ML19269E990
| ML19269E990 | |
| Person / Time | |
|---|---|
| Issue date: | 10/31/1977 |
| From: | Mcgrath P, Wall I, Yaniv S NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES), SANDIA NATIONAL LABORATORIES |
| To: | |
| References | |
| NUREG-0340, NUREG-340, NUDOCS 7911130480 | |
| Download: ML19269E990 (48) | |
Text
NUREG-030 OVERVIEW OF THE REACTOR SAFETY STUDY CONSEO.UENCE MODEL Paper Presented at the International Conference of Nuclear Systems Reliability Engineering and Risk Assessment I
held at Gatlinburg,. Tennessee June 19 - 25,1977 2195 223 bk $kI1kkm U. S. Nuclear Regulatory Commission
$iG11130
()
Available from National Technical Infonnation Service Springfield, Virginia 22161 Price: Printed Copy
- Microfiche $3.00 The price of this document for requesters outside of the North American Continent can be obtained from the flational Technical Information Service.
2195 224
NUR EG-0340 OVERVIEW OF THE REACTOR SAFETY STUDY CONSEQUENCE MODEL Paper Presented at the international Conference of Nuclear Systems Reliability Engineering and Risk Assessment held at Gatlinburg, Tennessee June 19 - 25,1977
- 1. B. Wall P. E. McGrath*
S. S. Yaniv H. W. Church
- R. M. Blond J. R. Wayland
- Manu cript Completed. September 1977 Date Published: October 1977
'Sandia Laboratories Albuquerque, NM 87115 Probabilistic Analysis Branch Office of Nuclear Regulatory Research U. S. Nuclear Regulatory Commission 2195 22"3 Washington, D. C. 20555
FORE'dORD This report consists of the paper " Overview of the Reactor Safety Study Consequence Model," which was presented by the authors at the International Conference of Iluclear Systems Reliability En6ineering and Risk Assessment held at Gatlinburg, Tennessee on June 19 through June 25, 1977 2195 226 Table of Content.s Page No.
FOREWORD.
i 1
INTRODUCTION 1
2 DESCRIPTION AND FUNCTION OF CONSEQUENCE MODEL.
2 2.1 Accident Description.
3 2.2 Atmospheric Dispersion and Weather Datn.
5 2.3 Cloud Depletion and Ground Contamination 7
2.4 Air and Ground 'ancentrations of Radioactive Materials 7
2.5 Dosimetry 9
2.6 Population Distributions.
11 2.7 Evacuation.
13 2.8 Health Effects.
15 2.8.1 Early and Continuing Somatic Effects.
16 2.8.2 Late Somatic Effects 19 2.9 Property Damage 2P 2.9.1 Evacuation Costs 23 2.9.2 Interdiction Model 23 2.9.3 Decontamination Model.
24 3
DISCUSSION OF ACCIDENT CONSEQUENCES.
26 3.1 Early Fatalities.
26 3.2 Latent Cancer Fttalities.
30 33 Property Damage 36 3.4 Accident Risks.
36 REFERENCES.
40 2195 227 List of Tables Table Page No.
1 Summary of Release Categories Representing Hypothetical Accidents.
4 2
ilumber of Reactors Assigned to the Composite Sites 11 3
Construction of a Composite Reactor Site.
12 4
Cumulative Population Versus Distance for Composite Sites 14 5
Effect of Medical Therapy on the Bone Marrow Dose Mortality Criteria.
18 6
Dose-Effectiveness Factors.
21 7
Effective Incidence of Radiation-Induced Latent Cancer Fatalities per Million Man-Rem of Whole-Body Radiation 31 8
Contribution of Different Exposure Pathways to Latent Cancer Fatalities for a PWR-1 Release Category.
32 9
Contribution of Different Exposure Pathways to Latent Cancer Fatalities for a PWR-2 Release Category.
33 10 Consequences of Individual Release Categories.
38 2195 22B
-111-
List of Figures Figure Page No.
1 Schematic Outline of Consequence Model.
2 2
Estimated Dose-Response Curves for Mortality in 60 Days with Minimal Treatment (Curve A),
Supportive Treatment (Curve B), and Heroic Treatment (Curve C) 17 3
Illustrative Decontamination Model for Ground Level Release.
25 4
Relative Doses to Bone Marrow at 0.5 Miles From Reactor.
27 5
Conditional Probability of Early Death As a Function of Distance from Reactor for Three Effective Evacuation Speeds Given a PWR-1A Release 29 6
Conditional Probability of Latent Cancer Death Given 9 PWR-2 Release (Apprcximately, Absolute Mortality Probabilities are 10-D per Reactor Year Times Stated Ones) 35 4
219" 229 3
-iv-
OVERVIEW OF THE REACTOR SAFETY STUDY CONSEQUENCE MODEL Abstract.
This paper describes the calculation of potential nuclear reactor accident consequences as perfor=ed for the Reactor Safety Study, an Assessment of Accident Risks in U.S. Commercial Nuclear Power Plants (WASH-1400) [1].
The objective of the study was to realistically assess the risk to society from commercial nuclear power plant accidents.
The engineering analysis of the plants, which is described in detail in the Reactor Safety Study, provides an estimate of the probability versus magnitude of the release of radioactive material.
The consequence model, which is the subject of this paper, describes the postulated accident after the release of radioactive material from the containment.
Separate models trace the released radioactive material through the environment and assess its impact on man.
1.
Introduction.
The specific purpose of the Reactor Safety Study (WASH-1400) [1] was to realistically assess the risk to society frco potential accidents in commercial nuclear power plants.
To perform the assessment, event trees and fault trees were used to estimate the probability of release of radioactive material to the human environment.
To estimate the magnitude of the release, phenem-enological models were used to describe the liberation of radioactive materials from the fuel and their transport to the exterior of the reactor containment building.
The assessment of the accident after release of the radioactive material from the containment is performed with the consequence model which is the subject of this paper.
2195 230 1
The first section of this paper provides a description of the consequence calculations from breach of containment and release of radioactive material to predicted consequences (early fatalities, property damage, etc.).
The second section of this paper provides a discussion of the individual consequences.
The purpose is to indicate how the consequences depend on the many parameters of the model.
This discussion also shows the relative degree of uncertainties in the consequences to the modeling considerations and selection of input data.
For a complete description of the consequence mode. and computed resultc, the reader is referred to Appendix VI of WASH-1400 [1].
2.
Description and Function of Consecuence Model.
A schematic outline of data and models of the consequence calculation is shown in Figure 1.
The following general description gives an idea of the makeup WEATHER DATA
' I DESCRIPTICN ATMCSPEERIC OF RADICACTIVE
+
DMFWCN RELEASE HEALTH y
EFFECTS CLOUD DOSI."ETRY
+
PCPULATICN DEPLETICN PRCPERTY y
DAMGE CON TICN Figure 1.
SCHEMATIC OUTLINE OF CONSEOUEUCE MODEL 2 95 231 2
of the data, the manner in which it is used, the output of the various individual models, and a general flow of the accident progression (consequence calculations).
The results of the consequence model is a set of complementary cumulative distribution functions (ccdf)* for specific consequences (e.g., early fatalities, property da= age).
Basically, these ccdfs are generated by calculating the magnitude of each consequence for each combination of postulated accident release magnitude, weather and population, and then, after ranking these magnitudes, by plotting the aggregate probability of all combinations which exceed a specified magnitude versus that magnitude.
2.1.
Accident Description.
The initiating point of the conse-quence calculation is the specification of the postulated accident in terms of the quantity of radioactive material that could be released to the environment, amount of energy associated with the release, the duration of the release, time of release after accident initiation, warning time for evacuation, elevation of release, and probability of the accident occurrence.
The range of postulated radioactive releases is characterized by 9 PWR and 5 BWR release categories as stated in Table 1.
