ML20235D960

Applicant Exhibit A-14,consisting of Paper Entitled, Modelling Evacuation Behavior During TMI Reactor Crisis. Copyrighted Matl

ML20235D960
Person / Time
Site:	Shoreham File:Long Island Lighting Company icon.png
Issue date:	07/08/1987
From:	Jerrica Johnson, Zeigler D CALIFORNIA, UNIV. OF, LOS ANGELES, CA, OLD DOMINION UNIV., NORFOLK, VA
To:
References
OL-3-A-014, OL-3-A-14, NUDOCS 8709250422
Download: ML20235D960 (9)

Text

_. . .

u,, 3 s>m twweer ,e teese Santaa, w.

w ,-- # mme m..

roern, of go- 3 2 2 -

x 3 7 g w4%ss,wi meg w wres hav -

' 'e

• Il[ $ WM~

mmumr a"hian i MODELLING EVACUATION BEHAVIOR DURING TIE ~

alule. Seu THREE MILE ISLAND REACTOR CRISIS-

'87 SEP 17 All :49 a a more J. H. JOHNSON. JR A over time Departmeni of Geography, Universny of Cahforma. Los Angeles, les Angel u re.

abihty n and h-

,umphon nomic ac, D. J. ZElGLER 0* ' Department of Pohucal hetence and Geography. Old Dominion Unwersil), Norfolk , ...

consumtw be weak- Werened tb Novemier 1%$)

depec of enzed the AbstracF-This naper reports on the results ofan empincal test of a causd ny model of evacuation ocemon-makmr which we have developed to asus) emergency managernent o enditures for the Nuc6 car Regulatory Commeon m the ahermath of the see: un s current Gent suppon all of the hmothemed

.itbacks in behavior is predictable on the baus of locauonal umuonhie. cycle cepuon for reactor silmg pohcies are docussed.vanables. !mpheauons for.and future-onemed ce sursest m defense 1. INTRODUCTION hold up The near catastrophic accident at the Three Mile Island a statistical standpoint the rehabihty of the mod ent). On (TM1) nuclear generating facihty near Harrisburg, a predictive tool will remain unknown unless an ac.

cks in de- l Pennsylvania, on 28 March 1979 evoked an unantic. cadent occurs at the Shoreham plant and actual be-inated and unpreccocmed level of spontaneous evac. havior is compared with the estimates denved from 3

irr than in ustion. Resulting in part from equipment failure and the model. We believe that the empirical test repo panly from human error, the accident prompted the on here, which is based on the actual behavior of governor of Pennsylvania to advise pregnant women southcentral Pennsylvania area residents dunng th '

hudreme and preschool children within 5 miles of the plant to TM1 reactor ensis, provides additional and st 81 evaruate, and everyone else within 10 miles to stay evidence of the explanatory power of our model. j l

!"'" 8" I indoors or sheher. This protective action advisory Moreover, the results suggest several chan j

'""""* should have prec.pitated the evacuation of only about need to be made m the policies govermnt both the

! g {

' M00 people [1]. Post accident surveys revealed, how-plants.siting and operation of commercial nucicar power g g,g, ,-

- ever, that as many as 200.000 people within a 25 mile J Oxford radius of the malfunctioning TMI reactor actually evacuated [2-5). While spontaneous evacuation has 2. THE MODEL L M ash.

E been known to occur in other types of emergencies. Our proposed model of nuclear reactor emergenc.

)>

O -j neither the magnitude nor the geographic extent ofits evacuauon occision making is depicted in Fig.1. In mer occurrence at TM1 could have been predicted on the oevelopm; this mode!. we rehed mainiv on works per.

der < 1,;)

O 5.y basis of what was known about human behavior taimnginto pubhc perceptions. behefs. a'nd attitudes te other ensis situations. ward the use of nuclear power to generate ciectncity y

g, y,.

Og The purpose of this paper is to repon on the results te.g.11G-16)). and on desenptive and expioratorw

,TI

<;- of an empirical test of a model of nuclear reactor quantitative studies of actual evacuation behavior emergency evacuanon decision-making wruch we have during the TM! reactor ensis p, IL21). As Fig. I iexpen-R 37- >7$ oeveloped to enhance our knowledge and understand- shows. the proposed modelis comprised of three eu i l$k

,_ ing of public behavior during the TMi reactor crisis. ogenous cateponcal vanables Oocation. stage-in-life -

!  % e have conducted one previous empincal test of the cycle, and social status) and three endogenous variabies  !

