMC's KMAP stnfrincludes full-time professionals inradiation medicine, nursing, healthphysics, radiochomistry, environmentalscience and emergency medical services.

A specialized accident radiobioassaylaboratory, including mohile whole bodycounters, is available at both offices toassist the staff in dose evaluation andpatient management. Associated with thisin-house accident response capability aretwo definitive clinical care centers:Hospital of the University of Pennsylvaniain Philadelphia and Northwestern MemorialHospital in Chicago. KMAP maintains itsclose affiliation with these centersthrough staff appointments, teachingresponsibilities and membership on therespective centers' Radiation AccidentCoordinating Committees. In addition, thetwo offices maintain and exercise thetwenty-four hour availability of aRadiation Kmergency Medical Team (RKM-Team)that can bo dispatched to a plant site orlocal hospital to assist in the initialevaluation of the patient(s) and in theorderly transfer to one of the clinicalcare centers. A typical REM-Team consists

345

346

of a radiation medicine physician, acertified health physicist, and radiationtechnicians who arc supplied with emergencyinstrumcnlation. A calibration laboratoryis maintained in Philadelphia to insurethat instruments are properly calibratedand functioning at all limes. Atwenty-four hour "hot-Hue" connects eachparticipating site and associated hospitalthroughout i he country with the uffic.es inChicago and Philadelphia. KMAT'participants are encouraged to use thisline to obtain advice and/or assistanceanytime a radiation injury occurs. RMC haslong recognized that even minor problemscan sometimes have medical-legalsignficance.

In addition to providing a centralizedemergency response capability, the 10MAI'staff assists personnel at nuclear powerplant sites and associated hospitals inestablishing and maintaining preparednessat the local level. This assistanceincludes the development of procedures,design of facilities, placement ofequipment and supplies, training ofpersonnel at plant sites in first aid andfescue of radiation accident victims,training ambulance service personnel intransportation; and training local hospitalpersonnel in emergency treatment. To theextent possible, procedures, equipment andsupplies are standardized at all sites.Specialized equipment has boon designed tofacilitate treatment and decontamination ofpatients while controlling thecontamination to a designated area of thehospi tal.

Preparedness is maintained at alllevels of care through training anddrills. Since 1970, KMAP has conductedover 500 drills and exercises. Thisincludes the development of scenarios,moulaging accident victims and evaluatingand/or controlling the drillr,. Most ofthese drills are videotaped from thebeginning of the simulated accident in theplant to the completion of the drill at thehospital. These video tapes arc used inthe critique at the hospital immediatelyfollowing the drill and are subsequentlyedited with narrative for use by the plantand hospital personnel for interim training.

Because of KHAP's emphasis onproviding the best patient care in the most-effective manner possible, the entireprogram is further strengthened andcoordinated through annual medical seminar;:and an educational newsletter,"Curie-Osity", special medical stafflectures, periodic meetings of the EHAPprofessional staff and monthlycommunication checks. The annual medicalseminar, sponsored by one of the definitiveclincial care centers, focuses onoccupational health and current issuesaffecting emergency medical planning. Thisseminar provides an opportunity for medicalstaff and plant health physics personnel toexchange ideas and learn about recent

accidents and developments in patientevaluation and can'.

During the past 10 years RMC hasevaluated and, when necessary, treated morethan 300 individuals with real or suspectedradiation injuries. Not all these patient::were from commercial nuclour power plants.In fact, only one of the paticnLs requiringmedical evaluation and treatment foroverexposure came from a nuclear powerplant. Most of those individuals hadradiation injury due to industrialradiography or medical radiotherapysources. (One of the patiunis undergoingtreatment, for example, received 100,000rads to his thumb and two fingers over nperiod of eight years.)

Because of major involvement withnuclear power plants, KMC's personnel havedeveloped considerable expertisn in dealingwith external and/or internal contaminationproblems. I'robably the most complexcontamination problem occurs when anindividual becomes ill or is injured in Acontaminated area of a nuclear facility,fifty-nine incidents such as this hoveoccurred in which patients were taken froma nuclear power plant to a nearby hospitalfor emergency treatment, eighty-onepercent of these cases involved injuriestypical of those occuring in other heavyindustries. These included falls (381),injuries due'to lifting and turning (197..),injuries caused by machine tool use 0%),injuries caused by chemical explosion/blast(6%) and others (9%). Contusions,lacerations, puncture wounds, open andclosed fractures, soft tissue injuries,blast injuries, burns, eye injuries, skullfractures and central nervous sytemsinjuries have bean some of the results ofthese accidents. The most serious injurieshave been due to falls. These haveresulted in two deaths and two instances ofparaplegia (i.e., paralysis of the lowerbody).

Nineteen percent of incidents involvedserious illnesses of endogenous andexogenous origins. In three instancesindividuals required treatment for cardiac,problems. Problems associated withdiabetes led to another needing hospitalemergency treatment. Two deaths occurredboth due to heart attacks. Illness ofexogenous (environmental) origin is notuncommon, and severe heat stress is thechief culprit.

Personnel illnesses occurred"round-the-clock", while the majority ofserious injuries occurred duringoff-hours. Protective clothing andrespirators were implicated in accident andillness causation in several instances.(Loose clothing can be caught in rotatingmachinery, can lead to tripping, andcontributes to retention of body heat, inaddition, peripheral vision is lost withsome respirators.)

Most of the accidents or illnessesthat occurred in contaminated areas

347

involved only one patient. Twenty percentof the instances involved two patients, butin no case were more than two individualsinvolved. In these incidents, the highestlevel of Contamination on a patient was20mR/hr. The highest level of attendantexposure from handling these patients was14mR received over a period of two andone-half hours. In no case did thecontamination compromise patient care. Inno case was contamination spread beyond thecontrolled area in the hospital. Patientswere taken to the hospital because of theirinjuries or illness - not because of theircontamination (which tended to be at"nuisance" levels).

Although no serious radiation injuriesoccurred at TMI, RMC's KMAP provided neededassistance during and after the accident.Support was provided in laboratoryanalysis, whole body counting, counselingemployees and their families, as well asmembers of the public.

In order to determine if a "systemsapproach" would be effective in themanagement of a multi-casualty situation,the accident at Chernobyl was carefully

studied by RMC personnel. The Russianmedical response 3ystem at Chernobylappeared to be well planned. It, too, wasbased on a 3 tiered "regional approach",with on-site care, local support hospitals,and definitive care centers. It differedfrom the EKAP program in that a permanentstaff of physicians, nurses and supportpersonnel were available in an on-site"clinic" having 115 beds. Although taxedby the magnitude of the medical emergencysituation, Chernobyl's two supporthospitals and both of their definitive carecenters responded appropriately to providecare to the sick and injured.

The strengths of R.iC's KMAP lies inits full-time, in-house, twenty-four hourcapability to respond and assist plantsites, local hospitals and definitive carecenters in the complete evaluation andtreatment of radiation injuries, in itsassociation with large university medicalcenters, in its familiarity with personneland specifics of nuclear power plants, andin its fifteen years of experience withradiation medical problems at nuclear powerplants and facilities.

Section VII

Accidents and the Public

Chairman: Russell Dynes

University of Delaware

ANS Topical Meeting on Radiological Accidents-Perspectives and Emergency Planning

Evacuation Behavior in Nuclear Power Plant Emergencies:An Alternative Perspective

John H. Sorensen

ABSTRACT - It is alleged that the Governor'sadvisory at TMI led to spontaneous evacuation andoverresponse. Several researchers have railedthis the "shadow phenomenon." This phenomenon,however, rests on assumptions that are challengedby this paper. Instead, it is concluded thatevacuation behavior at TMI was consistent withour general understanding of human behavior indisasters. Emergency planners should be moreconcerned with underresponse to a nuclear powerplant emergency than with overresponse.

I. INTRODUCTION

Since the accident at TMI, social scientistshave been attempting to predict how the publicwould respond to a future nuclear power plantaccident (Perry and Lindell, 1983; Johnson, 1983;Johnson and Ziegler, 1983; Johnson and Ziegler,1984). Much of their attention centers on theissue of evacuation behavior. The central thrustof the research is directed towards resolvingemergency response Issues. The results of theresearch have significant Implications for futurepolicy decisions concerning the feasibility andcontent of emergency planning in the nuclearindustry.

Several questions lie at the heart of thiseffort:

(1) How did people behave during the Three MileIsland accident?

(2) Why did people behave in the manner observed?(3) Can we infer future behavior based on the

observations at TMI?(4) What will cause people to evacuate In a

future nuclear power plant emergency?(5) Are nuclear power plant emergencies a unique

category unto themselves or may they inducebehavior similar to other types ofemergency?

The purpose of this paper Is to present adifferent view of this Issue other than that

being currently advanced. In doing so, the paperattempts to sort out the answers to thequestions. To initiate this debate, a review ofthe predominant line of current social sciencethinking about evacuation in nuclear power plantemergencies is presented.

II. THE "SHADOW PHENOMENON" AND OVERRESPONSE

A popular social science view holds that theexperience at TMI demonstrated a unique phenome-non never before witnessed on such a large scalein an emergency situation. Researchers labeledthis the "shadow phenomenon" (Ziegler et al.,1981). Stated simply, this phenomenon Is thatmore people evacuated at TMI than was recommendedin an advisory Issued by the Governor ofPennsylvania. It Is estimated that 144,000people within a 15-mile radius left the area(Flynn, 1979). It is pointed out, 144,000 peoplewithin a 15-mile radius left the area (Flynn,1979). It is pointed out, however, that only2,500 people met the evacuation criteria setby the Governor. Those criteria were that theperson either be pregnant or five years of ageand under, and live within a five-mile radius ofthe plant. Thus, the conclusion reached is that141,500 people "overresponded" to the advisory;that Is, evacuated without being told to do so bythe Governor's advisory. Presumably, the"shadow" extends for some distance away from TMI,that distance being established by the presenceof evacuees. Based on this rather arbitrary wayof measuring proper response, it Is concludedthat people will "overrespond" to a nuclear acci-dent. Overresponse is synonymous with takingaction when no recommendation from an officialsource is made. Often this Is referred to as"spontaneous evacuation." A corollary of theshadow phenomenon theory Is that spontaneousevacuation, hence, overresponse will always occurwhen a warning is issued for a nuclear plantaccident due to people's innate fear ofradiation. Currently, this theory Is being pro-moted in interventions concerning emergency pre-paredness at Atomic Safety Licensing Board

351

352

Hearings on several nuclear power plants. Itis the thesis of this paper that this theory ismisleading and based on unfounded assumptions.Adoption of this view may actually jeopardizepublic safety if an emergency does occur.

III. ASSUMPTIONS OF THE "SHADOW PHENOMENON"

Advocates of the shadow phenomenon resttheir proposition on several assumptions thatare never explicitly defined. These assumptionswarrant Identification

A. The Information Assumption.For this "overresponse" to occur, it is

implicitly assumed that the Governor's advisorywas the only information reaching the public thatcould be validly used to make a decision to eva-cuate. Since the advisory provides the con-ditions by which overresponse is measured, it canhe construed that in order to be "rational"people could only respond to the advisory.

B. The Insider Assumption.The phenomenon assumes that "overresponse"

is an objective condition that can be measuredusing universally accepted criteria. In doingso it implicitly assumes that the insiders, thatis, the people who are part of the system understress, hold the same definitions of appropriateresponse as those who observe response after theemergency. It also assumes that insiders definerisks and risk levels in the same way as the out-siders.

C. The Knowledge Assumption.The shadow phenomenon assumes that the

language for presenting the critarla and inform-ation in the Governor's advisory were compatiblewith the knowledge systems of people receivingthe Information. That is, people knew where the5-mile radius was, people could distinguish thehealth risk to pregnant versus non—pregnant womenand could comprehend the use of five years of ageas an evacuation criterion.

B. The Individualism Assumption.The shadow phenomenon assumes that people

make decisions and behave as individuals withinan undifferentiated social space. That Is, thatsocial groupings play no role In behavioralresponse to an emergency.

How valid are these assumptions? In thenext section I attempt to provide an answer tothis question by addressing some of the basicquestions posed in the Introduction to thispaper.

IV. ANALYSIS OF THE ASSUMPTION

A. The Information Assumption.While it is true that more people evacuated

than was recommended by the Governor's advisory,their actions are consistent with scientificknowledge on how people behave in disasters.People at TMI responded to information availableto them at that time, and their behavior isunderstandable given the nature and content ofthat Information.

First, the Governor's advisory did not tellthe entire population what to do, only a specificsubgroup. Without information directed at othergroups not included in the pregnant women andsmall children categories, It Is quite Inaccurateand misleading to say that people "did other thanwhat the Governor's advisory suggested." Withoutinformation from the Governor they acted on thebasis of other information. Support for this isfound in the NRC telephone survey (Flynn, 1979).Only 14% of those who evacuated said that theGovernor's advisory was critical in making theirdecision. Thus, they made decisions on the basisof other factors.

