VOL. 119, NO. 4, OCTOBER 2020

THE RADIATION SAFETY JOURNAL

The Official Journal of the Health Physics Society

www.health-physics.com Editorial

Introduction to the Trinity Nuclear Test Collection of Papers

THIS SPECIAL collection of papers in Heath Physics focuses Duringthelast6y,researchersattheNational on assessments of the environmental contamination, Cancer Institute undertook a risk projection study to radiation doses, and health risks from the Trinity nu- address the questions of interest for Trinity more fully. clear test. Trinity was the first detonation of a nuclear They received input from regional stakeholders and ex- device in the history of the world and took place on perts in New Mexico and derived five major 16 July 1945 in south-central New Mexico, three questions of concern, all which have been weeks before the bombing of Hiroshima and Nagasaki addressed in this collection of publications. These at the end of WW II. questions are: The test was kept secret, in particular, from the • What models and relevant input data are available to esti- residents of New Mexico, who received no advanced mate the doses to New Mexico residents? warning. It was not until 1987 that a detailed • What are the estimated magnitudes of organ doses received assessment of the Trinity fallout plume was published by location, age, and ethnicity in New Mexico from the in the scientific peer-reviewed literature. That analysis Trinity detonation? indicated a northeast-moving plume with little to no • contamination south of the detonation site. How many cancers are projected above baseline to have In the 1980s, there were extensive assessments occurred in population groups specified by location, age, conducted for Nevada nuclear testing, prompted by and ethnicity as a result of exposures to Trinity fallout? • the 10-y lawsuit of Allen et al. vs. the United States.To What evidence exists for a multi-generational cancer date, however, there has not been an assessment of risk among New Mexico residents from exposures to Trin- public exposures and health risks from Trinity. ity, and how large might that effect be? Residents of New Mexico have voiced for many • As it is well known that Trinity was relatively inefficient years their concern about lingering effects from the in the fission of its plutonium fuel, can the unfissioned Trinity test. In 2007, Congress requested the National plutonium be accounted for? Where is the unfissioned Cancer Institute (NCI) to provide expert guidance on plutonium, and what are the health implications of that exposures and risks from Trinity. The charge was given contamination? to an internal research group at the NCI who were studying health risks from radioactive fallout The authors of this collection of papers sincerely from nuclear testing in Nevada, Marshall Islands, hope that the information provided herein is found Kazakhstan, and elsewhere. Members of that group to be useful for those interested in the health and en- were also deeply involved in the study of radiation vironmental implications of the Trinity nuclear test. risk following the Chernobyl accident and on quantifying risk from medical radiation, which is STEVEN L. SIMON National Cancer Institute widely used in the U.S. and worldwide. While the NCI National Institutes of Health issued a brief report on Trinity to Congress in 2007, Bethesda, MD there were remaining issues that necessitated further detailed investigation. ■■

The author declares no conflicts of interest. Written work prepared by employees of the Federal Government as part of their official duties is, under the U.S. Copyright Act, a "work of the United States Government" for which copyright protection under Title 17 of the United States Code is not available. As such, copyright does not extend to the contributions of employees of the Federal Government. DOI: 10.1097/HP.0000000000001340 www.health-physics.com 389 Paper

Methods and Findings on Diet and Lifestyle Used to Support Estimation of Radiation Doses from Radioactive Fallout from the Trinity Nuclear Test

Nancy Potischman,1 Silvia I. Salazar,2 Mary Alice Scott,3 Marian Naranjo,4 Emily Haozous,5 André Bouville,6,7 and Steven L. Simon6

Americans. Meat was not commonly consumed in the summer Abstract—The Trinity nuclear test was detonated in south-central in most communities, and if consumed, it was among those aged New Mexico on 16 July 1945; in the early 2000s, the National 11-15 y of age or older who had relatively small amounts of Cancer Institute undertook a dose and cancer risk projection 100-200 g d−1. Most drinking and cooking water came from study of the possible health impacts of the test. In order to covered wells, and most homes were made of adobe, which conduct a comprehensive dose assessment for the Trinity test, we provided more protection from external radiation than wooden collected diet and lifestyle data relevant to the populations living structures. The use of multiple approaches to trigger memory in New Mexico around the time of the test. This report describes and collect participant reports on diet and other factors from the methodology developed to capture the data used to calculate the distant past seemed effective. These data were summarized, radiation exposures and presents dietary and lifestyle data and together with other information, these data have been used results for the main exposure pathways considered in the dose to estimate radiation doses for representative persons of all ages reconstruction. Individual interviews and focus groups were in the main ethnic groups residing in New Mexico at the time of conducted in 2017 among older adults who had lived in the same the Trinity nuclear test. New Mexico community during the 1940s or 1950s. Interview Health Phys. 119(4):390–399; 2020 questions and guided group discussions focused on specific Key words: exposure, population; exposure, radiation; intake, aspects of diet, water, type of housing, and time spent outdoors radionuclide; internal dose for different age groups. Thirteen focus groups and 11 individual interviews were conducted among Hispanic, White, and Native American participants. Extensive written notes and audio recordings aided in the coding of all responses used to derive ranges, prevalence, means, and standard deviations for each exposure variable for INTRODUCTION various age categories by region and ethnicity. Children aged 11–15 y in 1940s or 1950s from the rural plains had the highest THE TRINITY nuclear device was the culmination of the Man- milk intakes (993 mL d−1), and lowest intakes were among 11- to hattan Project to develop the first atomic bomb. The device −1 15-y-olds in mountainous regions (191 mL d ). Lactose intolerance called Trinity was tested at the Alamogordo Bombing and – rates were 7 71%, and prevalence was highest among Native Gunnery Range in south-central New Mexico on 16 July 1945. The communities near the Trinity test site were not 1Office of Dietary Supplements, Office of the Director, National In- given any notice or warning about the test since it was stitutes of Health, Bethesda, MD; 2Office of Communications and Public classified as top secret. To our knowledge, aside from a Liaison, National Cancer Institute, National Institutes of Health, Bethesda, MD; 3Department of Anthropology, New Mexico State University, Las brief report by the US National Cancer Institute (NCI) to Cruces, NM; 4Special Consultant on Native American communities, Congress in 2007, no research studies have been conducted Honor Our Pueblo Existence (HOPE), Espanola, NM; 5Pacific Institute to assess the health impact of the Trinity nuclear test on for Research and Evaluation: Albuquerque, NM; 6Division of Cancer Epidemiology and Genetics, National Cancer Institute, National Institutes the people of New Mexico. of Health, Bethesda, MD; 7Retired. From 2016 to 2017, the NCI conducted a field study For correspondence contact: Nancy Potischman, Office of Dietary to inform a study to estimate cancer risk from Trinity al- Supplements, National Institutes of Health, 6100 Executive Blvd., Bethesda, MD 20892, or email at [email protected]. ready underway. In collaboration with the NCI radiation (Manuscript accepted 21 April 2020) dosimetrists, a multidisciplinary team was assembled to col- The authors declare no conflicts of interest. 0017-9078/20/0 lect data and derive estimates of lifestyle factors that could Written work prepared by employees of the Federal Government as be applied to the dose assessment methods for the Trinity part of their official duties is, under the U.S. Copyright Act, a "work of test. Information needed included dietary practices, sources the United States Government" for which copyright protection under Title 17 of the United States Code is not available. As such, copyright does not of water, house construction materials, time spent outdoors extend to the contributions of employees of the Federal Government. in summer months, and other factors. It was necessary to DOI: 10.1097/HP.0000000000001303 obtain appropriate lifestyle and dietary input data for the 390 www.health-physics.com Estimating doses from the Trinity nuclear test c N. POTISCHMAN ET AL. 391 populations in New Mexico, including the sizeable proportion individual interviews (Kitzinger 1995). In terms of recollection of Hispanic/Latinos, as well as White and Native American of past events, groups were thought to have improved recol- tribal community members.8 To fully represent the lection compared to individual interviews. geographic and sociological diversity of New Mexico, Described in this paper are the methods and results of information was needed for both urban and rural settings data collection as part of the NCI study to examine poten- and different environment types (e.g., plains, mountains). tial radiation exposures and related health risks across the state Five cancer endpoints are being studied (Cahoon et al. of New Mexico from the Trinity nuclear test. Accordingly, the 2020): colon, lung, active (red) bone marrow, stomach, and objectives of this paper are to (1) describe the methodology thyroid. Generally, the major health risk associated with developed to capture lifestyle data through focus groups exposure to radioactive fallout is that of thyroid cancer and interviews in older adults who had lived in these resulting from internal exposures of the thyroid gland to ra- communities in the 1940s and 1950s and (2) to present dioiodine primarily through consumption of contaminated results obtained from the focus groups and interviews water, milk, other dairy products, and leafy vegetables. To on consumption of a variety of foods as well as lifestyle a lesser degree, consumption of other food products (e.g., parameters that can influence dose. vegetables, fruits, and meats from animals that grazed on MATERIALS AND METHODS contaminated vegetation) could contribute to the radiation dose to human beings. In terms of radioiodine exposure, Primary goals children are especially vulnerable because they often con- The primary goal of the data collection phase of the NCI sume large amounts of dairy products and have small thyroid Trinity study was to acquire diet and lifestyle data that were glands, which implies the energy from the radioactive decay relevant to all strata of the New Mexico population for will be concentrated in a smaller mass. Although there is which dose assessment was contemplated. The strata of the existing information about the date and time of radiation fall- study population were defined to be all combinations of the out exposure for each community as derived from fallout maps primary attributes that might serve to differentiate New (Beck et al. 2020; Bouville et al. 2020), the internal radiation Mexico residents in terms of the doses likely to have been exposure depends, to a large degree, on the consumption rate received. The basis of requiring data on each stratum is of contaminated food products, which generally varies by age. because data that differentiated population groups is needed External radiation exposure depends on other factors, includ- to support a reasonably reliable estimation of radiation dose. ing the magnitude of the local deposition, time spent outdoors, The main attributes that appeared to possibly differentiate and the type of construction materials for the home. In New groups included ethnicity; gender; age (1–4y,5–10 y, Mexico, the predominant construction materials were adobe 11–15 y, and 16+ y); general geographic location in the (earth mixed with water and an organic material such as state (north/south); environment type (also called ecozone), straw) and wood during the time of the test. which included plains, mountains, plains/mountains; and From research in the field of nutritional epidemiology, population density (urban and rural). In the analysis phase it is better to inquire about the past diet than to use current of the collected data from the field, data sets were derived diet as an estimate of a person’s past diet (Willett 2013). for the purpose of dose reconstruction that covered, to the Therefore, it is necessary to inquire about past diet around degree possible, the originally contemplated strata based on the time of the test. Based on similar methods and analyses consideration for the quality of data and the similarity of as part of a dose reconstruction and epidemiological study data by region and ethnicity. of thyroid disease related to nuclear testing in Kazakhstan In preparation for the data collection phase, we con- beginning in 1949 (Schwerin et al. 2010; Drozdovitch et al. ducted an extensive literature review to assess typical foods 2011), focus groups with individuals who were living in New consumed and dietary patterns in New Mexico in the 1940s Mexico in the years close in time to the detonation appeared among Native Americans and Hispanic/Latino populations. to be a useful research approach to address diet and other The review was focused on five main food groups with lifestyle factors. Details of the Kazakhstan study are described relevance to human exposure to radioactive isotopes of in the Discussion. elements that can be transferred via fallout-contaminated Data suggest that older adults benefit from group dis- foods: (1) dairy, (2) meat and organs of large and small cussions of topics from the past. Focus group interviews have animals, (3) plants, (4) fruit, and (5) drinking water. After all been shown to increase participant comfort when individuals potential sources of dietary contamination were reviewed are gathered into homogeneous groups (Kitzinger 1995; and the list of foods was complete, we ranked the foods by Krueger 2000; Lakshman et al. 2000), and focus groups three levels of potential exposure to human beings: high, with older adults generate a broader range of thoughts than moderate, and low. These rankings were based on the apparent frequency of consumption, whether an important 8 The findings in this paper do not apply to the Navajo Nation. pathway of radiation exposure might be involved, or if certain www.health-physics.com 392 Health Physics October 2020, Volume 119, Number 4 at-risk population groups, such as women and infants, differed participants to attend a focus group or individual interview in their consumption pattern. The key foods identified by this and provided details on the time, date, and location for the exercise were subsequently included in the data collection data collection. Senior centers’ directors were instrumental instruments so that information on the consumed amounts in bringing in qualified elders and organizing the timing of could be queried during focus-group meetings or individual the focus groups and individual interviews. interviews. Findings from the literature review and pilot All willing individuals who met the inclusion criteria interviews also indicated that the way of life in New Mexico were invited to participate in the data collection. Pilot testing did not change drastically between the 1940s and 1950s. in 2014 with nine volunteers aged 70-103 y using the indi- vidual interview method demonstrated the willingness and Participant recruitment ability of individuals to engage in recollection of the key in- Eligible participants were aged 70 y or older, had lived formation being queried. In the 2016-2017 field work, efforts in their New Mexico community during the 1940s and/or were made to construct homogeneous focus groups accord- 1950s, and were mentally competent to participate in the focus ing to ethnicity and region based on where participants re- groups or individual interviews. Although many parti- ported living in the 1940s and 1950s. The NCI’s Special cipants currently aged 70–80yweretooyoungin1945to Studies institutional review board (IRB) and the Southwest recollect their lifestyles that year, they were asked to report Tribal IRB approved the study. on their typical behaviors in the 1940s and 1950s, which should adequately represent their behaviors in the time period Focus group and interview procedures of interest. Because dietary recalls and estimates of portion One experienced individual (SIS) moderated all focus size are not reliable for those younger than 8 to 12 y at the time groups and co-facilitated all the logistics for the field work. being recalled (Livingstone and Robson 2000; Livingstone The emphasis of the focus group questions was on dietary et al. 2004), some of the younger participants were asked intakes of members of the family unit at all ages. Instruments to report on activities in the 1950s that they could recall. developed for data collection were translated into Spanish, In consultation with the Albuquerque Area Southwest and the focus group moderator spoke in Spanish or English Tribal Epidemiology Center (AASTEC) and the Albuquerque as needed. The moderator asked each Native American com- Area Indian Health Board (AAIHB), we followed protocols munity for their preferred language, and most of these partic- for obtaining approval of the Southwest Tribal IRB, provided ipants preferred speaking English, although an interpreter information to potentially interested Native American tribes, was available when needed. To aid with recollection of the met with tribal leaders, and obtained Tribal Resolutions, time of year of the Trinity test and to incorporate seasonal which granted permission for the study of a specific tribal changes in food intakes, respondents were queried about dietary group. In addition, we worked with local Native American intakes and activities in the summer and early autumn months. consultants and subject matter experts who facilitated con- In addition, data collection was conducted from July to October tacts with the tribes. Local New Mexican communities, tribal so that seasonal foods would have been recently consumed. leaders, and local Native American institutional review boards Plates, bowls, glasses, and a baby bottle from the 1940s were consulted regarding appropriate recruitment methods, were placed on the table of the focus group to help respon- cultural or linguistic practices and needs, as well as providing dents estimate serving sizes, while on the walls of the room suggestions of eligible and appropriate participants for focus in which the focus group was conducted, archival black and groups or interviews. white photographs were displayed of activities from the 1940s. For Hispanic and White groups, local academic collab- The bowls and cups had markings for different volumes to help orators contacted senior centers and community leaders to gauge amounts. For the focus groups of Native Americans, invite participation in the study and to schedule the focus some of the dishes on the tables were replaced with clay bowls groups. Advertisements were posted in local newspapers in- and cooking pots common in their communities in the 1940s viting possible respondents to join the study. Senior Centers and 1950s. To facilitate discussions and provide context for representing relevant eco-regions in the state also conducted questions, large charts on the walls showed different types outreach to seniors they served and provided input on who of foods and the various age groups of interest (1–4y, in the community might be able to participate. In addition, 5–10 y, 11–15 y, and 16+ y). flyers were posted in the Senior Centers to solicit volun- Each focus group participant was given a booklet show- teers. The most important criteria for eligibility were age ing a range of serving sizes as an additional aid for the esti- and cognitive ability among potential respondents. Local mation of the amounts of foods consumed. We developed academic partners and members of the study team con- the serving size booklet based largely on the Nutrition As- ducted screening interviews with potential participants in sessment Shared Resource Serving Size Booklet used at the the appropriate language to determine their eligibility to par- Fred Hutchinson Cancer Research Center (2000) for nutri- ticipate. The collaborators subsequently invited eligible tion and epidemiologic studies. The main modification was www.health-physics.com Estimating doses from the Trinity nuclear test c N. POTISCHMAN ET AL. 393 to limit the foods included to those of interest in the Trinity as- from small, medium, and large animals; (3) vegetables; sessment. For each food group, the booklet displayed four (4) grains; and (5) wild fruits and berries. color photographs of a sample food item on a standard-sized • Sources of water for drinking and cooking. plate with utensils on the sides and a written indication of • Hours per day spent outdoors in the summer. the portion size. The booklet also included one page with pic- • Type of building materials for the home. tures and hints to help respondents estimate the servings of various foods; for example, the booklet explained that 1 cup The focus group moderator offered detailed probes in the of mashed potatoes was about the size of a fist. A drawing form of open-ended questions related to each subject area. of a glass with delineations of 3, 6, 9, and 12 ounces was de- Then the moderator asked participants to estimate frequency signed to assist with estimates of fluid intake. and amounts of the foods consumed at different ages (1–4y, Before each focus group discussion, the moderator and 5–10 y, 11–15 y, and 16+ y). One study team member re- data collection team members greeted participants. The corded answers from participants into each cell on the wall moderator paraphrased the consent form and then gave par- chart. Because participants in the first focus groups did not ticipants additional time to read and sign the consent to in- look at the charts at the appropriate times, the moderator be- dicate their willingness to participate. Members of the study gan pointing to the food column at the beginning of each team worked with each participant to complete a form indi- new set of questions, and the note-takers documented an- cating family members that were living together in the same swers on spreadsheets or in written notes but not on the wall home with the participant during the1940s and 1950s, such chart itself. as parents, siblings, grandparents, and other extended fam- To the extent possible, answers were sought from each ily members. The individual interviews were for subjects participant for every question. Because focus groups were who preferred one-on-one interviews rather than the group limited to 2 h, there were occasions where some information sessions. Individual interview respondents used a printed was not collected. As a result, priority was given in the fol- sample-serving-size color booklet to determine food serv- lowing order: milks and other dairy, sources of water, the ing amounts. The responses from the individual interviews most commonly consumed meats in that group, time spent addressed community level behaviors and some quantita- outdoors, and building materials of homes. tive information about food amounts. Whenever appropri- Participants who preferred not to join a focus group or ate, the focus group sessions would open and close with a who found the timing or location of the focus group incon- prayer or blessing and sharing a meal, before or after the ses- venient were offered the option of participating in the study sion, as is common practice in this region. Two note-takers by completing an individual interview, typically conducted in who were members of the study team attended each focus the senior center or another convenient location. The type of group, and they identified each participant by number in their information that was solicited in individual interviews focused recorded notes. on the family of the participant as well as community-level The moderator began each session by asking all partici- information. To the extent possible, an interviewer and a pants for permission to make an audio recording of the discus- note-taker were present, but often the interviewer also had to sion and to use a Livescribe ballpoint smartpen (Livescribe, take notes. As with the focus groups, the interviewer asked Inc., San Francisco, CA) with an embedded computer and dig- about the participant’s diet, lifestyle, and home in the 1940s ital audio recorder. There were no recordings made if a single and 1950s and in 1945, if appropriate. The interviewer asked participant objected, which resulted in three focus groups with- about the participant’s recollection of the Trinity nuclear test, out audio recordings. Each participant then performed a card activities for summer celebrations and feasts, animals, and sorting exercise wherein 10 major food items (milk; cheese; home gardens. In addition to the questions asked of focus meat from large, medium, and small animals; dried meats; groups, the interviewer also asked about community character- fruits; leafy green vegetables; root vegetables; and grains) were istics in the 1940s and 1950s, such as the percentage of homes put in order from highest to lowest frequency of consumption that used water from each type of source, how milk was stored, by their families in the 1940s and 1950s. sources of foods, and degree of access to grocery stores, mar- The moderator continued with open-ended questions kets, or trading posts. Quantified estimates of intakes and about daily life at the time of the Trinity test and personal other relevant information from these individual interviews knowledge of the Trinity nuclear test. Next, the moderator were merged with the corresponding focus group’sdatafor asked questions about the following topics: a given community. Much of the information from the partic- ipant provided context for the differences in current lifestyle • How long women breastfed their babies. compared with that of the time period of interest as well as be- • Levels of consumption of (1) milk and other dairy products ing useful for coding or interpreting the focus group data. (including cows’, goats’,andsheep’s milk and home-made Study personnel offered all participants transportation cheeses); (2) meat (including organ meat and meat jerky) to the meeting location from their homes in addition to www.health-physics.com 394 Health Physics October 2020, Volume 119, Number 4 providing refreshments. Participants received $50.00 to presumed to have similar intakes and exposures to the compensate for their time based on the recommendations Hispanic populations by ecoregion. of social science researchers who had extensive experience conducting community-based participatory research with RESULTS Native American, Hispanic/Latino, and other communities in New Mexico. The senior centers received compensation Between July and October 2017, we conducted 13 fo- for their collaboration and for providing space for the focus cus groups—eight with Hispanic and White participants groups and interviews. and five with Native American participants. Focus group size ranged from 4 to 12, with a mean of 6 individuals per Data analysis group. We also completed 11 individual interviews in vari- We classified the geographic regions in which respon- ous communities, mostly at the same location as the focus dents had lived in the 1940s and 1950s as mostly plains, de- groups and conducted concurrently. Earlier focus groups sert, or mountains; northern, central, or southern regions; provided a greater cultural and linguistic narrative context and whether rural or urban. The urban centers were small that informed subsequent focus groups and interviews. cities that drew from populations living within the commu- Some senior centers did not respond to requests to par- nity or from the neighboring land. ticipate in the study, even after multiple visits by local collab- Data were coded based on the wall charts or spread- orators and NCI team members. Two other senior centers sheets completed during the focus groups or based on the scheduled focus groups, but either no participants arrived text from the note-takers. Spreadsheets were developed or potential participants had decided against the focus group. for each of the exposure topics by two coders. When both At one center, a few of those who decided not to participate in coders had been present at a focus group as note-takers, a focus group did agree to an individual interview. they each developed spreadsheets, which were then com- Both the focus groups and individual interviews began pared and harmonized based on audio recordings and rec- with a food card sorting activity. Each participant was given ollections of those present at the focus group. When only a set of 10 cards with text and pictures of broad food groups one or neither of the coders had been present, they would on each card. The participants used the cards to indicate read notes and listen to audio tapes or the audio from the which foods were the most and least frequently consumed Livescribe ballpoint smartpen, develop the spreadsheets, in their household and if a given food was not consumed and later together decide the best code for each partici- at all. After the card-sorting activity, the moderator contin- pant’s answer if there were discrepancies in the coding. ued the focus group and went through each food category In addition, to confirm or clarify some issues, the coders on the wall charts by age group starting with the youngest, consulted with the focus group moderator and team members and each participant provided their estimates of frequency who were present to obtain the best information for coding and serving size. The large wall charts had grids of age the data. Similar procedures were used to code the individ- groups on the rows and pictures of the food type in the col- ual interviews by two team members, and any disparities umns (e.g., color pictures of pigs, cows, vegetables, legumes, were reconciled. and fruits). The wall charts did work well with setting the Table 1 presents the six datasets that were created by stage. The moderator would often walk toward the charts merging similar data according to ethnicity, ecozone, and and point to each food group and asked about amounts of rural/urban status. Almost all of the regions had information food consumed by age. Participants used their individual gathered from two focus groups and two individual inter- serving-size booklets so that they could see examples of views. Based on discussions with academic consultants in varied serving sizes. Plates, bowls, and glasses were passed New Mexico, the small African American population was around between participants so that they could touch and