The data in this table represent the basic input to the consequence model and are generated by the engineering analysis of the PWR and BWR reactors as analyzed in the Reactor Safety Study.
xA ccdf is a plot of the probability of equalling or exceeding a specified value versus the specified "alue.
2195 232 3
Table 1.
SUIDfARY OF RELEASE CATEGORIES REPRESEllTIl1G llYPOTilETICAL ACCIDENTS Ilme of Duration Warning Time llevation Fraction of Core laventory Released {a)
I of Release 9I (neggy Release Delease Probatell s t y Release of Release for tvacuation
( s t egory (reactor-yr*I)
(hr)
(hr)
(Pr)
(meters)
(10 Stu/hr) te-Er 1(b)
CS-Rb Te-5b Sa-5r RuICI I8I ta I
P.tr I 9 a 10-II'I 2.5
- 0. 5 1.3 25 20 and 520 'I 0.9
- 0. /
0.4 0.4 0.05
- 0. 4 3 a 10'I Pea 2 8a 10~6 2.5 0.5 1.0 0
170 0.9
- 0. 7
- 0. 5 0.3 0.06 0.02 4 a 10-3 P.R 3 4a 10-6 5.0 1.5 2.0 0
6 0.8 0.2 0.2 0.3 0.02 0.03 3 a 10'3 P.it 4 5 a 10'I 2.0 3.0 2.0 0
1 0.6 0.09 0.04 0.03 5 a 10'3 3 a 10'3 4 a 10'8 P.R 5
- 7. a l0*I 2.0 4.0 1.0 0
0.3 G.3 0 03 9 a 10-3 5 a 10'3 1 a 10'3 6 a 10-4 7 a 10-5 PdR 6 6a 10-6 12.0 10.0 1.0 0
h/A L3 8 a 10*4 8 a 10'4 I a 10'3 9 a 10'I 7 a 10'S I a 10-5 P.R 7 4 a 10 5 10.0 10 0 1.0 0
N/A 6 a 10'3 2a 10-6 I a 10-6 2 a 10-5 Ia 10-6 1 a 10-6 2 a 10*I We 8 4a 10'S 0.5 0.5 h/A 'I O
h/A 2 a 10'I 1 a 10'8 5 a 10 1 a 10-6 g, 33 8 0
0 I
4 e
P.R 9 4 a 10-4 0.5 0.5 N/A 0
N/A 3 a 10-6 g, gg-7 6 a 10'I I a 104 I a 10'II O
O 6' h I Ia 10-6 2.0
- 0. 5 1.5 25 130 1.0 0.40 0.40 0.70 0.05
- 0. 5 5 a 10-3 8.h 2 6 a 10-6 30.0 3.0 2.0 0
30
- 1. 0 0.90 0.50 0.30 0.10 0.03 4 a 10'3 BdR 3 2 a 10^6 30.0 3.0 2.0 25 20 1.0 0.10 0.10 0.03 0.01 0.02 4 a 10'3 84R 4 2 a 10-6 5.0 2.0 2.0 25 N/A 0.6 8 a 10
5 a 10'I 4 a 10'3 6 a 10-4 6 a 10-4 I a 10-4 BdR 5 I a 10
3.5 5.0 N/A 150 N/A 5 a 10'"
6a 10*II 4 a 10-9 8 a 10'I3 8 a 10'I4 0
0 (a) Bac69ro.nd on the isotope groups and release mechanisms is presented in the Reactor Safety 5tudy. Appendia Vll.
(b) Organic todine is combined with elemental todines in the consequente calculations. Any error is negitgtble since its release fraction 6s relatively small fur all large release categories.
(c) includes Ru Rh Co. Mo. ic.
(d) Includes V. t e, f r. Nb. Co. Pr. hd. hp. Pe. Am. (m.
(e) Acc tdent -
within PWR I category have two distinct energy releases that af fecg consequences. PWR 1 category o PWR I A alth a probability of 4 a 1 *I is subdivided 6 per reactor-year and 20 a 10 8tu/hr and PWR 18 with a probability of a a 10'i per reac tor-year and 520 a 1 Stu-hr.
N (f) Nct applicat,le.
s (g) A 10 meter elevation 65 used ta place of zero representing the imid-point of a potential containment break. Any tv. pact on the results would be slight end conservative.
(D Nuu
2.2.
Atmospheric Dispersion and Weather Data.
The atmospheric dispersion of the released radioactive material is estimated with a Gaussian dispercion model.
The model is used to calculate ground level air concentrations, and subsequently ground concentrations of radioactive material to great distances from the reactor.
The Gaussian model is utilized in a fashion different from usual applications in that the model includes specification of thermal stability, wind speed, and precipitation occurrence on an hourly basis.
The calculated plume may change in characteristics every hour of travel time.
The model assumes no temporal variations in wind direction.
This simplify-ing assumption is justified on the grounds that the spatial representa-tion of population is relatively simple
- and that only single-station meteoro1.ogy is availabic at most reactor sites.
By propagating the unidirectional plume in all directions from the reactor, each geographi-cal 1ccation is potentially exposed.**
The weather data have been collected from six sites which were judged to be representative of all reactor sites with respect to variability of climatic or topographic features.
Each of the sites has one year of complete hourly recorded data, i.e.,
there were 8,760 read-of thermal stability, wind speed and precipitation occurrence.
n
- Representation of the population is simple in the sense that the data do not account for diurnal movements and that the population is assumed uniform within a spatial mesh having an azimuthal resolution of 22-1/2%.
- For an individual site, the relative probabilities of the wind blowing in each direction are represented by the wind rose.
In the calcula-tions for the 100 reactors at 68 sites, the consequence model utilized a uniform probability distribution of wind direction.
This simpli-fying assumption is justified for the calculation of aggregate risk from many reactor sites [2].
2195 234 s
Since the atmospheric dispersion of the radioactive material depends on the weather over a period of many hours, and since there is a large (almost infinite) number of combinations of hourly weather sequences, the year's worth of data is sampled in such a way that the true frequency distribution of accident consequences would be closely approximated.
This simulation is performed by systematically selecting (stratified sampling) during a year the time at which a postulated accident might be initiated.
The atmospheric dispersion of the radio-active material, therefore, is described by utilizing the hourly sequences of weather data following the selected accident starting time.
A 4-day sampling interval was found to cover the predominant weather cycles and to provide an acceptably small variance on the computed consequences.
Diurnal cycles are accommodated by a 13-hour shift in the starting hour.
This procedure of sampling from actual weather data allows one to calculate more realistically the dependence of the computed consequences on the type and sequence of weather conditions.
It provides a more realistic estimate of the probability distribution function of con-sequences than the commonly used procedure of as-ociating the conse-quence estimates from a time-invariant atmospFaric dispersion model with the frequency distribution of weather conditions [3].
For thermally hot plumes the centerline is modified to account for buoyant plume rise.
The vertical growth of the plume is constrained to the space under the mixing-layer depth, i.e.,
no penetration of the mixing-layer is presumed.
2195 235 6
2.3.
Cloud Deoletion and Ground Contamination.
As the plume of radioactive material travels outward from the reactor, competing mechanisms remove the airborne material.
In addition to radioactive decay, the radioactive material is removed by deposition processes, e.g.,
impaction on obstacles (dry deposition), and by precipitation scavenging (wet deposition).
These deposition mechanisms cannot be specified precisely.
There are significant dependencies of removal rates on, among other things, precipitation type and rate, particle density and size distribution, surface characteristics of the ground, and weather conditions.
For staplicity, the dry deposition velocity (ratio of the deposition flux to the air concentration at a particular distance frcm the surface) is assumed to be constant and characteristic of 1 micron diamecer aerosol which is thought to be a conservative selection.
Wet deposition occurs simultaneously with dry deposition when precipitation occurs.
Wet depositio is modeled by a sicple exponential removal rate.
When precipitation occurrence is specified by the weather data, it is assumed to occur uniformly within time and throughout the spatial interval.