Of the families that did not evacuate 71%stayed because they were not told to leave.Thus, many who stayed were waiting for guidancefrom official sources, and many who left actedon sources of information that filled the gapcreated by the lack of information in theadvisory. Had the Governor's advisory giveninformation to other groups of people, a differ-ent pattern of response would have likelyoccurred.

The theory of spontaneous evacuation assumesthat the Governor's advisory was the only infor-mation people acted upon. This simply was notthe case. According to the NRC survey the majorpiece of information that was critical in decidingto evacuate was the situation with the hydrogenbubble (30%). People did not behave contrary tothe advisory, but acted on the basis of theirsituational perceptions of the accident.

Further evidence that raises questions aboutthe validtty of this assumption Is provided by astudy of news media coverage of the accident(Stephens and Edison, 1982). The data In thestudy by Stephens and Edison show splits in bothlocal and national news coverage betweenreassuring and alarming statements on a range ofsafety issues including emergency preparedness,evacuation, radiation exposure, reliability ofInformation, plant conditions, the chance of ameltdown, and the hydrogen bubble. Some newssources were stating the need for a "totalevacuation of the immediate area" and many empha-sized the likelihood of an explosion ormeltdown.

Furthermore, people were exposed through themedia to the discustsons of other recommendationsand advisories by various officials. For exampleon Thursday, March 29, the second day of theaccident, a radio announcer mistakenly Inter-preted an Interview with Dr. Steinglass of theUniversity of Pittsburgh as an order for pregnantwomen and preschool children to evacuate ratherthan a cautionary recommendation. Fridaymorning, the Dauphin County Civil DefenseDirector Informed the public that an evacuationcould be ordered; however, directions on where togo and what to do also accompanied the statement.That same morning the Governor made a 10-milesheltering recommendation. Later in the day theofficial evacuation recommendation was made. OnSaturday, the Nuclear Regulatory Commission(NRC), one of the most credible Information sour-ces on the accident, went on record in a newsconference that a 20-mile evacuation order wasbeing considered.

353

In light of this evidence, the Governor'sadvisory is a poor standard by which to judge theappropriateness of response and was fairlylimited in its influence on evacuation behavior.People responded to the range of information inthe system. Furthermore, it appears that It wasthe uncertainty of the information that encour-aged evacuation and not deep-seated fears ofradiation.

B. The Insider Assumption.The preceding argument underscored the

inappropriateness of using arbitrary criteria tojudge behavior. Even if the imposition of a5-mile/pregnant women and small children cri-terion is accepted as valid, the shadow phenome-non still assumes that the response of the peopleto a disaster can be judged as being correct orincorrect by people who did not experience thedisaster. This outsider, however, has differentinformation than the insiders did at the time ofthe emergency. Often the situational factorsthat may lead to evacuation cannot be captured inthe research methodologies used to measure andunderstand behavior. Risk perceptions of theoutsider do not necessarily mesh with those ofthe insiders-

Rome empirical support for the questionablevalidity of this assumption is provided in afollow-up study of the TMI residents (Houtset al., 1980). In a reinterview of a sample ofpeople drawn from the NRC survey sample, respon-dents were asked if they would evacuate if asimilar accident occurred. The percentage ofpeople indicating they would evacuate was similarto the actual levels of evacuation behavior inthe accident. If people had thought they hadoverresponded we would have predicted much lowerpercentages for response to this behavioral*intent question. It seems likely, therefore,that researchers who judged people to be over-responders simply failed to appreciate thecultural and social context of the emergency. Atthe time TMI presented distinct risks with greatuncertainties. To use an analogy, if the possi-bility of a dam failure confronted flood plainoccupants would it be valid to label peoplemaking precautionary evacuations as"overresponders" when the dam did not fail?Would It be due to deep-seated fear of water?Unlikely. The mystique of radiation toresearchers is probably far greater than forexcess precipitation.

C. The Knowledge Assumption.The shadow phenomenon assumes that people,

in arriving at an evacuation decision, understoodthe context of and basis for the Governor's advi-sory. Research shows, however, that the publichas a poor technical grasp of nuclear power andionizing radiation concepts (Reed and Wilkes,1976). Furthermore, the public generally failsto even distinguish between various stages of thefuel cycle or between peaceful and militaryapplications of fission technology (Lindellet al., 1978). Thus, it is difficult to believethat the public would grasp the full meaning ofdifferentiating two high risk groups in the evac-uation, recommendation. While it may be ofsocietal value to protect pregnant women and

children, It usually is used as a means ofprioritizing protection and not to exclude othergroups from a margin of safety.

Furthermore, while it is much more likelythat people understand the concept of "miles" ofdistance, it does not necessarily follow thatthey are always good estimators of distance orview the five miles as a precise boundary.Considerable research has been conducted on howpeople perceive their environments and spatialorganization of these environments (Golledge andMoore, 1976; Saarinen, 1976; Rapoport, 1977;Lynch, 1960). One finding from these studies Isthat people are often poor estimators of distanceand, certainly, variable In the way they biasdistance estimates. Moreover people do notalways organize space in terms of concentricradii. Rather they use roads, landmarks, naturalboundaries and other natural and human featuresof the landscape as reference points. In addi-tion, it is unlikely that even with a goodreference point for the five-mile boundary, theradius was judged by people to be a fixed oraccurate barrier. People may not perceive thatairborne radiation would somehow obey this arti-ficial designation of a barrier that the Governorimplied was intransverslble by radiation.

D. The Individualism Assumption.The shadow phenomena assumes people will

evacuate as individuals and not in a largersocial context. This simply Is not borne out bya large range of evacuation studies in a varietyof disasters. The propensity of most people Isto evacuate In a group context (Drabek, 1969;Drabek and Stephenson, 1971; Perry et al.,1981;). For example, the Drabek studies indi-cated that of families intact during a flooddisaster, 92% evacuated together and 64X of thoseapart evacuated after reuniting. In othersituations people also behave In a collective orgroup context that does not necessarily Involvefamilies but friends, neighbors, or strangers aswell (Gruntfest, 1977).

In the case of TMI It Is difficult to assumethat people meeting the evacuation criteriaimposed by the Governor would behave as indivi-duals. A preschool child can hardly evacuatewithout assistance; an older sibling would not berequired to stay at home; and husbands andpregnant wives would be unlikely to part. Whileit has been acknowledged that in a family contextthere would have been at least 10,000 legitimateevacuees (Perry and Lindell, 1983), this stillignores the effects of extended family eva-cuation, peer group evacuation, and the Influenceof social processes in evacuation.

To gain a better picture of family evacuationat TMI, we reanalyzed the NRC survey data (Flynn,1979) to estimate how many households behaved asfamily units and how many were fragmented by theevacuation. The results show that fragmentationwas highest within 5 miles of TMI and decreasedwith distance from the plant. This suggests theevacuation recommendation may have contributed tomore fragmentation than would normally beexpected. Still, the majority of families didbehave as single units. Furthermore, as is docu-mented, a significant correlate of evacuation atTMI was having a friend or neighbor do so (Cutterand Barnes, 1982).

354

V. AN ALTERNATIVE EXPLANATION

The evacuation at TMI is largely explainablein light of scientific knowledge of how peoplerespond to warnings of an impending threat(Miletl, 1975; Mcluckie, 1973; Mileti et al.,1981; Quarantelli, no date). The two major fac-tors which seem to consistently affect behavior,regardless of hazard, are the situational percep-tion of risk and emergency information. Theself-reported data in the NRC survey support thispattern. The two major reasons cited by TMI eva-cuees for leaving were "the situation seemeddangerous" (91%) and "information on thesituation was confusing" (83%). At the time,people were hearing information that would leadthem to believe their safety was threatened.They were also hearing vastly conflictinginformation about the risks, what was being doneto remedy the situation, and what they should do.The confusing information was a signal thatemergency managers were not controlling the acci-dent. Evacuation was a prudent response that didnot need to follow from an "official order." Asequential or lexiographic model of decisionmaking in response to warnings provides an alter-native explanation to spontaneous evacuation dueto deep seated fear of the "shadow phenomenon"model.

In a simplified form, this model suggests thefollowing factors play a sequential role indetermining response: awareness of a risk,belief that the risk is real, personalizing therisk, evaluating alternative actions, anddeciding on a course of action. At TMI, astereotypical pattern of thoughts behind a deci-sion to evacuace could be characterized asfollows:

(1) A person became aware that a problemexisted.

(2) That person sought more information on theaccident situation.

(3) Information conveyed a sense of personalthreat to that person.

(A) The situation was not only threatening, butit was unknown whether the hydrogen bubblewould explode.

(5) A good means of protection was to get awayfrom the reactor.

(6) No other form of protection was seen to beeffective.

(7) The weekend situation gave that person mobi-lity to evacuate.

(8) Other people were perceived to beevacuating.

This decision process reflects a normal andpredictable course of behavior. People did notleave merely because they dread radiation.Moreover, they did not panic. In fact, the move-ment was slow and drawn out over a four- or five-day time frame, hardly what could be described asspontaneous. In sum, people responded in a wayconsistent with the disaster situation that wasconveyed to them.

VI. TMI AND PREDICTING EVACUATION BEHAVIOR

Based on the evidence presented in thispaper, it is certainly possible to use TMI topredict some patterns of behavior in a futureaccident. This is feasible because TMI was not atotally unique phenomenon. Rather, the patternsof response were largely consistent with thegeneral premises of human behavior in emergen-cies. People behaved in a manner consistent withthe information they received, the situationalperceptions of risk they formed, and the socialcontext of the accident situation. The notionsof "shadow" and of "overresponse" are largelyunfounded and best interpreted as creations ofresearchers rather than any social reality. Thisconclusion is both encouraging and disheartening.The conclusion that human response to a nuclearaccident is basically similar to a range of othertypes of emergencies is encouraging because weknow from experience how to issue warnings tominimize the problems that were experienced atTMI. While the inherent uncertainty in that par-ticular situation made emergency management func-tions difficult, the mere fact that a warning wasissued fails to account for the uncertainty andconfusion. An accident at the Ginna NuclearPower Plant outside of Rochester, New York, in1982 helps underscore the importance of an effec-tive warning system. Despite a release ofradiation, no evacuation ensued. This can beattributed in part to the successful response tothe steam tube rupture in which operatorsreturned the plant status to safety. It is alsoattributable to more effective communicationlinks to the local emergency response network andthe public, and, to the issuance of concise andcredible information. If evacuation behavior ina nuclear power plant accident were governed byinnate fears of radiation, some evacuation wouldhave been observed at Ginna regardless ofemergency management effort. Instead, thewarning system served to guide behavior In amanner consistent with the risks.

On the other hand, our conclusions aresomewhat disconcerting, as well. If emergencyresponse to a nuclear power plant emergency issubject to the same problems encountered in otheremergencies, a truly serious accident will resultin huraan casualty. This Is far more likely toresult from failure to heed a warning than fromoverresponse. Future litigations over emergencyplanning after a serious accident, assuming aviable: industry, would concern underresponse, nota sha.low of fear. Emergency planners should takesteps to develop effective warning systems whichseek to minimize both overresponse and under-response. This is feasible. Attempts to dealsolely wivth a shadow phenomenon could have dele-terious effects. That shadow is invisible; theconsequence of non-response will not be.

REFERENCESCutter, S., and K. Barnes, 1982, "Evacuation

Behavior and Three Mile Island," Disasters6, 116-124.

355

Drabek, T., 1969, "Social Processes in Disaster:Family Evacuation," Journal of Marriage andthe Family 30, 336-451.

Drabek, T., and J. Stephenson, 1971, "WhenDisaster Strikes," Journal of AppliedSocial Psychology 1, 187-302.

Flynn, C., 1979, Three Mile Island TelephoneSurvey: Preliminary Report on Proceduresand Findings, NUREG/CR-1093, U.S. NuclearRegulatory Commission, Washington, D.C.

Golledge, R., and G. Moore, 1976, EnvironmentalKnowing, Dowdon, Hutchingson, and Ross,Stroudsburg, Pennsylvania.

Gruntfest, E., 1977, "What People Did During theBig Thompson Flood," Working Paper #32,Institute of Behavioral Science, Universityof Colorado, Boulder.

Houts, P., H. Miller, G. Tokuhata, and K. Ham,1980, Health Related Behavioral Impacts ofthe Three Mile Island Nuclear Incident,Pennsylvania Department of Health, Harris-burg, Pennsylvania.

Johnson, J., 1983, "Planning for SpontaneousEvacuation During a Radiological Emergency,"Nuclear Safety 25 (March-April), 186-193.

Johnson, J., and D. Ziegler, 1983,"Distinguishing Human Response toRadiological Emergencies," EconomicGeography 59, 3-86-402.

Johnson, J., and D. Ziegler, 1984, "A SpatialAnalysis of Evacuation Intentions at theShoreham Nuclear Power Station," pp. 279-301in M. Pasqualetti and K. Pijawka (eds.),Nuclear Power: Assessing and ManagingHazardous Technology, Westview Press,Boulder.

