Are We Cleaning Up? 10 Superfund CaseStudies

June 1988

NTIS order #PB89-139018

Recommended Citation:U.S. Congress, Office of Technology Assessment, Are We Cleaning Up? 10 Super-fund Case Studies–Specia] Report, OTA-ITE-362 (Washington, DC: U.S. GovernmentPrinting Office, June 1988).

Library of Congress Catalog Card Number 88-600545

For sale by the Superintendent of DocumentsU.S. Government Printing Office, Washington, DC 20402-9325

(order form can be found in the back of this report)

Foreword

In a few years, Superfund became part of the American vocabulary becauseso many people feel so strongly about toxic waste. and cleanup of contaminatedsites. They remain worried about environmental and health effects, but a new con-cern has come to the fore: the enormous amount of money and the long times toclean up an ever-growing list of Superfund sites. Yet, even while the public de-mands effective cleanups, nearly everyone speaking and writing about Superfundseems to feel that serious problems exist. And the focus of public attention hasshifted from how much money ought to go to Superfund to how to achieve environ-mental results and efficiency. Right now there are more questions than answersabout diagnosing and f!xing Superfund.

Four committees of Congress asked the Office of Technology Assessment toassess how Superfund is being implemented under the 1986 Superfund Amend-ments and Reauthorization Act. They asked OTA to examine a number of technicalissues that arise near the beginning of the complex Superfund process. The studywas to assess the impacts of statutory provisions and program policies on environ-mental effectiveness and economic efficiency. The requesting committees were:the House Committee on Public Works and Transportation and its Investigationand Oversight Subcommittee; the House Energy and Commerce Committee andits Oversight and Investigations, and Transportation, Tourism, and Hazardous Ma-terials Subcommittees; the Subcommittee on Superfund and Environmental Over-sight of the Senate Environment and Public Works Committee; and the Subcom-mittee on Environment, Energy, and Natural Resources of the House GovernmentOperations Committee.

During our Superfund Implementation assessment we realized that we couldlearn much by finding out how sites progress through the Superfund program andhow—and when—critical decisions about their cleanup are being made. Before wecould answer tough but general questions about making Superfund work better,we had to know more about what was actually going on. This special report presents10 case studies of recent Superfund decisions at sites which OTA believes, fromsurveying over 100 recent cleanup decisions, to be representative of a broad rangeof contamination problems and cleanup technologies. We hope that everyone af-fected by Superfund can learn as much as we have from these case studies.

Many people have helped OTA with these case studies, especially EnvironmentalProtection Agency staff around the country who provided us with primary infor-mation about the sites. Several companies that are responsible parties at sites alsoprovided key documents. Responsibility for the contents of this document, of course,rests with OTA,

JOHN H. GIBBONSDirector

. . .Ill

Superfund Implementation Advisory Panel

David Marks, Panel ChairmanMassachusetts Institute of Technology

Kirk Brown Robert G. Kisse]]Texas A&M University E.I. du Pent de Nemours & Co., Inc.

Richard Brownell Steve LesterMalcolm Pirnie, Inc. Citizens Clearinghouse for Hazardous

William ChildWaste

Illinois Environmental Protection &encv William Librizzi“

Henry ColeClean Water Action and National

Campaign Against Toxic Hazards

E. William ColglazierWaste Management InstituteUniversity of Tennessee

William B, DeVilleLouisiana Department of Environmental

Quality

Deborah HankinsGeneral Electric Corp.

Bob HarrisENVIRON Corp.

William N. HedemanBeveridge & Diamond, P.C.

New Jersey Dqpt. of EnvironmentalProtection

George MuhlebachCiba-Geigy Corp.

Robert N. OggCH2M Hill

Robert B. PojasekChemCycle Corp.

Frank J. Veale, Jr.Texas Instruments,

James N. WelshThermal Kinetics

Inc.

NOTE: OTA appreciates and is grateful for the valuable assistance and thoughtful critiques provided by the advisorypanel members. The panel does not, however, necessarily approve, disapprove, or endorse this report. OTAassumes full responsibility for the report and the accuracy of its contents,

iv

OTA Project Staff–Are We Cleaning Up: 10 Superfund Case Studies

Lionel S. Johns, Assistant Director, OTAEnergy, Materials, and International Security Division

Audrey B. Buyrn, ManagerIndustry, Technology, and Employment Program

Joel S. Hirschhorn, Project Director

Kirsten U. Oldenburg, Deputy Project Director

David A. Dorau, Analyst

Karen L. Jordan, Research Assistant

Lynn M. Powers, Editor

Administrative Staff

Edna M. Thompson, Administrative Assistant

Diane D. White, Administrative Secretary

v

PageSummary and Analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1The Importance of the Record of Decision . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . zThe Usefulness of Case Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2Superfund’s Better Side. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5Summary of Trends From 10Case Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

IO Case Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16Format for Case Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16Case Study l: Chemical Control Corp., Elizabeth, NJ; EPA Region 2 . . . . . . . . . 18Case Study 2: Compass Industries, Tulsa County, Oklahoma; EPA Region 6 . . 25Case Study S: Conservation Chemical Company, Kansas City, Missouri,

EPA Region 7 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29Case Study 4: Crystal City Airport, Crystal City, Texas, EPA Region 6 ......, 34Case Study5: Industrial Excess Landfill Uniontown, Ohio, EPA Region 5 . . . 41Case Study6: Pristine, Incv Reading, Ohio, EPA Region5 . . . . . . . . . . . . . . . . . 44Case Study7: Renora, Inc; Edison Township, New Jersey, EPA Region 2... . 55Case Study 8:Sand Springs, Petrochemical Complex, Tulsa County,

Oklahoma, EPA Region6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59Case Study 9: Schmalz Dump Site, Harrison, Wisconsin, EPA Region 5...... 66Case Study 10: Tacoma Tar Pits, Tacoma, Washington, EPA Region 10.. . . . . 71

BoxesBox Pagel. How Does Superfund Operate? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32. 10)Case Study Sites With Capsule Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

TableTable Pagel. Times for Sites To Reach Points in the Superfund Process . . . . . . . . . . . . . . . . . 13

SUMMARY AND ANALYSIS

Introduction

Are we cleaning up the mess or messing upthe cleanup? In the eighth year of Superfund,this central question is still being asked. These10 case studies illustrate how the EnvironmentalProtection Agency (EPA) is implementing theSuperfund Amendments and ReauthorizationAct (SARA) of 1986. OTA has examined a greatmany more sites and believes these case studiesare representative of what is happening nation-wide in the Superfund program.

This report examines two fundamental ques-tions about using technology to cleanup toxicwaste sites. First, is the Superfund programconsistently selecting permanently effectivetreatment technologies which, according toSARA, are preferable because they reduce “tox-icity, mobility, or volume” of hazardous wastes?The answer OTA finds is that it is not.

Second, are land disposal and containment,both impermanent technologies, still being fre-quently used? The answer we find is yes. Fu-ture cleanups are likely for the wastes left inthe ground or shipped to landfills.

The Superfund program promised a lot. Peo-ple’s expectations have been high, perhaps toohigh for such a new, complicated, large-scaleeffort. Frustration often makes it difficult to seereal Superfund accomplishments. Since its in-ception at the end of 1980, Superfund has re-ceived a great deal of money, over $5 billionso far, to clean up the Nation’s worst toxic wastesites. But OTA’S research, analysis, and casestudies support the view shared by most ob-servers—including people in affected communi-ties and people in industry paying for cleanups—that Superfund remains largely ineffectiveand inefficient. Technical evidence confirmsthat, all too frequently, Superfund is not work-ing environmentally the way the law directs itto. This finding challenges all those concernedabout human health and the environment to dis-

cover what is wrong and fix it. Whether Super-fund will work cost-effectively over the longterm depends on how cleanup technologies areevaluated, matched to cleanup goals, selected,and implemented and how permanent the clean-ups will be. People want their cleanups—theones they live near or pay for—to last. Improv-ing public confidence in Superfund can be ap-proached from different directions, includingthe one taken in this report: making better de-cisions about cleanup technology.

Too much flexibility and lack of central man-agement control are working against an effec-tive, efficient Superfund program. EPA Regions,contractor companies, and workers have sub-stantial autonomy. In principle, flexibility canlead to benefits. But the case studies show theSuperfund program as a loose assembly of dis-parate working parts; it is a system of dividedresponsibilities and dispersed operations. Thereis no assurance of consistently high qualitystudies, decisions, and field work or of activeinformation transfer. The need for cleanups,the newness of the technological challenge, andthe growth of Superfund mask the inexperienceand mobility of the work force. Program man-agers have not offset inexperience in technicalareas and management with tight managementcontrols and intensive educational programsfor government and contractor workers. Over-simplified “bean counting” of results insteadof evaluations of what those results mean tech-nically and what they accomplish environ-mentally provides too little incentive for qual-ity work. The current decentralized system alsodoes not assure higher levels of program effi-ciency over time, even though some workersand offices may become much more effectiveand efficient.

A widespread belief among Superfund work-ers is that “every site is unique.” There is a ker-nel of truth to this belief. Yet uniqueness hasbeen carried to an extreme and has blocked un-derstanding of common site characteristics,

1

I . common cleanup problems, common solutions,and common experiences with site studies anddecisions. Identifying these commonalities isnecessary to understanding how Superfund isbeing implemented nationally and understand-ing how to improve the program. At the begin-ning, when only a few cleanups were addressed,sites looked very different from each other.Now, with hundreds of cleanups examined, itis easier to see the commonalities and to bene-fit from the experiences to date. The case studiesdiscuss similar experiences at various Super-fund sites and help illustrate the link betweenidentifying commonalities and achieving con-sistent cleanups.

Cleanup costs are major issues in the casestudies. In site cleanup decisions, many peo-ple in government and industry want to keepcosts as low as possible. Hence, there is atradeoff between environmental protectiongoals (How clean is clean?) and the cost of theremedy selected (Is it cost-effective?). There isalso a tradeoff between effective cleanup atsome sites versus no action at others. Thesetradeoffs are getting more difficult as more andmore sites requiring cleanup are identified.SARA’s preference for permanently effectivetreatment technologies—not a requirement thatthey always be used—makes these tradeoffseven harder; it also places more importance onthe accuracy of cost estimates and on evalua-tions of the permanency of different cleanuptechnologies. By understanding the capabilitiesof different cleanup technologies, it is easierto understand how compromises between costand environmental performance can lead ei-ther to “gold plated” or “band-aid” cleanups.

The Importance of the Record of Decision

A crucial step in the complex process of mov-ing a site from discovery to remediation (seebox 1) is the ROD’s technology selection.1Cleanup technology determines whether con-

IEpA hag said ‘gThe Record of Decision ., , is the centerpiece

of the administrative record against which the Agency’s deci-sionmaking maybe judged by the courts.” [U.S. EnvironmentalProtection Agency, “Interim Guidance on Superfund Selectionof Remedy,” Dec. 24, 1986.]

lamination will be eliminated or reduced to asafe level and environmental protection achieved,as well as determining cleanup cost. Technol-ogy selection is the primary focus of this OTAreport. But the ROD decision is not everything.Just as a map is not the territory, a ROD is notthe cleanup. Future analysis of the environ-mental results of cleanups is necessary to seehow the ROD strategic plan is implemented.Because cleanups have been fully implementedat so few sites and the data are so sparse, thisstudy does not fully examine actual cleanup ef-fectiveness and consistency with ROD goals.But the case studies examine the entire historyof the sites. And for some of the sites discussedhere, the technologies selected have failed orearly work to clean up immediate threats hasmade matters worse for final cleanup.

By examining RODS in detail, the function-ing of Superfund comes into focus becauseeverything that was done before the ROD mustbe considered and everything to come latermust be anticipated. Analysis of RODS offersenormous educational value to improve Super-fund implementation because they representthe critical junction between extensive studiesand expensive remedial cleanups. Cleanupcosts vary widely, from several hundred thou-sand dollars to tens of millions of dollars. Toput cleanup costs in perspective, consider thesimple concept of acreage. Data on 15 of thecleanups reviewed in this study indicate thattotal cleanup costs can reach $500,000 to $1million per acre,

The Usefulness of Case Studies

In Superfund, case studies are particularlyimportant because, even after 8 years, cleanuptechnology is a new and fast-changing field andthe work force is relatively young and inexperi-enced. Recent college graduates are often put incharge of multimillion-dollar projects at EPA.These people have had no direct experience andno coursework on cleanup, and they have almostno one to learn from, as turnover is high. Peoplein contractor firms also lack experience. Researchpapers and technical manuals have significantlimitations too. They are quickly outdated, are

3

Box 1.-How Does Superfund Operate?

The Superfund system is complex. Sites are identified and enter an inventory because they mayrequire a cleanup. At this point, or at any time, a site may receive a Removal Action because of emer-gency conditions that require fast action or because the site could get a lot worse before a remedialcleanup could be implemented. (Most of SARA’s requirements for remedial cleanups do not applyto removal actions, even though removal actions can cost several million dollars and resemble a cleanup.)In the pre-remedial process, sites receive a Preliminary Assessment (PA); some then go forward toa Site Inspection (S1), with some of those sites scored by the Hazard Ranking System (HRS). If thescore is high enough, the site is placed on the National Priorities List (NPL) and becomes eligiblefor a remedial cleanup paid for by the government, if necessary, or by responsible parties identifiedas having contributed to creating the uncontrolled toxic waste site. Under current procedures, onlyabout 10 percent of sites which enter the system are likely to be placed on the NPL. Some Stateshave their own lists of sites which require cleanup; these often contain sites not on the NPL.

NPL sites receive a Remedial Investigation and Feasibility Study (RIFS) to define contaminationand environmental problems and to evaluate cleanup alternatives. The public is given an opportunityto comment on the RIFS and EPA’s preferred cleanup alternative. Then, EPA issues a Record of Deci-sion (ROD) which says what remedy the government has chosen and the reasons for doing so; thedecision may be that no cleanup is necessary. A ROD may only deal with part of a site’s cleanupand several RODS may be necessary for a site. The ROD also contains a summary of EPA’s responsesto public comments. EPA chooses the cleanup goals and technology in the ROD. In actual fact a num-ber of actions involving different technologies are likely to be chosen for any but the simplest sites.The ROD is like a contract in which the government makes a commitment to actions which will ren-der the site safe. If responsible parties agree to clean up the site, they sign a negotiated consent decreewith the government; this stipulates the exact details of how the responsible parties will proceed.If the cleanup uses Superfund money, the State must agree to pay 10 percent of the cleanup cost.

In the post-ROD process, the site receives a Remedial Design (RD) study to provide details onhow the chosen remedy will be engineered and constructed. The whole process ends with the Reme-dia]Action (RA), the actual implementation of the selected remedy. Many cleanups include long-termmonitoring to determine whether the cleanup is effective and if more cleanup is necessary. A RODmay be reopened and amended because of new information discovered or difficulties encounteredduring the design and remedial action. When a cleanup is deemed complete and effective, the sitecan be delisted by EPA from the NPL.

too theoretical, assume substantial technical grants. These grants have not been available, how-knowledge, are either too detailed or too general, ever; EPA only began accepting applications inand may be biased to boot. Attending conferences April 1988.where new cleanup technologies are discussedin detail is difficult because of heavy workloadsand limited funds. Moreover, helping to informthe public is also critical, especially becauseSARA increases the participation of communi-ties in the program through technical assistance

The case studies examine the decisionmak-ingprocess, the quality of the information usedin it, and how well the decision and its techni-cal support are communicated by EPA to thepublic. Unlike “bean counting” statistics, whichgive quantitative program results for a largenumber of sites, case studies show how the

2For example, at EPA’s annual research SYmPOShm in MaY complex Superfund system really functions and1988 dealing-with treatment of hazardous waste only nine EPA illustrate the quality-of its environmental per-staff people who may be implementing Superfund (i.e., not in formance. Case studies cannot totally describethe Office of Research and Development) were registered outof a total of over 700 people. the extensive site studies (the RIFSS) which pre-

4

cede the ROD. Nor can they go behind thescenes to investigate all the reasons for deci-sions. But the ROD and its supporting RIFS areintended to stand alone in making the govern-ment’s case for the selected remedy and are theprimary information sources in the 10 casestudies.

This report does not aim to prove whethera technology is good or bad, or whether a deci-sion is unequivocally right or wrong. Cleaningup toxic waste sites is fraught with technicaluncertainties and surprises which cannot beeliminated entirely. The issue of quality ofRODS is not a black or white situations Eachone will have good and bad points. Any cleanuptechnology can be used effectively for some ap-plications, and every complex cleanup decisionhas strong and weak points. There is no prob-lem finding important, correct statements incase study RODS. Indeed, this report often usesstatements from one case study RIFS or RODto illustrate inconsistency or to underscore apoint about a problem in another ROD, Gener-ally speaking, the decisions made in these 10case studies are questionable because, forexample:

If different and readily available technicalinformation had been used, the decisionwould have changed significantly,The range of cleanup alternatives was toonarrow.The analysis was not comprehensive andwas not fair to different technologies.The study work was not internally con-sistent.Mistakes were made in calculations andestimates. .

Critical assumptions were false.Conclusions were stated without analysisand documentation.

3An experienced attorney advises responsible parties: “Legalissues, scientific and technical findings, plus the all-importantpolicy component all affect EPA decisions, Nowhere is this moreclearly shown than in the context of a Superfund Record of De-cision . . . the statute calls on EPA to make decisions based onwhich remedy is cost effective or which ‘adequately’ protectspublic health. Applying these terms entails a degree of subjec-tive judgment,” [P.H. Hailer, Zfazardous Mderids, January/Feb-ruary 1988,]

On a broader scale, other questions are im-portant: Are government policies and EPA’sorganization getting in the way of solid, defen-sible technical work? Is the timing of key piecesof work, such as testing technologies, poor?Looking across sites, are there trends for prob-lems in Superfund technology selection?

The last question is especially important. Itis crucial not to look narrowly at single sitesbut across sites. This is key to central, nationaloversight of Super fund. While individual casestudies can address technical soundness in aspecific ROD, all of them together show howconsistent the program is nationwide in under-standing the advantages and disadvantages ofcleanup technologies and in responding to thestatutory requirements on cleanup technologyselection. As does other information, RODSshow that Superfund is being implemented ina highly decentralized manner. There is incon-sistency in ROD format and presentation of in-formation, examination of cleanup alternatives,and technology selections. In itself, this is notnecessarily bad, but it does mean that centralmanagement oversight and controls by EPA arenecessary to avoid inconsistency leading to con-fusion, unnecessary costs and, for some sites,ineffective cleanup. Lack of consistency amonghundreds and, eventually, thousands of sitesis not an academic issue. Harm to human healthand the environment, loss of public confidencein government, and wasting money are what’sat stake.

The following case studies also show how asite moves through the Superfund system. Gen-eral perceptions about delays are documented.Rarely has so much information been assem-bled on individual sites, possible here becauseEPA has provided OTA with several databases.RODS do not contain such comprehensive in-formation, which itself is an important obser-.vation. On the other hand, there are many areasof interest which are not covered in these casestudies. Documents on a Superfund site can fillfile drawers. There are many legal and proce-dural aspects of Superfund; these case studiesfocus on technical areas and issues. While le-gal and liability issues get enormous attention,environmental protection is the reason for

Superfund and ultimately it is technology whichmust get the cleanup job done.

Superfund’s Better Side

A small fraction of RODS meet SARA’s re-quirements. Six recent well-done RODS arebriefly summarized below. While not perfect,each ROD sets a good remedial action plan,each selects what is likely to be a permanentlyeffective treatment technology, and each pro-vides adequate data and discussion to justifythe technology choice. These six RODS contrastsharply with the 10 case studies which are thefocus of this report.

Cooper Road Dump, Voorhees Township,New Jersey

EPA Region 2; NPL #473/7704–The ROD of 9/30/87decided to take no further action at the site. Adetailed technical case, based on substantial sitesampling, supported the conclusion that pre-vious removal actions at the site had 1eft it per-manently clean. The only question this RODraises is why the site scored so high on the HRSand wound upon the NPL. In hindsight, CooperRoad Dump illustrates a “false positive,” a sitethat went through the Superfund system un-necessarily. Indeed, in a survey of EPA Re-gional staff, this site was included on a list of“sites on NPL that should not be.”s No signifi-cant Federal or State money was spent to provethat no cleanup was necessary; the responsi-ble party paid for the work.

Davis Liquid Waste Site, Smithfield, Rhode Island

EPA Region 1; NPL #216/770; estimated cost, $28 mil-lion.—The ROD of 9/29/87 selected a compre-hensive remedial action plan. The plan included:1) onsite thermal destruction of 25,000 cubicyards of excavated raw waste and contami-nated soil with greater than 2 parts per million(ppm) of volatile organic chemicals; 2) place-ment of incineration ash and pollution control

4Ranking on National Priorities List and total number of rankedsites as of July 1987.

W.S. Environmental Protection Agency, unreleased contrac-tor report written by CH2MHill, November 1986.

residues that are found toxic through testingin an onsite RCRA hazardous waste landfill;3] provision of alternative water for affectedoffsite residents; and 4) restoration of ground-water by onsite treatment using air strippingand carbon adsorption.

The supporting Feasibility Study (FS) was atextbook example of careful analysis, which in-cluded alternative technologies and citationsof experiences at other cleanup sites. Most strik-ing was the early elimination of nontreatmentoptions, such as landfilling the hazardous waste,because, as stated in the FS, they “do not pro-vide for any treatment of contamination. ” Theanalysis also reviewed costs for substantial pi-lot treatability studies during the post-ROD de-sign phase (the RD) as well as acceptable can-cer risk levels as cleanup goals. However, a Iin 100,000 cancer risk level was used ratherthan the 1 in 1 million level more frequentlyused. Another, and probably related, reasonwhy this ROD is not perfect is that some un-treated hazardous material will be landfilledonsite instead of being treated. The higher risklevel seems to have been a.compromise madeto reduce cleanup costs. Also, the delay of thetreatability testing until after the ROD is un-desirable; although for this site there was moreinformation available to justify the technologyselection than in some of the case studies.

The Davis remedial plan used an excellentinterpretation of cost-effectiveness for makingtechnology choices: “an alternative which hasa similar public health and environmental ben-efit to other alternatives can be screened outdue to costs that are higher in order(s)-of-mag-nitude, ‘e

Love Canal, City of Niagara Fails, New York

EPA Region 2; NPL#142/770; estimated cost, about $30mifllon.—The ROD of 10/26/87 altered an earlier

Wompare this to EPA’s guidance which lacks the concept ofcomparable environmental protection: “[cost-effectiveness] re-quires ensuring that the results of a particular alternative can-not be achieved by less costly methods. This implies that for anyspecific site there may be more than one cost-effective remedy,with each remedy varying in its environmental and public healthresults.” [U.S. Environmental Protection Agency, “Interim Guid-ance on Superfund Selection of Remedy,” Dec. 24, 1986.]

6

decision at Love Canal to use onsite land dis-posal for dioxin contaminated sewer and creeksediments. Now, a mobile thermal destructionunit will be used onsite to destroy and removedioxin with an efficiency of 99.9999 percent.The cost for treatment will be twice that forland disposal, but the ROD selected thermal de-struction on the basis of its ability to meet stat-utory requirements by eliminating toxicity andmobility. In addition, several site demonstra-tions elsewhere had successfully destroyeddioxin-contaminated soil with mobile thermaldestruction units. EPA responded to extensivecommunity comments against landfilling thecontaminated material onsite and also decidednot to attempt to separate materials with lessthan 1 part per billion dioxin (EPA’s cutoff foracceptable contamination) because of uncer-tain reliability in doing so.

Operating Industries, Inc., Monterey Park,California

EPA Region 9; NPL #71/770; estimated cost: $4.8 mil-lion.—The ROD of 11/16/87 concerned an in-terim remedial action required to manage con-taminated Ieachate at the site, which had a long,complex cleanup history. The ROD selected anonsite Ieachate treatment system with severalproven technical steps that can reduce a diverseset of organic and inorganic contaminants tolevels low enough to permit discharge to a lo-cal water treatment plant. The key steps willbe gravity separation, coagulant addition, dis-solved air flotation, filtration, air stripping withvapor phase carbon adsorption, and liquidphase granular activated carbon adsorption.

The analysis of alternatives was first rate.Two constraints were applied that ruled outmore innovative approaches. First, the actionhad to be implemented easily and rapidly. Sec-ond, it had to be able to cope with major fluc-tuations in the composition of the leachate.Thus, some technologies that would actuallydestroy organic contaminants, such as plasmaarc thermal destruction and wet air oxidation,both followed by stabilization of solid residuescontaining toxic metals, were not consideredbecause they would probably face delays be-cause of State regulatory requirements and pos-

sibly public concerns. The disadvantage of theselected remedy is that the technologies usedrely almost entirely on separation. Therefore,significant amounts of concentrated hazardousresidues will have to be moved offsite for dis-posal or treatment.

There was some laboratory testing of siteleachate during the FS. Also, the process lead-ing up to the ROD was rigorous, including anextended public comment period with an un-usual opportunity for local citizens to reviewa draft ROD. (Normally, the public gets a verybrief statement of EPA’s preferred remedy toreview.) Although there was keen communityinterest, little of it dealt with the selection oftechnology, but rather with the specific loca-tion on which the leachate treatment facilitywould be built.

he-Solve, Inc., North Dartmouth, Massachusetts

EPA Region 1; NPL#206/770; Mimated cost, $19.9 mil-lion.—The ROD issued on 9/24/87 is one of themost technically detailed and complete RODSreviewed for this study. A previous cleanupbased on an earlier ROD was stopped when fouradditional hot spots of contamination werefound. The newly selected remedy consistedof: 1) the source control phase of onsite treat-ment of 25,500 cubic yards of excavated PCBcontaminated soils and sediments in a mobiledechlorination facility (volatile organic com-pounds will also be reduced); and 2) aquifer res-toration by pumping, repeated flushing, andtreatment involving air stripping and carbonadsorption, particularly for volatile organiccompounds. The site will be evaluated everyfive years because some hazardous substanceswill remain there; curiously, there are no landuse restrictions.

While dechlorination was considered an inno-vative technology, its selection was based onpositive pilot test results on an actual Super-fund site with similar contamination and cli-matic conditions.r (Other work by EPA showsthe approach effective in getting residual levels

The technology is sold by six vendors according to U.S. Envi-ronmental Protection Agency, “A Compendium of TechnologiesUsed In The Treatment of Hazardous Wastes,” September 1987,

7

of PCBS in soils down below 1 ppm.)a Additionalpilot study results will be obtained onsite priorto use, and if dechlorination is unsuccessful,the ROD specified that onsite incineration willbe used instead. Similar treatability and pilottests will be performed for the groundwatercleanup phase prior to full-scale use,

Cleanup goals at Re-Solve were based on riskanalysis on the basis of possible residential useof the site. A 1 in 100,000 excess (over back-ground) cancer risk level was chosen for thesoil and groundwater cleanup instead of themore common 1 in 1 million level. Accordingly,PCBS in the soil will be reduced to 25 ppm,which is a higher concentration than goals setat other sites.e For example, 20 ppm was cho-sen at the Ottari and Goss/Great Lakes Con-tainer Corp. site in New Hampshire; 5 ppm,at the Renora site in New Jersey; 1 ppm, at theTacoma Tar Pits site in Washington; and 1 ppm,at the Liquid Disposal site in Michigan (wherea 1 in 1 million risk was used). A recent EPAdocument refers to cleanup to “the desiredbackground levels (1 to 5ppm) or less.”lo In addi-tion, an assessment by EPA’s Office of Healthand Environmental Assessment concluded thata range from 1 to 6 ppm PCBS in soil is equiva-lent to 1 in 100,000 cancer risk.11 The Re-SolveROD, therefore, illustrates the compromise be-tween level of cleanup and acceptance of costby the government and responsible parties. TheFS noted that “the voIume of PCB contaminatedsoils increases exponentially as the cleanuplevels become more protective.” While the fi-nal decision may be disputed by some people,particularly on the issue of residual PCB level,

~A. Kernel et al,, “Field Experience With the KPEG Reagent,”paper presented at EPA’s Fourteenth Annual Research Sympo-sium,” May 1988.

Whe PCB concentration level corresponding to the 1 in 100,000risk level is 30 ppm, but EPA decided that the uncertainty ofthe approach allowed them to use 25 ppm as being representa-tive of that risk level. The PCB level for the 1 in 1 million risklevel was 3 ppm. Also, it was estimated that onsite groundwatermay contain 10 to 1S ppb PCB after cleanup, which is far in ex-cess of 0.08 ppb, the health-based cleanup level for a 1 in 100,000cancer risk for PCBS,

IOU. S. Environmental Protection Agency, “Report on Decon-tamination of PCB-Bearing Sediments,” January 1988,

llAs reported by EpA in its ROD for the Liquid Disposal Sitein Michigan, Sept. 30, 1987.

the decisionmaking process is clear and thereis public accountability.

Seymour Recycling Corp., Seymour, indiana

EPA Region 5; NPL #57/770; ostimateid cost, $18 mil-lion.—The ROD issued on 9/30/87 was the sec-ond one for the site. The selected remedy hasseveral key components: 1) a full-scale vaporextraction system to reduce the substantial pres-ence of volatile organic compounds; 2) the ex-traction and treatment of contaminated ground-water at and beyond the site boundaries; 3) theapplication of nutrients to remaining contami-nated soil to stimulate biodegradation; 4) theinstallation of a multimedia cap to restrict di-rect contact and limit water intrusion; 5) deedand access restrictions; and 6) a detailed mon-itoring program and technical criteria to de-tect failure and to plan future action if nec-essary.

A good technical analysis supported the selec-tion of this remedy over alternatives such asincineration and in situ soil washing. Inciner-ation would have cost $37 million and in situsoil washing would have cost $17 million, whilethe chosen plan will cost $18 million. But tech-nical impediments—the large size of the site (14acres), the large quantity of contaminated ma-terials (about 100,000 cubic yards), and thedangers of excavating soil with large amountsof volatile compounds—not cost, were the rea-sons for rejecting alternatives that may haveprovided more substantial treatment and de-toxification. In addition, the groundwater treat-ment is estimated to take from 28 to 42 years,but there is no faster alternative available. Ofsome concern is that treatability studies werenot done before the ROD. But the extractiontechnology is well proven and the final Sey-mour implementation plan is well thought out.

Summary of Trends From 10 Case Studies

As a rule, RODS are fraught with problems.The 10 case studies, chosen out of over 100RODS reviewed, illustrate in concrete wayssome disturbing trends among these problems—trends that compromise the ultimate protec-tion of human health and the environment (see

.

8

box 2 for capsule findings). These trends aresummarized below.

Evaluation and Selection of PermanentTreatment Technologies

Many good, permanently effective wastetreatment technologies are on the market but,too often, are not fully examined, or are notselected for use. A ROD may simply opt not totreat a site at all but rather to bury waste in alandfill or to cap the hazardous area, both im-permanent options. A site’s having too little ortoo much contaminated material is often citedas a reason for not choosing a permanent treat-ment technology. Too little material and toomuch material both mean high cost for treat-ment relative to costs for nontreatment alter-natives, but cost alone should not guide de-cisions.

Describing a cleanup technology as a 6’treat-ment” can be misleading. SARA sees a treat-ment as a technology “that, in whole or in part,will result in a permanent and significant de-crease in the toxicity, mobility, or volume” ofhazardous materials “to the maximum extentpracticable,” but SARA’s “treatment” allowsmuch interpretation. Furthermore, EPA has notestablished a hierarchy of preferred results andtypes of treatment.

Not all treatments accomplish the same things.For example, thermal destruction and some bio-logical and chemical treatment can irreversiblydestroy or detoxify nearly all of some toxic sub-stances and therefore reduce their mobility andvolume. But a number of physical and chemicaltreatments can separate organic and inorganicmaterials and release the hazardous materialcollected and concentrated to the environment(e.g., air stripping) or place it in a landfill (e.g.,carbon adsorption, precipitation, soil washing,solvent extraction). The preferred use of sepa-ration technology uses treatment to destroy thehazardous material collected.

Chemical fixation, stabilization, and solidifi-cation treatments usually only reduce mobil-ity, particularly for toxic metals, (but usuallyincrease volume) and they nearly always leavesome uncertainty about long-term effectiveness

because laboratory tests can neither fully dupli-cate field conditions over long periods norestablish what actually is happening to the con-taminants. 12 EPA has said that “There is, atpresent, no set protocol for evaluating the ef-ficacy of stabilization technologies.”13 The useof stabilization technologies for high levels oforganic contamination is particularly unproven.14

A recent EPA review of stabilization technol-ogy said:

Although S/S [solidification/stabilization]technologies have been used for more than 20years, there exists little information on long-term physical durability and chemical stabil-ity of the S/S mass when placed in the ground.., . Generally, S/S technology is recognizedeffective for inorganic waste, while organicwastes have the potential to cause problems. . . . The long term effects of organics on S/Sperformance are important, however, little re-search has been performed. . . . the capabilityof the technology to perform satisfactorily overlong periods of time has yet to be determined... , uncontrolled air emissions are a poten-tial problem to workers and the environment.15

These EPA views are inconsistent with currentEPA decisions that choose stabilization and callthem permanent remedies.

l~he attractiveness of stabilization type technologies is oftenedexpressed in noncost terms, such as: “Long term effectivenessof incineration, stabilization, and solidification are comparable.”IARCO Petroleum Products Co., “Critique of Sand Springs Oper-able Unit Feasibility Study,” Aug. 31, 1987,)

WLO Weitzman, L,EO Hamel, and E. Barth, “Evaluation of Solid.ification/Stabilization As A Best Demonstrated Available Tech-nology,” paper presented at EPA’s Fourteenth Annual ResearchSymposium, May 1908.