The removal rate is a function of the thermal stability.
The noble gases are assumed to ha insoluble and nonreactive, and therefore are not removed by either dry or wet deposition.
The ground concentration is calculated from the air concentration and the deposition rate.
The material deposited on the ground is subtracted from the airborne material.
2.4.
Air and Ground Concentrations of.sadioactive Materials.
The previous discussion describes how the consequence model calculates the air and ground concentrations of radioactive materials at various 2195 236 7
distances dow. wind after a postulated accident.
It is useful to reiterate briefly on the calculational flow which leads to estimates of air and ground contamination without some of the confusing mechanis-tic details.
For a postulated accident, the consequence mudel selects a starting hour for the accident from one year of weather data.
The dispersion model utilizes the weather data given at the starting time of the accident to describe the plume development for the first hour.
The distance of travel from the reactor within that hour is determined by the wind speed.
The plume development is changed each hour after the release to reflect the changing wind speed, thermal stability, mixing-layer height, and precipitation occurrence.
In an outward direction from the reactor, the air and ground concentrations of radioactive material are specified on a fixed polar grid.
Air and ground concentrations are assumed uniform over the interval and they reflect atmospheric conditions existing at the center point of the interval.
The released radioactive material is followed in the above manner, hour after hour, until the plume becomes essentially depleted or until 500 miles is reached.
If there is any remaining airborne material (except noble gases) at the end of the calculation, it is deposited uniformly in an interval of 500 to 2,000 miles within which a uniform population density of 78 people per square mile is ass <'ed.
Therefore, from the one selected accident starting time, spatia' distributions of air and ground concentrations of the released radio-active material are estimated.
By selecting another starting time, another spatial distribution of air and ground concentrations is estimated for the same postulated release.
This procedure is repeated 2195 237 8
often enough to approximate closely a frequency distribution of air and ground contamination isopleths following a postulated accident.
2.5.
Dosimetry.
The potential radiation dose to individuals and to populations is calculated from the previously described air and ground concentrations of radioactive material by using suitable dosi-metric models.
For this purpose, it is convenient to categorize the exposure pathways as those associated with the passing cloud and those associated with ground contamination.
The airborne radioactive material results in radiation doses through external radiation from the plume and radiation from inhaled radioactive material.
To receive the external radiation, the individual must either be immersed in the plume or in its general vicinity.
The model relates the air concentratfon of radioactive material to an external dose to various body organs, e.g.,
bone marrow, GI-tract.
The radiation dose from inhaled radioactive material is propor-tional to the exposure to airborne concentration of radionuclides at roughly 2 meters above the ground, and to the individual's breathing rate.
The dosimetric model describes the time dependent movement of the radioactive material within the body.
The model is essentially the ICRP Task Group Lung Model [4], modified to calculate a single short term exposure as opposed to long term inhalation and with some changes in the model parameters to reflect newer data.
The deposition site of the inhaled material within the respiratory tract depends on the aerosol size (assumed to be 1 micron).
The movement of a specific radionuclide after its depositicn is determined by its chemical form and particle size and shape.
Particle size is the major factor in determining the fraction of inhaled material retained in the deep lung.
Solubility determines the retention time of the particles in 9
910r 970 Li/J L.J J
the deep lung.
The dose from inhaled radioactive material is calculated over varying time periods up to 50 years.
The radioactive material deposited on the ground results in adiation doses to individuals through basically three pathways.
The radiation dose from direct external irradiation due to deposited material and from inhalation of resuspended radioactive material starts immediately upon depositing on tDe ground.
The third pathway is the ingestion of deposited radioactive material through food and water.
The ingestion of the radioactive material is a result of its depositing directly on vegetation which is consumed by man or by animals furnishing food for man.
The more indirect ingestion pathways involve the uptake of ground deposited radioactive material through the roots of vegetation.
The dosimetric models for ingested radio-active =aterial are similar to the models for inhaled material except that material is directly deposited in the gastrointestinal tract.
The quantity of ingested radioactive material is calculated in the ccusea'.. ace model with a relatively simple environmental model which considers the soil-grass-cow-milk pathway and an "other" pathway in which all other possible pathways are combined.
The soil-milk pathway model was derived from the large amount of experimental data that exists.
The results of this model, which agreed well with field results, were correlated with the many years of data collected from the nuclear weapon fallout studies.
This correlation yielded results for other possible environmental pathways to man.
The availability of these ingestion exposure models was determined from specific data on the agricultural characteristics for each reactor site.
2195 239 10
The radioactive material deposited on the ground is sur'ict to weathering.
The material may become bound to large soil particles and thereby be less susceptible to resuspension or to uptake by vegetation.
In addition, the material may be leached downward below root zones.
The downward movement also provides additiv.7al shielding to man from emitted penetrating radiation.
Therefore, the dose rate to man will decrease faster with time than that accountable by radioactive decay, 2.6.
Pooulation Distributions.
The population distributions within 16 radial sections around each of the 68 reactor sites on which the firct 100 LWRs are located were obtained from 1970 U.S. Census Bureau data.
In order to reduce the number of calculations necessary to estimate risk for 100 re ctors, each reactor site was assigned to one of six composite sites on the basis of comparable meteorology.
The makeup of these six composite sites is stated in Table 2.
Table 2.
NUMBER OF REACTORS ASSIGNED TO THE COMPOSITE SITES Composite Site Number of Number of Reactors Characteristics Sites BWR PWR Atlantic coastal site 10 5
9 Large river valley in northeast 10 6
8 Grec. Lakes shore 4
3 2
Southeast river valley influenced by Bermuda High 17 7
23 Central midwest plain 23 13 18 Pacific coastal site 4
0 6
68 34 66 2195 240 n
The population sectors for each composite site were generated from those of the assigned sites in the following manner which was designed to correctly represent both the average and the peak population sectors; the first site listed in Table 2 is discussed as an example.
Fourteen reactors at 10 actual sites are assigned to this composite site.
The actual population around each of these 14 reactors was described by sixteen 22.5* sectors.
These 224 sectors were then ranked from highest to lowest population based on the cumulative population within 50 miles of the reactor.
These 224 ranked population sectors were used to generate 16 representative population sectors in the manner indicated in Table 3.
For instance, the highest and second highest Table 3.
CONSTRUCTION OF A COMPOSITE REACTOR SITE Sector From Conditional Probability Sector of Ranked of Sector Being Composite Site Listing Exposed 1
1 1/224 2
2 1/224 3
3,4(a) 2/224 4
5,6(8) 2/224 5
Average of next 6 6/224 6
Average of next 6 6/224 7
Average of next 12 12/224 8
Average of next 22 22/224 9
Average of next 22 22/224 10 Average of next 23 23/224 11 Average of next 22 22/224 12 Average of next 22 22/224 13 Average of next 20 20/224 14 Average of next 20 20/224 15 Average of next 21 21/224 16 Average of next 22 22/224
(")Two reactors at one site.
2195 241 1,
ranked sectors (of the 224) were assigned to sectors 1 and 2, respec-tively, of the composite site.
The third sector of the composite site had a radial population distribution that is the average of the popula-tion distributions in the third and fourth sectors of the 224.
The population distributions for all 6 composite sites resulting from the above procedure are characterized in Table 4.
The most significant assumption implied by the above procedure is a uniform frequency distribution for wind direction.
A wind rose for a typical individual site shows a maximum difference in frequency of' about a factor of three.
When assessing the risk for 100 reactors, this variation is small compared to other modeling approximations [2].
Since airborne and ground concentrations of radioactive material and hence radiation doses are assumed to be uniform within a given annular section of a sector, the radiation dose to an individual or to a population is determined simply by their assigned annular sector.
2.7.
Evacuation.
Population evacuation was incorporated in the consequence model.
Evacuation should be distinguished from relocation.
" Evacuation" denotes an expeditious movement of people to avoid exposure to the passing cloud.
