ML20128Q448
Text
,
7- 11/84 PREPRINT: FOR PROCEEDINGS OF THE EIGHTEENTH MID-YEAR TOP.ICAL SYMPOSIUM (ENVIRONMENTAL RADIATION '85) 0F THE-HEALTH PHYSICS SOCIETY, JANUARY 6-10, 1985, COLORADO SPRINGS, CO.
OBJECT!vES OF EMEREENCY RESPONSE AND THE POTENTIAL BENEFITS OF
_ , _ EVACUATION AND SHELTER James A. Martin. Jr.
Of vision of Risk Analysis and Operations Office of Nuclear Regulatory Research U.S. Nuclear Regulatory Comission Washington. 0.C. 20555 i
A8STRACT o
Basic radf ation protection objectives for the controlled enviroment are transferrable verbatte for use in the uncontrolled or emergency situation. These '
objectives are: avoid near-term injury or fatality and reduce individual risks and total health eFEcts to levels as low as reasonably achievable (ALARA). The potential benefits of sneltering and evacuation in meeting these objectives during
- a response to a severe LWR accident was investigated. It could be very difficult to reduce total health effects because collective dose (man-rem) increases mono-
' tonica11y for scores, pertiaps hundreds of miles from a release ootnt. In this regard, shelter and ad hoc respiratory protection appear to be the only rational and feasible near-tem protective actions that would be available to most people.
However, the first two objectives can be met by early, precautionary evacuation, within two to three illes imediately upon the declaration of a General Emergency (e.g., a core melt accident), and sheltering elsewhere. In the event of an actual major release later relocation from highly contaminated areas would be an integral part of the mergency response. Analyses performed to investigate the various emergency response options and potential benefits are described. The perspectives obtained should be reflected in emergency plans.
INTR 000CT!0N Radiologf cal emergency plans for ffzed nuclear released to the atmosphere) (US81; US82a). Al though fact 11 ties should be constructed to achieve spe- these " source terms" are being intensively inves-ctffc radf ation protection objectives in the event tigated (Ba84) and may be revised in the future.
of a future radiological release. Basic radiation the NRC set was used as the best current estimates protection objectives for controlled envirornents, of severe accident source tems. The important such as a tark place, can be succinctly stated as characteristics of this set are Ifsted in Table 1.
follows:
o These would all be core melt accidents. The AVOID serious'non-stochastic radiation effects estimated probabilities of these accidents occurring (i.e.. near tem or early, injuries and are very low, but the estimated release fractions
' fatalities), and (of the core inventory) to the atmosphere would be o large, especially for SST1. In general tems, REDUCE individual stochastic risks and total SST1 would correspond to a coincident early.
latent health effects to levels as low as massive failure of contaf ruent, with little or no reasonably achievable.
scrubbing of the release by engineered safety features in a plant. S$T2 would correspond to a These basic objectives are transferrable directly coincident major contafruent failure with degraded to the uncontrolled or accident envirorment, so perfomance of engineered safety features. SST3 long as it is recognized that there can be no 4
guarantees proffered that all objectives can be could f avolve a late basemat melt-through accident.
met in all conceivable ciretastances in an uncon-with efficient scrubbing of particulates and a trolled envirorment (US78; US80s In84). For this smaller release of noble gases, as well.
paper the potentf al berefits and practicality of 8ACKGROUNO cabinations of evacuation and shelter to achieve these objectives in response to severe Ifght water nuclear poter plant (LWR) accidents was investigated. Clearly, even without an emergency plan see off-site mergency response by the pubi fc would be expected in the event of these accidents. Examined The LWR accidents considered was the set
=
SST1. $$T2 and SST3. suggested by the U.S. Nuclear here tes the question: iftet would be a practical Regulatory Comission (NRC) as representative emergency protective action schee involving groupings of severe acctdent source tems (frac- evacuation and shelter, idifch would satisfy the tions of the core inventory of radionucifdes basic radf ation protection objectives? It was recognized that indf vf dual entrapment situations 8507130434 850426 PDR FDIA )J BELAIR85-199 PDR ,h
can be readily visualized; thus, the perspective .o was to investigate evacuation and sheltering In the further ever.t of an actual major tenefits that could accure for most of the people atmospheric release from containment, rnost of the time given the postulated accidents. people in shelters would relocate from highly contminated areas lef t in the Certain leading clues were available from we e of the m b ase, previous studies. Three are especially noteworthy.