Lindell, M., T. Earle, J. Hebert, and R. Perry,1978, Radioactive Wastes: Public AttitudesTowards Disposal Facilities. HARC Report411-004, Battelle Human Affairs ResearchCenter, Seattle.

Lynch, K., 1960, The Image of the City, MITPress, Cambridge.

McLuckie, B., 1973, JThe Warning Systems: A SocialScience Perspective, National WeatherService, Washington, D.C.

Mileti, D., 1975, Natural Hazard Warning SystemsIn the United States, Institute ofBehavioral Science, University of Colorado,Boulder.

Mileti, D., J. Hutten, and .T. Sorensen, 1981,Earthquake Prediction Response and Optionsfor Public Policy, Institute of BehavioralScience, University of Colorado, Boulder.

Perry, R., and M. Lindell, 1983, "Nuclear PowerPlant Emergency Warnings: How Would thePublic Respond," Nuclear News (February),49-57.

Perry, R., M. Lindell, and M. Greene, 1981,Evacuation Planning in Emergency Management,Lexington Press, Lexington, Massachusetts.

Quarantelli, E., undated, Evacuation Behavior andProblems, Disaster Research Center, OhioState University, Columbus.

Rapoport, A., 1977, Human Aspects of Urban Form,Pergamon, Oxford.

Reed, J., and J. Wilkes, 1981, "Technical NuclearKnowledge and Attitudes Before and AfterTMI," paper presented at Annual Meeting ofthe Society for Social Study of Science,Atlanta, Georgia, November, 1981.

Saartnen, T., 1976, Environmental Planning:Perception and Behavior. Houghton Miffltn,Boston.

Stephens, M., and N. Edison, 1982, "News MediaCoverage of Issues During the Accident atThree Mile Island," Journalism Quarterly(Summer), 199-259.

Ziegler, D., S. Brunn, and J. Johnson, 1981,"Evacuation from a Nuclear TechnologicalDisaster," Geographical Review, 1-16.

ANS Topical Meeting on Radiological Accidents-Perspectives and Emergency

Warning Human Populations of Technological Hazard

George O. Rogers and Jiri Nehnevajsa

ABSTRACT1 Warning people of an impending hazardseeks to make them aware of the threat and to elicit actions thatwould minimize the dangers to life and property.2 Becausetechnological and natural hazards differ in important ways, thealerting and notification process for technological and naturalhazards is also different. One of the differences rests in theability of people to detect many natural hazards in a directsensory manner; technological hazards often make suchdetection more difficult. For example, detection of radiologicalreleases without instrumentation is nearly impossible, but evenwith icrnadoes where warning is notoriously difficult, people areat leaut able to use their senses to detect the potential forhazard. Hence, warning for technological hazards is in someways more problematic, generally representing a rather rapidshift from normalcy to emergency. This paper builds on thesignificant foundation of natural hazard warning research indeveloping a model of warning suitable for technologicalhazards. This model specifically examines immediatecascading, or networking, of the warning signal and message, sooften reported in the natural hazard literature. The implicationsfor technological and natural hazard warnings systems areexamined.

I. INTRODUCTION

Warning people of an impending danger may bepartitioned into two somewhat distinct aspects. The first dealswith alerting the public that something is wrong, that somehazard is imminent. The second concerns the ability ofemergency officials to communicate the warning message toprompt appropriate action. The primary issues of alertingrevolve around the ability to make people aware of the threat.This alerting often involves the technical ability to develop,construct, maintain and use a warning system, which mayconsist of sirens, bells, whistles, television and radio broadcasts,telephone systems and even social organization. The primarynotification issues center on the public's interpretation of thewarning message. The interpretation of the warning message isfundamentally important in the selection of appropriate action inresponse to warning. The focus of this paper is on the socialprocesses associated with alerting the public to potential danger.

Hazards are broadly cast as technological and natural.Technological hazards are generally characterized by the failure

'The research on which this paper is based was partially supported by theFederal Emergency Management Agency (Cooperative Agreement No. EMW-K-1024). The authors accept full responsibility for the contents herein andgratefully acknowledge the support, comments and criticisms offered bycolleagues at the University of Pittsburgh.

^ e former is the alerting function of warning systems, and the latterconstitutes their notification function.

of a technological system(s). While technological hazards aregenerated by human interaction with the environment, naturalhazards may be viewed as "acts of God." Hence, questions ofculpability are often associated with technological hazards. Thedevelopers and operators of the technological system becomelegally responsible for the safety of its operation. Suchculpability is seldom attributed to natural disasters. At least forsome kinds of technological crises the potential impact area ispredictable. While even the best meteorologist has difficultydetermining the pathway(s) and point(s) of impact for anapproaching tornado, hazardous facilities that are geographicallyfixed, thus there is the advantage of being able to establishemergency planning zones in proximity of the fixed facility. Onthe negative side, at least some technological hazards are lessdetectable than any natural hazard. Nuclear exposures, forexample, are not detectable by any of the five human senses,some toxic chemical releases are similarly undetectable.However, other technological hazards make possible the use ofmany of the same detection criteria as natural disaster threats.For example, the danger of dam failures is often brought OR byheavy rains and results in flooding that is detectable by the samemechanisms as other forms of flooding. This paper focuses onthe social aspect of the alerting process, primarily for hazards ofa fixed, or at least known geographic location. However, thesocial principles applied to these fixed geographic locationsapply equally well to other technological hazards and evennatural hazard alerting situations.

II. BACKGROUND

Warning messages pass through a variety of pathwayswhich may color their meaning. Some of these pathways involvecognitive functions, others have to do with social structuralconsiderations. An individual's interactions with others formsocial networks. Even though these networks have many forms,their routine and established nature has led to widely accepted

empirical generalizations about how they function in society atlarge (e.g., Parsons 1951, Coleman et al. 1957, Granovetter1973, Blau 1977) and in particular how they function duringemergency warning. Two general propositions are stronglysupported by the disaster literature (Williams 1964): First, thatpeople respond to emergency warnings in a context of their priorexperience, extant social and physical environment and existingconditions which interact with the warning message. Andsecond, that the degree to which the warning message isreceived depends on the nature of the message, taken in thecontext of the social network, and the prior behaviors of all socialactors in processing such information. Hence, people in socialnetworks in specific locations have extant estimates of the threatpresented by the environment in which they live. Theseestimates and their experience vector provide the data base fromwhich the selection of behavior is derived-the decision toaccept, ignore, disseminate, challenge, or confirm the warningmessage (Baker 1979).

357

358

WARNING ISSUED

No Response

WARNING MESSAGE CONTENTAmbiguity/CertaintyEstimated Magnitude,Timing, t Location

No-^Null Response

WARNING MESSAGE PROCESSEDDisaster Experience/Proximity

Confidence in SourceFamily Members/StructureObservation/Interaction

THREAT PERCEPTION

BEHAVIOR SELECTIONDo Nothing/DisseminateWarning/Confirm Warning

Respond to Warning

-St.EVALUATE BEHAVIOR

Figure 1 - Warning Process

Emergency warnings may result in the recognition of threatwhich creates a psychological discomfort. One importantmechanism to alleviate this involves efforts to reduceuncertainty. The warning process (Figure 1) involves bothfactors affecting the message and characteristics of the receiver.Once the warning is received the content is evaluated in terms ofthe certainty and ambiguity of the estimated severity, timing andlocation of impact. Essentially, "Is it likely to effect me? Whenwill it occur?" The evaluation of the warning message results inthe determination of its relevance. If the message content isdeemed irrelevant, no emergency response is likely. However,should the warning message be considered relevant, themessage is further processed in the context of prior disasterexperience, relative proximity, confidence in the source ofwarning, interpretation and discussion with members of thesocial network. The warning message is processed in thecontext of the existing social structure, which results in, at least,the initial perception o! threat. The aimulalive process providesthe foundation for the selection and evaluation of emergencybehavior.

The social process which is then triggered also serves tofurther disseminate the warning message. When an individualreceives, recognizes, verifies and believes the warning messageand deliberately disseminates it to others in the social network, apurposive warning dissemination takes place. An incidentalwarning takes place when the individual in the process ofseeking confirmation of the warning message, inadvertently getsin touch with someone who has not been alerted yet. Warningconfirmation and dissemination through the social network helpswarn previously unwarned people, if indirectly, and confirms themeaning of the original warning for those having received itpreviously. "Instead of trying to stop...people [from] calling oneanother...ways ought to be found to take advantage of such caiis

so as to improve the dissemination of warning messages..."(Kendrick 1979:346). Furthermore il should be added that bothconfirmation and dissemination of warning through the socialnetwork improve response to emergency warnings (Quarantelliand Dynes 1376, Perry et al. 1980, Drabek and Boggs 1968,Mileti and Beck 1975, and Rogers and Nehnevajsa 1984).

Effective warning messages have been described by Janis(1958) as requiring a balance between fear-arousing and fear-reducing statements. Fear-arousing statements describe theimpending danger in sufficient detail as to evoke vivid mentalimages of the crisis, reducing the possibility tor surprise as thedisaster evolves. Fear-reducing statements realistically presentthe mitigating factors of the impending situation, and provideinformation regarding realistic actions to be taken by authoritiesand individuals, both independently of one another and jointly.The fear-arousing conient of fhe warning message alerts thepublic to the potential for harm, while the fear-reducingstatements consist of notification of appropriate mitigating action.

III. THE ALERTING PROCESS

People are alerted to the potential for danger by a varietyof sources. These sources of warning are broadly classifiable:warning by authorities, from the mass media, and thosetransmitted through the social network. Drabek (1969) and Perryet a!. (1981) refer to warning from authority as those messageswhich are generated from and disseminated by emergency-services organizations (e.g., police or fire departments, civildefense organizations or the national guard), while mass mediawarnings usually come from radio and television, although inslowly-evolving disasters, the print media also play a crucial role.The social network provides warning through relatives, friendand neighbors. Perry el al. (1981) report that 41.2% of firstwarnings for riverine floods in lour communities came fromauthorities. While there was apparently no time for mass mediawarnings in two communities, the mass media accounted for8.1% of warning alerts. Nearly half of the first warnings in thesefour communities stemmed from the social network--37.6% fromfriends and neighbors, and 13.0% from relatives. While thedistinction between social network alerts that purposelydisseminate the warning, and those that incidentally warn othersis not possible on the basis of the present evidence, the receiptof first warning is often made through personal contacts withmembers of the oocial network (Perry 1981, and Mileti 1974).

Around Three Mile Island more people received their firstwarning (were alerted) via social networks than expected toreceive such warnings. Flynn (1979) found that only 6%expected to be alerted by friends, neighbors and relatives, butBrunn et al. (1979) and Barnes et al. (1979) report 18% to 25%actually claimed to have received their first warning from socialnetwork sources. Of those people receiving warning on the firstday of the accident, 22% were alerted through the socialnetwork, and for those with the highest saliency (living within 6miles), 43% were alerted by people in their social network(Barnes et al. 1979). This is consistent with other researchwhich suggests that "word-of-mouth" warnings are more likelyamong people most likely to be affected by the impendingdanger (e.g., closest to the threatened area (Diggory 1956).Hence, for fixed-site technological hazards, where the saliencyfor nearby residents is fairly clear, social networks may be moreeffective than in situations where the proximity of hazard is lessclear. In natural disasters, in which the probable impact areacan be ascertained reasonably well, significant proportions ofpeople are also alerted through the social network. For example,Perry (1981) reports 31.7% and 38.6% of the people receivedtheir first warning from others in their social network inconnection with the volcanic activity and floods respectively.

3f>9

The overall emergency alerting process can be considered ascomprised of two basic processes. The warning alert processdetermines the capability of the warning technology (e.g., sirens,bells) to deliver the warning message to the public. Theeffectiveness, of course depends on factors of the physicalenvironment and the system technology, both constrained bynatural laws. Siren sound coverage, ambient noise levels,warning signal attenuation, biological hearing capability,acoustical properties oi the alerting signal are among the salientconsiderations. To the extent that human activities alter suchparameters, such as sleeping, operating equipment or listeningto music, social behavior is clearly critical to the actualized initialreceipt of the message. The dissemination of the warning alerttakes place through the household and neighborhood alertprocesses. The household or "area" process involves the inira-household dissemination of the warning message, while theneighborhood process represents the inter-householddissemination of warning.

The warning alert process results in some householdsbeing completely alerted by the initial warning signal, others mayhave at least one person alerted, and in some households noone will be alerted by the initial warning signal (Figure 2). Thehousehold alerting process characterizes the distribution of thewarning signal within a household that is partially alerted by theinitial warning signal (i.e., at least one person). Householdswhere everyone is aleited, either by the initial warning signal orthrough intra-househcld dissemination, become the potentialwarning message transmitters in the neighborhood process.