ItFor example, a recent EPA study found “large losses of or-ganics during the mixing process” [L. Weitzman et al,, op. cit.].Another EPA study showed that stabilization was not competi-tive with thermal and chemical treatment technologies and soilwashing for organic contamination [R,C. Thurnau and M.P. Es-posito, “TCLP As A Measure of Treatment Effectiveness: Re-sults of TCLP Work Completed on Different Treatment Tech-nologies for CERCLA Soils,” paper presented at EPA’s FourteenthAnnual Research Symposium, May 1988]. A demonstration ofa stabilization technology under EPA auspices concluded that“for theorganics, the leachateconcentrations were approximatelyequal for the treated and untreated soils” [P.R. de Percin andS. Sawyer, “SITE Demonstration of Hazcon Solidification/Stabili-zation Process,” paper presented at EPA’s Fourteenth AnnualResearch Symposium, May 1988].

IsC,C. Wiles and H,K. Howard, “U.S. EPA Research in Solidifi-cation/Stabilization of Waste Material,” paper presented at EPA’sFourteenth Annual Research Symposium, May 1988.

9

Box 2.-10 Case Study Sites With Capsule FindingsCase Study 1Chemical Control Corp., Elizabeth, New JerseyEPA Region 2; NPL rank: 223 out of 770Estimated cost: $7,4 million

Unproven solidification (chemical fixation) technologywas selected to treat in situ highly contaminated subsur-face soil, which previous removal actions had left belowthe water table and covered up with gravel. No treatabilitystudy was used. The cost of incineration was overestimated.The cleanup will leave untreated contamination onsite.

Case Study 2Compass Industries, Tulsa County, OklahomaEPA Region 6; NPL rank: 483/770Estimated cost: $12 million

Capping (containment) of waste was chosen over inciner-ation. Capping was called a cost-effective, permanent clean-up even though it does not provide permanent protectioncomparable to incineration. No commitment was made totreat contaminated groundwater.

Case Study 3Conservation Chemical Co., Kansas City, MissouriEPA Region 7; NPL rank pendingEstimated cost: $21 million

Capping of the site and a hydraulic containment systemto pump and treat some contaminated groundwater werechosen over excavating and treating contaminated soil andburied wastes, which was recommended in an EPA studyand by the State. Water treatment cannot remove all thediverse contaminants at the site. The ROD said that no esti-mate could be made for the duration of the cleanup.

Case Study 4Crystal City Airport, Crystal City, TexasEPA Region 6; NPL #639/770Estimated cost: $1.6 mijlion

Excavation of contaminated soils and wastes (which wereburied in a previous removal action) and their disposal inan unlined landfill with a cap over it were selected overincineration. No treatability study supported the conclu-sion that the selected remedy is permanent on the basisof the adsorption of diverse contaminants to site soil. Ma-jor failure modes for the landfill were not examined.

Case Study 5Industrial Excess Landfill, Uniontown, OhioEPA Region 5; NPL #164/770Estimated cost: $2 million

Providing alternate water to houses that have or are likelyto have contaminated wells was a satisfactory interim re-medial action. However, actions to address the source ofcontamination and to stop and treat contaminated ground-water are long overdue.

Case Study 6Pristine, Inc., Reading, OhioEPA Region 5; NPL #531/770Estimated cost: $22 million

In situ vitrification was developed originally for radio-active soils, but its use for chemically contaminated sitesis still unproven. In situ vitrification was selected—withouttreatability test results-chiefiy because its estimated costwas about half that of onsite incineration. But the estimatedcost for incineration is probably high by a factor of 2. in-cineration offers more certainty and probably would costno more than the chosen remedy. Groundwater will bepumped and treated by air stripping and carbon adsorption.Case Study 7Renora, Inc., Edison Township, New JerseyEPA Region 2; NPL #3781770Estimated cost: $1.4 million

The selected remedy makes use of offsite landfilling forsoils contaminated with PCBS. Also, biological treatmentwas selected for soils contaminated with diverse organiccompounds and toxic metals and for contaminated ground-water, but no treatability study supported its selection.Case Study 8Sand Springs Petrochemical ComplexTulsa County, OklahomaEPA Region 6; NPL #761/770Estimated cost: $45 million

EPA originally said that solidification technology wasineffective for the high organic content wastes and thaton site incineration was effective. EPA then reversed itselfand selected solidification for most of the cleanup, whichthe responsible party had claimed effective based on itstreatability study. Incineration is to he used if solidifica-tion technology is not successfully demonstrated or failsafter solidified material is landfilled on the floodplain site,but criteria for failure are unspecified.Case Study 9Schmalz Dump Site, Harrison, WisconsinEPA Region S; NPL #190/770Estimated cost: $800,000

A simple compacted earth cover over the soil contami-nated with lead and chromium was selected. Solidifica-tion/stabilization treatment was rejected, although this wasa textbook example of appropriate use of the technology.Voluntary well abandonment and monitoring was chosenover pumping and treating contaminated groundwater.

NO treatabilty study results supported the selection ofchemical stabilization. Significant amounts of untreatedcontaminants as well as the treated materials will be leftonsite. The effectiveness of the treatment is uncertain. In-cineration was said to offer no better protection and wasreiected because of its higher cost.

“

10

*

Moreover, a cleanup may consist of manydifferent operations in which treatment maybe only a small part. Removal actions may sendhazardous waste to landfills, perhaps muchmore than may be treated subsequently. Or ac-tion may be taken on contaminated soil but noton contaminated groundwater or vice versa.Too many RODS assume that any use of anytechnology is a treatment that meets the letterand spirit of the statutory requirement. Gen-eral Superfund statistics on treatment can bemisleading because they do not distinguishamong different technologies used at a site fordifferent amounts of material.

There is no clear line between sufficient andinsufficient technical and economic data for se-lecting among cleanup technologies. A RODmay choose an unproven or inappropriate tech-nology or both with the claim that it is a per-manent remedy, or a ROD may eliminate a tech-nology because it remains untried on a largescale. It is not uncommon to have a multimil-lion-dollar cleanup decision made without anytechnical data to support it, either from the tech-nical literature or from tests done on site ma-terial.

Information used to compare treatment tech-nologies is often inaccurate and incomplete.Poor information compromises the RIFS, theselection of remedy, and public support of cer-tain remedies. Alternative treatment technol-ogies that are practical are sometimes ignoredor not chosen. Costs for innovative technologiesmay be unreliable, either too low or too high.Good or bad experiences at other sites are notstudied. An example is the failure, discoveredin 1985, of chemical stabilization treatment atthe Conservation Chemical Co. site after onlya few years of use; nevertheless, RODS are se-lecting chemical stabilization for similar prob-lems more than ever before.

Contractors may quote a wide range for di-rect costs per unit of material treated for anygiven treatment technology. For example, quotedunit costs of onsite incineration ranged froma low of $186 per cubic yard for Seymour Recy-cling to $730 per cubic yard at Pristine for thesame amount of treated material; both sites arein the same EPA Region. 2 Le unit cost quoted

for mobile, onsite incineration in the Chemi-cal Control case in New Jersey and at the Pris-tine case in Ohio (where the technology wasrejected) was twice the unit cost used at theDavis Liquid Waste site in Massachusetts (wherethe technology was selected). At the ChemicalControl site, both $500 and $750 per cubic yardunit costs were quoted for two cleanup alter-natives using the same onsite incineration. Inboth cases, the material burned was essentiallythe same and the type of incineration technol-ogy was the same (the difference in the optionswas where the residuals were disposed).

Such variations make it hard to establish atechnology’s cost-effectiveness—or lack of it—relative to other technologies. Even when a con-tractor uses the same burden rate (see below)among ROD cleanup alternatives, inaccurateunit costs can distort the comparative analy-sis. For example, with Pristine, if direct costhad been $186 per cubic yard instead of $730(with the same 83 percent burden), the total costfor incineration would have been $15 million,not $51 million; Pristine had rejected inciner-ation and selected in situ vitrification for $22million. If total estimated costs have any effecton post-ROD activities, then actual cleanupcosts for clients—and profits to contractors—may vary substantially and some may be muchgreater than they could be.

Contractors estimate cleanup costs by addingto direct costs substantially different levels ofindirect cost (burden or markup). In the Pris-tine case, the burden—various contingencies, con-struction services, and design costs—amountedto 83 percent of direct costs, while for Davisand Re-Solve, involving the same RIFS contrac-tor, the burden was 35 percent; the Davis andRe-Solve indirect costs explicitly included pi-lot study work, while the costs for Pristine didnot. For Seymour Recycling, the burden was60 percent; for Chemical Control, 56 percent;and for Crystal City, 29 percent. The range inburden rates over different sites and across andwithin contractors illustrates an importantmanagement problem in Superfund.

RODS cannot always depend on the resultsof tests done for other sites. Treatability studiesrefer to tests on site material and are supposed

- J

.

11

to bridge the gap between general informationabout the technology and the more specific in-formation needed for technology selection inthe ROD. Results of treatability studies on onesite, particularly for innovative technologies,do not necessarily mean that a given treatmentwill work or not work for some other waste site,unless the conditions are nearly identical or thetechnology’s performance is not waste specific.The problem is that some technologies are verywaste specific, and it is impossible to accuratelyextrapolate positive test results from one wasteto another, especially because Superfund sitesoften have very complex, site-specifc wastes.Incineration of organic contaminants is non-specific, whereas biological treatment is quitewaste specific. Onsite treatment technologies(in which the waste is brought to the technol-ogy) perform more predictably than in situ tech-nologies (in which the technology is broughtto the waste) because the latter’s effectivenessdepends on site conditions, such as chemical,physical, and biological properties of the soil.These can vary widely from site to site.

When they are done, most treatability studiesare not done early enough. It is critical that theybe done during the RIFS before the ROD, butmost are done during the design phase after Ithe ROD. Treatability studies will improve theRIFS by providing technical data to support theROD’s analysis of cleanup alternatives and toensure that the ROD’S cleanup choice is effec-.tive and satisfies statutory requirements. How-ever, EPA now often speeds up RODS, appar-ently to meet fiscal year goals; thus treatabilitytests during the RIFS are sacrificed. This sac-rifice can backfire. Negative test results afterthe ROD would indicate the wrong technologychoice and the waste of a lot of time and money.Worse, altering a ROD at this point, even forgood reasons, may meet some resistance. Fi-nally, when responsible parties or technologycompanies conduct these tests, EPA may needto assure their objectivity by independentlyverifying the results.

Some RODS choose technologies that are inEPA’s Superfund Innovative Technology Evalu-ation (SITE) program, an indication that a tech-nology has not yet been proven. For example,

the Chemical Control ROD chose a new typeof in situ stabilization, the Pristine ROD chosein situ vitrification, and the Sand Springs ROD

&chose a stabilization technique in the S E pro-gram. If, as EPA says, the SITE program existsto obtain “sound engineering and cost data”and to “resolve issues standing in the way ofactual full-scale application, ” then how cansuch ROD selections be justified? If they arejustifiable, are the SITE demonstrations reallynecessary?

The chemical character and complexity of sitecontaminants and how they affect the use ofsome technologies do not get enough attention.A few indicator compounds, used to representall site contaminants for risk assessment, maybe inappropriate for technology evaluation be-cause physical and chemical properties maydiffer from the way health effects vary. The re-sult can be a poor technology choice. Also, sitesampling may be insufficient to detect hot spotsof contamination that would facilitate usinglimited treatment to cut cleanup costs. In ad-dition, groundwater monitoring may not bereliable.

Impermanent Technologies

When wastes are left in the ground or ingroundwater or are redisposed in a landfill, aROD may claim that the remedy is permanentwhen, in fact, it is not. Permanence may beclaimed even when technical factors suggesta high probability of failure, that is, of releaseof hazardous substances, and of another cleanup.In such cases, the ROD would be more credi-ble if it acknowledged the remedy as imperma-nent and defended it on its own merits relativeto truly permanent ahernatives. Moreover, animpermanent remedy and a false sense of secu-rity could lead! for example, to land use thatwould only complicate future cleanup and poseunacceptable risks.

Contrary to the law, containment/land dis-posal decisions seldom analyze the risk of fu-ture failure, damages, and further cleanup.While some RODS claim that containment/landdisposal techniques are proven and reliabletechnologies with no implementation problems,

.

.

.

..

.....

...

.

...

.

...

12

there is evidence to the contrary. For example,the RCRA clay cap being installed at the Win-throp Landfill Superfund site in Maine failedin September 1987 before its construction wascompleted. The ROD of November 1985 saidthe technology was proven, routinely used, andposed no construction difficulties. There hadbeen no analysis of potential failure; under theoriginal Superfund statute-the ComprehensiveEmergency Response and Liability Act of 1980—such analysis was not required. Under SARAit now is.

Sometimes a ROD does not commit to a defi-nite outcome even though it appears to haveselected a technology. Contingencies, uncer-tainties, and multiple future options do not as-sure the public that there will be a permanentremedy and that it will be fully implementedin a timely and effective way. Often, the RODdoes not provide specific technical criteria forsubsequent decisions, such as for groundwatercleanup or land use, nor are there necessarilyassurances of independent validation of dataand effective EPA oversight of activities by re-sponsible parties and contractors. Specificgroundwater monitoring requirements are par-ticularly important because recent EPA re-search has found that “low sampling frequencycoupled with the generally smaller samplingnetworks suggest that efforts to characterizegroundwater contamination at [Superfund]sites may be inadequate.

Impermanent remedies, which provide lessprotection than permanent ones and do not as-suredly meet cleanup goals, are often selectedpurely because they are cheaper in the short run;in the long run they are very likely to be moreexpensive. Regarding cost-effectiveness, whentwo or more cleanup options offer the samelevel of environmental protection and can meetestablished cleanup goals (from risk assessmentor existing regulatory standards), everyone willagree that the lowest cost option should be cho-sen. Impermanent technologies are not cost-

l@R.H. Plumb, Jr., “A Comparison of Ground Water Monitor-ing Date From CERCLA and RCRA Sites,” Ground Water Mon-itoring Research, fall 1987, pp. 94-100.

effective remedies and do not satisfy SARA,therefore, when permanent technologies arepractical. The average estimated cost of thecleanups in the six good RODS noted earlierwas $20 million. In contrast, the average esti-mated cost of not-so-good cleanups in the 10case studies below was $12 million. (In the 10case studies, the average for the five treatmentremedies is $16 million and the average for thenontreatment remedies is $7.5 million.) It is truethat a permanent cleanup based on treatmenttechnology is likely to require a larger initialoutlay than an impermanent cleanup based onland disposal. Even a modest cost differencecan mean a lot added up over thousands of sites.

EPA is less responsive to community concernsabout a remedy being impermanent than to in-terests which favor a lower cost impermanentremedy. Thus community concerns about im-permanence are not very likely to lead to a moreexpensive cleanup technology. There are manyincentives for various parties to keep cleanupcosts low by using onsite containment/land dis-posal or even some relatively inexpensive formsof treatment, such as stabilization and separa-tion technologies. These parties include poten-tially responsible parties (PRPs) that may haveto pay for the cleanup, States that have to pro-vide 10 percent of the cost (unless PRPs pay),and EPA which wants to distribute availablefunds as broadly as possible and which wantsto obtain settlement agreements with PRPs toreduce calls on Superfund money.

In selecting cheap, impermanent remedies,claims of comparable estimated costs may hidethe truth that low cost was the key deciding fac-tor. Getting accurate costs to compare cleanupalternatives is crucial. Overestimates or under-estimates may be used to justify a choice or arejection. For example, at the ConservationChemical Co. site in Missouri, where a settle-ment with PRPs was involved, an EPA contrac-tor and the State recommended one remedy (re-jected) which was said to cost $24 million overanother remedy (selected) which cost $21 mil-lion. But available EPA data suggest that therejected remedy would actually cost from $40million to $150 million.

13—

Program Efficiency

EPA pushes most RODS to completion by theend of the fiscal year and this kind of bureau-cratic pressure can lead to poor cleanup deci-sions. To meet deadlines, EPA may reempha-size public comments that would otherwise leadto reevaluation of facts and technologies; EPAmay make a hasty, technically unsupported de-cision as it did at the Sand Springs site in Okla-homa. Typically, there is less than one monthbetween the end of the public comment periodand the issuance of the ROD. (See table 1 forsummary data from the 10 case studies on timesto reach certain stages in the cleanup process.)The RIFS may also suffer from hurried reviewby EPA because of pressure to issue a ROD bythe end of the fiscal year or quarter.

The pm-remedial process has received littleattention even though sites can be releasing haz-ardous substances into the environment and,during the time they are unexamined and un-attended, get worse. The time from site iden-tification through placement on the NPL isabout 3 years for the case studies (and oftenmuch longer for other sites examined by OTA).

The time between a site’s placement on theNPL and the start of the RIFS varies greatly,averaging about 16 months. Nationwide, there

is no apparent relationship to the site’s HRSscore; a high score does not necessarily speedcleanup (e.g., three sites with similar high HRSscores waited 39, 15, and 3 months). For siteswithin an EPA Region, however, the HRS scoredoes seem to matter; this time the waitingperiod decreased with decreasing score or haz-ard level (e.g., in Region 6, the HRS scorehimeto RIFS start were 47/39, 32/12, and 29/-3).17

That is, the more hazardous the site accordingto the HRS, the longer it takes to start the RIFSon the site. This seems opposite to what mightbe desirable; but in Region 6, the French Lim-ited site ROD said that “The position (rank) ofa site on the [National Priorities] list is incon-sequential.”

The RIFS process, from start of the studiesthrough issuance of the ROD, takes from 2 to3 years. Within this time, early decisions toeliminate some technology alternatives and per-form treatability studies for others could be, butusually are not, made. Studying more technol-ogies than necessary increases the time and costof the RIFS, makes it more difficult to decideto do treatability testing on the most viable tech-

l~he last scoreltime is an example of a site for which the RIFSwas started 3 months prior to the site’s placement on the NPL,

Table I.-Times for Sites To Reach Points in the Superfund Processa

Average RangeFrom entry into Superfund inventory until:

Preliminary Assessment completion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 months 1-45Site Inspection completion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .21 months 1-44Placement on National Priorities List . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .......36 months 4-75Start of RIFS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 months 20-68Completion of RIFS . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75 months 47-103

Signing of ROD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81 months 50-104Completion of ROD remedy (ESTIMATED) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 years 6-20

Between Preliminary Assessment completion until:Site Inspection . . . . . . . . . . . . . . . . . . . . , . . , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 months O-39Placement on NPL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 months 3-73Start of RIFS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42 months 13-68

Between placement on NPL and start of RIFS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 months –3-39Duration of RIFS:

Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 months 21-38Total period (studies through ROD) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 months 24-39

Between signing of ROD and ROD estimate of completion of remedial action . . . . . . . . . . . .......38 months 20-120Duration of public comment period. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .......33 days 24-44Time between end of public comment period and signing of ROD. . . . . . . . . . . . . . . . . . . . . . .......34 days 15-122aB~sgd on the lo cage studies in this OTA special report.

nologies, and sometimes contributes to poorRODS.

After the ROD, actual cleanup action, includ-ing remedial design, takes 2 to 3 years, Some-times there are repeated RODS and new actionson different parts of the cleanup (called oper-able units) and sometimes on the same part ofthe cleanup.

The entire process from site identificationthrough final (estimated) remedial cleanup canfrequently take about 10 years. Unexpectedfindings sometimes complicate the process. Forexample, remedial cleanup stopped at the Con-servation Chemical site in Missouri and at theRe-Solve site in Massachusetts when new infor-mation about the sites’ contamination showeda need for more studies, another ROD, and newcleanup strategies. Some risks to health andenvironment are likely during such long re-grouping periods. Contaminants are likely tomigrate from areas of high to low concentra-tion, increasing the extent and complexity ofcleanup, particularly for groundwater.

Risk Management and Cleanup GoalsThere are often problems with how risks are

assessed and how cleanup goals are met. Differ-ent levels of risk maybe used and very differ-ent cleanup technologies may be said to be com-parable, because EPA allows a broad rangefrom 1 in 10,000 to 1 in 10 million excess life-time cancer risk.la Sometimes compromises aremade to reduce cleanup cost by allowing ahigher risk than the 1 in I million cancer riskcommonly used in Superfund. A cleanup canbe deemed complete even though significantcontamination remains onsite or migrates off-site. Regarding cleanup goals, a cleanup tech-nology can be justified in superficial ways. Haz-ards (the source of the risk) may not beeliminated through permanent technologies butexposures to the hazard—i.e., the risk—may bereduced through impermanent actions, such as

IWancer risk assessment is not the only way cleanup goalsare established. Current regulatory standards for acceptable levelsof contaminants are also used, but these are not available formany contaminants. When risk assessment is used, probable,worst case, or other levels of risk are calculated. Sometimes pre-cleanup risks are also calculated.

capping a site, or institutional controls, suchas deed restrictions that have uncertain futureimplementation.

RODS do not consider cumulative exposuresand risks from multiple sources of similar haz-ardous substances. Cleanup levels may lookacceptable on a site basis but might not whentwo or more Superfund sites are close together.An example is the two Superfund sites in Okla-homa on opposite sides of the Arkansas River;neither ROD evaluates risks from the other site.Environmental risks seem to take a back seatto bureaucratic definitions of Superfund sitesand to constraints imposed by seeking fundsfrom responsible parties.

The risks of transporting hazardous materi-als offsite for land disposal or even treatmentare not considered. Furthermore, SARA’s re-quirements to use permanent treatment tech-nologies are not applied by EPA to waste sentoffsite. The ROD can say that the cleanup willbe permanent, even though the site was origi-nally a land disposal facility, and the wastesare slated for a landfill that itself might becomea Superfund site. Moving hazardous waste fromone hole in the ground to another is the nonso-lution that was behind SARA’s preference forpermanent cleanup. For the purpose of manySuperfund cleanups, EPA’s assumption seemsto be that hazardous waste sent to a regulatedlandfill will never fail and require cleanup eventhough there is widespread agreement, evenwithin EPA, that landfill technology will ulti-mately fail. There are also many widely recog-nized uncertainties about regulatory compli-ance and future corrective action.

Most RODS seem uncertain about or do notaddress future land and water use in judgingwhether a selected remedy will be safe and per-manent. In some cases, there is a lot of interestin reusing the land for productive purposes. Forexample, at the Schmalz site in Wisconsin,where contaminated soil is to remain in place,the ROD makes no land use restrictions. Anyremedy that leaves hazardous waste in placeor caps it suggests the need for explicit atten-tion to future land and perhaps groundwateruse.

15

The Record of Decision Document

The technical content and quality of RODSvaries substantially across and within EPA Re-gions. Supporting RIFSS generally lack cita-tions to the technical literature, important data,and discussions of actual experiences, good andbad, at sites that have used the technologies un-der consideration. Multimillion-dollar decisionsare often made without any significant techni-cal data to support them. A ROD may drop orchoose a cleanup technology with little or nodiscussion or justification.

Probable causes for the meager level of tech-nical detail are: enormous public pressure toclean up sites sooner; attempts to compensatefor delays; bureaucratic pressures to produceRODS faster; poor contractor performance; lackof central, national oversight; and some at-tempts to carry out activities after the RODwhen there is less public scrutiny. Conflicts ofinterest also may be a problem. Does the RIFScontractor own a cleanup technology or willit or some affiliated company stand to profitif a particular cleanup technology is selected?Is the RIFS contractor also a responsible partyat the site? Does a responsible party own thecleanup technology selected for the cleanup?

EPA Regions are not using a standard formatfor RODS. Lack of uniformity makes RODS dif-ficult to analyze and compare for oversight andquality control purposes. Of particular impor-tance is the way alternative cleanups are evalu-ated. Different criteria are used.lg Sometimes

IOA july Ig87 directive from EPA’s Assistant Administratorfor Solid Waste and Emergency Response outlined nine “keycriteria which should be considered in evaluating and compar-ing alternatives. ” An earlier directive contained essentially the

the evaluation focuses on each alternativeseparately with very little comparison. Whencomparative analysis is used, it often is super-ficial and qualitative or semi-quantitative withonly rankings for alternatives.

Even for a technical expert, the basis for acleanup decision is often hard to understand;the public has an even greater problem. RODSoften lack much key information, such as testdata, other nearby sources of contamination,earlier actions, or even an earlier ROD. In hind-sight, earlier actions are frequently ineffectivefrom a longer term perspective and often makesubsequent attempts to permanently clean sitesmore costly and difficult. At the Crystal Citysite in Texas, for example, a previous actionburied hazardous materials which must nowbe excavated and re-buried onsite in a finalcleanup. The ROD offers an opportunity—notyet used—to evaluate past site actions and tolearn from them.

Sometimes a remedy and its implementationconstitute a research or demonstration projectbecause there is no treatability study data orthe technology isn’t proven for the site. But thecleanup is not publicly presented as experi-mental or highly uncertain. While the technol-ogy selected may, in some cases, make sense,the public may ultimately think it unfair of thegovernment to hide the uncertainty and risk.Moreover, making the claim that a permanentremedy has been selected is questionable if thetechnology is experimental.

same evaluation criteria, although they were not presented asclearly. [U.S. Environmental Protection Agency, Office of SolidWaste and Emergency Response, directives 9355.0-21 (July 24,1987] and 9355.0-19 (Dec. 24, 1986)].

10 CASE STUDIES

Each of the 10 case studies in this report isdrawn from a Record of Decision (ROD), whichis EPA’s official understanding of the facts fromsite studies and EPA’s explanation of how thefacts support its selection of a cleanup tech-nology. Each ROD also includes a summary ofhow EPA responded to public comment andgenerally includes a summary of the adminis-trative record (related documents) for the site.EPA can and sometimes does reject, change,and supplement the findings of the contractor(s)who prepare the RIFS and draft the ROD.

Methodology

These 10 case studies were selected from re-cent RODS, from September 1987 through De-cember 1987, which EPA has acknowledgedcame under SARA. OTA identified issues inits 1985 study Superfund Strategy and in theongoing Superfund Implementation study ofwhich the case studies is just one part. Thisother work helped in the selection of represent-ative case studies. The 10 cases here were cho-sen to illustrate different technology selectionproblems, none of which are unique to thesesites, and to illustrate different types of sitesand hazards. This report discusses about 10 per-cent of all recent RODS to which EPA has ap-plied SARA. OTA examined nearly all recentRODs—over a hundred—to verify that the casestudy sites are representative.

Format for Case Studies

The case studies are presented in a standardformat. Following is a sample of the format withexplanations, where necessary, of the catego-ries and of the terms and sources used in thecase studies. Unless noted otherwise, the quo-tations in the case studies are from the site’sROD. Statements from other RODS, from othercase studies, and other Superfund sites are oftenused in the case studies to illustrate programinconsistencies. To complete each site’s history,two EPA data management systems were needed:1) the CERCLIS inventory of all sites reported

to EPA, and 2) the Superfund ComprehensiveAccomplishments Plan (SCAP). The latter is abudget management system; OTA used datafrom SCAP NPL Site Summary reports datedOctober 27, 1987.

Sample Format

Name, location of site, and EPA region:

Capsule OTA findings:

Key dates:

Entered Superfund system: EPA maintainsan inventory of sites called CERCLIS. The datewhen the region gets notification of a site isrecorded is the site discovery date. Many sitesnew to Superfund come with long histories ofcontamination and cleanup efforts.

Preliminary Assessment: The PreliminaryAssessment is the first screening step in the pre-remedial process; it consists mostly of exami-nation of existing records. It is done by EPAcontractors or by States. Sometimes a PA isdone after other actions which are supposedto come before it, apparently to satisfy the re-quirement that it be done.

Site Inspection: The Site Inspection involvessome field work and testing to define the na-ture and scope of the hazard. The S1 is the sec-ond screening step in the pre-remedial proc-ess and leads to Hazard Ranking System (HRS)scoring of the site. The S1 is done by EPA con-tractors or States; the initial scoring is by EPAcontractors, EPA regional staff, or States.

National Priorities List. proposed date:● final date:. site rank:

Sites that get an HRS score of 28.5 or morego on the NPL and become eligible for reme-dial cleanup. Initially EPA proposes a site forthe NPL, and, after an opportunity for publiccomment, the site can become a final NPL site.Final sites are ranked by their HRS score; theranks in the case studies are from the NPL asof July 1987. Then the NPL had 770 ranked sites

16.

17

and more than 200 proposed sites. EPA revisesthe NPL only periodically, approximately oncea year. The significance of the rankings for tak-ing action has not been made clear by EPA. Asite can receive various Superfund actions with-out being on the NPL.

RIFS start and completion: The Remedial In-vestigation/Feasibility Study (RIFS) providesthe information base for the ROD. RIFSS aredone by private engineering consulting firms,paid by EPA, responsible parties, or States.

Public comment period before ROD: EPA isrequired to make available certain documentsfor public review for 21 days prior to the ROD;the period can be extended.

Signing of ROD: EPA Regional Administra-tors officially sign RODS, although in a fewcases the EPA Headquarters Assistant Admin-istrator, Office of Solid Waste and EmergencyResponse, may do so,

Estimated complete remediation: The RODnormally estimates when the final action or,if the action is not final, when the wholecleanup will be done.

Total time: The total elapsed time of abovedates.

Brief description of site:

Major contamination/environmental threat:

HRS scores: EPA’s policy is that a site score aminimum of 28.5 to be placed on the NPL.(Once only, States can nominate one site forthe NPL regardless of its score.) The maximumsubscores for groundwater, surface water, andair are 100, and a formula is used to combinethe subscores so that the maximum total scoreis also 100. (This calculation applies to the ver-sion of the HRS used for the case study sites;a newer HRS version, required by SARA, maychange this methodology.) There are many con-cerns about the accuracy of HRS scores andtheir use in ranking NPL sites; an HRS scoremay not paint an accurate picture of a site’soriginal or current environmental threat. Sitesare not rescored after removal actions or in-terim remedial measures.

Removal actions: Removals are site actions onnon-NPL sites and on NPL sites before (or dur-ing) a remedial cleanup. They are usually han-dled by a different office within the Superfundprogram than that which handles the remedialcleanup. A variety of removal actions can betaken as emergency or time-critical measures.SARA authorizes more time and money forremoval actions than did CERCLA.

Cleanup remedy selected:

Satisfaction of SARA statutory requirements:

1) Selection of permanent cleanup.—TheSuperfund Amendments and ReauthorizationAct of 1986 (SARA) states that EPA shall: 1)“select a remedial action that . . . utilizes per-manent solutions and alternative treatmenttechnologies or resource recovery technologiesto the maximum extent practicable,” and, 2)if this is not done, “publish an explanation asto why a remedial action involving such reduc-tions [in the toxicity, mobility, or volume of thehazardous substance, pollutant, or contami-nant] was not selected.” (SARA Section 121)

Z) Accurate assessment of land disposal andcontainment alternatives.—SARA states thatEPA shall “take into account:” . . . the long-term uncertainties associated with land dis-posal; . . . short- and long-term potential for ad-verse health effects from human exposure;. . .the potential for future remedial action costsif the alternative remedial action in questionwere to fail . . . “ (SARA Section 121)

RIFS contractor: Information on time, cost, andcompany is normally available from the ROD;if not, OTA obtained it from other EPA sources.

State concurrence: Only information reported byEPA in the ROD was used.

Community acceptance: The ROD’s responsive-ness summary was chiefly used. It does not nec-essarily reflect the full range of public opinionabout a site because it only describes direct in-teractions through the official public commentprocess between the community and EPA.

Special comments:

General conclusions:

18

Case Study 1Chemical Control Corp., Elizabeth, NJ;

EPA Region 2

Capsule OTA findings. -Unproven solidification(chemical fixation)technology was selected totreat in situ highly contaminated subsurfacesoil, which previous removal actions had leftbelow the water table and covered up withgravel. No treatability studywas used.The costof incineration was over estimated. The cleanupwill leave untreated contamination on the site.

Key dates:

Entered Superfund system: 5/1/79Preliminary Assessment: 5/1/79Site Inspection: 4/1/79 - 8/1/82National Priorities List–proposed date: 10/1/81–final date: 9/1/83—site rank: #223 out of 770RIFS start and completion: 12/31/84 -6/30/87 (ROD says it began in 11/86, butSCAP has earlier date)Public comment period before Record ofDecision: 7/6/87 - 8/14/87Signing of ROD: 9/23/87Estimated complete remediation: 28 to 32months after R-OD (around 4/90)

Totai time.—ll years

Briaf description of site.—“The site consists of this2.2-acre property and the portion of theElizabeth River adjacent to the property, . . . thewater table aquifer at the site [is] saline and ti-dally influenced. The site is flat and barelyabove sea level. Chemical Control Corp. oper-ated from 1970 until 1978 hauling, treating, anddisposing of a wide variety of industrial wastes.Throughout its operations, it was cited for vio-lations thet included discharging liquids ontothe ground adjacent to the Elizabeth River andaccumulating thousands of drums of incompat-ible wastes.”

Major contamination/environmontai tlnat.–” . . . soilsbeneath those removed by the NJDEP [New Jer-sey Department of Environmental Protection]are highly contaminated with a variety of or-ganic compounds and to a lesser degree with

metals. . . . these contaminants are strongly ad-sorbed to the soil and are present in the ground-water in relatively low concentrations. The con-taminants found in the [river] sediments ., . arenot all attributed to the Chemical Control site.”The contaminated layer “averages approxi-mately ten feet thick. . . . some of the moremobile chemicals continue to leach into thegroundwater. Significant health threats areposed by direct contact, fugitive dust emission,and volatilization. Contaminants are only leav-ing the site via the groundwater. . . . direct con-tact with sediments as well as ingestion of con-taminated shellfish are both potential exposureroutes. Flooding happens occasionally at thesite now . . . “ The ROD indicated a volume ofcontaminated material of 18,000 cubic yards.

