AD-A240 444IBl~iJlhIN~hIlllhiII(

Defense Nuclear AgencyAlexandria, VA 22310-3398

... 'DNA-TR-89-20-V1

Target Area StudiesViume I-Combustible Fuel Loads in Nashville, Tennessee

Russell A. GajMichael A. Dore.Richard D. SmallPacific-Sierra Research Corporation12340 Santa Monica BoulevardLos Angeles, CA 90025-2587

September 1991

Technical Report

CONTRACT No. DNA 001-88-C-0119

Approved for puli elae;distribution is unlimited.

-112 91-10792-4i31 i

Destroy this report when it is no longer needed. Do notreturn to sender.

PLEASE NOTIFY THE DEFENSE NUCLEAR AGENCY.ATrN: CSTI. 6801 TELEGRAPH ROAD. ALEXANDRIA. VA22310-3398. IF YOUR ADDRESS IS INCORRECT. IF YOUWISH IT DELETED FROM THE DISTRIBUTION LIST. ORIFTHEADDRESSEE IS NO LONGER EMPLOYED BYYOUR

ORGANIZATION.

REPORT DOCUMENTATION PAGEP e=: b-f io __= cogs= ofI" c I -S f: 10 Za~ e I h=WMr ects -,e~ U in refwawV 91r W SexciI CErun ea sz-ws.

C!. ---. it s wl~nrý Hea==W e~es D~eae t, 0es =4Re~s 12,51,11~03's ~ Sum 1204 A.- , vA 20-93= aa I* -.%e 0!e cl V ý =4 &.49m Pe k Rf P.d ( 9-0153). W M O DC 2N0S

1 AGENCY USE ONLY (Leave blank) 2. REPORT DATE 3. REPORT TYPE AND DATES COVERED

7 910901 Technical 850519 -8812234. TITLE AND SUBTITLE 5. FUNDING NUMBERS

Target Area Studies C - DNA 0O1-88-C-0119Volume I-Combustible Fuel Loads in Nashville, Tennessee. PE - 62715H

6. AUTHOR(S) PR - RMTA -RA

Russell A. Gaj. Michael A. Dore. and Richard D. Small WU - DH047780

7- PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES) a. PERFORMING ORGANIZATION

REPORT NUMBER

Pacific-Sierra Research Cornoration12340 Santa Monica Boulevard PSR Report 1842Los Angeles. CA 90025-2587

9- SPONSORINGMONITORING AGENCY NAME(S) AND ADDRESS(ES) 10. SPONSORINGJMONITORINGAGENCY REPORT NUMBER

Defense Nuclear Agency6801 Telegraph Road DNA-TR-89-20-V1Alexandria. VA 22310-3398RARP/Hartley

.1. SUPPLEENTARY NOTES

This work was sponsored by the Defense Nuclear Agency under RDT&E RMC CodeB4662D RM RA 00001 RARP 3500A 25904D.

12a- DISTRIBUTIONIAVAILABILITY STATEMENT I12b. DISTRIBUTION CODE

Approved for public release: distribution is unlimited.

13. ABSTRACT r"axv=ru 2W0 words)

Standard land use and census data is used to construct a detailed combustible fuel loadmap for a typical medium-sized U.S. city (Nashville. Tennessee). Our method recognizes elev-en land use categories. Each category is assigned fuel loading values based on analyses ofthe total burnable fuel mass and areal densities of the buildings and vegetation present Wefind that the distribution of fuels is non-uniform and highly dependent on the internal struc-ture of the city. The highest fuel load densities (>40 kg/rn 2 ) are found in portions of the cityoccupied by densely-packed multiple-family residential buildings. liquid fuel storage termi-nals, and in industrial/commercial complexes. Much lower fuel loads (<5kg/mi2) are found inoutlying suburbs and rural areas. This uneven fuel distribution results in a several-fold dif-ference in the net fuel availability for potential urban targets spaced only a short distanceapart- We demonstrate that this may in turn strongly influence the intensity of nuclearburst-generated fires, the amount of smoke emitted, and the altitudes to which the smoke Isinjected.

14. SUBJECT TERMS 15. NUMBER OF PAGESFuel Loads Urban Fires Fire Plumes 36Nuclear Winter Urban Geography Land Use 16. PRICE CODESmoke

17 SECURITYCLASSIFICATON 18. SE-CURITYCCLASSIFICATION 19. SEICURITY CLASSIFICATION 20. LIMITATION OF ABSTRACTOF REPORT OF THIS PAGE OF ABSTRACT

UNCLASSIFIED UNCLASSIFIED UNCLASSIFIED SAR

Stan 1azd Fo--m 298 (Rev2-89)i -

UNCLASSIFIEDSEEM~r-y CLASS!-ICA-.1'd 0= TH~S PAGE

CLASSIFIED BY:

N/A since Unclassified

DECLASSIFY ON.

NI/A since Unclassified

SEU Ly ASS5ý=71IVOF THS AG~

PREFACE

This report continues the Pacific-Sierra Research Corporation (PSR) study of the ef-fects of fire generated by nuclear weapons. In this volume, we map the distribution ofcombustible fuels in a representative medium-sized U.S. city (Nashville, Tennessee).The relation between this distribution and the internal structure of the city is dis-cussed. In Vol- 2. these fuel loads are used to calculate the atmospheric smoke injec-tion for specific urban targets and collateral areas.

This research was supported by the Defense Nuclear Agency under contractDNA001-88-C-01 19 and was monitored by Dr. David Auton.