I WARMING ALERT PROCESS |

HouseholdAlert Complete

[HOUSEHOLD ALERTING PROCESS

NEIGHBORHOOD ALERTING PROCESS|

Neighborhood Alert Complete

Incomplete Warning Alert

Figure 2 - Alerting Process

Households that initially remain unalerted may receive thewarning through the neighborhood alerting process. Thehousehold alerting process is consistent with family re-unificationfor household emergency response (Rogers and Nehnevajsa1984, Frazier 1979, and Drabek" 1969), w.iich finds thathouseholds prefer to respond to crises as a unit. Both thehousehold and the neighborhood alerting proctsses provideconsistent confirmation of warning that often lakes olace duringemergency warnings (Rogers 1985).3

IV. WARNING ALERT MODELS

Considering a late-night (i.e., 12 midnight to 6 a.m.)warning, Nehnevajsa (1985a) incorporates three major lacfors inassessing warning alert: 1) The effect ol non-sleeping activities.2) the effect of intra-househokJ networking, and 3) the offset ofinter-household networking. Activity probabilities are based on adetailed study ol lime use conducted by the Univfirsity ofMichigan (Jusier et al. 1983) which results in iaienightprobabilities of being awake of .092 and .058 between midnightand 2 a.m., and 2 a.m. and 6 a.m. respectively (Humrnon et al. InPress). Intra-household networking is considered on the basis ofhousehold by size and composition, peak alerting signal levelsas a function of distance from the warning signal source,attenuation rates for different types of housos ami residentialconditions. Somewhat conservatively assuming that about half ofthose alerted by the initial warning signal wiii rrjke a singlecontact with another household (ever though 87.5% ofrespondents in a recent University of Pittsburgh study expecttheir neighbors to contact others, even in the middle of the night,to warn them of impending danger), initial "acoustic" alerting of69.0%, would be augmented to 79.3% of the residents, giventhese basic considerations. A second acoustic signal, resultingin 72.8% of the people alerted would be enhanced to 82.2%alerted.

Even an elementary model, which only accounts forarousal probabilities by household size significantly reduces theproportion of people left unwarned (Nehnevajsa 1985b).Assuming that only one in four people alerted by the initial signalwould attempt to make contact with others in their social networkthe initial warning signal leaves 15.5% of the people unalerted.However, a single social network contact decreases theproportion of unalerted people to 11.6%, and a secondnetworking attempt reduces the unalerted proportion to 8.6%.

One significant limitation of these models revolves aroundthe timing ot the networking process (Landry and Rogers 1982and Nehnevajsa 1985b). The warning process that incorporatesboth the initial alerting system, which is technologically (e.g.,sirens, bells, television, or radio) based, and diffusion of warningthrough the social network, involves initial alerting via a"broadcast" process,4 and subsequent alerting via a "birth"process5 (Lave and March 1975). Both processes are timeoriented (t) and limited by the size of the population to be alerted(N). The broadcast and birth processes are representedrespectively by

dn/dl - a,(N-n), and dn/dt - a2n(N-n),

where n, is the alerted population at the beginning of each period(t0, t,, t2..., and a, and a2, are proportions summarizing thediffusion properties of the respective processes. Combiningthese two processes into a single warning system

3This papor does not examine tho significant issues associated with thenotification process, which involves the belief and interpretation ol tho warningmessage and the selection of behavior. The significant issues ol behavioralcontagion, will people take action independently or in conjunction with thebehavior of Iriends. neighbors and relatives, and the ellcct of source of warningconfirmation, or type ot warning notification on behavior selection are excluded.

4Thls process rests on tho broadcast ot the warning message by tochnicalmechanisms, such as television, radio, sirens, bells, whistles or a combination ofthese specific technologies in combination with organizational assistance. It isrolerred to as broadcast because tho message is broadcast from a rolativolycentralized source to tho public.

5This procoss rests on the dissemination of the message among people. Itrelies on a less centralized warning dissemination, whore each recipient passesor at least attempts to pass Iho message to others in tho network.

3K0

dn/dl. k{a,(N-n)} + (i-k){a2n(N-n)},where k, is the proportion receiving the warning alert signal, and(1-k) represents the proportion not alerted by the broadcastsignal.

Using this classical model of the dissemination of warning,the timing of warning over the initial period can be examined.Suppose ihe broadcast warning system only operates in the firstthree minutes of the warning period, even though no warningsystem that we are aware of operates only in the first fewminutes without being reactivated in later periods. Furthersuppose that k is equal to the proportion of non-sleepers. This isequivalent to saying that arousal from sleep need not beconsidered for those that are not asleep. Finally consider abroadcast process efficiency (a,) of only .5, and a birth processeffectiveness (a2) of .3. This broadcast efficiency is well belowthe acoustical warning rate reported by Nehnevajsa (1985a), andthe contagion effect is substantially below people's expectationsand reported incidents. Even assuming these conservativesystem parameters, the warning system alerts 76.2% in thedead-ol-night (2 a.m. to 4 a.m.) in the first 15 minutes (Figure3).6 Given the drastically larger proportion of non-sleepingpeople between 8 a.m. and 10 a.m., and 8 p.m. and 10 p.m., theproportion warned exceeds 80% in the initial periods of warning,resulting in approximately 88% being alerted in the first fifteenminutes. Given quite different broadcast alerting probabilities,reflecting the period of the day differences, the results at the endof fifteen minutes are remarkably similar, but the trajectory withinthe period is very different. Hence for technological hazards withvarious onset times, the broadcast system requirements will besomewhat different. For technologies either with long onsettimes or where warning systems can be activated early, systemsmay place greater dependence on the birth process inemergency warning.

V. POLICYCONCLUSIONS

IMPLICATIONS AND

Figure 3 - Timing of Warning

sMinutes is perhaps better described as steps, inasmuch as we remainuncertain about the exact duration required lor message passing. While givenadequate saliency household contagion probably take less than a minute,message passing in the neighborhood process is loss certain. Hencfc, what welabel as minutes, reflects steps that probably range in duration Irom somewherenear 15 seconds upwards to 3 to 5 minutes The actual duration of these timeintervals is almost certainly dependent on the nature of the hazard. Its saliency,timing and action requirement.

Social network alerting deserves full recognition as a validaspect of the overall emergency warning process. It cannot beassumed that everyone, especially younger children, is able loproperly interpret the meaning of the warning signal, even if it is"heard," and thereby recognize the impending threat. Therefore,Ihe cascading effects of networking become an integral part ofthe system. The dissemination of warning provided by Ihecascading of the warning message through the social networksignificantly enhances coverage of the warning alert. Hence,emergency warning systems can effectively alert residents inadjacent areas by taking advantage of the social nelworkdissemination of warning. This is particularly true for fixed-location technological hazards, such as nuclear power plants.However, it is incumbent upon risk managers of such facilities toincrease public awareness of the potential for hazard, ability lorecognize and interpret the alerting signal, and awareness ofwhat actions to lake.

All emergency warning systems take advantage of both analert signal and a further dissemination of the warning throughthe social network. The trade-off between the two processesrests on considerations of cost and timing of adequate coverage.Because the birth process depends on alerted people todisseminate the warning message, Ihe more expensivebroadcast ol emergency warning is inherently faster. Forhazards with onset trajectories similar to hurricanes, the warningsystem can place more reliance on the networking process.Relatively slowly evolving emergencies not only provide time forthe social networking process to be highly effective, but thesehazards also allow people to become attuned to the impendinghazard, which "pre-charges" the network for further alerting andnotification as information about the hazard gets to be moreintensive, and the danger becomes more acute. On the otherend of the hazard spectrum, rapidly evolving hazards requiregreater reliance on a broadcast system, even though such asystem can never be completely effective on its own. Hence,one key factor in determining the extent to which the less costlysocial network dissemination of warning can be employedconcerns the technologies which permit an early detection olparticular hazards. Can the hazard be detected with sufficientlead time to alert the public? Another factor in the selection of anefficient emergency warning system (i.e., obtaining coveragewith an optimum mix of the broadcast and social networkprocesses) is the reliability of early warning, and appropriatepolicy decisions tc warn at early stages of a possible disaster.This involve? another trade-off between the issuance of earlywarning and the probability of a false alarm.

To the extent that there is actable time, any warningsystem can be improved, in the sense of alerting more peoplewith less time, both through improvements in the broadcastsystem and by enhancing the social nelwork process. Thebroadcast system can alert more people by enhanced coverage(e.g., louder signals, or morb complete distribution of warningdevices among the population). While these systemimprovements are desirable, the social network process can alsobe enhanced by encouraging people to contact others when theyare alerted. By encouraging people to become involved in Iheemergency warning process, emergency preparedness beyondbetter warning is improved, because people are more likely todevelop an understanding of the potential hazards, the nature ofpotential threats, the kinds of available protective actionsneeded, and take an active role in assuring their own safety, andin enhancing the safety of their relatives, friends and neighbors.

361

VI. REFERENCES

Baker, E. (1979) "Predicting Response to Hurricane Warnings:A Preanalysis of data from Four Studies," MassEmergencies. 4,9-24.

Barnes, K. et al. (1979) "Responses of Impacted Population tothe Three Mile Island Nuclear Reactor Accident," RutgersUniv., Dept. of Geography.

Blau, P. M., (1977) Inequality and Heterogeneity: A PrimitiveTheory of Social Structure. The Free Press, New York.

Brunn, S. D. et al. (1979) "Social Survey of Three Mile IslandArea Residents," Michigan State Univ., Dept. of Geography.

Coleman, J., et al. (1957) "The Diffusion of Innovation AmongPhysicians" Amer. Sociology Review. 20, 253-270.

Diggory, J. C. (1956) "Some Consequences of Proximity to aDisaster Threat," Amer. Sociology Review, XIX, 47-53.

Drabek, T. (1969) "Social Processes in Disaster: FamilyEvacuation." Social Problems, 16, 336-346.

Drabek, T. and K. Boggs (1968) "Families in Disaster: Reactionsand Relatives." J. of Marriage and the Family, 30,443-451.

Flynn, C.B. (1979) Three Mile Island Telephone Survey,"Mountain West Research, Tempe, AZ.

Frazier, K. (1979) The Violent Face of Nature, William Morrow &Co., New York.

Granovetter, M. S., (1973) 'The Strength of Weak Ties," Amer.J. of Sociology. 78,1360-80.

Hummon, N. P. et al. (In Press) "Time Budget Analysis and RiskManagement: Estimating the Probabilities of EventSchedules of American Adults," Enhancing RiskManagement and Decision Making. R. Waller et al. (eds),Plenum, New York.

Janis, I. L, (1958) Psychological Stress, John Wiley and Sons,New York.

Juster, F. T., et al. (1983), "1975 - 1981 Time Use LongitudinalPanel Study", The University of Michigan, Institute forSocial Research, Survey Research Center, Ann Arbor.

Kendrick, F., (1979) J_he Violent Face of Nature: SeverePhenomena and Natural Disasters. William Morrow andCo., New York.

Landry, T., and G. Rogers, (1982) "Warning Confirmation andDissemination," Univ. of Pittsburgh, Univ. Center for Socialand Urban Research.

Lave, C. A. and J. G. March (1975) An Introduction to Models jnthe Social Sciences. Harper & Row, New York.

Mileti, D. (1974) A Normal Causal Model Analysis of WarningResponse. Univ. of Colorado, Depl. of Sociology (Ph.D.dissertation).

Mileti, D. and E. M. Beck (1975) "Communication in Crisis,"Communications Research. 2,24-49.

Nehnevajsa, J. (1985a) 'Testimony of Jiri Nehnevajsa," Hearingbefore the U. S. Nuclear Regulatory Commission, AtomicSafety and Licensing Board, in the matter of Carolina Powerand Light Co. et al. (The Shearon Harris Nuclear PowerPlant, Unit 1), November 4 and 5.

Nehnevajsa, J. (1985b) "Some Effects of Social Networking onWarning Message Delivery," for International EnergyAssociates, Washington, D.C.

Parsons, T., (1951) The Social System. The Free Press, NewYork.

Perry, R. (1981) "Citizen-Evacuation in Response to Nuclear andNon-Nuclear Threats," Fed. Emergency ManagementAgency, Washington, D. C.

Perry, R. et al. (1980) J_he Implications of Natural HazardEvacuation Warning Studies for Crisis Relocation Planning,Battelle, Human Affairs Research Centers, Seattle, WA.

Perry, R. W. et al. (1981) Evacuation Planning jn EmergencyManagement. Lexington Books, Lexington, MA.

Quarantelli, E. L., and R. R. Dynes (1976) "The Family andCommunity, Context of Individual Reactions to Disaster,"Emergency and Disaster Management. H. Parad et al.(eds), Charles Press Pub., p. 231-244, Bowie, MD.

Rogers, G. O. (1985) "Human Components of EmergencyWarning," Univ. of Pittsburgh, Univ. Center for Social andUrban Research.

Rogers, G. O., and J. Nehnevajsa (1984) Behavior and AttitudesUnder Crisis Conditions: Selected Issues and Findings.U.S. Government Printing Office, Washington D. C.

Williams, H. (1964) "Human Factors in Waming-and-ResponseSystems," J_h§ Threat of Impending Disaster. G.H. Grossinet al. (eds), MIT Press, Cambridge, MA.