" Relocation" denotes a post-accident response to reduce exposure from long term ground contamination.
The evacuation model is basem on an analysis of evacuation data assembled by the U.S. Environmental Protection Agency [5].
This statistical analysis, which is reported in Appendix VI of WASH-1400 [1],
found that evacuations occur as a uniform mass of people all moving with the same speed and that the effective evacuation speed is log-normally distributed.
The effective evacuation speed is defined as the total dista. ice travelled divided by time since warning.
The 13
f f
f f
f 7
00 00 00 00 00 00 11 11 11 11 11 11 0
xx xx xx xx xx xx 0
n 5
53 06 01 18 91 92 r
i o
h 42 88 81 61 65 11 t
t c
i S
e w
E f
f f
f f
f s
T 00 00 00 00 00 00 n
I 11 11 11 11 11 11 a
o S
i 0
xx xx xx xx xx xx f
t E
0 o
a T
2 66 96 19 15 85 17 l
I y
u S
11 38 91 19 14 16 t
p O
i o
P l
p M
f f
5 f
6 6
i O
00 00 00 00 00 00 b
e C
11 11 11 11 11 11 a
v b
i R
0 xx xx xx xx xx xx o
t O
5 r
a F
81 31 67 54 43 87 p
lu E
12 58 94 77 23 11 l
m C
a u
N n
c A
f f
f f
f f
o T
00 00 00 00 00 00 i
y S
11 11 11 11 11 11 t
b I
i D
0 xx xx xx xx xx xx d
d 2
n e
S 40 59 95 51 22 21 o
n U
c i
S 11 49 13 12 21 15 f
R e
e E
h d
V 3
f
- j 3
3 f,
t 00 00 00 00 00 00 s
N 11 11 11 11 11 11 y
a O
b I
0 xx xx xx xx xx xx r
T 1
d o
A 76 21 75 74 16 20 e
t L
t c
U 37 81 37 25 51 22 h
e P
g s
O i
P j
- j 3
f 3
3 e
k 00 00 0
o0 00 00 w
a E
11 11 1
l1 11 11 e
V s
p I
5 xx xx x0 xx xx xx r)
T o3 n A
81 11 7
23 85 49 t
i L
ce U
92 23 2
34 91 21 el M
sb d U
a e C
))
))
))
))
))
))
6T t ab ab ab ab ab ab 1
h
((
((
((
((
((
((
e g ne i 4
d i s e y
e h( w e
e ch t
n l
l l
e ng id u b
a l
r rei t
l we a
t at o
eui s
a s
l T
s vs h vl e
t no n a
a s
ifa w
s op o o
re rnd d
a ix is c
eh e
iu i
o te te vt k t
m m
c a
al c
ir a
syr l g li i
ro L
aee l n c
un um t
n elB ai i
pi p n
e t
hl ra f
oe o0 a
gn a
t ay tl i
Pb P5 l
ri e
uvb np c
)
)
t a
r o
e a
b N4U
=
A L
G S
C P
(
(
Nn L
distributions are very skewed towards very low speeds.
In the conse-quence model, t~1e distribution is approximated by three discrete speeds.
The population within 25 miles is assumed to move radially outward from the reactor with a 307. probability of having an effective speed of 0.0 mph, a 40% probability of 1.2 mph, and a 30% probability of 7.0 mph.
The time available for each individual's evacuation is the sum of the warning time and the time required for the radioactive plume to reach the individual.
It is assumed in the evacuation model that the population movement is always radially outward from the reactor, i.e.,
there is no cross-wind movement.
During the evacuation, the population is assumed to be unshielded from exposure to airborne radioactive material both exter-nally and through inhalation and to be shleided from exposure to ground contamination due to surface roughness and the automobile.
If the vacuating population is overtaken by the cloud of radioactive material, it is assumed that people will have moved outside of the contaminated area within a 4-hour period.
Beyond 25 miles, the people are assumed to be relocated within 7 days if the chronic dose due to exposure to ground deposited radio-nuclides exceeds a specified value.
However, if the dose accumulated within the first 7 days due to exposure to contaminated ground exceeds 200 rads, then the people are assumed to be relocated within 1 day.
2.8.
Health Effects.
Three categories of potentiel health effects are calculated:
early and continuing somatic effects, late somatic effects (cancers), and genetic effects.
Early and continuing somatic effects manifest themselves within a year of exposure.
In contrast, latent cancers would be observed 2 to 40 years after exposure 2195 244
and genetic effects in succeeding generations.
Late somatic and genetic effects stemming from a major release of radioactive material would manifest themselves as an increase in the spont:neous incidence of cancer or genetic effects in the exposed population.
Since early and continuing somatic effects are usually observed after large, acute doses of radiation (e.g., whole-body doses of 100 rads), they would be limited to persons within 50 miles or so of the reactor even for the largest conceivable release.
Conversely, late somatic and genetic effects may result from very low doses albeit with low incidence.
Consequently, these effects may occur at long distances. rom the reactor.
2.8.1.
Early and Continuing Somatic Effects.
Early and continuing mortalities may result frca radiation damage to the bone marrow, lung or gastrointestinal tract with radiation haage to the bone marrow being the most important contributor.
Under the conditions.:ikely to exist as a result of reactor accidents, radiation damage to the lung or to the gastrointestinal tract is not likely to be lethal unless accompanied by bone marrow damage.
The medical advisors to the Reactor Safety Study proposed three bone marrow dose-mortality criteria [6], t;pending on the degree of medical therapy given to the exposed individual.
These curves are reproduced in Figure 2 and are denoted by A, B and C for minimal, supportive, and heroic therapy, respectively.
Mortality criteria are of ten stated in terms cf the dose that would be lethal to 50% of the exposed population within 60 dcys (denoted by LD50/60).
In Figure 2, the LD may be read on the abscissa opposite the 507. value on the 50/60 ordinate.
2195 Pa5 16
99.99 99.8 99 98 95 o
90 ae E
80 A
B C
E 70 E
60 50
~
40 d
30 5
20 o2 10 E
5 wo 5
2 a
1
- 0. 5 f
- 0. 2 0.05
'J
'I 0.01 0
200 400 600 800 1000 1200 1400 DOSE (RADS)
Figure 2.
ESTIMATED DOSE-RESPONSE CURVES FOR MORTALITY IN 60 DAYS WITH MINIMAL TREATMENT (CURVE A),
SUPPORTIVE TREATMENT (CURVE B), AND HEROIC TREATMENT (CURVE C) nA /
2193
/e n
The Reactor Safety Study judged that, in the event of a serious reactor accident in the United States, the Federal and State govern-ments and the utility involved would mobilize medical resources through-out the nation to aid the exposed population; a maj or constraint would be the availability of specialized resources.
The type of medical therapy, the estimated value of LD and the number of people who 50/60 could potentially be treated are summarized in Table 5.
Table 5.
EFFECT OF MEDICAL THERAPY ON THE BONE MARROW DOSE MORTALITJ CRITERIA Medical Therapy LD rads No. of People 50/60 Minimal None 340 (large number)
Supportive Barrier nursing, copius anti-510 2500-5000 biotics, transfusions of whole blood, packed cells and plate-lets.
Start within 20 days Heroic Supportive treatment plus bone 1050 50-150 marrow transplantation.
Start withit 10 days The critical dose range for supportive treatment is 350 to 550 rads to the bone marrow.
In the event of the worst calculated accident (a probability of about 10~9 per reactor-year), the expected number of people receiving such a dose would be about 5,000.
For less severe accidents, the number would be much smaller.
Since this number is consistent with the constraint on supportive treatment, the number of early fatalities is estimated on the basis of curve B in Figure 2.
2195 247 18
The study also c;timated the number of prenatal deaths and cases of early morbidities including hypothyroidism, temporary sterility, congenital malformations, growth retardations, cataracts and prodromal vomiting [1].
2.8.2.
Late Somatic Effects.
