The Assessment of Radiation Exposures in Native American Communities from Nuclear Weapons Testing in Nevada

Abstract

Native Americans residing in a broad region downwind from the Nevada Test Site during the 1950s and 1960s received significant radiation exposures from nuclear weapons testing. Because of differences in diet, activities, and housing, their radiation exposures are only very imperfectly represented in the Department of Energy dose reconstructions. There are important missing pathways, including exposures to radioactive iodine from eating small game. The dose reconstruction model assumptions about cattle feeding practices across a year are unlikely to apply to the native communities as are other model assumptions about diet. Thus exposures from drinking milk and eating vegetables have not yet been properly estimated for these communities. Through consultations with members of the affected communities, these deficiencies could be corrected and the dose reconstruction extended to Native Americans. An illustration of the feasibility of extending the dose reconstruction is provided by a sample calculation to estimate radiation exposures to the thyroid from eating radio-iodine-contaminated rabbit thyroids after the Sedan test. The illustration is continued with a discussion of how the calculation results may be used to make estimates for other tests and other locations.

Full text

Risk Analysis, Vol. 20, No. 1, 2000
101
0272-4332/00/0200-0101$16.00/1 © 2000 Society for Risk Analysis
The Assessment of Radiation Exposures in Native American
Communities from Nuclear Weapons Testing in Nevada
Eric Frohmberg,
1
Robert Goble,
2
Virginia Sanchez,
3
and Dianne Quigley
4
Native Americans residing in a broad region downwind from the Nevada Test Site during the
1950s and 1960s received significant radiation exposures from nuclear weapons testing. Be-
cause of differences in diet, activities, and housing, their radiation exposures are only very
imperfectly represented in the Department of Energy dose reconstructions. There are impor-
tant missing pathways, including exposures to radioactive iodine from eating small game. The
dose reconstruction model assumptions about cattle feeding practices across a year are un-
likely to apply to the native communities as are other model assumptions about diet. Thus ex-
posures from drinking milk and eating vegetables have not yet been properly estimated for
these communities. Through consultations with members of the affected communities, these
deficiencies could be corrected and the dose reconstruction extended to Native Americans. An
illustration of the feasibility of extending the dose reconstruction is provided by a sample calcu-
lation to estimate radiation exposures to the thyroid from eating radio-iodine-contaminated
rabbit thyroids after the Sedan test. The illustration is continued with a discussion of how the
calculation results may be used to make estimates for other tests and other locations.
KEY WORDS:
Native American; dose reconstruction; radiation; nuclear testing; Nevada
1. INTRODUCTION
The Nevada Test Site (NTS) was selected in 1950
as the prime U.S. continental location for nuclear
weapons testing. In part it was chosen for “its climate,
remoteness, the low population density in the area,
and the fact that the adjoining Nellis Air Force Base
Bombing and Gunnery Range . . . minimized risk to
public safety while providing added security.”
(1)
Sub-
stantial amounts of radiation were released from nu-
clear weapons tests and this radiation traveled long
distances. Hence remoteness and low population
density did not mean there were no significant expo-
sures. Faced with public concerns and lawsuits, the
Department of Energy (DOE) began in 1979 its esti-
mates of radiation doses from nuclear weapons test-
ing at NTS, with the Off-Site Radiation Exposure Re-
view Project (ORERP).
(2,3)
The ultimate goal was to be able to estimate the potential
dose to any person who lived in an area where fallout
from the NTS was deposited, based on that person’s age,
occupation, and place of residence. Both external and in-
ternal doses were to be considered and the calculations
were to use actual data (as opposed to assumed values) if
appropriate data could be found. (3, p. 470)
This was the first of DOE’s major dose recon-
structions, estimates of the radiation exposures that
people may have received from living downwind of
nuclear production and testing.
(4)
As such it was a
major accomplishment and had considerable influence.
The models developed were used as templates for
1
Maine Bureau of Health, Augusta, ME (present address) and
Clark University, Worcester, MA.
2
Clark University, Center for Technology, Environment, and De-
velopment of the George Perkins Marsh Institute, Worcester,
MA 01610.
3
Citizens Alert–Native American Project.
4
Clark University, Worcester, MA.

102 Frohmberg
et al.
other dose reconstructions, such as the Hanford En-
vironmental Dose Reconstruction Project, and pro-
vided experience in collecting and evaluating a wide
variety of historical data.
(2,5,6)
Because it was early in
the learning process there were also difficulties and
deficiencies in the effort. The computational effort
was very complex and is thus difficult to adapt to an-
swering new questions. The importance of public in-
volvement in obtaining information, in reviewing the
models, in interpreting results, and in developing
public confidence in the results
(4)
was not recognized.
No concerted effort was made for identifying sub-
populations with special lifestyle characteristics, de-
spite the strong influence lifestyle, especially diet, had
on exposure. Some recent efforts at dose reconstruc-
tion for native populations are described by Harris
and Harper and by Simon and Graham.