* - /

%- -

111-~1

CONVERSION TABLE

Convers'on factors for U.S. customary to metric (SI) units of measurement

To Convert From To Multiply

angstrom meters (m) 1.000 000 X E-10

atmosphere (normal) kilo pascal (kPa) 1.013 25 X E+2

bar kilo pascal (kPa) 1.000 000 X E+2

barn meter 2 (M2) 1.000 000 X E-28

British Thermal unit (thermochemIcal) jouie (J} 1.054 350 X E+3

calorie (thermochemical) joule [J} 4.184 000

cal (thermochemical/cm-2 mega jule/mIMJ/m2-) 4.184 000 X E-2

curie gMga becquerel (GBqr 3.700 000 X E+Idegree (angle) radian (rad) 1.745 329 X E-2

degree Fahrenheit degree kehlin (KW tr,=(tof + 459.67)/1.8

electron volt joule (J) 1.602 19 X E--19

erg joule (JM 1.000 000 X E-7

erg/second watt (IV] 1.000 000 X E-7

foot meter (ml 3.048 000 X E-I

foot-pound-force joule MJ} 1.355 818

"gallon (U.S. liquid) meter3 (M

3) 3.785 412 X E-3

inch meter (m) 2.540 000 X E-2

jerk jouleiJ) 1.OW O0 X E+9

Joule!kilogram IJ/Kgl (radiation doseabsorbed) Gray (Gy) 1.000000

kilotons terajoules 4.183

kip (1000 Ib)b newton MN) 4.448 222 X E+3

kip/Inch2- (ksl) kilo pascal ,kPa) 6.894 757 X E+3

ktap newton-second/m 2 (N-s/re2 ) 1.000 000 X E+2

micron meter (m) 1.000 000 X E-6

rIl meter (m) 2.540 000 X E-5mile (international) meter (m) 1.609 344 X E+3

ounce kilogram (kg) 2.834 952 X E-2

pound-force [lbf avoirdupois) newton (N) 4.448 222

pound-frce inch newton-meter (N-m) 1.129 848 X E-I

pound-force/inch newton/meter (N/m) 1.751 268 X E+2pound-force/foot 2 kilo pascal (kPa] 4.788 026 X E-2

pound-force/lnch2 (psI) kilo pascal (kPa? 6.894 757

pound-mass (Ibm avoirdupois) kilogram (kg) 4.535 924 X E-Ipound-mass-foot2 (moment of Inertia) kilogram-meter 2 (kg-m2) 4.214 011 X E-2

pound-mass/foot 3 kilogram/meter 3 (kglm3) 1.601 846 X E÷ Irad (radiation dose absorbed) Gray (Gyr* 1.000 000 X E-2roentgen coulomb/kilogram (C/kg) 2.579 760 X E-4

shake second (s) 1.000 000 X E-8

"slug kilogram (kg) 1.459 390 X E+ I

tort (mm Hg. 0&C) kilo pascal (kPa) 1.333 22 X E--I

*The becquerel (Bq) is the SI unit of radloactUIty: Bp - i event/s.

**The Gray MGy) is the SI unit of absorbed radiation.

iv

TABLE OF CONTENTS

Section Page

PREFACE ...................................................... i-

CONVERSION TABLE ........................................... iv

FIGURES .................................................. vi

1 INTRODUCTION..............................................1

2 NASHVILLE: GEOGRAPHY AND LAND USE ..................... 3

2.1 GENERAL DESCRIPTION .................................... 3

2.2 LAND USE .................................................. 6

3 FUEL LOAD CALCULATION ..................................... 14

3-1 NONRESIDENTIAL LAND ................................... 14

3.2 RESIDENTIAL FUELS ....................................... 16

3.3 FUEL LOADING MAP.....................................17

4 IMPLICATIONS FOR SMOKE INJECTION ........................ 20

5 CONCLUSIONS ................................................. 26

6 LIST OF REFERENCES ......................................... 27

S... . .

FIGURES

Figure Page

1 Metropolitan Nashville. Tennessee ............. _ ..................... 4

2 Land-use summary, Nashville-Davidson Country .................... 5

3 Residential land, metropolitan Nashville ............................. 7

4 Single-family housing density (structures per square kilometer) ...... 8

5 Multiple-family building density (structures per square kilometer) .--- 9

6 Industrial land, metropolitan Nashville .............................. 11

7 Commercial/service land, metropolitan Nashville .................... 12

8 Agricultural land, metropolitan Nashville ............................ 13

9 Fuel load density (kilograms per square kilometer),metropolitan Nashville .............................................. 19

10 Hypothetical targets. metropolitan Nashville ......................... 21

11 Cumulative fuel mass for three targets .............................. 22

12 Fuel load density as function of radius for three targets .............. 23

13 Simulated smoke plume after 2 h for two target zones ................ 25

vi

SECTION 1

INTRODUCTION

Cities contain large amounts of combus- representative of all U.S. citi.,. and is of-tible material- The wood. plastic. paper. ten applied to European and Soviet citiescloth, and asphalt used in the structure as well.and contents of most buildings, togetherwith the oil and gasoline distributed But cities are not all alike. Even citieswithru u the urbandgasoncomplex disred which appear similar according to thethroughout the urban complex. repre-

sent a concentrated fuel source which. if convntional measures of rank (popula-

ignited. could inject enormous quantities tion. population density. built-up area

of smoke into the atmosphere. Nuclear fraction) may be quite different from the

weapons provide an efficient incendiary standpoint of fuel loading. This

trigger for city fires. It is estimated that a city-to-city variation is linked to differ-full-scale nuclear attack against the U.S. ences in urban land use-the relativefraction of a citv devoted to residential.would yield approximately 40 Tg of commercial, and industrial purposes.smoke, primarily from urban targets Each category is characterized by a par-[Small. Bush. and Dore. 19891. Yet many ticular building .ype and density-. Theof the specifics about this urban smoke amount of flammabie material ,ithin(e.g.. the net particulate mass. its optical structures is also dependent on land