Table 1. Datasets applied to population strata for dose reconstruction as derived from merging focus groups and individual interviews. Dataset Region Ethnicitya Number of participants from focus groups and interviews

A Rural/Plains Hispanic 16 B Rural Mountains or Plains/Mountains Hispanic & White 16 C Urban Mountains or Plains/Mountains Hispanic & White 23 Rural/Plains White D 12 Urban/ Plains Hispanic & White E Mountains & Plains/Mountains Native American 7 F Plains Native American 12

aAfrican Americans were assumed to have similar habits to the Hispanics in each region.

www.health-physics.com Estimating doses from the Trinity nuclear test c N. POTISCHMAN ET AL. 395 feel each item and continue to become familiar with serving and White communities and from 33% to 71% in Native sizes on actual serving dishes. After the first focus groups, American communities. Cheese consumption was very answers from respondents were documented by entering low, although participants did report the homemade soft the data into spreadsheets or through long hand notes and cheese that would be consumed on special occasions. Some not entered onto the wall charts. The variety of answers sug- participants reported buying hard cheeses from a store and gested that participants were fairly independent in their as- information about this source of cheese was not used in sessments. To ensure the focus groups were balanced, the the dose estimations (i.e., coded as zero). Wild greens were moderator made sure to avoid having one person influence consumed during the day outdoors and were difficult to or dominate the responses provided by the rest of the group. quantify, but wild or garden greens collected for a meal were This ensured that each participant gave independent intake reported as quantity of cooked greens consumed per age responses. category. Many respondents reported gardens with a variety Estimated mean intakes and standard deviations for the of vegetables, including squash, carrots, turnips, corn, green groups are presented in Table 2. The dosimetrists were pro- beans, peas, potatoes, and legumes. It was not possible to vided with these data as well as the minimum and maximum obtain information on all these items, but it appeared that for each age group cell. The lowest intakes of cows’ milk they were consumed in small amounts and that many repre- were in the urban (small city) mountain communities in sented a minor pathway of exposure. Based on information Dataset C and in the mountains/plains communities of obtained in focus groups, it was estimated that these vegeta- Dataset E. The communities with the highest consumption bles were consumed at about 20% the rates observed for amounts (Datasets A and D) were largely rural and with leafy greens. Wild berries, plums and other fruit were con- many ranching families that owned milking cows or re- sumed during the day while outdoors, and participants es- ceived milk from neighbors. Intakes of milk from other an- timated consumption quantities for each type of fruit. imals were low for all respondents, although a few with Respondents often mentioned having very limited economic lactose intolerance reported consuming goats’ milk. Lactose resources. They explained how they grew their own food and intolerance rates ranged from 7% to 24% among Hispanic had simple diets of beans (legumes), corn, and homemade

Table 2. Mean (standard deviation) consumption data in mL d−1 or g d−1 by age group. Consumption rate (mL d−1 or g d−1) Foodstuff Age, y A B C D E F

Cows’ milk 1–4 273 (98) 246 (154) 117 (75) 538 (257) 118 276 (225) 5–10 617 (175) 471 (300) 172 (179) 531 (319) 237 395 (252) 11–15 695 (264) 533 (386) 191 (201) 993 (418) 0 629 (373) 16+ 592 (424) 182 (256) 182 (256) 615 (540) 0 392 (288) Cows’ 1–4 37(45) 37(45)a 0000 cheese 5–10 49 (20) 49 (20) a 13 (2.4) 0 0 7 (16) 11–5 50(18) 50(18)a 21 (7) 0 0 11 (22) 16+ 50 (18) 50 (18) a 15 (2) 0 0 12 (23) Beef/large 1–4 40 (6) 18 (15) 38 (16) 42 (33) 1.2 (2) 3 (4) animals 5–10 78 (17) 39 (27) 71 (72) 88 (41) 9 (13) 5 (9) 11–15 206 (87) 54 (28) 99 (116) 204 (86) 12 (14) 11 (10) 16+ 77 (93) 56 (32) 107 (125) 198 (64) 15 (17) 9 (16) Mutton/pork (small 1–4 4 (6) 0 8 (6) 22 (61) 0 11 (8) animals for 5–10 5(6) 0 13(10) 43(74) 0 14(10) Data Set F) 11–15 8 (9) 53 (71) 16 (9) 109 (230) 0 26 (21) 16+ 8 (9) 53 (71) 8 (6) 200 (210) 0 23 (19) Leafy 1–4 52 (35) 11 (21) 77 (67) 31 (21) 45 (64) 39 (52) vegetables 5–10 122 (104) 23 (42) 220 (120) 90 (0) 71 (51) 88 (67) 11–15 217 (174) 45 (42) 251 (121) 103 (81) 122 (57) 170 (104) 16+ 263 (210) 143 (38) 290 (110) 165 (68) 193 (62) 279 (188) Fruit and 1–4 28(20) 0 547(410) 28(20)a 5 (5) 109 (87) berries 5–10 44 (47) 0 722 (344) 44 (47) a 9 (9) 173 (184) 11–15 76 (77) 0 1034 446) 76 (77) a 14 (11) 218 (165) 16+ 41 (36) 0 773 (337) 41 (36) a 20 (17) 284 (238)

aData imputed from similar group (dataset A).