Q 'y proposed model, using loglinear causal modelhng and tattitude to nuclear power. nuclear reactor accident 4

'(Z O-

<O

~' survey data obtained in imerviews with a representative sample of households within the vicinity of the Shore-nsk perception, and the evacuation decision, that is.

to evacuate or not to evacuate). The structure of the ham Nuclear Power Station in New York State, and proposed model revolves around f ve maior research

~" the results strongly suppon the hypotheses developed hypotheses. which have been discussed iri detail cise-C in this paper [5). However, that empirical test was based where 15). For the purpose of this stud h they can be on mrended rather than actual evacuation behavior, summanzed as fo!!ows:

b. Ahhough there are a number of sound reasons to be-m

- lieve that people will act out their iehavioral intention 5Hmothesu !

in tne event of a nuclear reactor emergency {6-9). fromThe locanon of a household's residence in relati l ,

to the piam will c:recth influence the evacuation de

' Autnor to whorn corresponcenet snodd tv mieressed _ casion. Two vanables. distance and direction of tne house from ine piant. are postulated to capture-in i 8709250422 870708 n PDR ADOCK 05000322 g PDR

- @)

w ) H JomN # an(, L J . h K. n

'8' i ax,d Ars y POWER r ' m r7vor I j, 1 OC.41Km .

[

h N

.i

$7AG($N N N.KLLAh ACCIDE NT RiSA jIDCn0M' TION L tFF j 4

CYCLE /

a

)

< g I

E V AOUA TION

. DECISION

> SOCIAL --

CLASS ~-

STATUS Fig.1. A generahzed model of nuclear reactor ernergency evacuation decision-making, i

statistical sense-the effect of the local geography of of radiation will be more likely to evacuate than those the reactor site on the household evacuation decision who perceive that they are not in danger.

in a radiological emergency, in addition to the foregoing hypothesized direct ef-H.ipothesis 2 fects, we also specify in the model a number ofindirect influences on the evacuation decision. As Fig. I shows, The household's stage in-life cycle will directly in- the household's physical location in relation to the fiuence the evacuation decision. Families in their plant stage-in-life evele, and social status are hypoth-childbeanng years, especially those who already have esized to influence the evacuation decision indirectly young children in the home, will be more likely to through both preaccident attitude toward nuclear evacuate than older famihes without children in the power and nuclear reactor accident risk percepuon.

home. l.

Young, high status families who live close to the plant l will De more likeiy to oppose nuclear power and to Hipothesa 3 perceive that they are at nsk of an accident than older.

'The head of household 3 social status will directiv wer status households who live farther from the re-influence the evacuation decision. More specihcalh'. actor site. Also. preaccident attitude toward nuclear high status households will be more likely to evacuate EP#[ '8 hypothesized to influence the evacuation Inan lower status famihes. occision indirectly through reactor accident nsk per-ception. Those who oppose the use of nucitar power H.ipotnesis 4 to generate eicetricitv will be more likeiy to perceive i

Tne head of household's preaccident stutude toward that thev are at nsk'of an accident than those who nuclear power will directly influence tne evacuation favor its'use.

decision. Those who oppose the use of nuclear power in sum, our proposed model of nuclear reactor to generate electricity will be more melined, than those emergency evacuation decision-making incorporates who favor its use, to evacuate in a radiological emer- attriDutes of both the location and the local population peney. as predictors of evacuation. In the section that follows we apply the model in a case study of evacuation be-H.ipothesis 5 havior dunng the TM1 reactor ensis, in the fmal sec.

The head of household's percepuon of the nsk as' tion, we discuss the implications of our findmps for sociated with a nuclear reactor accident will directly future-onemed radiological emergency preparedness inliuence the evacuation decision. Irrespective of the and response planning, and for reactor siting policies.

distance of their residence from the plant and of the emergency information they receive from crisis man 1 MPIRICAL ANALYSIS agement ofhcials. those heads of household who per. 3.1 D ::: and methods ceive themselves and other members of their family to Data used in this test of the proposed model were be at nsk of exposure to a potential or actual retease taker: from the NRC sponsored Tnrce Afi/c Island J

q t

Woelons evacumuon nenn.or 7eicpume Sunri (hereane relence te as g the NRC -

survey)l2) Designec to e trer m nwcholopcat pioved m this apphcation of it -

and econorme impacts of the 7 M1 accioent. meiuding nucacar reactor emergene) occis on.

the extent of evacuation. the W survey was aamirr fr -

in path analysis ordman icast squates regress.on (OLS) istered to a stratihed random sample of !$(O houw- procedures are used to solve the structural cauations holds residmg withm a $$ mile radius of the site ar* . correspondmF to the hypothesned hnkages m a causal proximately three months after the immediate ensis model, the loghnear equivalent uses the lopt formu.

was over. lation to accompissh this task [22). LOP 1 ts a statistical j Eleven variables were selected from the NRC survey techmoue which has been developed specihcally for I for the purpose of this analysis. These vanables, ahmg analyzmr relationships between dichotomous de' pen.

with the specific ouesuons from whi:h they were de- cent vanables and either cateponcal or independent vanables. nierval level I nved, are hsted m Table 1. Consistent with the structure

)