properties. and the altitudes to which it uses is alependentrialand

is injected) are highly uncertain. These houses contain more fuel per unit area

uncertainties result from arbitrary as- than most commercial establishments.sumptions regarding tl.2 amount. type. and both are more heavily loaded thanand distribution of combustible fuel residential structures.) To some extent.loads in urban areas. the fraction of an urban area occupied

by each land use category is a function

Most nuclear winter studies have char- of geographical region-at least withinthe U.S. [Bush and Small. 1987]. Butacterized the fuel loading in cities by a considering the diversity of city types in

single overall average value. These val- the U.S. (and the even -geater differ-ues are derived either from statistics of grences betwveen U.S.. European. and So-the production, consumption. and accu- nebtwnU..Euoa.adS-viet cities). it is unlikely that a singlemulation of flammable materials [Crut- generic model of urban structure couldzen, Galbally. and Briihl 1984: Bing,. accurately represent all cities. Likewise,1985] or from limited surveys of the fuel it is doubtful that a single "real- cityloads in various building types in U.S. could serve as a good model for all oth-cities [Federal Emergency Management ers. Yet the correlation between fuelAgency. 1982: Turco et al.. 1983: Nation- loading and land use suggests a formalal Research Council. 19851. In the latter method for estimating fuel loads in realapproach. each building type is assigned cities.to one of several functional categories(citv center, suburban residential. etc.). Another characteristic of cities is theirA highly idealized model of the geograph- highly uneven internal distribution ofical distribution of these urban subdivi- land use types (and. therefore. of fuelsions (such as the zonal or -concentric loads) which results from their uniquering" model) is then used to average the topography. historical r_. .-,l,_pment. andsurvey data over the entire city. The re- socioeconomic growth. This lack of uni-sultant weightea mean is assumed to be formity is important in evaluating the

I

fuel loads associated with a given target- bined with the cartographic land useing scenario. If. for example, the ignition data. The result is a composite fuel loadzone around a particular urban target map for metropolitan Nashville. Sincecontains mostly industrial land or a our approach relies on readily availableheavily developed central business dis- land use and demographic data, it istrict (CBD). the fuel loading in the area more generally applicable than thoseburned could be far greater than the used in previous fuel load estimates forareal average value for that city. More individual cities (e.g., Bracciaventi, Feld-smoke would be produced and, because man, and Newman, 1968, for Detroit; Si-the fires would likely be very intense, the monette et al., 1986, for San Jose).plume would reach higher into the atmo- These calculations were based on sur-sphere and contain a greater elemental veys of at structures within the areascarbon (soot) fraction. On the other studied. This approach. though detailed,hand, another target zone in the same cannot be easily extended to other cities.city might encompass mostly low density On the other hand, oar approach can besuburbs, park land, or even water. The readily applied to any U.S. metropolitansmoke injection there would be relatively area. We present our Nashville calcula-minimal. Again, land use is the primary tion as an example.

determinant.

In this report we present a calculation of In Sec.2, we describe the physical geog-the fuel load for Nashville. Tennessee, raphy and land use characteristics of theand the surrounding area. Our basic metropolitan Nashville area. In Sec. 3.data set consists of fuel load estimates we discuss our methodology for comput-for 11 land vse categories (five urban ing fuel loads from land use data andand six nonurban). These estimates are present a fuel load map for Nashville. Inderived from extensive surveys of the Sec. 4. we choose several potential tar-structure, contents, and densities of gets. demonstrate the sensitivity of fuelbuildings across the United States. Cen- loading to target position, and discusssus data are used to weight residential the implications for smoke emission andbuilding densities within Nashville. plume injection altitudes. ConclusionsThese fuel load estimates are then com- are presented in Sec. 5-

2

SECTION 2

NASHVILLE: GEOGRAPHY AND LAND USE

2.1 GENERAL DESCRIPTION. dissected hills frames the city. Most ofNashville lies on the south bank of the

Nashville is located on the Cumberland Cumberland River. which makes a seriesRiver in north-central Tennessee. It is of wide meandering loops as it flowsthe state capital and the center of a large through the region. About 40 km up-metropolitan area that encompasses stream from the city center. Old Hickorymost of Davidson County. as well as Dam blocks the river and forms a largeparts of neighboring Williamson. Wilson. slack-water lake. Another. even largerSumner. Cheatham. and Rutherford man-made lake. the J. Percy Priest Res-counties (Fig. 1)- It is the 26th largest ervoir. lies on the Stones River southeastU.S. citv. with an estimated population of the city. The dam and two lakes(as of julv 1986) of 473.670 [U.S. Bu- (created in the 1930s by the Tennessecreau of the Census. 1988l. Perhaps best Valley Authority) supply metropolitanknown as the center of the country mu- Nashville with most of its electrical pow-sic industry (it is the home of the Grand er and are the focus of much of its recre-Ole Opry). Nashville also serves as the ational activities.headquarters for several large finance.insurance, and publishing companies.Leading industries include printing. re- Davidson County and the cit" of Nash-cord production. and the manufacture of vfile form a single political entity. Theclothing. shoes. glass. heating and cook- governing unit. known as the Nashvilleing equipment. and tires. In addition. Metropolitan Government. has beentwo automobile plants are located widely regarded as a major innovation innearby: the Nissan assembly plant in local government since its inception inSmyrna. and ilhe General Motors Saturn 1963 Its county-wide jurisdiction hasplant in Spring Hill (about 50 km south made it possible to extend urban ser-of downtown). The area is served by vices such as sewage disposal far beyondthree major highways. Interstates 24. the original city limits, and has contrib-40. and 65. which intersect near down- uted to the rapid growth of outlying sub-town and ioop around the CBD: several urbs. Yet. as shown by Fig. 2. aboutrailroad lines (it is a major regional rail- two-thirds of Davidson County is still ru-road center): and Nashville Metropolitan ral. In addition. the county encompassesAirport. an American Airlines hub. Nash- several communities (e.g.. Goodlettsville.vi'.:e is also the location of sb-iteen col- Berry Hill) which. although under the ju-leges .and universities, the most notable risdiction of the metropolitan govern-of which include Vanderbilt and Fisk ment and usually lumped together withUniversities. Nashville for census purposes. are ac-