HRS scores.—groundwater O; surface water18.18; air 79.49; total 47.13

Removai actions. —State removal of large quan-tities of wastes began in March 1979 and wasinterrupted by a major fire in April 1980. Afterthe fire, the State removed more material,including 3 feet of surface soil which wasreplaced with gravel. Also, from November1980 until July 1981, the State operated agroundwater recovery and treatment system.This action plus groundwater movement andnot just the adsorption of contaminants to soilmay explain why the subsequent Remedial In-vestigation found little contamination and whythe HRS groundwater score in 1982 was zero.Overall, the State of New Jersey has spent $25million on the site.

After the site became a Superfund site, fouradditional initial remedial measures were car-ried out (in 1984, 1985, 1986, and 1987) to re-move more materials from the site.

Cieanup remedy selected. —Other than contain-ment, treatment alternatives considered weresoil washing, solvent extraction, and incinera-tion, The selected remedy was in situ fixation(chemical fixation, stabilization, and solidifi-cation often are used to describe similar treat-ments). Fixation chemicals would be injectedthough an expandable bit drill which wouldpass through the gravel layer: “A series of over-

19

lapping columns would be formed convertingall of the contaminated soil at the site into asolid mass. This would inhibit water from flow-ing through the site, thereby preventing theproduction of leachate. In addition, some con-taminants may be chemically altered and in-corporated in the solid matrix formed by thisaction, reducing the toxicity as well as the mo-bility of the contaminants. The potential for ex-posing the contaminated soil would be elimi-nated. The treatment] will create a solid matrixthat will have extremely low permeability. . . .because it is implemented primarily belowground, [the treatment] offers protection againstreleases during a flood. ”

The estimated cost for the selected remedyis $7.4 million, while the cost for excavation,onsite treatment, and onsite disposal of residuesis $14.5 million for fixation and soil washingand $22.3 million for incineration. The RODalso commits to some other relatively minorremoval actions and environmental monitor-ing, “including an evaluation after five yearsto assess its protectiveness to public health andthe environment. ”

Satisfaction of SARA statutory requirements.

1) Selection of permanent clean up.—’’The re-medial alternative presented in this documentis a permanent solution for closure of the Chem-ical Control site. . . . this remedy satisfies thepreference for treatment that reduces toxicity,mobility, or volume as a principal element. . . .this remedy utilizes permanent solutions andalternate treatment technologies to the maxi-mum extent practicable. Based on the input re-ceived during the public comment period, thisalternative has been selected by the EPA andthe NJDEP as the final permanent solution forthe site. [The selected remedy] also utilizes analternative treatment technology that offers amore cost-effective remedy. ”

The fact that the ROD met SARA’s require-ment for a five-year review indicates that EPArecognizes that the selected remedy would leaveuntreated, undestroyed, and toxic wastes on-site. EPA normally responds to the statutoryrequirement for review for land disposal/con-

tainment remedies and when only partial de-struction treatment technologies are used at asite. The review has implications for future landuse because use of the land might interfere withreviews and because results of reviews mightreveal hazards that would block land use.

A major issue with the selected remedy ofin situ fixation is that it is not a proven tech-nology; no data exist to show effectiveness forcleanup of a hazardous waste site comparableto Chemical Control. No treatability study wasconducted prior to the ROD to provide evidenceof effectiveness in terms of resistance to long-term leaching or actual toxicity reduction. Thediverse set of contaminants at the site wouldpose a challenge to conventional chemical fix-ation techniques. The use of in situ fixation be-neath the water table in saline conditions mayexacerbate the difficulty of achieving an effec-tive cleanup. Various ROD statements on thisissue include:

●

●

●

●

●

●

“Although in-situ fixation is not yet a stand-ard construction practice, several vendorsare available that provide this service. ”64 .*, the in-situ process described in thisdocument is currently being studied for useat other hazardous waste sites. ”“The long-term reliability of this alterna-tive is especially promising” (emphasisadded).“A treatability study and field test will berequired during design to prove the tech-nology . . . “64 environmental samples will be col-lected to monitor the effectiveness of theremedy.” (No specific technical criteria aregiven.)“Although such an application of this tech-nology is fairly new, promising results havebeen obtained in laboratory tests, and it isbeing tested at other hazardous waste sitesand evaluated under the Superfund Inno-vative Technology Evaluation program.The Chemical Control project will benefitfrom the experience gained at these sites.”

Despite the last comment, OTA has been onlyable to identify the evaluation within the SITEprogram. A vendor in conjunction with Gen-

20

eral Electric has demonstrated its technologyat a GE site in Hialeah, Florida, in April 1988.PCB contaminated soil was treated. A demon-stration report is supposed to be available ap-proximately four to six months after the dem-onstration. OTA contacted the vendor of thein situ chemical fixation technology for the GEsite and was told that no tests have been con-ducted on materials from Chemical Control butpreliminary laboratory tests have been success-ful on material from the Hialeah site. GE willbe using the technology to actually clean upthe Hialeah site; the cleanup will constitute theSITE demonstration. GE plans to do this clean-up even though EPA has not officially sanc-tioned its use at Hialeah; the agency has prob-lems in issuing regulatory permits for aninnovative treatment technology, necessary be-cause the cleanup is not at a Superfund site.Given the timing of the Hialeah demonstration,it is difficult to understand how the selectedremedy can be already judged to satisfy require-ments for Chemical Control. Officially, thereare no performance criteria to prove the effec-tiveness of the technology. Moreover, one ofthe objectives for the SITE demonstration is todetermine the “integrity of the solidified soilover a period of five years.” (U.S. Environ-mental Protection Agency, “Superfund Inno-vative Technology Evaluation (SITE) Program,”HWERL Symposium, May 9-11, 1988.)

But when will the technology be completelyevaluated? Some innovative technologies havebeen demonstrated several times and still haveuncertainties for broad use. The informationavailable on the SITE demonstration is incon-sistent with the schedule for implementationof the selected remedy given in the ChemicalControl ROD and seasonal constraints may alsodelay cleanup.

Moreover, the SITE demonstration will notbe performed on a comparable contaminatedmaterial. The presence of volatile organics andmetals, for example, makes the Chemical Con-trol project significantly different. A positiveSITE demonstration result will not, therefore,substitute for a treatability study on ChemicalControl material nor should it justify perform-ing a pilot study at Chemical Control.

The EPA decision to use incineration at theSouthern Crop Services site in Florida at aboutthe same time of the Chemical Control deci-sion undercuts the use of stabilization forChemical Control. In justifying its selection ofincineration, EPA said: “Solidification and fix-ation of pesticide contaminated soil was foundto be technically unacceptable due to the highdetected concentrations and because organicpesticides tend to leach from solidified mate-rial. This technology was deemed unaccept-able.” (U.S. Environmental Protection Agency,Region 4, memorandum, Sept. 8, 1987.) Pesti-cides are among the many different types oforganic contaminants present at Chemical Con-trol. The FS for the Re-Solve site in Massachu-setts rejected stabilization because “there hasbeen limited success in chemically fixing organiccontaminants such as solvents and PCBS. ”

The ROD for the Liquid Disposal site in Mich-igan, which also selected stabilization for soilcontaminated with organic chemicals, said thatthe hazardous substances “will not be perma-nently destroyed” and “hazardous chemicalsstill remain in that [treated] mass. ” And the FSfor the site said: “Considerable research dataexists demonstrating the effectiveness of thistechnology in immobilizing a wide range ofcontaminants, primarily inorganic. A substan-tial amount of data does not exist, however, toaccurately judge the long-term reliability of theprocess.” Of particular significance to the useof stabilization at Chemical Control, the Liq-uid Disposal ROD also selected a slurry walland impermeable cap around and over thetreated material, in part because it is necessaryto “protect the solidified soil/waste from degra-dation by upgradient ground water that isslightly contaminated with chemicals” (empha-sis added).

A recent EPA study found “large losses oforganics during the [stabilization] mixing proc-ess.” (L. Weitzman et al., “Evaluation of Solidifi-cation/Stabilization As A Best DemonstratedAvailable Technology,” paper presented atEPA’s Fourteenth Annual Research Sympo-sium, May 1988.) Another EPA study showedthat stabilization was not competitive with ther-mal and chemical treatment technologies and

21

soil washing for organic contamination. (R.C.Thurnau and M.P. Esposito, “TCLP As A Meas-ure of Treatment Effectiveness: Results ofTCLP Work Completed on Different TreatmentTechnologies for CERCLA Soils,” paper pre-sented at EPA’s Fourteenth Annual ResearchSymposium, May 1988.) A demonstration of astabilization technology under EPA auspicesconcluded that “for the organics, the leachateconcentrations were approximately equal forthe treated and untreated soils.” (P.R. de Per-cin and S. Sawyer, “SITE Demonstration ofHazcon Solidification/Stabilization Process,”paper presented at EPA’s Fourteenth AnnualResearch Symposium, May 1988.)

Senior EPA people have made an importantobservation about in situ stabilization:” . . . thedecision to use a stabilization technique shouldbe made only after the chemical and physicalproperties of the solidified waste have been ex-tensively tested to insure that the required prop-erties have been developed. ” (D.E. Sanning andR.F. Lewis, “U.S. EPA Research on In-SituTreatment Technology,” Anatomy of Super-fund, proceedings of the 8th National GroundWater Quality Symposium, September 1986.)The ROD for Chemical Control does not meetthese requirements.

An educational short course and two recentEPA documents on cleanup technologies makeno reference to the selected technology. (U.S.Environmental Protection Agency, “RemedialEngineering of Hazardous Waste Sites,” TheNational Hazardous Materials Training Cen-ter, October 1987; “A Compendium of Tech-nologies Used in the Treatment of HazardousWaste,” September 1987; “RCRA/CERCLATreatment Alternatives for Hazardous Wastes,”October 1987.) The latter EPA document is usedto teach people implementing hazardous wasteprograms about waste treatment and says:“Solidification technologies are designed to beused for final waste treatment. This means thetechnology should be applied only after othertreatment techniques have been applied, i.e.,incineration, chemical treatment or other.”

Another observer has commented on this ap-proach: “Experimental studies have been con-

ducted in the field. The level of treatmentachievable is variable, depending on the wasteand soil conditions. The potential for long-termimmobilization is unknown at this time. Thereliability of the treatment is unknown sincethere is no information on its long-term effec-tiveness.” (R. Sims et al., “Contaminated Sur-face Soils In-Place Treatment Techniques,”Noyes Publishers, 1986.)

The Feasibility Study for the Crystal City siterejected in situ chemical stabilization: “Im-mobilization, chemical treatment, and physi-cal treatments have not been shown to be fea-sible for in situ treatment of these contaminantsas it is not possible to get a good, uniform, welldistributed treatment.” The focused FS for theConservation Chemical Co. site in Missouri re-jected in situ stabilization: “Technology [was]attempted and [was] found not feasible at othersimilar sites. Technology is not sufficiently de-veloped.” An addendum study also rejected insitu immobilization (which cannot be differen-tiated from stabilization): “Technology wouldnot effect organic and other non-metallic con-tamination; thus, these substances would con-tinue to be a source of contamination. Immobili-zation reactions are reversible.” The FS for thePristine site in Ohio rejected in situ chemicaldetoxification: “Treatability study is requiredto assure effectiveness. It is difficult to ensureproper reactant mixing and verify effectiveness.”

The same contractor that prepared the Chem-ical Control FS has tested in situ chemicalstabilization elsewhere. (“Feasibility Study Salt-ville Waste Disposal Site, Smyth County, Vir-ginia,” August 1986.) Successful laboratory andpilot tests led to the fieldtesting of a particularin situ treatment; however, the fieldtests failedand the approach was dropped. This twist il-lustrates the uncertainty of a technology, evenafter successful laboratory tests.

The ROD said that the selected remedy “offersa level of long-term protection comparable toor exceeding that of any of the other alterna-tives.” However, if the selected remedy has notbeen shown to destroy the organic contami-nants the way incineration could, is this asser-tion of comparable permanence correct? In-

22

cineration followed by chemical fixation of theresidue to immobilize toxic metals offers ahigher level of protection. Therefore, the addi-tional estimated cost for the incineration op-tion (three times more than fixation) does noteliminate its cost-effectiveness. Also, the costestimate for the selected remedy is unreliablebecause the technology has not been used be-fore on such a site.

This ROD illustrated the benefit of examin-ing the supporting Feasibility Study. In thiscase, the FS introduced several new elements:

1. The ROD did not reveal several facts aboutthe site and the selected remedy that were inthe FS: a) the curing time for the fixation mate-rial is about one month; b) there is a volumet-ric increase in the waste after treatment thatdepends on site materials and conditions; c) “Itis unlikely that solidification can be effectedin contaminated areas at the interface betweenthe river and the site. This residual contami-nation will continue to flush from such areassurrounding the solidified mass”; d) “This alter-native will not reduce any potential humanhealth or environmental impacts associatedwith the contamination detected in the gravelcover atop the site”; e) “ . . . even under non-flood conditions, the water table is quite closeto the surface of the site”; and f) The estimatedcosts for the bench test and pilot test for thein situ fixation alternative are $770,000.

Z. The FS analysis of the selected remedy sug-gested that there is reliable information onwhich to base conclusions. The text containedphrases such as: “has been demonstrated,” “theavailable literature, ” and “it is reported,” sug-gesting that technical literature and EPA re-ports were used. But, all the information camefrom a single vendor (identified in figure 3-6in the FS). The FS also referred to the GE Hia-leah site as a Superfund site, which it is not,and the FS said (in June 1987) that “a field ap-plication of the emerging technology is pres-ently underway, ” which it was not.

3. The FS analysis of incineration is poor. Theunit cost for incineration only (with other costsfigured separately) was $750 per cubic yard forthe combination of onsite incineration and off-

site disposal (where a baseline of 21,000 cubicyards of soil was used) and $00 per cubic yardfor the combination of onsite incineration andonsite disposal (where a baseline of 27,000 cu-bic yards of soil was used). These differencesin unit cost and soil volume do not make sense.Other vendors are now quoting less than $3OOper cubic yard for the volume of work at thissite and, indeed, $3OO per cubic yard was quotedin the FS for the Davis Liquid Waste site inRhode Island (where onsite incineration wasselected) and $186 per cubic yard was quotedin the FS for the Seymour Recycling site. TheChemical Control FS also stated that “Rotarykiln and fluidized bed incinerators are the onlytypes of mobile units currently available.” Thisstatement is not true. According to EPA, thereare three other types of full-scale mobile ther-mal technologies available: circulating bed, in-frared, and wet air. (U.S. Environmental Pro-tection Agency, “RCRA/CERCLA TreatmentAlternatives for Hazardous Wastes,” October1987.) If $300 per cubic yard were used (withother costs factored in separately) for onsite mo-bile incineration of 18,000 cubic yards (the sameamount of material as for the selected remedy),then the cost would be about $14 million. Thiscost compares to the ROD cost for onsite in-cineration with onsite disposal for 27,000 cu-bic yards at $22 million. The ROD omitted FSlow and high cost estimates for the options.Since there is no field experience within situfixation, its high cost estimate of $14 millionis significant as an estimate. Thus it is possiblethat incineration at $14 million might be aboutthe same cost as the selected remedy and notthree times more. More recently, after the ROD,a news story reported that the cleanup project’sestimated cost is $10 million, with $750,000 ahlocated for the one-year design job, an amountwhich could not account for the cost of the treat-ability study. (Superfund, Feb. 1, 1988.)

4. The analytical framework used to evalu-ate alternative cleanup approaches is inconsist-ent with commonly accepted practice and with EPA’s recommendations. A July 1987 EPAdirective clearly recommended the use of ninecriteria; an earlier directive was less clear. Onlyfour criteria were used in this FS: technical fea-

23

sibility, public health and environmental con-cerns, institutional considerations, and cost.However, analysis using additional criteria,particularly factors such as reliability and im-plementability which are normally stressed,would have worked against the selected rem-edy. (Indeed, this analytical result happened inthe ROD for the Conservation Chemical Co. sitein Missouri, in which the only soil treatmentalternative faired poorly on reliability and im-plementability and thus was rejected.) TheChemical Control ROD evaluated each cleanupalternative separately. An explicit comparisonof alternatives weighing relative advantagesand disadvantages was not done in the FS orthe ROD. (At Conservation Chemical, the RODcomparative analysis of cleanup alternativeswas done within individual discussions ofevaluation criteria, a very useful approach com-pared to discussing each alternative for allcriteria.)

To sum up, an analysis of the FS shows thatthere was little basis to select the in situ chemi-cal fixation alternative and that the FS analy-sis was biased in favor of the selected remedyand against using onsite incineration. Onsiteincineration is a proven, more cost-effective,and more reliable cleanup alternative than theFS and ROD indicated.

2) Accurate assessment of land disposal andcontainment alternatives. —To a significant ex-tent, the selected remedy is a land disposal/con-tainment approach because, unless shownotherwise with positive test results, chemicalfixation cannot be assumed to detoxify all con-taminants. Leaving the contaminated soil on-site and beneath the water table raises ques-tions about future failure. No technical criteriawere established to determine failure of theselected remedy. Failure is a real possibility,since the same generic treatment failed at theConservation Chemical Co. site. A volume in-crease in the treated waste with the additionof fixation chemicals raises questions about theintegrity of the resulting solidified mass withinthe site with in situ use. Costs were not esti-mated to repair a failure of the selected remedy.

A traditional containment wall was exam-ined, but the ROD said that a containment wallmight need replacement in the future. But theROD noted that a wall “would offer effectiveprotection if institutional controls were im-posed to prohibit any future digging at the site.”

Interestingly, in the Chemical Control situa-tion, a containment approach might have madesense—as an interim measure—because thereare many other sources of river contaminationaround the site that could reduce the effective-ness of a remedy that leaves site material vul-nerable to recontamination. The ROD noted therisk of recontaminating clean material back-filled into the site. Contaminants from the rivermight also affect the effectiveness of chemicalstabilization.

RIFS contractor. —State-led; NUS Corp. undersubcontract to Ebasco Services, Inc.; about $1million obligated. SCAP indicated a RIFS from9/28/83 - 2/15/85 at a cost of $208,000, a subse-quent ROD on 2/15/85, a Remedial Design at$504,000, and a Remedial Action at $485,000.There was no information in the 1987 ROD’ssite history on the earlier RIFS, ROD, and re-medial action, and SCAP listed separately theinterim remedial measures and removal actionswith their costs. To confuse things still more,a master EPA list of all Superfund RODSshowed an earlier ROD on 9/19/83.

State concurrence.—New Jersey agreed with theselected remedy.

Community acceptance.—The ROD said: “Themain concern of local officials is that a thor-ough, permanent remedy be expeditiously im-plemented.” The responsiveness summary alsoindicated some public concern about the effec-tiveness of the selected remedy, particularly be-cause of its inattention to toxic metals.

Special comments.—The ROD did not addressthe problem of highly contaminated river resi-dues. The chief reason is the other sources ofcontamination: “ . . . remediation of the riversediments is premature. ” If the river residuesare considered part of the Chemical controlsite, then this ROD does not offer a final per-manent cleanup of the entire site.

24 ●

Why did the State leave the contaminated soilin place in a previous site action and cover itup with permeable gravel? This action compli-cates a permanent remedy and will contributeto continued leaching of contaminants into theriver and its sediments for about 10 years. Eventhough EPA said that the contaminants adhereto the soil, not all the contaminants could be-have so ideally nor, as discussed in the FS, willall contaminants be treated to reduce their mo-bility. The ROD gave no data to support the con-tention that all the contaminants are tightlybound to the soil. The ROD noted: “The NJDEPhas also indicated that contaminant concentra-tions in the soil at Chemical Control exceedState guidelines.” It is not clear why an interimcontainment action at the site, such as a slurrywall and cap, was not implemented years ago.

Other innovative in situ treatment technol-ogies—including biological treatment andvitrification—could have been considered intreatability studies as viable candidates, butwere not. The FS rejected in situ vitrification(ISV) on grounds that also could have been usedto reject the selected remedy. Yet ISV wasselected for the Pristine site, and the CrystalCity FS evaluated ISV favorably, although it didnot select ISV or any treatment alternative:“[ISV] has been successfully demonstrated inlaboratory and bench testing. IISV] was deter-mined to be feasible given the existing infor-mation available and is retained for furtherevaluation. ” A recent NJDEP report’s discus-sion on innovative/alternative technologiessaid: “various technologies presently existwhich can adequately address contaminatedsoil and other contaminated media. For exam-ple, waste vitrification (imbedding waste inglass) can immobilize organic or inorganic con-taminants while generating residuals that aredelistable and environmentally safe.” More-over, the NJDEP report also noted that treata-bility studies are done during the RIFS to “filldata gaps . . . and supply information neededto select a design alternative.” (New Jersey De-partment of Environmental Protection, “Com-prehensive Management PIan 1988-1992,” Oc-tober 1987.) The point here is not whether ISVis the best cleanup technology for the site but

that a case could have been made to evaluateit as the ROD did for in situ stabilization.

The ROD had no summary of the adminis-trative record which, because of the inconsist-ent information on dates, would have been veryuseful.

General collusions.–EPA’s high confidence andcertainty about the selected remedy is unsup-ported by analysis. In this case as in severalother case studies, the ROD did not followEPA’s guidance that ROD analysis “must bebased on a specific process within [a] technol-ogy category . . . to ground the analysis in harddata.” (U.S. Environmental Protection Agency,“Interim Guidance on Superfund Selection ofRemedy,” Dec. 24, 1986.) The ROD packagecontained a letter of August 31, 1987, in whichEPA told the City of Elizabeth’s Director ofHealth, Welfare and Housing:” . . . we feel thatin-situ fixation will protect public health andthe environment from any hazards posed bythe site.” As with the statements in the RODabout the selection being permanent, the state-ment in the letter was inconsistent with theneed to prove the effectiveness of the remedy,through a treatability study, after it was selectedbut before it is fully applied. In the same letterto the city official, EPA said: “In the event thatthese tests show that in-situ fixation would failto offer protection of public health and theenvironment, the ROD would be amended asnecessary. ” Examining alternatives, selectinganother remedy, amending the ROD, and im-plementing another remedy would, of course,take considerable time. Indeed, this scenariohas happened at other Superfund sites, includ-ing the Re-Solve site in Massachusetts, the Con-servation Chemical Co. site in Missouri, andat Love Canal in New York.

A systematic bias against incineration wassuggested in the site FS evaluation, particularlyfor cost. Use of mobile incineration might notcost significantly more than the selected remedy,but it would offer more certain effectiveness.

The Chemical Control site illustrates the prob-lem of delaying a treatability study until the de-sign phase. While chemical fixation is consid-

25

ered a treatment, it cannot be assumed todetoxify all contaminants. A treatability studyduring the design phase should be limited toobtaining data necessary for the detailed, engi-neering design of the selected technology andalso to develop technical criteria to guide po-tential bidders on the project. If a treatabilitystudy is necessary to show effectiveness, as inthis case, then it should be done, as it some-times is, during the RIFS process. If it is de-layed, then “Such a deferral may result in apremature (and administratively ‘irreversible’)commitment to a technology that may not beappropriate for a given site.” (D. Truitt and J.Caldwell, “Evaluation of Innovative WasteTreatment Technologies,” Waste ManagementConference-Focus on the West, Colorado StateUniversity, June 1987.)

The selected remedy for Chemical Controlcan be considered a land disposal/containmentapproach. OTA does not mean to challenge themerits of the in situ chemical fixation technol-ogy but does question the decisionmaking proc-ess used at this site. Making the remedy selec-tion before treatability test results are availablemay mean that EPA was in a hurry to promoteinnovative treatment technology and to issuethe ROD.

The ROD also did not assure a permanentremedy for the site because it ignored thecleanup of the highly contaminated river sedi-ments, ignored the contamination in the gravel,and ignored the untreated material at the river’sedge. The ROD over estimated the cost of on-site incineration, which could achieve morepermanent, more complete, and more certaincleanup at a cost of about $14 million, insteadof the ROD’s $22 million estimate.

The serious complication of other nearbysources of contamination shows that Superfundsites cannot be seen in isolation. The ROD notedthat recontamination of the site is a p~tentialproblem. Therefore, a case could have beenmade for coordinating this cleanup with othersto assure an overall, permanently effective so-lution for all of them.

There seems to be an unusual interest in theRIFS and ROD process in reusing the site and

constructing something on it, despite the un-certainty of the selected cleanup, despite thecontaminated materials to remain onsite, anddespite the other nearby sources of contami-nation. The FS indicated that the State of NewJersey owns the land and that, with the selectedremedy, New Jersey’s own law regarding realestate transfer would be violated if the site wasput into commercial reuse “Since some resid-ual contamination will exist in the gravel coverunder [the selected remedy], and since the sub-surface contamination will still be present (al-though immobile), it is unlikely that this alter-native ,will comply with the concentrationrequirements of ECRA [New Jersey’s Environ-mental Cleanup Responsibility Act]. It is feltthat this alternative will be consistent with theintent of ECW, however.” The FS noted a can-cer risk above 1 in 1 million for contact withthe contaminated gravel, a risk that has impli-cations for future use of the site and onsiteworkers.

Case Study 2Compass Industries, Tulsa County, Oklahoma;

EPA Region 6

Capsule OTAfindings.-Capping (containment) ofwaste was chosen over incineration, cappingwas called a cost-effective, permanent cleanupeven though it does not provide permanent pro-tection co~arable to incineration. Treatmentof contaminated groundwater is not yet planned.

Kay datas:●

●

●

●

a

●

●

●

●

Entered Superfund system: 10/1/80preliminary Assessment: 4/1/80Site Inspection: 7/1/82National Priorities List–proposed date: 9/83–final date: 9/84—site rank: #483 out of 770RIFS start and completion: 6/29/84- 7/13/87Public comment period before Record ofDecision: 7/22/87 - 8/31/87Signing of ROD: 9/29/87Estimated complete remediation: 9/90

Totai time.—lO years

.

26

Brief description of site.-The site is” . . . an aban-doned landfill located west of Tulsa, Oklahoma.The site occupies an abandoned limestonequarry. From 1972 to 1976 the site was per-mitted and operated as a solid and industrialwaste landfill. physically, the site is situatedon a bluff approximately one-quarter mile southand zoo feet above the Arkansas River. Anelementary school lies within one-half mile anda major regional park is immediately adjacentto the site.”

Major contamination/environmental threat.—” . . . alarge number of organic and inorganic prioritypollutants were detected. They include a totalof 12 inorganic priority pollutants and at least33 organic priority pollutants . . . pathways ofpossible off-site contaminant migration are sur-face water, groundwater, and air. The possi-bility also exists for direct contact at the sitewith contaminated source materials, such assludge, soil, or sediments. The majority of thecontamination in the groundwater is confinedto the upper aquifer. Samples of groundwaterfrom monitoring wells on the site are highlycontaminated. This indicates a degradation ofgroundwater quality due to waste disposal inboth the perched and deep aquifers. The vol-ume of waste was determined to be approxi-mately 620,000 cubic yards. The average ground-water flow rate of both aquifers is 720 gallonsper day or an estimated 263,000 gallons of waterper year [into Arkansas river]. The most recentfire burned for several years before it appar-ently burned out in 1984. . . . there exists a po-tential for future fires. . . . [During fires] ele-vated levels of air contaminants may presenta health hazard.”

HRS scores .—groundwater 11.05; surface water18.46; air 59.49; total 36.57

Removai actions.—None indicated.

Cieanup remedy selected.—Two major alternativeswere considered: 1) leaving waste in the ground,capping the site, and treating groundwater; and2) incineration of excavated wastes. There arethree parts to the selected remedy: 1) cappingthe site; 2) if deemed necessary through com-pliance, monitoring after installation of thecover material, collecting and treating onsite

the contaminated groundwater in the upper,perched water bearing zone; and s) installingfences and signs along the perimeter of the cap.“This alternative consists of site grading, capplacement, diversion of surface water, and airemissions monitoring. The site cap will be re-quired to meet RCRA specifications. Ground-water will be treated at a later date if found tobe necessary. The site will be monitored for aperiod of at least so years . . . to ensure that nosignificant contaminant concentrations mi-grate from the site” (emphasis added).

Estimated cost: $12 million.

Satisfaction of SARA statutory requirements:

1) Selection of permanent cleanup.—TheROD said that “ . . . [the selected] remedy uti-lizes permanent solutions and alternative treat-ment technologies to the maximum extent prac-ticable.” The ROD acknowledged that cappingonly reduces the mobility of contaminants;groundwater treatment would also “reduce thevolume and toxicity of wastes on site to somedegree.” Hazardous residues from water treat-ment would be shipped offsite to a landfill.

Full onsite thermal destruction was examinedas an alternative and was given highest ratings“because this process would destroy the or-ganic compounds in the waste.” Partial onsitethermal destruction of 2 percent of the waste,coupled with capping and groundwater treat-ment, would have been an improvement overcapping and groundwater treatment alone. Re-garding full onsite thermal destruction: “ . . .this remedy is not cost-effective ($339 millionvs. $12 million). ” However, the ROD acknowl-edged that full incineration would give the bestoverall environmental protection. Regardingpartial thermal destruction: “ . . . the increasein cost does not justify the negligible increasein protection to human health and the environ-ment.” Also, regarding implementability: “On-site incineration remedies . . . will require rela-tively more attention during design than otherremedies . . . and were therefore rated lowerthan the other alternatives.” That is, treatmentrequires more work than waste containment.

2) Accurate assessment of land disposal andcontainment alternatives. —Regarding short-

27

term health effects for capping, there are” . . .essentially no risks to workers or residents.”For long-term effectiveness, the remedy “ . . .will successfully reduce long term risks to hu-man health and the environment.” Standardoperation and maintenance for site and cap wasplanned. Regarding future actions: No futureremedial actions are anticipated. The selectedremedial action is considered permanent” (em-phasis added).

The selected capping remedy, was given twohigher levels of ranking for reduction of toxic-ity and volume of waste as compared to the noaction alternative, even though capping doesno more to waste than no action. Ratings forthe reduction of mobility for the selectedremedy are probably too high, especially be-cause it is not certain whether or not ground-water movement would be affected.

Several other areas of uncertainty remain:●

●

●

“Future land use considerations will beevaluated in the upcoming design phasebased on the needs of protection of thecap. ”That some water infiltration through thecap, which would cause migration of con-taminants into groundwater, might happenis indicated by the possible use of a syn-thetic liner in the cap: “The long termadvantage to the liner is that less waterwould be generated from the seeps” (em-phasis added).With regard to long-term impacts: “The po-tential for future fires and continued off-site migration of contaminants pose ad-verse human health and environmental im-pacts. Other impacts which the site maypose cannot be effectively predicted, ARCRA cap and groundwater treatmentwould mitigate these problems as well asmost of the unseen, long term problems”(emphasis added).

In the FS for the Pristine site in Ohio, land-filling the contaminated soil was rejected “Be-cause there is no treatment of soils to reducethe mobility, toxicity or volume [it] is not a per-manent remedy. . . and [it] is the least preferredunder SARA” (emphasis added). In the FS for

the French Limited site (in the same EPA re-gion as Compass), use of a slurry wall and capto contain hazardous waste was described asa “temporary solution” for which the “volumeand toxicity would not be affected” and “thepotential would always exist for failure of ei-ther the cap or the slurry wall allowing for themovement of unstabilized wastes containedonsite. ”

RIFS contractor. –State led, $624,000; JohnMathes & Associates.

State concurrence.—“The State . . . has concurredwith the capping portion of this remedy. . . . theState did not support any of the other proposedremedies.”

Community acceptance.—” . . . the public was infavor of [capping] over thermal treatment of thewaste. . . . the public concern was that the ther-mal treatment unit would create hazardousemissions and increase the potential for ex-posure.”

Special comments:

●

●

●

●

No treatment technology other than ther-mal destruction was considered in the fi-nal analysis, although other possibilitiesexisted.No commitment to using a liner was madeeven though “ . . . Subtitle C of the Re-source Conservation and Recovery Act,which requires a cap with liner, is relevantand appropriate. ”No specific technical criteria were used fordeciding what types and levels of contami-nation found via groundwater monitoringwould trigger actual groundwater treat-ment. There was no comment on level ofcertainty that groundwater monitoringwould in fact detect plumes of contami-nation.No consideration was given to the effectof leaving wastes in the ground and to theeffect of contaminants that have alreadymigrated into the subsurface. These sub-surface contaminants can cause futurecontamination of groundwater that movesinto and through the site area and eventu-ally into the Arkansas river, even though

28

capping reduces water infiltration throughthe site surface.

● Although the ROD acknowledged the SARArequirement to review the chosen remedy,which leaves waste onsite, every five years,there is no explicit commitment to doingso.

General conclusions.-The remedy selected (cap-ping) and its supporting analysis do not satisfystatutory requirements on remedy selection.The selected remedy is not, as the ROD as-serted, a permanent remedy. A number of state-ments in the ROD contradict the claim of per-manency. For example, the possibility clearlyexists for future remedial action because wastesare left untreated in the ground: “If however,future migration does occur appropriate reme-dial actions will be taken.” The long-term un-certainties, the potential environmental risks,and future cleanup costs for capping have notbeen examined. Moreover, the perspective onland disposal and capping in this ROD is in-consistent with work at other Superfund sites.

EPA said that the selected remedy is less envi-ronmentally effective than thermal destruction;therefore, the chosen remedy is less cost-effec-tive. Despite the extremely high cost of totalincineration, the issue of the environmental ef-fectiveness of capping remains. If capping isnot effective, then its lower, more attractive costdoes not make it cost-effective and does notmake it a permanent remedy. It is not an either-or situation.

To reduce cost, the partial incineration op-tion of hot spots of contamination could havebeen a compromise option. Perhaps spendingtwo to three times more money than capping,instead of 20 to 30 times more for complete in-cineration, could have provided a permanent,cost-effective remedy. The ROD suggested thatthe site area is some 100 acres, but a statementin the responsiveness summary refers to 32acres for the cap. In either case, the amountof soil sampling at the site—28 locations—wasinsufficient to accurately characterize contami-nant distribution. (Assuming there are 32 acres,sampling is about one location per acre. Forcomparison, at the Renora site in New Jersey,sampling was done in 12 locations per acre;

at the Seymour Recycling site in Indiana, it wassix locations per acre; and at the Tacoma TarPits site in Washington, it was one-and-a-halflocations per acre.) Hence, there was insuffi-cient data to consider how partial excavationand incineration for the most contaminatedareas might be cost-effective.