www.health-physics.com 396 Health Physics October 2020, Volume 119, Number 4 flour and corn tortillas. Most of New Mexico is at high Table 4. Percentage of participants reporting types of housing materials. altitude, and the population experiences seasonal differences Construction material of participant homes (%) in food intake. A seasonal subsistence and barter-based diet Data set Adobe Wood Other was common. A (Rural Plains) 78 11 11 Breastfeeding was very common in all communities, B(RuralMountains) 91 9 − with rates of 100% in most communities and of 82% in only C (Urban Mountains) 73 27 − one community. Although many men had difficulty reporting D (Rural & Urban Plains) 50 17 33 on the typical duration of breastfeeding in their communities, E (Mountains &Plains) − 100 − most participants reported that breastfeeding lasted at F(Plains) 100 −− least 12 mo and often up to 24 mo (Table 3). Respondents also reported that lactating mothers did not eat special di- ets, and babies were introduced to cows’ milk only after although they did use rain water for bathing, watering the gar- lactation ended. den, and washing clothes. The moderator deliberately probed As Table 4 shows, most community members lived in multiple times about the collection and use of rainwater to en- adobe homes during the study period, although one Native sure there was no misunderstanding and to obtain specific American community in a mountains/plains region built information about this potential source of contamination. their homes with wood (Dataset E). In the preliminary in- Often, the focus groups began with discussions of con- vestigation, pica emerged as a possible exposure route for trasts between current daily life and daily life during the study participants in adobe homes. In response to this possibility, period. The foods were largely from a local source, and some we asked questions about primary exposure to the external foods were no longer available. Many groups discussed the surfaces of homes (licking the outer walls of their adobe fact that resources were scarcer during the study period than houses after a fresh external mud plater application). Intake they are now, how they managed with limited resources, and of adobe surface was determined to be minimal. that nothing was wasted. Participants reported that each fam- Respondents were asked about the amount of time ily member ate all the food provided on the plate, unlike their spent outdoors during the summer when they had no school. current grandchildren who have many food options and often Most individuals reported being outside from sunrise to do not finish all the food on their plates. All family members sunset in summer, except for some young children and girls, helped with needed family activities, such as working in the who were busy with indoor activities or chores. Respon- fields, house chores, and taking caring of siblings. dents from three communities reported that able children One of the hallmarks of an effective focus group is cre- and adults would sometimes sleep on the roof of their home ating open communication that is free of judgement and feels at night because the air was cooler than inside. This was like a conversation instead of answering a series of questions. then specifically queried in subsequent focus groups but By comparison, an unsuccessful focus group is one in which was not a common habit in other areas. there is no open communication environment that allows par- The large majority of communities reported having ticipants to share and contribute to discussions. The success well water for drinking and cooking, and many wells were of the focus groups and individual interviews were demon- covered and had windmills. Buckets of water were brought strated by the level of detail participants provided about life into the home and covered until the water was used. Animals in New Mexico when they were growing up and their owned by the families would drink fresh water from the ace- willingness to answer questions. For instance, participants quias (open water ditches or canals). One Native American were able to recall the locations and types of wild spinach community used water piped in from springs to parts of the and fruits they picked and ate during the summer months. village, and families would bring buckets of this water They described the location of the wells in their community home for drinking and cooking. Almost no communities and even the name of the family that lived next to the well. reported collecting rain water for cooking and drinking, Among participants that drank cow's milk, they shared whether the milk came from a local dairy or a cow that Table 3. Breastfeeding duration by data set. lived in their community. Participants that drank milk from a goat shared which members of their family were Data set Mean number of months (standard deviation) not able to digest cow’s milk. The variety of reported intakes A (Rural Plains) 10.5 (4.5) from participants across food types further demonstrates that B (Rural Mountains) 16 (4.4) all participants were able to provide individual responses. At C (Urban Mountains) 13.9 (5.2) the end of each focus group, participants shared how much D (Rural & Urban Plains) 13.9 (5.5) they enjoyed talking about what life was like when they E (Mountains and Plains) 12 were growing up. Data from all focus groups and individual F (Plains) 21 (5.6) interviews were used.

www.health-physics.com Estimating doses from the Trinity nuclear test c N. POTISCHMAN ET AL. 397 Data from appropriate earlier studies were used to de- asking about frequency of a standard portion, and using rive uncertainty factors for use with the dose estimations detailed methods for estimating portion size increases cor- to characterize what we believe to be the credible range of relations between original and recalled diets. These results mean values of consumption rates and lifestyle factors. Al- suggest methods used in the focus groups are consistent though no previous studies that we relied upon for estimat- with suggested practices for obtaining information on past ing uncertainty used focus groups or had such a long recall diets. In addition, having had very limited variety in diets period, the mean differences and standard errors from studies in the time period of interest may have helped with the rec- with 10 or more years between recorded and recalled data ollection for the items being queried. were used to derive best estimates of the uncertainty factors The specific amounts of foods that participants re- (Friedenreich 1992). We applied smallest uncertainty factors ported were representative of the typical diet for children to the data from the best focus groups, based on the modera- of specific ages and ethnicities with which they were famil- tor, nutritional expertise, and evaluation of spreadsheet data. iar. We based our methodology on previous research in Then more conservative uncertainty factors were derived for which we compared self-report against focus group data in the focus groups that were deemed less reliable. Further dis- a similarly exposed sample. In 1998, in an epidemiological cussion of the use of uncertainty factors in the dose estima- study of individuals exposed to radiation from nuclear test- tion is provided in Simon et al. (2020). ing during their childhood and adolescence in Kazakhstan between 1949 and 1962, the NCI, the Semipalatinsk State DISCUSSION Medical Academy, and the Kazakh Research Institute for Radiation Medicine and Ecology similarly collected food The data collection team’s observations suggested that and behavior information using a basic questionnaire. In that having open discussions about life during this period ap- study, the outcome of interest was the effect of radiation expo- peared to have aided recollection among all participants sure on the development of thyroid nodules. In 2007, the NCI within each focus group. Engaging in collective discussions conducted focus groups with older adults in Kazakhstan to during the focus groups seemed instrumental in setting the obtain detailed information on milk consumption and other stage and potentially improving recollections. Participants exposures of interest in 1949 and early 1950s (Schwerin builtoneachother’s memories, which allowed most partic- et al. 2010). In that study, the response data from the focus ipants to report on summer practices as well as details of groups improved the analyses of the thyroid radiation dose how foods were prepared and consumed. The team’s prac- estimates for participants from the previous epidemiologi- tice of translating all instruments, speaking in the preferred cal study compared to the self-reports on the basic question- language (Spanish or English), or even simply asking the naire (Drozdovitch et al. 2011). Based on lessons learned in participants which language they preferred likely aided in that study, we used similar strategies and methods in the their engagement in the activities. Other manners in which Trinity study. the team demonstrated respect for the participants’ culture Previous research indicates that the card sorting method included opening with a blessing, sharing a meal, and use used in this study could facilitate memory recall among se- of time-period-appropriate dishes, clay bowls, and cooking nior citizens (Craik et al. 1990). In our study, participants pots. Collecting consistent answers across similar but sepa- would often talk out loud to themselves as they sorted rate focus groups suggested some indirect validity and their individual set of cards and comment on the types of should improve the estimation of dose, also confirming that foods. For example, participants quickly identified which the attention to culture was effective in all groups who en- foods they did not eat and took more time to sort through gaged in this research. the foods that they consumed more often. Performing the Review of previous studies on recall of the past diet, card sorting before engaging in the focus group discussion particularly distant past, revealed several consistent conclu- appeared to help orient the participants to the time period sions (Willett 2013; Friedenreich 1992), although no studies of interest and perhaps reduce the influence that may have were reviewed that had a 70-y time-lapse. More reliable re- resulted from fellow participants. As a result, imple- call was observed for studies that used interviews rather menting the card-sorting approach afforded a reliable way than self-administered questionnaires and that recall is most for participants to remember which foods they ate most often reliable for foods eaten rarely or with a high stability over and focus on the time period of interest. time. Although advancing age may be associated with de- Providing participants with serving bowls, plates, cups, creased ability to recall past dietary intakes, long-term memory and utensils from the 1940s and 1950s offered a tactile way often remains intact despite loss of short-term memory in older to remember the foods they ate at the time. As participants adults (Krall et al. 1988). Recollections may be improved by engaged in the discussion, they often held a bowl or plate asking about specific foods rather than grouping food items and compared it with the types of plates their families had together, asking about individualized portion size is superior to used. Many participants reported having used the same styles www.health-physics.com 398 Health Physics October 2020, Volume 119, Number 4 as those on the focus group table or in the pictures on the wall. group, we did not measure household consumption, and none Furthermore, participants reported on community celebra- of the surveys conducted during our study period collected tions, which were highlights of the summer and featured information from the unique ethnic and racial groups unique foods (e.g., meat from cow or pig) that were shared populating New Mexico or similar surrounding states at and involved preparation, planning, and collection of re- the time. sources. The fact that focus group participants were able Although the main purpose of the initial literature re- to recollect summer activities and lifestyle differences helps view was in preparation for data collection, it later indirectly provide confidence in the many estimates, particularly for validated data collected in our pilot testing, focus groups, time spent outdoors for each age group. and interviews. For example, water sources, types of feasts and celebrations, and types of foods consumed during the Validations, limitations, and uncertainty summer in our study were similar to those found in the liter- Given that the year of exposure (1945) is more than ature review and in pilot testing. Moreover, similarities were seven decades in the past, there are clearly limitations of found in the responses both in the one-on-one interviews our data collection, issues of data validation, and uncer- and the focus groups. The same types of foods were con- tainty that require discussion. One important point is to rec- sumed among similar ethnic groups across the state of ognize that the data we collected are for a risk projection New Mexico. Picking a flowering wild plant commonly study rather than an individual dose assessment and, accord- called “cota” or thelesperma filifolium and using it to make ingly, are not intended to represent any identifiable individ- tea was a common practice among all Native Americans. ual but, rather, represent typical behaviors of the identified Whites consumed more milk in comparison with Native group. In some instances, certain types of information were Americans and Hispanics. Collected exposure data were found to be more difficult for participants to report on com- recalled with various degrees of difficulty. There is uncertainty pared to other information. Hence, the reliability of reported of the reported milk consumption rates, though the consistency values is not necessarily equal across all food types and life- within and across focus groups from the same region style parameters. While it might be possible to argue against supports validity. Some groups had high intakes because specific data that we report, the values we present can be at- of the availability of milking cows, and low intakes were tributed to the participants in our study and for that reason related to the reverse. The reporting on lactose intolerance are believed to have a firm basis in genuineness. Moreover, seems quite reliable because participants knew exactly because questions in the focus groups and interviews were who in the family could or could not have consumed fresh phrased to elicit responses about families and general com- milk. Therefore, these prevalence estimates can be used munity members, the responses obtained are expected to with confidence. Intakes of meats were low in most cases have a degree of representativeness of the communities and clearly reported, so these estimates should have good at large. reliability as well. Intakes of leafy green vegetables were We had originally planned to compare the collected recollected well, but the participants had more difficulty data to appropriate national survey data as a means of vali- quantifying fruit intakes. Many participants had little difficulty dation, but we found no comparable data from such surveys reporting the prevalence and duration of breastfeeding in their conducted during the time period of interest (Tippett et al. families, although others could not provide any information 1999). Specifically, there was no survey on diet that isolated on this topic. Almost all participants easily reported on the the state of New Mexico during our study period, although the sources of water their families used, building materials of 1955 survey data (USDA 1957) included 11 diverse western their houses, and time spent outdoors in summer. These states that spanned from the south to the north of the data are presumed to have been reported with the greatest country. There are household-level dietary data collected accuracy of any of the questions, suggesting that they can in a nonrepresentative nationwide study of homemakers in be used for the dose reconstruction with high level of 1942 (USDA 1944), which queried about quantities of listed confidence. As discussed earlier, uncertainty factors were foods used by the household the previous week. This “7-day derived for each of the data sets depending on the subjective food list” method was used in a nationally representative assessment of reliability by the moderators and previous survey in 1955 (USDA 1957). Based on the average of studies in the literature. data from the 1942 and 1955 surveys, household milk con- −1 −1 sumptionaveraged456mLd for urban, 452 mL d for rural CONCLUSION non-farm, and 870 mL d−1 for rural farm communities. These data are consistent with our findings of higher intakes in The methods and logistics applied as described appeared the rural farm communities. However, we could not use to have worked well in that they resulted in clear engagement the 1942 and 1955 survey data to derive measures of valid- of the participants and ultimately in delivering data useful for ity for our data because our findings differed greatly by age the dose estimations. Effective methods used by the study www.health-physics.com Estimating doses from the Trinity nuclear test c N. POTISCHMAN ET AL. 399 included enlisting the help of Senior Centers and interested Cahoon EK, Zhang R, Simon SL, Bouville A, Pfeiffer RM. Native American tribes, screening participants, involving Projected cancer risks to residents of New Mexico from exposure to Trinity radioactive fallout. Health Phys 119: local consultants, and the approaches used in the focus 478–493; 2020. groups. The props and pictures were of interest to the partic- Craik FIM, Morris LW, Morris RG, Loewen ER. Relations between ipants and helped with orienting them to the distant past as source amnesia and frontal lobe functioning in older adults. evidenced by their comments. The audio recordings and in- Psychol Aging 5:148–151; 1990. Drozdovitch V,Schonfeld S, Akimzhanov K, Aldyngurov D, Land dependent coding of the data helped ensure reliable esti- CE, Luckyanov N, Mabuchi K, Potischman N, Schwerin MJ, mates based on the information provided by participants. Semenova Y, Tokaeva A, Zhumadilov Z, Bouville A, Simon Much of the data collected from the individual interviews SL. Behavior and food consumption pattern of the population were similar to those collected from the focus groups, and exposed in 1949-1962 to fallout from Semipalatinsk nuclear test site in Kazakhstan. Radiat Environ Biophys 50:91–103; 2011. this overlap adds confidence in the reliability of the data, Fred Hutchinson Cancer Center. Sample serving size booklet [online]. particularly at the community level. Although some focus 2000. Available at https://sharedresources.fredhutch.org/content/ groups were more reliable than others, grouping the data sample-serving-size-booklet. Accessed 8 January 2019. by ethnicity and ecoregion helped provide more stable esti- Friedenreich CM, Slimani N, Riboli E. Measurement of past diet: review of previous and proposed methods. Epidemiol Rev 14: mates of the exposure-related variables, and the standard 177–196; 1992. deviations provided information on the variability in the an- Kitzinger J Qualitative research. Introducing focus groups. Br Med J swers. Finally, similarities in answers within region across 311:299–302; 1995. focus groups and individual interviews was useful and im- Krall EA, Dwyer JT, Coleman KA. Factors influencing accuracy of dietary recall. Nutr Res 8:829–841;1988. proved our confidence in the estimates. The data collected Krueger R Focus groups: a practical guide for applied research. and presented here are further summarized in the context Thousand Oaks, CA: Sage Publications; 2000. of development of exposure models (Bouville et al. 2020) Lakshman M, Charles M, Biswas M, Sinha L, Arora NK. Focus and are subsequently used to derive estimates of radiation group discussions in medical research. Indian J Pediatr 67: 358–362; 2000. dosesacrossthestate(Simonetal.2020).Thecollected Livingstone MBE, Robson PJ. Measurement of dietary intake in data are of the best quality possible given the long passage children. Proc Nutr Soc 59:279–293; 2000. of time and sparse published information on diet and life- Livingstone MB, Robson PJ, Wallace JM. Issues in dietary intake style in New Mexico in the mid-1940s. assessment of children and adolescents. Br J Nutr 92(suppl): S213–S222; 2004. Acknowledgments—This research was supported primarily by the Intramural Schwerin M, Schonfeld S, Drozdovitch V, Akimzhanov K, Research Program of the National Cancer Institute with partial support from Aldyngurov D, Bouville A, Land C, Luckyanov N, Mabuchi K, the Intra-Agency agreement between the Radiation Nuclear Countermeasures Semenova Y, Simon S, Tokaeva A, Zhumadilov Z, Potischman Program of the National Institute of Allergy and Infectious Diseases with the N. The utility of focus group interviews to capture dietary National Cancer Institute, NIAID agreement #Y2-Al-5077 and NCI agreement consumption data in the distant past: dairy consumption in #Y3-CO-511. The authors wish to acknowledge the contributions from past Kazakhstan villages 50 years ago. J Dev Orig Health Dis and present Trinity study team members: Ruth Pfeiffer, Elizabeth Cahoon, 1:192–202; 2010. Lauren Houghton, Cheryl Deaguiar, Abigail Ukwuani, Kayla Myers, Jessica Lopez, and the Southwest Tribal Epidemiology Center, particularly Kevin Simon SL, Bouville A, Beck HL, Melo DR. Estimated radiation doses received by New Mexico residents from the 1945 Trinity English. We appreciate input we received from public interest and expert – groups in New Mexico including Honor Our Pueblo Existence (HOPE) nuclear test. Health Phys 119:428 477; 2020. and the Tularosa Basin Downwinders Consortium. We are particularly Tippett KS, Wilkinson Enns C, Moshfegh AJ. Food consumption grateful to the many study participants across the state of New Mexico surveys in the US Department of Agriculture. Nutr Today 34: for the insights, time, and commitment, including the Native American Tribal 33–46; 1999. leaders and tribal members who participated, and the numerous staff, US Department of Agriculture. Family food consumption in the management, and volunteers at many Senior Centers that participated. United States. Washington, DC: USDA; Miscellaneous Publi- cation No. 550; 1944. REFERENCES US Department of Agriculture. Dietary levels of households in the United States. Washington, DC: USDA; Household Food Con- Beck HL, Simon SL, Bouville A, Romanyukha A. Accounting for sumption Survey, Report No 6; 1957. unfissioned plutonium from the Trinity atomic bomb test. Willett W Recall of remote diet. In: Willett W,ed. Nutritional epide- – Health Phys 119:504 516; 2020. miology. New York: Oxford University Press; 2013: 140–149. Bouville A, Beck HL, Thiessen KM, Hoffman FO, Potischman N, Simon SL. The methodology used to assess doses from the first nuclear weapons test (Trinity) to the populations of New Mexico. Health Phys 119:400–427; 2020. ■■