Late somatic effects are limited to latent cancer fatalities plus benign and cancereus thyroid nodules.
The following discussion is limited to latent cancer fatalities since the methodology for all these effects is similar and it is the most significant effect.
Following the irradiation of a large number of people, there is a latent period during which no increase in cancer incidence is detec-table. After this period, the radiation-induced cancers appear at an approximately uniform rate for a period of years, which is termed the plateau.
The risk of latent cancers is normally stated either in terms of the incidence rate during the plateau period (cases per million exposed population per year per rem) or in terms of the expected number of cases (cases per million man-rem).
The latter value is merely the integral under the curve, or the incidence rate times the plateau period.
As a starting point, the study uses the estimates stated in a report issued by the National Academy of Sciences on the Biological Effects of Icnizing Radiation (the BEIR report) [7].
The BEIR report estimates risks on both.an absolute and relative basis and by using 30-year and lifetime plateaus.
For the reasons given in Appendix 'll of WASH-1400 [1], the study accepts the absolute basis and a 30-year plateau as being more appropriate for the evaluation of reactor risks.
2195 ?^8 19
The BEIR report relied heavily on the ongoing study of the Japanese atomic bomb survivors, who received very high dose rate exposure of gamma, beta and neutron (high-LET)* irradiation.
Further-more, the dose magnitudes were estimated to range from 10 to over 300 rem.
Those survivors receiving less than 10 rem were used as a control population group for the BEIR estimates.
The doses from a reactor accident would be almost exclusively due to low-LET radiation (i.e.,
no neutrons and less than 17. due to alpha radiation).
Except for a few individuals who might be irradiated by the passing cloud very close to the reactor, the dose rates to the whole body would be less than 1 rem per day which, with respect to latent cancer induction, is a low dose rate.
For all these reasons, the exposures resulting from a reactor accident would be different from the exposures on which the BEIR report bases its estimates with respect to quality of radiation, dose rate and dose magnitude.
The risk estimates generated in the BEIR report are based on a linear extrapolation from the aforementioned data to ze: o doses without any threshold dose (i.e., a dose magnitude below which r.here would be zero induction of cancer).
Both the BEIR and United Nations [8] reports caution that this linear hypothesis is likely to overestimate the risks for low doses and/or low dose rates of low-LET radiation and that, in cases of low exposure, it cannot be ruled out that the risk may actually be zero, k
Linear energy transfer (LET) is a measure of the rate of energy loss along the track of an ionizing particle.
High-LET radiation includes alpha particles and neutrons.
2195 249 20
Since the objective of the Reactor Safety Study is to make as realistic an assessment of risks as is possible and to place bounds on the uncertainty, the study makes three estimates of the number of latent cancers from a reactor accident.
The upper bound is based on the BEIR estimates with some small changes reflecting recent data.
For the central (most realistic) estimate, the upper bound is modified by dose-effectiveness factors which are stated in Table 6.
Table 6.
DOSE-EFFECTIVENESS FACTORS Total Dose Dose Rate (rem ner day)
(rem)
<1 1-10
> 10
< 10 0.2 0.2 0.2 10 - 25 0.2 0.4 0.4 25 - 300 0.2 0.4 1.0 These factors, which are based on recent experimental data for animals, reduce the expected incidence of latent cancers for small doses and/cr low dose rates.
In the opinion of the study, these central estimates would represent a more realistic assessment of latent cancer fatalities arising from a reactor accident; the study's medical advisors were of the unanimous opinion that these dose effectiveness factors still probably overestimate the true risk.
The overall pattern of data shows no observable difference from an unirradiated control population for persons receiving either an acute dose of less than 25 rem or a chronic dose of less than 1 rem per day to the whole body.
As an approximate indication of a possible nonzero lower bound, the study estimates the population dose received by individuals in excess of a threshold and applies the incidence rate used for the central estimate.
219b 2r.30 21
The BEIR report estimates the incidence of radiation-induced latent cancer fatalities for individual organs and summarizes the overall effect in terms of whole-body radiation.
The latter approach was appropriate since the BEIR report was primarily concerned with external radiation to the whole body.
In the event of a reactor accident, inhalation of radioactive material from the passing cloud may result in a nonuniform dose distribution in the body; certain organs (e.g.,
lung) may receive much higher doses than others.
In order to accommodate this nonuniform dose distribution, the doses and the expected radiogenic latent cancer deaths are calculated for indi-vidual organs and su==ed to determine the overall risk.
For reference purposes, the whole-body values are also calculated.
There is a detailed discussion in Appendix VI of WASH-1400 of the changes to the risk coefficients and plateaus recommended in the BEIR report; the changes are very small.
The impacts of the dose effective-ness factors and of the organ-by-organ dose calculation will be dis-cussed in Section 3.2.
2.9.
Property Damage.
Property damage following a postulated reactor accident is not of the same nature as that esulting from other potential catastrophic events, i.e.,
there is no physical damage to the property offsite.
The property damage arises as a result of contamination by radioactive material and the possible radiation dose that could be received if the property were utilized in its intended manner.
The restriction in the property use results in economic loss.
The components of property damage, as assumed and modeled in the consequence model, are evacuation costs, loss of agricultural products, decontamination costs and population relocation costs.
2195 251 22
2.9.1.
Evacuation Costs.
The evacuation model designed for low probability high consequence accidents assumed that all persons within a 5-mile circumference of the reactor and within a 45' arc out to 25 miles centered on the prevailing wind direction at the time of the accident would participate in evacuation.
An analysis of EPA evacuation data [5] provided an estimate of the costs, corrected for inflation, of evacuating large numbers of persons and ft.nishing or paying for temporary food and shelter.
2.9.2.
Interdiction Model.
The radioactive contamination of a large area may. result in the contamination of milk produced by cattle grazing on contaminated pastures, in the external contamination of crops and/or excessive radiation doses to man.
In such events, the milk and crops may be impounded and/or the people relocated for a period of time.
All of these actions are termed " interdiction."
The interdiction model is based upon maximum acceptable doses in the unlikely event of a reactor accident.
The dose criteria used by the Reactor Safety Study are based upon the recommendations of the U.S. Federal Radiation Council [9] and the British Medical Research Council [10].
The dose criteria are translated into corresponding contamination 2
levels (Ci/m ) of different radionuclides by the dosimetric models described in Section 2.5 and an environmental model which incorporates the grass-cow-man or soil-root-crop-man pathways.
The milk interdiction level is the most restrictive so that the area over which milk would be impounded would be the largest.
Conversely, the people interdiction level is the least restrictive so that the area from which people would be relocated would be the smallest.
The " weathering" of deposited 2195 252 23
radionuclides is properly incorporated so that as the cumuiative lifetime dose decreases, the interdiction distance slowly moves towards the reactor.
The cost of agricultural losses is calculated on the basis of current prices and average land fractions devoted to agricultural use.
Relocation costs are composed of the loss of income during the period of relocati.on (90 days), moving costs (to a distance of 100 miles),
and the econcaic loss of the physical property.
Property value is amortized over 30 years. For relocation periods greater than 10 years, the total value of the property is assessed.
2.9.3.
Decontamination Model.
Decontamination is defined as the cleanup and removal of radionuclides.
The possible decontamination methods include physical removal of the radionuclides, stabilization of the radionuclides in place, and environment management.
The particu-lar procedure utilized in a given case would depend on many factors, including (1) the type of surface contaminated, (2) the external environment to which the surface is exposed, (3) the possible hazards to man, (4) the costs involved, (5) the degree of decontamination required, and (6) the consequences of the decontamination operation.
A measure of effectiveness of decontamination operations is the decontamination factor DF, which is defined as the original contaminant concentration (in curies per square meter) divided by the contaminant concentration after decontamination.
The decontamination model is illustrated in Figure 3.
Without decontamination, the interdiction criterion (L) translates to a distance R.