' of the proposed model these data are dmded mia en The proposed causal model of esacuation decision-makmg reduces to four structural equations or rnodels. {

openous and endogenous vanables. "The exoFenous variables consist of the survey respondents' locational one for each of the endogenous vanahies general at-

! (distance and direction of their homes from the planti, utudes toward nucicar power. attitude toward the TMI stage-m life cycle (age, mantal status, and the presence piant. TMl accident nsk perception. and the evacuation of a female who was pregnant and/or young children occision. In standard notation form (22j. the four couanons are as follows:

in the home at the time of the accidentl. and socth!

status (years of school completed by the head ofhouse-l hold) charactensues. The endogenous vanables consist 4 ,g4; 4;y of the survey respondents' preaccident attitude ioward nuclear power in general and the TM1 plant in partic- W. 23 . .u ) p, } (2) ular, their perception of the nsks associated with the 5Ai. Ab Avhko}

accident, and their behavioral response. (For a detailed D) discussion of the theorcucal rauonale for includmg :wo b8' 4 * # "M # " ! N measures of nuclear attitudes m this empincal test, see [5].)

As in the earlier empirical test [5]. loghnear causal where the terms are defmed m Table I and in Fig. 2.

modelling, a modihed version of path analysis. is em- All of the hypothesized direct and indirect causal Imkages in the proposed model are embodied in these Table 1. Concepts. vanabies and specihe measures Concepts and Vanables Desenpuon A. Locational

1. Perceived distance (xil Exactly how many miles do you hve from the Three Mile Island

2. Direction (2d Nuclear Power Piant? Exact mileage recorded.

B. Stage-irrfamily hfe cycie Coded into four categones: (1) East (2) % est (3) North M) South

3. Age of household head up

, Absolute age recorded.

( Mantal status ud '

Coded into 5 categoner (1) Single (21 Mamed (3 Divorced W 5eparated ($1 Wacowed I

5. Yount children in the home up L Any one pregnant m tne home ud Absolute numoer recoroed. i l Coced mto two categones: (1i yes (0, no l C. Sonal Class 5tatus

7. ) cars of school completed u l Absolute number recoroed D. Psychological

5. General attitude owarc nucicar power ud

', How do the advantages of nuclear power m general compare to its disadvantages. Coded into 3 catesones: (1) aovantages much )

{ ,

greater than disadvantages (2) advantages somewhat greater than  ; i

i j

tne disadvantages (3) advantages and disadvantages about the same 4) advantages somewhat less than disadvantages (5)

9. Atutude to TMi plant (W advantages much less than disadvantages.

before the accioent did you thmk the advantages of havmg the TMl  !

j nuescar stauon m your area were greater thar.. less tnan or about {

l the same as the disadvantages. Coced mto same 5 categones as in

10. TM) accident nsk percepuon u J Number 5 above.

How senous a inrest did you feel the TMI nuclear station was for i you and your famih's safet) at tne ume of the accident? Coded mte four catepones: (1) Ve*v senous (2) serious (3) somewhat of a tt:reat I4 I no threau  ;

E benavsor

11. Evacuauon Der: son W D.*1 ant rnemner of your nousenold eva;uate cunn; the weeks o' l tne TM macenC Cooed mio Iwe categoner U i ves (3 no (;

i 1

N' V W _ _

u,>

I H Jom W >> sn:Dl 20aJs

~~'

'. I r

x N o.27.// " * ' , ' ~N ',, ...

h w

lX m, %s.

/

/

m i

%--~.

~

\

\

s

\

\

I' l

l

! '"' es. \

l NA 3,12 1N j Y ^"

i N'"

! .,x l

]' N on 4

g ST

~

- ,7 - su -

d" 1

""m  ;

$ /

g

,./

s /

/,..............

. . . ........i.., . . . ...........,,, 1 a,.

o".<ig ;.3

, ,. . . . . .. ... s e .o n,. . .. i . z u s,. i v . ... .. i c .a e,. . . ..e t., o ey a s. . o re.. e e r i..a a n o n,, es e n ino ,

,,a n. .. . a s ,.a

,,. ...........o Fig. 2. Final path model for the Three Mile Island Nuclear Generstmg Facility.

four equations. Using logit analysis to solve in succes- which oftentimes results in a loss of potentially valuab sion these four equations the analytic objectives here mformation. In the present research context this is i are hrst. to derive estimates of the direct and indirect perceived to be not a senous limitation. mamly because efie ts (as measured by the beta coefhcients) of vanables the focus is on the validity of the hypoinesized causal !r Aithrough xm on the evacuation decision (x n

). and sequences among the exogenous and endogenous van.

second, to determine how wc!! the overall model hts ables, rather than the magnitude of the efie:ts along the observed data. It should be noted here that since the paths in tne model wht:b would recuire more pre-the logimear version of path ana}ysis is unable to assign single values to causal paths when polytomous vanabies a cise measurements. For ine specific purpose of this re-s r:r.. m short. we beheve we have scie ted the best are involved [22) it was necessan to di:notomtze tnose asailaoic cata and methods.