www.health-physics.com Paper

The Methodology Used to Assess Doses from the First Nuclear Weapons Test (Trinity) to the Populations of New Mexico

André Bouville,1 Harold L. Beck,2 Kathleen M. Thiessen,3 F. Owen Hoffman,3 Nancy Potischman,4 and Steven L. Simon1

Key words: dose assessment; fallout; Trinity test; New Mexico Abstract—Trinity was the first test of a nuclear fission device. The test took place in south-central New Mexico at the Alamogordo Bombing and Gunnery Range at 05:29 AM on 16 July 1945. INTRODUCTION This article provides detailed information on the methods that were used in this work to estimate the radiation doses that were THE TRINITY nuclear test was the culmination of the Manhattan received by the population that resided in New Mexico in 1945. Project that began in 1942 to develop the atomic bomb. The The 721 voting precincts of New Mexico were classified according to ecozone (plains, mountains, or mixture of plains and mountains), nuclear device that became known as Trinity was designed and size of resident population (urban or rural). Methods were and fabricated at the Los Alamos Scientific Laboratory in developed to prepare estimates of absorbed doses from a range of northern New Mexico and tested in south-central New 63 radionuclides to five organs or tissues (thyroid, active marrow, Mexico at the Alamogordo Bombing and Gunnery Range stomach, colon, and lung) for representative individuals of each voting precinct selected according to ethnicity (Hispanic, White, at 05:29 AM on 16 July 1945. Trinity was the first test of Native American, and African American) and age group in 1945 a nuclear fission device ever and resulted in the first (in utero, newborn, 1–2y,3–7y,8–12 y, 13–17 y, and adult). nuclear explosion. The device simulated the Fat Man type Three pathways of human exposure were included: (1) external plutonium implosion device used less than 1 month later in irradiation from the radionuclides deposited on the ground; (2) inhalation of radionuclide-contaminated air during the passage of the bombing of Nagasaki, Japan. the radioactive cloud and, thereafter, of radionuclides transferred For the first time, a comprehensive dose assessment of (resuspended) from soil to air; and(3)ingestionofcontaminated the radiation doses to New Mexico residents resulting from water and foodstuffs. Within the ingestion pathway, 13 types the detonation of the Trinity test has been prepared (Simon of foods and sources of water were considered. Well established models were used for estimation of doses resulting from the three et al. 2020). The overall goals and purpose of this work pathways using parameter values developed from extensive literature included developing and documenting dose assessment review. Because previous experience and calculations have shown models for the target populations, providing the necessary that the annual dose delivered during the year following a nuclear data for those models, as well as the means to evaluate test is much greater than the doses received in the years after that first year, the time period that was considered is limited to the dosimetric uncertainty for those target populations. first year following the day of the test (16 July 1945). Numerical Because the NCI Trinity study is a risk projection study estimates of absorbed doses, based on the methods described in (see Cahoon et al. 2020), it is only necessary to estimate this article, are presented in a separate article in this issue. doses to representative persons in subgroups in which the – Health Phys. 119(4):400 427; 2020 dose might be differentiated. This contrasts with other dose reconstructions in which doses must be estimated to 1National Cancer Institute, Bethesda, MD; 2US Department of Energy (retired); 3ORRISK, Oak Ridge, TN; 4National Institutes of Health, identifiable persons. Bethesda, MD. For our purpose, we have defined the primary target The authors declare no conflicts of interest. populations, or strata, for dose estimation to be combina- For correspondence contact: André Bouville, 5450 Whitley Park Terrace, Bethesda, MD 20814 or email at [email protected]; tions of location, ethnicity, and age. Other determinants of [email protected]. dose [e.g., environment type (ecozones of plains, moun- (Manuscript accepted 17 June 2020). 0017-9078/20/0 tains, etc.) and population density (e.g., urban and rural), Written work prepared by employees of the Federal Government as as discussed] are modifiers of the basic strata. part of their official duties is, under the U.S. Copyright Act, a “work of Consistent with our goal of estimating doses to represen- the United States Government” for which copyright protection under Title 17 of the United States Code is not available. As such, copyright does not tative persons, we provide a strategy and means for estimating extend to the contributions of employees of the Federal Government. uncertainty to the same representative persons. While the form DOI: 10.1097/HP.0000000000001331 of the methods for estimating uncertainty is like more detailed 400 www.health-physics.com c A. BOUVILLE ET AL. 401 studies of individual doses, the purpose here was only to (Ng et al. 1990); (2) the US Department of Health and Human bound the doses to the representative- person doses. Services (US DHHS) assessment of the doses and risks from fallout in the United States from tests conducted worldwide EXPOSURE PATHWAYS CONSIDERED (US DHHS 2005); and (3) the review by the United Nations Scientific Committee on the Effects of Atomic Radiation Estimates of absorbed doses from a range of radionu- (UNSCEAR) of the doses arising from fallout at the global clides were prepared for representative individuals (I) of the scale (UNSCEAR 1993). The exact composition of the fis- 721 precincts (L) of New Mexico classified according to 2 sion products generated by a nuclear weapons test depends ethnicity [Hispanic (H), White(W), Native American (NA), on the specific fissile nuclide (235U, 239Pu, or 238U) and on and African American (AA)], age group in 1945 (in utero, the neutron energies involved. Therefore, the composition – – – – newborn, 1 2y,3 7y,8 12 y, 13 17 y, and adult), with some of the fission products generated by the Trinity test cannot distinctions made for ecozone [plains (P), mountains (M), be expected to be the same as those for the tests considered or mixture of plains and mountains (P/M)], and population in the other studies. However, the same fission products are density [urban (U) or rural(R)]. For all representative generated in either a fission or a thermonuclear test because individuals, absorbed doses were estimated for five organs thermonuclear tests are triggered by the explosion of a fis- or tissues (thyroid, active marrow, stomach, colon, and sion device, and differences in the distribution of the fission lung). These specific organs and tissues were selected products are mainly observed for products with low fission because they are expected to give rise to the largest yield. The selection of 63 radionuclides was deemed to be numbers of cancers resulting from radiation exposure to large enough to include all radionuclides of some impor- fallout (Simon et al. 2010a): (1) the thyroid gland tance. Analysis of the dose distributions for the populations concentrates radioiodine, which induces thyroid cancer, of three representative voting precincts in New Mexico the predominant health effect investigated in epidemiological presented in a companion paper (Simon et al. 2020) shows studies related to fallout (US DHHS 2005); (2) irradiation of that only 38 of the selected radionuclides contributed to 95% active marrow is expected to increase the risk of leukemia, of the dose resulting from either external irradiation to all another health effect investigated in epidemiological studies; organs and tissues of the body or from internal irradiation to (3) stomach and colon are highly exposed after ingestion of any of the five organs or tissues that were considered. fallout because most of the fission products are highly Among the 63 selected radionuclides, 54 have radioac- insoluble; and (4) lung is for many fission products the most tive half-lives that are less than 3 mo, and only nine have ra- exposed organ following inhalation of fallout. dioactive half-lives longer than 9 mo. Previous studies of Three pathways of human exposure were included: (1) NTS fallout have shown that as a result of the preponderance external irradiation, arising mainly from the radionuclides of short-lived radionuclides most of the doses from external deposited on the ground, and for a small part from radionuclides irradiation within a few 100 km of the test site are delivered in the passing cloud; (2) inhalation of radionuclide-contaminated during a few months following a nuclear test (Beck 2005), air during the passage of the radioactive cloud and, thereaf- while the annual doses from internal irradiation are much ter, of radionuclides transferred (resuspended) from soil to greater in the year following the test than in any subsequent air; and (3) ingestion of contaminated water and foodstuffs. year (Simon et al. 2010b). Thus, the doses from Trinity were To the extent possible, well established models were used to calculated for only the first year following the day of the test calculate the doses resulting from those three pathways. (16 July 1945) at all precincts and for all three exposure path- More than 150 radionuclides were produced in the Trin- ways under consideration (external irradiation, inhalation, ity test. The 63 radionuclides that were selected for the estima- and ingestion). These first-year doses also were used in the tion of the doses from internal irradiation are listed in Table 1. risk analysis, and the very small doses that were received af- These radionuclides, consisting mostly of fission products, are ter the first year were taken to be negligible. essentially the same as those previously selected for the study of doses and risks in the Marshall Islands, where 66 tests, ABSORBED DOSES FROM EXTERNAL many of them thermonuclear, were conducted between 1946 IRRADIATION and 1958 (Simon et al. 2010a). The list of 63 selected radionu- clides is also generally consistent with those considered in The absorbed doses from external irradiation, Dext(I, other important fallout studies: (1) the Off-Site Radiation L), in mGy, to representative individual I in precinct L were Exposure Review Project (ORERP), which was tasked to calculated using the following equation: estimate the doses from fallout resulting from the nuclear KIðÞ DextðÞ¼I; L ½ððÞþXoutðÞL Tout ðÞI ðÞXin ðÞL TinðÞI 1Þ weapons tests that were conducted at the Nevada Test Site 24

2The research findings in this paper do not explicitly apply to the people of where: the Navajo Nation. - www.health-physics.com 402 Health Physics October 2020, Volume 119, Number 4 Table 1. Lista and radioactive half-lives of the radionuclides considered in the study.