After a maximum decontamination factor of 20, the land area t
between d and R will beccme available for reoccupation.
Subsequent i
2 2195 253 aa
weathering of the radionuclides will reopen the land area between R2 and R.
In fact, the model is slightly more sophisticated since two 3
decor.tamination factors of 2 and 20 are incorporated; the decontamina-tion costs for the former are significantly lower than for the latter.
These costs were estimated on a per capita basis so that for the population within a spatial interval requiring decontamination, the total decontamination costs could be easily calculated in the conse-quence model, a
=]
f-GROUND CONTAMINATION AT t = 0 a
AFTER DECONTAMINATION AT t = n YEARS (RADI0 ACTIVE S
DECAY AND WEATHERING FORCES) o e
5
\\
DF o
w s
MAX s
._ _ _ _ _ _'_ g _ N,
8 LAND INTERDICTION LEVEL L s
I I
i
~~
I l
I s
1 3
I a
R R
3 2
i DISTANCE FROM REACTOR Figure 3.
ILLUSTRATIVE DECONTAMI"ATION MODEL FOR GROUND LEVEL RELEASE 2195 254 as
3 Discuccion of Accident Concequences.
Thic section of the paper provideo a discuccion of the individual ectimated consequences, i.e.,
"arly fatalitiel, latent cancer fatalitiec, and property damage.
The purpone of thic cection ic to provide the reader with an indication of how tneae individual concequences depend on the various models in the concequence model and their input data.
3.3.
Early Fatalities.
Before exploring 9 sensitive parametera, it in instructive to consider the cauce of early deaths and the princi-ple radionucliden involved.
Doces to three organc are considered fe" t h *: calculation of early fatalities.
On the basis of clinical and experimental data described in the Reactor farety ftudy, it is evident that the dominant twchanism for early death in the radiation damage to the bone marrow.
Figure 4, which is for a BWR 1 release, shows the relative deaec at 0.5 miles from the reactor to the bone marrow from the maior expocure modea associated with the passing cloud and the rolative contributionr from the different radionuclidec to the doses.
It is evident in Figure 4 that for thic release and distance the external doce from the pasaing cloud, the external doce from the contaminated ground within 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br />, and the dose from inhaled radio-nuclider concribute equally to the overall bone marrow dose.
The principle radionuclides resulting in bone marrow dose are Te-132, I-132, I-133, I-135, 3r-89 and Ba-140.*
The calculation of early fatalitiec is one of the most sencitive results of the consequence model.
The reason for this sensitivity is due to a number of important factors.
First, clearly the number of early fatalities la sensitive to the accumed value for ID 50/60*
- The dose contributions are for radionuclides either in the pascing cloud, deposited on the ground or inhaled and include an3 contribu-tions from their radionctive daughtars.
26 2195 255
10 i
CUMULATIVE
~
EXTERNAL DOSE FROM DOSE FROM 3
/ PASSING CLOUD INHALED 10
[
EXTERNAL DOS $ FROM RADIONUCLIDES :
/
'/
/_
GROUND WITHl,N 4 HR.
l'I35
/
ft-1)?
La,J 1 13) m Sr-N Q
E r 'JI C
B+-10 r.m g
, i33 W 10
'f:,
- s tu
[
",,.,,3
- ' -ili
t
$ 3,;3 E
I' l > 4.,
-i
~'
u:
an Att Of 44
- 4DIC'.LCLINS k'
1 10 j
.u cre, pa HC*.'JCLIM S
~
"(,%
ALL CTQR R ADIC'.L:CilMS m
y
_ g y
l i
0 __
L 10 0
50 100 150 200 250 300 DAYS AFTER ACCIDENT F_irure 4.
RELATIVE DOSES TO BONE MARROW AT f
0.5 MILES FROM REACTOR 2195 256 27
Early mortality from radiation is a threshold effect.
If an individual receives less than the threshold dose, he may be very sick and have a high probability of latent cancer but he will not be classified as an early death.
As a measure of this sensitivity, a reduction in LD50/60 from 510 to 340 rads would increase the expected number of early fatalities by a factor of 3 or 4 depending upon circumstances.
Second, the natural forces in the environment are constantly at work to disperse and dilute the released radioactive material.
Even for very large releases of radioactive material, there is only a finite distance over which the material will remain sufficiently concentrated to result in a large dose.
The results of the calculations with the consequence model demonstrated that early fatalities were generally limited to within 10 to 15 miles from the reactor even for the largest release magnitude.
This fact is illustrated in Figure 5 where the probability of early death given a PWR-1A release category (defined in Table 1) is plotted against distance from the reactor.
By combining this graph with the probability of a PWR-1A release category and a given population distribution, one can estimate the number of early fatalities.
Third, it is evident in Figure 5 that the early fatalities are sensitive to the effective evacuation speed.
Finally, the ccncentration of radioactive material deposited on the ground within 20 miles of the reactor is influenced by the initial behavior of the cloud of radioactive material.
If there is sufficient thermal energy associated with the release, the cloud will be buoyant and will nat intersect the ground until atmospheric dispersion has spread it.ignificantly.
The q ttity of thermal energy will determine the initial
.^vation that the cloud will reach.
Under the same atmospheric conditions, the higher the initial elevation of the cloud 2195 257 23
I=
i i
i i
i i
i i
~I 10 3
3
}s INEFFECTlVE EVACUATION (O MPH)
_~
D
\\
/
10'2r\\
2 w
s m
g o
s R
A.
u 10'
-s 2
r s
m g
o s
cc c.
s EFFECTIVE ' 's 10 e
EVACUATION 5
0 SPEED 1. 2 MPH
\\
c
\\
..\\,
o
-5
\\
10 g
g 1
i
-6 i
t i
e i
i t
i 10 0
- 2. 5
- 5. 0
- 7. 5 10 12.5 15 17.5 20 22.5 MILES FROM REACTOR Fi re 5.
CONDITIOiAL PROBAB,ILITY OF E.iF.LY. DEATH AS.A-m m...
v oin.4C 3,o.,...t~,n,a
-m
. ; s t.
. d:a r avTION 0:-
1:: st v
- v r.
.aasn EVACUATION MEDS CIVE"T~G2-1A IGLEAfE (a)
Approximately, absolute mortality probabilities are 10-6 per reactor vear times quoted values (b)
The error bars denote the variatien in the mean values for the six meteorological data sets (c)
For effective evacuation speeds of l. 7 and 7 mph, the conditional probability of early death is effectively zero within 25 miles.
2195 253 29
the greater the distance frcm the reactor that the cloud will intersect the ground.
When the cloud intersects the ground at increasing distance, the air concentration at ground level will be decreasingly smaller.
It is for this reason that, generally, thermal energy of the release
..as an important effect en the calculation of early fatalities.
The exception occurs when radioactive caterial is deposited on the ground by precipitation.
Since precipitation scavenging removes airborne material from all parts of the cloud, the important parameter is the total quantity of airborne caterial and not the grcund level air concentratien.
For this reason, the maximum n: nber of early fatalities is fairly independent of release energy (see Figure VI 13-25 of Appen-dix VI of WASH-1400 [1]'.
3.2.
Latent Cancer Fatalities.
The calculation of latent cancer fatalities is perforced on the basis of population doses.
Since no threshold is assumed for cancer induction from radiation exposure, it is reasonable to expect that the number of latent cancer fatalities would not be very sensitive to the various assumptions made in the consequence model.
The single parameter of primary importance is the total quantity of radioactive caterial released in the accident.
Table 7 states the average effective incidence per million man-re= to the whole-body for different calculations. methods.
Tables 8 and 9 state the percentages of latent cancer f.talitics attributable to each exposure pathway calculated from bot crgan-by-organ doses and whole-body doses.
Tables 7 and 8 are diff rent from their counterparts, Tables VI 13-3 and VI 13-4, in Appendix VI of WASH-1400 [1]. First, during preparation of this paper, a programming error in the censequence model was discovered shcreb:. dose effectiveness factors were ceing erroneously applied to ne -alculation 0' k-
- v. east cancer Correction 2195 259 30
Table 7.