vanabies m Table I with more than two response cat.

epones (xi through 4, M tnrougn xml. 3.2 Tmdmg5 Before proceedmg to repc t on the results of in:

Ftgere 2 0:p::ts in: hnal path model of evacuation empi-ical test, we should address two concerns which cecision makm; dunng the TM1 reactor ensis. Tne conceivably could be raised noout ine data and meth.

ods employed in this paper. The hrst pertams to our discussion which fobows is organized around the fwe reliance on a sun'ey which was conducted three months ma.ior research hypotneses advanced in tne preceding section of this paper.

c/ter the TMl accident and yet purports to measure Tne log.istic regression analyses of the NRC sun'e3 precccedent attitudes toward nuclear power. We feel data indicate that the evacuation decision auring the confident that the NRC survey provices a reasonably reliable indicator of the pt:ateident attitudes of south- TMl reactor ensis was directly influenced by the ic-canon (distance and direction) of the resioents' homes central Pennsylvania area resioents for the following in relation to the plant and this lends empin::al suppor 3 two reasons. Erst. recent analyses of colietted survey to our first hypotnesis. Figure 2 shows that peopie closer research on nuclear power have shown that public at-to the plant (within 3 miics) were more likely to have tituces have remained fairly stabie smee the early 1970s evacuated inan people fa ther from the enppled reactor

[12. 23. 24); and second. cons: stent with these studies, (beyond 5 miics). This suggests that evacuation from a chi-square test revealed that the post-accident atti.

TM) confo med to the classic distance decay phenom.

1 tudes of TM1 area resioents. also measured in the NRCenon which has been shown to charactenze public be-

.lg survey. were not stattsucalh difierent from attitudes havior in response to other kind of environmental pnor to the accident. The second concern whien might stressors [25. 26). As Fig. 2 indicates. in addiuon to its j be rmsed is related to our need to cenotomize variab:es.

- direct snfluence, residenna; proximit.s also infiuenced 9

F V

.. _ _ _ _ _ . _ . _ _ _ i

i l i' 1 Munehms occuanor renevio*

is

/ tne evacuanon occiuor, monecto intoup T Ml acci-oent nsi perception up more haeh to have evacuated than tnose who were s supponers of botti .

Along with distance, direction from the piant ano -

directly mfiuenced the evacuation o* cision (Fag. 21. Fmah3 . tne NRC surve3 lends empincal support to llouseholds residmg south and cast of the piant were our hfth hypothests: that the evacuation occasion will less likely to have evacuated than those hvmg north be directly mfspenced by the head of household's ac.

and west of the sitc. This hndmg is consistent with the cioent nsk perception. Those household heads in south resuhs of earlier behavioral studies. which indicate that central Pennsylvama who felt that the malfunctioning the pattern of evacuation from TMI was directionally TMI reactor posed an immediate threa' to personal biased [2,4J. As Fig. 2 shows, the direction of the place and family health.t;afety.and welfare were more likely of residence from the plant 2(2 ) also influenced tne to have evacuated than those who perceived that the

'i evacuation decision indirectly through the household enppied reactor posed no threat at all (Fig. 2).

head's perception of the nsks assocsated with the TMl in sum the results of the logistic regression analyses accident (x,,). on the NRC survey data suppon nearly all of the hy-The NRC survey data also support our second by-pothesized direct Imkages in our model of nuclear re-pathesis, which states that m a nuclear reactor emer- actor tmergency evacuation decision tr.2 king. But ex-Leney the evacuation decision will be directly milu, aetly how well does the proposed model ht the data?

enced by the life-cycle stage of the household. Two of . In the loglinear 2 version of path analysis the goodness-the four stage-in-life cycle indicators included m this of-ht stausue is r 1which is analogous to r in OLS),

analysis, age of household head3(A ) and the, presence and for our model the value is .21. which is indicative of a relativel 3 good ht, in fact, Henser and Johnson )

of children m the home (4L mfluenced the evacuation decision directly (Fig. 21. Famihes in which the head tremelC7) consioer r* salues between .2 and .4 to be cA.

of household was under age 35 and/or those with young 3 good his. Also indicative of the goodness-of-children at the time of the accident were more likelyht. when the overall model was apphed to the mdividual to have evacuated than famihes in which the head was cases to see how well it could predict the known re.

over 35 and/or no children were present in the home, sponses,715 were correctly reclassihed into the groups

{

in addition, both of these vanables (A to which they were origmally assigned on the basis of 3 and 4) influ- their actual behavior dunng the TMl ensis.

enced the evacuation decision indirectly through l preaccident attitudes toward nuclear power (4). As hypothesized, young famihes with children were more ( SUMMARL CONCLUSIONS.