Doing enough soil sampling to assess a siteaccurately enough to detect hot spots has beenstudied by EPA. Soil sampling is a major ef-fort: “Systematic sample site selection is nor-mally used when attempting to determine arealextent of contamination or when evaluatingspatial variations. Sampling locations are de-fined by a grid or coordinate system and sam-ples are collected at preselected locations ina uniform pattern. ” (R.J. Bruner, “A Reviewof Quality Control Considerations in Soil Sam-pling,” Quality Control in Remedial Site Inves-tigation, American Society for Testing and Ma-terials, 1986, pp. 35-42.) The critical tradeoffbetween the cost of taking more or less cleanupaction has been summed up by EPA: “If thecost of a false positive (incremental cleanup ofadditional area) is less than the cost of a falsenegative (health risk due to not cleaning anarea), then the larger probability of false posi-tive is acceptable. If the [contaminant concen-tration action level] were raised, the probabil-ity of false positives (unnecessary cleanup]would be lessened, but with an increase in theprobability of a false negative (leaving a ‘dirty’area).” (G.T. Flatman, “Design of Soil SamplingPrograms: Statistical Considerations,” QualityControl in Remedial Site Investigation, Amer-ican Society for Testing and Materials, 1986,pp. 43-56.) The latter happens when averagesite concentrations are used to decide whatcleanup to perform, because the average valueis below the action level. If hot spots are found,their concentrations will be above the actionlevel and false negative (dirty) areas, as wellas false positive (clean) areas, are avoided; thatis, dirty areas are cleaned, but clean areas arenot.

At the Compass site, another strategy couldhave been to delay cleanup or to see cappingas an interim remedy until more work couldbe done to fully examine alternative treatment

29

technologies. The large amount of waste at thesite poses a difficult problem for which someinnovative cleanup technologies, including insitu techniques to avoid excavation, could havebeen considered and examined in treatabilitystudies.

Why the State and the community chose cap-ping, an impermanent and incomplete remedy,over incineration and groundwater treatment,which are permanent and complete, is not en-tirely clear. But avoidance of higher costs andpessimism about the safety of incineration seemto be the critical factors. Concerns about in-cineration can be addressed through effectivecommunication of state-of-the-art incinerationtechnology including effective pollution con-trol technology. The concern about enormouscost for incineration could have been addressedthrough either the hot spot or search for alter-native treatment strategies.

Actually, the groundwater contaminationproblem did not get enough attention, as cap-ping cannot completely stop further contami-nation. (The ROD had a tacit acknowledgmentof a carcinogenic risk factor of 1 in 100,000 forcontaminant migration through groundwaterseeps entering the Arkansas River. It is not cer-tain that the cap alone, or even in combinationwith groundwater cleanup, would permanentlyreduce this to EPA’s typical goal of limiting riskto 1 in 1 million.)

This site also illustrates a subtle and largelyignored issue in Superfund cleanups—cumula-tive risk. To what extent is a cleanup at onesite planned relative to neighboring cleanupsites that can contribute environmental risk tothe same population? Assessment of environ-mental risk at only one Superfund site seldomacknowledges human exposures from anothersite and, therefore, what seems for one site tobe a safe level of contamination, exposure, andrisk may not be so cumulatively. The CompassIndustries site and the Sand Springs petrochem-ical Complex Superfund site face each otheracross the Arkansas River. The migration ofcontaminated groundwater into the river fromboth sites would increase the danger to the samedownstream users.

Case Study 3Conservation Chemical Company, Kansas

City, Missouri, EPA Region 7

Capsule OTA findings. —A hydraulic containmentsystem to pump and treat some contaminatedgroundwater and capping of the site were cho-sen over the alternative of excavating and treat-ing contaminated soil and buried wastes, whichwas recommended in an EPA study and by theState. Water treatment cannot remove all thediverse contaminants at the site. No estimatewas said to be possible for the duration of thecleanup.

Key ~atm:

Entered Superfund system: 1/1/79Preliminary Assessment: 3/1/79Site Inspection: 3/1/79 - 11/1/80National Priorities List–proposed date: 4/1/85—final date: none—site rank: noneRIFS start and completion: Complex his-tory of studies by PRPs and EPAPublic comment period before Record ofDecision: 3/26/87 - 5/8/87Signing of ROD: 9/30/87Estimated complete remediation: nonepossible

Total time.—Unpredicted but probably a verylong time–decades.

Brlaf description of site.–” The site is approxi-mately 6 acres in size and is situated on thefloodplain of the Missouri River near the con-fluence of the Missouri and Blue Rivers, on theriver side of the levee. [The aquifer under thesite] is used as a source of drinking water byboth private residents and public water sup-ply companies.”

“Waste disposal operations began [in 1960]and continued until approximately 1980. CCCemployed a variety of waste handling practices,including but not limited to solvent incinera-tion, solvent resale, pickle liquor neutralization,cyanide complexation, chromic acid reduction,and ferric chloridelferric sulfate recovery. Re-sidual materials from the various treatmentprocesses were generally disposed of on site

30

in the basins. Drums, bulk liquids, sludges, andsolids were buried at the site. Some wastes, suchas drummed cyanide wastes and arsenic andphosphorus containing wastes, were disposedof on site without treatment. . . . approximately93,000 cubic yards of materials are buried onsite. ”

Major contamination/environmmtal threat.—There are21 substances “substantially in excess of appli-cable criteria or standards for water quality.These include six metals, cyanide, four pheno-lic compounds, and 10 volatile organic com-pounds (VOCS).” Other substances “cause con-cern for aquatic life.” Also, dioxin was detectedat levels up to 29 parts per billion (ppb). (A levelof 1 ppb has been EPA’s guideline for soilcleanup.)

The greatest risk comes from the use of con-taminated groundwater. Next is the risk fromcontaminated soils, which may be “transportedby precipitation runoff into surface water bod-ies or the groundwater. Contaminated soils alsopresent hazards from direct contact and winddispersion of particulate.”

T~he] groundwater is considered to be a cur-rent drinking water source since groundwateris used for drinking water within a two mileradius of the site.”

HRS scores .—groundwater 51.02; surface water9,45; air 0.00; total 29.99.

Removal actions.–None indicated in ROD; SCAPindicated over $2 million spent on removal,starting in 1985.

Cleanup remedy selected.—This remedial cleanupis the third selected for the site. The first clean-up, done by the original owner, was discoveredin 1985 to have failed. The State had approveda closure “which called for the addition of ab-sorbents and cementing materials to the wastein the uppermost 5 feet of each basin. Wasteacids, predominantly pickle liquor, and fly ashwere mixed with the upper layer of waste ma-terials in the basins. Tests conducted in 1985indicated that the desired pozzolonic cement-like properties have not formed. Also there areindications that this material has deterioratedand will continue to deteriorate. ”

The 1987 ROD indicated a previous ROD inmid-1985 that adopted a circumferential con-tainment approach with interior pumping. Butits implementation was stopped in 1986 whengeotechnical investigations found that the depthto bedrock ranged so high (to 160 feet) that “theconstruction of a circumferential impermea-ble barrier could be more difficult than origi-nally believed, ” (Neither SCAP or EPA’s mas-ter list of all RODS indicates an earlier RODfor this site.)

Many cleanup alternatives for the Conserva-tion Chemical site have been examined, andmost have been eliminated. Because the enact-ment of SARA came after the initial studies,EPA performed two more studies in 1987. How-ever, the current ROD evaluated only threemain cleanup alternatives in what is a well-structured and well-presented analysis: 1) the1985 remedy, 2) onsite containment of contami-nants by onsite pumping and groundwatertreatment, and 3) excavation followed by soiltreatment.

The 1987 ROD chose a remedy that includes:1) the use of a permeable cap to allow waterintrusion to assist groundwater cleanup; 2) awithdrawal well system to achieve an inwardgroundwater gradient; 3) a groundwater treat-ment system based on several unit operations,including “at a minimum, such treatment proc-esses as metals precipitation (utilizing bothhydroxide and sulfide precipitation), filtration,biological treatment, and carbon absorption”;and 4) offsite groundwater monitoring. Somedescriptions of the chosen remedy from theROD on the selected remedy are:

●

●

●

●

66. . . relies on hydraulic, rather than struc-

tural, containment to prevent migration ofcontaminants from the site. ”“Although designed primarily for contain-ing the on-site contaminants, [it] would alsoclean up a portion of the off site contami-nation. ”“while the treatment technologies that willbe employed provide high levels of treat-ment, they do not remove 100 percent ofthe contaminants. ”“This cleanup process could take a sub-stantial time period. ”

31

●

●

●

●

“There is no methodology available to esti-mate the length of time required forcleanup.”“This cleanup would include the dischargeof acceptable levels of contaminants re-maining in the groundwater after treat-ment to surface waters and the need to dis-pose of solid wastes resulting from thegroundwater treatment processes.”<6 .0. because an active pumping systemrelies upon the use of currently availabletechnology, which can be constructed inthe shortest time frame, this alternativewould provide expeditious implementationof the remedial action with substantial cer-tainty as to its effectiveness in protectingpublic health and the environment.”66 . . . may prove to be the least costlyremedy that would meet the environmentalgoals and requirements of CERCLA.”

Satisfaction of SARA statutory requirements:

1) Selection of permanent cleanup.—TheROD said the selected remedy “is cost-effective,consistent with a permanent remedy [and]applies permanent solutions and alternativetreatment technologies to the maximum extentpracticable.” On the issue of permanency: “Con-taminants present at the site will be containedat the site, thereby eliminating further uncon-trolled releases into the environment.” Also:“Treatment of hazardous substances to reducetheir volume, toxicity and mobility by treatingthe extracted groundwater is the principal ele-ment of [the selected remedy] . . . the uncertain-ties of [the soil treatment alternative’s] techni-cal feasibility at this site raise substantialquestion as to its practicability. Extensive re-search would be necessary prior to its imple-mentation to resolve this question. For thesereasons, [the selected remedy] offers treatmentto the maximum extent practicable.”

The selected groundwater treatment part ofthe remedy, however, does not involve destruc-tion technology to a large extent. Most of theunit processes are separation technologies. Theresult is the generation of hazardous sludgesrequiring management and the discharge ofcontaminants to the air. Moreover, no attemptwas made to estimate the duration of the ground-

water cleanup. Such estimates have been madeat other Superfund sites. If there is insufficientdata to make such an estimate, then there isa remarkable degree of uncertainty about thefunctioning of the groundwater pump and treatapproach.

The important feature of the selected remedyis that it does not directly deal with the con-taminated soil and buried wastes on site. Draw-ing water through the site, or flushing, is notlikely to remove all contaminants. Dependingon soil conditions and what chemicals arepresent, some contaminants are difficult to re-move by flushing. It is difficult to conceive ofwater drawn through the site’s hazardous ma-terials being able to dissolve or otherwise re-move all the diverse contaminants at the Con-servation Chemical site. And what will happenwhen the pumping is stopped? Indeed, the RODdid not claim complete removal of the site’s con-tamination.

The ineffectiveness of flushing was shownat tests at the Volk Field Air Force site in Wis-consin. EPA laboratory research on the use ofsurfactants to remove organic contaminantsfrom soil had been successful. However, thefield study found that in situ soil washing withaqueous surfactants “was not measurably effec-tive,” (U.S. Environmental Protection Agency,“Project Summary—Field Studies of In Situ SoilWashing,” February 1988.)

The selected remedy leaves a very largeamount of untreated hazardous material on thesite. The ROD said that protection against therisk posed by contaminated soil will be ad-dressed first by the permeable cap on the siteduring the pump and treat stage and secondby the placement of a RCRA cap upon comple-tion of the groundwater cleanup. Limiting siteaccess is also offered as a means of minimiz-ing risk. Regarding the dioxin contamination:“Since all the [dioxin] containing samples wereobtained from sludge and surface soil samples,the waste containment strategy and surface capwill minimize possible contact with TCDD. ”

A major issue for the Conservation Chemi-cal site is the rejection of the alternative of ex-cavation followed by soil treatment, which an

32

EPA contractor recommended in a specialstudy for Superfund’s enforcement program(Jacobs Engineering, “Analysis of AlternativeRemedial Action For The Conservation Chem-ical Co. Site, Kansas City, MO,” March 1987.).The study concluded: “The only treatmentmethod which would meet the environmentalprotection goal (of permanent removal or de-toxification of contaminants), and therefore theonly method likely to gain public acceptance,is excavation followed by soil treatment. There-fore, we recommend that excavation followedby soil treatment be considered as an alterna-tive treatment technology, under the require-ments of SARA, for implementation at the CCCsite.” Admittedly, the study dealt with soil treat-ment or washing in general terms. Moreover,even this treatment is a separation technologywhich would, like the selected groundwatertreatment, produce concentrated residues whichwould have to be managed. Overall, the studywas an excellent analysis of cleanup alterna-tives and carefully considered the pros and consof a number of options including in situ bio-reclamation, in situ soil flushing, and excava-tion followed by landfilling. The ROD includedthe recommended alternative but consistentlyevaluated it more negatively than the selectedremedy. Some relevant ROD comments sup-porting rejection of soil treatment are:

●

●

●

●

On reliability: “while [it] could be imple-mented, extensive testing and studies wouldbe necessary to verify this prior to imple-mentation. ”On implementability: “[it] applies a newtechnology and, as a result, there are sub-stantial uncertainties associated with im-plementation of this alternative which maytake considerable time to resolve beforethis alternative could be implemented”On technical effectiveness: “There are stillunresolved uncertainties associated with thetechnical effectiveness of [soil treatment].”On environmental concerns: The ROD notedthat there would be short-term impacts be-cause of excavation which could be mini-mized but not eliminated. “The option alsoinvolves the discharge of low levels of con-taminants and the generation of treatmentplant sludges requiring disposal.”

On safety: “The potential safety risks for[it] appear to be greater.”On public acceptance: “the alternatives aregenerally equivalent based on anticipatedpublic acceptance.” (Avery different state-ment than one in the Jacobs Engineeringreport.)On cost: The ROD’s estimated cost of theselected remedy is $21 million and for soiltreatment $24 million. “While there are anumber of uncertainties for each alterna-tive . . . [they are] the greatest for [soiltreatment].”On operation and maintenance: “[Soiltreatment], if feasible, should require a’sub-stantially shorter period of operation andmaintenance than [the others].”

Z) Accurate assessment of land disposal andcontainment alternatives.—Although there isa groundwater treatment component to theselected remedy, the cleanup rests on contain-ment of the hazardous materials that are thesource of the groundwater contamination.There was no significant analysis of the long-term uncertainties and possible failures of thecontainment and capping aspects of the clean-up. Considering the proximity of the site to bothsurface and groundwater, this lack of analysisis a major shortcoming of the selected remedy.

In the FS for the French Limited site in Texas,use of a slurry wall and cap to contain hazard-ous waste was described as a “temporary solu-tion” for which the “volume and toxicity wouldnot be affected . . . [and] . . . the potential wouldalways exist for failure of either the cap or theslurry wall allowing for the movement of un-stabilized wastes contained onsite. ”

RIFS contractor. —Information on the complexRIFS history is missing, However, the SCAPnotes that on 5/21/87 there was an EPA takeoverof the RIFS, but the takeover came after theRIFS reports were completed.

State concurrence.— The front of the ROD said:“The State of Missouri has been consulted onthe selected remedy.” In the responsivenesssummary at the end of the ROD, reference wasmade to a written comment by the Director,Division of Environmental Quality, Missouri

33

Department of Natural Resources: “The Mis-souri Department of Natural Resources recom-mends the alternative incorporating excavationof the wastes and soil washing with downgra-dient groundwater pumping and treatment beutilized for the remedial action at the CCC site.The commentor stated that this recommenda-tion is consistent with the final recommenda-tion contained in an EPA contractor’s reporton alternative remedial action technologies atthe CCC site.” The ROD contained a copy ofan internal EPA memo stating that EPA head-quarters and the Department of Justice supportthe selected remedy.

Cominunity acceptance. —The ROD said that therewas a “low level of community concern. Nomajor public concerns have been received atthis time.” One comment noted in the respon-siveness summary is from the Coalition for theEnvironment, Kansas City. The questions touchedupon the length of time for the groundwatercleanup.

Special comments. —The ROD says: “Total riskfrom all carcinogens should be between onein ten thousand to one in ten million. ” This isa very broad range which could mean that ac-tual risks associated with cleanup goals mightbe 100 times as great as the 1 in 1 million 1evelused most frequently by EPA.

The site’s HRS groundwater score seems lowin the context of the information in the RODon what contaminants are present and the useof the aquifer for drinking water.

General conclusions.–Contrary to what the RODconcludes, the selected remedy does not offera permanent remedy which effectively reducesthe toxicity of the site’s contaminants. Contami-nant volume would decrease somewhat throughgroundwater treatment, because groundwatermoving through the site would flush some con-taminants from the soil. Groundwater treat-ment, for the most part, removes some un-known amount of contaminants which maybelandfilled somewhere else or may be dischargedinto the air.

Although the soil excavation and treatmentalternative is not a true destruction approach,it is more consistent with the intent of SARA,

as the EPA contractor report also concluded.The cleanup selected for the Seymour Recy-cling site in Indiana sets a better example be-cause it includes two components to treat sitecontaminants in addition to the pump and treatcomponent for the groundwater.

The rejection of the soil excavation and treat-ment alternative seems to be based on uncer-tainty about its effectiveness. This uncertaintyexists because no treatability study was con-ducted as part of the RIFS. But this uncertaintymust be balanced against the uncertainties ofthe selected pump and treat remedy: How longwill water be pumped and treated? What con-taminants in what amounts will be removedand what will remain on site? How protectiveis the capon the site? What is the ability of thehydraulic containment system to prevent con-taminants from moving off site in the ground-water? Also, although the pump and treat canbe started sooner, the ROD acknowledged thatthe soil treatment remedy could be completedin a much shorter time.

Many uncertainties weaken the claim that theselected remedy is cost-effective. Even if a treat-ability study for soil treatment was successful,the selected remedy—with its comparable un-certainties—would not offer the same overalllevel of long-term environmental protection.Therefore, regardless of cost, it would not becost-effective.

The rejected soil treatment alternative ($24million) was estimated to cost about the sameas the selected remedy ($21 million). However,because soil treatment was only discussed ingeneral terms, its cost is highly uncertain, espe-cially when compared to the selected remedywhich uses off-the-shelf equipment, The esti-mated cost of the soil treatment alternative wasprobably underestimated. Indeed, for a simi-lar cleanup at the Chemical Control site the costof soil treatment was about the same as for Con-servation Chemical, even though the amountof material treated at Chemical Control (some20,000 cubic yards) was a small fraction of thatat Conservation Chemical (the ROD indicatedabout 100,000 cubic yards).

EPA’s contractor report on alternative reme-dial action at Conservation Chemical used a

34

figure of 298,000 cubic yards for contaminatedsoil to be treated and included about $5 mil-lion (out of about $20 million) for groundwaterpumping and treatment. Thus, the average costfor cleaning the excavated soil was about $0oper cubic yard, a very low price for any kindof contaminated soil treatment. (The ROD didnot contain any theoretical or experimentaldata to show that a very simple form of soilwashing could be effective in removing a di-verse set of contaminants.)

The cost of the rejected alternative can berecalculated in two ways. First, the unit costof the treatment can be altered and a range con-sidered for the amount of material to be treated.A figure of $5OO per cubic yard is comparableto an estimate for a similar cleanup in the FSfor the Chemical Control site and for the Re-Solve site. The FS for the Crystal City site hada cost of over $1,000 per cubic yard for a soilwashing alternative for both organic contami-nants and arsenic. And a company with a mo-bile soil washing technology, not applicable tometal contaminants, has indicated a cost of$450 to $1200 per cubic yard. (Tufts Univer-sity, “Transportable Treatment Unit Technol-ogies,” July 1986.]

The ROD acknowledges the uncertainty aboutonsite contaminants. If it is assumed that about100,000 cubic yards of material would be ex-cavated and treated (a figure consistent withinformation in the ROD), then, at a unit costof $5OO per cubic yard, the total cost would beabout $5o million, about twice what was esti-mated. For the figure of 298,000 cubic yardsfrom the 1985 and 1987 studies and the costof $5OO per cubic yard, the ● otal cost is $150million. Thus, the range is $5o to $150 million.

A second way to recalculate is to ask whetheran estimate closer to the ROD’s can be ob-tained? If a volume of contaminated materialhalfway between the two estimates in the re-ports is assumed (200,000 cubic yards) and thecalculation is based on a un”it cost of soil treat-ment at $2OO per cubic yard (between the op-timistic value of $5o per cubic yard and theabove $5OO per cubic yard), the total is $4o mil-lion. In this conservative scenario the cost isstill about twice that used in the ROD.

The attractiveness of the selected remedy,therefore, rested in part on its certain cost of$21 million relative to the underestimated costof $24 million for the soil treatment option. Thecomparable ROD costs appear to remove lowcost as a deciding factor, Would the selectedremedy seem less attractive from the SARA per-spective of preferring a permanent remedy ifthe soil treatment option was significantlyhigher in cost? True, there is a legitimate issuefor excavating materials and the risks associ-ated with it. But such excavation has beenselected at other sites because there are estab-lished techniques to mitigate such risks (e.g.,wetting materials to avoid dust). A technicalcase for not excavating materials, given for theSeymour Recycling site in Indiana because oflarge amounts of volatile chemicals, was notmade for Conservation Chemical.

Finally, this case may illustrate the lack ofmanagement oversight of RIFSS and RODS inthe Superfund program (even if the case doesnot indicate a high level of interest in reachinga settlement with the PRPs). Even a cursory ex-amination of the data for and of the uncertain-ties about the volume of treated material andcost would probably have spotted the cost un-der estimate for the soil treatment alternative.

Case Study 4Crystal City Airport, Crystal City, Texas,

EPA Region 6

Capsule OTA findings. —Excavation of contami-nated soils and wastes (which were buried ina previous removal action) and their disposalin an unlined landfill with a cap over it wereselected over incineration. No treatability studysupported the conclusion that the selectedremedy is permanent on the basis of the adsorp-tion of diverse contaminants to site soil. Majorfailure modes for the landfill were not ex-amined.

Key dates:

. Entered Superfund system: 8/1/83● pre~iminary Assessment: 3/1/87● Site Inspection: 9/1/84● National Priorities List

35

●

●

●

●

–proposed date: 10/84–final date: 6/86—site rank: #639 out of 770RIFS start and completion: 9/28/85 to7/13/87Public comment period before Record ofDecision: 8/10/87 to 9/14/87Signing of ROD: 9/29/87Estimated complete remediation: 8/89

Total time.—6 years

Brief description of site.—” The site is comprisedof approximately 120 acres of land. Surround-ing the airport property . . . is land used forgrazing animals . . . a municipal landfill . . . anelementary and high school as well as a resi-dential area . . . Since 1949 the city has oper-ated the facility as a municipal airport. Severalprivate companies conducted aerial pesticideapplicating businesses at the airport untiI1982.”

Major contamination/environmental threat.—’’The esti-mated volume of contaminated soil exceeding100 parts per million (ppm) total pesticide is12,000 cubic yards.” Although a large numberof contaminants have been detected, “The con-taminants of greatest concern at the site (tox-aphene, DDT, and arsenic) were chosen fromthe compounds detected based on their wide-spread distribution over the entire site as wellas the relative toxicity and concentration. ”There are also buried materials from an earlierremoval action and contaminated buildings. Di-rect contact, surface water, and air emissionsare major routes of exposure. The ,worst caseexposure scenario is for residents of a.nearbyhousing project.

HRS scores.—ground~ater 33.01; surface water12.92; air 43.08; total 32.26

Removal actions.— Immediate removal: 10/31/83to 11/8/83 for $33,000; 40 cubic yards of wasteand between 50 and 70 drums of material wereburied in two onsite landfill cells. Second re-moval: 4/24/84 to 4/25/84 for $25,000; 19 drums(FS says 21 drums) were buried in an offsitelandfill, the site was fenced, warning signs wereposted, and according to the FS: “eroded areasof the clay caps were repaired. ”

Cleanup remedy selected. —In addition to a num-ber of containment alternatives, incinerationand critical pressure fluid extraction wereevaluated. The ROD described the remedy as:“Onsite consolidation of alI material which ex-ceeds the health-based criteria of 100 milligramsper kilogram (mglkg) total pesticides. Placementof a RCRA cap over the consolidation cell. Mon-itor site for a minimum of 30 years followingconstruction bf selected remedy. Deep-well in-jection of decontamination liquids. Five yearreview of selected remedy. ” The argument wasmade that “By consolidating the contaminatedsoil away from the runway and taxiways, landuse could be maintained.” Estimated cost: $1.6miIlion.

Satisfaction of SARA statutory requirements:

1) Selection of permanent cleanup.—’’Theselected remedial action is considered perma-nent. consolidating this ‘naturally treated’waste under a hazardous waste cap is . . . con-sidered permanent. ” However, while the RODdid not say that an alternative treatment tech-nology was selected, which it was not, the RODdid suggest that alternative treatment would be“inappropriate,” No reduction in toxicity or vol-ume was claimed and the ROD correctly notedthat these reductions are “not a requirementof the [SARA] provision. ” The ROD said thatincineration ’did not conform with the Super-fund statute as well as the consolidation/cap-ping remedy” and that health and environ-mental protection is equal for incineration andconsolidationl/capping alternatives.

The ROD’s case for permanency for consoli-dation/capping rested on these facts:

●

●

●

●

". . . soils [are] characterized by high clay

content and extremely low permeabilities.”" . . . [the] aquifer is located 750 feet belowthe surface of the site and is isolated fromthe contaiminated surface soils of the siteby thick clay layers.”"

. . . contaminants are already highly im-mobilized and fixed within a solid soilmatrix."" . . . arsenic and organic pesticides [are]locked into [the] top foot of the alkaline soilsat the site. ”

36

Ž “The degree that contaminents are boundup is of the same degree that would havebeen achieved if the pure contaminants hadbeen’ processed by a solidification tech-nology.”

The permanence of consolidation/capping isuncertain for two reasons. First, no sound tech-nical case was made that all of the diverse rangeof contaminants would be adsorbed tightly tothe site soil. A treatability study could have beenconducted to demonstrate whether significantleaching of contaminants is likely. No liner willbe used to separate the waste from underlyingsoil. Data on the contaminants and the soil arepertinent. For example: “The primary indica-tor used to determine the degree to which anorganic contaminant binds to soil particles isthe organic carbon partition coefficient (Koc),.A higher Koc for an organic compound indi-cates a greater tendency to adsorb to organicparticles in soils, although migration may stilloccur throughout the site.” [Feasibility Studyfor the Renora site in New Jersey.) K.OC ‘data

bare in the FS for Crystal City ut were not fullydiscussed; site contaminants vary remarkably(by a factor of a million) and some contaminantshave low Koc values which suggest poor ab-sorption and the ability to migrate.

Mineral surfaces can also greatly affect themobility of an organic contaminant. A researchpaper that found no adsorption of phenol toa mineral has noted the problem of uncertaintyabout adsorption of organic compounds toclays: “Hemphill and Swanson found phenoladsorption on untreated kaolinite, montmoril-lonite, and illite. Others did not find any phe-nol adsorption on untreated clays. Luh andBaker, however, did discover significant ad-sorption by clays for substituted phenols . . . “(E.C. Yost and M.A. Anderson, Environ. Sci.Technol., vol. 18, pp. 101-106, 1984.) Anotherresearch paper that examined the interactionsbetween organic compounds and clay mineralsconcluded that in relation to ideal, laboratoryconditions “rates and selectivity maybe differ-ent and difficult to predict under environmentalconditions.” (E.A. Voudrias and M. Reinhard,Geochemical Processes, at Mineral Surfaces,ACS Symposium Series 323, September 1985.)

This uncertainty about adsorption is why atreatability study on actual site contaminantsand site materials would be necessary to ver-ify that some form of effective natural stabili-zation would take’ place at’ the site.

(’A complication at the Crystal City site is the

presence of solvents that can affect be adsorp-tion of other contaminants. EPA research con-cluded that” . . . the effects of solvents in haz-ardous waste contaminated soils may includetwo fadtors: 1) decrease in total sorption’to soils,

alnd 2) increase in leaching potential throughchanges in soil struture’U.S: EnvironmentalProtection Agency, Review Of In- Place Treat-ment Techniques for (Contaminated SurfaceSoils, vol. 2, Novembern1984] Also, the FS didnot say whether any work was done to ideni-tify.the forms) of arsenic at the site. The abil-ity of arsenic to remain immobilized becauseof adsorption to soil is not’straightforward. Ofthe two Chemical forms of arsenic, the moretoxic arsenite is more mobile,than arsenate andadsorption,is affected by the presence of cer-tain metals in the soil and by the pH which canchange overtime. (U.S. Environmental Protec-tion Agency, Review of In-Place TreatmentTechniques for Contaminated Surface Soils,vol. 2, November 1984.)

‘Second, not all the site’s contaminants areplanned to be consolidated. The FS indicatedthat only half the site’s contaminated soil mightbe capped. The choice of the cleanup criterionof 100 ppm of total pesticide in combinationwith the decision to continue to allow the siteto be used as a municipal airport is question-able. The ROD said that the cancer risk ap-proaches 1 in 100,000 for onsite exposure of22o days a year, which is possible for onsiteworkers and which is a higher risk than is the1 in 1 million usually sought by EPA. More-over, the FS indicated that a significant healththreat would persist if site use was not “limitedto 10 to 15 days per year.” This issue is impor-tant because of the absence of future land userestrictions.

Moreover, no cleanup criterion was estab-lished for arsenic, which is significant becausein correspondence to government officials EPA

— —

37

said: “Of the two types of contaminants, arsen-ic compounds predominate, are more toxic,and more persistent in the environment [thanorganic pesticides].” EPA also described arse-nic as “the most toxic contaminant” at the.site.Also, risks for inhalable dust particles maybeincorrect because such small particles were nottested to determine actual level of contami-nation.

On the subject of incineration, the ROD said:“Treatment ‘will not significantly reduce ‘themobility of the contaminants due to both thecharacteristics of the contaminants as well asthe impermeable nature of the soils." In fact,incineration would have offered more certain

permanence. The case against incineration isflawed for several reasons.

The ROD said: “A secondary treatment tech-nology (soil washing) would be necessary tore-move the arsenic compounds from the ‘treated’soil.” However, no consideration was given tothe proven feasibility of using chemical fixa-tion or stabilization for arsenic in incineratorash followed by landfilling. There is consider-able information to support this approach. InAugust 1987, EPA published extensive infor-mation in the Federal Register on effective treat-ment of arsenic in the context of the ResourceConservation and Recovery Act regulatory pro-gram. EPA said that “all the available data showthat the [Extraction procedure] regulatory levelof 5.0 [milligrams per liter] for arsenic can beachieved.” (52 Federal Register 29992; Aug. 12,1987.)

Treatability tests for solvent &traction andchemical fixation of arsenic contaminated soiland sediment from the Vineland Chemical Co.Superfund site in New Jersey have been suc-cessful. (These tests were done for the sameRIFS contractor as at Crystal City. The reportswere filed in December 1987, although the testswere probably planned and executed ‘muchearlier.) For some time, a commercial chemi-cal fixation company has made available ex-tensive data on the effectiveness of its treatmenton relatively high levels of arsenic in incineratorash. Treatment costs were said to be between$30 and $55 per ton. (Chemfix Technologies,Inc., testimony before House Subcommittee on

Transportation, Tourism; and Hazardous Ma-terials, Dec. 7, 1987.) Arsenic is a contaminantat the Tacoma Tar Pits site in Washington,where stabilization was, selected. Moreover,within the same EPA region as Crystal City,the ROD for the French Limited site in Texassaid: “The PCBs and arsenic can be controlledby stabilization of the treatment residues.”

Biological treatment for arsenic is anotheralternative that could have been examined ina treatability study. Then it would be unneces-sary to design a “custom [innovative) system, ”A recent report said: “Arsenic compounds tendto be converted by bacteria into volatile formsthat disperse to harmlessly-low concentrations. ”(R.U. Ayres, et. al., Toxic Chemicals, Health,and the Environment, The Johns Hopkins Uni-versity Press, 1987, pp. 38-70.) However, notall bacteria could treat arsenic and some de-velopment would probably be necessary.

A third alternative is a thermal treatment andrecycling facility in Louisiana which has beenused to treat a number of cleanup waste soilsand sludges. The treatment facility can handlevery large volumes of hazardous waste. Itsunique process would result in a residual ma-terial which appears to safely contain residualmetals, such as stabilization does. Moreover,the cost is reported to be relatively low; trans-portation costs must be added, but even thenthe total costs might be competitive to mobileincineration with the added advantage that thecleanup at Crystal City might be done quickly.

Moreover, OTA has examined a removal ac-tion at Southern Crop Services in Delray Beach,Florida, (not an NPL site) in which mobile in-cineration was selected for cleanup of the sametype of pesticide and arsenic contamination.The Florida kite has the same history as Crys-tal City Airport. EPA noted that a “naturallyrich organic layer near the surface of the soil”explains why the pesticides are concentratedand localized and why downward migrationof the pesticides into groundwater has beenslowed; (U.S. Environmental Protection Agency,Region 4, “Action Memorandum” for South-ern Crop Services site, Sept. 8, 1987.) As muchas 5,000 cubic yards of contaminated soil over2.5 acres will be incinerated at a maximum cost

38

of $2.5 million. The EPA document made nomention of any problem with arsenic in the in-cinerator ash. The cost of the incineration isestimated at $300 to $500 per cubic yard. Setupof the mobile incinerator was to begin in No-vember 1987 with completion in August 1988.An EPA Headquarters review on October 9,1987, of the memorandum by the Regional of-fice added the following: “It was determinedthat on-site incineration was the most suitablemethod to cleanup the site because it destroysthe hazardous waste and eliminates the needto transport the waste off-site.” This memoran-dum also added that, relative to offsite land dis-posal at half the cost of incineration, the “ad-ditional cost is considered reasonable becauseincineration provides a permanent solution.”Moreover, EPA said: “The disposal of the wastein a landfill represents a less permanent solu-tion to the problem than incineration, and istherefore less desirable in light of SARA em-phasis on more permanent solutions.”