EFFECTIVE INCIDENCE OF RADIATIO:;-I::DUCED LATE.NT CA.NCER FATAL LTIES PER MILLIO: F.A:i-REM OF WHGLE-BODY RADIAlIO:4 Upper Lower Bound Release Bound Central Threshold Method Category (BEIR)
Estimate 10 rem 25 rem Whole Body 122 48 46 31 Sum of Individual FWR-1 203
- c7 103 37 Organs PWR-2 143 75 73 62 of this error (i.e., no dose effectiveness factors for breast cancer) increases the everage effective incidence of latent cancer calculated on an organ-by-organ basis by a few percent (Table 7) and enhances the relative importance of breast cancer (Table S),
Second, the calcula-tional method was improvcd for Tables 8 and 9 ccrpared to Table VI 13-3.
For example, the original calculations for the distribution were based upon an actual population distribution which cr ated a misleading result whereas those for Tables S and 9 used a eniform population distribution.
Further, the centribution of each exposure pathway in Table VI 13-3 was estimated by setting its conversien factors to zero which methed is only approximate since the central estimate is based upon a piece-wise linear T.odel; this approximation was circurvented for Tables 3 and 9.
Third, Tables 7I 13-3 and V 13-~ were based uper a BWR-1 (PWR-1) release category.
The results of calculations based upon a PRR-2 release categcry have been added to Table 7 and in Table 9.
The very low probability
(< 10-3/ reactor-year), high latent cancer fatality estimates are dcminatec by the PWR-1 and BWR-1 release categories but the other release categories (PWR-2 through FWR-9 and BWR-2 through BWR-5) dominate the higher probability
(.
10-S/ reactor-year) end of the spectrum The PWR-1 and BWR-1 release categories e
m 31 2199 tou
Table 8.
CONTRIBUTION
. DIFFERENT EXPOSURE PATHWAYS TO LATENT CANCER FATALITIEo FOR A PWR-1 RELEASE CATEGORY ( )
Percentages 0
2 9
9 e
u v
v v
o e
a u
w:
x so a
o o
o 2c S
S 8
5
-0 CE
$i 08 a
a m
m oa
<a w <
a External Cloud
.1
.1
.3
.1
.1
.1 1
1 Inhalation frca Cloud
.3 22
.4
.1
.5
.2 24 5
External Grcund
(< 7 days) 2 2
5
.9
.7 2
13 18 External Ground
(> 7 days) 3 5
19 2
2 7
43 64 Inhalation of Resuspended Contamination
.2 13
.2
.3
.2
.2 14 4
Ingestion of Contaminated Foods 1
.6 2
.6
.5
.7 5
8 SUBTOTALS 12 43 27 4
4 10 100 100 T' _.<c ep t thyrcid cancer which is calculated separately as discussed in Appendix VI of WASH-1400 [1]
(b)The gastrrintestinal tract includes stomach and the rest of alimentary canal
( )"All other" denotes all cancers except those specified in the table (d)Whole-body values are proportional to 50-year whole-body man-rem 2195 261 32
Table 9.
CONTRIBUTION OF DIFFERENT EXPOSUEE PATHWAYS TO LATENT CANCER FATALITIES FOR A PWR-2 RELEASE CATEGORY (a)
Percentages
=
w U
3x e
o e
a e
a e
az a
oe 3
3 S
3N E$
0 5S External Cloud
.2
.1
.5
.1
.1
.1 1
1 Inhalation from Cloud
.5 4
.7
.2
.4
.2 6
3 External Ground
(< 7 days) 3 2
7
.7
.9 3
16 16 External Ground
(>
7 days) 12 8
28 3
4 11 66 68 Inhalation of Resuspended Contamination
.2 1
.2
.4
.2
.1 3
2 Ingestion of Contaminated Foods 2
1 3
1 1
1 9
10 SUBTOTALS 18 16 39 5
6 16 100 100 l )All footnotes to Table 8 apply have a relatively very large ruthenium content compared to the other release categories.
This high Ru-106 content results in a preferential exposure of the lung due to inhalation from the passing cloud and from resuspended contamination and in the dominance of lung cancer in the latent fatalities as shown in Table 8.
For the other release cate-gories, typified by PWR-2, the dominant exposure pathway is from ground 2195 2c2 33
contamination which gives a more uniform dose to the different organs and is comparable to a whole-body dose.
The more prevelant single cancers are estimated to be breast, leukemia and lung.
The more uniform dose associated with the PWR-2 release category also results in a less significant difference between the organ-by-organ and the whole-body dose calculations for PWR-2 than for the PWR-1 release category as shown in Table 7.
Finally, it is also evident in Table 7 that the use of threshold doses of 10 or 25 rem would not greatly reduce the estimates of latent cancer fatalities.
This low sensitivity is due to the already low effectiveness of small doses in the central estimate.
Figure 6 shows the average probability of late death versus distance from the reactor given a PWR-2 release.
(This figure is different from its counterpart, Figure VI
-26, in Appendix VI of WASH-1400 [1] which was based upon a PWR-ia release.)
On the average, the probability of a latent cancer fatality per individual will be roughly proportional to the airborne concentration of radioactive material with distance from the reactor.
The difference between the
" total" and " chronic" probabilities in Figure 6 is the conditional probability of late death due to radiation exposure within the first week; the discontinuity at 25 miles is caused by evacuation.
The chronic risk of latent cancer fatality per individual is roughly constant between 5 to 80 miles since decontamination and/or interdiction actions will limit an individual's dose to approximately 25 rem.
Within a radius of 5 miles, the average chronic risk is reduced since this area will be permanently interdicted under extreme accident conditions.
Beyond 80 miles, the chronic risk falls rapidly with distance since the airborne concentration decreases.
By combining 2195 263 34
10-3 i
l I i ilIii i
i i 4 l l lll l
l 1 I ii1a f-TOTAL g
/
85 8?
~
>- M h$
@x 8Z 5E 10-4 d5 EW p5 E,_
CHR011C 55 u - 5 10-5 1
10 102 103 MILES FROM REACTOR Figure 6.
CONDITIONAL PROBABI'.ITY OF LATENT CANCER DEATH GIVEN A PWR-2 RELEASE (APPROXIMATELY, ABSOLUTE MORTALITY
-0 PROBABILITIES ARE 10 PEP. REACTOR YEAR TIMES STATF9 ONES) this graph with the probability of PWR-2 release and a given popula-tion distribution, one can estimate the number of late fatalities.
Since the population 4s predominantly located beyond 25 miles, the number of late fatalities is roughly proportional to the population living between 25 and 100 miles from the reactor.
21 c)5 204 35
3.3.
Property Damace.
The estimation of costs associated with 4 nuclear reactor accident involves a consideration of costs in four categories, i.e.,
evacnation, crops and milk impoundment, decontamina-tion, and interdiction (relocation of population from interdicted area and loss of value of interdicted area).
For the very large releases of radioactive material, it was found on the average that evacuation contributed 9% to the total costs, milk and crop impoundment 13%,
decontamination 19i. and interdiction 59%.
The need to interdict or decontaminate an area was determined primarily by the concentration of cesium-134 and cesium-137 deposited on the ground.
Milk impoundment was governed primarily by the level of iodine-131, and crop impoundment by strortium-89 and strontium-90.
3.4.
Accident Risks.
A summary of accident risks due to 100 commercial light water reactors is given in Table 5-6 through 5-8 and Figures 5-10 through 5-16 of the Main Report af WASH-1400 [1].
The risks are stated in terms of early fatalities and illnesses, latent cancer fatalities, thyroid nodules, genetic effects, and property damage. Correction of the programming error noted in Section 3.2 increases the average number of latent cancer fatalities stated in the Reactor Safety Study
- by roughly 15%.