likely, prior to the TMl accident, to oppose the use of AND IMPLICATIONS nuclear power to generate electricity than those over Dunng the emergency at the Three Mile Island nu-age 35 without young children. As Fig. 2 shows, neither clear generatmg station, evacuation becime a common marital status (x.) nor the presence in the household adaptive response among the local population. The at the time of the accident of a woman who was preg- spaual behaviors that manifested themselves, however, l i l

nant (xs )inDuenced the evacuation decision directly; did not conform to the norms established on the baris however, both variables influenced the decision indi- of evacuations from natural disasters. Nevertheir.ss, rectly through interaction cEects with age of household planning for nuclear emergencies has proceeded as if heads (x3) and the presence of young children in the there were no diKerence between the Three Mile Island home (As). expenence and other types of disasters requiring evac-  ;

uation, in the NRC survey, we also fmd support for our I third hypothesis, that the evacuation decision in a nu- The geographi; dimensions of the evacuation process cicar reactor emergency will be directly inGuencedatbyThree Mile Island, that is, the magnitude of the evacuanon and its large geographical extent, have con-ine social status of the household head. From the data sc we selected one indicator of social status years of vmeed us that radiological emergency response plan-senool completed (x,). and the results of the logisur mng must incorporate wnal we know about the nu-regression analyses indicate that famihes in which the cicar response behavior of human populations. In this ..

household head had completed 12 or more years of paper we have drawn upon the TMl behavioral data -

school were more likely to have evacuated than families compiled for the NRC in the aftermath of the accident m which the head-of household had less than twelve and employed loglmear causal modelling to test a prc-vears of formal education. In nearly all of the post. posed model of nu: lear reactor emergency evacuation accident beha vioral studies, this indicator of social sta- cecision-making and behavior. Developed speci6cally tus (x,) has emerged as a statistically signincant cor- to assist emergency management ofheials better un.

j relate or determinant of evacuation. Cutter and Barnes derstand public behavior during the TMI reactor crisis, 6 l

[18) concluded. for example, that "tne more educated the model specihes the main factors that influenced the individual, the greater the pro emergency deersion-making at the individual house-dividual to undertake evacuauon "pensity for that in-hold level and how these factors. through a fairly com-  !

And Hu and Slavs-man's [19] analysis showed better educated households piex web ofinteractions, prompted a large number of l

' to have evacuaied at a rate 10% higher than less edu- tne local residents to evacuate, even though they had t.

cated households. not beet advised to take such action The model ex- I TiAt the evacuation will be directly inguen:ed by hibits A relatively good fit to the NRC data which leads the head of household's preaccident attitude toward us to ton:lude that nuclear emergency evacuation be-nuclear power is also conbrmed by the NRC survey havws is predictable on the basis oflocational, life- '

cata. Figure 2 shows that those residents who. pnor to ;ycle, social status, attitudinal, and risk percepuon vanabies. ,

the accident. opposed nuclear power in penera! (Ad I anc ine TMI piant in parucular txa (b = 1.04 were if we a 0ept the IM] accident as th0 most appro-pnate analog for t redictmg response to future nucicar m i

.Es snow-

n J h >ve. >6- undD.J 21cak- i 1

i reactor accioents. Inen tne results oftnn study suypest inat the guiceisnes which have tieen cuannsned for tiott. ac!!are should ar; accioen occur once tne facihty =

constructed. heensed. and bepns to generate peer Ine preparation and esaiuation of obsite emergency

/ preparedness and resronse pians in support of custing ,

i I

• nuclear power plant operations need to be revawd to R EERIACIA -

account for the behavioralpredispositions of the local

• I. M K Goldnaner and J. E. Lenman. Cnses esacuation population m the event of a radiological emergenet cunng the Three Mde Island nuclear accident. Paper pre.

Fonunately, in revising the federal regulations m the 5#"'#d *""' d *" ##""# "U"C d *#'" #" I"M"'

aftermath of the TMl accident. the NRC and the Fed-

" # " '"" # ""' D 'h*' "d"E3I g eralEmergency Management Agency (FEM A) laid the , C b Hvnn. Three Mac h/and 7e.spnnne Surver Pre-basic groundwork for such a behaviorahy based an- f,,n,no,,. 3,p,,,u,f p,,ggy,e5 and /mdsnri U.S. Nucicar t

proach to nuclear reactor emerFency preparedness and kegulatory Commission. washmrton. D.C 00?9t response planning [28). Under the revised regulations 3. K. barnes. J. bro 9us. S. Cuner and J. K. Mstebell Re.

spon3c3 of/mpacied hipulanon to the Tnree Mile hIund licensees are required to distribute fat least artnually) Auc/cor Aeonor Accadent An /nneal Asacssment Dmus-i to people living with 10 miles of their facilities infor. sion Paper No.13. Departrnent of Geogranh>. Rutgers 1

mation pertainmg to the nsk of radiation and the range U "ff58'F New brunswiet. New Jersey 0979). l of protective action they might be asked to take in case 4. S. D. brunn J. H. Johnson. Jr. and D. J. Zeigler. I' mal of a nuclear reactor emergency. It is this population Report on a 5ncsa!Surver of Three Mt/c Island Arca Arv within 10 miles which probabihstic nsk assessment sacnr3 De anment of Geograpn) Michigan State UnL versity, Last Lansing. Ml(1979).