(7,8)
Further-
more, at that time stakeholder involvement in plan-
ning and design of the study as contemplated, for in-
stance, in the NAS report
Understanding Risk
(9)
and
explored to some extent in later Hanford studies,
(7)
was not even considered.
Fig. 1. Traditional lands of the Western Shoshone and Southern Paiutes. The Nevada Test Site is shown in the
center; two arrows indicate the most frequent wind directions for nuclear tests, chosen to avoid transport in
the direction of major southwestern cities.

Radiation Exposures in Native Americans 103
Among the populations with the largest expo-
sure to nuclear testing were Western Shoshone and
Southern Paiutes. Their traditional lands are shown
in Fig. 1. The lands surround the Nevada Test Site and
occupy the areas that were downwind for most of the
tests. Isopleths for external radiation exposures from
nuclear weapons testing at NTS to date are shown in
Fig. 2. The two figures together suggest that radiation
exposures to Western Shoshone and Southern Paiute
communities may be of concern and raise the further
concern of whether the ORERP dose reconstruction
provides an adequate description of such exposures.
This paper explores these two concerns by at-
tempting to answer the following questions:
1. Were significant radiation exposures received
by members of Western Shoshone and South-
ern Paiute communities as a result of nuclear
weapons testing in Nevada?
2. Does DOE dose reconstruction provide an
adequate basis for assessing these exposures?
3. Can the inadequacies of the dose reconstruc-
tion be remedied by using information that
DOE has already?
4. On the basis of presently available informa-
tion, approximately how severe were the ra-
diation doses received by community mem-
bers, particularly those doses not adequately
described in the DOE dose reconstruction?
5. Assuming additional information will be
needed for any dose reconstruction, what is a
suitable approach to obtaining such informa-
tion and to designing the dose reconstruction?
Fig. 2. Spatial distribution of the cumulative external exposures from nuclear weapons testing at the Nevada Test Site through the period of
testing (1951–1972) as estimated in the ORERP. The heaviest exposures occur downwind in the directions indicated by the arrows in Fig. 1.

104 Frohmberg
et al.
Although our analysis directed to the first four
questions is conducted by using methods compatible
with the DOE reconstruction, we expect that cre-
ation of a dose reconstruction that will be useful to af-
fected communities will involve a reformulation of
the questions asked and, correspondingly, adjust-
ments to the analytical methods and the information
base. Such reformulations have not been made by
communities around the test site, and our concern in
question (5) is with who participates in designing and
conducting studies.
2. KEY RESULTS FROM THE OFF-SITE
RADIATION EXPOSURE REVIEW
PROJECT FOR THE NEVADA TEST SITE
The DOE Dose Reconstruction for the Nevada
Test Site was the first major effort by DOE to compile
estimates of radiation doses resulting from its facili-
ties. Like other efforts, it was developed in response
to expressed public concerns and actual or potential
lawsuits. The methods used and the models devel-
oped have strongly influenced subsequent dose re-
construction efforts. The initial challenge was to as-
sess the historical information available for its
suitability in estimating doses. As described by Thomp-
son and McArthur:
(3)
Ideally, dose calculations for areas affected by fallout
from nuclear tests at the NTS would be based on exter-
nal radiation exposures and on levels of radiation in
the air, food, and water consumed by residents in the
days and weeks after a test. Although these quantities
were measured in some places and for some tests, there
were not enough such data to allow all the necessary
doses to be calculated. Instead a computational scheme
was developed that used historical data on fallout dep-
osition as the fundamental data set and used computer
models of ingestion, inhalation, and external exposure
as the mechanism for calculating dose . . . (p. 470)
Accordingly, the historical data base included
(1) extensive but scattered measurements of radia-
tion and radioactivity after the tests; these measure-
ments were, however, taken for other purposes than
assessing doses to members of the population; and
(2) survey information on agricultural practices and
markets and on lifestyles of the exposed population.
The computational scheme used this information to
develop estimates of external and internal doses. Ex-
ternal doses were based on measurements of total
radioactivity after each test. Internal doses used the
same deposition data and depended on models for
food production and consumption based on the sur-
vey data. The DOE dose reconstruction did not pro-
vide for significant public participation in the design
of the study or in the development of its informa-
tion; nor did it provide for any participation in the
review of the methods and models or in the assess-
ment of results.
Native Americans were not included in the sur-
veys and the dose reconstruction did not specifically
identify Native American lifestyle characteristics, in-
cluding agricultural, hunting and gathering practices
and diets. When asked to provide dose estimates to
representative members of Native American commu-
nities, DOE gave estimates for a “shepherd lifestyle”
as representative of an outdoor lifestyle.