In NUREG-0396 (US78), risk y,,s, distance was displayed It is noted that most core melt accidents In a simple manner in a figure reproduced here as would not involve early containment failure (US75 Figure L This figure serely indicates that dose U h84h WG f t h WmW wrenth (and risk) y,s distance decreases monitonically that only about one out of ten core melt accidents from a source point for an atmospheric rol ase. might lead to an SST1 type of reletse. on the Theoretically, the decrease varies as r- . per% For h m% W Me "u%
approximately, where r is the dowmsind distance. precautionary evacuation...in the event of a core Because of obstructions, wind meander & wind ,
shifts during a release period, dose might more realistically vary as r 8 (inverse squared). The A version of the CRAC2 code which provides r * (inverse distance) curve in Figure 1 was included as an aid to the reader. Without further the calculations (US83a; US83b). People in elaboration or caveats, the information in Figure 1 g can be taken to indicate that protective actions for each of the three major pathways external within a few miles of a release point would be gammas from the plume, inhalation, and ground most beneficial because the risk is clearly greatest contamination. These protection factors are within this distance. typical for residences with basenents (US75).
Two other clues were provided in NUREG/CR-2239 (U582b) especially in two figures reproduced here Evacuation was modeled in the first or as Figures 2 and 3. Figure 2 displays the probabi - near, zone by assuming a one hour delay af ter an ity of exceeding various nebers of early fatalities initial warning (e.g. declaration of a General given an $5T1 accidental release and various Emergency by the plant), and a 10 mph evacuation emergency response assumptions. The bottom curve speed radially away from the plant. Under this in the figure clearly illustrates the potential assumption, people begin to leave one-half hour benefits of a minimal delay before evacuation. af ter the SST1 release begins. The assumption For the sianmary evacuation curve; delays of 1. 3 that people leave one half hour af ter a major and 5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br /> were assumed to occur 0.3. 0.4 and 0.3 release begins is a generally pessimistic one.
of the time, respectively. thfortunately, for the This certainly should not be the planned sequence, bottom curve in this figure an evacuation within This has been noted previously (Ma77. Ma80 Ma82 twenty-five miles was assumed, at a speed of bub 4).
2 10 miles per hour. This implies that a large area could be cleared in 2.5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br /> - hardly a practical The second, or mid, zone extended from the assumption for most people most of the time, inner zone to 10 miles. People in this area were assumed to relocate af ter four hours exposure to Figure 3 contains two important clues. Shown ground contamination left in the wake of the puff.
here is the conditional probability of incurring The third, or far, zone extended from 10 miles, an early fatality beyond various distances assain9 Hare. people were assumed to relocate af ter eight the SST1 accidental release and various energencY hours exposure to ground contamination. These responses. Again, the importance of a minimal relocation times were estimates of the time it delay before evacuation is clear from the bottom would take to locate hot spots, provide notifi-curve which indicates that with a short delay time cations, and for people to move a short distance the early fatality distance should not exceed two (Ma77. USB4) away from the hot spots.
miles. Again, the impractical twenty-five mile evacuation distance is noted. Further, as indicated The New York City meteorological set in the In this figure, all shelterees were assumed to CRAC2 data files was used for the calculations.
stay on contaminated ground for a full day before This set contains rainfall about eight percent of relocation to uncontaminated areas. This is the time. Rain can cause heavier than normal hardly a realistic assumption considering that ground contamination. Also, the population distri-dose rates in many areas could exceed 10 rem per butions in the CRAC2 data sets were used.
hour in the wake of the puff (US84). Relocation RESRTS fras shelter would be expected to occur soon af ter radiological monitoring teams identified such " hot spots". Individual Risks CALCMATIONS Principal results of the calculations are displayed in Figure 4 and Table 2. As shown in Following these clues, consequence estimates Figure 4. for an 800 megawatt - electrical LWR at were performed using the following protective a coastal site in the U.S. Zero early fatalities action assumptions: was calculated for the most severe SSTI accident postulated for an early evacustion distance of o Early, precautionary evacuation within three miles. The uppermost curve in this figure
- 1. 2 or 3 miles and innediate shelter shows that the predominantly sheltering protective elsewhere. In the event of a core r alt action strategy clearly suffers by coeparison to accident (General Emergency). the early, short range evacuation strategy. These results bear out the intuitive, qualitative perspec-tive illustrated in Figure 1.