ANS Topical Meeting on Radiological Accidents—Perspectives and Emergency Planning

Emergency Preparedness Training for Local CommunitiesMichael J. Cooley and Karen K. Thompson

ABSTRACT

Detroit Edison, in cooperation with MonroeCounty, has developed a comprehensive train-ing program for local emergency workers inthe area surrounding the Fermi 2 NuclearPower Plant. Using expertise from bothorganizations, a program consisting of twovideotapes, two slide-tapes and nine narratedslide series was produced to address theworker-specific training needs of countyemergency workers. In June of 1985, theprogram was approved by Detroit Edison andthe Monroe County Board of Commissioners.To date, Monroe County has trained more than1,000 emergency workers.

This program has been so well received thatthe county staff has developed and presenteda modified version of this program to thegeneral public. The result of this coopera-tive effort is increased public confidencein emergency preparedness at the state, localand utility level and a renewed spirit ofcooperation and trust between the utility andlocal units of government.

Over the last few years, the issue ofemergency planning has received increasingattention, particularly in communities whichsurround nuclear power plants. Althoughlocal governments in Michigan are responsiblefor developing an emergency plan, the res-ponsibility for approving training programsfor emergency workers in local communitiesrests with the state. This division ofresponsibility sometimes requires a highdegree of cooperation by all concerned todevelop a comprehensive, yet community-specific program.

According to Appendix E of 10CFR50, aradiological orientation training programmust be made available to off-site emergencyworkers. In the State of Michigan, theEmergency Management Division of the StatePolice is the responsible agency for ensuring

that emergency workers are trained and pre-pared to handle emergencies relating to anaccident at a nuclear power plant. TheState has developed a four-part trainingprogram using NUREG-0654 as a guide. Theprogram encompasses basic information on theoperation of a nuclear power plant, effectsand detection of radiation, and the emer-gency plans at the state, local and utilitylevel. This program was implemented inMonroe County in 1984. Monroe County is thelocation of Detroit Edison's Fermi 2 PowerPlant, 30 miles south of Detroit. Fermi 2is powered by a boiling water reactor capableof generating 1100 megawatts.

The state-developed training programwas originally presented jointly by statepolice and radiological health personnel,the local emergency preparedness coordinatorand utility personnel. Overall, the programlength was nearly four hours'—much too longfor the purpose intended. Furthermore, theprogram did not address specific duties andresponsibilities of the local emergencyworkers.

To improve the program, Detroit Edisonand the Monroe County Office of CivilPreparedness began discussions to explorealternatives. The first, and perhaps thesimplest, was to revise the existing pro-gram. The county wanted a program with anew focus, one which related directly totheir emergency plan and addressed theirneeds. A second alternative was to engagean independent consultant to analyze theneeds of the county and to develop anddeliver a program. The local coordinatorsapproached the consultant who had assistedin the preparation of their emergency plan.The consultant offered a package which in-cluded a stand-up instructor, supplementedby slides and overhead transparencies.

There were several disadvantages thatdeveloped in exploring the use of a consul-tant for this project. For one, theconsultant proposed a two to eight hour pre-sentation. This was considered again to

363

364

to be much too long by the county coordina-tors. In addition, the instructors from theconsulting firm were stranscrB to thecommunity. The Office m Civil Preparednesswanted to increase their Involvement withthe emergency workers and the public bymaking the presentations themselves.

Another disadvantage in havingInstructors from a consulting firm was theirlimited availability. The county wanted aninstructor who would be able to schedule andconduct presentations throughout the year.

Finally, the'contract proposed by theconsultant was for development and initialimplementation of the program. Any revisionsand presentations in subsequent years wouldbe at an added cost. The county wanted tominimize their long-term reliance on outsideassistance. Ideally, one of their own staff,an individual familiar with the emergencyplan and with the local community, would bebest suited to make the presentations.

The Monroe County emergency plannersdecided on another alternative—to askDetroit Edison to work with them in develop-ing a new program. The emergency planninggroup at Detroit Edison had a close workingrelationship with the staff at the CountyOffice of Civil Preparedness. The utilitymet with the county staff and agreed to workcooperatively on the new program. The countystaff provided experts on their emergencyplan throughout the development phase of theproject. The Fermi 2 staff provided tech-nical expertise on plant operations and basicradiation protection. Detroit Edison alsoprovided overall project scheduling andcoordination of production activities includ-ing photography, videotaping and artwork.

In planning the program, Detroit Edisonand Monroe County personnel first agreed uponsome parameters. The County wanted a programno more than one hour and 20 minutes inlength. They also wanted it tailored tospecific groups of emergency workers, toappear realistic and be a high qualityproduction. The realism was achieved throughextensive use of local workers, equipment andfacilities In producing the program. Thequality was heightened through the use of avariety of audio-visual media and a pro-fessional actor. Because this new programwas being developed for Monroe County, theDirector of the Office of Civil Preparednesswas responsible for review and approval ofall program materials.

A schedule was developed so that com-pletion of the training program coincidedwith preparation for the annual Fermi 2emergency exercise. Six months were allowedto develop the program from start to finish.

The original four-part program developedby the State Police Emergency Management

Division formed the basis for the newMonroe County Training Program. The subjectmatter was adapted to specifically meet theneeds of Monroe County.

The new program evolved into 4 majorsegments. Part I, a 15 minute slide-tapeprogram, deals with plant operations andemergency planning at Fermi 2. Initialreview and approval of this segment wasperformed by the supervisor of emergencyplanning for Detroit Edison. Final approvalwas given by the Director of the Office ofCivil Preparedness.

Part II is a 15 minute videotape whichpresents Information about radiation in-cluding definitions, discussion of types ofradiation, how it is detected and measured,and how decontamination is performed. Thispart of the program was developed andapproved cooperatively by the Fermi 2emergency planning and health physics groups,the State Division of Radiological Healthand the County Health Department.

Part III, a 20 minute slide-tape pro-gram, was developed by the State PoliceEmergency Management Division and describesthe Michigan Emergency Preparedness Plan andresponsibilities at the state level.

Part IV covers the Monroe Countyemergency plan and includes specific learn-ing modules for each of the major areas ofresponsibility under the plan. These areasare: Emergency Operations Center (EOC)Staff, Public Information, Warning, LawEnforcement, Public Works, Fire Services,Health/Medical, Social Services, EmergencyMedical Services/Transportation and SchoolServices. EOC Staff Orientation, a12-mlnute videotape, was developed for EOCstaff officers. A 10-minute slide programwas prepared on each of the other ninemodules.

Parts I through III were presented asthe basic package for all county emergencyworkers. In addition, one of the tenspecific modules of Part IV were presentedto the individual groups. For example,firemen received Parts I, II, III and theFire Services module while law enforcementpersonnel received Parts I, II, III and theLaw Enforcement module.

The program was completed within theprescribed six-month time frame and pre-sented to the Monroe County Board ofCommissioners. Upon their approval, pre-sentations to county emergency workers began.To date, over 1,000 emergency workers havebeen trained. The vast i jority of thosewho have seen the program have found itbeneficial in outlining their emergencyduties and in providing the needed back-ground information.

365

Currently, a short review program isunder development for emergency workers whocompleted the initial training. As a resultof the high level of interest generated bythis program, the county has independentlydeveloped an abbreviated version of theprogram for the general public. To date,they have presented it to over 200 countyresidents.

In addition to providing the neededinformation to emergency workers, thisprogram has many other benefits. It hasincreased the credibility of the CountyOffice of Civil Preparedness with the public.It has also greatly increased public aware-ness of emergency preparedness at the localand utility level. Finally, there is anIncreased spirit of cooperation and trustbetween Detroit Edison and Monroe County.

The cost to Detroit Edison for theentire program was significantly less thanit might have been for some of the otheralternatives--well worthwhile for aninnovative approach to training and for theexcellent relationships built with thesurrounding community.

ANS Topical Meeting on Radiological Accidents-Perspectives and Emergency Planning

Emergency Planning, Public Information, andthe Media: Some Recent Experiences

Steven B. Goldman

ABSTRACT Within the last two years, events atseveral nuclear plants generated national mediaattention. Most were classified as "UNUSUALEVENTS" or "ALERTS". One event necessitated a"SITE AREA EMERGENCY" classification. None ofthese events posed a hazard to the generalpublic, yet each posed a serious threat to thecredibility of the utilities. These and otherevents show that (1) media response to a nuclearplant accident is not necessarily related to thetechnical severity of the accident and (2)utilities must have in place effective andcoordinated emergency public informationprograms, particularly for "lower level"emergency classifications.

I. INTRODOCTIOH

American electric utilities with nuclearunits have developed public information programsspecifically designed for nuclear emergencies.Regulations for nuclear emergency publicinformation CEPI) are few, but the consequencesof a mismanaged nuclear EPI program are farreaching.

This paper will recount the publicinformation problems of the March 1979 accidentat the Three Mile Island Nuclear Station,followed by a brief review of resultant EPIregulations. Utility and media response duringmore recent emergencies will be presented anddiscussed; finally, lessons learned from theseevents will be identified.

To identify the role of emergency publicInformation, it is helpful to divide nuclearemergency response into two broad efforts:

1) Placing the plant into a safe mode2) Protecting the health and safety of the

general publicThe former consists of the various

operational, technical and maintenanceevolutions. The latter also consists oftechnical aspects (dose assessment, protectiveactions). Public response to an emergency — andtherefore the protection of the public's healthand safety — is a direct function of the publicinformation effort. Lindell and Perry (1983)

show that public response to an emergency — anyemergency — depends upon:

o Pre-existing beliefs about the hazardo Interpretation of information

disseminated by official sourceso Credibility of those sourceso Distortion or addition of information

by unofficial sourceso Evaluation of the threato Choice of available responsesThus, in order to meet the .goal of

protecting the health and safety of the generalpublic around a nuclear facility — indeed, anyfacility — an effective emergency publicinfom-ation program must be in place before theemergency occurs.

Media response to a nuclear plant accidentmay also be evaluated in relation to the abovesix Items with perhaps some additions such as"amount of conflict present" and "other newsstories occurring". A key lesson for emergencyplanners as well as utility managers 13 thatmedia Intere3t in a nuclear plant accident ianot necessarily proportional to the severity ofthe accident.

II. REVIEW OF PUBLIC INFORMATION PROBLEMSDOBIMG THE THREE MILE ISLAWD ACCIDEWT

It Is instructive to review the technicalaspects of the accident at the Three Mile Island(TMI) Nuclear Station Unit 2 here because, asGeorge Santayana wrote, "Those who cannotremember the past are condemned to repeat it."

A. Prior to the Accident

From a public information point ofview, the TMI accident probably could havehappened to almost any U.S. nuclear utility.The Utility, and in fact the U.S. nuclearIndustry in general was not prepared for anongoing accident, where time was of theessence. The only previous msjor U.S. nuclearaccident was the fire at Browns Ferry in March1975 which was essentially over by the time thesituation was made public. The Windscaleaccident in Britain during October 1957 had farreaching offsite consequences, but with regard

367

368

to the media and public perception of nuclearpower at that time, It occurred In an entirelydifferent era. TMI was the first of its kind.

Additionally, the Utility's PublicInformation 3taff had, on the average, less thanone year's experience with nuclear power, and noone person on that staff had any technicalbackground. Complicating the situation was thefact that, with rare exception, media personnelalso had little technical background. Virtuallyno one was prepared for a major nuclear powerplant accident because so few believed that onewould happen.

B. During the Accidentf

The Public Information problems during theTMI accident were severe. The overall problemwas that a chain of information flow was notestablished from the plant to the public.Information was not coordinated among the keyplayers — the Utility, the state, the federalgovernment, local officials — and was notpresented to the media and public In a cohesivemanner. Because information flow was notestablished, the Utility's Public Informationstaff was not kept informed by plant technicalpersonnel; the Public Information staff was attimes hours behind the accident sequence. Thus,initially it appeared that the Utility PublicInformation staff consciously "minimized thesignificance of the accident. In actuality,they did not know how far the accident hadprogressed.

At a press conference during the first dayof the accident the Utility failed to say thatthere had been small offsite radiation releasesand that more could be expected. The state'sLieutenant Governor, acting as state spokesman,knew that radiation levels above background hadbeen detected. He asked the Utility'sspokesperson why the Utility had not told thisto the media. The Utility's spokespersonreplied, "Because it did not come up." Fromthat point on, the state was suspicious of theUtility's information and motives, and reliedupon its own staff. At a press conference laterthat day, the Lieutenant Governor toldreporters, "Metropolitan Edison has given youand us conflicting information."

This statement — particularly in the post-Watergate, post-Vietnam era — was a red flag tothe media. Reporters may not have known thedifference between a neutron and a coolingtower, but they knew that "conflictingInformation" made a great story.

What followed for the next week or so waspublic Information chaos. Official news sourceswere confused. Technical Jargon was used,overused and misused. Sources of informationwere everywhere and nowhere. "Experts" appearedand were quoted extensively. A representativeof the U.S. Nuclear Begulatory Commission aotedas the official spokesperson for the federal •government and the Utility stopped talking.However, this did not 3top the "Hydrogen Bubble"story from occurring. That, you may recall, was

when speculation led to perceived Imminentdisaster.