This change in the latent cancer fatalities varies from a small increase at higher probability, low consequence events to abcut 25% increase at low probability, high consequence events.
Therefore, the maximum calculated latent cancer fatalities increase from 1500 to about 1900/ year per reactor-year.
- Latent cancer fatalities are stated in Tables 5-5, 5-6, and 5-8 and Figures 5-5 and 5-12 of the Main Report. and in Figure VI 13-33 of Appendix VI.
2195 265 3s
The reader was cautioned (page 13-26) in Appendix VI of WASH-1400 "that a given trial is unlikely to maximize each consequence.
That is, one trial may result in maximum early fatalities but another may result in maximum property damage." As a result of additional work, it can now be stated more emphatically that, in the probability range of 10-8 to 10-9 per reactor-year, there is a very low probability of any early fatalities occurring simultaneously with a large number of latent cancer fagalities.
Thus, it is incorrect to suggest that the total risk is represented by adding the stated early and larent cancer fatalities for two reasons.
First, for a given probability, e.g.,
10~9/ reactor-year, the events resulting in the corresponding early and latent cancer fatalities are disjoint.
Second, early and latent cancer deaths have different characteristics; the former would occur within a short time period after exposure and involve an average of about 40 years life-sbartening for each fatality, whereas most of the latter would occur within the time period 10 to 40 years after exposure and involve an average life-shortenfig of about 10 years for each cancer fatality [11].
The contriLations of the various postulated release categories to the overall risk are stated in Table 10.
It is evident in this table that the PWR-l through FWR-3 and BWR-1 through BWR-3 release categories are the major contributors to accident risks.
The release categories not involving a core meltdown, PWR-8, PWR-9 and BWR-5, are negligible contributors to the public risk.
2195 2<6 37
Table 10.
CONSEQUENCES OF INDIVIDUAL RELEASE CATEGORIES
-9(b)
Accident Risk (a) 10
=
Probability Per Reactor Early Latent Damage Early Latent (C)
Damage Type Year Fatality Fatalities Fatalities Fatalities
($X103)
PWR-1A 4 x 10-7 4.4 x 10-6 6.2 x 10-N 420 500 500 10 PWR-1B 5 x 10-7 1.0 x 10-5 5.5 x 10-4 290 500 660 10 PWR-2 8 x 10-6 1.7 x 10-5 6.1 x 10-3 6700 1000 625 10 PWR-3 4 x 10-6 9.8 x lo-b 5.2 x 10-3 2700 300 625 10 10-4 87 0
200 2
PWR-4 5 x 10-7 0
2.1 x PWR-5 7x 10-7 0
9.7 x 10-5 53 0
83 0.5 PWR-6 6 x 10-0 0
1.2 x 10-4 320 0
12 0.1 PWR-7 4x 10-5 0
1.3 x 10-5 2100 0
2 0.07 4
PWR-8 4x 10-5 0
3.0 x 10 6 34 0
7 0.01 PWR-9 4 x 10-4 0
5.1 x 10-o o
o o,o BWR-1 1x 10-6 4.3 x 10-6 8.7 x 10-4 440 500 660 10 o$
BWR-2 6x 10-6 3.2 x 10-7 3.2 x 10-3 2900 30 825 10 BWR-3 2 x 10-5 0
5.8 x 10-3 2900 0
625 10 BWR-4 2x 10-6 0
7.8 x 10-5 21 0
41 0.1 BWR-5 1 x 10-4 0
1.8 x 10-7 0
0 0
0 (a)
Risk is defined as probability of release category times the integral of the product of probability and consequence magnitude.
The risk values are weighted by the fraction of the particular light water reactor type in the 100 reactor pop 21ation, i.e., 0.66 for PWR and 0 34 ror BWR.
The average consequence can be obtained by dividing the risk by the accident probability per reactor year.
(b)
These values indicate the approximate consequence for a 10-9 per reactor-year probability.
The overall consequences as shown in Appendix VI of WASH-1400 [1]
ps) were obtained by the summation of probabilities for a given consequence magnitude.
These values cannot be reproc :ed by the summation of the 10-9 reactor-year).
sc) consequences at a given probability (e.g.,
/
LJ7 (c) 10-9 latent fatalities are per year values.
N CN N
ACKNOWLEDGEMENTS A work of the magnitude of the Reactor Safety Study consequence model represents the contributions of numerous people.
The authors wish to acknowledge, in addition to their outstanding leadership of the overall study, the major contributions to the consequence model of Norman C. Rasmussen and Saul Levine.
The authors received the support of a very talented group of medical advisors:
Victor P.
- Bond, James F. Crow, Marvin Goldman, Leonard D. Hamilton, George G. Hutchison, Clarence C. Lushbaugh, Roger O. McClellan, Harry R. Maxon, Samuel C.
Morris, James V. Neel, Dean R. Parker, William L. Russell, Eugene L.
Saenger, leonard A.
Sagan, Maurice '. Sullivan, Niel 'dald, and Joseph A. Watson.
A complete list of acknowledgements is contained in Appendix VI of WASH-1400 [1].
2195 2oB 39
REFERENCES
[1]
Reactor Safety Study, "An Assessment of Accident Risk in U.S.
Commercial Nuclear Power Plants," U.S. Nuclear Regulatory Commission, WASH-1400, NUREG-75/014 (1975).
[2] Sprung, J. L. and G. P. Steck, " Correlations between wind direction probability and population at nuclear reactor sites" (Sandia Laboratories report to be published).
[3]
McGrath, P.
E.,
Ericson, D.
M.,
and Wall, I.
B., "The Reactor Safety Study (WASH-1400) and its '.mplication for Radiological Emergency Response Planning," IAEA International Symposium ni Handling of Radiation Accidents, February 28 - March 4, 1977, Vienna, Austria.
[4]
ICRP Task Group on Lung Dynamics, " Deposition and Retention Models for Internal Dosimetry of the Human Respiratory Tract,"
Health Physics 12, 173-207 (1966).
[5]
Hans, J.
M.,
Jr., and Sell, T.C.,
Evacuation Risk - An Evaluation, U.S. Environmental Protection Agency, EPA-520/6-74-002 (1974).
[6]
Wald, N.
and Watton, J.
A.,
" Medical Modification of Human Acute Radiation Injury," Fourth International Conference of the Inter-national Radiation Protection Association, April 24-30, 1977, Paris.
2195 269 a
[7]
The Effects of Exponure to Lo*> Levels of Ionining Radiation (BEIR Report), Advisory Corren!.ttee on the Biological Effects of Ionizing Radiation, Division of Medical Sciences, :Tational Academy of Sciences, National Research Council, 'a'ashington, D.C.
(1972).
[8]
United :iati.ons Scientific Committee on Atomic Radiation (UNSCEAR),
Ionizint, Ra di a t i.on :
Leve_19 and Effects, Report E 72 IX 18, Vols. I and II, l'ni ted :ia tions, New York (1972).
[9]
Federal Radiation Council, Baelground Material for the Development of Protectfve Action Guides for Strontium-89, Strontium-90 and Cesium-132, FRC Staff Report No. 7.(1965).
[10] !!edical Research Council, Criteria for Controllinr_ Radiation Doses to the Public After Accidental Escanes of Radioact.ive Material, lier !!ajesty's Stationery Office, London (1975).
[11] Davis, II. T, et al.,
" Estimates of life-shortening due to latent cancer fron potential nuclear reactor accidents" (Sandia Labora-tories report to be published),
2195 270 u
UNitiD ST ATES
]
F Nuct. tan etcutAfony couMissioN g
W ASHINGTON, D. C.
20555
,g,,,c,,,,,,,,,,,,
b==d
...EA<....E, PEN ALTY FOR PRIVATE U,E. $300 i
.i
=
l h
I i
"E
'Il
]
a H.
s.
=
=.
a
-=
e E'
i
.w' r'
2195 271
~
_