studies indicate would be in the greatest danger if an 5 J. H.Jnnn on. Jr. A model of evacuahon decision-making accidental release of radioactive matenah were to occur m a nuclear re.netor emergency. 6 corr Act. 71. 40.b4 Ik

~ [29). The regulations also instruct hcensees to arrange NWS) for an independent review of-their emergency pre. O M. Fishbem and 1. AJzen, /fc/Jel. Annudc. /nsentmn and paredness program to be conducted at least every 12 henarmr An introducithn to Tiscore and Research Ad.

! y months. The results of the present study and of our d'50n

• c5) RC8d'ng. M A 0 97$).

1 previous research (S] suggest hrst, that heensees should 7 1. Aizen and M. Fishbein, Onderstandmr Annudes and be required to expand their pubhc education and in- j

("y w enuce Hau, WeM formation program to encompass the population 9 j P within the 50 mile. instead of the more hmited 10 mile, E. J. H. Johnson. Jr. and D.J. 2eigler. Disunguishmg human emergency planning zone of their facihties; and second. resrvonses to radiological emergencies. Leon 6 corr 59 36t.-102 (1983 L l

' that the independent evaluation should be designed to 9, D.J. Zemier and J. H. Johnson. Jr.. Evacuauon behavior assess the effects of this information on not only pubbe in response to nucicar power plant accioents. Prof Geor

w. awareness and understanding of the emergency plan .% 207-215 (1984).

but also nuclear attitudes and reactor accident risk 10. J. R. Eiser and J. van der Phgt, behefs and values in the perceptions. Elsewhere we have argued that the data nuclear oebat:./. A/>phed50c. PJvck 9. 524-536 (1979L necessary to conduct such an evaluation would have h, y'd R e c.

to be gathered via social surveys, which also would be uba A J Stolu designed to ascertain information on the locational. Risk and beneht percepuons. acceptabihtv .iudgements..

social background, and demographic charactenstics of and self-reponed actions toward nuclear power. / Soc 4

the 50 mile emergency planning zone population (8). Psrrh 186. IA197 (1982L

~l Tne application of our model of evacuation decision. 12. E. D. Melber. 5. M. Neate>. J. Hammersta and %'. L Rankin. Aucicar hwcr and the Pubirc. Analtso of Cv/.

making to these data would assist emergency planning /ccicd5urvei Arsearch. banche Human Afiairs kescarch ofhetals in estimating the magnitude of spontaneous Centers.5eatuc WA 0977L evacuation and assessing its hkely impact on the time l '- H. J. O:way. D. Maurer and K. Thomas. Nucicar poact; a

n would take to evacuate the 10 mile plume exposure tne cuesuon of pubh: acceptance. hnure.i 10. IM-; t t pathway zone population. U Ob)

Our fmdings also suggest strongly that locational K P. Sm B Fasennof. and S. Lientenstem. Facts and fea s.-

factors othec than population density and distnnution. unoersunem; perceived nst. pp IM 214. In n C.

5:nong and W . A. Albers decsl Socmal LM Assrso the two enteria which have been tne primary focus of ment: HowSa/c ts $d Lnouch'henum Press. New i ort recent geographical research on reactor siting pohcies 0480L in the U.S. and the UK [30. 3!. 32]. should be taken 15. P. Slovie and B. Fischhof. How safe is safe enougM pp.

into account in future sitmg accisions. In our model 112M in C. A % alter et al Ieds.). 70s Hot w handec' we have used the distance and drection of the indi. Yale Umversii.v Press. New Haven. CT (19f 3L vidual s residence from the plant as surrogate measures an oer .

and R W Puk annuces of the impact of the local geography of the area sur' to nucicar energ). Enert.r toher 12. 302-305 0 0W 17 D. J. Zeis)er. 5. D. brunn and J. H. Johnson, Jr.. Evac-rounding the reactor on nuc} car emergency decision-making and behavior. In resht), however, a much untion from a nucicar te:nnological disaster. Geog Acr.

71.1-16 09 E I).

broader range oflocational attributes, including pnys-

!L S. Cutter and K. Barnes. Evacuauon behavior and Tntre ical features in the local environment and loc.al land. Miie 1siand. Ihscsiers 6. I16-124 09E2 L use patterns, influenced evacuauon decision-making 19. T. Hu and K. 5 Slaysman, heahh Aciaird Econorruc dunng the TMl reactor crisis D7). We believe. there. Costs ofthe Tnrer Mitchland Accident. insutute for Poh fore, that in the future the evaluation of proposed re, ic> hesearch and Evaluauon. Center for Research on actor sites should consist of a very aetailed assessment Human Resources. The Pennsylvania State Universit3 of Inc potential impact of tne geography of tne local University Park. PA U981L area, including the use of socia? surveys to forecast m L S. Miher. Ascosumnaland5nuauonal Foreab/c.< Ar-jord ic Docuanon at 7nrec Mile Island Unnubbsned evacuation behavior as descriDed above, on eme gency planners' c!Iorts to protect pubhc heshh. safety, and Pr..D. Tnesu.. Depanment of Geograpny. Temnie L nr sersin. Pnnacemnk. PA f!4Fi t j r

i n.i

s.