There are several other examples of TMIpublic information problems, but the basicproblems can be summarized as follows:

1. No information flow path wasestablished from the plant technicalstaff to the public information staff

2. There was little coordination ofInformation among the organizationsinvolved

3. Each organization did not have aprincipal, designated spokesperson withaccess to appropriate information in atimely manner

^. There was no 3ingle location from whichthe media and public could findofficial and timely Information

5. Neither the Utility's PublicInformation staff nor the media had thetechnical background sufficient tounderstand or explain the accident.

6. There were no plans or procedures todeal with the Public Informationaspects of a major nuclear plantemergency

C. After the Accident

Studies show no real evidence of theutility or others consciously conspiring to hideInformation. The basic problem was that no onewas prepared for the accident. In theaftermath, however, several regulations weredeveloped to ascertain that utility andgovernmental agency Public Information staffswould be prepared In the future.

III. U.S. MPCLEM EMEHGEWCT PUBLIC IMFOfi-HATIOHREGOLATOBT RBOOIRPCITtS

As a result of TMI, several U.S. governmentregulations were issued. The basic planningdocument for U.S. nuclear utilities Is calledNUREG-0654/FEMA-REP-1 (1980). Other guidance,particularly from the Federal EmergencyManagement Agency, Is being developed.

Although not regulatory, the Institute ofNuclear Power Operations (INPO) has included EPIas one of its emergency preparedness evaluationareas (INPO 85-001, 1985). As in otherevaluation areas INPO developed performancecriteria for EPI programs to aerve as a basisfor evaluation. And as with the NUREGrequirements, the INPO criteria are general.

Thus, EPI regulations are not specific.This lack of specificity has proved to be amixed blessing for the nuclear Industry. On theone hand, It allows utilities the flexibility toprovide quality, comprehensive publicinformation programs that serve well during anemergency as well aa during non-emergencies. Onthe other hand, the lack of specifics alsoallows utilities to meet only the letter of tholaw and go no further. Those utilities whochoose the latter course will cause the nuclear

369

industry problems, for they are not prepared tohandle a major nuclear emergency. They have notlearned from the past.

IV. BECEHT UTILITIES EIPEHIEMCE

There has been much improvement in utilityEPI efforts since TMI. While many agree thatquality programs are necessary, few agree aboutwhat constitutes a quality program. A review ofrecent emergencies provides some clearindications of what utilties may expect shouldan emergency occur.

Within the Ia3t 2 years, events at U.S.nuclear plants generated national mediaattention. None of these events posed a hazardto the public yet they po3ed a serious threat tothe credibility of the utilities. The followingis a brief summary of those events:

A. Event A

At the Vermont Yankee Nuclear Power Plantin June, 1981, a source of radiation wasaccidentally withdrawn from its shielding; thisresulted in abnormally high radiation levels inone area of the plant. An "ALERT" was declaredat about 9:00 a.m. and within four hours theutility received over 250 telephone calls fromthe media, public, and financialorganizations. Those calls were placed to thecorporate headquarters where there were onlyfour public information personnel. Althoughapproximately 20 people were assigned to the EPIorganization, this organization was notactivated. Many people who tried to call couldnot get information because the phone lines werebusy.

Many media reports were inaccurate. Amajor television network reported that there hadbeen an explosion, which was untrue. A numberof reports stated that plant workers wereoverexposed to radiation; this was alsountrue. A post-accident review of media reportsindicated that local media were generallyaccurate in reporting information. Nationalmedia, however, were often inaccurate and tendedto sensationalize the stories.

B. Event B

An "ALERT" was declared at a boiling waterreactor because of an equipment malfunction in acomponent associated with the reactor coolantsystem. The plant was shut down with no hazardto the public. However, the media began totelephone the corporate headquarters soon afterthe "ALERT" was declared. Although the utilitydid not count the number of calls, the volumewas so large that additional personnel wereassigned to answer those calls. One majorproblem was an inability to get information tothe local media in a timely manner. EPIprocedures called for EPI personnel to first

telephone wire services, and then notify mediawith offices close to the plant. However, sinceso many phone calls were received at thecorporate headquarters, the staff did not havetime to call the local media. Thus, many of thelocal media went to the plant for- information.A public information representative was on dutyat the plant, but was assigned to the plantTechnical Support Center. As a result, no oneat the plant was available to talk to the localmedia.

A second problem resulted when localgovernment officials notified industries nearthe plant to be prepared to evacuate. When themedia became aware of this, they becameskeptical of the utility's statements sayingthere was no hazard. The utility was tellingthe truth, but this conflicting informationresulted in a loss of utility credibility andcreated an environment to cause the public to beconcerned about its safety.

C. Event C

The two previous events occurred duringnormal working hours. At Sacramento MunicipalUtility District's Rancho Seco Nuclear PowerStation, a hydrogen explosion occurred in theelectrical generator at night. This event wasclassified as an "UNUSUAL EVENT". This wasfollowed later by an unrelated loss of all non-nuclear instrumentation, which necessitated an"ALERT" declaration for under 10 minutes.Again, as with the other two events, the publicwas not endangered. In a yet further unrelatedevent off site, a pipe bomb was set off at aschool near Sacramento. The explosion was heardthroughout the county, and the media beliovedthat the sound of the explosion came from RanchoSeco. When the media began to call the utilityheadquarters for information, the telephoneswitchboard called public information personnelat their homes. The utility did not activatetheir EPI staff; two public informationpersonnel attempted to answer media inquiriesfrom their homes.

Responding to the explosion rumor, a numberof reporters went to the plant forinformation. The plant had a public informationofficer position, however, no one had beenassigned to fill that position. Thu3, there wasno one for the media to talk to at the plant.

The following day a number of media reportssharply criticized the utility for not providinginformation. The local government held ahearing where officials similarly criticized thecompany public information efforts. One Statecongressman criticized the utility and comparedthe accident to the one at Three Mile Island.There was, of course, no relationship betweenthe two eve-its but the congressman madeheadlines implying there was. The key here isthat the utility's response - not the accident- became the story.

370

D. Event D

Late in January, 1986, Cleveland ElectricIlluminating Company's Perry Nuclear Power PlantUnit 1 was undergoing several test and pre-fuelloading evolutions. Unit 2 was in a deferredconstruction status. Fresh nuclear fuel wasonsite in the Fuel Building; it was expectedthat the low power license would be grantedwithin a month.

At 1148, on Friday, January 31, the PerryNuclear Power Plant, (PNPP) experienced anearthquake measuring approximately 5.0 on theRichter Scale. As a result of the earthquake,the PNPP Emergency Plan wa3 implemented.Emergency response actions went into effect atapproximately 1206 when the emergency wasclassified as a "SITE AREA EMERGENCY". As partof the Site Area Emergency response, all non-essential personnel were evacuated from the sitefor accountability purposes. The situation wasdowngraded to an "ALERT" at 1300 andsubsequently terminated at 1130. Recovery wentinto effect and continued until Saturday at 1100when PNPP and NRC conducted a meeting and agreedon an Action Plan. There were no injuries toworkers. There was also never any risk to thepublic.

The Perry Plant has a Community RelationsSection which handles site media relations; theCleveland Electric Illuminating corporate officehas its own Public Information Department. Thetwo work together during PNPP events.

By 1215, Community Relations Sectionpersonnel had learned of the event andestablished contact with both the Control Roomand Site Security. By 1230, a PublicInformation Emergency Response Team had beenactivated and was answering media calls. AnInformation Liaison had reported to theTechnical Support Center to establish aninformation flow from the TSC to the CommunityRelations staff. Within 3 hours of theearthquake, 2 media statements were issued.That afternoon eight media outlets (2television, 3 radio, 3 newspaper) arrivedonsite. Hundreds of media and industry callswere logged onsite. Over the next two days, twomore media statements were released withresultant inquiries. By Monday, things wereback to "normal".

During the first afternoon, there was somefriction between the site and corporate PublicInformation staffs. Site thought Corporate wastoo demanding, Corporate thought Siteunresponsive. As it turned out, during the siteevacuation the telephone system was secured andall incoming telephone calls went through oneSecurity Officer. Thus several people includingthe media and the state could not get through tothe Community Relations Staff. They in turncalled the corporate offices. This sounds likea typical post-accident critique, but here isthe key point: corporate logged over 2000 phonecalls - this for a plant not yet in operation.

Overall the Perry Public Informationresponse was good. At no time did the PNPPresponse become the story.

E. Event E

At approximately 3:50 a.m. on July 29, 1986a secondary side steam line ruptured near thecondenser of the Rochester Gas and Electric(BG&E) Company's R. E. Ginna Nuclear Station.In accordance with procedures, the incident wasclassified as an "UNUSUAL EVENT" at 1:00 a.m.The unit was manually shutdown and there was norelease of radiation, nor were there anyinjuries. The UNUSUAL EVENT was terminated at5:14 a.m.

RG&E's public information people went intoaction. At 8:30 a.m. that morning RG&E and thetwo area counties held a media briefing.Approximately 15 reporters covered thebriefing. Additionally, RG&E received over onehundred phone calls over two days. Thisincident had all the makings of a major newsstory - morning accident, pipe failure, steamrelease, plant shutdown, same state as Shoreham,same reactor vendor as Seabrook, major accidentthere is 1982, etc - but it was not. Althoughone T.V. station gave RG&E its usual hard time,the other media outlets were fair. This storydid not make national news.

F. Event F

The tragedy at Chernobyl carries severalpublic information lessons. It is clear thatthere are philosophical differences in nuclearaccident response and public Informationpractices between the Soviet Union and theUnited States. Still, western media coveragewas extensive. It is of interest to note thatthe lack of official information about theaccident did not deter the press in gatheringits stories. As seen with American nuclearplant accidents, the media will go around thesystem to find Information. The quality of thatinformation may be suspect at times, but theinformation will be found.

One Chernobyl lesson is that when officialinformation is scant or not available,unofficial information Increases in value andcredibility. Recall the incident where a Dutchamateur radio operator allegedly monitored aRussian transmission telling of hundreds ofcasualties and appealing for help. This storyso far has proved untrue, but It was a majorstory of the accident.

Another lesson concerns informationwithholding or "processing". The Soviets aremasters of information processing, whether it befull disclosure, partial disclosure, nodisclosure, filtering, embellishment oreuphamlsing. But when the efforts of the mediaare concentrated as they were with Chernobyl,the official subterfuges did not work. Yet

371

there are U.S. utility managers today who thinkthey can succeed where the Soviets could not;they prefer not to release information to thepublic or even to their Public Informationpeople. Event after event 3hows that thisphilosophy, although once in a while lucky,harms the utility in the long run.

V. LESSORS TO BE LEARMED

A. What You Should Expect

I have di3CU3sed media response to severalnuclear incidents. There were two UnusualEvents, two Alerts, a Site Area Emergency andTMI and Chernobyl, which clearly would have beenGeneral Emergencies. Of the American plants,two were located in the north, one each in thesouth, west, east and midwest. Despite allthese differences, there are common threads tothe media's response to all of theseaccidents. Based upon these and otherexperiences, this i3 what you should expect.First, some groundrules:

o Media interest in a nuclear plantincident is not necessarilyproportional to the accident'stechnical severity.

o A utility's technical perception of anuclear plant incident is not relatedto nor will it influence the media'sperception of that incident.

o Media interest in nuclear event3 13high.

o The general public has a fear ofradiation - any amount of radiation.Microcuries and millirems mean nothingcompared to cancer.

o The media'3 coverage of nuclear issuesand events is in a class by itself,

o Simplistically, all nuclear powertopics for the general public and mediamay be categorized into two issues:radiation and money.

o When politicians become Involved, themedia interest and perceivedseriousness of the problem doubles,

o Even a minor event can generateconsiderable media interest.

This is what you 3hould expect if you havea nuclear incidents

o There will be much interest in theevent,

o You will receive dozens if not hundredsof phone calls.

o The media will most likely show up atyour site or corporate headquarters orboth. The use of helicopters andsatellite transmission facilitatesthis.

o The media will want a spokesperson; ifyou don't supply one, they will findone.

o If the media are not satisfied withyour pre3s operation, they will goaround your system.

o Inaccuracies will occur; rumors will begenerated.

o Local media will be more accurate thannational media; national media will bevery demanding.

o All media want a steady stream ofinformation. If they don't get it fromyou, they will get it elsewhere.

o The media may not be fair. Reportersare human, they too have pressures;"bad new3 sells"; and the media -particularly TV - want to beat thecompetition. Objectivity is a worthygoal but in nuclear coverage it is notalways attained; some reporters andeditors do have axes to grind.

o If your company is privately owned, thefinancial community will be extremelyinterested in the accident.

o Your company's image will re3t upon howwell your Public Information staffdoes.

o "Low level" events can cause a utilityas many informational problems as amajor accident.

o Media response will be a function ofyour response. The better you do, thebetter they do. If you shoot yourselfin the corporate foot, the media willduly record it; they will not call adoctor. If so, your response - and notthe accident - will become the story.