At Crystal City Airport, the estimated costof the incineration option of $10.8 million mayhave been over estimated; $2.9 million was forthe arsenic treatment (about $200 per cubicyard). Without the special arsenic treatment,the cost of the alternative would be $7.9 mil-lion; thus the incineration option would costabout $575 per cubic yard (total, not unit cost),which is more than the cost range used for theFlorida site but which agrees with the FS datafor the Davis Liquid Waste site, where inciner-ation was also selected at a cost of about $600per cubic yard. If the expensive arsenic treat-ment could be substantially cut, the cost for theincineration alternative at Crystal City mightdecrease by more than $1.5 million. As notedabove, use of the facility in Louisiana wouldprobably reduce costs significantly more.

Moreover, there are instances where the RODwas biased against incineration and its com-monly accepted benefits. For example: “organiccontaminants may be reduced through an in-tegrated incineration system,” and “incinera-tion would remove the organic contaminantsfrom the solid” (emphasis added). The truth isthat incineration would definitely destroy theorganic contaminants.

Curiously, the alternative using chemicalstabilization was analyzed for cost assumingthat a lined landfill would be used for the treatedmaterial. The selected remedy uses an unlinedlandfill for untreated material. Use of a linedlandfill adds an extra $700,000,j a cost that wasincluded in the stabilization option but ex-cluded in the selected remedy of onsite con-solidation.

2) Accurate assessment of land disposal andcontainment alternatives.—“Failure of thisremedy is unlikely as long as proper mainte-nance of the cap is conducted. ” Nevertheless,many failure modes are possible but were notexamined, including the gross, disruption oflandfilled material (disposal cell will be about190 by 190 feet) and its dispersal due to an air-plane crash, perhaps with fire and explosion;a natural disaster such as an earthquake, flood,drought, and cracking of the soil that are appli-cable to the site; the uptake of contaminantsby biota, bioaccumulation, and ingestion by ani-mals in the food chain; the undetected or un-corrected erosion of the cap; greatly increaseddownward transport through highly permea-ble soils due to large-scale pathways such ascracks and root holes; and the perforation ofcap liners by animals and bugs and the subse-quent intrusion by water. Without consideringthese possibilities, the ROD overrates the tech-nical feasibility of the selected remedy,

This ROD and its FS also illustrate anothercommon problem in technology selection—thetechnical literature, including EPA’s own, israrely researched and cited to support conclu-sions. For example, in a recent’ report, EPAsummed up good practice with caps over land-fills: “Major storm events must also be consid-ered, since even an arid region can be subjectedto infrequent but major storms that cause anom-alous ground saturation and percolation to adepth ordinarily not reached. Accordingly, arather complete review of expectable stormevents and their frequencies should be requiredin preparing the background on the hydrologi-cal system.” (U.S. Environmental ProtectionAgency, “Project Summary—Design, Construc-tion, and Maintenance of Cover Systems for

39

Hazardous Waste: An Engineering GuidanceDocument,” November 1987.)

The ROD was ambiguous about the perma-nence of the selected remedy: “If however, fu-ture migration does occur, appropriate actionswill be taken. ” The term “significant unfore-seen offsite contamination” was also used.

In contrast, the ROD for the Pristine site wasrealistic: “The lifetime of a RCRA multilayercap is finite, and the contaminated soils willbe left in place to contribute to groundwatercontamination at some future time should thecap fail. . . . there are no data available on thelong term effectiveness and permanence ofRCRA caps.” In the. FS for the French Limitedsite (in the same EPA region as Crystal City),use of a slurry wall and cap to contain hazard-ous waste was described as a “temporary solu-tion” for which the “volume and toxicity wouldnot be affected . . . [and] . . . the potential wouldalways exist for failure of either the cap or theslurry wall allowing for the movement of un-stabilized wastes contained onsite. ” In compar-ison with Crystal City, these are prime exam-ples of inconsistency across Superfund sites.

RIFS contractor.—State led; $218,000 ($726,000obligated); Ebasco Services Inc.

State concurrence. —The ROD said the State “hasremained silent. ”

Community acceptance.—The ROD indicated thatthe community favored incineration. There isalso a lot of other evidence, because of a Con-gressional hearing in Crystal City on this issue(Apr. 11, 1988), that the community and othersstrongly opposed and continue to oppose theremedy chosen by EPA. A large number of lo-cal, State, and national government officialsand organizations have requested EPA to changeits decision.

Special comments. —The Preliminary Assessmentwas completed several years after the Site In-spection, according to CERCLIS; but SCAP in-dicates that the preliminary assessment startedon 9/26/84, slightly after the site inspection,

The first removal action (in 1983) that buriedhazardous materials set the precedent for an

impermanent remedy and contributed to theneed for remedial cleanup today.

Although the ROD said that “the organic com-pounds will continue to degrade under the capinto less toxic compounds,” no actual data oranalysis was given to support natural biodegra-dation under the conditions expected at the site.

The FS gave data that “suggests that the con-taminants are migrating offsite through water/sediment transport.” This observation meritsmore attention and an explanation of the ex-act mechanism of transport.

The case for concluding that RCRA is notapplicable was that “the contaminated mate-rial will be consolidated in the unit or area ofcontamination from which they originated. ”This conclusion is inconsistent with decisionsat other Superfund sites and means that cer-tain relevant aspects of RCRA on regulatoryrequirements for hazardous waste landfills,such as liners and leachate collection, were notapplied as required by SARA.

The data on contaminant detection frequencyin table 2 of the ROD were different than thedata given in the FS.

The ROD said no groundwater was encoun-tered, yet the HRS groundwater subscore is notzero.

General conclusions. —No sound technical casesupported the conclusion that containing thewastes onsite constitutes a permanent remedyaccording to the intent of SARA, All of the con-taminants may not bind tightly to the site soil,relevant regulatory requirements will not bemet, health risks may be greater than normallyacceptable levels, and a number of major fail-ure modes of the containment system were notexamined;,

The cost of the incineration’ alternative wasover estimated because of the residual arseniccontamination in the ash. In fact, stabilizationof such a contaminant has been successfullydemonstrated and is relatively low cost; bio-logical treatment is also known to be feasible.The advantages of incineration over the se-lected remedy for the organic contaminants

40

were discounted. Moreover, in comparison tothe decision to use mobile incineration at theSouthern Crop Services site with a nearly iden-tical type of pesticide and arsenic contamina-tion, the negative view of incineration at Crys-tal City Airport seems inconsistent and evencontrived.

Since incineration is proven for the organiccontaminants at the Crystal City site and “pro-vides better overall protection than consolida-tion/capping—contrary to the ROD’s claim th`atthe two choices are equal—a cost-effectiveremedy was not chosen. The justification usedby EPA for picking incineration at SouthernCrop Services, in terms of its greater benefitsover land disposal, particularly regarding per-manency of remedy, undercuts the evaluationby EPA at Crystal City Airport.

The FS analysis of treatment technologies forthe Crystal City Airport illustrates a nationwideproblem–current technology evaluations andthe decisions based on them are not explaina-ble by site-specific conditions. Several technol-ogies rejected for Crystal City could have beenjustified as well as they were at other sites wherethey were chosen (and are discussed in othercase studies in this report). For example, atCrystal City, in situ chemical stabilization wasrejected: “Immobilization, chemical treatment,and physical treatments have not been shownto be feasible for in situ treatment of these con-taminants as it is not possible to get a good,uniform, well distributed treatment.” (Thistechnology was selected for the Chemical Con-trol site in New Jersey and elsewhere.) Biologi-cal treatment was rejected: “[it] is generally in-effective for destroying these wastes as thetreatment is not performed in a controlled envi-ronment. Several processes are being developedwhich show potential However, none of theseprocesses have been developed past the labora-tory stage. Therefore, biological treatment hasbeen ruled out.” (This technology was selectedfor the Renora site in New Jersey.) In situ vitrifi-cation was rejected; it was selected for the Pris-tine site in Ohio.

The Crystal City site illustrates the problemof using a small number of indicator contami-nants not just for risk assessment but also for

technology selection. The effectiveness of somecleanup technologies depends on specific phys-ical properties which can vary substantiallyamong contaminants. There were a number ofcontaminants identified-at Crystal City that arenot likely to adsorb tightly to the soil. For ex-ample, toxaphene is far less likely to bind tightlyto soil than DDT. Both DDT and toxaphene posea problem’ for safeguarding water quality be-cause both have laboratory detection limitswhich are above their water quality limits. (R.H.Plumb and J.R. Parolini, “Organic Contamina-tion of Ground, Water Near Hazardous WasteDisposal Sites: A Synoptic Overview,” paperpresented at a Geological Society of Americaconference, Phoenix, October 1987.) Moreover,some chemicals—particularly solvents-may af-fect the adsorption of others present. Adsorp-tion of contaminants to soil was asserted tomake the case that containment was similar tosolidification treatment, but no analysis or treat-ability tests were made to confirm the hypothe-sis. Without such efforts, it is not reasonableto assume adsorption of all the contaminantsunder all future conditions.

Keeping the cost of remedial cleanup lowseems to have been an important goal. The RODindicated that no responsible party is availableto pay for, cleanup. The cost for the selectedremedy was estimated at $1.6 million, while in-cineration was estimated at $11.4 million, TheFS contained an unusual statement: “The costof a cleanup technology is also a factor of con-cern in the primary screening step.” Alterna-tives were kept in the analysis “if their esti-mated costs are not more than an order ofmagnitude higher than an alternative technol-ogy which performs to the same approximateextent.” Generally, FSS do not cut cleanup alter-natives from preliminary screening on the ba-sis of cost.

Invoking cost is done in the ROD when asound case can be made for equivalent envi-ronmental protection among different alterna-tives. Then, the issue of cost and when to esti-mate it takes on new importance because ofSARA’s requirements on technology selection.“It is difficult enough to estimate costs at thisearly [screening] stage of the feasibility study

41

when ‘old’ technologies are involved; it is hardlyprudent to try to estimate the costs of innova-tive technologies before a much more detailedanalysis (not to mention extensive pilot testing)is performed.” (D. Truitt and J. Caldwell; “Eval-uation of Innovative Waste Treatment Tech-nologies,” Waste Management Conference-Fo-cus on the West, Colorado State University,June 1987.) At Crystal City, the big differencebetween the RIFS obligation ($726,000) and theactual money spent ($218,000), if the data arecorrect, may also indicate that less work, suchas treatability studies, was done than could havebeen done and should have been done to bet-ter evaluate cleanup alternatives.

Case Study 5Industrial Excess Landfill, Uniontown, Ohio,

EPA Region 5

Capsule OTA findings.—Providing alternate waterto houses that have or are likely to have con-taminated wells was a satisfactory interim re-medial action. However, actions to address thesource of contamination and to stop and treatcontaminated groundwater are taking a verylong time.

Key dates:

Entered Superfund system: 12/1/80Preliminary Assessment: 12/1/83 (fromCERCLIS); 12/9/83 (ROD)Site Inspection: 8/1/84 (CERCLIS); 3/5/84(ROD)National Priorities List–proposed date: 10/1/84–final date: 6/1/86—site rank: #164 out of 770RIFS start and completion: 12/28/84 to 8/87(final focused FS)Public comment period before Record ofDecision: 8/12/87 to 9/10/87Signing of ROD: 9/30/87

Estimated complete remediation: 12/89 forthis interim remedial action. (According to theROD, it will take 5 years to design and imple-ment an aquifer restoration remedy. If a RODfor the final cleanup is issued by October 1988

as scheduled, the complete remedy would endin late 1993, but this estimate maybe optimistic.).

T o t a l t i m e . — 1 3 y e a r s

Brief description of site. —This ROD addresses anoperable unit or interim remedial action of theoverall remedy. “The Industrial Excess Land-fill is a closed sanitary landfill . . . From 1968to 1980 the site was operated ., . for the dis-posal of a variety of solid waste materials. Dur-ing this time, the landfill accepted municipal,commercial, industrial, and chemical wastesof substantially undetermined and unknowncomposition, primarily from the rubber indus-try in Akron, Ohio. Large quantities of chemi-cal and liquid waste were dumped onto theground either from 55-gal@ drums or fromtanker trucks. Although much of the liquidwastes were listed as latex and oil at the timeof disposal; witnesses have described the dis-posal of solvents and volatile industrial chem-icals with foul odors. ” The county ordered astop to the dumping of chemical wastes in Jan-uary 1972. It was not until 1980 that a courtordered closure and a closure plan was engi-neered and implemented; the site was coveredand seeded.

“The IEL [Industrial Excess Landfill] site islocated on a tract of approximately 30 acreswhich had previously been the site of miningoperations (sand and gravel and possibly coal).The landfill has a relatively pervious soil cover.”

Major contsfnination/environmental threat. -’’About80 percent of the site is believed to be under-lain by buried solid waste materials. There areover 400 residential homes located within a 0.5mile radius of the landfill. . . . over long periodsof time, the sand and gravel and immediatelyunderlying bedrock at IEL will act as a singleaquifer.” The landfill “is located in permeablesoils without an impermeable liner.”

Citizen complaints prompted testing in 1983that verified contaminated drinking water. ” . . .EPA discovered contamination of several pri-vate drinking water wells near the site. TheAgency determined that the cause of the con-tamination was the migration of hazardous sub-stances from the Industrial Excess Landfill. . . .contaminants have migrated approximately 600

42

feet from the western edge of the landfill, im-pacting the groundwater of 10 homes. Someof the residential wells sampled contained or-ganic contaminants (vinyl chloride and chlo-roethane) which are attributable to the landfilland inorganic contaminants (barium, copper,cadmium, and nickel) above background levels,also attributable to the landfill. In March 1987,U.S. EPA found levels of vinyl chloride and bar-ium exceeding federal drinking water stand-ards in approximately ten residential wells nearthe landfill.”

Contamination was also found in samplesfrom shallow monitoring wells onsite near thesite borders at “levels which exceed standards

the observed levels of vinyl chloride [2 to7 parts per billion (ppb)] in 3 of the 51 wellssampled are equal to or exceed the Safe Drink-ing Water Act Maximum Contaminant Level(MCL) of 2 ppb ..."

A risk assessment found risks greater than1 in 1 million excess lifetime cancer risk.

HRS scores.—groundwater 88.46; surface water0.00; air 0.00; total 51.13.

Removal actions.-EPA performed interim emer-gency actions to protect residents in the shortterm. The Superfund Comprehensive Accom-plishments Plan (SCAP) showed that workstarted on 12/2/85 at a cost of $973,000. T h eROD contained no summary description of ex-actly what was done, but there were indicationsof several actions, including air stripper treat-ment for contaminated groundwater, methaneventing, and some evacuation of houses. “Whilethe air strippers effectively deal with vinyl chlo-ride contamination, they will not remove otherhazardous substances, such as heavy metalsand semi-volatile organics, which threaten tomigrate from the IEL site. . . . the Agency de-termined to go forward with a permanent alter-native water supply, rather than continuing toproceed on a piecemeal basis with air strippers,whose long-term liability to protect publichealth cannot be guaranteed.”

Cleanup remedy selected.–’’Provide alternatewater to an area comprised of approximately

100 homes . . “ The cost was estimated to bearound $2 million.

“The primary objective . . . is to protect hu-man health by providing a reliable supply ofsafe, potable water to residents whose ground-water is currently contaminated or has the po-tential for being contaminated by IEL beforethe site itself is remediated. If unchecked, con-tamination will continue to migrate westward,affecting the groundwater of approximately 100homes in a 15 year time period. U.S. EPA ex-pects to implement a remedy for the IEL sitebefore contaminants can migrate beyond thisprojected area.”

Satisfaction of SARA statutory requirements:

I) Selection of permanent cleanup.—Becauseof its goal to provide alternative water as aninterim measure, this ROD did not examine orselect treatment technologies. “A permanentremedy at IEL will almost certainly involvesome sort of groundwater treatment to reducethe level of contamination.”

To support the plan to provide new water forhomes not yet contaminated, the ROD correctlystated that " . . . groundwater flow and contami-nant migration predictions are not exact sci-ences, and that predictions concerning the tim-ing and effectiveness of remedial action are notalways fulfilled ., . “

But why wasn’t anything done to stop themovement of the contaminated groundwater?Beyond actual source control or treatment andgroundwater treatment, several interim meas-ures would have been consistent with a final,permanent remedy. Examples include: 1) vacuumextraction and destruction of volatile organicchemicals from the site, 2) testing for and ex-cavation of hot spots of contamination, 3) in-stallation of a containment wall or barrier, 4)plume stabilization pumping, and 5) placementof a more impermeable cap or cover. Land userestrictions could also have been considered.Such actions might well have been taken earlier;the site has contaminated local water suppliessince it was closed by court order in 1980. Thesite’s HRS groundwater score, determined in1984, is exceptionally high.

43

The ROD granted that “The statutory prefer-ence for treatment is not satisfied because thisaction constitutes an operable unit for the over-all site remedy. Treatment alternatives for theoverall site will be addressed in the compre-hensive RI/FS documents.”

Z) Accurate assessment of land disposal andcontainment alternatives.—The ROD did notconsider these types of alternatives, except thata de facto no action on the buried wastes meantthat the original land disposal remained ineffect.

RIFS contractor. —Camp Dresser and McKee;about $1 million obligated.

State concurrence. —The State of Ohio concurredwith the selected remedy.

Community acceptance. —It was the communitythat called on government officials to take seri-ous action in the first place. For this ROD, in-tense community concerns focused on gettingnew water for more houses and on identifyingthe exact source of the water.

A local newspaper, the Akron Beacon Jour-nal, November 15, 1987, reported: “Recently

the EPA agreed to connect 110 homes inUniontown to a new water supply when 1,500were in possible danger. Fortunately, Gov.Celeste saw no reason to threaten the rest andagreed to spend state money to build a newwater system for all the homes. Gov. Celesterightly saw that any error should be on the sideof safety. The federal EPA’s approach seemsto be the opposite. A telling example was therelease of documents showing the air tests hadbeen bungled. The documents became publiconly after citizens invoked the federal freedomof information law to obtain them. Informationabout mistakes in the air tests had been knowto the EPA since Oct. 21. A responsible EPAwould have informed the public immediately. ”

Special comments.–Any ROD signed on 9/30/87,the end of the Federal fiscal year, may have beena rush job; the ROD for IEL may have suffered

accordingly. It contained an incomplete sum-mary of the administrative record, and it didnot decide on the actual alternate water source,a contentious issue locally. EPA was respond-ing to community concerns by deferring thewater decision.

In the responsiveness summary, EPA madea very interesting statement on cost-effective-ness: “U.S. EPA is required to select the lessexpensive alternative, given that effectivenessand implementability are equal” (emphasisadded). Adding some factor other than effec-tiveness is new. It could, however, act againstselecting newer cleanup technologies, whichusually have not been used on a large scale.

The argument for not relying on air stripperswas technically sound in view of the chemicalcomplexity of the site contaminants.

General conclusions. —It probably will be closerto 20 years than 13 between the time the IELsite was closed and the time some form ofsource control and groundwater treatment isapplied to the site. Is EPA correct that the se-lected interim remedy “is fully consistent witha permanent remedy?” Although the ROD re-ferred to aquifer restoration, it does not men-tion removal of the source of contamination.While providing alternate water does not standin the way of a permanent remedy, neither doesit do anything to make the permanent cleanupeasier, Indeed, this ROD acknowledged that thegroundwater contamination will get worse.Thus, because additional interim measureswould stop or slow down the migration andhelp to alleviate the source of the problem, thisinterim measure is not fully consistent with anultimate permanent remedy.

The assertion that it will take 5 years to de-sign and implement an aquifer restoration rem-edy was probably too optimistic and is notlikely to build community confidence. Experi-ence at similar sites suggests that groundwatercleanup can take considerably longer. For ex-ample, at the Seymour Recycling site in Indi-ana the groundwater treatment was estimated

44

to take from 28 to 42 years; at the Davis LiquidWaste site in Rhode Island, it was estimatedto take from 5 to 10 years; and at the Re-Solvesite in Massachusetts, it was estimated to take10 years. There is, however, no obvious alter-native for groundwater cleanup at IEL, EPAhas said: “The actual performance of a groundwater remedial action is difficult to predict untilthe remedy has been implemented and opera-tional data have been assessed.” (U.S. Environ-mental Protection Agency, “Guidance on Reme-dial Actions for Contaminated Ground WaterAt Superfund Sites,” draft, October 1986.)

Before a permanent remedy is implemented,more wells may become contaminated. Also,contaminants can migrate at substantially dif-ferent rates, and, therefore, the nature of thespreading contamination can change over time.Such changes are well verified by research.(See, for example, R.L. Johnson et al., GroundWater, September/October 1985.) Wells that arecontaminated early can get worse as new, moreslowly moving contaminants reach them; wellsnot yet contaminated eventually see the effectsof the most rapidly moving contaminants, Onthis point the ROD noted: “Since the publica-tion of the FFS [Focussed Feasibility Study], re-cent data revealed that levels of nickel exceededAmbient Water Quality Standards.” Also, “vi-nyl chloride has migrated off-site quickly . . .while its parent compounds pose a threatenedrelease from the site because they are migrat-ing at a slower rate. “ The ROD also noted: “theshallow and deep aquifers are continuous andlinked to one another.” Therefore, the complex-it y of groundwater contamination and its clean-up could worsen significantly. The possibilityof upgradient chemical migration should notbe ruled out. (See R.H. Plumb, Jr., ProceedingsSecond Canadian/American Conference onHydrogeology, 1985, pp. 69-77.)

The long history of the site, the extensivegroundwater contamination offsite, and the de-lay in addressing the source of the problem isfeeding community lack of confidence in gov-ernment efforts and demands for new waterto more houses.

Case Study 6Pristine, Inc., Reading, Ohio, EPA Region 5

Capsule OTA findings.—In situ vitrification wasdeveloped originally for radioactive soils, butits use for chemically contaminated sites is stillunproven. Without treatability test results, insitu vitrification was selected for this site chieflybecause its estimated cost was about half thatof onsite incineration. But the estimated costfor incineration is probably high by a factor oftwo. Incineration offers more certainty andprobably costs no more than the selected remedy.Groundwater will be pumped and treated byair stripping and carbon adsorption.

Key dates:●

●

●

●

●

●

●

Entered Superfund system: 4/1/79Preliminary Assessment: 1/1/83Site Inspection: 9/1/82National Priorities List–proposed date: 12/1/82–final date: 9/1/83—site rank: #531 out of 770RIFS start and completion: 9/5/84 to 11/87Public comment period before Record ofDecision: 11/13/87 to 12/11/87Signing of ROD: 12/31/87

Estimated complete remediation: 8/91; 2 yearsfor source control, 5 to 10 years for ground-water cleanup (August 1991 is given in ROD,but this seems optimistic and inconsistent withother groundwater cleanups; 10 years for ground-water cleanup is more realistic)

Total time.—2O years

Brief description of site. —The site is in a suburbof Cincinnati. The site is 2.2 acres and “is bor-dered by residential and industrial areas. Thereare two aquifers under the site.” In the late1970s, a liquid waste incinerator was operatedat the site. “In April 1979, as many as 8,000to 10,000 drums and several hundred thousandgallons of bulk liquids were on site, consistingof acids, solvents, pesticides, PCBs and otherchemicals. ” A consent order shut down the fa-cility in September 1981.

45

Major contamination/environmental threat.—” . . . over90 compounds were detected in the ground-water, soil, sediment, and surface water. ”

“Groundwater in the upper aquifer is con-taminated primarily with volatile organic com-pounds (VOCs) such as benzene, vinyl chloride,tetrachloroethene (TCE), and 1,2-dichloroethane.Semi-volatile compounds (semi-VOCs) and pes-ticide compounds occurred in relatively lowerconcentrations. The lower aquifer is contami-nated with benzene and 1,2-dichloroethane.There are also elevated levels of lead andfluoride.”

" . . . the presence of VOCs in the [Readingmunicipal wells] indicates that the groundwaterquality in the vicinity is compromised and con-tinued monitoring is recommended . . . Thelower aquifer is the source for the regionalwater supply [13,000 people]. ” This route of ex-posure poses the largest risk.

“Sediment . . . and soil in the upper two feetof the site are contaminated with a variety ofVOCs semi-VOCs, and pesticides. Principalcontaminants in surface soils are benzene, di-eldrin, and DDT.” Low levels of dioxins andfurans were also found.

“Subsurface soil contained VOCs . . . Therewere also elevated levels of cadmium, lead, mer-cury and zinc. ”

“Surface water was contaminated with VOCs,semi-VOCs and pesticides . . . There were alsoelevated levels of inorganic compounds (cad-mium, chromium, and mercury).”

A good risk assessment established cleanupgoals at 1 in 1 million risk level. The RIFS calcu-lated how much soil would have to be removedto “eliminate both the risk associated with ad-sorption and ingestion of soils and ingestion ofgroundwater contaminated through leachingfrom the soil.”

HRS scores.—groundwater 60.00; surface water10.91; air 0.00; total 35.25

Removal actions. —The site operator removedwaste from June 1980 to November 1983 un-

der a consent decree. Some responsible partiesremoved waste and soil from March 1984 toJuly 1984 under an administrative order. TheROD did not say how much material was re-moved, nor its disposition, but it was probablylandfilled.

Cleanup remedy selected. —This ROD was a finalsource control remedial action but also in-cluded groundwater cleanup. Another RODmight be issued for additional groundwatercleanup.

The key component of the selected remedyis in situ vitrification (ISV) for 37,700 cubicyards of contaminated soil and sediment. ISVwas chosen over onsite incineration. ISV is aninnovative technique that uses electrodes in theground to pass electricity through soil, melt it,vaporize and at least partially destroy organicchemicals, and leave in place a chemically in-ert, stable, glass-crystalline mixture. Temper-atures in the range of 2,000 to 3,6000 F are pos-sible. Different cells of soil are melted in orderto cover a site. The melt grows downward andoutward as power is applied. As the vitrifiedzone grows, it incorporates nonvolatile elementsand destroys organic components by pyrolysis.The pyrolysis byproducts migrate to the sur-face of the vitrified zone, where they combustin the presence of oxygen. “The estimated timerequired to complete the vitrification processis two years assuming the use of one vitrifica-tion unit.”

Groundwater will be pumped and treatedwith an air stripper and carbon adsorption.That is, separation, not destruction, technol-ogy was selected. Groundwater monitoring wasset up. The possibility of deed restrictions wasraised. "It is estimated that it will take five toten years to extract and treat the contaminatedgroundwater,”

Estimated cost: $22 million.

Satisfaction of SARA statutory requirements:

1) Selection of permanent cleanup.—In its ini-tial two screening stages, the FS examined alarge number of treatment technologies. How-

46

ever, after more detailed screening, many ofthe treatment technologies said to be applica-ble were dropped without much justification.For example, three in situ treatment technol-ogies passed the initial screening, but solutionmining and soil vapor extraction were droppedand only vitrification was retained for moredetailed analysis. Onsite treatment technologiesthat passed the initial screening included fixa-tion/solidification, soil washing, and dechlori-nation, but only incineration was analyzedfurther.

The ROD said that the selected ISV remedy“will significantly reduce the mobility, toxic-ity, and volume of hazardous substances in thesoil through treatment. The mobility of the con-taminants will be reduced significantly, suchthat no leachate is expected to be producedfrom the vitrified material. This is a permanenttechnology, the results of which are expectedto last for a million years. The toxicity of or-ganic components will be decreased becausethe organics are destroyed or changed to otherforms by pyrolysis or vaporization. The volumeof the soil will be reduced by 25 to 30 percentbecause the vitrification causes the soil massto consolidate.”

A chief issue is whether or not ISV is a proventechnology. ISV is an alternative treatment andan innovative technology that was developedoriginally for treating radioactive contaminatedsoils, but its use for chemical contaminationraises new questions. How should ISV be clas-sified? Is it thermal destruction or stabilization?EPA’s SITE technology demonstration pro-gram categorizes it as stabilization, as do others.(N. Nelson et al., Toxic Chemicals, Health, andthe Environment, The Johns Hopkins Univer-sity Press, 1987, pp. 205-279.) Stabilization isa reasonable label because metal contaminantsremain in the final glass-like material and be-cause the leaching of metals and the completedestruction and removal of organic contami-nants are uncertain. Although very high tem-peratures are reached, not all organic contami-nants will either be destroyed or be able toescape and be captured. However, EPA alsocalls ISV thermal destruction. (U.S. Environ-mental Protection Agency, Technical Resource

Document: Treatment Technologies for Halo-genated Organic Containing Wastes, vol. 1, Jan-uary 1988.) This EPA report describes ISV as:“Not commercial, further work planned. No[performance] data available, but DREs [de-struction-removal efficiencies] of over six ninesreported. ”

ISV’s developer, Battelle Pacific NorthwestLaboratories, which supports ISV’s inclusionin SITE, has commented on treatability test-ing: “While the results are promising, feasibil-ity testing to confirm applicability is stronglyrecommended prior to any commitment to de-ploy the process on a site that contains signifi-cant quantities of organics that are unconfinedin the soil column. . . . feasibility testing is rela-tively inexpensive [a few thousand dollars]. Thefocus of the feasibility testing is the perform-ance requirements for the off-gas treatment sys-tem and the type and quantity of secondarywaste generated. Experience with low boilingpoint organics that are uncontained in the soilcolumn is very limited, and feasibility testingwith actual site samples prior to application isstrongly recommended” (emphasis added).(V.F. Fitzpatrick, “In Situ Vitrification–A Can-didate Process for In Situ Destruction of Haz-ardous Waste,” Proceedings of the 7th Confer-ence on the Management of UncontrolledHazardous Waste Sites, December 1986, pp.325-332.)

ISV depends on the effectiveness of the col-lection and treatment system for released gasesto keep undestroyed organic contaminants (orproducts of incomplete combustion) from en-tering the environment, This off-gas system islike a separation technology; hazardous residuescan be either destroyed or landfilled after car-bon adsorption. The greater the volatility of con-taminants, the greater their release into the off-gas collection system. At Pristine, many of thecontaminants are highly volatile at relativelylow temperatures. By the time the soil is melted,therefore, many contaminants have moved.

What happens to organic contaminants inISV is crucial to understanding its cleanup ef-fectiveness relative to other technologies, suchas incineration. A published paper reported on

47

a test of ISV on PCB contaminated soil: “Smallquantities of PCBs, furans, and dioxins weredetected in the untreated off-gas, but none weredetected in the vitrified mass. A few samplesdirectly adjacent to the block contained meas-urable concentrations up to 0.7 ppm PCBs. ”(R.R. Battey and J.T. Harrsen, “In Situ Vitrifi-cation for Decontamination of Soils Contain-ing PCBs,” Proceedings of the Oak Ridge ModelConference, February 1987, pp. 233-245.) Inanother report on the same experiment, the“process destruction was slightly greater than99.9 percent. The small amount of material re-leased to the off-gas system was effectively re-moved, yielding an overall system DRE of>99.9999 percent.” (V.F. Fitzpatrick, “In SituVitrification—A Candidate Process for In SituDestruction of Hazardous Waste,” Proceedingsof the 7th Conference on the Management ofUncontrolled Hazardous Waste Sites, Decem-ber 1986, pp. 325-332.)

The ROD responsiveness summary said:“with worse case conditions, 97 percent of allorganics are destroyed. Most tests indicate that99 to 99.99 percent destruction is achieved.”Another ROD statement is more optimisticabout destruction versus removal: “The test re-sults [on PCBs] indicate that the organics aredestroyed and not merely collected in the off-gas system.” All this information shows thatISV might be very effective but that the issueof total destruction of organics through boththermal treatment and off-gas collection, re-moval and possible treatment needs clarifica-tion. Lateral migration of vaporized organicsinto adjacent soil or perhaps downward intogroundwater is also important and needs de-tailed resolution for application of ISV to anylarge, uncontained site.

This last issue has received major attentionby Larry Penberthy, who calls it vapor retreat.While Penberthy is a competitor of ISV, hemakes a good technical arguement: “Insteadof being destroyed, the vaporizable chemicalcontaminants simply move away from the hotcore melt by Vapor Retreat, unaltered. Theymove downwardly below the melt core as wellas horizontally away from the melt core. Thisvaporizing/condensing action is progressive,

building up concentration in the isothermlayers corresponding to each chemical’s boil-ing point. This writer expects the DRE to beonly 25-50 percent.” (Larry Penberthy, letterto Laura A. Ringenbach, attorney for respon-sible parties, Mar. 28, 1988; Pyro 32A and 32newsletter of Penberthy Electromelt Interna-tional, Inc., Apr. 7 and 13, 1988.) This vaporretreat phenomenon could lead to increasedcontamination of groundwater. Moreover, inorder to test for this effect it would be neces-sary to test a rather large volume of soil so thattemperatures away from the molten zone arelow enough to have condensation of vaporizedcontaminants, However, most testing is doneon too small a volume of contained materialto see this effect.

Penberthy has a number of other criticismsof the tests performed by Battelle which, evenafter examination of Battelle’s comments onPenberthy’s analysis, seem important enoughto require additional study and testing. More-over, Penberthy has raised important safetyquestions, such as effects from soil heating andsubsidence, about using ISV at such a heavilyindustrialized area as the one around Pristine.No significant examination of the risks posedby ISV has been made.