3 ' Moothing evees.auori retavoor.

+'

_,. :ll P. D. Cecan and f' 5 Huuts ine kenosupscai smtmet. I2t NRC'YEM A. Cwerw uor treparanon o

-t: of the Tnere Mne hunc inisoen: .l truman horen 86.

2b-34fl9Mi kaanovico' E.mergenes Aesponse hans anc Freparec. . '.

ness en Support of bucteo* h*mvr hann U.S hucscar 22 D. Knoke and P.J. Burke Leonea Hugeh Saar Puh -

heations, heverh Hahs CA (t480) . . hermatorv Cornmesuon and Feoeral Emergenc3 Man.

23. B D Meiner. The impact of TM1 upon the pubut ac-'

asement Agene) % ashington, D.C. (140).

aptence of nucicar powet, Prog m buct Energy EU. 387- .29. H. L Colims. b.. K. Gnmes and F. Gaspin ieds.). Mannme 398 iI983).' .

' hasu lov the development c/5 tate and bra! Governmew '

. RadioiopacalEmergency krsponse hona U.S. hvetcar

24. R. C. Mitchell, The Public Aespan.se to 7perer A(t/c lstand ,

Compilanon ofPubhc Opmion Data About hucleo' En - kesulatory Commisuon. Washmgion. D.C. (1978). .

ergr. Discussion Paper D 58. Resources for the Future.; 30. J. Ftrns and $ Openshau. Nuebear power in the US and Washmpton. D.C. I1979). .

UK.The role of sning m safety philosoph). pp. 192-214..

In F.J. Calzonetts and B. D. Solomon teds.). 6eorraphsco!. .

1

25. M. K. L.indell and T. C. Earte. Hou close es close enoughL pubhc perceptions of the nsks ofindustnalfacihties. AuA himensums of E'nergr. D. Reidel Dordrecht. Hohand Anaksu 3,24 b253 (1983). 7(1985).

26. M. Harve). J. W. Frazier and M. Matuhonis.' Cognition - :31. 5. OpenshaA The geography of reactor unng poheses in of a hazardous environment reactions to bullato airport . , the UK. Institute of bntish Geographers. 7tansactmni 7, ! $0-162 (l982). .

noise. Econ. Geogr. 52, 263-286 (1979)

27. D. A. Henserand L W. Johnson. App /ws/Meserete Chmce 30. S. Openshnu. The siting of nuclear power stationfand.

Afode/hng. John Wiley and Sons. New bd (1981). - pubhe safety m the Uk. Aegsonal5tudies 16.1S.bIVf '

(1982).

i O

l a t

') . . .

hq

[

8djs ~

^

d M  %% B

&h

$D4

$Mf$ MQ@%Q NUCLE

&h&

AR

?Y?$2*% hf i h  ; REGULATOR  % COMMISSION  % $N&

QWg% C Mtg?WW wgA  !%@@AW"&E&fh$$%)&$$kik$h d pan $@hy Q$PWp WG%MWdW $$h &dQd S g &Rm pf&Mm fA My M m-N &W &n d udMM %albw +& M p$:

n QM6%MQW"pbmWd

&upw m :;m&,y m $ mr e4p f

aw M

f ME MO k y R A N DLUMif QM4fdQ 0 RM$%d h 6 h N W

@ @y $P$gnMkm$$a M$$ i MW ffg%p@h h 4Wddf%m@%@b

$5W$4fROM@%@^%M@h a n , t e tin gic hje fgheb$pppQg@ Emi 1leN LMu lfi h

$hkNkkdh$hkhkM@kkkbkM$MhkkhNhhfhh N hM Mb@@[MAnys

@4 W WR.g F iceed.i.nglandimaonsidere.

' M Wexhib m. ~

filed d o c ume art %fytherofficialthearingMe ro n t sa'd sproi$N$g$ eiasNW m p$ gg e,w r GWraS t a t.,.i t?. i sM, . i g _h tT. A, ,b t,d,v,e xem pet.d.4arf

'n w n> g ith the.opy~

o 7 t'w esEC i

U c o u r tt

+

e x p bxu~nlessd

'r e sXbes l,yn wc~a t9wa n u~s hofscopy.righ+ttconcer.ns s%o e r,i g i n a 1 iundernseah 1

47a y+Rf,1 y.