tuallv distinct cities. In this report weNashville is situated in the northwest therefore distinguish between Nashvillecorner of a large lowland region known proper (an arbitrary division which in-as the Nashville Basin. The terrain con- cludes most of the central built-up area)sists of gently rolf'ng land punctuated by and Davidson Countv (the wider polidcalsmall rounded hills. or knobs. To the unit). The outline of Nashville proper iswest and north, a chain of low. highly shown in Fig. I.

3

2i

04

00

0 0

0 %n

0 E0

00

010

0

01.0

z

4

Wholesaleand retail

Transportation, tradecommunications,utilities Manufacturing

Consumer

services -

Communityservices Water (excluding

u J. Percy Priest

productionReservior)and extraction

-- 2Single-familyhousing

Undeveloped.•,

Muitpia-family

housing

Scums: Met,-poft Planng Conso f1983]_

J Figcure 2. Land-use su,,-ary, ,Nashville-Davidson County.

5

2.2 LAND USE. time [Marshall, 19751: (1) linear or secto-ral growth along the main thoroughfares

Residential land occupies 49 percent of extending southwest and northeast fromthe urban area within Davidson County downtown, and (2) discontinuous growth(Fig.2). (By comparison, in U.S. cities associated with the establishment of thewith populations greater than 100.000, first outlying -job-oriented" suburbsthe average is 41 percent [Northam. (e.g.. West Nashville. Old Hickory. and1979].) Figure 3. produced from United the area around the Radnor TrainStates Geological Survey (USGS) digital Yards)- Thus, downtown Nashville wasland use data. shows the distribution of gradually abandoned as a residentialland areas devoted primarily to residen- center, thereby setting the stage for thetial purposes. Most of the residential development of the CBD and the pres-land is concentrated within Nashville ent-dav suburban sprawl.proper in two zones located southwestand northeast of downtown. Other con- Today, residential Nashvlle continues tocentrations are found in the Goodletts- Today, rapid rate.lAecording toville area, along the northern shores of extand at a rapid rate. According to sta-Old Hickory Lake (Hendersonv-ille). and tistics compiled by the MetropolitanOintecitieskof Lak VerndersandilSe)yaad Planning Commission (MPC). this growthin the cities of La Vergne and Smytrna. is occurring primarily in outlying areasSmaller rural centers are found through- along the interstate highways [MPC.out the remainder of the metropolitan 19861. Interstate 40 is a particular focusarea. of this expansion: most residential build-The distribution of residential land in ing permits issued in recent years haveTashvile dstribuiong reflectsid hentiabeen for districts along thih highway.Nashville strongly" reflects the historical from Bellevue on the west to Hermitage

development of its transportation net- f Bl on the et se Fg H)rTisagmwork Inthi sese t i tyica ofanyH~ills on the cast (see Fig. 1). This corn-

work. In this sense it is typical of any muter-oriented growth pattern is typicallarge U.S. city of similar age [Bush and of suburban expansion in modern U.S.Small. 19871. Founded in 1779 as a mili- cities [Northam. 19791.tary outpost (Fort Nashborough). Nash-ville quickly became established as amajor riverport. For nearly a century. Another indication of the suburbanriver traffic dominated the development character of residential Nashville is theof the young city, and. as a result, popu- dominance of single-family detachedlation. commerce, and industry grew (SFD) homes over multiple-family (MF)mainly north and south along a line par- housing structures. Only 15 percent ofalleling the Cumberland. Residential and all the residential structures in Davidsoncommercial zones were not well differen- County are MF. and these account fortiated during these early years, with re- less than 10 percent of the total residen-tail businesses scattered among resi- tial land area (see Fig. 2 and MPC.dences within a few city blocks of the 1987). (The average for all U.S. citiesriver [Marshall. 19751. This pattern con- with populations greater than 250.000 istinued even into the 1850s with the 17 percent [Northam. 1979].) The arealcoming of the railroads and the develop- density of SFD homes by census tract isment of a more diversified economy. In- shown in Fig. 4. Note that the highestdeed. significant expansion of residential densities are found in the older residen-Nashville did not begin until the 1890s. tial zones immediately adjacent to thewhen the streetcar (and, later, the auto- city center and in West Nashville. whilemobile) made quick transportation to the lowest densities are found in theand from the crowded central city possi- newer, more affluent outlying suburbs.bie. Two growth patterns emerged at this The distribution of MF structures (Fig. 5)

6

S-

-S -3

U-;

7L

cn

4-3-

an

T_ C~ ? 3

C) C) C31

Nn V. M C

EM,

* /n

4-3

C-,

+)

I IA

I 4-1 3n

I 4T

MV C'I3 AS

11 El 1

shows a similar pattern, although the town-a pattern typical of most U.S. ur-densities are lower. The distribution of ban areas [Bush et al.. 19881. Figure 7structure ty-pes is important, since it de- also shows a large downtown commercialtermines the distribution of fuel loads in zone comprising the heavily developedresidential zones. CBD and Vanderbilt University. There

are also commercial zones around theIndustrial zones and commercial/service airport, and a large shopping center neardistricts also display definite distribution Smyrna.patterns. The most prominent industrialareas are clustered along the Cumber- Agricultural land comprises one of theland River. mainly close to downtown largest land use types in the Nashville(Fig. 6). Several large industrial parks area. Figure 8 shows the distribution ofare also present; the largest are near cultivated land of all types in the region.Smyrna and in Rayon City near Old Most of the agriculture is confined toHickory Dam. On the other hand. the somewhat small plots on the relativelydistribution of commercial/service areas flat land east of Nashville proper. West of(which include shopping centers. corn- the city. most of the nonurban land ismercial strips, office buildings. schools, undeveloped wildland. The only signifi-and hospitals) is often more linear. In cant agricultural areas here are concen-Nashville (Fig. 7). there are prominent trated along the Cumberland apd withincommercial/service strips along major the many stream-cut "hollows- which in-surface streets extending from down- tersect the river valley.