The ROD did not focus on the depth of ISV.The plan is to go down to 8 feet for half thesite and 12 feet for the other half. The Batteyand Harrsen article (see above) noted that thegreatest efficiencies for ISV occur when it isused to depths of 10 to 20 feet. The technologydoes not work well when contaminants are onthe surface and therefore, soil covers are some-times used, The ROD also noted: “The equip-ment must be specially designed and produced, ”The depth and equipment issues are possiblecauses of underestimated costs.

The other big issue for the Pristine site is therejection of onsite incineration. The ROD ac-knowledged: “incineration is a proven technol-ogy. “ “Incineration ., . is fully protective of hu-man health and the environment since theingestion and leachability threats are elimi-nated.” The ROD did not acknowledge thatsome stabilization of the incinerator residues

48

might be necessary because of toxic metal con-taminants. But overall, the ROD did not noteany disadvantages of incineration: “The use ofmobile incinerators is common and the per-formance of these systems has been demon-strated. It is relatively easy to operate the systemalthough a trained operator will be needed.”

Nevertheless, incineration was not selected.The ROD said: “vitrification is the lower costalternative. Therefore, incineration is not rec-ommended for implementation at the Pristine,Inc. site.” However, the costing for incinera-tion seems too high. The ROD’s total cost esti-mate of $51 million for incineration was basedon a unit direct cost of $730 per cubic yard (for37,700 cubic yards). Meanwhile, the FS said thatthe unit cost ranges from $350 to $500 per cu-bic yard. Several other recent FSS (for the DavisLiquid Waste and Re-Solve sites) by the sameRIFS contractor provided detailed vendor costsand analysis for onsite incineration. From thosetwo FSS, OTA used the cost data for three differ-ent technologies for a range of contaminatedsoil to be treated (4,300 to 57,000 cubic yards)and obtained a (conservative) estimate of a unitcost of about $300 per cubic yard for the levelof effort at Pristine. This range is consistentinformation that other vendors gave to OTA.

EPA said recently that mobile infrared in-cineration of contaminated soils costs “from$120 to $225 per ton [which could be as highas $180 to $340 per cubic yard], depending onthe number of tons incinerated per day.” (U.S.Environmental Protection Agency, memoran-dum from John H. Skinner, Office of Research& Development, Dec. 10, 1987.) The FS for Sey-mour Recycling gave costs for onsite incinera-tion over a very broad range of amounts of con-taminated soil: for 35,000 cubic yards the costwas $186 per cubic yard; at the smallest scale(18,000 cubic yards) the unit cost was $349 percubic yard. In the Crystal City FS the unit costfor onsite incineration was $240 per cubic yardfor about half the amount of material at Pris-tine. In a recent decision for a Superfund re-moval action at the Southern Crop Services sitein Florida, where mobile incineration was se-lected, EPA said that it expected bids at from

$300 to $500 per cubic yard for less than 5,000cubic yards of soil.

Actual (bid) costs for incineration can varybut do not explain the high estimate used forPristine. A recent report showed a cost of about$750 per cubic yard (comparable to the Pris-tine ROD estimate) for a cleanup, but the soilquantity was under 10,000 tons and there wasmore than just soil to clean. (J.F. Frank et al.,“Use of Mobile Incineration to Remediate theLenz Oil Site,” paper presented at Superfund’87, 1987, pp. 459-464.) At another site, a ven-dor (a subsidiary of the Pristine FS subcontrac-tor) got $250 per ton under a turnkey arrange-ment; total costs probably were about $450 percubic yard for cleaning between 7,500 and10,000 tons of PCB contaminated soil. (J.W.Noland, remarks at Weston Environmental Fo-rum, Washington, DC, February 1988.) Mobileincineration was used to incinerate materialsat the Nyanza Superfund site in Massachusetts.The vendor charged about $600 per cubic yardfor a very small quantity, about 200 cubic yards.At the Prentiss Creosote Superfund site in Mis-sissippi, a vendor charged about $200 per ton($300 per cubic yard) for mobile incinerationof 7,500 tons; at the Southern Crop ServicesSuperfund site in Florida, the vendor charged$360 per ton for 3,000 tons,

The Pristine ROD cites no technical factorsto explain a $730 per cubic yard cost the site.For example, no mention has been made ofburied drums. Even if the ash were to be chem-ically stabilized because of toxic metal content,the additional cost would not account for thecost discrepancies noted above. Moreover, inan internal inconsistency the FS calculation fora cleanup of only 8,100 cubic yards used a unitcost of $658 per cubic yard; instead of the ex-pected higher unit cost for a smaller volume,a lower figure was used.

In the groundwater cleanup, the ultimate dis-position of the collected hazardous substancesis uncertain because it is not clear how the car-bon that becomes contaminated by removingorganic substances will be managed. The RODsaid: “Bench scale studies will be done to de-

49

termine the need for metals treatment.” Airstripping only removes volatile organics andcarbon adsorption is not likely to be effectivefor the metals. Maximum reported values oflead in groundwater are 178 ppb and 148 ppbin the upper and lower aquifers; the drinkingwater Maximum Contaminant Level (MCL) forlead is 50 ppb and the proposed MCL Goal is20 ppb. For cadmium the corresponding ground-water levels are 39 ppb and 9.4 ppb and thestandards are 10 ppb and 5 ppb. The complex-ity and intensity of groundwater contaminationare great enough to warrant more detailed anal-ysis of groundwater treatment, as was done forthe Operating Industries site in California.

Any cleanup of contaminated surface waterat Pristine is left uncertain by the ROD.

2) Accurate assessment of land disposal andcontainment alternatives.-The ROD containedan excellent rationale for not leaving contami-nated material onsite: “It may leach into thegroundwater at levels that will exceed ARAR’s[regulatory standards] at some future time andthus increase the groundwater treatment timeor require additional future remedial action.The lifetime of a RCRA multilayer cap is finite,and the contaminated soils will be left in placeto contribute to groundwater contamination atsome future time should the cap fail.” Theresponsiveness summary said: “there are nodata available on the long term effectivenessand permanence of RCRA caps.”

By its very nature, ISV leaves treated contami-nated material onsite. The ROD acknowledgeduncertainties about ISV:

●

●

●

“Vitrification is expected to be effective onthe soil type present. . . “(emphasis added).“Air monitoring will be conducted to en-sure that the hood is collecting and treat-ing the gases.” (Such monitoring is not rou-tine when mobile incinerators are used.)“Some limited monitoring of the vitrifiedmass will be required to assure that it isa reliable and permanent remedy.” (Suchmonitoring is not routine when stabiliza-tion is used.)

●

●

●

●

“[There is] limited demonstrated perform-ance. 99.9999 percent DRE [destructionremoval efficiency] are expected for dioxinand PCBs.”“Because this is not a proven technology,prior to implementation of this remedialaction, bench and/or engineering pilotscale studies will be required to confirmthe effectiveness and applicability of thistechnology to site conditions.”“If this treatment method is found to beineffective, this Record of Decision mayneed to be reopened.”“Monitoring will be conducted during thetreatment process to determine if contami-nation is migrating through the soil as aresult of the treatment.” (Such monitoringis not routine with in situ techniques.)

ISV is not easy to implement—at least, thetechnology costs more—when water content ishigh. Thus the ROD noted: “Because of con-cern over the effectiveness of vitrifying the up-per outwash lens, consideration will be given,during these bench and/or pilot studies, towhether the lens should be drained prior tovitrification. ” Also, the responsiveness sum-mary said twice that the site’s soil has “highmoisture content” when it defended why vacuumextraction of VOCs was not feasible. However,when the selected ISV remedy was defended,the responsiveness summary—four pages later—said “the moisture content . . . is not high. ”Either high is high, or actual measured valuescould be used to show it is high for one tech-nology but not too high for the other technol-ogy, if that was the case; however, no actualdata were used. (This may illustrate a lack ofROD quality control and ROD rushing at theend of a fiscal year quarter.)

To its credit, the Pristine ROD specified: “TheToxicity Characteristic Leaching Procedure(TCLP) is the testing mechanism that shouldbe used to verify the complete treatment. If thistreatment method is found to be ineffective, thisRecord of Decision may need to be reopened. ”However, a recent technical report said thatvitrified contaminated soil performed poorly:

50

“The vitrified product as evaluated with thesestandard leaching tests did not perform well.The reason for this is not known. It maybe thatthe nature of the tests maybe inappropriate formonolithic, vitrified masses, or vitrificationmight not be as effective as chemical stabiliza-tion for simple metal systems.” (J.J. Barich etal., “Soil Stabilization Treatability Study at theWestern Processing Superfund Site,” paper pre-sented at Superfund ’87 conference, 1987, pp.198-203.) According to the report, for example,leaching of zinc–over a short period in a stand-ard EPA leach test—was about 10 times greaterthan for conventional stabilization. The respon-siveness summary portrayed much more cer-tainty: “the metals are encapsulated and boundup in the ISV process.” The responsivenesssummary also said: “ISV has been tested on haz-ardous waste and has been successful.” Thesetwo statements contradict the Barich findings.

Battelle has said that the TCLP is technicallyinappropriate for monolithic waste forms.Their principal concern is that reducing treatedmaterial to fine particles—and exposing un-bonded contaminants–is based on the false as-sumption that treated material may not main-tain monolithic properties. Does this mean that,if Battelle is successful in making its point aboutthe TCLP, the Pristine ROD will be reopenedbecause the TCLP will not be used to test ISV’seffectiveness?

A recent EPA study that examined eightemerging treatment processes for decontami-nation of PCB contaminated sediments rankedISV last by using two sophisticated methodol-ogies. The report said: “all the processes ex-cept In Situ Vitrification appear to merit fur-ther development for this application.” (U.S.Environmental Protection Agency, “Report onDecontamination of PCB-Bearing Sediments,”October 1987.) While PCB contamination is notthe dominant problem at Pristine and sedimentspose a special problem for ISV unless they aredewatered, this report is important becausealmost all previous information on ISV hascome from its developer, including a lot of em-phasis on tests on PCB material. The PristineROD said: “An additional application [of ISV]is being planned by EPA for a PCB contami-

nated site.” In August 1987 it was reported thatEPA Region 5 had “conditionally accepted”ISV for an emergency response action for PCBcontaminated materials at the Greiner’s Lagoonsite in Fremont, Ohio. (Hazardous Waste News,Aug. 24, 1987.) As of May 1988, OTA was in-formed by EPA that no date had been set forthe test—the actual removal action at the site—and EPA confirmed that this site action wouldconstitute the ISV test for EPA’s SITE program.Only if new test data confirm the presence ofhigh concentrations of PCBs will ISV be usedat the Greiner’s Lagoon site and, even then,probably not before Spring 1989. If the PCBsare low, then another site would probably beselected, delaying the SITE demonstration stillmore. The Pristine responsiveness summarysaid that the demonstration will be performedprior to use of ISV at Pristine.

Failure of the vacuum off-gas collection sys-tem is possibile and of concern because of thehigh population density near the site. The re-sponsiveness summary said: “Should this oc-cur, the organics will be rapidly dispersed inthe air, allowing for a very low probability ofany adverse impacts through inhalation down-wind of the site. ”

ISV received very detailed examination in theFS for the BF Goodrich and AIRCO site in Ken-tucky, which was completed several months af-ter the Pristine ROD. ISV was not selected atthe BF Goodrich/AIRCO site, primarily becauseits high cost made it not cost-effective. Severalof the comments about ISV in that FS (from adifferent EPA region and contractor) are im-portant relative to the decision to use ISV atPristine:

●

●

●

“The effectiveness of off-gas collection andtreatment is not known. ”“The complexities of this repetitive proc-ess [incremental movement across a site]are not known since it has not been fullydemonstrated on a large site.”“There is very little data available as towhether vitrification is a reliable technol-ogy.” (Compare this to the Pristine ROD:“results indicate vitrification to be a relia-ble technology.”)

1--- ----- --

51

“The mass could take several years to cool,depending on the size; therefore, a tem-porary fence should be constructed to pre-vent physical contact with the cooling ma-terial.”“Implementation of the vitrification proc-ess has yet to be demonstrated on a com-mercial scale . . . It is probable that thefrequency of operation and maintenanceproblems would be higher than for a proventechnology. As a result, reliability wouldbe lower than for proven technologies.Emissions during implementation wouldrequire extensive control for both VOCsand dust. ” These concerns are especiallyrelevant to Pristine because of its locationin a highly industrialized area.

Similarly, ISV was rejected at several Super-fund sites in Colorado in 1987 (Denver RadiumOperable Units: Open Space, Card Property,1000 West Louisiana Properties, and 12th andQuivas Properties). What is especially interest-ing is that contamination by radioactive mate-rials would seem to be an ideal application ofISV because it was originally developed as acleanup technology for radioactive materials.However, all three RODS say the same thing:“[ISV] was eliminated during the initial screen-ing because its implementability for this par-ticular application is unproven. [ISV] has notbeen demonstrated on a large scale of utilizedin a highly-populated urban area like that ofthe Card property.” Moreover, in these casesISV was also rejected because of the possible“escape of radon gas and associated radon de-cay products.” This would seem to also be appli-cable to escape of trapped gaseous organic con-taminants at a site like Pristine, and the concernundermines the belief that monolithic, solidi-fied ISV material offers secure, tight, and per-manent encapsulation.

RIFS contractor. –Camp Dresser and McKee(CDM); the cost was at least $500,000. The RODdid not indicate who did the FS, but a copy ofthe FS shows it was CDM with Roy F. Westonperforming the FS as a subcontractor. The RODsaid that several figures in the FS were wrongand recalculated figures were used in the ROD.The ROD also said that after the RI was com-

pleted “several gaps were identified” and ad-ditional work was done which took anotheryear.

State concurrence.—The ROD said that Ohio’sletter of concurrence is forthcoming. This sug-gests that the ROD was rushed to get it out bythe end of the fiscal year quarter.

Community acceptance.–“The community andPRPs are generally in agreement with the ground-water extraction and treatment component ofthe alternative. Some members of the commu-nity have fully supported U.S. EPA’s recom-mended alternative, while the PRPs rejectedvitrification and have proposed installation ofa RCRA cap with soil gas venting. The City ofReading prefers that U.S. EPA fund a less ex-pensive remedial action and give it the remain-ing funds to build a new municipal treatmentplant.”

Special comments.—The FS had an initial discus-sion of the two main treatment alternativeswhere unit costs and other data were presented.For in situ vitrification the sole technology de-veloper, Battelle Pacific Northwest Labora-tories, was acknowledged and was the sourceof the data. For onsite incineration, the sourceof information is described as “Firm A.” Onlyrotary kiln incineration is used in the detailedanalysis. Because a wholly-owned subsidiaryof the FS subcontractor, Roy F. Weston, Inc.,does cleanups with a transportable rotary kilnincinerator it seems likely that subsidiary wasFirm A.

The ROD acknowledged some uncertaintyabout the groundwater cleanup: “The extentof contamination from Pristine, Inc., will bedetermined by additional studies during the re-medial design.” However, this ROD addressedthe final source control remedy and should notbe expected to be definitive about the ground-water cleanup. The ROD also noted that con-tamination in the lower aquifer may be theresult of “a multi source groundwater contami-nation problem in the area. ” According to theROD, a variety of other types of actions mightbe used, including RCRA corrective action.

The responsiveness summary had an inter-esting interpretation of the provision in SARA

52

Section 121(b)(2) that says that a technology canbe selected even if it has not “been achievedin practice” elsewhere. The interpretation wasthat the provision allows a technology selec-tion without directly applicable data on its ef-fectiveness before the ROD. Legally, this posi-tion seems correct because congressional intentwas not to prevent full-scale application justbecause there has been no prior full-scale ap-plication of a new technology. But traditionalengineering practice does not condone choos-ing or using a technology without supportivesite-specific test data. The issue is that ISV hasnot been tested sufficiently at Pristine nor onlarge, unconfined contaminated soil not the ab-sence of successful full-scale use of it at a simi-lar site.

The responsiveness summary would havebeen more useful had it provided the sourcesof specific comments, which is normally done.

General conclusions.—There area number of out-standing aspects to the Pristine ROD. The com-mitment to meeting SARA’s preference for per-manent treatment technology is excellent. Thecommitment to an innovative technology iscommendable. The risk assessment is excep-tionally good, detailed, and well presented. Forexample, using a 1 in 1 million risk for inges-tion and direct contact, the cleanup targets forsoil are 3,182 ppb and 15,041 ppb for benzeneand trichlorethylene; but, when removing thethreat to groundwater from leaching of the con-taminated soil was used, the targets for the twochemicals decreased to 116 ppb and 175 ppbto meet drinking water standards.

Unlike most FSS, the Pristine FS presenteda preferred alternative that was recommendedfor implementation. It is OTA’s understanding,from speaking to RIFS contractors, that EPAusually directs them not to give a recommendedcleanup. Why was ISV so strongly supportedfor Pristine, especially before the use of it inthe removal action planned by Region 5?

As reported in Pristine’s responsiveness sum-mary, EPA’s selection of ISV got a poor recep-tion, apparently from responsible parties, be-cause of its unproven state, high cost, and

preferential handling in the FS and ROD: “Be-cause of its obvious bias in favor of ISV . . . theFS does not properly evaluate all existing rele-vant technologies . . . “ EPA was accused of be-ing “arbitrary and capricious. ” EPA defendedits selection at great length, including a discus-sion of why vacuum extraction of VOCs, appar-ently considered a viable alternative by the re-sponsible parties, is not applicable to the site.(The technology was not analyzed in the FS.)However, EPA’s discussion did not resolve thequestions raised here about ISV, nor did EPAgo into any technical depth in discussing vacuumextraction. Extensive work done for the respon-sible parties indicates that vacuum extraction,which has been selected for cleanups else-where, may be feasible and cost-effective atPristine (see below). To the extent that ISV re-duces risk mainly through removal of volatileorganic contaminants, it performs functionallylike vacuum extraction removal of volatile or-ganics. But vacuum extraction is intrinsicallya lower cost technology that uses less capitalintensive equipment and energy than does ISV.

The problem is the decisionmaking processand the accuracy of crucial data upon whichit is based, As with so many sites, treatabilitytests for Pristine were postponed to the post-ROD Design Phase even though test data arenecessary to fully support the selection ofremedy. This criticism is not directed at the ISVtechnology itself, which might eventually workat the site and which is an important newcleanup technology. But the wisdom of choos-ing ISV for the Pristine site remains question-able for several reasons.

Consider the following initial criterion to becomplied with for a technology to pass the ini-tial screening in a feasibility study: “There mustbe a demonstrated history of successful use ofthe technology in environments similar to the

. site. All technologies of a research and de-velopment nature, and which cannot be rea-sonably said to be in common use, are rejected.”This criterion is from the 1987 ROD for theNorthern Engraving Corp. site in Wisconsinin the same EPA Region as Pristine. ISV tech-nology could not meet that criterion, nor is the

—

53

criterion consistent with SARA, but the pointhere is inconsistency within the Superfundprogram.

Either EPA’s SITE program needs to provein situ vitrification or the technology shouldbe released from the demonstration programand accepted as proven technology in RODselections. Without treatability study results,the uncertainty about ISV’s effectiveness, whencoupled with no cost advantage over provenincineration (see below), weakens Pristine’sROD and the government’s attempts to get re-sponsible parties to take over the cleanup. Al-though SARA allows the use of new innovativecleanup technologies, there is little engineer-ing or public support for selecting an unproventechnology without supportive data. Data arenecessary to substantiate a technology’s abil-ity to meet specific cleanup goals, especiallywhen other proven and cost-effective perma-nent treatment technologies are available. More-over, in the Pristine FS, treatment technologiesother than incineration were eliminated fromdetailed evaluation with no technical basis. Al-though the FS and ROD were clearly commit-ted to using a treatment technology, the alter-natives became very restricted. Aside from ISVand conventional incineration technology, noconsideration was given, for example, to aboveground vitrification of excavated soil in a fur-nace. This technology was said, by its developer(Penberthy Electromelt International, Inc.), tocost $180 per ton ($280 per cubic yard) in di-rect unit costs; a transportable furnace wouldtake about 500 days to perform the work at Pris-tine. These figures are competitive with ISV.

Is EPA prepared to alter its decision in theDesign Phase if test results are negative? TheROD said that manufacture of the specially de-signed equipment “will occur concurrentlywith the remedial design. ” In other words, abig investment will be made before test resultscan support the selection.

The cost of the ISV choice may have beenunder estimated for several reasons discussedearlier (depth of treatment and building spe-cial equipment). The range of unit cost for ISV

given in the FS is $250 to $350 per cubic yard.The FS used a mid-range value of $290 per cu-bic yard. If a unit cost of $350 is used, it leadsto a total cost of $26 million instead of the RODestimate of $22 million. (Total costs include bur-dens and groundwater cleanup.)

The FS contained initial screening costs forthe two treatment options: ISV and incinera-tion. At the screening stage, the two costslooked comparable: ISV at $13 million and in-cineration at $18 million. These figures seemto come from using the upper range of the unitcosts given in the FS ($500 for incineration and$350 for ISV, with no burden added). But theISV cost increased by 70 percent in the finalcost calculations in the FS, while the incinera-tion cost increased by 183 percent, with no ex-planation provided. There is more uncertaintyabout the cost of the ISV option, however, thanabout incineration and just the opposite changein cost would have seemed more plausible.

As discussed earlier, the cost of onsite in-cineration was seriously over estimated and be-yond EPA’s standard allowable range of +50/–30 percent. (This allowable range, itself, islarge enough to invalidate a technology deci-sion based on cost.) The FS did not use the $425per cubic yard mid-range value from its ownincineration data ($350 to $500); data which ap-pears higher than reliable cost estimates. OTArecalculated the cost of the onsite incinerationoption. Instead of using $730 per cubic yard(which the ROD used to calculate incinerationcosts), OTA used $300 per cubic yard (see dis-cussion above). Including all the other costs forthe incineration alternative, such as ground-water cleanup, as done in the FS, the total costfor the incineration alternative then becomes$23 million (close to ISV’s cost). If the cost ofthe ISV option is also recalculated to reflect thehigh end of the cost range supplied by the ven-dor, then its cost is $26 million. The conclu-sion is: onsite incineration is not likely to bemore expensive than ISV at Pristine.

In these recalculations the indirect or bur-den costs (83 percent) are those used in the Pris-tine FS; however, these are much higher than

54

those in the FSS for the Davis and Re-Solve sites.At Pristine, various contingencies, construc-tion services, and design costs amounted to aburden of 83 percent, while in the Davis andRe-Solve FSs the burden is 35 percent; the lat-ter explicitly included costs for pilot studywork, while the figures for Pristine did not. ForCrystal City, the burden was 29 percent; forChemical Control, 56 percent; and for SeymourRecycling, 60 percent. For the BF Goodrich/AIRCO site in Kentucky where ISV receivedextensive examination, the total overhead onthe ISV direct costs was 60 percent (and theunit cost was $275 per cubic yard). The markupat Pristine was substantially higher than themarkups in these other recent FSs. That is, bothtreatment estimates at Pristine would have beenlower if a lower burden were used; with 35 per-cent, used by the same RIFS contractor in otherFSs, the cost for ISV becomes $16 million in-stead of $22 million and the incineration op-tion (at $730 per cubic yard) becomes $38 mil-lion instead of $51 million.

Indeed, mobile incineration might be lesscostly; if the low $186 per cubic yard figure fromthe Seymour site (which is in agreement withother data given above for recent actual con-tracted costs) is correct and is used for Pris-tine, with the same high burden rate as the otheralternatives, the total cost of the rejected in-cineration option is about $15 million [insteadof the estimated $51 million) versus $22 mil-lion for the selected remedy of in situ vitrifica-tion. With the lower, more typical burden rateof 35 percent, the incineration option comesto about $11 million (versus $16 million for ISVat the same, lower burden rate). The Seymoursite is in the same EPA Region as Pristine. Thesetwo sites illustrate that more regional oversightis necessary to catch inconsistencies in datacritical to technology selection.

The lower figures for incineration are impor-tant in the context of the government obtain-ing a settlement with the responsible parties;if this estimate proves correct on closer scru-tiny and true costs of incineration at Pristineare indeed much lower than the cost of ISV,then incineration becomes the more attractivecost-effective permanent remedy. However, on-

site incineration is likely to cost more thanvacuum extraction of volatile organic contami-nants, which the responsible parties favor. Butvacuum extraction is a separation technology,and an important issue (as it is for ISV) is whatis done with the extracted contaminants. If theyare destroyed rather than landfilled after car-bon adsoption, the costs increase. Moreover,the diversity of contaminants at Pristine re-quires careful analysis of vacuum extraction’sability to remove them; it might be able to do so.

Although at Pristine ISV was rated compara-ble in effectiveness to incineration and betterfor cost, its implementability is lower than thatfor incineration because ISV has not been rou-tinely used for chemical waste cleanups. ISVhas a higher level of uncertainty with regardto site conditions, and there is a need for site-specific design. Indeed, the FS said, “There ismore data to support incineration . . . Of thesoil contaminant destruction alternatives, only[incineration] has a demonstrated performanceand reliability, ” Incineration offers consider-ably more certainty as to effectiveness, relia-bility, and cost.

Cost aside, incineration is a less risky selec-tion at this time in the absence of treatabilitystudy data that could remove uncertaintiesabout ISV for the Pristine site, especially withregard to off-gas collection and treatment, themigration of contaminants into surroundingsoil, the degree of destruction of all organic sitecontaminants, and safety uncertainties for thesurrounding community.

There is another uncertainty about the im-plementation of ISV. Battelle has exclusiverights from the Department of Energy to mar-ket ISV for nonradioactive sites. The ROD ac-knowledged that “ISV is a patented processwhich requires a license.” Other companieswho do actual cleanup work are not familiaror experienced with the technology. However,the Pristine responsiveness summary said:“The selection of ISV is not patently unfairsince the developer will be licensing firms tocarry out the process and the bid process willbe competitive.” Subsequently, Battelle changedthe way it offers the technology such that its

55

availability and, hence, competition may be re-stricted. Rather than supplying the technologyto cleanup companies, which has not provena successful strategy, Battelle has helped forma new company with startup capital. The newcompany, the GeoSafe Corp., will seek addi-tional venture capital and will enter the haz-ardous waste cleanup business directly withISV; it has the exclusive worldwide rights forthis market. No competitive bid process now

‘ appears possible.

The ROD omitted any commitment to SARA’srequirement for 5-year reviews when hazard-ous material remains onsite, a requirementwhich applies in this case because of the stabili-zation aspect of the technology. For example,the 5-year review was called for in the Chemi-cal Control ROD, which selected in situ stabili-zation, and at the Tacoma Tar Pits, which se-lected stabilization.

The Pristine case illustrates how differentoffices of the same EPA contractor and howdifferent Superfund contractors can use sub-stantially different data. One contractor’s of-fice used a cost for incineration for Pristine ofabout twice what another of the contractor’soffices used for the Davis Liquid Waste and Re-Solve sites. A close examination of the calcu-lations for estimated costs at Pristine revealsthat a very high indirect or cost burden wasused, compared to indirect costs in FSs for sev-eral sites. Such cost variations have no techni-cal basis.

Case Study 7Renora, Inc., Edison Township, New Jersey,

EPA Region 2

Capsule OTA findings:-The selected remedy makesuse of offsite landfilling for soils contaminatedwith PCBs. Biological treatment was selectedfor soils contaminated with diverse organiccompounds and toxic metals and for contami-nated groundwater, but no treatability studysupported its selection.

Key dates:

● Entered Superfund system: 5/1/81

●

●

●

●

●

●

●

Preliminary Assessment: 8/1/82Site Inspection: 9/1/81 to 8/1/82National Priorities List–proposed date: 12/1/82–final date: 9/1/83—site rank: #378 out of 770RIFS start and completion: 5/85 to 8/87Public comment period before Record ofDecision: 8/18/87 to 9/10/87Signing of ROD: 9/29/87Estimated complete remediation: 1 to 2years after signing

Total time.—8 years

Brief description of site.–The site “is an approxi-mately one acre parcel of land in an area zonedfor light industrial use. The surrounding areais residential with three sensitive uses (a nurs-ery school, senior citizens center, and an apart-ment complex) within two thousand feet of thesite. . . . two residential developments [were]built in close proximity to the site during theperiod of time the RI/FS was conducted. From1978 to 1982 Renora Inc., transported and ac-cepted materials containing hazardous sub-stances for transfer, storage, blending and ulti-mately, disposal through abandonment at thesite [in 1982].”

Major contamination/environmental threat.—Evidenceof contamination problems started in 1978. In1985, the Remedial Investigation (RI) said: “Sur-ficial soils (0 to 2 feet) are primarily contami-nated with polychlorinated biphenyls (PCBs)and polynuclear aromatic hydrocarbons (PAHs)and to a lesser extent with volatile organic com-pounds (VOCs), acid extractable compounds(AECs), other base/neutral organic compounds(BNCs) and heavy metals. Shallow groundwaterbeneath the site is contaminated with low levelsof chloroethane, (a volatile organic compound)and heavy metals. Surface water and sedimentsamples show levels of heavy metals, tetra-chloroethene, phenols and pesticides. No evi-dence of air contamination was found at thesite, No buried drums were found at the site. ”

The RI concluded that the significant path-ways of exposure are direct contact and sub-sequent incidental ingestion by children tres-

56

passing the site, future onsite workers, andfuture site residents.

According to the ROD:” . . . there are no off-site impacts directly attributable to site opera-tions. Therefore, no management of migrationmeasures were selected as part of the overallremedy for any environmental media. Althoughgroundwater does not pose a public health risk,achievement of target treatment/residual levelswill result in restoration of groundwater qual-ity to potable water standards. ”

HRS scores.-groundwater 69.32; surface water9.4; air 0.00; total 40.44

Removal actions.-The ROD said: “A removal ac-tion was initiated in October 1984 and con-tinued through April 1985. During the cleanup,approximately 33,000 gallons of liquid wasteand 28,000 gallons of PCB contaminated wasteoil along with approximately 500 cubic yardsof non-PCB contaminated soils and 560 cubicyards of PCB-contaminated soils were shippedoff-site for proper disposal [presumably in alandfill].” The SCAP shows a Federal removalaction 10/23/84 to 10/31/84 at a cost of $27,000and that the responsible parties performed one,9/28/84 to 4/16/85. Data from the New JerseyDepartment of Environmental Protection indi-cates that the responsible parties spent $4 mil-lion for their removal action.

Cleanup remedy selected.—The remedy has fourkey components:

1. “ . . ● excavation of all PCB-contaminatedsoils containing concentrations above 5ppm [parts per million] (approximately1,100 [cubic yards]) and off-site land filldisposal . . . “

2. “. . . biodegradation of all PAH-contam-inated soils containing concentrationsabove 10 ppm (approximately 4400 [cubicyards]) . . .“

3. “ . . . use of groundwater as an irrigationmedium for the bioremediation system. . . “ [and]

4. “ . . . backfilling, grading and revegetation.”

The cost of the selected remedy was estimatedat $1.4 million.

A number of cleanup alternatives were ex-amined, including containment approaches,treatment of less material, use of incinerationinstead of landfilling, and conventional ground-water treatment.

Satisfaction of SARA statutory requirements:

1) Selection of permanent cleanup.—The RODsaid: “Overall, [the selected remedy] is protec-tive of public health and the environment. Aninnovative treatment technology would be uti-lized as a major portion of the remedy. Thereis complete reduction of the toxicity, mobility,and volume of the contamination. The remedyis permanent and would not require long-termmanagement” (emphasis added). More cau-tiously, the ROD said the remedy “significantlyreduces the toxicity, mobility and volume ofcontaminants” (emphasis added). Also: “Uponcompletion of the remedy future site uses willbe unrestricted.”

There is no specific technical information inthe ROD or Feasibility Study to support theselection of biological treatment for the Renorasite. There are no test data, no citations to thetechnical literature, nor reference to previoususe at specific sites. The ROD stated: “A prereq-uisite to implementation of the bioremediationportion of the alternative is a pre-design treat-ability study to refine parameters of the opera-tion.” There are a large number of contami-nants, and many of the organic contaminantsand heavy metals are considered difficult to bi-odegrade. The biological approach is not off-the-shelf cleanup technology, except for a fewsimpler types of cleanups.

A key issue is the extent of destruction by bi-otreatment. While it can be easy to get somedestruction, it can be very difficult to get com-plete destruction or as much, for example, asrequired for incineration (99.99 percent de-struction). Finding ways to enhance biodegra-dation for a complex set of chemicals and forrecalcitrant contaminants can be difficult. Forexample, a recent research paper discussed the“degradation of Benzo[a]pyrene and otherrecalcitrant PAHs” and explained its failed at-tempt to foster biodegradation by noting that

I

57

“organic amendments which are readily uti-lized for carbon and energy are often ineffec-tive in stimulating degradation of recalcitrantorganic compounds.” (M.P. Coover and R.C.Sims, Hazardous Waste & Hazardous Materi-als, vol. 4, No. 2, 1987, pp. 151-158.)

The current state of technical knowledge andexperience does not support the cleanup selec-tion in the absence of site-specific data to proveeffectiveness in meeting the cleanup goals.There are substantially different forms of bio-logical treatments, ranging from simple landtreatment to the sophisticated use of bioreac-tors using a variety of additions to promote andsustain biological destruction to desired resid-ual levels of contaminants, but the ROD dealtwith the technology only in its simpler, genericterms.

Biodegradation was selected for the FrenchLimited site in Texas, but the ROD emphasizedthat “biodegradation of PCBs to the criterion(23 ppm) has not been demonstrated.” EPA re-quired, therefore, that a secondary stabilizationtreatment be used on the residue from the bio-treatment.

The FS for the Liquid Disposal site in Michi-gan examined biological treatment in more de-tail than most studies and did not select it. Achief reason was: “The level of effectivenessof the biodegradation technologies on a non-homogeneous waste stream is unknown.” Thestudy noted that extensive testing would be nec-essary to prove the technology effective for thesite.