B. What You Should Do

If you could prevent accidents at yourplants, this conference and your Jobs would notbe necessary. It should be clear as to whatthis means to nuclear emergency planners.First, you will have accidents and incidents.Second, your emergency public information effortshould be top quality. Just like your onsiteorganization, your Public InformationOrganization should be able to transfer fromnormal operation to emergency operationssmoothly and efficiently.

Space limitations prevent a listing here ofkey elements of a successful emergency publicinformation program. However, Just as we canlearn from mistakes, we also should learn fromsuccesses. The Perry Plant successfully handledthe media aspects of their earthquake because ofthe following reasons:

o A detailed emergency public informationplan was in place.

o The Community Relations Sactionestablished a chain of informationflow, outlined responsibilities,coordinated information, and providedspokespersons. The Public Informationstaff controlled the situation.

372

o The Public Information staff not onlypartcipated In exercises, they heldseparate Public Information drills.These proved extremely valuable,

o CEI and PNPP Management support thePublic Information effort. Theydevoted time and resources to PublicInformation - a wise investment whichhas reaped dividends.

A successful and credible emergency publicinformation operation is the culmination of muchplanning, training and coordination of effortsamong those affected: the utility; federal,state and local officials; special interestgroups; the media and the public Within theutility alone, coordination is required amongemergency planning, plant information staff,operations, plant management and corporatemanagement. Nuclear Emergency PublicInformation la much more than being able towrite news releases.

An effective emergency public Informationprogram Is an integral part of a site emergencyoperation as well as being the cornerstone of atotal corporate emergency communicationeffort. Emergencies at nuclear plants haveproven that the media, governmental officials,the financial community, employees, stockholdersand special interest groups will actively seekinformation about emergency activities. Torespond to this interest and to service thecompany's own Interests, a company must developa comprehensive corporate emergencycommunication program. This requires theinvolvement of the total company, not justpublic information personnel.

The challenge to emergency planners is tomake emergency public information programseffective. Each event that occurs reinforcesmore strongly the necessity of a good EmergencyPublic Information program. In the minds of thepublic, the Emergency Public Information effortsare the face, soul, and conscience of theutility. The performance of a utility'sEmergency Public Information program willdetermine how well public safety is protectedand will determine the public's verdict on theutility. In a larger sense, Emergency PublicInformation has and will impact the future ofnuclear power in the United States. This Is oneof Three Mile Island's lessons learned.

3. "Performance Objectives and Criteria forOperating and Near-Term Operating LicensePlants", INPO 85-001, January, 1985,Institute of Nuclear Power Operations,Atlanta, Georgia.

REFERENCES

1. M. K. Lindell and R. W. Perry, February,1983, "Nuclear Power Plant EmergencyWarning: How Would the Public Respond",Nuclear News, American Nuclear Society.

2. "Criteria for Preparation and Evaluation ofRadiological Emergency Response Plans andPreparedness in Support of Nuclear PowerPlants". NUREG-O65VFEMA-HEP-1, Revision1, November, 1980, U. S. Nuclear RegulatoryCommission and U. S. Federal EmergencyManagement Agency. ;

» IA

ANS Topical Meetiag ra Radiological AccMeats—Perspectives aid Eaergeacy Plaaaivg

Factors Influencing Media Coverage of a Radiological IncidentRoy K. Bernhardt and Layton J. O'Neill

ABSTRACT Most organizations have an existingpolicy for interactions with the media. Thispolicy often requires that interactions be withor through a professional group of publicinformation officers or the Office of PublicAffairs. This policy tends to give individualmembers of an organization the belief that theyare not responsible or in some instances, evenallowed to interact with the media. To achievegood media relationships and/or coverage,individual interactions are necessary andrequired. The guidelines for media interactionsprovided in the Federal Emergency ManagementAgency (FEMA) sponsored Radiological EmergencyResponse course are relatively straight forwardand simple to adopt.

The Nevada Operations Office (NV)Radiological Assistance TEAM (RAT) is Team No. 4under the Department of Energy Region VIICoordinating Office. In addition to the normalU.S. Department of Energy (DOE) RAT standbyradiological instrumentation and health physicscapabilities, the NV RAT has a number of uniquefeatures.

Beginning in 1970, NV was given theresponsibility by the Regional CoordinationOffice for the Radiological Assistance Programin the state of Nevada due to the mountainbarrier separating the California teams fromNevada.

From 1972 to the present, the NV has had aformalized agreement with the state of Nevada,which identifies the NV 24-hour emergency phonenumber as the primary phone number to call inthe state of Nevada for radiological assistance.The agreement provides for the NV team toperform the initial response and control at theincident scene until a state representativetakes charge. NV provides a supply of smallbillfold cards (Illustration 1) to the StateOffice of Preparedness for their distributionthroughout the communities of Nevada.Approximately 2,000 have been distributed todate.

A comprehensive NV RAT NotificationProcedure was developed, which calls for routine

NV-1S5 (10*5) FORRADIOLOGICAL

ASSISTANCECONTACT

NEVADA OPERATIONS OFFICE

U. S. DEPARTMENT OF ENERGYLaa Vagaa, Navada24 HOUR TELEPHONE

(702) 295-3343OR — Region 7 Coordinating Office:

San Francisco Operations Office, U.S.O.O.E.Telephone: day or night (415) 273-4237

IT IS ESSENTIAL TO INFORM THE PARTIES CONTACTEDOF THE EMERGENCY NATURE OF THE CALL

The Radiological Assistance Plan it a service provided to tirepublic by the U.S. Department of Energy. Its function is to adviseand assist local authorities in dealing with accidents involvingradioactive materials.Radiological Assistance learns are available to local authoritiesat ad time*. Collect calls WIN be accepted if necessary.For specific instructions regarding initial procedures in thehandling of radioactive incidents refer to:EMERGENCY ACTION GUIDELINES FOR INCIDENTS INVOLV-ING RADIOACTIVE MATERIAL: "This Instruction le found Inthe NV Radiological Aaeletance Team NotmcaUon ProcedureManual, which la Issued by the Nevada Operations Office,USOOE, Laa Vegas, NV."

BILLFOLD CARD

notification of both the State Office ofEmergency Management and the State Division ofHealth, Radiological Section, if a response ismade to a state controlled location.

For an expeditious response, the NVmaintains a RAT Captain Duty Officer roster withbeeper call-out capability. To further enhancethe team captain's response capabilities, theteam captain has a dedicated vehicle having thecustomary radiological equipment along withminor clean-up equipment, such as a HighEfficiency Particulate Air (HEPA) filteredindustrial vacuum cleaner, A/C generator,brooms, shovels, flood lights and air sampling

373

374

capability.The DOE Aerial Measurements Systems (AMS)

is operated by an NV contractor. The NV RAT hasbeen able to rely upon the readily available AMScapabilities. These capabilities include thetransportation of the team IT embers and equipment,plus aerial monitoring of the airborne releases,surface deposition and lost radioactive sources.

In 1985, NV RAT again enhanced it's responsecapability to include the clean-up oftransportation spills of defense low-level wasteshipments, by developing a plan and equipment fora contractor operated Clean-up Team.(Illustration 2). The cornerstone of thiscapability being the specially designed HAZVACtruck. It was the first major exercise of the

equipment used during the exercise.The education of the news media and local

authorities is another of the FEMA guidelinesaddressed in the Radiological Emergency ResponseCourse offered at the Nevada Test Site. By beingpresent at the exercise, the media had anopportunity to view first hand, and thus reporton its use and operation, equipment andmaterials that can be made available to stateand local emergency response organizations toassist in an actual emergency if one shouldoccur. The capabilities of a complex piece ofequipment, the "HAZVAC", was explained to themedia. In the resulting media coverage, onestory notes the equipment could also be used inthe event of toxic spills.

NV RADIOLOGICAL RESPONSE CLEAN-UP TEAMON-SCENE ACTION

Clean-up Team that created the situation forstimulating the news media involvement.

The fact that the news media was invited toview the exercise caused some concern among teammembers and observers. The concern seemed to becaused by the "bad press" usually given toradiation and situations Involving radioactivematerial. The decision to invite the media is inkeeping with the first FEMA guideline concerningpublic relations and media interactions. The newsmedia treatment of an accident and publicreaction to that accicent depend to aconsiderable extent upon what was done before theaccident to educate the public about the specifichazards of the particular operation and whatsafeguards have been provided.

The media made the public more aware of theequipment and trained personnel available in theevent of a transportation accident by reportingthe exercise, the response team's actions and

One of the "lessons learned" fromconducting this exercise was the need for anon-scene media spokesperson with the authorityto issue statements to the media without firstchecking with headquarters. Provisions andguidance for this media representative should beincluded in emergency response plans. Adesignated spokesperson as part of the emergencyresponse team will be in a position to ensurethat information released is provided as quicklyas possible, in the proper perspective, factualand consistent. A trained spokesperson will beable to safeguard classified information, yetstill release facts about the incident. Thesefacts should be released even if they proveembarrassing or admit to errors.

This attitude and commitment will developan understanding and rapport between the mediarepresentatives and the informationspokespersons. However, each individual member

375

of the emergency response organization shouldreceive training to be able to function as themedia interface, should the need arise. A programof periodic exercises to test not only theemergency response team's abilities andequipment, but also the interface and liaisonwith the media is necessary and prudent.

Indexes

Author Affiliations ai J Addresses

Addis, R. P.E. I. du Pont de Nemours and CompanySavannah River LaboratoryAiken, SC 29808

Adler, Martha V.Energy DivisionOak Ridge National LaboratoryBldg. 4500-NP.O. Box XOak Ridge, TN 37831-6190

Adler, VernonFederal Emergency Management AgencyWashington, DC 20472

Athey, G. F.Pacific Northwest LaboratoryP.O. Box 999Richland, WA 99352

Ayer, J. E.U.S. Nuclear Regulatory CommissionWashington, DC 20555

Baggcnstos, MartinSwiss Nuclear Safety InspectorateCH-5303 WUrenlingenSwitzerland

Ballinger, M. Y.Pacific Northwest LaboratoryP.O. Box 999Richland, WA 99352

Bander, T. J.Pacific Northwest LaboratoryP.O. Box 999Richland, WA 99352

Benjamin, Richard W.E. I. du Pont de Nemours and CompanySavannah River LaboratoryAiken, SC 29808

Berger, Mary EllenRadiation Management Consultants5301 Tacony St.BoxD5Philadelphia, PA 19137-2307

Bernhardt, Roy K.Reynolds Electrical & Engineering Co., Inc.Environmental Sciences TrainingP.O. Box 14400Las Vegas, NV 89114

Berry, Hollis A.EG&G Energy Measurements, Inc.Aerial Measurements OperationsP.O. Box 1912Las Vegas, NV 89125

Blackburn, James A.Illinois Department of Nuclear Safety1035 Outer Park DriveSpringfield, IL 62704

Borchert, Harold R.Division of Radiological HealthDepartment of Health301 Centennial Mall SouthP.O. Box 95007Lincoln, NE 68509

Boyns, Philip K.EG&G Energy Measurements, Inc.Aerial Measurements OperationsP.O. Box 1912Las Vegas, NV 89125

Bradley, Mark J.Babcock & Wilcox CompanyNuclear Power DivisionP.O. Box 10935Lynchburg, VA 24506-0935

Burke, Richard P.Sandia National LaboratoriesDivision 334Albuquerque, NM 87185

379

380

Burson, Zolin G.EG&G Energy Measurements, Inc.Aerial Measurements OperationsP.O. Box 1912Las Vegas, NV 89125

Cain, David G.Electric Power Research InstitutePalo Alto, CA 94303

Calabrese, Richard V.Department of Chemical and Nuclear EngineeringUniversity of MarylandCollege Park, MD 20742

Carter, Thomas F.International Energy Associates Limited2600 Virginia Ave., NWSuite 1000Washington, DC 20037

Cheok, Michael C.NUS Corporation910 Clopper Rd.Gaithersburg, MD 20878

Cooley, Michael J.Detroit Edison CompanyRadiological Emergency Response Preparedness6400 N. Dixie HighwayNewport, MI 48166

Cybulskis, P.Battelle Columbus Division505 King AvenueColumbus, OH 43201-2693

Dahlstrom, Thomas S.EG&G Energy Measurements, Inc.Aerial Measurements OperationsP.O. Box 1912Las Vegas, NV 89125

DeBusk, Richard E.Management Systems LaboratoriesVirginia Polytechnic Institute and StateUniversity

Blacksburg, VA 24060

Denning, R. S.Battelle Columbus Division505 King AvenueColumbus, OH 43201-2693

Dickerson, M. H.Lawrence Livermore National LaboratoryP.O. Box 808Livermore, CA 94550

Dinitz, IraU.S. Nuclear Regulatory CommissionWashington, DC 20555

DiSalvo, R.Battelle Columbus Division505 King AvenueColumbus, OH 43201-2693

Doyle, John F., IllEG&G Energy Measurements, Inc.NV Program OfficeP.O. Box 1912Las Vegas, NV 89125