10

c,

4-3

0

4-3

S.-

L.-

411

400

001ýC

12w

4-3

00is-

co

S.-

13

SECTION 3

FUEL LOAD CALCULATION.

3.1 NONRESIDENTIAL LAND. Such areas have low building densitiesand thus relatively low fuel loads.

We assign all land within the metropoli-tan Nashville area to one of eleven land Vegetation within cities also contributesuse categories. These categories, based to the combustible fuel load. Followingon the definitions given by Anderson et Bush et al. [19881. a nominal burnableal. 119761. are listed in Table 1. Also fuel load per vegetated urban area of 5shown are the assumed fuel loads per kg/iM2 . regardless of land use category isunit occupied area F. the areal percent- assumed. Such loading is at least twiceages occupied by buildings and vegeta- as large as that expected for nonurbantion A. and the area-weighted net fuel vegetation, but is justified by the greaterload densities FD for each category. degree of plant desiccation that can oc-These values are derived from several cur when vegetation is in close pro.mitysources. For the two main nonresidential to intense, long-burning structural fires.urban categories (commercial/services Blast effects, such as the breaking andand industrial), we use the structural scattering of tree limbs and shrubs, arefuel load and building density estimates also likely to enhance the amount ofderived by Anno et al. 119881. These esti- burnable vegetation in urban areas.mates are based on analyses of thestructure and contents of 33 nonresi- The Nashville area also includes largedential building types (representing over portions of nonurban land. We use six6000 individual structures) within the nonurban land use categories (Table 1).United States. Aerial photographs and Fuel loads have been assigned to thosetopographic maps of five representative categories based on estimates developedU.S.urban areas were used to determine by Bush and Small (19851. For the Nash-building densities. The results represent ville region. we assume an agriculturalaverages for the entire United States. Al- fuel loading appropriate to the predomi-though intercitv differences no doubt ex- nant crops grown in the area--corn. soy-ist. we find that. in general. the variation beans, and tobacco. Fuel loads for thesein nonresidential land use characteris- crops are seasonal. The crop loading ap-tics is small within the U.S. (see Bush plies to midsummer when the crops areand Small. 1987): therefore. these aver- near maturity: in -inter the loading isages are not further weighted in our cal- minimal. The fuel load estimates forculations. rangelands and forests also represent

summertime conditions. For forests. weThe values for the two remaining nonres- choose a value applicable to the decidu-idential urban categories (transporta- ous woodland endemic to the Nashvilletion/communication/utilities and other region (comprised mainly of oak. ash.urban) are from Bush et al. [19881. The and maple trees). The assumed fuel loaddesignation -other urban- refers to devel- for forests is less than one-third the val-oped portions of the city that include ue chosen for agricultural land. This dif-large amounts of open space. Examples ference results from the higher standinclude urban parks. golf courses. ceme- densities and more abundant water sup-teries. and zoos [Anderson et al.. 19761. ply qypical of cultivated land iBush and

14

Table i. Fuel load by land use type in Nashville.

Fuel Load per Percent Area Net FuelArea Occupied Occupied Load Density

Land Use Type F(kg/m ) A FD (kg/m )

Urban

Residential 130 (structures) 7.5 (structures)5 (vegetation) 40.0 (vegetation) 11.7

Commercial, services, 110 (structures) 21.6 (structures)and mixed urban 5 (vegetation) 10.0 (vegetation) 24.3

Industrial and mixed 225 (structures) 19.6 (structures)5 (vegetation) 5.0 (vegetation) 44.4

Transportation, communication, 60 (structures) 5.0 (structures)and utilities 5 (vegetation) 15.0 (vegetation) 3.8

Other urban 60 (structures) 5.0 (structures)5 (vegetation) 50.0 (vegetation) 5.5

Nonurban

Agriculture 2 . 0 a 70.0 1.4

Rangeland 0.2 100.0 0.2

Forest 0 . 6 b 100.0 0.6

Water 0.0 100.0 0.0

Wetlands 0.0 100.0 0.0

Barren 0.0 100.0 0.0

aMidsummer value for grain croplands [Bush and Small, 1985].

bSummer estimate for deciduous forests [Bush and Small, 1985].

15

Small. 19851. Wintertime fuel loads in where FSFD and F.%,,: are the mean fueldeciduous forests would be highf r by loads per foundation area for SFD andabout 30 percent. This follows from MF structures, respectively, andsmaller live loadings in the forest canopyand greater concentrations of leaves and NSFD * aSFDtinder on the ground. ASFD= (3)

ares

Given the fuel load per foundation areafor nonresidential buildings (Fstructures)-the burnable vegetation load per unit andarea (F..egetation). and the mean fractionalareas occupied by each (As:;ctures andA•egetationi. the net nonresidential fuel ..MF - aMFload densitv FD is AMF -- (4)

- ~ares

FD = Fstructur es AstructuresFD F .tr s * Asgtactu. (1) are the corresponding fractional arealF1egeta-ion * Avegetation, (1) building densities. Here Ns5 -D and NXIF

denote the numbers of SFD and MF

where. Fstpuctures = 0 is assumed for all buildings within the tract. ars, and aMF

nonurban categories. The resultant val- are their mean foundation areas andues for FD are listed in Table 1. ares is the total residential area withinthe tract (including structures, streets.