Biological treatment was rejected in the Fea-sibility Study for Crystal City because “[it] isgenerally ineffective for destroying these wastesas the treatment is not performed in a controlledenvironment. Several processes are being de-veloped which show potential. However, noneof these processes have been developed pastthe laboratory stage. Therefore, biological treat-ment has been ruled out. ” Biodegradation wasalso rejected in the ROD for the Tower Chemi-cal Superfund site in Florida: “Biodegradationdoes not address the metals contaminationfound at the site and would require long term

operations before full clean-up is effective.Other technologies, e.g., incineration, wouldprovide equal destruction efficiencies in ashorter time frame. ”

The Renora ROD said:●

●

●

". . . bioremediation of soils is considered

an innovative treatment technology in thefield of hazardous waste management.”“Although available scientific literature in-dicates implementation of the bioremedi-ation portion of the alternative is feasible;a pre-design treatability study would be re-quired to confirm the operational reliabil-ity of the alternative. ”" . the bulk of the contaminated soils (ap-proximately 60 percent of the total) whichremain are amenable to onsite bioremedi-ation. Available scientific literature and itsuse in the oil refining industry indicate thatthe bioremediation aspect of the selectedalternative will achieve the target treat-ment/residual levels.”“The prospect for long-term reliability ofthe alternative would be established by thepre-design treatability work and subse-quent verification sampling. However, asthis remedy is permanent and substantiallyreduces the toxicity, mobility and volumeof contamination the likelihood of remedyreplacement is low” (emphasis added; com-pare to different statement above).

The choice of offsite landfilling over inciner-ation was not discussed in detail in a direct wayfor the alternative selected. Other alternativeswhich would include more use of incineration,because no bioremediation would be used, weresaid to “not result in providing any greater pro-tection of public health or the environment thatwould justify the incremental cost increase. ”

Z) Accurate assessment of land disposal andcontainment alternatives.—The use of offsitelandfilling for the PCB contaminated soil in-stead of treatment is contrary to the intent ofSARA, even though the amount is relativelysmall. A recent EPA study on PCB cleanup con-cluded: “Landfilling of such materials, wherelegal, is a potential source of groundwater con-

58

lamination, and only a temporary measure atbest.” (U.S. Environmental Protection Agency,Office of Research & Development, “Bengart& Memel PCB Site Soil DecontaminationProject,” undated but apparently 1987.) The FS

for the Liquid Disposal site in Michigan said:“However, moving wastes from one site toanother does not constitute a permanent re-medial action” (emphasis added). For Renora,there was no discussion of the negative aspectsof using offsite landfilling. The Renora RODlooked strictly from the perspective of this site:“Excavation of PCB contaminated soils and off-site landfilling will physically remove hazard-ous substances, pollutants and contaminantsfrom the site.”

RIFS contractor.—The RIFS was paid for by agroup of potentially responsible parties andconducted under contract by BCM Eastern Inc.;$250,000; an endangerment assessment wasdone by Camp Dresser and McKee under con-tract to EPA.

State concurrence.—The State of New Jersey con-curred with the selected remedy.

Community acceptance.—The responsiveness sum-mary gives little information on what the com-munity felt about the selected remedy.

Special comments. –The analysis of cleanup al-ternatives was somewhat confusing becausesome options, including the selected one, re-ferred to offsite disposal consisting of eitherlandfilling or incineration. Except for cost,there were no distinctions made within an alter-native for the use of landfilling versus inciner-ation. Therefore, the selected alternative mighthave received overly high evaluations becauseincineration was included as an option but ulti-mately not selected.

The ROD contained a good statement on cap-ping: “[it] would not be considered permanentsince the toxicity and volume of contaminantsin the soil would remain essentially unchanged.”

The ROD did not commit to groundwatermonitoring after the selected remedy is imple-mented, which seems relevant, since it says thatthe groundwater will be restored to a potable

condition and since the HRS groundwaterscore was quite high. However, the responsive-ness summary did include a monitoring stepin its description of the selected remedy.

General conclusions. —A key issue is the choiceof offsite landfilling over offsite incinerationfor the PCB contaminated soil. A major driv-ing force behind SARA’s requirements for per-manently effective treatment technologies wasthe long-term ineffectiveness of moving buriedhazardous waste from Superfund cleanup sitesto other land disposal sites. This ROD, consid-ering only this site, implied that offsite disposalresults in maximum protection. It did not con-sider the long-term consequences at another siteof landfilling materials transported from thissite. In addition to the two EPA sources alreadynoted, the FS for the Pristine site in Ohio re-jected the option of sending contaminated soilto an offsite landfill because “ . . . there is po-tential for the contaminated soil to cause a prob-lem at the off-site facility. . . . the alternative isnot permanent and is the least preferred underSARA.” The reasons for ruling out offsite in-cineration of the PCB contaminated soil atRenora were not given.

The Renora remedy also perpetuates a trendstarted by the major removal action completedat the site in 1985. A viable alternative, as ex-amined in the ROD, was to incinerate the wasteinstead of landfilling it. Indeed, the ROD con-tained an important statement on this point inthe discussion of the alternative that was se-lected: “If the excavated PCB contaminatedsoils are incinerated instead of landfilled, therewould be a permanent reduction in the toxic-ity, mobility and volume of contaminants insoils.” The clear implication is that landfillingis not comparable in meeting SARA’s require-ments. The chief reason for not selecting in-cineration of the PCB contaminated soils ap-pears to be its greater cost, an additional $4.6million. The ROD noted: “it is likely that the[potentially responsible parties] will implementthe selected remedy.” In other words, approvalof offsite landfilling by EPA may have facili-tated getting an agreement from the responsi-ble parties to clean up the site.

59

The interest in using innovative treatmenttechnology is commendable, and some biologi-cal remedy may, in fact, be found effective. Butthe ROD decision was made without support-ing technical information. The ROD and theFS contained no details about the technology.There are many forms of biological treatment,and it is still too new a cleanup technology tomake assumptions on effectiveness at a sitewith so many organic contaminants, some atrelatively high concentration. Technologyspecificity is high for biological treatment,which means that it is difficult to extrapolatesuccess from one waste to another. Delay oftreatability testing until after the ROD createsconsiderable uncertainty and the potential foractions which are not fully protective of pub-lic health and environment because of eithersubstantial loss of time or a compromise ofcleanup goals if the testing shows problems inthe biotreatment. Indeed, because of the needfor extensive treatability testing, the estimatedtime for complete implementation seems overlyoptimistic. Such biological treatment (both aer-obic and anaerobic processes) of contaminatedsoil was rejected at an early screening stage inthe Pristine FS because “Mixed wastes and lowconcentrations (less than 100 ppm) are difficultto treat. ” The same condition exists at Renora.

Because of the use of landfilling and the selec-tion of an unproven treatment technology, theselected remedy cannot be assured to be per-manent. Moreover, compliance with SARA’srequirement on the reduction in toxicity, mo-bility, or volume was described in three differ-ent ways: complete, substantial, and significant.This puzzling situation may indicate end-of-fiscal-year ROD rushing or confusion over thecapabilities of the selected remedy.

Case Study 8Sand Springs Petrochemical Complex,

Tulsa County, Oklahoma, EPA Region 6

Capsule OTA findings.–EPA originally said solid-ification technology was ineffective for the highorganic content site wastes and incinerationwas effective. Nevertheless, EPA reversed it-

self and selected solidification for most of thecleanup, which the responsible party had foundeffective in its treatability study. Incinerationis to be used if solidification technology is notsuccessfully demonstrated or fails after solidi-fied material is landfilled on the floodplain site,but the criteria for failure are unspecified.

Key dates:●

●

●

●

•●

●

●

Entered Superfund system: 8/1/80Preliminary Assessment: 6/1/80Site Inspection: 11/1/80National Priorities List–proposed date: 10/84–final date: 6/86—site rank: #761 out of 770RIFS start and completion: 6/29/84 to 5/4/87Public comment period before Record ofDecision: 7/29/87 to 9/1/87Signing of ROD: 9/29/87Estimated complete remediation: 11/91

Total time.—11 years

Brief description of site.–”The site operated as arefinery from the turn of the century throughthe 1940s. The property has since been devel-oped as an industrial area and consists of anabandoned solvent and waste oil recycler, anactive transformer salvage/recycler, activechemical manufacturers and various other in-dustries. . . . the site is located on the northernbank of the Arkansas River, immediately westof Tulsa, Oklahoma. The site encompasses ap-proximately 235 acres [and] includes unlinedacid sludge pits, a surface impoundment, sur-ficial sludge contamination, solvent and wasteoil lagoons and contaminated sediments. The[site] is located in the alluvial floodplain of theArkansas River.”

Major contamination/environmental threat.—’’Totalknown waste volume is approximately 130,000cubic yards. During the period of operation haz-ardous substances were stored or disposed ofin drums, tanks, unlined pits and lagoons orburied on-site. These substances include vari-ous volatile and non-volatile organics, chlori-nated solvents, and sludges containing heavymetals. Waste pits have contaminated local

60

groundwater and caused migration of surfacecontaminants. ”

EPA concluded that there are four majorsources of risks: direct contact with organic car-cinogens and highly acidic wastes and surfacewaters; air emissions of acid fumes and vola-tile organic compounds; surface waters pol-luted by runoff during heavy rains; and ground-water being contaminated directly by lagoonsand indirectly from site runoff: “in heavy rainsthe site is submerged.”

According to the FS: “It is believed that con-taminants from the pits, ponds, and lagoons areleaching into the alluvial aquifer, therefore, amajor pathway for migration is probably ground-water. However, by definition this pathway hasbeen excluded from consideration during theFS for the Operable Unit.” The same contami-nation problem exists for surface water migra-tion offsite. The FS also said: “Several on-siteponds and lagoons have a history of breachingtheir containment structures: there have beenincidents of dike walls breaching for one of theGlen Wynn lagoons, as well as flow of materi-als from the river acid sludge pits into the Ar-kansas River, which have occurred in the past. . . . the contents of the large and small acidsludge pits had breached their dike walls onseveral occasions.”

HRS scores.-groundwater,44 .90; surface water21.82; air 0.00; total 28.86

Removal actions.—A private party performed aremoval action in 1984; there are no details inthe ROD.

Cleanup remedy selected.—EPA designated thiscleanup as a source control operable unit thatcovered surface liquids, sludges, and heavilycontaminated solids but not minimally contami-nated soil or groundwater. The latter is to beaddressed in a subsequent ROD. Originally,before the ROD was officially signed, EPAselected onsite incineration of wastes andsolicited public comment on it as part of theRIFS public comment period; the agency hadalready evaluated solidification and onsite land-fill and solvent extraction, all of which wererejected. EPA changed its mind and selected

solidification, accepting a five-part proposal byAtlantic Richfield Company (ARCO), a respon-sible party, that included:

1. excavation and offsite thermal destructionof some unspecified volume of surficialsludges;

Z. solidification and/or stabilization of all re-maining sludges and containment of theresulting material in an onsite RCRA haz-ardous waste landfill;

3. demonstration that solidification technol-

4

5

ogy meets EPA approved criteria and,should it not do so, use thermal destruc-tion [apparently onsite];no liability release for the site or from fu-ture maintenance and monitoring; andrepair or restoration of the landfill to en-sure no migration or destruction or treat-ment of all or a portion of its contents, asEPA deems appropriate, should monitor-ing show that the solidification/stabiliza-tion remedy fails.

It appears that ARCO is anticipated by EPAto sign a consent decree, agreeing to pay forthe cleanup.

Cost data on the selected remedy is absentbecause the combination of solidification andincineration was not evaluated in the FS. It isunclear how much material will be incineratedoffsite initially. But if solidification is used itwill cost less than incinerating all the waste.Incineration was estimated in the ROD to cost$67 million and complete solidification wasestimated at $38 million (the comparable figuresin the FS are $54 million and $31 million). OTAestimates that the probable comparable cost ofthe five-part remedy is $45 million, but this fig-ure is highly uncertain because there are manydifferent forms of solidification.

Satisfaction of SARA statutory requirements:

I) Selection of permanent cleanup.—EPA saidin the ROD that the selected remedy, basedmostly on solidification, fulfills the statutorypreference. However, as EPA stated: “on-sitethermal destruction of wastes . . . appears tomeet more statutory selection criteria than theother remedies evaluated. ” With solidification,

61

“unlike on-site thermal destruction, the toxic-ity of wastes would not be reduced and the vol-ume of wastes would be increased.”

EPA views in the ROD on the chosen solidifi-cation option at this site included the following:

●

●

●

●

●

●

●

●

●

●

●

●

●

●

". . . [there was a] lack of demonstrated

permanence.”" . . . the capability of solidification orstabilization techniques to permanentlybind with high organic wastes, such asthose found at Sand Springs, has not beendemonstrated in the pilot studies conductedon-site. ”" . . . without further treatment free liquidcontaminant concentrations were not re-duced to meet RCRA land ban restrictions.”" . . . the unconfined compressive strengthof the stabilized material . . . does not meetthe recommended disposal criteria. ”" leaching tests conducted by EPAsCincinnati laboratory show that the solid-ified material leaches contaminants.”" . . . leaching of contaminants, and incom-plete encapsulation [small globules of wastewere seen] raises questions about the longterm effectiveness and permanence of theprocess.”the waste” . . . contains 50 percent organiccompounds raising doubts about the abil-it y of stabilized or solidified waste to meetRCRA requirements in the long term.”samples “ . . . show obvious degradationof the solidifying matrix following analy-sis for total organic content. ”" the net assessment is that solidifica-tion or stabilization processes present dif-ficult problems with respect to meetingARARs [standards].”" . . . possible air emissions.”" . . . volumetric increase of 50 [to 200percent].”" . . . the potential for failure was deter-mined to be greatest for the on-site solidifi-cation remedy.”"

. . . the source of the contamination willnot be destroyed.”additional “ . . . studies will need to beperformed on the subsurface petroleumwastes. ”

The FS summed up its evaluation of onsitesolidification: “Not a proven technology forhigh organic waste. Contaminant source iso-lated, may not be rendered nonhazardous. Maynot meet ARAR.” It remains unclear whetherthe test results supplied by ARCO removed allof the above concerns for EPA; but the inclu-sion of the third provision in the five-partremedy which requires demonstration of solid-ification technology suggests that EPA was notfully convinced by the ARCO test data or thatit did not have enough time to fully evaluateit prior to signing the ROD.

EPA has tied the environmental acceptabil-ity of the solidification remedy to two condi-tions: “if the effectiveness of this concept is ade-quately assured or if ARCO undertakes thecorrective actions deemed appropriate by EPAshould the remedy fail.” No such conditionswould have been attached to the originally cho-sen thermal destruction remedy, which ARCOalso examined in its treatability study andwhich was found to work effectively. Cause forEPA rejecting the incineration option in theROD was said to be its “serious implementa-tion problems,” but EPA’s FS analysis also saidthat all the processes that would treat waste on-site “are judged to each have the same degreeof implementability.” Moreover, the ROD stated:“Actual implementation time for solidificationand thermal destruction is comparable . . . “

EPA said: “The proposed remedy is consid-ered permanent.” And that it “is cost-effectivecompared to equally environmentally protec-tive alternatives.” But EPA also said that thethermal destruction alternative offered moreoverall protection than solidification. The FSsummed up its case for onsite incineration:“Proven technology destroys hazardous mate-rial, Containment source worker health andsafety addressed in remediation. Meets ARAR.”

2) Accurate assessment of land disposal andcontainment alternatives.—With regard to fu-ture operation and maintenance, the ROD saidthat these “will be minimized since the sourceof the contamination will be removed.” But,as the ROD also stated, solidification does notdestroy the source of contamination that will

62

be left onsite. The ROD contained a replace-ment cost of $100 million should failure occurfor the onsite solidification and landfilling op-tion but no cost for the incineration alterna-tive. For these two pure (single technology) op-tions, the reduction in cost of about $30 millionfor solidification is offset by a possible futurere-remediation cost of $100 million. And, thattradeoff still exists.

The onsite landfill is supposed “to reducegroundwater infiltration and the chances of anycontaminants migrating off-site. ” But if “ . . .significant, unforeseen, off-site migration orcontamination occurs as a result of the site,appropriate remedial measures will be taken, ”No detailed analysis of future failures wasgiven. The ROD did not express concern abouthaving the onsite landfill in a location that issubmerged in heavy rain and that is in a flood-plain adjacent to the river. The ROD stated fur-ther: “While a hazardous waste landfill of so-lidified waste would protect health in the shortterm, the long term stability of this material isnot proven. ”

For Sand Springs, the importance of thewater level to remedy selection is striking. TheROD for the Tower Chemical Superfund sitein Florida, for instance, commented on the useof chemical stabilization followed by onsiteland disposal, a remedy the ROD rejected. Thecomments apply directly to the Sand Springssite: “Although this process is effective in ad-dressing inorganic contamination, the volumeof materials would increase, thus causing in-creased disposal facility requirements. In addi-tion, the soils being solidified contain signifi-cant amounts of organic compounds whichcould affect the integrity of the cement mono-lith. The presence of organics will require con-tainment of the monolith within an on-site land-fill built above the land surface due to the locallyhigh water table. This technology would alsorequire long-term (30 years) monitoring whichis less favorable than technologies which pro-vide permanent destruction of wastes. . . . ahigh water table at the site makes it infeasibleto solidify or build an on-site landfill which

meets the design specifications outlined inRCRA.”

The issue of effectiveness of solidificationtechnology for organics is critical for SandSprings. The FS for the Re-Solve site in Massa-chusetts rejected stabilization because “therehas been limited success in chemically fixingorganic contaminants such as solvents andPCBs.” The ROD for the Liquid Disposal sitein Michigan, which also selected stabilizationfor soil contaminated with organic chemicals,said that the hazardous substances “will notbe permanently destroyed” and “hazardouschemicals still remain in that [treated] mass. ”And the FS for the site said: “Considerable re-search data exists demonstrating the effective-ness of this technology in immobilizing a widerange of contaminants, primarily inorganic.A substantial amount of data does not exist,however, to accurately judge the long-termreliability of the process. Long-term leachingand volatilization can be expected for solubleand volatile organic wastes.” Although, stabili-zation was selected for Liquid Disposal, so wasthe use of a slurry wall and impermeable caparound and over the treated material, as a sec-ond level of control. The ROD for the FrenchLimited site in Texas (same EPA region asSands Spring) said: “Fixation is questionabledue to high organic content of untreated soils.”

An EPA report’s observations on halogenatedorganic wastes also apply to the selection ofchemical stabilization for Sand Springs: “thearea of solidification/encapsulation is one re-quiring additional study before it can be con-sidered viable technology. ” (U.S. EnvironmentalProtection Agency, Technical Resource Docu-ment: Treatment Technologies for HalogenatedOrganic Containing Wastes, vol. 1, January1988.)

Another EPA document, used to teach cleanupworkers about waste treatment says: “Solidifi-cation technologies are designed to be used forfinal waste treatment. This means the technol-ogy should be applied only after other treatmenttechniques have been applied, i.e., incineration,

63

chemical treatment or other. ” (U.S. Environ-mental Protection Agency, “RCRA/CERCLATreatment Alternatives for Hazardous Wastes,”October 1987.)

A specific type of solidification tested forSand Springs was mentioned in the FS for Crys-tal City, but the solidification/landfill alterna-tive was not selected at Crystal City. Data pro-vided by the vendor, on waste from some othersite, and reported in the Crystal City FS on twocontaminants also present at Sand Springs (2-methylnaphthalene and phenanthrene] showedhigh levels in the leachate. A demonstration ofthe same stabilization technology under EPAauspices concluded: “for the organics, theleachate concentrations were approximatelyequal for the treated and untreated soils.” (P.R.de Percin and S. Sawyer, “SITE Demonstra-tion of Hazcon Solidification/Stabilization Proc-ess,” paper presented at EPA’s Fourteenth An-nual Research Symposium, May 1988.)

A recent EPA study found “large losses oforganics during the mixing process.” (L. Weitz-man et al., “Evaluation of Solidification/Stabili-zation As A Best Demonstrated Available Tech-nology,” paper presented at EPA’s FourteenthAnnual Research Symposium, May 1988.)Another EPA study showed that stabilizationwas not competitive with thermal and chemi-cal treatment technologies and soil washing fororganic contamination. (R.C. Thurnau and M.P.Esposito, “TCLP As A Measure of TreatmentEffectiveness: Results of TCLP Work Com-pleted on Different Treatment Technologies forCERCLA Soils,” paper presented at EPA’sFourteenth Annual Research Symposium, May1988.)

RIFS contractor. —State led; $1.1 million; JohnMathes & Assoc.

State concurrence. —The State of Oklahoma fa-vored solidification over incineration,

Community acceptance. –EPA judged that thecommunity was more in favor of solidificationthan incineration. The community was veryconcerned about the future use of an incinera-tor for waste from other sites, the worsening

of the area’s air pollution, and harm to the lo-cal economy. (Building an incinerator for in-dustrial waste is frequently sold by industry tocommunities as a local economic advantage;several efforts to site a hazardous waste inciner-ator in Oklahoma are underway.) AlthoughEPA tried to allay the community’s concernsabout incineration, ultimately the communitypreferred the uncertainties of the solidificationtechnology and accepted the assurances thatincineration would be used if solidification wasless effective.

On this issue of the safety of mobile inciner-ation, Sand Springs can be compared to theDavis Liquid Waste site in Rhode Island wherethere also was substantial, documented com-munity concern about onsite incineration, con-cern to which EPA responded with good tech-nical points but, unlike Sand Springs, did notalter its choice of incineration. Also, in the RODfor the French Limited site (in the same EPAregion as Sand Springs), EPA defended mobileincineration: “Performance standards for airemissions from incinerators would be met, min-imizing the risk from these emissions. EPA con-siders the implementation of an incinerator tobe relatively simple in comparison to the otheralternatives evaluated in the summary. ”

Moreover, it is not clear that the communitywas totally aware of air pollution problems withsolidification. EPA’s responsiveness summarysaid: “Pilot studies have shown that some vola-tile compounds are driven off during excava-tion and mixing of the waste with the solidify-ing agent. Mass emission rates have not beenquantified.”

Special comments.—The ROD’s analysis of clean-up alternatives said that for any alternative itwill be necessary to pump and treat surface im-poundment liquids and to discharge them intothe Arkansas” River; no details were given.Moreover, the accepted ARCO proposal madeno mention of these needs.

General conclusions.—Sand Springs has somegood points: 1) pilot treatability studies wereused to evaluate treatment technologies; 2) alter-

64

native treatment remedies were analyzed; 3)EPA responded to the concerns and interestsof responsible parties, the community, and theState; and 4) some preferred treatment technol-ogy was selected by EPA.

But Sand Springs has many problems too, in-cluding a lot of confusion about what test datawere used by EPA and when. The ROD selectednot a reliable permanent remedy but a plan withseveral contingencies, with no assurance of apermanent remedy. Over a short period of time—about one month—EPA reversed a well-sup-ported, technically sound decision to use inciner-ation rather than solidification/stabilizationtechnology. In the ROD, EPA said: “Solidifi-cation was considered in detail during the Fea-sibility Study and actual pilot studies. Adequateinformation is available on which to base a de-cision.” Was EPA talking about the informationavailable when it signed the ROD, includingthe ARCO test results, or information obtainedby EPA prior to the RIFS public commentperiod? The ROD suggested that EPA had con-ducted its own tests on solidification of site ma-terials. But it may have used a very differenttype of solidification technology. In the RODresponsiveness summary EPA said: “The [ARCO]pilot studies had a major influence on the rem-edy selected. ” If that was so, then why did theROD still contain so many negative commentsabout solidification? Part of the answer maybe revealed in another statement by EPA in theROD’s responsiveness summary: “Solidifica-tion pilot tests [presumably ARCO’s] were onlyconducted on the surficial acid sludge waste.Additional waste characterization and pretreat-ment studies will need to be performed on thesubsurface petroleum wastes.” Another part ofthe answer is that EPA had conflicting test datafrom two different sources on several solidifi-cation technologies. Therefore the question per-sists: Was there enough test data to justify theROD’s selection of remedy?

The ROD contained no details on how EPAwill assure that independent, detailed, andtimely testing will track progress on the selectedremedy and detect ineffective performance inthe long term, if that occurs. If the treatmenttechnology is ineffective, contaminants will

leach out of the solidified mass, because thetreated material will be placed into a landfillon a floodplain adjacent to the Arkansas River.Landfill failure was not considered, nor is itslocation compatible with regulated use of landdisposal. Moreover, cause for concern aboutindependent testing and verification of solidifi-cation’s effectiveness is driven by ARCO’s po-sition that any form of waste treatment is un-necessary: “Improved site security and a claycap would mitigate this [accidental direct con-tact] potential risk.” (ARCO, letter to Carl Ed-lund, EPA Region 6, Aug. 31, 1987.) ARCO’scritique of EPA’s FS of August 31, 1987, saidthat fencing and a cap “could be a sufficientremedy.” Moreover, in this document ARCOalso said: “Long term effectiveness of inciner-ation, stabilization and solidification are com-parable.” These views of ARCO suggest thatthe selection of solidification, with costs muchlower than incineration, was a compromisemade by both EPA and ARCO and that future,post-ROD actions require close EPA scrutiny.

ARCO’s effort to get EPA to retrench fromits original decision to use incineration prob-ably was helped by its apparently successfulcriticism of the quality of the RIFS. Indeed, anumber of ARCO’s comments are consistentwith OTA’s observations in this report forRIFSsin general. For example, ARCO said: 1)“significant gaps exist in the data presented andconsidered in the FS”; 2) “The analysis reflectedby the FS is cursory and of limited detail”; 3)“The lack of back-up, the limited detail and thelack of references suggested that the analysismay not have involved the development of anyadditional information beyond that providedby the authors’ experience”; and 4) “The FSis characterized by an over-reliance on assump-tions rather than actual performance data.”

A big question still remains. Who bears theburden of proof that a treatment technologyworks before EPA officially endorses its useat an actual cleanup? There is no basis in thetechnical literature for concluding that solidifi-cation/stabilization technology is likely to beeffective for wastes with so much and so manydifferent kinds of organic contamination. Thepresence of negative laboratory results, which

65

EPA suggested it had prior to the ARCO testdata, would normally prevent application of acleanup technology at the site in question. Itis true that unusual conditions might justify anunproven technology; for example, in an emer-gency situation and where all other treatmenttechnologies are less applicable. But this siteis not generally considered to require emer-gency attention, and the pilot study on inciner-ation was successful. (Note that stabilizationof incinerator residue contaminated with me-tals only is proven technology.) It may well bethat, as EPA says, the technology proposed byARCO is “a promising innovative technology,”but sanctioning its full-scale applicationthrough a ROD on the basis of limited data ob-tained by the responsible party that conflictswith data obtained by the government, is a bigstep, especially because the data are inconsist-ent with what is generally understood aboutthe capabilities of the technology. In fact, thereare a number of very different proprietaryforms of solidification and stabilization (ARCOactually tested two and got similar results), andit is not clear that either ARCO or EPA has con-sidered or evaluated enough of them and theirperformance relative to the technologies usedin the ARCO treatability study.

The degree to which the ARCO data supportEPA’s decision also raises the issue of how ac-curately current EPA tests—in this case for haz-ardous waste treated by stabilization—predictlong-term environmental effectiveness. Muchmore rigorous testing appears necessary tomake the case that stabilization of hazardousorganic material, such as at Sand Springs, as-sures insignificant leaching of organic contami-nants under long-term conditions at the site.Therefore, although the ARCO test data do lookgood, they are limited by the test proceduresthemselves. Nor were the ARCO test data ob-tained by using standard test protocols andquality assurance procedures to assure the pub-lic and the government that the data are relia-ble. Doing this is a major effort and an impor-tant characteristic of EPA SITE program andmost treatability testing done by or for the gov-ernment.

After the ROD was issued, a news publica-tion reported: “Solidification poses ‘very littlerisk whatsoever,’ says an EPA headquarterssource, who is encouraged that the agency iswilling to allow its use at a Superfund site. Thetechnology has not been proven to ‘truly bindorganics,’ but any release of organic substanceswould be slight, the source explains. ” (InsideEPA, Nov. 27, 1987, p. 13.)

Moreover, the solidification technology eval-uated by ARCO appears to be one which is inEPA’s SITE technology demonstration pro-gram to “resolve issues standing in the way ofactual full-scale application, ” The demonstra-tion was conducted October 13-16, 1987, at theDouglassville Disposal Superfund site in Penn-sylvania, but results of the test were not madeavailable before the ROD and EPA made noreference to them for Sand Springs. Results re-cently made available were negative for organiccontaminants.

The way the remedy selection was made il-lustrates what can happen when there is muchpressure to issue a ROD by the end of the fiscalyear. ARCO submitted its treatability study re-sults to EPA on July 15, 1987. As late as Au-gust 21, 1987, results from ARCO’s pilot testswere still being obtained and disseminatedamong EPA staff. Several formal ARCO reportsare dated August 31, 1987, including one criti-cizing EPA’s RIFS. This site, like a very highpercentage of all Superfund sites, had its RODissued in the last days of the fiscal year (Sep-tember). In this case, EPA might have stayedwith its original, technically supported decisionand kept on schedule or delayed issuing theROD while it: 1) designed more tests for ARCOto carry out to convincingly demonstrate, be-fore actual use, the long-term effectiveness ofsolidification/stabilization technology for thediverse wastes at the site; and 2) developeddetailed protocols for future testing and moni-toring as well as technical criteria which wouldtrigger the switch to incineration, if solidifica-tion failed. In this case, the responsiveness ofEPA to local pressures seems to be related toselecting a lower cost technology and facilitat-

ing the government’s attempts to have a pri-vate party fund the cleanup.

As with the Compass Industries Superfundsite directly across the Arkansas River fromSand Springs, little attention was given to pos-sible groundwater contamination from the re-medial action for the present hazardous mate-rial and to the risks to downstream water usersfrom cumulative discharges into the river.

Case Study 9Schmalz Dump Site, Harrison, Wisconsin,

EPA Region 5

Capsule OTA findings.-A simple compacted earthcover over the soil contaminated with lead andchromium was selected. Solidification/stabili-zation treatment was rejected, although thiswas a textbook example of appropriate use ofthe technology. Voluntary well abandonmentand monitoring was chosen over pumping andtreating contaminated groundwater.

Key dates:●

●

●

●

●

●

●

●

Entered Superfund system: 5/1/83Preliminary Assessment: 5/23/83; (5/1/83 inCERCLIS)Site Inspection: 5/23/83; (5/1/83 in CERCLIS)National Priorities List–proposed date: 9/1/83–final date: 9/1/84—site rank: #190 out of 770RIFS start and completion: 12/28/84 to8/14/87Public comment period before Record ofDecision: 8/17/87 to 9/8/87Signing of ROD: 9/30/87Estimated complete remediation: 20 months(around 6/89)

Total time.—6 years

Brief description of site. –The ROD said: “TheSchmalz Dump Site is located on the northshore of Lake Winnebago in the Town of Har-rison, The site occupies approximately 7 acresin the Waverly Beach Wetlands area. The neigh-boring city of Appleton, with a population of60,000, has its drinking water intake approxi-mately 1200 feet from the shore of Lake Win-

nebago, in close proximity to the site.” Vari-ous types of waste were disposed there from1968 to 1979.

Major contamination/environmental threat.—Accord-ing to the ROD, there was “Initial on-site sam-pling by the State of Wisconsin and the U.S.Army Corps of Engineers in early 1979 . . . Leadand chromium were also detected in relativelyhigh concentrations at several sampling sta-tions. Lead and trivalent chromium [the leasttoxic form of chromium] were found through-out [soils at] the site at concentrations rangingfrom detection limits to 1940 milligrams perkilogram (mg/kg) and 964 mg/kg respectively. . . . lead and [trivalent] chromium in soils

. with concentrations greater than 14 mg/kgand 100 mg/kg respectively, pose an unaccept-able lifetime risk from direct contact.” The RODindicated a volume of contaminated materialof 8,000 cubic yards.

The second environmental threat was con-taminated groundwater. “ . . . the silty sandaquifer beneath the site appears to be separatedfrom the lower aquifer by a fairly thick, con-tinuous clay layer. It is therefore unlikely thatcontaminants from the site would enter thelower aquifer and reach residential wells. Theshallow aquifer beneath the site contains levelsof trivalent chromium above background butnot above the MCL [Maximum ContaminantLevel]. ., . the water table is three to five feetbelow the land surface and direction of flowis , . . towards Lake Winnebago.” Regarding theability of lead and chromium to leach from thesoil, the ROD said: “Results of the tests showthat very low levels of both lead and chromiumare leachable. . . . trivalent chromium has anaffinity to fine grained, silty soils like thosefound in the site area.” A groundwater model-ing study “indicates that in fifty years, ground-water containing chromium would have mi-grated just beyond the site boundary.”

A complex technical issue for this site is thatthere are two different chemical forms of chro-mium: the more toxic hexavalent form and theless toxic trivalent form. The analysis for soilcontamination recognized that even trivalentchromium is toxic. However, the analysis of

67

the groundwater threat discounted the hazardfrom trivalent chromium or from it turning intohexavalent chromium.

HRS scores.-groundwater 27.62; surface water80.00; air 0.00; total 48.92

Removal actions.—Earlier operable unit cleanupinvolved removal of PCB contaminated mate-rials in 1987; Remedial Design, $81,000 and Re-medial Action, $2.6 million.

Cleanup remedy selected. —This ROD is the sec-ond one for the site. The first ROD in August1985 addressed PCB contamination at the siteand selected “removal of construction debrisand sediments containing elevated concentra-tions of PCBs. ”

Alternatives examined included: 1) ground-water pumping with treatment, slurry wall andcap; 2) a RCRA cap; and 3) solidification/stabili-zation treatment of soil and onsite disposal ofresidues. The ROD said: “The preferred alter-native involves the installation of a low permea-bility, compacted earth material cap over ap-proximately seven acres of lead and chromiumcontaminated soils, and implementation ofgroundwater monitoring for lead and chro-mium. A voluntary well abandonment programfor nearby wells is also proposed.”