(10)
The DOE estimates indicate that external whole
body exposures were significant, on the order of 10
mSv (1 Rem), for residents of many Native American
communities such as Duckwater, Ely, and Moapa,
and external exposures at some locations off-site in
Nevada and Utah were substantially greater.
(3,10)
In-
ternal dose estimates for most organs were smaller;
however, doses to the thyroid were as much as .1–.2
Sv (10–20 Rem) for adults, and in excess of 1 Sv (100
REM) for small children.
(3,10)4
There is substantial un-
certainty in these estimates as well as considerable in-
terindividual variability in the doses received by indi-
viduals residing in the same location at the same
times.
5
However, quantitative estimates of the uncer-
tainty and variability are not routinely provided in
the ORERP materials.
How well do the dose estimates for the shepherd
lifestyle represent what happened to Native Ameri-
cans at the various locations? The lifestyle may de-
scribe external exposures with reasonable accuracy.
Before concluding this, however, several questions
should be answered. One is to what extent the mobil-
ity of these populations, who traveled extensively to
hunt, gather pine nuts, and visit relatives, affects ex-
posure estimates based on residence. In particular,
hunting and gathering pine nuts may have led people
to travel relatively close to the NTS. A second ques-
tion is whether terrain and special meteorological
conditions (such as fog) will have systematic effects
on deposition at locations—particularly high altitude
4
The next highest internal doses were to the lower digestive tract,
which reflects the passage of a variety of insoluble radionuclides.
5
The importance of distinguishing uncertainty from variability is
discussed at length in the National Academy report Science and
Judgment in Risk Assessment.
(11)
Variability represents actual dif-
ferences in people’s experience, whereas uncertainty represents
lack of knowledge. The two have very different implications in
most practical situations.
(12)

Radiation Exposures in Native Americans 105
locations—which were used by people but remote
from the roads where radiation measurements were
made. It is clear from experience with Chernobyl
(13)
and from acid deposition studies
(14)
that there can be
very substantial local variation in amounts deposited.
A third question is whether other specific aspects of
lifestyle affect external exposures significantly: did
the use of fur and fresh basket material contribute
significantly to exposure? The effect of taking these
issues into account will surely be to increase the esti-
mated variability in doses; it may increase the mean
dose estimates as well.
For internal exposures, the shepherd lifestyle is
clearly unrepresentative: it ignores the contribution
of hunting to the diet; this represents a missing path-
way from the DOE analysis. It is very likely that the
DOE assumptions about how cattle and other live-
stock were fed during the year do not apply,
6
and the
DOE-assumed seasons for vegetables may not de-
scribe the collecting and eating of both fresh produce
and wild vegetation within the native communities.
Furthermore, the assumptions about variability in
diet and in other individual characteristics must be
examined anew for these populations. As a result the
DOE estimates for internal doses cannot be relied
upon; they are almost certainly underestimates be-
cause they neglect the important pathway of hunting
and because the communities’ cattle feeding prac-
tices rely less on stored feed.
3. FEASIBILITY OF CORRECTING THE DOE
DOSE RECONSTRUCTION TO ACCOUNT
FOR NATIVE LIFESTYLES
Two kinds of adjustments are needed in the dose
reconstruction models: (1) models for missing path-
ways must be added, and (2) the included pathway
models must be corrected to reflect Native American
practices. The initial step must be a pathway analysis.
This can be done effectively only in collaboration
with the affected communities. Simon and Graham
(8)
observed at the conclusion of their work on dose as-
sessment from nuclear weapons testing in the Mar-
shall Islands:
The single most important conceptual requirement for
conducting valid assessments is recognizing important
exposure pathways. . . . One lesson learned repeatedly
in Marshall Islands studies has been to rely on local ex-
pertise to provide information important to acquiring
an understanding of pathways. No better information
can be produced than that provided by the population
whose quality of life is under examination. (p. 453)
The ORERP modeling approach can be adapted
naturally to an appropriate revision of the pathway
analysis. This is because the DOE dose reconstruc-
tion creates an estimate of the amount of fallout for
each test over the geographical areas of concern and
because amounts of fallout are the starting point for
estimating doses through all pathways. New informa-
tion that pertains to each pathway is needed on how
members of the Native American communities lived
during the periods of testing. The required informa-
tion can be developed reliably only by the commu-
nities. Furthermore, it is important to remember
that lifestyles may well be different in different
communities and among different tribes. Such in-
formation will vary by season and includes key ele-
ments of diet including practices in hunting and
preparing game, agricultural practices, livestock
management, and the collection of wild plants. Fac-
tors affecting external exposures include housing
characteristics and outdoor living practices and the
use of various natural materials. Members of the
communities may have traveled substantially to
hunt and to gather pine nuts, and many community
members may have moved during the periods of
testing. Therefore, information from the communi-
ties about people’s movements may also be needed
to relate to the geographical distribution of fallout
from the various tests.