yards, etc.). Values for NsF, NMF. and3.2 RESIDENTIAL FUELS. ares are determined uniquely for each

census tract, while asm. , FSF. andResidential zones comprise the largest F.,IF (Table 2) are assigned values consis-fraction of developed land in metropoli- tent wit.h analyses of the structure, con-tan Nashville. Housing types and build- tents, and size of resiriences throughoutmng densities vary widely across the the southern U.S. and are assumed notregion and. as a result, so do residential to vary from tract to tract. Some intracitvfuel loads. We account for these varia- variaticn in these values no doubt exists.tions by weighing the densities of both due primarily to differences in land val-SFD and MF buildings within individual ues, socioeconomic status, and histori-census tracts. For tracts inside Davidson cal growth patterns. However, by ac-County. we use the housing counts de- counting for residential structure densi-veloped by the Nashville-Davidson ties (NsFD/aes and NF/ares) on a tractCounty Metropolitan Planning Commis- by tract basis, the most important varia-sion [MPC. 19831; outside Davidson tions in residential fuel loads are ac-County. residential building densities counted for. Our estimate for a,• (348are estimated from U.S. Census Bureau In 2 ). is also assumed invariant withindata [U.S. Bureau of the Census, 19831. Nashville: it is derived from a statisticalThe area-weighted residential fuel load analysis of the number of floors andper foundation area for a particular cen- units contained in MF structures builtsus tract is calculated between 1980 and 1984 [U.S. Bureau of

the Census, 19851. This value compares

FSFD * ASFD + FMF*AMF well with the measured mean rooftopFstme = + , (2) area (equivalent to foundation area) of

ASiD + AMP MF buildings (305 m 2 ) found in a

three-city survey [Fretz. n.d.I

Table 2. Residential structure characteristics.

Fuel Load per Mean FoundationFoundation Area Area

Structure Type F (kg/m 2 ) a (m2 )

Single-familydetached (SFD) homes 1 1 5 a 1 5 5 b

Multiple-family(MF) buildings 16 7 a 3 4 8 C

aMean value for southern U.S.. from Anno et al. [19881.bFrom Ransohoff et al. [19871, assuming 68.6 percent of SFD

homes in southern U.S. are one story.CU.S. average based on analyses of construction reports (see text).

Using the relation A,,., Asm + AM in densities in U.S. cities are focund in theEq. (2) and applying the result to Eq. (1). northeast and the west [Anno et al..FDres is computed for the residential 19881).land in each census tract. We find thatthe distribution of FDres is highly vari- A simih."L weighting of fuel loads forable. with values ranging from less than non-residential urban land could also be3 kg/m 2 in some outlying areas to over performed. The densities and types of90 kg/im 2 in the heavily developed MF structures in each commercial and in-housing district on the edge of downtown dustrial zone could be determined fromnear Vanderbilt University. For Davidson aerial photographs. maps, and economicCounty as a whole. FDres = 11.7 kg/m 2. census data. Breakdowns of fuel loadsThis is about 20 percent less than the for various building types could then beaverage residential loading for all U.S. ci- used to derive detailed fuel density esti-ties (15.0 kg/im2 including vegetation) mates for these areas.derived by Anno et al. 119881 and about20 percent greater than the average for 3.3 FUEL LOADING MAP.the southern U.S. (9.7 kg/m 2). These dif-ferences are due largely to regional dif- The fuel loading for the 60 km by 38 kmferences in the density of residential Nashville area shown in Fig. I is devel-structures. In particular. we find that oped on a 1 km by I km grid. (The reso-Ares = 7.5 percent for Davidson County. lution is arbitrary and limited only bywhile Ares = 9.7 percent for all U.S. cities the resolution of the land use maps-i.e..and Ars = 6-3 percent for the southern 0.04 km 2 for all urban categories and bo-United States. Thus. residential Nash- dies of water and 0.16 km2 for all otherville is somewhat more densely built land use types: a much finer grid couldthan average for a southern city. but is be used). The areal fraction f, of eachstill less dense than the average U.S. grid square occupied by each of the 11city. (The highest esidential building land use categories is determined, and

17

the net fuel load density FDnet for the usually assumed for cities [Penner.square calculated from 19861. Moreover. the distribution of fuel

11 is clearly nonuniform with some (target)areas having substantially greater or

I (FD)i (5) lesser than average loadings.

where (FD)i is taken from Table 1 for all Major bulk liquid fuel storage sites inNashville are concentrated in the indus-nonresidential land use categories and trazoealntesuhbnkfte

FD,,is cmpued a desribd abve. trial zones along the south bank of theFDres is computed as described above. Cmein erXetNsvle te

The fuel load map is shown in Fig. 9.sites are located along the river immedi-

Although the highest fuel loads are lo- atelv north and south of downtown and

cated near the downtown area. the dis- at the airport- The most vulnerable of

tribution around the central city is these sites are those in which the prod-

highly asymmetric. Generally. higher ucts are stored in above-ground tanksfuel loads follow the distribution of resi- easily identifiable in aerial photographs

dential land-i.e.. along a southwest- and topographic maps. (In Nashville.

to-northeast axis through the city center- these tanks contain only refined prod-

The largest ioadings (> 40 kg, m 2) are as- ucts: there are no refineries in the area.)

sociated with industrial zones and hea-v There are 87 above-ground tanks rang-MF residential concentrations. Outside ing in size from about 18 to 36m diame-

the built-up area. the highest fuel loads ter with an average capacity of 4550 m 3 .