-- ~
,c -
r -
V' 's M "
~M' Mw _~
g Q' W.f_'"__""_'____~_~-~---~-~~---------------
-- ^ ~
Identical calculations were performed for These perspectives are illustrated in Table 3 four other LWR sites in the U.S. Principal results and Figure 5. Table 3 is a summary of pertinent fir these five sites are shown in Table 2. Although infomation in NUREG-0340 (US77), dich illus-the CRAC2 population distributions of these sites trates the pathway and temporal contributions to were used for the calculations, the results we w total calculated latent cancer fatalities for the normalized to 500 persons per square mile within PWR-1 and PWR-2 accident categories of the Reactor 10 milet. The powe levels of these LWRs ranges safety Study (US75). These accidents would be of ever a factor of two. The results for the lowest the ilk of SST1, and would include substantial power ?evel are indicative of what results for the releases of long half-lived cesium radienuclides, highest power level would be with a factor " 'wo It is readily apparent frca this table that long reduction in the SST1 source ters. term exposure pathways would dominate the total number of latent cany rs. Only extensive and Several interesting aspects of the results expensive decontmeination and condemnation are shown in, or may be inferred from, the infoma- processes over the long tem would be efficacious tion in Table 2. There were no residents within in reducing total latent cancers.
three miles of site neber one, and zero early fatalities was calculated. This illustrates the Figure 5 illustrates how collective dose and potential benefit of a very early evacuation of total latent cancer estimates increase with distance, n:arby areas, i.e. before a release given a core This figure was taken from NUREG/CR-2239 (US82t).
mel t. For the highest power levels, a few early Calculated total latent cancers accrue to large fatalities were calculated for SST1 for site distances, regardless of the composition and number two even with a three mile early evacuation magnitude of the release. Indeed, for the average assumption. These people were caught by the front site fully half the latent cancers could accrue ci the plume, in most cases. This was a high outside of fif ty miles from a release point.
population density site. In all cases, several This phenomenon has been noted previously for tens of persons suffered early injuries (e.g. routine atmospheric amissions of purely noble prodrcznal vomiting). These calculated injuries gases (M474). A corollary is that near field occurred at various azimuths and to distances to emergency protective measures would provide little 12.5 miles, but mostly well within 10 miles. The benefit as regards the objective of reducing collective total neber of early injuries calculated is an dose and total latent health effects. However, artifact of the CRAC code, dich adds up calcu* for a purely noble gas release the cloud (external) lated injuries wherever they are located. At any cansna dose pathway would dominate. In this case particular azimuth, no more than a few early Immediate sheltering to large distances would be injuries was calculated, dich would be the case an effective protective action to achieve the for a single puff. Further, these injuries occurred objective of reducing total latent cancers ALARA, at low conditional probablitties. especially where sheltering would not be inconvenient anyway. This would be little different from an In all cases, peak early fatalities and air pollution alert.
Injuries were associated with rainfall, a sudden calm af ter transport, or stable meteorological The interlay between estimates of stochastic-conditions (low wind speed, narrow plumes, nightime latent cancer fatalities and costs of condemnation conditions). Emergency planners should be espe- of contaminated property is illustrated in Figure 6 cially aware of the import of these particularly for the SSTI accident at the Indian Point site.
adverse, low probability weather conditions, dich The data points for this figure were obtained lead to heavy ground contamination by particulates. from NUREG/CR-2239 (US82b). These are for various Separate calculations, not shown, indicate that for dose projection criteria for land interdiction the SST1.elease, calculated early injuries could be (condemnation). Nornelly in CRAC2, people are Gliminated by slightly snaller particulate source allowed to remain in contaminated areas where the terms (lower ground contamination), better shieldin9 projected whole body dose is less than 0.25 Sv (25 f aster relocation or combinations thereof, rem) in 30 years. As illustrated in Figure 6. for SST1 a very large increment in costs would be No early fatalities or injuries were calculated incurred in reducing total latent cancer fatalities for the SST2 and SST3 accident scenarios, for the by interdicting property at a lower dose projection.