55Fe (2.7 a) 93Y(10h) 117Cd (2.5 h) 133I(21h) 147Nd (11 d) 60Co (5.3 a) 95Zr (64 d) 117mIn (2.0 h) 135I (6.6 h) 149Pm (53 h) 64Cu (13 h) 95Nb (35 d) 121Sn (27 h) 137Cs (30 a) 149Nd (1.7 h) 77As (39 h) 97Zr-97mNb (17 h) 125Sb (2.8 a) 139Ba (83 min) 151Pm (28 h) 83Br (2.4 h) 99Mo (66 h) 127Sn (2.1 h) 140Ba (13 d) 153Sm (46 h) 88Rb (18 min) 99mTc (6.0 h) 127Sb (3.9 d) 140La (1.7 d) 237U(6.8d) 89Sr (51 d) 103Ru-103mRh (39d) 129Sb (4.4 h) 141La (3.9 h) 239Np (2.4 d) 90Sr (29 a) 105Ru (4.4 h) 129Te (70 min) 141Ce (33 d) 239Pu (24,000 a) 90Y(64h) 105Rh (35 h) 131mTe (30 h) 142La (91 min) 240U(14h) 91Sr (9.6 h) 106Ru-106Rh (370 d) 131I (8.0 d) 143Ce (33 h) 240mNp (7.4 min) 91mY(50min) 109Pd (14 h) 132Te (78 h) 143Pr (14 d) 240Pu (6,600 a) 92Sr (2.7 h) 112Ag (3.1 h) 132I (2.3 h) 144Ce-144Pr (280 d) 92Y(3.5h) 115Cd (53 h) 133mTe (55 min) 145Pr (6.0 h)

aThe list includes pairs of radionuclides (for example, 106Ru-106Rh) that are treated together because the radioactive half-life of the decay prod- uct (106Rh) is much shorter than that of its precursor (106Ru) and the two radionuclides can be considered in radioactive equilibrium for all practical purposes of the study.

K = the conversion coefficient from exposure3 to dose • To determine the Urban/Rural status, data from the 1940 (mGy mR−1) for representative individual I; and 1950 census for the largest cities and towns in New 24 = the number of hours in a day; Mexico were used to estimate the 1945 population of Xout = the outdoor exposure (mR) in precinct L during each. Cities/towns with more than 10,000 residents were the first year after the test; classified as Urban. Table 2 shows the eight cities/towns Tout = the number of hours spent outdoors in a day by and the estimated 1945 populations that were greater than representative individual I; 10,000 persons, which would qualify as urban centers. Xin = the indoor exposure (mR) in precinct L during the The population of the eight most populous cities/towns first year after the test; and represents about 30% of the state population. Forty-three, Tin = the number of hours spent indoors in a day by rep- or about 6% of the state’s 721 voting precincts in 1954, resentative individual I. were classified as Urban based on the above scheme. Two types of data are needed to estimate the values of Therefore, 94% of the precincts were classified as Rural; Dext(I, L): (1) the non-radiation data identifying the charac- • The ethnic groups considered in the 1940 and 1950 cen- teristics of representative individual I, and (2) the radiation suses were identified as: Native White; Foreign born data used to estimate the values of the exposure, Xout and White; Negro; Other races. We assumed that: (1) the Xin, in each precinct L, as well as the values of the conver- white population in our classification corresponded to sion coefficient K from exposure to dose. 58.3% of the Native White + Foreign born White; (2) the Hispanic population in our classification corresponded to Non-radiation data 41.7% of the Native White + Foreign born White; (3) the For each of the 721 precincts (L)ofNewMexico,the population of Native Americans in our classification approximate coordinates (longitude and latitude) of the corresponded to that identified as Other races; and (4) geographic centroids, a metric of population density [urban the population of African Americans in our classification (U)orrural(R)], ecozone classification [plains (P), mountains Table 2. List of urban centers and number of precincts classified as (M), or combination of plains and mountains (P/M)], and the Urban. total population in 1945 of each ethnic group were assessed. Number of precincts Population numbers of four ethnic groups [White (W), County City/Town 1945 population classified as Urban Hispanic (H), Native American (NA), and African American (AA)] were derived from the New Mexico sections in the US Bernalillo Albuquerque 66,132 16 1940 and 1950 censuses according to seven age groups Santa Fe Santa Fe 24,162 9 Chaves Roswell 19,610 2 (in utero, newborn, 1–2y,3–7y,8–12 y, 13–17 y, and adult). Curry Clovis 13,692 2 Interpolations and assumptions were needed to convert the San Miguel Las Vegas 13,063 4 census data into the required grouping for our assessment. Eddy Carlsbad 12,546 1 For example: Lea Hobbs 12,247 2 Doña Ana Las Cruces 10,355 7 3 1 Although SI unit for exposure is C kg , the traditional unit, mR, is used in this Sub-total = 171,807 43 paper to facilitate the comparison with data and results of the 1950s and 1960s.

www.health-physics.com c A. BOUVILLE ET AL. 403 corresponded to that identified as Negro in the census; Table 4. Types of building material used in residences according to and data set (Potischman et al. 2020) and corresponding values of the shielding factor. • In the 1940 and 1950 census, population data are given for 5-y age groups (<5 y, 5 to 10 y, etc.). Those data were Housing construction materials (frequency usage, %) interpolated to fit the age distribution selected for the Shielding Data set Adobe Wood Other factor, SF purposes of this paper. See Table 1 of Simon et al. (2020) for the prepared data. In that preparation, we assumed that: A7811110.17 (1) the in-utero population was equal to three quarters of B919— 0.12 the population of infants <1 y of age in 1945; and (2) the C7327— 0.19 age distribution for the entire population of New Mexico D5017330.29 — — was representative of the age distribution of any ethnic E 100 0.5 F 100 ——0.077 group in any ecozone or type of residence.

To determine the number of representative individuals months, when most of the external dose was delivered I in each precinct L, consideration was also given to the dif- and children were out of school. ferences in lifestyle and diet, which depended on variables including ethnicity, ecozone, and population density of the precinct. The primary lifestyle data that have an influence Radiation data on the absorbed dose from external irradiation are the build- The radiation data that were used to estimate the values ing materials of the residences and the fraction of time spent of the absorbed doses from external irradiation, Dext(I,L), in mGy, in each precinct L, are: (1) the exposure rates at H + 12 in and outdoors. Data from a sample of New Mexico −1 residents alive at the time of Trinity, along with the dietary h, in mR h , in each precinct; (2) the times of arrival of fall- information, were collected in focus-group meetings and out in each precinct in hours counted from the time of deto- key-informant interviews conducted within the framework nation; and (3) the values of the conversion coefficient K of this study with participants classified into 18 possible from exposure to dose, in mGy per mR, which, in this paper, strata (Table 3) according to ethnicity, ecozone, and type of are taken to only depend on the age group of the representa- residence (Potischman et al. 2020). As dietary and lifestyle tive individual. These radiation data are described in the fol- data for some of the strata were considered to be relatively lowing paragraphs. similar, for our purposes, six data sets (A, B, C, D, E,and ExposureratesatH+12h F) were derived to describe the population groups of New The values of the outdoor and indoor exposures, Xout Mexico (Table 3). and Xin, cumulated over the first year after the test were de- The average values of the lifestyle data obtained from rived from data described by isopleths of exposure rates, in the focus-group meetings and key-informant interviews are units of mR h−1, at 1 m above ground, shown in Fig. 1. The presented in Tables 4 and 5: fallout pattern presented in Fig. 1 is adapted from Quinn (1987) and from Cederwall and Peterson (1990). Both stud- • Housing construction materials: as indicated in Table 4, ies, which were carried out within the framework of the adobe and wood were the predominant building mate- ORERP, used a methodology that had been developed for rials for the residences in 1945; and the analysis of the nuclear weapons tests conducted at the • Time spent outdoors: the values of Tout (hours per day) Nevada Test Site in the 1950s and 1960s (Church et al. were adjusted to represent the same age groups used in 1990). Quinn (1987) performed an analysis of the available the dose calculation. The results, presented in Table 5 ac- exposure-rate data, which, for the most part, were collected cording to data set and age group, show relatively small during the 3 wk following the Trinity test; the measure- differences from one data set to another and from one ments were taken with the help of a variety of instruments, age group to another. These values apply to the summer at or near ground level, mainly at populated locations and

Table 3. Data set assignment for each of the 18 strata. Stratum Data set Stratum Data set Stratum Data set Stratum Data set

W(U, P) D H(U, P) D NA(P) F AA(U,P) D W(U,P/M) C H(U,P/M) C NA(P/M) E AA(U,P/M) C W(R, P) D H(R, P) A NA(M) E AA(R, P) A W(R, M) B H(R, M) B AA(R, M) B W(R,P/M) B H(R,P/M) B AA(R,P/M) B

www.health-physics.com 404 Health Physics October 2020, Volume 119, Number 4 Table 5. Time spent outdoors (h) per day, according to data set and descending fallout before decreasing to a value at the end age group, in summer months (based on Potischman et al. 2020). of cloud passage over the site that was due only to activity Data Set In uteroa Newborna 1–2y 3–7y 8–12 y 13–17 y Adult on the ground. The exposure rate decreases after it reaches A 9 910.59.38.38.49 its maximum primarily because the rapid radioactive decay B 9 910.59.38.38.49 of nuclides deposited on the ground more than offsets the C221313138.62additional fallout. D 5 5 5 6.2 7.6 7.1 5 Quinn (1987) derived the times of arrival of fallout ei- E10101010101010ther from the sequential readings of exposure rate, covering Fb 11 11 11 11 11 11 11 the entire period of descending fallout, which were available aThe values for “in utero” and “newborn” are taken to be the same as those for for some locations, or, in the absence of such readings, from the adult. a meteorological model in which the particle trajectories bIn addition, 85% of the people in this data set slept on the roof at night. and fall times were calculated for each layer of the radioac- tive cloud. The time of maximum rate of fallout, computed along roads connecting those populated locations. Cederwall using the meteorological model, was defined as the time of and Peterson (1990) used a meteorological model to expand arrival of fallout, denoted as TOA in this paper. The time of Quinn’s results to areas without available exposure rates in maximum rate of fallout occurs between the time of initial the northern part of the state. Because the measurements arrival of fallout and the time of peak activity, as observed of exposure rate were made at various times after the test, in the sequential readings of the exposure rate. Within the they were, as has been done in other fallout studies, framework of the ORERP, Quinn (1990) compared the re- normalized by Quinn (1987) to 12 h post-detonation (H sults obtained from the sequential readings of exposure rate +12 h) by means of a multi-exponential function based on and from the meteorological model for several tests conducted calculated radionuclide inventories and exposure rates as a at the Nevada Test Site and found reasonable agreement function of time for Trinity (Hicks 1985). In this work, the between the two sets of values. exposure rate at H+12 is referred to as X˙ (12).4 Because The times of arrival of fallout, counted from the time of only exposure rates greater that 0.2 mR h−1 at H+12 h detonation at H + 0, were extracted from Quinn (1987) but were considered in Quinn (1987), we added an additional isopleth of 0.1 mR h−1 and considered that the exposure rate in any precinct below 0.1 mR h−1 wasequalto0.05mRh−1 at H+12 h. This assumption implicitly assigns some degree of exposure to persons living in every precinct of New Mexico in 1945. We note, however, that for most of the state the exposure from Trinity was small compared to the subsequent radiation exposure from NTS and global fallout (Simon et al. 2020). Values of X˙(12, L) for every precinct centroid inside the pattern were obtained by interpolation between isopleths. Times of arrival of fallout The radioactive cloud produced by the nuclear detona- tion extended to an altitude of several kilometers. The nuclear debris, containing a combination of fission and activation products, fell to earth from the different layers of the cloud over a duration which, as a rule of thumb, was on average nu- merically equal to the initial time of arrival of fallout (Quinn 1990). During the passage of the radioactive cloud over a par- ticular location, the exposure rate at ground level initially be- gan to increase as a result of radioactive decay of the nuclides in the air (descending fallout), followed soon after by the combination of the radioactive decay of descending fallout Fig. 1. Map of New Mexico with isopleths of exposure rates (mR and of nuclides already deposited on the ground. The expo- −1 sure rate usually reached a maximum while there was still h ) at H+12 h after the test (Quinn 1987; Cederwall and Peterson 1990). These isopleths are derived from measurements of exposure rate conducted during the first few days following the test. Fuzzy 4In a companion paper (Beck et al. 2020), the exposure rate at H + 12 h is spots show the locations of the 8 cities with 1945 populations greater denoted as E12. It was also referred to as ĖR(12), ̇Ė(12), ̇Ė12,XE,andẊ12 than 10,000. The approximate centerline of the fallout trace is repre- in other papers related to fallout. sented by two segments.