are found on agricultural plots (for ex- Assuming each tank is half full. the esti-ample. along the Cumberland River west mated total fuel inventor: is 197.900 in 3 .

of the city). These results demonstrate For a density of 740 kg/m 3 (typical of

the highly uneven distribution of com- gasoline), the estimated stored fuel massbustible fuels in and around urban is 0 146 Tg. If distributed uniformlyareas. throughout Nashville proper (232 k•n 2 ).

the loading is increased by 0.63 kg/mi2 :

We find that the mean fuel load for the if distributed over Davidson Countyentire map shown in Fig. 9 is 5.0 kgcim-: (1187 km2 exclusive of J. Percy Priestfor Davidson County. it is 6.9 kg,'m-: Reservoir). the additional loading wouldand for Nashville proper. 18.9 kg/r 2 . be 0. 12 kg/m 3 . Fuel stored in vehicles'Such values are considerably smaller and pipelines enhance these values onlythan the approximately 40 kg/mr2 mean slightly.

L ,'Ve estimate a net gasoline invcntorv of 0.005 Tg stored in -chicles in metropolitan Nashville as-sumning 0.67 vehicles per person IU.S. Bureau of the Census. 19861. an average tank capacity of 12gallons. and each tank half full.

18

~~CD

c0

cc c

CD,

00

Co

01C.

.010

311

SECTION 4

IMPLICATIONS FOR SMOKE INJECTION

To demonstrate how the nonuniform dis- areas at greater radii. This is quite un-tribution of fuel loads in a city influences like the idealized uniform fuel load dis-the smoke source, we consider area fires tribution assumed in most smoke plumecollateral to three hypothetical target and nuclear winter calculations. The ra-sites: (1) the Tennessee State capitol (a dial dependence in FDnet implies that thepolitical target); (2) the Tennessee Air fire intensity or duration may not beNational Guard station (a military site); constant throughout the target area.and (3) the power generating station lo- Since the smoke injection is close!ycated at the J. Percy Priest Dam (an eco- coupled to the burning dynamics, mark-nomic target). The locations are shown edly different smoke injections can re-in Fig. 10. Circles of 10-kmn radius, sult even in the same city. Target zoneswhich define the optimum 3-psi over- 2 and 3. on the other hand, have a morepressure ignition zone for a 1-MT air- uniform fuel load distribution whichburst, are drawn around each target slowly increases with radius as moreThere is considerable overlap; the maxi- high-density, built-up land is involved.mum distance between any two targetsis only 16 km. Yet, despite their proximi- The fuel load densities (Fig. 12) havety. the amount of fuel contained within been used to determine the fire plumeeach target zone is quite different. Figure and smoke injections for each target11 shows the variation in net fuel area. The fire intensity is assumed pro-amount as a function of radius. Target portional to FDet, the total burning timezone 1 encompasses the most fuel (mean is 2 h. and the smoke emission factordensity 18.5 kg/m 2-about equal to the (smoke mass generated per unit fuelaverage for Nashville proper) since it lies mass burned) is 0.03. For target area 1.near the geographical center of the ur- both the smoke generation rate and theban area. Target 2. located at the airport heat release rate are greatest between 1-near the edge of the built-up area, con- and 4-km. For targets 2 and 3. however.tains the next highest loaung (8.2 kg/ the fuel densities are greatest 10 kmIn2 ). and target 3. near the edge of the from the burst and thus the intensitycity. contains the least (4.4 kg/mr2). and smoke production are largest at theThus, a nuclear burst over target 1 could fire edge. (We assume a 10-km fire ra-burn more than four times the fuel of an dius for a 1-Mt burst) A mean Julyidentical burst over target 3. sounding for Nashville was used. The at-

mosphere was warm and moist at lowerlevels, and the troposphere conditionally

These differences in f11l availability af- unstable.fect not only the amount of smoke pro-duced, but also the fire behavior. In Fig. Figure 13 shows the simulated smoke12 we show the variation in fuel load plumes associated with targets 1 and 3density FDe. as a function of radius for after 2 h. (The complete set of results isthe three targets. Note the strong radial presented in Heikes et al. 1989.) Clear-dependence. especially for target zone 1. ly. the difference in the amount of fuelwhich encompasses mostly heavily contained in these target zones has aloaded urban land within about 4-km ra- dramatic impact on the plume. With thedius and more lightly loaded suburban relatively low and uniform fuel loads

20

-- -,

CD cn

CiC

ej

14-,

CO,

V15 101Z

-4

21

5

Target 1

-4i

0

co 3

-0

E

(/22

0E 2 Target 20

w

00- 1

6 7 8 9 1

Ignition radifus (km)

Fgure 11. Cumiulative fuel mass for three tarqets.

22

30

S~Target 1

N 20

C12

E00Ca

Cm 0 Target 2

CO 0

C

Target 3

0 1 2 3 4 5 6 7 8 9 10

Radius (kmn)

Figure 12. Fuel load density as function of radius for three tarqets.

• 23

associated with target 3. the plume re- availability in the target 1 burning zone.mains confined to the lower troposphere but also to its heavy concentration nearand spreads radially at about a 3-km al- the center. This increases the heat re-titude (Fig. 13a) despite the conditionally lease rate near the middle of the fire andunstable sounding. The target I loading, provides the plume with enough buoyan-however, produces a plume that devel- cy (and therefore momentum) to pene-ops explosively in the conditionally un- trate to very high altitudes. Clearly. thestable atmosphere. penetrates the behavior of the smoke plume is highlytropopause at 15 km. and extends far sensitive to the target locations in andinto the stratosphere (Fig. 13b). The dif- around the city.ference is due not only to the greater fuel

24

a. Target zone 3.

20-

18-

16-

14-

";12-

-10-

6-

2 • • C.. ... ...

b. Target zone 1.