noted energency response assumptions, at all power levels. It should be clear from Figure 1 that he additional perspective is important in early, precautionary evacuation of nearby areas in this regard. For a core melt accident with fatture the event of a core melt accident would significantly of containment, the constituents of a release may reduce individual latent. cancer risks in the event not be known for some time (Ma80; Ma82). Thus, cf a core melt accident. the shelter to large distat.ce (where convenient) option should be predetermined, as well as the Collective Dose evacuation to short distance option, as an insnediate response to the declaration of a General Emergency In contrast with individual risks of non. (core melt). Indeed, it was in this Itght that stochastic effects, dich would be relatively the protective action assumptions listed under near-ff eld or close range ef fects, estimated total CALCULATIONS, above, were made.
latent cancers would increase monttonically with distance. Further, for releases which include a be caveat is impertant here. These collective substantial abundance of long half-1:ved particu- dose perspectives derive from the asseption of a lates, the collective risk would be associated proportional relationship between risk and dose.
with long ters (years) exposure to grea:nd contamina.
tion. Thus, protective actions during the energency phase wuld provide little benefit in reducing total latent health effects.
DISCUS $104 The calculations discussed above shcne that in the event of a core melt accident early evacuation of relatively small areas near a LWR SPd sheltering elsewhere would provide significant roouctions in individual risks of stochastic and non-stochastic health effects. The simple emergency response scheme suggested can be predetemined for specific in-plant mergency action levels appropriate for the General Emergency class. Because the actions are so simple and easily understood, there is an excellent chance the plan would work. If need be.
The early, inunediate evacuation area suggested by these calculations is the size of many low copulation zones around LWRs in the United States, and the early evacuation radius (2-3 miles) is less than the distance to the nearest population center Cf 25.000 persons or more, at most LWR sites in theU.S.(US79). Thus, there should be few impedi-ments to evacuation in these areas most of the time.
A further conclusion is that reduction in collective dose and latent cancers would be very difficult to achieve in the early (emergency) response phase. Sheltering during the passage of a predominantly nable gas re-lease would be efficacious, tnat only if sheltering to long distances were undertaken.
A few caveats to these conclusions are
, noteworthy: There are large wicertainties in the absolute values of the results of the calculations.
Nevertheless, the relative potential benefits of
,various evacuation / sheltering / relocation protective action strategies should be clear, especially whert large differences in results are obtained. .
At a few LWR sites in the U.S., and at many foreign sites, heavily populated areas entst in the l near vicinity. Witch could make early, feediate i evacuation difficult impractical or impossible, J For these sites, better shielding protection factors j cay pertain and smaller early evacuation distances I may be justified. On the other hand, some low l
population zones persist for many miles, and early. l insnediate evacuation of such areas may be a reason- I able objective for a core melt accident.
j It is acknowledged that entrapment situations can exist for some people or many people at some ]
time. Early. imriediate evacuation may be physically [
, impossible or extremely hazardous during a snow or I ice storm, for example. Special arrangements I should be made in emergency plans for identified persons in early evacuation zones who suffer from l significant impediments to mobility. In the event j of a core melt during a highly famebile situation. '
e.g., the ice storm, remaining in shelter and relocation from het spots (if a release occurs) as quickly as possible would be the only reasonable and practical alternatives. These are highly unlikely combinations of unlikely situations and ;
the suggested protective action schee should satisfy the basic radiation protection objectives l for most people, most of the time.
Finally, as compared to the suggested predetermined protective action plans, protective action decisions can be made on an ad hoc basis at any time. The distinction between preTeTermined actions and ad hoc actions is very important for emergency planners.
REFERENCE 5 8684 Satte11e Columbus Laboratories 1984.
- Radio- U578, U.S. % clear Regulatory Commiission,1978, nJclide Release thder LWR Accident Conditions.*
- Planning Sasts for the Development of State and 4 BM! 2104 Vols I-VI, July 1984 (Columbus OH Local Goverruent Radiological Emergency Response <
Plans in Support of Light Water % clear Pomer ;
43201). Plants *, NUREG-0396. December 1978 (Washington.
I sus 4 Surke R.P. . Helsing. C.D. . Aldrich. D.C.. D.C. 20555).