1 ign A

. . . 'y.> mwi . y . ;gj ' ' A

• L,Ja w.e

' ~ ~

, pf f < ~ <

g - , .

. . g pgggggggg m ay . .wvm p 4. opp /%gQ my 7 A l h o ff,s.t h e,idco;u m e n.t sg s e n t3t o6T Iyf o rlp r o c'e,s s i.n gMa, l Wi t h.ip,t,h e @4.n#,, ;W W

' Fe x. e. m. , p t, ?c.l a s s i f i, c a.t i'o'.n.D r "

M

.1 A x. W&

x ,,

< P.xp. _, ,. i -@,

og Q' NqA; j yM h WQf~ ,W:h

,3 t vQ myQ Qlp Q._f Q >' ' '

ghg;;lQ':p r4 ( p  ! q.g> 7Q 7 m, g.,. 4."  :' sy i r .4 . A_ ..!

,. ;3 ,

? .

, 'e s%. . u s.  ;;1 ~,,w'%,.n g ;. Q i;/ s .

_L 6r A

}v ip% g

..y . s 3 w,., a

" i

, 'J; f '  ; ,(} d: a?t \

06 p;,

g a' Mi .-

• qp

. g;: .. s

-~ $m:Q.3 1 .,

i

^

>,%Q J. .g f +< r  : Q,{

.a

,Q, A

[~.' g,d g

?.! ( AiT ( uT-p# 4  !

' ' g 1 g g, V <

c(

Q' .- l a! MQ , * '

'Q7, C $1g{.

i , -,

1. .M;3 ;, y a

</,

~&~' 1 .. 4n n.4.

rc.yb 4;. g:b Wby 7:;Q u .c t '"

i ig 2 4 Iw ,

k- 3

, m 4,

y.:

,4: -

$ -w,,. l 'I '

6 v '

r:M! $Nm

{ 3Ny

• 'qa t i. ,

S

),}4As

, ,n xy;! d ,  ?

d S - [f t:1.J.

'.e sn y ( ,

4b N j~q i

., :. ft .y )

, ^,

, f, . 's f' i A ,

'. > J i'Q &)

, , , 4 s

ds M_ +

sp*s t, i T #

i

' , :f a q t <

, , , 4

1. I ,3 p

la ,..$.,-4 s t x

3

~

a .

e v:v .- .

l -

Al a "

w

,, 4,

- y, ' yK.14

.f

e .

wx . -

g g@.?IQ i

.y y' f;19 y u ,

sm '

y 1

V t  %(

5kwq,c {! f w , t 1 1,

3 l'

~ u f.y 1

e x - ,

w

,7-,

m

" 1. q

+ '.t .; i ,'

a g :f t

u ', >

Ri ,

<Q v

CM . . .. < lldj B 2

~ < , " Mf o.,,-.

. c $1 4

v

a

,ju; M'.

p% <g Q i r ,

n,:

-n n.,:n'. f.2:n

, ~

o a wv+v ig r+

y hl/

17' y.w :,

e r

m&g fj

v w: , .y , w , ~,rs.

mu u

,) -. 'O l b R, j , '

.l g e F '

, e s

,~.9 4~',.4;. s4

,4 g< -.

g#q qq.-

7 p:- ,, w y ,,u 4 r W,s 'y "

M " '

M .^ '

W OQb ,i::LQ, ,

e .7u ,

tf D.. < g jpy, p 4 i i ,

< 1 .ir " pn i .

- ~; R uiO' . . !]

3; M.a rp b

>4:

e -

4 w

', ,y; 8

S h l- >

, t 4

cp  :%gg,1 t

.g k

3 ,. k()

h/v gm,o- Yi@ b .

, O S ', ' t

,, ., M,

) h- g

, e.

'N I 'J p , II. * ' b *# '

Oi > > i ,

'i s i

M,rj w GP <i.

f.. ,i h'

l r

- g 3

< c

'7p 's, y t M

' , 4 e r b i v. a 9

t , I,[.

'y .,j t,

1. ,J

[ 91 < r ' i f fq

\ - ~;f' ( [ N i

+ ,

g; w c{

- v i, "

n

',i 4

'p ;r; ,

y

+

m i }/ a.

(

- .; ,  % ...I' '

, F T i {\y C g e b 710, < p.g.Q t q;; '

-s Q '. .

(p pyn;; r < g: <; , s

,4 po

.sr , y . m r i i; h Y '

M .

g if 3$M4 %Q:1 QpQ:2: Lll' Y , NA +

\Q Q , t lW., J

'. , @. ? W'?A . "

e ,

i? , E M-tm >F m r F ~ ~

M l%.W! 7,g f

% +11 '

, , a %

M. ;

• W f @j$ g@% 4g P k;e f M g' hl /Q$,g ,;g ,

i i,

,;g ,

,s m" " ,

/ N.

%:i;MMQp s%iM s Wg '

1. g@viiWA%gi t

4 O