www.health-physics.com c A. BOUVILLE ET AL. 405 are not shown in Fig. 1. The TOA values varied from 0 h at outside the pattern were given the same values of X˙(12) and the test site to about 40 h when the main radioactive debris TOA, there is no difference between the lowest, average, and cloud reached the northeast border of New Mexico. In a highest values of X˙(12) and TOA in those counties. manner similar to what was done for the exposure rates, we Conversion coefficients K from exposure to dose extended, by simple interpolation, the TOA values beyond Photon energies of a few hundred keVare typically emit- the area of New Mexico considered in Quinn (1987) to ted by the fission products during their radioactive decay. assess approximate doses statewide. Values of TOA(L) for Based on a similar work related to the nuclear weapons tests every precinct centroid inside the pattern were obtained by conducted in the Marshall Islands (Bouville et al. 2010), the interpolation between isopleths. Values of TOA(L) outside conversion coefficients from exposure to dose for photon the fallout pattern were assigned values consistent with energies of a few hundred keV were taken to vary with age those obtained for the nearest counties inside the pattern. but to be independent of the organ or tissue of the body and All precincts of a given county partially or entirely outside of the geometry of irradiation (ICRP 2010; NCRP 2018). the pattern were assigned the same value of TOA. Lowest, ˙ With respect to the variation with age, the conversion average, and highest precinct values of X (12) and TOA coefficient K was estimated to be 6.6 x 10−3 mGy mR−1 from Trinity fallout in each of the 31 counties of New for adults (Bouville et al. 2010) and to be slightly higher Mexico are presented in Table 6. Because precincts entirely for younger ages (Bellamy et al. 2019; Jacob et al. 1990). The selected values of K for the age groups considered in − − this study were 6.6 10 3 mGy mR 1 for the in-utero − − subjects (same value as for adults), 8.6 10 3 mGy mR 1 Table 6. Lowest, average, and highest precinct values of X˙ (12) and TOA from Trinity fallout in each of the 31 counties of New Mexico. for babies less than 1 y old and for infants less than 3 y, −3 −1 ˙ 1 7.9 10 mGy mR for children aged 3 to 7 y and 8 to X (12) (mR h at H+12 h) TOA (h) − − 12 y, and 7.3 10 3 mGy mR 1 for adolescents aged 13 County Lowest Average Highest Lowest Average Highest to 17 y. Bernalillo 0.18 0.37 1.40 8.60 10.2 11.2 Outdoor exposure Catron 0.05 0.08 0.13 5.30 9.6 12.0 The outdoor exposure in precinct L, Xout(L), in mR, cu- Chaves 0.05 1.30 0.91 15.8 21.3 24.0 mulated over the first year after the test, was obtained as: Colfax 0.23 1.56 2.0 26.9 33.1 38.0 Curry 0.05 0.06 0.11 18.1 23.4 24.0 X ðÞ¼L ∫1 y Xt˙ ðÞ; L dt ð2Þ De Baca 0.17 1.04 2.48 8.84 12.1 15.5 out TOA Dona Ana 0.05 0.05 0.05 24.0 24.0 24.0 where TOA is the time of arrival of fallout in precinct L,in Eddy 0.05 0.05 0.05 24.0 24.0 24.0 hours since the detonation, and X˙ ðÞt; L is the exposure rate Grant 0.05 0.05 0.05 24.0 24.0 24.0 −1 Guadalupe 0.86 22.7 104 6.9 10.5 14.1 at time t in precinct L,inmRh , which was expressed an- Harding 0.11 0.16 0.24 23.8 26.2 29.5 alytically in the form: Hidalgo 0.05 0.05 0.05 24.0 24.0 24.0 R −l R; Lea 0.05 0.05 0.05 24.0 24.0 24.0 ˙ ; ˙ ; ∑i¼10 ; ðÞV i t XtðÞ¼L X ðÞ12 L i¼1 a i e ð3Þ Lincoln 0.05 8.86 109.6 2.8 13.3 22.0 V Luna 0.05 0.05 0.05 24.0 24.0 24.0 McKinley 0.05 0.06 0.12 17.1 23.2 24.0 where R/V is the ratio of the total activities of the refractory Mora 0.73 1.58 2.00 17.6 21.6 25.5 (R) and of the volatile (V) radionuclides in the radionu- Otero 0.05 0.05 0.05 24.0 24.0 24.0 clide mix deposited on the ground at TOA.FortheTrinity Quay 0.05 0.13 0.29 13.5 21.5 32.0 test, values of R/V were estimated to range from 0.5 at dis- Rio Arriba 0.12 0.23 0.62 21.0 28.8 38.4 tances far away from the detonation site to 3 in precincts Roosevelt 0.05 0.08 0.13 15.6 22.1 24.0 close to the detonation site (Beck et al. 2020). Detailed in- Sandoval 0.13 0.20 0.53 11.6 16.8 25.6 formation on how R/V was determined for each precinct is San Juan 0.05 0.06 0.11 30.8 36.2 39.8 discussed by Beck et al. (2020) and is summarized in this San Miguel 0.26 3.3 19.3 10.0 14.0 20.5 paper in the section “Absorbed Doses From Internal Santa Fe 0.22 1.06 4.14 9.9 15.4 21.4 Irradiation.” Sierra 0.05 0.05 0.05 8.0 8.0 8.0 Equation (3) is based on calculated exposure rates, nor- Socorro 0.15 35.5 423 1.0 3.42 6.49 malized at H + 12 h, for the Trinity test (Hicks 1985). For Taos 0.34 0.62 1.04 23.5 29.0 36.5 these calculations, the vertical profile of the activity depos- Torrance 1.3 68.0 481 4.3 6.7 9.1 ited on the ground was assumed to be exponential with a re- Union 0.12 0.21 0.71 28.7 35.5 42.4 laxation length of 0.16 g cm−2 to take the surface roughness Valencia 0.05 0.27 1.28 6.19 9.44 12.4 effects into account (Beck 1980). Exposure rates at 1 m www.health-physics.com 406 Health Physics October 2020, Volume 119, Number 4 above the ground surface are provided for the time of deto- Indoor exposure nation (H + 0) and for 30 decay times ranging from 1 h to 50 y. The indoor exposure Xin(L), in mR, from TOA until Hicks’ results are calculated assuming an instantaneous depo- 1 y after the test, is calculated as the product of the out- sition of fallout at H + 0, whereas the actual conditions after a door exposure Xout(L), in mR, and of the shielding fac- nuclear detonation involve a combination of descending fall- tor SF(L): out and of activity depositing on the ground over a relatively extensive length of time. However, comparisons of the calcu- : lated results and of the measured exposure rates, either for XinðÞ¼L XoutðÞL xSFLðÞ ð4Þ Trinity (Quinn 1987) or for the tests detonated at the Nevada Test Site (Hicks 1982), show a good agreement for times after The value of SF(L) varies according to the construction of the peak activity, probably because the rapid radioactive decay residences (e.g., brick, adobe, wood, stone or concrete, of nuclides deposited on the ground more than offsets the ad- etc.). The frequency of different housing construction mate- ditional fallout. For times preceding the peak activity, the def- rials, obtained from the focus-group meetings and the indi- inition of TOA being at the maximum rate of fallout results in vidual interviews conducted in New Mexico (Potischman the inclusion of the exposure due to descending fallout in the et al. 2020), is shown in Table 4 for each of the six calculated exposure arising from the activity deposited on the data sets. Adobe and wood were the main housing ground, because the increased calculated exposure between construction materials in New Mexico in 1945. The the times of the maximum rate of fallout and of the peak ac- shielding factors were taken to be 0.077 for adobe houses tivity is approximately compensated by the elimination of the (Gordeev et al. 2002) and 0.5 for wooden and other houses, exposure between the times of initial arrival and maximum grouped together (Dunning et al. 1957). The average rates of fallout. A description of the various ways in which values of the shielding factor for each of the six data sets, TOA could be defined in order to take the descending fallout considering the mixtures of housing construction materials, into account is provided in Thompson et al. (1994). are also presented in Table 4. The values of a(R/V, i) and of l (R/V, i), in h−1,forR/V = Organ absorbed doses 0.5 were obtained by fitting the exposure rates (normalized to The organ absorbed doses from external irradiation −1 1mRh at H+12 h) calculated for 30 times post shot by Hicks [Dext(L, I) in mGy] received by a representative individual (1985). Henderson (1991) fitted Hicks’ data for R/V = 0.5; I who resided in precinct L during the entire first year after Beck (2009) calculated the exposure rates (normalized to the test are taken to have the same values for the five princi- 1mRh−1 at H+12 h) for values of R/V > 0.5 and fitted those pal organs and tissues of interest (lung, thyroid, active mar- data. The resultant values of a(R/V,i)andofl(R/V,i)forR/V row, stomach, and colon): equal to 0.5, 1, 1.5, 2, and 3 are presented in Table 7. These fitted values assume only radioactive decay. In the calcula- KIðÞ tions of the exposure rates, we assumed that the activity de- D ðÞ¼L; I x ½ðÞþX ðÞL x T ðÞI; L ðÞX ðÞL xT ðÞI; L ext 24 out out in in posited on the ground varied exponentially with soil depth −2 KIðÞ with a relaxation length of 0.16 g cm from the time of de- ¼ X ðÞI; L ð5Þ 24 tot position at TOA until the end of the first year after the test (Hicks 1985; Beck 1980). Weathering of radioactivity into where the soil column due to rain will cause the exposure rate to Tout(I,L) = the number of hours per day spent outdoors decrease more rapidly than radioactive decay alone, partic- by the considered representative individual I. Its value, which ularly during the several months following the deposition. is determined in the data set in which precinct L belongs, de- Since most of the exposure occurred in the first few weeks pends on the age and ethnicity of the considered representa- after the arrival of the fallout, we assumed that weathering tive individual, as well as on the ecozone and type of in the arid climate of New Mexico would have had a minimal residence in precinct L. The values of Tout (I, L) were derived effect on the integral exposure for the first year and have from the focus-group meetings and the individual interviews conservatively neglected weathering. Our calculations (Potischman et al. 2020); they are presented in Table 5; show that accounting for weathering would decrease the Tin(I,L)=24-Tout(I,L) = the number of hours per day 1-y exposure by about 10% for TOA = 4 h and by about spent indoors by the representative individual I; 15% for TOA = 40 h, for an R/V value of 0.5 in both Xtot(I,L) = the total exposure, in mR, from TOA to the cases. The variation of R/V from 0.5 to 3 only has a small end of the first year after the test, associated with the repre- effect on the 1-y exposure; for example, the 1-y exposures sentative individual, I, in precinct L;and for R/V =3wouldbeabout5and8%smallerthanthe K(I) = the conversion coefficient from exposure to ab- corresponding values for R/V=0.5, for TOA = 4 and 40 h, sorbed dose in any organ according to the age of the repre- respectively. sentative individual, in units of mGy mR−1. www.health-physics.com c A. BOUVILLE ET AL. 407 Table 7. Values of the parameters a(R/V,i) and l(R/V,i) used to estimate the variation of the exposure rate with time, according to equation 3 (Henderson 1991; Beck et al. 2020). Values of a(R/V, i) R/V =0.51 1.52 3

i=1 1.05 × 102 1.02 × 102 9.75 × 101 9.44 × 101 9.00 × 101 i=2 3.21 × 101 3.14 × 101 2.64 × 101 2.31 × 101 1.96 × 101 i=3 2.98 × 100 1.64 × 100 4.64 × 100 4.56 × 100 3.04 × 100 i=4 7.77 × 101 4.52 × 100 4.95 × 106 1.32 × 100 9.24 × 101 i=5 2.03 × 101 1.74 × 101 1.46 × 100 3.74 × 101 2.53 × 101 i=6 2.38 × 100 3.62 × 101 3.92 × 101 1.22 × 101 1.32 × 102 i=7 3.18 × 102 8.11 × 102 7.69 × 102 5.79 × 103 3.00 × 103 i=8 3.45 × 103 8.66 × 103 7.60 × 103 2.84 × 102 8.51 × 106 i=9 2.37 × 105 2.75 × 104 2.21 × 105 2.29 × 105 1.00 × 108 i=10 6.02 × 106 6.41 × 106 1.00 × 108 3.12 × 106 2.57 × 106 Values of l(R/V, i) R/V =0.51 1.52 3 i=1 2.00 × 100 2.16 × 100 2.12 × 100 2.09 × 100 2.02 × 100 i=2 6.84 × 101 7.61 × 101 7.31 × 101 7.06 × 101 6.11 × 101 i=3 6.84 × 101 9.24 × 102 3.22 × 101 3.13 × 101 2.17 × 101 i=4 5.24 × 102 3.37 × 101 5.58 × 104 8.76 × 102 4.78 × 102 i=5 9.79 × 103 3.37 × 101 8.83 × 102 2.26 × 102 9.52 × 103 i=6 1.57 × 101 1.88 × 102 1.87 × 102 8.24 × 103 7.31 × 104 i=7 2.25 × 103 4.33 × 103 4.15 × 103 4.04 × 104 2.08 × 103 i=8 4.14 × 104 6.84 × 104 5.11 × 104 2.57 × 103 1.49 × 105 i=9 1.92 × 105 4.46 × 105 1.24 × 105 2.42 × 105 3.38 × 107 i=10 1.00 × 106 1.40 × 106 3.46 × 104 1.87 × 106 3.76 × 106

Uncertainties in the estimation of the doses from accuracy of the instrument used. Therefore, the interpo- external irradiation lated values of X˙(12) at each precinct centroid have sig- Extensive information on the main sources of uncertainty nificant uncertainties. For the purposes of this work, encountered in the assessment of doses in environmental stud- the uncertainty in X˙ (12) was assumed to be described ies, as well as on the methods that can be used to treat them, by a triangular probability distribution, [TRI(min, is provided in NCRP Reports (NCRP 2007, 2009a and b). In mode, max)], as follows. The precincts were classified this article, uncertainties in the doses received by each popula- into three categories: (1) those inside the fallout pat- tion group in each voting precinct were derived, primarily, from tern with a TOA less than or equal to 10 h; the uncer- the uncertainties on the estimated exposure rates normalized at ˙ tainty distribution around the best estimate is TRI H+12 h, i.e., X ð12). For a given representative individual, I,and (0.33, 1, 3); (2) those inside the fallout pattern, with a given precinct, L, the external dose can be expressed as: a TOA greater than 10 h; the uncertainty distribution ; ; around the best estimate is TRI (0.5, 1, 2); and (3) ; ˙ ; XoutðÞL XtotðÞI L DextðÞI L : DextðÞ¼I L XðÞ12 L x ˙ x x ð6Þ ̇ ; X ðÞ12; L XoutðÞL XtotðÞI; L those outside the fallout pattern, where XðÞ12 L = 0.05 mR h−1; the uncertainty distribution around the The uncertainties in individual parameters were assessed to best estimate is TRI (0, 1, 1.2); ˙ be as follows: • XoutðÞL = XðÞ12; L : the uncertainty in the ratio of the out- door exposure Xout(L), in mR, from TOA to the end of 1 y − • X˙ ðÞ12; L : as discussed in Beck et al. (2020), the estimates after the test, and of the exposure rate at H+12 h, in mR h 1, − of X˙ ðÞ12 ,inmRh 1, at each precinct centroid were ob- depends essentially on the uncertainty in TOA,which,just tained by interpolation of the published fallout pattern like XðÞ̇12; L , was interpolated from the TOA pattern (Fig. 1) that was constructed from an analysis of all published by Quinn (1987). Quinn (1990) estimated that, post-shot monitoring data supplemented by meteorological depending on the nuclear test and on the distance from data (Quinn 1987). Unfortunately, since the monitoring the test site, TOA could vary from 15 min to several hours. data were limited to areas with roads, the published fallout Taking as an example an uncertainty of 30 min for TOA = 1 ˙ pattern is itself somewhat uncertain. In addition, the pre- or 2 h, the uncertainty in the Xout ðÞL =XðÞ12; L ratio would cision and accuracy of X˙ (12) vary depending on the be about 20% for TOA = 1 h and about 10% for TOA = 2 h.