1984. *In plant considerations for off-site emergency response to reactor accidents." Health U579. U.S. % clear Regulatory Cosmission 1979, .
- Demographic Statistics Pertaining to Nuclear I Phystes J,r g . 46,763-773. Pomer Reactor 5ttes". MUREG-0348. October,1979 t f
In84 International Commission on Radiological (Washington. D.C. 20555).
1 Protection. " Protection of the Public in the Event U580. U.S. % clear Regulatory Commission.1980 of Major Radiation Accidents: Principals for
- Planning,* Report for Casuittee 4. Adopted by the " Criteria for Preparation and Evaluation of main Cosetssion May 1984. ICRP/84/C4-5/2. Radiological Emergency Respo#se Plans and Preparedness in Support of Nuclear Poner Plants *, ,
Ma74 Martin. Jr.. J. A., " Calculations of Doses, NUREG-0654. Rev 1. November 1980 (Washington, D.C.
Population Doses and Potential Health Effects No 20555).
to Atmospheric Releases of Radionuclides from U.S. % clear Poer Reactors During 1971*, U581. U.S. Nclear Regulatory Commission 1981.
' Technical Bests for Est1 eating Fission Product Radiation Cata~~ and Geoorts. ~34 309-319. June 1974. Behavior During LWR Accidents *, NUREG-0772. June Ma77 Martin. Jr., J.A.,1977.
- Doses while 1981(Weshington.D.C.20555).
traveling under mell estabitshed plumes.* Health Physics -Jr. , 32, 305-307. U582a, U.S. % clear Regulatory Cosmission.1982
- The Development of Severs Reactor Accident source Terms- 1957-1981*, NUREG-0773. November 1982 Ma80 Martin, Jr.. J.A.,1980,
- Perspectives on the ro's of radiological monitoring in an emergency.= (Washington, D.C. 20555).
Trans. Am.
~~~
Nuct . -Soc., 34, 727-729. U582b. U.S. Nclear Regulatory Cosmission,1982
' Technical Gutdance for $tting Criteria Develop-Ma82 Martin, Jr.. J.A.,1982,
- LWR acc1Jent ment
- MUREG/CR-2239, December 1982, prepared by spectrtsn. release characteristics and consequences.= Sandia National Laboratories (Washington. D.C.
in Proceedings of the nortshop on Meteorological aspects of energency response plans for naclear 20555).
power plants. NUREG/CP-0032, Aupst 1982, U.S. U583a, U.S. kclear Regulatory Commission.1983, Nuclear Regulatory Commisston. Washington, DC
- Calculations of Reactor Accident Consequences 20555. Verston 2, CRAC2: Computer Code-User's Guide".
NUREG/CR-2326. February 1983, prepared by U575 U.S.hciear Regulatory Cos=1ssion.1975 Sandia National Laboratories (Washington. D.C.
- Reactor safety Study *. IRJREG-75/014. Appendix VI.
October 1975 (Washington. DC 20555). 20555).
U577. U.S. mclear Regulatory Commission.1977, U5836. U.S. helear Regulatory Commitssion.1983
- 0verview of the Peactor Safety Study Consequence *CRAC Calculations for Accident Sections of Model*. NUREG-0340, October 1977 (Washington. Envirorenental Statements'. MUREC/CR-2901. March 1983, prepared by Sandia National Laboratories D.C.20555). (Washington, D.C. 20555).
L584. U.S. % clear Regulatory Commission.1984
- Dese Calculations for severe LWR Accident scenarios." MUREG-1062. May 1984 (Washington.
D.C.20555).
C
- '+
6._e
I 1
. l l
. J l
'able 1: Characteristics of Postulated ,
Severe Accident Scenarios '
Accident Type Release Characteristics 551 I ssT z ssT 3 Warning Tine Before 0.5 1.0 0.5 Release (hr)
ReleaseDuration(hr) 2 2 4 b
Radioguclide Invengory Fractica teleased to Group (EBe) the Atmosphere Xe.cr 13. 1.0 0.9 0 006 I 27. 0.45 0.003 24? d Cs-Ab 0.6 0.67 0.009 1 -51 Te.56 8. 0.64 0.03 251 Sa 5r 14 0.07 0.001 141 Ru 21. 0.05 0.002 241 Le 110. 0.009 0.0003 1 6h
- a. As defined in the Reactor Safety Study (U575).