Hunting and eating game is at least one major
pathway that does not appear in the present version
of the model. To demonstrate the feasibility of adjust-
ing the dose reconstruction, we present an illustrative
model developed with particular Western Shoshone
and Southern Paiute communities for one particular
hunting pathway.
4. A MODEL FOR RADIO-IODINE DOSES
FROM THE HUNTING AND EATING
OF RABBITS
7
Small game, especially rabbit,
8
was an important
staple in the Western Shoshone and Southern Paiute
6
For instance, Shoshone cattle supplying milk to community resi-
dents—the most important ORERP pathway—were grazing for
more of the year than the ORERP assumptions indicate.
7
Details on this model and the numerical assumptions appear in
the MA thesis of Eric Frohmberg.
(15)
8
Other important animals hunted and eaten include Yellow-Bel-
lied Marmot (
Marmota flaviventris
, locally called groundhogs),
various species of ground squirrel (
Spermophilus spp.
), porcu-
pines, sage hen, blue grouse, and dove.
(16)

106 Frohmberg
et al.
diets during the period of testing. We have learned
from the affected communities that very little of the
animals was wasted, and, in particular, that the ani-
mal thyroids were routinely eaten.
9
On the basis of in-
formation from community members on rabbit hunt-
ing and on a set of measurements of radio-iodine in
vegetation and in rabbit thyroids collected by DOE
after the Sedan nuclear test in July 1962, we have con-
structed an illustrative model for radiation doses to
the thyroid from eating jackrabbits.
10
The model has three components: (1) a vegeta-
tion submodel, which describes variations in the
amount of radio-iodine in vegetation and its disap-
pearance with time; (2) a rabbit thyroid submodel,
which relates the amount of radio-iodine in the rabbit
thyroid to the amounts of radio-iodine in vegetation;
and (3) a human dose model, which relates the dose
to the human thyroid to the amount of radio-iodine
in the rabbit thyroid. The third submodel uses Dun-
ning and Schwart’s ingestion dose coefficients
(17)
to
estimate human doses. The three submodels are illus-
trated schematically in Figs. 3, 4, and 5.
After the Sedan test in 1962, DOE collected
data on radio-iodine concentrations in vegetation
and in rabbit thyroids at four locations (Groom Val-
ley, Penoyer Valley, Railroad Valley, and Currant)
for several weeks.
(18–21)
These data were used to de-
fine the vegetation and rabbit submodels. The ap-
proximate locations where vegetation and rabbit
thyroid data were collected after the Sedan test are
shown in Fig. 6.
11
The vegetation submodel has three components:
(1) a lognormal distribution fit to the data of Turner
and Martin
(21)
describing the initial concentrations of
radioactive iodine in vegetation at each of the four lo-
cations; (2) a lognormal distribution again represent-
ing a fit to the data of Turner and Martin
(21)
and describ-
ing weathering rates of the iodine from the vegetation;
and (3) the radiological half-life of iodine-131, which
is approximately 8 days.
The rabbit submodel uses the vegetation sub-
model as an input. It has four additional components:
(1) a vegetation ingestion rate for the rabbits eating
vegetation, chosen to be a narrow lognormal distribu-
tion following an argument of Turner;
(20)
only limited
data are available on this factor; (2) a thyroid uptake
factor representing the amount of ingested iodine
that is absorbed and stored in the thyroid; again little
data are available. We created a lognormal distribu-
tion by modifying Turner’s argument
(20)
to use only
absorption data pertaining to jackrabbits; (3) a bio-
logical half-life or excretion rate; again we modified
Turner’s estimated distribution
(20)
by using only data
for jackrabbits in creating a lognormal distribution
for this quantity; and (4) the half life for iodine-131.
The human submodel has two components: (1) a
distribution describing a range of estimates for the
number of rabbits people ate per week—this was
based on a small survey of Western Shoshone and
Southern Paiute community members and should
not be regarded as having general application; and
(2) dose-conversion estimates relating amounts of
iodine ingested to doses; the coefficients and the log-
normal distribution describing their variability were
Fig. 3. Schematic illustrating the vegetation submodel.
9
Cleaning methods minimize waste. A rabbit, marmot, or ground
squirrel is gutted by removing the intestines through a slit either
in the stomach or under the armpit. The intestines are discarded.
Organs not discarded include the heart, lungs, liver, and upper
gastrointestinal tract (including the esophagus and associated or-
gans like the thyroid). When a larger mammal (such as a deer) is
harvested, the esophagus would be removed and discarded be-
cause of its large size and ease of removal. As the esophagus is re-
moved the thyroid would likely also be removed.
(16)
10
The Western Shoshone eat both black-tailed jackrabbits (
Lepus
californica
)
and white-tailed jackrabbits (
Lepus townsendii
).
They also eat various species of cottontail rabbits (
Sylvilagus
spp.