www.health-physics.com 408 Health Physics October 2020, Volume 119, Number 4 If the uncertainty in TOA is 1 h for TOA = 5 h, the un- ABSORBED DOSES FROM INTERNAL certainty in the ratio is about 5%. Other sources of un- IRRADIATION certainty include those related to the variation of the The most common and important pathways of internal outdoor exposure rate with time (model, value of R/V, exposure are ingestion of radioactively contaminated food- weathering conditions), which are deemed to be minor stuffs and inhalation of radioactively contaminated air. The compared to the uncertainty related to TOA. The uncer- doses from internal irradiation (ingestion and inhalation) were = ˙ ; tainty in XoutðÞL XðÞ12 L was subjectively taken to be derived from the radionuclide-dependent activities estimated log-normally distributed around the best estimate with to have been deposited at TOA on the ground (Agd)oron a GSD of 1.2 for any precinct, though it is recognized −2 vegetation (Aveg), in Bq m , derived directly from outdoor − that the uncertainty may be greater at close-in distances exposure rate, X˙ ðÞ12 ,inmRh 1 at H + 12 h. than at far away locations; The general expression that was used for the estimation • Xtot(I,L)/Xout(L): the uncertainty in the ratio of the total of the internal dose, Dpw,inmGy,toanorgan,m, of a represen- exposure, Xtot(I,L), in mR, from TOA to the end of the tative individual, I, resident of precinct, L, from a radionuclide, first year, associated with the representative individ- Z, and an exposure pathway, pw, was expressed as: ual, I,inprecinct,L, and of the corresponding outdoor exposure, X (L), depends on the age, ethnicity, and out A ðÞZ; L; TOA ecozone of the representative individual. The parame- D Z; L; I; m X˙ 12; L pw pwðÞ¼ðÞ˙ ; ters of interest are the fraction of time spent outdoors, X ðÞ12 L −1 Tout,inhd , and the shielding factor, SF, associated ICpw ðÞZ; L with the building materials used in the residence in ApwðÞZ; L; TOA 1945. This information was derived from the focus groups and the key informant interviews (Potischman Qpw ðÞZ; L; I DpwðÞZ; I; m ð7Þ et al. 2020). Based on the variability of the responses ICpw ðÞZ; L QpwðÞZ; L; I provided by the participants of the focus groups and the key informants and taking into account the uncer- where −2 tainties related to the reliability of the responses pro- Apw(Z,L,TOA)= the deposited activity, in Bq m ,rele- vided by a small group of persons to represent much vant to pathway pw, of radionuclide, Z,atTOA,inprecinctL; larger groups, the uncertainty in Xtot(I,L)/Xout(L)was ICpw(Z, L) = the time-integrated concentration, from subjectively taken to be log-normally distributed TOA until 1 y after the test, of radionuclide Z,inprecinct around the best estimate with a GSD of 1.3 for any L, in air for inhalation pathways and in the foodstuff of in- age, ethnicity, and ecozone of the representative indi- terest for an ingestion pathway. ICpw(Z, L) is expressed in − − vidual, in precincts with TOA greater than 3 h. For Bq d m 3 for inhalation of contaminated air, in Bq d L 1 the precincts with TOAs less than 3 h (that is, between for the ingestion of contaminated water or milk, and in Bq − 05:29 and 08:29 AM on the day of the test), as the dkg 1 for the ingestion of any other contaminated foodstuff age-dependent values of Tout may be substantially dif- of interest; ferent from the representative values for the entire day, Qpw(Z, L, I) = the activity intake, in Bq, from TOA until the GSDs of the log-normal uncertainty distributions one year after the test, of radionuclide Z, in air for inhalation around the best estimate were taken to be 1.5; and pathways and in the foodstuff of interest for an ingestion ; − • Dext ðÞI L ¼ KIðÞ: the uncertainty in K(I), in mGy mR 1, pathway, by representative individual, I, resident of precinct L. Xtot ðÞI;L depends on the geometry of irradiation, on the energy The methods used to estimate the values of each term ˙ spectrum of the incident g rays, on the organ that is on the right side of the equation, except for XðÞ12; L ,are considered, and on the age of the representative indi- discussed in turn. vidual. In our analysis, only the age of the representa- Normalized activities deposited on the vegetation and on ˙ tive individual was considered in the estimation of the the ground [Apw/X (12, L)] central values of K(I). However, all factors mentioned The methodology used to estimate the normalized ac- above were considered in the evaluation of the uncer- tivities deposited on the ground is based on factors and rela- tainty of K(I), which was subjectively estimated to be tionships derived by Hicks (1985) for each radionuclide and distributed log-normally around the best estimate with for a range of TOAvalues for a mixture of radionuclides cor- a GSD of 1.2. responding to R/V =0.5.Hicks’ values were extended to values of R/V > 0.5 by Beck et al. (2020). A model devel- The way the uncertainties in the individual parameters were oped in collaboration with Russian scientists was then used combined to evaluate the uncertainties in the dose estimates to estimate the normalized activities deposited on the ground is provided in Simon et al. (2020). for values of R/V other than 0.5 and also to estimate the www.health-physics.com c A. BOUVILLE ET AL. 409 normalized activities deposited on vegetation as a func- of the total fallout activity that was on small particles at dis- tion of R/V5 (Beck 2009; Beck et al. 2020). tances close to the detonation site and to the trace axis was very small but increased to 1.0 as one reached distances at Estimation of R/V which all particles larger than approximately 50 mm would The radionuclides created during the explosion of a nu- have fallen out by gravitational settling (Lindberg et al. clear device are usually classified into refractory or volatile 1959; Larson et al. 1966). For values of TOA ≥ Tcr, the rel- according to whether their melting point is higher or lower ative R/V ratio is equal to 0.5 and all activity deposited will than 1,500 oC (Hicks 1982). For example, isotopes of iodine be on particles <50 mm. Even though it is believed that the and cesium are classified as volatile, and isotopes of zirco- particle-size distribution and therefore the activity distribu- nium are classified as refractory (Beck et al. 2010). The var- tion changed with time of deposition post detonation (that iation of R/V with location reflects the fact that refractory is, increased with time for volatile radionuclides and de- nuclides condense from the vaporized nuclear debris onto creased for refractory radionuclides), the crude assumption condensation nuclei at earlier times after detonation com- was made that the activity distribution remained constant pared to volatile nuclides and, thus, tend to be incorporated (excluding radioactive decay) in locations where TOA was into large particles, i.e., those greater than 50 mmindiame- greater than, or equal to Tcr. ter. Since the larger and more massive particles deposit ear- For the Trinity test, the value of Tcr, obtained as the lier due to gravitational settling, the earlier the TOA,the quotient of the maximum height of the radioactive cloud, larger the proportion of large particles deposited and the 10.7 km, and of the sedimentation velocity of 50 mmparti- greater the proportion of total activity deposited that is on cles, taken to be 0.73 km h−1 (Beck et al. 2020), is found to large particles, i.e., R/V is higher at close-in distances and be 14.7 h. Using the joint US–Russian model, described in lower at more distant locations. This phenomenon, termed Beck et al. (2020), the values of R/Vat the centroids of each fractionation, reflects that the R/V ratio in the deposited fall- precinct L were determined from N50 (L), which is the frac- out differs from its “unfractionated” value in the debris cloud; tion of total beta activity on particles <50 mm at the time of that is, without any depletion due to deposition on the arrival of fallout (TOA(L)): ground. Hicks (1982) considered that all elements of a given • for precinct centroids along the axis of the trace of the fission chain, that is, with the same mass number, are either radioactive fallout, N (L) is denoted as N (L)and refractory (R)orvolatile(V), except for the chains 91, 140, 50 50a calculated as: and 141. For these intermediate chains, the refractory frac- tions were determined by Hicks (1982) for each chain at 6 − : − 3 3 20 s post detonation. N50aðÞ¼L 1 0 987 x exp d Tr ð8Þ In this analysis, it is only of interest to compare the −1 level of fractionation in each precinct to an unfractionated where Tr = TOA(L)/Tcr and d =1.6h .Thisequation, radionuclide mix. Therefore, the degree of fractionation is which was developed in association with Russian scientists expressed as a relative R/V ratio, where a ratio of 1.0 represents (Beck et al. 2020), is very similar to that used by Beck et al. unfractionated fallout and a ratio of 0.5 represents fallout (2020), the only difference being in the value selected for d, where one half of the atoms of each refractory radionuclide which reflects the spread of the radioactive cloud, and has been removed, typical of fallout at long distances from • for precinct centroids that are off-axis, N50 (L)iscalcu- the detonation site (Beck et al. 2010; Hicks 1982). Because lated as: of the interdependence of fractionation and particle size and pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Âà ˙ ˙ because the time of deposition varies for particles of differ- N50ðÞ¼L N50aðÞL −0:13 x N50aðÞL xlnXðÞ12; L =XðÞÞ12; axis; L ent sizes, use was made of the critical time, Tcr, defined as ð9Þ being the time since detonation for all particles greater than ˙ −1 50 mm in diameter, with an assumed density of 2.5 g cm−3, where X(12, axis, L) is the exposure rate, in mR h ,atH+12h to be deposited. The choice of 50 mm is partly based on the along the axis of the trace, at the location where TOA = TOA(L). observation from post-detonation test data that in general, It follows from equations 8 and 9 that the value of R/V ˙ ; only particles of less than approximately 50 mmindiameter in precinct L is derived from the values of TOA(L),XðÞ12 L , ˙ ; ; ˙ ; were originally retained on vegetation and that the fraction and XðÞ12 axis L .SinceXðÞ12 L and TOA(L) were deter- mined for all precincts, X˙ ðÞ12; axis; L is the only value that ˙ ; ; 5A peer-reviewed publication on this methodology is under preparation. remains to be calculated. The variation of XðÞ12 axis L 6For example, if for R/V = 1, 40% of the chain is R and 60% is V,thenfor along the axis of the trace was estimated using Quinn’s fall- R/V = 2, the refractory fraction of the chain is multiplied by 2 while the vol- atile fraction remains the same so that the refractory fraction is now 80% and out pattern as follows: the volatile fraction 60%. But since the total is now 140%, normalization • needs to be done for 100%, resulting in the refractory fraction of this chain The axis of the trace was assumed to consist of two linear for an overall R/V = 2 of 57% and in the volatile fraction to be 43%. segments (Fig. 1): the first segment for the first 10 h after

www.health-physics.com 410 Health Physics October 2020, Volume 119, Number 4

the Trinity test, beginning at the detonation site (geograph- Table 8. Derivation of the values of R/V from those of N50 or N50a. o − o ical coordinates: 33.68 Nand 106.48 W) and ending at R/V N50a or N50 location with coordinates 35.00o Nand−104.91o W; and 0.5 >0.83 the second one, from 10 to 40 h after the shot, beginning 10.43– 0.83 o − o at the end of the first segment (35.00 N, 104.91 W) and 1.5 0.23 – <0.43 o − o ending at 37.00 N, 104.91 W; 20.09– <0.23 • ˙ ; ; The values of XðÞ12 axis L and of TOA(L) were estimated 3<0.09 for a number of points along each segment, and the var- iation of X˙ ðÞ12; axis; L with TOA(L)wasfittedusinga variety of functions; • For the first segment, the best fit of the variation with ˙ ; ; = ; = ˙ ; TOA of the exposure rate along the axis, XðÞ12; axis; L , Agd<50ðÞZ L R V TOA X ðÞ12 L −1 in mR h , was obtained, if TOA(L)<1.9h,as: b =b ; ¼ ˙ N50 ðÞL ðÞZ 0:5;12 FgdðÞðZ TOA 14Þ ˙ 2:5 3 X R;12 XðÞ¼12; axis; L EXPð4:9–14:9 TOAðÞ L þ 13:0 TOAðÞ L Þ ð10Þ V and, if 1.9 h ≤ TOA(L)<10h,as where  2 4 6 A (Z,L,R/V,TOA)andA (Z,L,R/V,TOA)=theac- −128 þ 304 TOAðÞ L −7:2 TOAðÞ L þ 0:045 TOAðÞ L gd gd<50 X˙ ðÞ¼12; axis; L tivity on particles of all sizes and the activity on particles 1−0:47 TOAðÞ L 2 þ 0:0085 TOAðÞ L 4−0:0021 TOAðÞ L 6 þ 0:000013 TOAðÞ L 8 less than 50 mm, respectively, of radionuclide Z deposited ð11Þ; and − on the ground at time TOA (h) at location L (Bq m 2), where • For the second segment, the variation with TOA of the ex- the degree of fractionation is R/V; − posure rate along the axis, X˙ ðÞ12; axis; L ,inmRh 1,was  b obtained as ˙ = the ratio of the total b activity deposited on X R;12 ˙ ; ; : − : : 2− : V XðÞ¼12 axis L 22 8 1 9 TOAðÞþ L 0 022 TOAðÞ L 0 0012 the ground to the exposure rate at H+12 h according to the − TOAðÞ L 3 þ 480 e TOAðÞ L : degree of fractionation R/V (Bq m−2 per mR h−1). Selected b ð12Þ values of ˙ are shown in Table 9; X R; V 12 =b b For each precinct L, the value of N50a(L)wasdeter- ðÞZ R; and (Z/ )0.5,12 = the ratios of the activity of V 12 mined using the value of TOA(L) in eqn (8). The value of radionuclide Z to the deposited total b activity at time H N (L) was obtained from eqn (9), using the values of N (L), 50 50a +12 h, according to R/V, for particles of all sizes (R/V ≠ ˙ ; ˙ ; ; XðÞ12 L ,andXðÞ12 axis L . Finally, the value of R/V is de- 0.5) and for the fraction of particles <50 mm(R/V = 0.5), re- rived as presented in Table 8 from the value of N50a or N50 spectively (unitless). Selected values of ðÞZ=b R; for the 63 V 12 (Beck et al. 2020). The R/V ratios range from 3 in precincts radionuclides under consideration are presented in Table 10, close to the detonation site to 0.5 at distances where TOA is in which (1) all isotopes of Fe, Co, Cu, Zr, Nb, Mo, Tc, Pr, greater than 14.7 h. Values of R/V =0.5werefoundtoapply Nd, Pm, Sm, U, Np, and Pu were treated as refractory, as to 667 of the 721 precincts. Values of R/V>0.5 were found well as 142La, 143Ce, and 144Ce;(2)allisotopesofAs,Br,Rb, for 54 precincts located near the first segment of the axis Ru,Rh,Pd,Ag,Cd,In,Sn,Sb,Te,I,andCsweretreatedasvol- of the trace (Fig. 1). atile, along with 90Sr and 90Y; and (3) 91Sr, 91mY, 140Ba, The activities of radionuclide (Z) deposited at TOA on 140La, 141La, and 141Ce were treated as intermediate be- ground and vegetation at location (L) depend on the expo- ˙ tween volatile and refractory (Hicks 1982). The values for sure rate X(12, L) and on the degree of fractionation (R/V). 239Pu and 240Pu, considered to be refractory radionuclides, Normalized activity of nuclide Z deposited on the were calculated assuming a 240Pu/239Pu atom ratio of 0.025 ˙ 137 239 240 ground at TOA [Agd/X (12)] and a Cs/( Pu + Pu) activity ratio of 30 for R/V =0.5 From the model described earlier, the activity of radionu- (Douglas 1978; Beck et al. 2020); and −2 clide Z deposited on the ground at TOA, Agd in Bq m ,nor- ˙ Fgd(Z,TOA) = the value at TOA of the function describ- malized to X ð12), is expressed as follows: ing the variation of the activity of radionuclide Z on the ˙ AgdðÞZ; L; R=V; TOA =X ðÞ12; L ground, normalized to t = 12 h, for the range of TOA values relevant to the Trinity study (1 to 44 h). This function, which b =b ; is independent of the R/V value, considers the build-up of ¼ ˙ ðÞZ R ;12 FgdðÞðZ TOA 13Þ X R ; V activity resulting from the radioactive decay of the precur- V 12 sors of radionuclide Z and the loss of activity resulting from including: the radioactive decay of radionuclide Z. www.health-physics.com c A. BOUVILLE ET AL. 411