Estimated costs for the groundwater treat-ment alternative were $3.4 million; for solidifi-cation and/or stabilization, $2.8 million; for aRCRA cap, $2.4 million; and for the selectedremedy of a soil cover $800,000. (In the bodyof the ROD the selected remedy is called a soilcover, but in the beginning of the ROD it iscalled a cap. Soil cover is more accurate be-cause a cap implies a more complex, engineeredapproach to containing hazardous waste.)

Satisfaction of SARA statutory requirements:

1) Selection of permanent cleanup.—TheSchmalz ROD was straightforward: “The stat-utory preference for treatment set forth in . . .SARA is not satisfied because treatment wasfound to be impracticable due to questionabletechnical feasibility, inadequate short-term pro-tection, and inappropriate site conditions.”With regard to overall environmental protec-

tion, the ROD said the selected remedy “wouldprovide adequate protection from contami-nated soils on site.” Treatment was rejected forboth contaminated groundwater and soil. TheROD noted for both treatment approaches:“Treatability or compatibility testing is re-quired . . . prior to design and construction.”

EPA has said, however: “Toxic metals rep-resent a long term threat in the soil environ-ment. This threat can be reduced considera-bly if the heavy metals can be permanentlyimmobilized by either chemical or physicalmethods.” (U.S. Environmental ProtectionAgency, Review of In-Place Treatment Tech-niques for Contaminated Surface Soils, vol. 2,November 1984.)

The ROD acknowledged that solidification/stabilization of excavated soil could be “a per-manent remedial action to limit the off-site mo-bility, volubility [an unusual term instead of vol-ume] and toxicity of the heavy metals.” Theselected remedy “is expected to significantlyreduce the mobility of lead and chromium bycontainment in the site soils, but do nothingto reduce toxicity or volume of contaminants.”The chief reasons for rejecting solidifica-tion/stabilization were that:

●

●

●

✌

●

Excavation is risky “due to potential air-borne migration of dusts from the site.”“The reliability ., , is unknown principallydue to the lack of data documenting long-term success or failure of similar projects.”It “ . . . is not expected to significantly min-imize risks associated with ingestion ofsoils without additional restrictions on useof the site (e.g., additional fencing).”" . . . there is considerable research datato suggest that silicates used together witha cement setting agent can stabilize a widerange of materials including metals. How-ever, the feasibility of using silicates forany application must be determined on asite specific basis, particularly in view ofthe large number of additives and differ-ent sources of silicates which maybe used, ”(Interestingly, the paragraph is verbatimfrom the book Hazardous Waste TreatmentTechnologies, by Gerald Rich and Kenneth

68

●

●

●

Cherry, 1987; the statement is a good ex-planation of the need for a treatability studyfor solidification.)“Based on the content of soils on the site

. [the alternative] maybe difficult to im-plement. Contaminated soils consist ofsolid waste, wood, brick, and car bodies,which would make implementation dif-ficult.”its “ . . . reliability. . . is unknown. ”it” . . . is not conducive to a wetlands envi-ronment. Capping and vegetation of thesite is. ”

With the groundwater problem, there are twosubtle issues. First, the ROD emphasized thesampling data for chromium concentration inthe shallow aquifer, which ranged from 14 to48 micrograms per liter (µg/l) and the fact thatthe values are below the MCL of 50 µg/l. How-ever, the highest value is close enough to theMCL value to worry about the precision of sam-pling and the possibility of future increases inconcentration. Indeed, EPA work indicates thatchromium can be underreported by more thanenough to make the 48 µg/l observation abovethe standard. (K.A. Aleckson et al., “InorganicAnalytical Methods Performance and QualityControl Considerations,” Quality Control in Re-medial Site Investigation, American Society forTesting and Materials, 1986, pp. 112-123.) TheROD said: “Groundwater was determined notto be a public threat because chromium con-centrations are below [Safe Drinking Water Act]drinking water standards. However, leachingof chromium and/or lead to groundwater couldpotentially cause drinking water standards tobe exceeded. Based on the above discussions[of small amount of leaching according to stand-ard tests and affinity to soil], onsite soils arenot likely to ever increase chromium and leadconcentrations in the ground water to greaterthan the drinking water MCLs of 50 µg/l. How-ever, because there is a remote possibility thatthis pathway could later become a concern, itwas determined that groundwater should bemonitored over time. In addition, residents inthe vicinity of the site will be asked to volun-tarily abandon any existing wells. This is a

precautionary measure to ensure that no po-tential for exposure exists should contaminantlevels in groundwater increase in the future.”

The ROD’s case is supported by the resultsof leachability tests which found low levels ofboth lead and chromium; however, the test em-ployed does not necessarily describe long-termeffects under actual site conditions. The RODalso noted that the background groundwaterlevel for chromium is 5 µg/l; therefore, thereis contamination from the site. However, theROD also said: “groundwater contaminationis not above MCLs and there is no leachaterelease.”

The ROD for the Tower Chemical Superfundsite in Florida, where there is chromium con-tamination in the groundwater, had a targetcleanup level for the treatment of 100 milliongallons of water of 0.05 µg/l even though thegroundwater standard is 50 µg/l. The publichealth threat there is considered minimal.

The second groundwater issue is a secondplume of contamination which the ROD de-scribed as “an isolated off-site anomaly westof the Schmalz Site.” Two concentrations re-ported for the location are 185 µg/l and 1140µg/l of dissolved chromium. The contamination“appears to emanate from a localized pointsource. Based on the history of dumping in thearea, this phenomenon is not unusual.” It seemsthat a narrow definition of the site boundarykept these higher concentrations from beingconsidered a significant factor in selecting thecleanup remedy. Information in the ROD clearlyindicates that the second plume would also flowinto Lake Winnebago.

Overall, the critical ROD conclusion was:“Based on the rate of groundwater movement,and taking into consideration the dilution thatwould occur once ground water discharges tothe Lake, the levels of chromium in the ground-water should never pose a threat to Appleton’swater supply. ” Groundwater treatment wasalso rejected because “several problems can oc-cur at each component stage. This could resultin delays or inability to implement the alter-native. ”

69

In summing up its comparison of alternatives,EPA said that groundwater treatment “does notprotect against direct contact” and that thesolidification/stabilization treatment of the soil“would be protective upon implementation,however, there are several problems associatedwith implementation of this alternative thatmake it undesirable.” The use of a slurry walland cap or a RCRA cap are “not cost effectivebecause they provide excess protection forgroundwater.”

2) Accurate assessment of land disposal andcontainment alternatives. —The word perma-nent was not used to describe the overall se-lected remedy. The ROD said that the cap isprotective and that the groundwater monitor-ing “will provide protection from potential fu-ture releases. ”

Regarding permanency of the selected remedy,the ROD said that” . . . the only potential needfor replacement is seen to be that of the capor soil cover. This need could occur if the origi-nal cap was washed out by some storm event,if heavy equipment were to abrade the cover,or if unforeseen subsidence were to occur.” Norestrictions on future land use were set, andthe ROD noted that EPA has legal authority “toissue an order for corrective action, should theowner make an attempt to damage the cap.”A letter from the State said, “the cap could bedamaged by the landowner, who has indicateda desire to build on the site.” The ROD for theTower Chemical site in Florida eliminated thealternative of using a soil mixture cap because:“This technology is unproven and has exten-sive monitoring requirements. Development ofdessication cracks could cause failure. Highfailure potential” (emphasis added).

In the FS for the French Limited site (in EPARegion 6), use of a slurry wall and cap to con-tain hazardous waste was described as a “tem-porary solution” for which the “volume andtoxicity would not be affected . . . [and] . . . thepotential would always exist for failure of ei-ther the cap or the slurry wall allowing for themovement of unstabilized wastes containedonsite. ”

The decision not to use a RCRA cap was in-consistent with statutory requirements aboutsatisfying current regulations, and it raises sig-nificant uncertainties about future failure, Offi-cial EPA guidance notes: “A key task of coverdesign is the selection of suitable materials. Thecover usually will include a synthetic mem-brane and a large volume of soil or soil-like ma-terial . . . “ (U.S. Environmental ProtectionAgency, “Project Summary-Design, Construc-tion, and Maintenance of Cover Systems forHazardous Waste: An Engineering GuidanceDocument,” November 1987.) The selectedcover is very simple and poses substantial un-certainty about future risks. In the ROD for thePristine site EPA defends its rejection of cap-ping by saying” ., . there are no data availableon the long term effectiveness and permanenceof RCRA caps. ” Clearly the situation is worsefor a simple soil cover.

With regard to groundwater monitoring:“Any increase in existing levels of chromiumor lead will be evaluated as to whether correc-tive action is necessary based on levels found.”The ROD did not give any specific technicalcriteria for deciding when other remedial ac-tions will be necessary.

There was no discussion of the possible fu-ture oxidation of trivalent chromium to themore toxic and mobile hexavalent chromium.EPA research has noted: “under conditionsprevalent in many soils, Cr(III) can be oxi-dized.” (U.S. Environmental Protection Agency,Review of In-Place Treatment Techniques forContaminated Surface Soils, vol. Z, November1984,) Such oxidation constitutes a potentialmode of failure for the selected remedy, espe-cially in the context of future land use. AlthoughOTA does not know whether the site soil posesthis problem, it is an important enough issueto have been examined by EPA. Another po-tential effect is that trivalent chromium in watercan be oxidized to the hexavalent form in cer-tain types of chlorination treatments; therefore,leaching of trivalent chromium into ground-water that eventually enters a drinking watersupply can be a problem, especially if it is un-expected.

70

RIFS contractor.—Camp Dresser and McKee;over $600,000 for all of the RIFS work.

State concurrence.— “The State of Wisconsinsupports our preferred alternative, however ithas several concerns related to implementationof the remedy. . . . due to the excess cost in-volved, they do not feel that a groundwatertreatment alternative is warranted. The Statehas concerns over whether adequate cap pro-tection is available for alternatives involvingcapping the site. . . . the State has agreed to at-tempt to obtain a voluntary agreement from thelandowner.”

Community acceptance. —The ROD said the “com-munity does not perceive the site as an imme-diate danger. . . . at least some residents feelthat a full RCRA Subtitle C Cap should be in-stalled.”

Special comments.—Although the ROD said theselected soil cover had a low permeability, un-like some other RODS no specific permeabilityvalue was given.

Documents obtained from EPA on the useof treatability studies show that a study wascompleted for the Schmalz site in April andMay 1987; the study focused on “surface watercontaminated with PCBs, Cr [chromium], andPb [lead].” This attempt to verify physical andchemical treatment may have been done as partof the earlier operable unit cleanup for PCBs,but the results are relevant to groundwatercleanup for the Schmalz ROD, which does notmention the study.

In the RIFS, the detection limit for hexava-lent chromium was 20 times higher than fortrivalent chromium, which might explain whyhexavalent chromium was not found, if it waspresent.

Some of the estimated costs for the rejectedcleanup alternatives look high. For example,the estimate of $2.8 million for stabilization ishigher than estimates at other Superfund sites.With a figure of $200 per cubic yard for totalstabilization costs, consistent with data at theSand Springs and Liquid Disposal sites wherestabilization was selected, the correct value forSchmalz is probably about $1.6 million.

The body of the ROD includes a part of theselected remedy that is not in the remedy’sdescription at the beginning of the document.That is the recommendation that “adjacentproperty be evaluated under the pre-remedialprogram.” Presumably this action refers to thehot spot of contamination just outside the siteboundaries used by EPA. Such an examination might, therefore, open up the possibility ofanother site cleanup, but the ROD only recom-mended the evaluation instead of requiring it.

General conclusions.-The Schmalz site is defi-nitely not one of the worst Superfund sites. Buteven though the environmental threat from thesite is not severe, the handling of the remedialcleanup raises important questions. The degreeof certainty expressed by EPA for the long-termeffectiveness of its selected remedy is incon-sistent with the technical limits of the remedy.A good example of EPA’s over optimism is inthe responsiveness summary: “Following im-plementation of the selected remedy, exposureto contamination from the Schmalz Site willbe eliminated” (emphasis added). This state-ment is inconsistent with the technical limita-tions of a soil cover and with uncertain failures,responses to monitoring results, and land use.

Moreover, ROD statements that the “ground-water is not contaminated” are incorrect. Theissue is whether the risks estimated by EPA arecorrect and stable or whether groundwatercleanup is warranted now or whether it maybecome necessary. EPA seemed to place heavyemphasis on a technicality, namely that con-tamination within the bureaucratic boundariesof the site was slightly below the current regu-latory standard for allowable chromium con-tamination in drinking water. What if the meas-ured chromium contamination goes up the 5percent necessary to bring it over the stand-ard? Moreover, the ROD’s selected remedyomits the statutory requirement of reexamin-ing the site at least every 5 years because un-treated hazardous waste will be left onsite.

The volume of contaminated soil (8,000 cu-bic yards) at Schmalz is relatively small for acleanup site. Nevertheless, various ROD state-ments indicate that minimizing cleanup costs

I

71

was a significant motivation in the remedyselection. Selecting both soil and groundwatertreatment was estimated to cost eight times asmuch as the soil cover. The ROD said alterna-tive treatment remedies for both contaminatedsoil and groundwater were rejected because“they are more costly while achieving the samedesired results” as the selected remedy. Onlyby using a narrow, short-term objective of limit-ing exposure to the hazardous waste are therecomparable environmental results of soil treat-ment and a soil cover.

For the soils problem, the rejection of thesolidification/stabilization alternative is incon-sistent with generally understood capabilitiesof the technology. Indeed, having soil contami-nated with only two toxic metals, lead and chro-mium, offers a textbook example of when chem-ical fixation works best. None of the argumentsgiven in the ROD against using this soil treat-ment describe especially unique or difficultproblems. For example, the presence of largeburied objects is faced routinely; they can bewashed and reburied. Regarding implementa-tion risks, wetting excavated materials for dustsuppression is routine and, in exceptionalcases, inflatable domes have been used. Thetechnology has been selected for Superfundsites posing much more challenging kinds andlevels of contamination for which solidificationis unproven.

The argument that stabilization technologyneeds to be verified for a site is an argumentfor conducting a treatability study, preferablyduring the RIFS, not for rejecting the alterna-tive. If, as the ROD acknowledged to be the case,the soil is contaminated enough to pose a trueenvironmental risk, then the selected remedyof a soil cover is not permanently effective. Theabsence of land use restrictions is particularlyworrisome.

The rejection of groundwater treatment doesnot consider several factors: 1) values for chro-mium contamination very close to the MCL;2) the hot spot of high chromium contamina-tion apparently just outside the site boundary;and 3) the exact environmental and health ef-fects, which might not be eliminated by dilu-

tion, resulting from likely continued leachingand migration of contaminants into Lake Win-nebago, even if water intrusion is curtailedthrough the surface of the site. The problemis that groundwater movement will still occurbeneath the surface where the contaminatedsoils reside. At the Liquid Disposal Superfundsite in Michigan, a RCRA hazardous waste capand a containment wall will be built aroundchemically stabilized material, landfilled onsite,to prevent just this type of leaching and move-ment of contaminated groundwater.

Moreover, in the Schmalz ROD, no specificcriteria were given for groundwater monitor-ing or for triggering a decision that a ground-water remedy and a better soils remedy areneeded. This site highlights a problem foundby recent EPA research: “many [Superfund] in-vestigations are not producing sufficient datato adequately characterize ground water con-ditions near these sites.” (R.H. Plumb, Jr., “AComparison of Ground Water Monitoring DataFrom CERCLA and RCRA Sites,” Ground WaterMonitoring Research, fall 1987, pp. 94-100.)

Case Study 10Tacoma Tar Pits, Tacoma, Washington,

EPA Region 10

Capsule OTA findings:–No treatability study re-sults supported the selection of chemical stabili-zation. Significant amounts of untreated con-taminants as well as the treated materials willbe left onsite. The effectiveness of the treatmentis uncertain. Incineration was said to offer nobetter protection and was rejected because ofits higher cost.

Key dates:●

●

●

●

●

●

Entered Superfund system: 2/1/82Preliminary Assessment: 4/1/82Site Inspection: 3/1/83National Priorities List–proposed date: 10/81–final date: 9/83—site rank: #347 out of 770RIFS start and completion: 11/84 to 9/87public comment period before Record ofDecision: 11/6/87 to 12/6/87

72

● Signing of ROD: 12/30/87● Estimated complete remediation: Assume

2 years after ROD

Total time.—7 years

Brief description of site.-The site is approximately30 acres and “within a heavily industrializedarea . , . [the] site is part of the CommencementBay-Nearshore/Tideflats Superfund site locatedwithin the Tacoma Tideflats industrial areanear Commencement Bay.” (The ROD does notdescribe this action as an operable unit, but thatis what it appears to be. However, Tacoma TarPits is listed as a separate site in CERCLIS, threeyears after the large Commencement Bay siteentered the system.) “A coal gasification plantwas in operation on site from 1924 through1956. The study area currently contains a metalrecycling facility . . . a natural gas transfer sta-tion . . . a rail freight loading yard . . . a meatpacking plant . . . and a railroad switchingyard . . . “

Major contamination/environmental threat.—"The sitecurrently contains two ponds, a small tar pit,and various surface-water drainage ditches.The study area is located near several majorsurface water bodies [waterways]. . . the Puyal-lup River, and Commencement Bay. Althoughnone of these water bodies are used for watersupply, the bay and river do support extensivefish and shellfish populations. Several portionsof Commencement Bay have been identifiedas being severely contaminated, resulting in ad-verse biological effects . . . contamination of thelocal groundwater resource is also of concern.Many local industries use groundwater fromon-site wells . . . “

With regard to site contaminants, first foundin 1981: “Many of these organic compoundsare toxic and several are considered to be car-cinogenic. These compounds include aromatichydrocarbons (i.e., benzene, toluene), poly-nuclear aromatic hydrocarbons collectivelyknown as PAHs (i.e., napthalene, benzo(a)py-rene), as well as numerous other classes ofhydrocarbons and cyanide. Heavy metals . . .include arsenic, mercury, and lead.” The au-tomobile recycling facility has also caused leadand PCB contamination. The estimated volume

of tar is 5,000 cubic yards; it is mostly in threeareas at depths of several feet and more. PCBsare found up to 204 parts per million (ppm),and lead in soil in the 2,000 to 8,000 ppm range.Three shallow aquifers have varying degreesof contamination, and the ROD noted the “cur-rent lack of understanding of local groundwaterhydrology.”

●

Four indicator compounds were used to esti-mate risks and establish cleanup goals: ben-zo(a)pyrene, PCBs, benzene, and lead. For themost part, the cleanup levels are consistent witha 1 in 1 million cancer risk level, except for leadin soil which seems high relative to the MCLvalue. The major emphasis was correctly put onexposure of onsite workers over short periods.

HRS scores.—groundwater 6.12; surface water10.91; air 71.92; total 42.20

Removal actions.—None indicated.

Cleanup remedy selected. –This ROD apparentlyis for an operable unit of the larger Superfundsite of which it is a part, even though it did notuse the term. Besides the selected remedy,cleanup alternatives examined included con-tainment and landfilling, incineration, in situvitrification, and groundwater treatment.

“The preferred remedial alternative . . . is acombination of source control measures, meas-ures to control contaminant release, and alsomeasures to reduce human exposure to con-taminants. This alternative consists of the ex-cavation of the most severely contaminatedsoils, stabilization of these soils using a tech-nique which immobilizes contaminants, cap-ping of the stabilized material [with asphalt],treatment of the surface water, continued ground-water monitoring, regulatory controls on waterusage for both surface and groundwater, andrestrictions on site access. ” Thoroughly mixedexcavated materials will be “fed to a mixingvessel where silicate polymers, cement, andwater from the site ponds is added.”

However, in site areas that are not severelycontaminated with PAHs, soils and sediments“will be excavated to a depth not to exceed 3feet.” This requirement means that significantamounts of contaminants may not be excavated

—

73

and treated. Exactly what amount will not beexcavated cannot be judged from the informa-tion given in the ROD, but the site’s history andcomplex contamination suggests that this maybean important limitation to the selected rem-edy. Many of the contaminants have beenpresent long enough to have migrated down-ward a significant distance. “The total esti-mated volume of material to be excavated is45,000 cubic yards.” The total cost for theselected remedy is $3.4 million. This cost im-plies a rather low cost for the stabilization partof the cleanup of about $50 per cubic yard.

Satisfaction of SARA statutory requirements:

1) Selection of permanent cleanup.—TheROD said: “This remedy satisfies the prefer-ence expressed in SARA for treatment that re-duces toxicity, mobility, and volume. . . . it isdetermined that this remedy utilizes permanentsolutions and alternative treatment technol-ogies to the maximum extent practicable.”

The uncertainty about contaminants left un-treated onsite means the selected remedy maynot be permanent. The intent to comply withSARA’s requirement for 5-year review confirmsthat the selected remedy is not assuredly per-manent. This situation raises questions aboutfuture land use. Further uncertainty about per-manence is indicated by the ROD’s comment:“If as a result of this frequent reassessment,the remedial action is shown to have decreasedperformance, the nature and extent of addi-tional actions will be considered.”

An important issue for Tacoma Tar Pits con-cerns the cleanup of contaminated ground-water. Although the ROD contained cleanupstandards for groundwater, “the remedial ac-tion does not currently provide for groundwaterextraction and treatment.” The basis for thislack of cleanup was that the selected remedywill reduce surface water intrusion and con-taminant flow into the water, that existing con-tamination will be swept away and into its ulti-mate discharge, and that “Action levels ofcontaminants in groundwater have not beenconsistently exceeded at off-site locations. ” Aclaim of permanence at this time is prematurebecause the ROD said that it may become nec-

essary to evaluate and implement an alterna-tive remedial action that includes groundwaterextraction. A commitment to groundwatermonitoring was made, but given the acknowl-edged complexity of local hydrogeology andgiven EPA’s interest in minimizing cost, it ne-cessitates a major, carefully planned effort.

With regard to the selected stabilization treat-ment technology: “No bench or pilot studieshave been performed to date, these being leftuntil the Remedial Design is commenced . . . “The diverse and highly concentrated contami-nants pose a major challenge for a chemicalstabilization technology. The current state ofknowledge and experience does not support anassumption of effectiveness for the Tacoma TarPits site. The ROD noted: “Laboratory experi-ments will be performed to ensure that thestabilization process effectively immobilizescontaminants. Following this activity, a largerscale ‘pilot study’ will stabilize a larger volumeof contaminated material from the site. Thispilot study will determine the effectiveness ofthe stabilization process.” The uncertaintyabout effectiveness means that the selectedremedy did not merit the high (maximum pos-sible) rankings it received for effectiveness inthe ROD’s analysis of cleanup alternatives.Moreover, there is no basis for saying: “Thechemical stabilization process should signifi-cantly reduce the toxicity and leachability ofsite soils,” The ability of any chemical stabili-zation technology to reduce toxicity of a widerange of organic and inorganic contaminantshas not been proven nor is it generally acceptedin the technical community, But it is reason-able to claim that the mobility of the contami-nants might be reduced through stabilization;to claim more than that would require actualtest data on site materials, Elsewhere the RODsaid: “Permanent treatment can be providedthrough the immobilization of contaminants.”This statement is, however, overly emphatic atthe current stage of knowledge about the site.

An important statement on the selectedstabilization technology was in the responsive-ness summary, given in response to a concernabout its effectiveness: “Although the cement/polymer stabilization process is a proven tech-

74

nique for immobilization of heavy metals, thistechnique has not been conclusively proven tobe effective in immobilizing organic contami-nants in coal tars, Therefore, both laboratoryand bench scale treatability studies will beperformed during the design phase of the re-medial action to ensure the process will be ef-fective and permanent. . . . the soils/tars con-taining the highest tar content . . . may beconsidered for an alternate type of treatment/disposal (i.e., incineration) if the stabilizationprocess is found to be ineffective for the wastematrix” (emphasis added). Also: “Criteria to beused to evaluate the effectiveness of the stabili-zation process during laboratory and benchscale studies . . . “ will be addressed in the de-sign phase.

The FS for the Re-Solve site in Massachusettsrejected stabilization for organic contaminationbecause “there has been limited success inchemically fixing organic contaminants suchas solvents and PCBs,” The ROD for the Liq-uid Disposal site in Michigan, which alsoselected stabilization for soil contaminated withorganic chemicals, said that the hazardous sub-stances “will not be permanently destroyed.”And the FS for Liquid Disposal said: “Consid-erable research data exists demonstrating theeffectiveness of this technology in immobiliz-ing a wide range of contaminants, primarilyinorganic. A substantial amount of data doesnot exist, however, to accurately judge the long-term reliability of the process.”

The ROD for the Tower Chemical Superfundsite in Florida said the following about chemi-cal stabilization processes, which it rejected:“This technology would also require long-term(30 years) monitoring which is less favorablethan technologies which provide permanent de-struction of wastes.” The inference is that thetechnology does not provide permanent de-struction of wastes.

Another EPA document, used to teach peo-ple about waste treatment said: “Solidificationtechnologies are designed to be used for finalwaste treatment. This means the technologyshould be applied only after other treatmenttechniques have been applied, i.e., incineration,

chemical treatment or other. ” (U.S. Environ-mental Protection Agency, “RCRA/CERCLATreatment Alternatives for Hazardous Wastes,”October 1987.)

A recent EPA study found “large losses oforganics during the mixing process.” (L. Weitz-man et al., “Evaluation of Solidification/Stabili-zation As A Best Demonstrated Available Tech-nology,” paper presented at EPA’s FourteenthAnnual Research Symposium, May 1988.]Another EPA study showed that stabilizationwas not competitive with thermal and chemi-cal treatment technologies and soil washing fororganic contamination. (R.C. Thurnau and M.P.Esposito, “TCLP As A Measure of TreatmentEffectiveness: Results of TCLP Work Com-pleted on Different Treatment Technologies forCERCLA Soils,” paper presented at EPA’sFourteenth Annual Research Symposium, May1988.) A demonstration of a stabilization tech-nology under EPA auspices concluded that “forthe organics, the leachate concentrations wereapproximately equal for the treated and un-treated soils.” (P.R. de Percin and S. Sawyer,“SITE Demonstration of Hazcon Solidifica-tion/Stabilization Process,” paper presented atEPA’s Fourteenth Annual Research Sympo-sium, May 1988.)

The rejection of other treatment technologiesfor Tacoma Tar Pits did not have much techni-cal analysis behind it. The analysis in the RODrests mainly on a very simple rating system.For example, all 10 alternatives, including cap-ping the waste and incinerating it, received thesame rating of high for technical feasibility (in-cluding effectiveness, useful life, operation andmaintenance requirements, possible failuremodes, constructability, implementation time,worker safety, and neighborhood safety). Allbut the two no action or nearly no action op-tions received the same high rating for publichealth impacts (including minimization ofchemical releases, exposures during remedialaction, and exposures after remedial action).But sound technical bases exist for finer dis-tinctions among such abroad range of alterna-tives. For example, the incineration optionsoffered substantially greater effectiveness, relia-bility, permanency, and certainty of destruc-

tion of toxic substances than capping the wasteor the selected remedy.

The estimated costs of the alternatives prob-ably weighed heavily: two incineration options(including stabilization of residue) had total

costs of $17 million (only surface soils) and $243million (all soil with contamination with a riskgreater than 1 in 1 million cancer risk); bothoptions included groundwater pumping andtreatment. In comparison, the selected remedy(stabilization) would cost $8 million if ground-water treatment is included in the calculation.The cost will be only $3.4 million, however, be-cause groundwater treatment was excluded.

2) Accurate assessment of land disposal andcontainment alternatives.—There is a statementon future land use: “Land use restrictions willbe imposed to prevent or require stringent con-trol of future excavation on the site, to preventfuture use of surface water and shallow ground-water, and to prevent site access by personnelother than site workers. ” However, there is nodetailed analysis of possible future failures ofthe landfill in which the contaminated materi-als will be re-buried after stabilization.

The ROD for the Tower Chemical site madea good point about concrete or asphalt caps (asselected for Tacoma Tar Pits): “The risk of fail-ure . . . is high due to the potential for fractureformation.” The FS for the Pristine site says:“Asphalt is photosensitive, and subject to crack-ing due to settling, chemical action, and vege-tation. Frequent inspections are required to en-sure cap integrity. ” The asphalt cap option wasrejected at both of these sites.

RIFS contractor. –The studies were paid for byresponsible parties. The ROD noted that, al-though EPA and the State found the documentsacceptable, “EPA has prepared an addendumfor each document addressing issues that thestudies have inadequately or incompletely ad-dressed.” Geotechnology, Inc., performed theRI; Envirosphere Company (Ebasco) performedthe FS. For the entire Commencement Bay/Nearshor/Tideflats site, the SCAP indicatesthree different RIFSs with the last one labeledTar Pits started on 9/23/83 and then taken overby the PRPs on 11/1/84. It is not clear what ac-

tions resulted from the two earlier RIFSs andno completion dates for them are indicated; $2.5million was spent by the government on thefirst one in 1982 and 1983.

State concurrence.—” The State of Washingtonhas been consulted and has verbally concurredwith the selected remedy.” Verbal concurrencemay indicate a rushed ROD at the end of thefiscal year quarter.

Community acceptance.—The ROD noted thatcommunity interest “has not been actively dem-onstrated.” The reasons given for the lack ofcommunity interest are the site’s location withinthe larger Commencement Bay Superfund site,the lack of private residences nearby, and anumber of cleanup actions already taken in thearea.

Special comments.—Although there are state-ments about restricting future land use, thereare also statements that suggest that those re-strictions may be applied only in the short term,For example: “The [stabilization] reagent com-position is formulated to provide a high-strengthsurface capable of supporting trucks and othervehicles. ”

Cleanup goals were set for indicator contami-nants. While these goals make risk assessmentmore manageable, there can be problems withusing them for analysis of the effectiveness ofa cleanup technology that is chemical specific,such as chemical stabilization. Moreover, thecleanup very much depends on data that re-veal areas of high tar concentration, wherethere is no excavation depth limit. However,there may be other areas that have high con-centrations of other contaminants and that maybe either overlooked or fall under the provisionof the excavation depth limit. Compounding theproblem is the relatively small amount of soilsampling that has been reported, averaging onlyabout 1.5 locations per acre.

The administrative record indicated a largenumber of contractors have performed studieson the site. It is not clear whether EPA had in-dependent work done to verify work done forthe responsible parties. There is also some con-fusion about the relationship between Tacoma

76

Tar Pits and the larger Commencement Baysite because the former is listed separately inCERCLIS and the ROD does not use the termoperable unit.

General conclusions.—The technical informationobtained prior to and used in the ROD did notsupport claims that the selected remedy ispermanent nor even that it will be effective. De-laying testing of the chosen stabilization tech-nology and setting of criteria for its effective-ness until after the ROD undercuts the claimthat a permanent remedy has already beenselected. The scope and depth of analysis ofalternative cleanup technologies was less thanseen in any other ROD examined in this report.This shortcoming directly affected technologyselection; it maybe related to the strong involve-ment of the responsible parties, particularly inconducting the RIFS. (An experienced attor-ney advises responsible parties: “Participationin the IRIFS] study provides an opportunity togenerate information that can sway EPA deci-sion makers on important issues. We all knowthat one can interpret the same data a numberof ways. Differing conclusions can be made andsupported from the same data. A company thatuses its experts to argue convincingly in favorof one conclusion often can influence the ulti-mate decision.” [P. H. Hailer, Hazardous Ma-terials, January-February 1988.])

While the ROD’s interest in alternative treat-ment technology is commendable, the chiefdriving force for selecting the remedy appearsto be cost: “The final selected remedy meetsthe requirement of cost-effectiveness as thisalternative provides for permanent treatment,and contaminant release minimization for acost significantly less than other alternativesexhibiting a similar level of protection. Addi-tional cost of these [other alternatives] is theresult of the use of more costly technologiessuch as incineration . . , or the excavation oflarger volumes of soils coupled with off-sitelandfilling.” But no data support the conten-tion of similar or equal levels of protection for

stabilization and incineration; therefore, theclaim that the selected remedy is cost-effectiveis unsupported.

However, incineration for a comparable vol-ume of contaminated soil definitely would bemuch more expensive ($242 million) and, there-fore, this site, like many others, shows howimportant it is to examine the issue of compara-ble environmental protection for a cost-effec-tiveness decision. Moreover, Tacoma Tar Pits .illustrates the need to consider a broader rangeof treatment technologies to reduce cleanupcosts. Biological treatment for such a sitedeserves attention. The case for its considera-tion at Tacoma Tar Pits, for example, was asgood, if not better, than for the Renora sitewhere it was chosen without treatability testdata to support the decision. A research pro-gram on developing biotechnology for clean-ing up old manufactured gas plant sites suchas Tacoma Tar Pits is underway at the Univer-sity of Tennessee with support from the GasResearch Institute, The choice at Tacoma TarPits could have included postponing the reme-dial action or conducting treatability studiesfor biological treatment.

Moreover, the cost of the selected remedymay have been significantly under estimated.Data from a vendor of the stabilization tech-nology most likely to be effective on this sitesuggests a cost of about $150 per cubic yardinstead of the $50 indicated in the ROD. If ma-jor costs for treatability studies are added in,the cost of the selected remedy could be about$5 million more than the $3.4 million estimatedin the ROD for a total of $8.4 million. The highercost matches the low range of the ROD’s inciner-ation options without groundwater treatment.

The lack of a commitment to groundwatercleanup may be linked to the location of thesite in a highly contaminated area, includingcontaminated major water bodies. This is a sit-uation where analyzing the cleanup in isola-tion may be misleading and ultimately in-efficient.