). Most of the measurements from the Sedan test were on
jackrabbits and we use properties of jackrabbits (such as weight
and weight of thyroid) in constructing the model. Cottontails and
other small mammals would give similar results per weight of
animal (amount of food provided).
11
The Sedan test on July 6, 1962, was one of the largest tests at NTS.
It was an approximately 100-kton test designed to produce a large
crater. Sedan was part of the Plowshare project testing the use of
nuclear weapons for excavation.
(22)
Although large, the test was not
considered one of the dirtier tests (in terms of total U.S. fallout).
(6)
Fig. 4. Schematic illustrating the rabbit submodel.

Radiation Exposures in Native Americans 107
taken from Dunning and Schwarz.
(17)
These estimates
and their variability have been reviewed by Ng;
(23)
they are similar to those found in ICRP 56 and Sny-
der et al.
(24,25)
Predicted values from the vegetation and rabbit
submodels were compared with measured concentra-
tions of radio-iodine in vegetation and in rabbit thy-
roids for the four locations and for the various times
of measurements (extending over 30 days after the
Sedan test) with the data of Turner and Martin.
(21)
On
average—for each location and time period—the
predictions and data agreed within a factor of 3 and
there were no obvious biases. The models thus pro-
vide a consistent representation of these data.
Developing illustrative dose estimates in steps is
convenient. The dose to an individual from eating one
rabbit after a test such as the Sedan test will depend on
the location from which the rabbit was taken—because
the amount of fallout was different at different loca-
tions; it will depend on the amount of time that elapsed
after the test before shooting the rabbit—because
radio-iodine first is added to the thyroid as the rabbit
eats contaminated vegetation but then declines as the
concentrations on vegetation decline from weathering
and radioactive decay and as radio-iodine is lost from
the rabbit by excretion and radioactive decay. The dose
also depends on the age of the individual. Similarly,
there will be considerable variability in the dose re-
ceived from one rabbit compared with another, which
reflects the variability described in each of the three
submodels. An interesting initial step is to calculate the
dose to an individual from eating one rabbit at the time
a few days after the test when the concentrations of io-
dine in the rabbit thyroid were at their highest.
• The average (mean) such dose to an adult
(17
1
years old) from eating a Groom Valley
rabbit a few days after the Sedan test was ap-
proximately 10 mSv (1 REM) whereas 5% of
adults would have had doses of approximately
30 mSv (3 REM). The mean dose to a small
child (
½
–2 years old) from one Groom Valley
rabbit would have been 90 mSv (9 REM),
whereas 5% of children would have had doses
of 0.3 Sv (30 REM).
• At Currant the corresponding (mean and 5th
percent) a doses a few days after the Sedan test
were .6 mSv (60 mrem) and 2 mSv (200 mrem),
whereas the small child doses were 5 mSv (0.5
REM) and 16 mSv (1.6 REM), respectively.
There was even more variability in the expected
doses people received from eating rabbits after the test
because there was variability in the time after the
Fig. 6. Approximate locations where data were collected on con-
centrations of I-131 in vegetation and rabbit thyroids after the
Sedan nuclear test in 1962.
Fig. 5. Schematic illustrating the human submodel.

108 Frohmberg
et al.
test when a particular rabbit may have been con-
sumed, and there was also variability in the frequency
with which people ate rabbits. As an illustration of
the range of doses expected and their dependence on
time after the test, we show in Figs. 7 and 8 daily dose
estimates for a 3-year-old child (
½
–2 years old) con-
suming rabbit portions over time after the Sedan test,
again for rabbits taken from Groom Valley (Fig. 7)
and from Currant (Fig. 8). We use the survey results
to estimate frequency of eating rabbits.
Accumulated doses from eating rabbits after the
Sedan test may be found by summing the daily values.
The result is approximately a factor of 3 times the peak
dose from one rabbit described previously. Thus,
• The accumulated dose to be expected for typical
consumption of rabbits taken from Groom Val-
ley after the Sedan test is approximately 30 mSv
(3 REM) for an average adult and .1 Sv (10
REM) for the most exposed 5%; the accumu-
lated dose for small children is approximately
0.25 Sv (25 REM) for the average child and .8
Sv (80 REM) for the most exposed 5th percent
• The accumulated dose to be expected for typical
consumption of rabbits taken from Currant
after the Sedan test is approximately 1.8 mSv
(180 mrem) for an average adult and 6 mSv (.6
REM) for the most exposed 5%; the accumu-
lated dose for small children is approximately
15 mSv (1.5 REM) for the average child and 50
mSv (5 REM) for the most exposed 5th percent.
5. EXTENSION OF THE MODEL TO OTHER
TESTS AND OTHER LOCATIONS
The model results may be used at other loca-
tions and for other tests by multiplying the results by
the ratio of the amount of radio-iodine in fallout at
the location and test of concern to the amount of
radio-iodine in fallout for the Sedan test at Groom
Valley or Currant.