- b. For a 1000 MW(e) the one. half hour after shutdown at end of core 11fe (3 gars) (U575). Table 2: Results asseing $$T1 and five
- c. 1 [84 = 1 Esa84
- 10 disintegrations /sec. Noble population distributions.
9asplusIactigityequals1.06billioncuries,
- d. 1( 5)
- 1 a 10' Early mean Mean Power Evacuation Numer of Number of Site Level Radius Early Early a Nueer (W-e) (miles) Fatalttles' Injurios b
1 1100 1.2. & 3 O 60 2 1100 1 50 200 2 20 100 3 4 %
3 800 1 40 200 2 4 %
3 0 M 4 650 1 20 130 2 C 70 Table 3: Tempore) and pathway contributions to 3 0 %
latent canceg fatalttles for severe 5 550 1 0 60 source terms
,I"8'y hp1 Percent Contributionb
" 'I N -3
- a. Normalized to 500 persons /sq mile witYn 10 miles.
No residents utth three miles - equivalent to b.
Esternal Cloud g g evacuation before a release.
Inhalation from Cloud 24 3 taternal Ground
((7 days) g3 16 taternal Ground
( > 7 days) 43 64 Inhalation of Resuspended Contamination 14 2 Ingestten of Contaminated Foods 5 10 gg pg
- a. From NUREG-0340 (U577).
- b. PWE.1 and PWR-2 are severe accident release categories from the Reactor Safety Study (U575).
The releases are of the order of $5T 1.
....C. e ,--$,
I i
LO -
0.s - 1 Relatwe 0.6 -
Does i Rate 0.4 -
t/r 0.2 -
if,2 0
O 0.5 1 2 3 4 5 Distance (Miles)
Figure 1. Relationship of dose rate anJ distance for a low level atmospheric release.
18, _____ _ 100 7 22: :- T NO EMERGENCY RESPONSE ggg)
I.P. 51TE
_ _ _ _ _ w 10'I.
I O-e.
s b '
1HR DE1AY g 10 MPH v
5
- g SHELTERING I
O '
E
- g2 g
I 10'*-
~
$HRDELAY 10 MPH o . No emergency response o . Susanary Evacuation. 55T1 within 10 et 1123 * (E) e . I hr delay.10 siph. '
25 Mt. Evac. Radius >
within 25 at 24 HR. nelocation Time 16'100 ' '
- 3 4
'10I 11 '10' . , ,10' 10' l's . Id y gygygg X.EARLY FATALITY DISTANCE (Mit Figure 2. Impact of a range of energency response asseptions Figure 3. sensittwity of early fata11ty distances to on Calculatetl early fatality probao111 ties. amergency response assumptions.
4 10' x \
3 SST1
- 800 M W-e M*h M
gal -
l!"l go 3 -
X 3
5 t8o 1e 8,2 -
a : >
8 .
1-Early Evacuation M'e1 3 Radoue (Mileel
=
2 10** .
.is...g . .....ng . . iii ng ....n.
i i s i s ing' i M' M Ma g3 g4 ye X, ACUTE FATAUTIES Figure 4. Conditional probabilities of vertous numbers of acute f atalities, assuming 55TI accident, early evacuatica of small areas, and a slow relocation from highly contaminated areas.
1.0 a 85 a t:
g 0.0 -
SST3 (F" a.
0.0 - SST2 l y .,
5: 0.4 -
i ;
5 SSTi ;
3 3
0.2 Uniform Population j 3 h! Distribution 2 m.
' I ' ' ' ' ' ' ' j sm um) . tateneatten O'
0 40 30 120 3 8" ' ' ' 3' 100 200 W # 7""3 DISTANCE IMILES)
I..
(0.8)
Figure 5.
Increase in calculated latent cancer fatalities with distance fo* three source tems. #" (0.21)
I33 (0.05) 4 ====. . %
e # 3 m e a Va&M4 0F tA8S -- i. ,teluJosual figurg 6. taltaisted esen latent Cancer fatalities and cost treteeffs f 8P leveral lateretttien date levels
($$Il 444140#1 at IAdits P9 tat $1te).
e
_ - - -,-