Gant, Kathy S.Energy DivisionOak Ridge National LaboratoryBldg. 4500-NP.O. Box XOak Ridge, TN 37831-6190

Giangi, George J.GPU Nuclear Corporation2574 Interstate DriveHarrisburg, PA 17110

Gladden, Charles A.EG&G Energy Measurements, Inc.Aerial Measurements OperationsP.O. Box 1912Las Vegas, NV 89125

Goldman, Steven B.NUTECH Engineers, Inc.7910 Woodmont Ave.Bethesda, MD 20814

Greenaugh, Kevin C.National Bureau of StandardsCenter for Fire ResearchGaithersburg, MD 20899

Gunter, Alan DaleTechnology Applications, Inc.6621 Southpoint Drive N.Jacksonville, FL 32216

Hardt, Todd L.Babcock & Wilcox CompanyLynchburg Research CenterP.O. Box 11165Lynchburg, VA 24506-1165

Hayes, David W.E. I. du Pont de Nemours and CompanySavannah River LaboratoryAiken, SC 29808

381

Hickey, E. E.Pacific Northwest LaboratoryP.O. Box 999Richland, WA 99352

Hockcrt, John W.International Energy Associates Limited1717 Louisiana, NESuite 202Albuquerque, NM 87110

Hoel, Doris D.E. I. du Pont de Nemours and CompanySavannah River LaboratoryAiken, SC 29808

Hogan, RosemaryEmergency Preparedness BranchOffice of Inspection and EnforcementU.S. Nuclear Regulatory CommissionWashington, DC 20555

Hoover, Raymond A.EG&G Energy Measurements, Inc.Washington Aerial Measurements DepartmentP.O. Box 380Suitland, MD 20746

Hull, Andrew P.Safety & Environmental Protection Div.Brookhaven National LaboratoryUpton, NY 11973

Isaak, Hans-PeterSwiss Nuclear Safety InspectorateCH-5303 WiirenlingenSwitzerland

Jamison, J. D.Pacific Northwest LaboratoryP.O. Box 999Richland, WA 99352

Jaske, Robert T.Federal Emergency Management AgencyWashington, DC 20472

Jobst, Joel E.EG&G Energy Measurements, Inc.Aerial Measurements OperationsLas Vegas, NV 89119

Johnson, Robert W., Jr.The MAPSS Corporation8906 Kingston PikeKnoxville, TN 37923

Jones, Troyce D.Health and Safety Research DivisionOak Ridge National LaboratoryBldg. 4500-SP.O. Box XOak Ridge, TN 37831-6101

Kaiser, Geoffrey D.NUS Corporation910 Clopper Rd.Gaithersburg, MD 20878

Knox, J. B.Lawrence Livermore National LaboratoryP.O. Box 808Livermore, CA 94550

Knox, Wayne H.Knox Consultants4794 Springfield DriveDunwoody, GA 30338

Kurzeja, R. J.E. I. du Pont de Nemours and CompanySavannah River LaboratoryAiken, SC 29808

Linnemann, Roger E., M.D.Radiation Management Consultants5301 Tacony StreetBox D-5Philadelphia, PA 19137-2307

Logsdon, Joe E.Guides and Criteria Branch (ANR-460)Office of Radiation ProgramsEnvironmental Protection AgencyWashington, DC 20460

Long, Robert L.GPU Nuclear Corporation100 Interpace ParkwayParsippany, NJ 07054

Martin, G. F.Pacific Northwest LaboratoryP.O. Box 999Richland, WA 99352

Martin, James A.Reactor Risk BranchOffice of Nuclear Regulatory ResearchU.S. Nuclear Regulatory CommissionWashington, DC 20555

382

Mateik, Dennis E.EG&G Energy Measurements, Inc.Washington Aerial Measurements DepartmentP.O. Box 380Suitland, MD 20746

McKcnna, ThomasIncident Response BranchOffice of Inspection and EnforcementU.S. Nuclear Regulatory CommissionWashington, DC 205S5

McNces, James L.Radiological Health BranchAlabama Department of Public HealthMontgomery, AL 36130

McPhcrson, G. DonaldU.S. Department of EnergyWashington, DC 20545

Meieran, Harvey B.HB Meieran Associates458 S. Dallas Ave.Pittsburgh, PA 15208

Meitzler, WayneBattelle Pacific Northwest LaboratoriesRichland, WA 99352

Miller, Steven R.Attorney-AdvisorU.S. Department of EnergyWashington, DC 20585

Mishima, J.Pacific Northwest LaboratoryP.O. Box 999Richland, WA 99352

Moeller, D. W.Harvard UniversityBoston, MA 02138

Moeller, M. P.Pacific Northwest LaboratoryP.O. Box 999Richland, WA 99352

Morentz, James W.Research Alternatives, Inc.966 Hungerford DriveRockville, MD 20850

Morris, M. D.Engineering Physics and Mathematics DivisionOak Ridge National LaboratoryBIdg. 9207-AP.O. Box YOak Ridge, TN 37831-2009

Nehnevajsa, JiriUniversity Center for Social and Urban

Research and Department of SociologyUniversity of PittsburghPittsburgh, PA 15260

O'Neill, Layton J.U.S. Department of EnergyHealth Physics DivisionP.O. Box 14100Las Vegas, NV 89114

Owczarski, P. C.Pacific Northwest LaboratoryP.O. Box 999Richland, WA 99352

Parker, Michael C.Illinois Department of Nuclear Safety1035 Outer Park DriveSpringfield, IL 62704

Phillips, Trudy E.Babcock & Wilcox CompanyNuclear Power DivisionP.O. Box 10935Lynchburg, VA 24506-0935

Phillipson, D. S.EG&G Energy Measurements, Inc.Aerial Measurements OperationsP.O. Box 1912Las Vegas, NV 89125

Podolak, Edward M., Jr.U.S Nuclear Regulatory CommissionDivision of Emergency Preparedness and

Engineering ResponseOffice of Inspection and EnforcementMS 521 EW/WWashington, DC 20555

Price, Howard A., Jr.The MAPSS Corporation8906 Kingston PikeKnoxville, TN 37923

Ramsdell, J. V.Pacific Northwest LaboratoryP.O. Box 999Richland, WA 99352

Reilly, Susan M.NUTECH Engineers, Inc.7910 Woodmont Ave.Bethesda, MD 20814

Ritchie, Lynn T.Sandia National LaboratoriesDivision 6415Albuquerque, NM 87185

383

Rogers, George O.University Center for Social and Urban

Research and Department of SociologyUniversity of PittsburghPittsburgh, PA 15260

Roush, Marvin L.Department of Chemical and Nuclear EngineeringUniversity of MarylandCollege Park, MD 20742

Sadler, Larry D.The MAPSS Corporation8906 Kingston PikeKnoxville, TN 37923

Sakenas, CherylIncident Response BranchOffice of Inspection and EnforcementU.S. Nuclear Regulatory CommissionWashington, DC 20555

Schcrpelz, R. I.Pacific Northwest LaboratoryP.O. Box 999Richland, WA 99352

Schubert, J. F.E. I. du Pont de Nemours and CompanySavannah River LaboratoryAiken, SC 29808

Seeman, Mary JoU.S. Nuclear Regulatory CommissionWashington, DC 20555

Selby, J. M.Pacific Northwest LaboratoryP.O. Box 999Richland, WA 99352

Sigg, R. A.E. I. du Pont de Nemours and CompanySavannah River LaboratoryAiken, SC 29808

Sjoreen, Andrea L.Oak Ridge National LaboratoryBldg. 4500-NP.O. Box XOak Ridge, TN 37831-6243

Sorensen, John H.Oak Ridge National LaboratoryBldg. 4500-NP.O. Box XOak Ridge, TN 37831-6206

Speck, Samuel W.Muskegum CollegeNew Concord, OH 43762

Stephan, J. G.Pacific Northwest LaboratoryP.O. Box 999Richland, WA 99352

Stratton, William R.Stratton & Associates2 Acoma LaneLos Alamos, NM 87544

Tawil, Jack J.Pacific Northwest LaboratoryP.O. Box 999Richland, WA 99352

Thompson, Karen K.Detroit Edison Company2000 Second AvenueDetroit, Ml 48226

Thorpe, James M.EG&G Energy Measurements, Inc.Aerial Measurements OperationsP.O. Box 1912Las Vegas, NV 89125

Touchton, Robert A.Technology Applications, Inc.6621 Southpoint Drive N.Jacksonville, FL 32216

Ullrich, R. L.Biology DivisionOak Ridge National LaboratoryBldg. 9220P.O. Box YOak Ridge, TN 37831

Vallario, E. J.U.S. Department of EnergyWashington, DC 20585

Walker, J. AndrewManagement Systems LaboratoriesVirginia Polytechnic Institute and State

UniversityBlacksburg, VA 24060

Weber, A. H.E. I. du Pont de Nemours and CompanySavannah River LaboratoryAiken, SC 29808

384

Weiss, Eric W.U.S. Nuclear Regulatory CommissionEvents Analysis BranchDivision of Emergency Preparedness andEngineering Response

Office of Inspection and EnforcementMS MNBB 3302Washington, DC 20555

White, John R.U.S. Nuclear Regulatory Commission, Region 1631 Park AvenueKing of Prussia, PA 19408

Wick, James A.Safety, Health & Environmental Affairs DivisionE. I. du Pont de Nemours and CompanyWilmington, DE 19898

Wilson, R. H.Health Physics DepartmentPacific Northwest LaboratoryP.O. Box 999Richland, WA 99352

Wolff, William F.Office of Nuclear SafetyU.S. Department of EnergyWashington, DC 20545

Young, R. W.Division of Radiation PolicyDefense Nuclear AgencyWashington, DC 20305-1000

Author Index

Addis, R. P.Adler," Martha V.Adler, VernonAthey, G. F.Ayer, J. E.

B

Baggcnstos, MartinBallinger, M. Y.Bander, T. J.Benjamin, Richard W.Berger, Mary EllenBernhardt, Roy K.Berry, Hollis A.Blackburn, James A.Borchert, Harold R.Boyns, Philip K.Bradley, Mark J.Burke, Richard P.Burson, Zolin G.

13529123323763

30763

2371233453731091613991

145253109

Gant, Kathy S.Giangi, George J.Gladden, Charles A.Goldman, Steven B.Greenaugh, Kevin C.Gunter, Alan Dale

H

Hardt, Todd L.Hayes, David W.Hickey, E. E.Hockert, John W.Hoel, Doris D.Hogan, RosemaryHoover, Raymond A.Hull, Andrew P.

I

Isaak, Hans-Peter

291313

101, 223367241247

14514111319312720943149

307

Cain, David G.Calabrese, Richard V.Carter, Thomas F.Cheok, Michael C.Cooley, Michad J.Cybulskis. F.

D

Dahlstrom, Thomas S.DeBusk, Richard E.Denning, R. S.Dickerson, M. H.Dinitz, IraDiSalvo, R.Doyle, John F., Ill

24719719329736367

8531967

21527367

287

Jamison, J. D.Jaske, Robert T.Jobst, Joel E.Johnson, Robert W., Jr.Jones, Troyce D.

Kaiser, Geoffrey D.Knox, J. B.Knox, Wayne H.Kurzaja, R. J.

Linnemann, Roger E.Logsdon, Joe E.Long, Robert L.

11317579

181331

297215149135

34527947

385

386

J*

M

Martin, G. F.Martin, James A.Mateik, Dennis £.McKenna, ThomasMcNecs, James L.McPherson, G. DonaldMeieran, Harvey B.Meitzler, WayneMiller, Steven R.Mishima, J.Moeller, D. W.Moeller, M. P.Morentz, James W.Morris, M. D.

N

Nehnevajsa, Jiri

O

O'Neill, Layton J.Owczarski, P. C.

Parker, Michael C.Phillips, Trudy E.Phillipson, D. S.Podolak, Edward M., Jr.Price, Howard A., Jr.

R

Ranwdell, J. V.Reilly, Susan M.Ritchie, Lynn T.Rogers, George O.Roush, Marvin L.

1137343

73,2091192995

1752716319

113169331

357

37363

1611452233031 O 1181

237187253357197

Sadler, Larry D.Sakenas, CherylScherpelz, R. I.Schubert, J. F.Sceman, Mary JoSelby, J. M.Sigg, R. A.Sjoreen, Andrea L.Sorensen, John H.Speck, Samuel W.Stephan, J. G.Stratton, William R.

TTawil, Jack J.Thompson, Karen K,Thorpe, James M.Touchton, Robert A.

U

Ullrich, R. L.

V

Vallario, E. J.

W

Walker, J. AndrewWeber, A. H.Weiss, Eric W.White, John R.Wick, James A.Wilson, R. H.Wolff, William F.

YYoung, R. W.

1812952371372651913122935151953

259363225247

341

19

31913532514911103291

331

*U.S. GOVERNMENT PRINTING OFFICE: 1987-748-168/40122