(26)
The ORERP data base con-
tains this information.
(2,27,28)
Making these adjust-
ments will require combining the ORERP estimates
of aggregate radioactive deposition for each nuclear
test with test-specific multipliers developed by
Hicks
(26)
for estimates of the amount of each radio
nuclide deposited. A coherent development of this
method, which should be pursued in the context of a
revised dose reconstruction, would compare Martin
and Turner’s measurements and measurement
methods
(21)
at the four locations with the ORERP
and Hicks measurements, measurement methods,
and interpolation, both in that region and over the
Fig. 7. The distribution of daily dose estimates (mSv) for a 1-year-old child eating rabbits taken from Groom
Valley after the Sedan test according to the survey of rabbit consumption. Shown are 5th percentile, 50th per-
centile, mean, and 95th percentile estimates.

Radiation Exposures in Native Americans 109
time periods and regions of interest. Such analysis
should provide an indication of the deposition mea-
surement component of the uncertainty in dose esti-
mates; however, that analysis was beyond the scope
of this project. The following calculation for accu-
mulated exposures at Duckwater illustrates the ap-
proach and will give a reasonably good approxima-
tion to the results that can be expected to be
obtained. Again, this approach should be based on
community-specific information.
• At Duckwater, the fallout amounts from the
Sedan nuclear test were similar to the
amounts at Currant. This can be observed by
comparing both external doses and thyroid
doses from the ORERP database for both
Currant and Duckwater for the Sedan test.
• The ratio of the DOE-estimated median adult
thyroid doses (through other pathways), which
are proportional to the amount of iodine in fall-
out summed over 11 tests, to the DOE thyroid
dose estimate for the Sedan test is approxi-
mately 90. Two of these tests, George (June
1951) and Apple II (May 1955), had thyroid
doses approximately 24 times that for Sedan.
• So to estimate doses from eating rabbits from
a number of tests at Duckwater, one multi-
plies the Currant dose estimates for Sedan by
a factor of 90. This yields adult doses of 160
mSv (16 REM) and 500 mSv (50 REM) (mean
and most exposed 5%, respectively) and child
doses of 1.3 Sv (130 REM) and 4 Sv (400
REM), respectively, for the 11 tests.
• To obtain dose estimates for George or for
Apple II, one multiplies the Currant dose esti-
mates for Sedan by a factor of 24. This yields
adult doses of 40 mSv (4 REM) and 120 mSv
(12 REM) (mean and most exposed 5%, re-
spectively) and child doses of 0.3 Sv (30 REM
and 1 Sv (100 REM), respectively, for each of
these tests.
6. DISCUSSION OF DOSE ESTIMATES
These thyroid dose estimates are substantial.
The range of adult dose estimates for eating rabbits
from Duckwater extends above the DOE dose esti-
mate covering all pathways. The small child doses ex-
ceed the average doses observed in the Utah epide-
miological study, which found increased rates of
thyroid cancer.
(29,30)
They are also greater than the
lowest levels at which thyroid cancers have been at-
tributed to exposure from Chernobyl.
(31)
The whole
body doses estimated by DOE overlap the range in
which leukemias were found in Utah.
(32,33)
Fig. 8. The distribution of daily dose estimates (mSv) for a 1-year-old child eating rabbits taken from Currant
after the Sedan test according to the survey of rabbit consumption. Shown are 5th percentile, 50th percentile,
mean, and 95th percentile estimates.

110 Frohmberg
et al.
7. CONCLUSIONS
In the introduction we listed five questions to be
answered in the paper. The answers given may be sum-
marized as follows:
1. Native American community members living
in their ancestral lands received substantial
exposures from Nuclear Weapons testing at
the Nevada Test Site.
2. These exposures are not adequately de-
scribed by DOE’s dose reconstruction: (1)
missing pathways include iodine exposures
from the hunting of small game; (2) the model
assumptions about the feeding of cattle that
provide dairy products do not generally apply;
and (3) assumptions about other food path-
ways may also not apply.
3. The information base collected by DOE is not
adequate to address these important aspects
of a dose reconstruction, and this failing
stems from the lack of participation by Native
community members in the collection, inter-
pretation, and planning for the use of key in-
formation about lifestyles and concerns.
4. Approximate estimates of radiation doses to
thyroids received from eating rabbit thyroids
can be estimated on the basis of scattered data
from DOE studies and information selectively
gathered from affected communities. We have
presented illustrative calculations showing
that these exposures were severe.
5. A successful dose reconstruction, i.e., a dose
reconstruction that is both useful and reason-
ably accurate, can be achieved only with ac-
tive participation by members of the affected
communities.