24-22 P20'.: .^18-

16

......... ...

10

8 .6 .

2

0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40Radial distance (kin)

Fioure 13. Simulated smoke plume after 2 h for two target zones.

25

SECTION 5

CONCLUSIONS

We have estimated the combustible fuel and in industrial/commercial complexes.loads for metropolitan Nashville. Tennes- The nonuniformity of the fuel distribu-see using land use and census data. tion results in a several-fold difference inNashville's fuel loads are not uniformly the net fuel availability for targetsdistributed and are highly dependent on spaced only a few kilometers apart. Thisthe internal structure of the city. The in turn strongly influences the intensityhighest fuel load densities are found in of nuclear weapon fires. the amount ofportions of the city occupied by densely smoke emitted. and the altitudes topacked multiple-family housing struc- which the smoke is injected.tures. hydrocarbon storage terminals.

26

SECTION 6

LIST OF REFERENCES

Anderson. J.R.. et al.. A Land Use and Federal Emergency Management Agency.Land Cover Classification System for Use -What the Planner Needs to Know Aboutwith Remote Sensor Data. U.S.Govern- Fire Ignition and Spread.- Attack Envi-ment Printing Office. Washington. DC. ronmental Manual. Chap. 3. FederalUSGS Professional Paper 964. 1976. Emergency Management Agency. CPG

2- IA3. 1982.

Anno. G-H.. et aL.. Nuclear Winter - Fretz. R.K.. Metropolitan Rooftop AnalysisSource-Term Studies. Vol. 4. Fuei Loads in for Three Cities: Los Angeles. Hartford.U.S. Cities. Pacific-Sierra Research Cor- and Miami. Jet Propulsion Laboratory.poration. Report 1628. 1988. Pasadena. California.

Bing. G.. Estimates of Total Combustible Heikes. K.E.. Gaj. R. A.. and R. D. Small.Material in NATO and Warsaw Pact Coun- Target Area Studies. Vol. 2. Plumes fromtries. Lawrence Livermore National Labo- Urban Target-Area Fires. Pacific-Sierraratorv. Livermore. California. UCRL Research Corporation. Report 1909.93192. August 1985. 1989.

Bracciaventi. J. S Feldman. and R. Marshall. J. M.. -Residential ExDansionand Central City Change.- in J. F.

Newman. Fuel Array Survey of Detroit Blumstein and B. Walter (eds.). GrowingMichigan. Naval Applied Science Labora- Metropolis. Vanderbuilt University Press.toi;. Brookl-n. New York. NASL-940- Nashille.. Tennessee. 1975. pp. 33-63.90-PR-3. 1968.

Metropolitan Planning Commission.Bush. B.W.. and R.D. Small. Smoke Pro- 1982-83 Land Use and Housing Survey,duced by Nonurban Target-Area Fires fol- Unpublished Study. Nashville-Davidsonlowing a Nuclear Exchange. Pacific- County Metropolitan Planning Commis-Sierra Research Corporation. Report sion. Nashville. Tennessee. 1983.1515. 1985. •_ . Housing Statistics. 1986. Unpub-

"lished Study. Nashville-Davidson Count"• "Nuclear Winter Source-Term Stu- Metropolitan Planning Commission.

dies. Vol. 2. The Classification of U.S. Nashville. Tennessee. 1987.Cties." Pacific-Sierra Research Corpora-

tion. Report 1628. 1987. National Research Council. The Effectson the Atmosphere of a Major Nuclear Ex-

Bush. B.W.. et al.. ,Nuclear Winter change. National Academy Press. Nation-Source-Term Studies. Vol. 6. Smoke Pro- al Academy of Sciences. Washington.duced by a Nuclear Attack on the United DC. 1985.States. Pacific-Sierra Research Corpora- Northam. R.M.. Urban Geography. 2dtion. Report 1628. 1988. ed.. John Wiley and Sons. Inc.. 1979.

Crutzen. P. J.. I. E. Galbally. and C. Penner. J.E.. -Uncertainties in theBrfahl. "Atmospheric Effects from Smoke Source Term for "Nuclear Winter"Post-Nuclear Fires.- Climatic Change. No. Studies.- Nature. Vol. 324. 1986. pp.6. 1984. pp. 323-364. 222-226.

27

Ransohoff. L. M.. et al.. Topics in Nuclear U.S. Bureau of the Census. 1980 CensusWinter Source-Term Research. Vol. 1. of Population and Housing, Nashville-Da-Composition of Residential Structures in vidson. Tenn._ U.S. Government Printingthe United States. Pacific-Sierra Re- Office, Washington, DC, PHC80-2-252,search Corporation. Report 1761. 1 No- August 1983.vember 1987.

Simonette. D. S.. et al.. Magnitude and , Characteristics of New Housing:Spatial Distribution of Urban Flammable 1984. in cooperation with the U.S. De-Materials in the San Jose Area. Calffor- partment of Housing and Urban Devel-nia.- Department of Geography. Univer- opment. U.S. Government Printingsitv of California. Santa Barbara. 1986. Office. Washington. DC. Construction

Report C25-84-13. June 1985.Small. R. D.. B. W. Bush. and M. A.Dore, -Initial Smoke Distribution for Nu- . Statistical Abstract of the Unitedclear Winter Calculations.- Aerosol Sci- States: 1987, 107th ed.. U.S. Govern-ence and Technology. Vol. 10. No. 1. ment Print Office. Washington. DC. De-1989. cember 1986.

Turco. R. P.. et al.. Nuclear Winter:Global Consequences of Multiple Nuclear . unpublished data. Los Angeles Re-Explosions.- Science. Vol. 222. 1983. pp. gional Office, Van Nuys. California, Sep-1283-1292. tember 1988.

28