It is worth noting some of the questions we have not
answered in this paper:
• We did not provide a corrected dose recon-
struction. The dose estimates we present are il-
lustrative on the basis of small samples and
special cases of test site experience. Our pur-
pose here was simply to display the gap be-
tween the present dose reconstruction and ac-
tual experience, to show that the actual
experience was radiologically significant, and
to show that a technical basis for developing
the needed exposure models is available.
• We did not describe a detailed methodology
for revising the dose reconstruction. That task
includes (1) defining how to usefully character-
ize dose information, including uncertainties;
and (2) establishing sources and procedures
for obtaining pathway and other needed data.
• We did not address community concerns about
the legacy of radioactive contamination that
extend beyond questions about historical ex-
posures immediately following nuclear tests.
Information developed from dose reconstruc-
tion will be part of the history of the affected commu-
nities. Individuals may wish to know about their own
exposures and about the nature of the exposures that
occurred in their communities from nuclear weapons
testing. Even at this late date, knowledge of radiation
doses may be helpful in developing mitigation mea-
sures such as medical monitoring. And the information
could be used as a basis for assigning compensation to
people or communities injured by the exposures.
These possible uses exemplify the importance of hav-
ing the communities maintain an active role in the
creation of the study. When and what sort of medical
monitoring could help depends on community char-
acteristics and community medical needs. The suit-
ability and appropriateness of compensation for past
wrongs vexes many Native communities and differ-
ent communities with different histories may react
differently. The fact that the initial dose reconstruc-
tions were constructed while the U.S. government
was denying legal responsibility for harm caused by
nuclear testing has not helped create confidence in
the usefulness of dose reconstruction efforts.
Dose reconstruction for these communities is
feasible as demonstrated by the illustrations we have
presented. But such reconstruction will require par-
ticipation and collaboration in developing informa-
tion appropriate to the variety of affected communi-
ties. Furthermore, successful completion of dose
reconstructions for communities will occur only with
substantial participation and collaboration with the
communities in formulating the important questions
to be answered and the design and implementation of
the studies along with the development of the needed
information base.
ACKNOWLEDGMENTS
This research was supported in part by funds
from the National Institute of Environmental Health
Sciences (NIEHS) Environmental Justice–Partnerships
in Communication (RFA-ES-94-005), and The Com-

Radiation Exposures in Native Americans 111
prehensive Environmental Response Compensation
and Liability Act (CERCLA) Trust Fund through a
cooperative agreement with the Agency for Toxic
Substances and Disease Registry (ATSDR) Tribal
Environmental Health Education Program. Addi-
tional funding was provided by the Childhood Cancer
Research Institute, the Ruth Mott Fund, the Public
Welfare Foundation, the W. Alton Jones Foundation,
the North Shore Unitarian Universalist Veatch Pro-
gram, and the Ben and Jerry’s Foundation.
We are grateful to the Ely-Shoshone Tribe and
the Native American Nuclear Risk Management
Community Advisory Committee for information,
advice, and encouragement. Kim Townsend and
Maurice Frank provided detailed information on
hunting and cleaning of game. David Wheeler and the
Las Vegas Office of DOE provided ORERP results
and helped us understand them. Dan Handy, Do-
minic Golding, Dale Hattis, Doug Brugge, and Peter
Ford have advised and assisted us. We also appreciate
the observations of three anonymous referees.
REFERENCES
1. U.S. Department of Energy.
Profile of the NTS
(http://
www.v.doe.ov/nts.html, 1995).
2. B. W. Church, D. L. Wheeler, C. M. Campbell, R. V. Nutley, and
L. R. Anspaugh, “Overview of the Department of Energy’s
Off-Site Radiation Exposure Review Project,”
Health Physics
59
, 503–510 (1990).
3. C. B. Thompson and R. D. McArthur, “Challenges in Develop-
ing Estimates of Exposure Rate Near the Nevada Test Site,”
Health Physics
71
, 470–476 (1996).
4. C. W. Miller and J. M. Smith, “Why Should We Do Environ-
mental Dose Reconstructions?”
Health Physics
71
, 420–424
(1996).
5. A. M. Bouville, M. Dreicer, H. L. Beck, W. H. Hoecker, and B. W.
Wachholz, “Models of Radioiodine Transport to Populations
within the Continental U.S.,”
Health Physics
59
, 659–668 (1990).
6. H. L. Beck, I. K. Helfer, A. M. Bouville, and M. Dreicer, “Es-
timates of Fallout in the Continental U.S. from Nevada Weap-
ons Testing Based on Gummed-film Monitoring Data,”
Health
Physics
59
, 565–576 (1990).
7. S. G. Harris and B. L. Harper, “A Native American Exposure
Scenario,”
Risk Analysis
17
, 789–796 (1997).
8. S. L. Simon and J. C. Graham, “Dose Assessment Activities in
the Republic of the Marshall Islands,”
Health Physics
71
, 438–
456 (1996).
9. National Research Council,
Understanding Risk
(National
Academy Press, Washington, DC, 1996).
10. D. Wheeler,
personal communication
(1995).
11. National Research Council.
Science and Judgement in Risk As-
sessment
, pp. 1–651 (National Academy Press, Washington,
DC, 1994).
12. D. Hattis and R. Goble,
Needs for Improvement in Priority-Setting
Methodologies
(Resources for the Future, Annapolis, Mary-
land, 1992).
13. C. Hohenemser, “The Accident at Chernobyl: Health and En-
vironmental Consequences and the Implications for Risk
Management,”
Ann. Rev. Energy
13
, 383–428 (1988).
14. D. Bennett, R. Goble, and R. Linthurst,
The Acid Deposition
Phenomenon and Its Effects: Critical Assessment Document
(U.S. EPA, 1985).
15. E. Frohmberg,
Aggravation of Environmental Equity in the De-
sign and Culture of Risk Assessment
(Clark University, 1997).
16. M. Frank,
Personal Communication
(1995).
17. D. E. J. Dunning and G. Schwarz, “Variability of Human Thy-
roid Characteristics and Estimates of Dose from Ingested
Iodine-131,”
Health Physics
40
, 661–675 (1981).
18. W. E. Martin, “Interception and Retention of Fallout by
Desert Shrubs,”
Health Physics
11
, 1341–1354 (1965).
19. W. E. Martin,
Early Food-Chain Kinetics of Radionuclides fol-
lowing Close-in Fallout from a Single Nuclear Detonation
, 1-758-
782 (Atomic Energy Commission, 1965).
20. F. B. Turner,
Uptake of Fallout Radionuclides by Mammals and
a Stochastic Simulation of the Process
, 1-800-820 (Atomic En-
ergy Commission, 1965).
21. F. B. Turner, and W. E. Martin,
Food-chain Relationships of
Iodine-131 Following Two Nuclear Tests in Nevada
(UCLA,
1963).
22. C. R. Schoengold, M. E. Demarre, and E. M. McDowell,
Ra-
diological Effluents Released from Announced U.S. Continental
Tests 1961–1981
(U.S. Department of Energy, Nevada, 1990).
23. Y. C. Ng,
et al.
, “59 ORERP Internal Dose Estimates for Indi-
viduals,”
Health Physics
5
, 693–713 (1990).
24. ICRP.
Age-Dependent Doses from Intake of Radionuclides
(International Commission on Radiation Protection [ICRP],
1990).
25. S. F. Snyder, W. T. Farris, B. A. Napier, T. A. Ikenberry, and
R. O. Gilbert,
Parameters Used in the Environmental Pathways
and Radiological Dose Modules (DESCARTES, CIDER, and
CRD Codes) of the Hanford Environmental Dose Reconstruc-
tion Integrated Codes (HEDRIC)
(Battelle, Pacific Northwest
Laboratories, 1994).
26. H. G. Hicks, “Calculation of the Concentration of Any Radio-
nuclide Deposited on the Ground by Offsite Fallout from a
Nuclear Detonation,”
Health Physics
42
, 585–600 (1982).
27. T. B. Kirchner, F. W. Whicker, L. R. Anspaugh, and Y. C. Ng,
“Estimating Internal Dose due to Ingestion of Radionuclides
from Nevada Test Site Fallout,”
Health Physics
71
, 487–501
(1996).
28. F. W. Whicker, T. B. Kirchner, L. R. Anspaugh, and Y. C. Ng,
“Ingestion of Nevada Test Site Fallout: Internal Dose Esti-
mates,”
Health Physics
71
, 477–486 (1996).
29. R. A. Kerber,
et al.
, “A Cohort Study of Thyroid Disease in Re-
lation to Fallout from Nuclear Weaspons Testing,”
JAMA
270
,
2076–2082 (1993).
30. J. E. Till,
et al.
, “The Utah Thyroid Cohort Study: Analysis of
the Dosimetry Results,”
Health Physics 68, 472–483 (1995).
31. E. E. Buglova, J. E. Kenigsberg, and N. V. Sergeeva, “Cancer
Risk Estimation in Belarussian Children due to Thyroid Irra-
diation as a Consequence of the Chernobyl Nuclear Accident,”
Health Physics 71, 45–49 (1996).
32. S. L. Simon, et al., “The Utah Leukemia Case-Control Study:
Dosimetry Methodology and Results,” Health Physics 68, 460–
471 (1995).
33. W. Stevens, D. C. Thomas, J. L. Lyon, J. E. Till, R. A. Kerber,
S. L. Simon, R. D. Lloyd, N. A. Elghany, and S. Preston-Martin,
“Leukemia in Utah and Radioactive Fallout From the Nevada
Test Site: A Case-Control Study,” JAMA 264, 585–591 (1990).