Table 9. Selected values of the total b activity and of the exposure rate at vegetation as a result of deposition via wet processes, that b= ˙ H+12 h, X ðÞ12 R; , according to the degree of fractionation, R/V. V 12 is rainfall, that occurred during the passage of the radioac- tive cloud at location L. Although eqn (16) could also be ap- b= ˙ 2 1 R/V X ðÞ12 R; (Bq m per mR h ) V 12 plied to wet deposition (Kinnersley and Scott 2001), it was 0.5 4.11 × 106 considered important to use different models for dry and 14.92×106 wet fallout because the main parameters influencing the 1.5 5.43 × 106 fraction of the activity that is intercepted and initially 25.79×106 retained by vegetation are different in dry and wet condi- 36.24×106 tions. For dry deposition, important factors are particle size, standing biomass, leaf area available for dry deposition, and Normalized activity of nuclide Z deposited on the ˙ whether vegetation is dry or wet, whereas for wet deposition vegetation at TOA [Aveg /X ð12)] important factors are rainfall amount, R in mm, chemical The normalized activity of radionuclide Z deposited on ˙ form of the deposit, stage of development of the plant, and vegetation at TOA at location L, Aveg/X ð12), expressed in its water storage capacity (Pröhl 2009; Hoffman et al. Bq m−2 per mR h−1 at H + 12 h, is: 1992, 1995; Horton 1919). It is essential to note that fwet R R Aveg Z; L; ; TOA Agd<50 Z; L; ; TOA fdry V ¼ V only applies if the radioactive cloud passes over the location ẊðÞ12; L X˙ ðÞ12; L L considered at the same time as rain falls. Because rainfall R in New Mexico in the summer occurs almost exclusively as Agd Z; L; ; TOA fwet þ V ð15Þ X˙ ðÞ12; L transient and short-lived thunderstorms, we have estimated that each local rainfall event might last about 30 min, thus where fdry (unitless) is the fraction of the activity attached to giving the probability of coincidence of the radioactive cloud particles <50 mm that is intercepted and initially retained by and the rainfall event at the same time over L to be 1/48 ≅ vegetation as a result from deposition via dry processes. It is 0.02. This translates to a 98% probability that fwet =0anda calculated according to eqn (16), which is a modification by 2% probability that the value of fwet is in the range from 0 to Vandecasteele et al. (2001) of the original formulation by 1. Recorded daily values of the precipitation, expressed Chamberlain (1970): in tenths of millimeter, were collected from the NOAA Global Historical Climatology Network (Menne et al. 2012) fdry ¼ M 1− exp −a Ydry ; ð16Þ for all measuring stations of New Mexico for 16, 17, and where M is the maximum interception (unitless), a is the fo- 18 July 1945. The recorded values for the day of the test 2 −1 (16 July 1945) are shown on Fig. 2. An interpolation liar interception constant [m kg (dry mass)], and Ydry is the standing crop biomass [kg (dry mass) m−2]. The values procedure, using inverse distance weighting of the three o of M may vary according to the type of vegetation and local closest observations within a radius of 0.3 and no spatial plant growth (e.g., row crops vs. pasture with no rows), correlation other than distance, was used to estimate the while the values of a may depend on particle size and chem- daily precipitation at all precincts of New Mexico. In cases where rainfall occurred during the passage ical form, and the values of Ydry may vary according to the ecozone and the type of vegetation (Thiessen and Hoffman of the radioactive cloud, as assumed in the following cal- 2018). It is assumed here that the values of a are independent culations, fwet could be calculated (Thiessen and Hoffman of particle size in the range from 1 to 50 mm. This assump- 2018) as: tion is based on the summary by Pröhl (2009) of interception fwet ¼ minðð1; LAI k S=R ½1– expðÞ−R lnðÞ2 =c k S 17Þ measurements obtained from field experiments for dry deposits; in that summary, the interception fractions or the absorption co- where: efficients do not show a clear variation with particle sizes from LAI = the leaf area index, a dimensionless quantity that characterizes plant canopies (unitless). It is defined as the 1to50mm. The selected values of M, a, and Ydry are: one-sided green leaf area per unit ground surface area. LAI • M =0.85, a =2.8,Ydry = 0.3 for vegetables grown in gardens ranges from 0 (bare ground) to over 10 (dense conifer forests); for human consumption; k = a unitless constant that quantifies the ability of an • M =1, a =2.8,Ydry = 1 for fruit, berries, etc. directly exposed element to be attached to the vegetation. Pröhl (2009) pro- during fallout; and posed values for k of 0.5 for anions (Br, I), 1 for monovalent • M =1, a = 2.8, Ydry = 0.3 for pasture grasses and other cations (Rb, Cs), and 2 for polyvalent cations (Sr, Ba). The vegetation grazed by cows, sheep, and swine. available information for the other elements under consider- ation (Fe, Co, Cu, As, Y, Zr, Nb, Mo, Tc, Ru, Rh, Pd, Ag, fwet is the fraction of the activity attached to particles of Cd, In, Sn, Sb, Te, La, Ce, Pr, Nd, Pm, Sm, U, Np, Pu) is all sizes that is intercepted and initially retained by very sparse; the proposed value for those elements is 1.25, www.health-physics.com 412 Health Physics October 2020, Volume 119, Number 4 Table 10. Selected values of the ratios of radionuclide Z and of the deposited total b activity at time H+12 h, according to the degree of fractionation, R/V. Nuclide R/V =0.5 R/V =1 R/V =1.5 R/V =2 R/V =3

112Ag 2.45 × 103 1.61 × 103 1.20 × 103 9.61 × 104 6.86 × 104 77As 8.99 × 105 5.94 × 105 4.44 × 105 3.55 × 105 2.54 × 105 139Ba 2.45 × 103 1.61 × 103 1.20 × 103 9.61 × 104 6.86 × 104 140Ba 5.74 × 103 4.47 × 103 3.84 × 103 3.46 × 103 3.04 × 103 83Br 2.38 × 103 1.57 × 103 1.17 × 103 9.36 × 104 6.69 × 104 115Cd 4.22 × 104 2.78 × 104 2.08 × 104 1.66 × 104 1.18 × 104 117Cd 3.59 × 104 2.37 × 104 1.77 × 104 1.41 × 104 1.01 × 104 141Ce 1.45 × 103 1.43 × 103 1.43 × 103 1.43 × 103 1.43 × 103 143Ce 2.07 × 102 2.73 × 102 3.06 × 102 3.25 × 102 3.48 × 102 144Ce/144Pr 2.21 × 104 2.92 × 104 3.27 × 104 3.48 × 104 3.72 × 104 60Co 9.00 × 107 1.19 × 106 1.33 × 106 1.41 × 106 1.51 × 106 137Cs 9.90 × 106 5.89 × 106 3.09 × 106 2.18 × 106 1.28 × 106 64Cu 7.23 × 103 9.54 × 103 1.07 × 102 1.14 × 102 1.22 × 102 55Fe 2.16 × 107 2.85 × 107 3.20 × 107 3.40 × 107 3.64 × 107 131I 7.60 × 103 5.01 × 103 3.74 × 103 2.99 × 103 2.13 × 103 132I 2.22 × 102 1.47 × 102 1.09 × 102 8.73 × 103 6.23 × 103 133I 8.73 × 102 5.76 × 102 4.30 × 102 3.43 × 102 2.45 × 102 135I 9.45 × 102 6.23 × 102 4.65 × 102 3.71 × 102 2.65 × 102 117mIn 1.04 × 103 6.88 × 104 5.14 × 104 4.10 × 104 2.93 × 104 140La 1.09 × 103 8.47 × 104 7.27 × 104 6.56 × 104 5.76 × 104 141La 4.26 × 102 4.21 × 102 4.20 × 102 4.19 × 102 4.19 × 102 142La 3.19 × 103 4.20 × 103 4.70 × 103 5.00 × 103 5.36 × 103 99Mo 1.59 × 102 2.10 × 102 2.35 × 102 2.50 × 102 2.68 × 102 95Nb 6.23 × 106 8.21 × 106 9.20 × 106 9.78 × 106 1.05 × 105 147Nd 1.71 × 103 2.25 × 103 2.52 × 103 2.69 × 103 2.88 × 103 149Nd 1.56 × 103 2.05 × 103 2.30 × 103 2.45 × 103 2.62 × 103 239Np 2.12 × 101 2.80 × 101 3.14 × 101 3.34 × 101 3.57 × 101 240mNp 4.44 × 103 5.86 × 103 6.56 × 103 6.98 × 103 7.48 × 103 109Pd 7.89 × 103 5.20 × 103 3.88 × 103 3.10 × 103 2.21 × 103 141Pm 4.71 × 103 6.21 × 103 6.95 × 103 7.39 × 103 7.92 × 103 151Pm 4.36 × 103 5.76 × 103 6.45 × 103 6.86 × 103 7.34 × 103 143Pr 5.71 × 104 7.54 × 104 8.44 × 104 8.98 × 104 9.61 × 104 145Pr 2.74 × 102 3.62 × 102 4.05 × 102 4.31 × 102 4.62 × 102 239Pu 3.16 × 107 4.15 × 107 4.67 × 107 5.09 × 107 5.41 × 107 240Pu 2.88 × 108 3.78 × 108 4.26 × 108 4.64 × 108 4.93 × 108 88Rb 1.25 × 102 8.25 × 103 6.16 × 103 4.91 × 103 3.51 × 103 105Rh 3.64 × 102 2.40 × 102 1.79 × 102 1.43 × 102 1.02 × 102 103Ru/103mRh 2.65 × 103 1.75 × 103 1.31 × 103 1.04 × 103 7.44 × 104 105Ru 6.19 × 102 4.08 × 102 3.05 × 102 2.43 × 102 1.74 × 102 106Ru/106Rh 3.58 × 104 2.36 × 104 1.76 × 104 1.41 × 104 1.00 × 104 125Sb 1.06 × 106 7.03 × 107 5.26 × 107 4.20 × 107 3.00 × 107 127Sb 1.53 × 103 1.01 × 103 7.53 × 104 6.01 × 104 4.29 × 104 129Sb 1.67 × 102 1.10 × 102 8.24 × 103 6.57 × 103 4.69 × 103 153Sm 1.48 × 103 1.95 × 103 2.18 × 103 2.32 × 103 2.48 × 103 121Sn 8.18 × 104 5.39 × 104 4.03 × 104 3.21 × 104 2.29 × 104 127Sn 5.63 × 104 3.71 × 104 2.77 × 104 2.21 × 104 1.58 × 104 89Sr 9.06 × 104 5.43 × 104 4.05 × 104 3.23 × 104 2.31 × 104 90Sr 4.44 × 106 2.66 × 106 1.99 × 106 1.59 × 106 1.13 × 106 91Sr 4.61 × 102 3.59 × 102 3.01 × 102 2.67 × 102 2.29 × 102 92Sr 1.12 × 102 1.48 × 102 1.66 × 102 1.77 × 102 1.89 × 102 99mTc 1.09 × 102 1.44 × 102 1.61 × 102 1.71 × 102 1.83 × 102

Continued next page

www.health-physics.com c A. BOUVILLE ET AL. 413 Table 10. (Continued)

Nuclide R/V =0.5 R/V =1 R/V =1.5 R/V =2 R/V =3

129Te 1.90 × 102 1.25 × 102 9.35 × 103 7.46 × 103 5.32 × 103 131mTe 4.87 × 103 3.21 × 103 2.40 × 103 1.91 × 103 1.37 × 103 132Te 2.15 × 102 1.42 × 102 1.06 × 102 8.45 × 103 6.03 × 103 133mTe 4.15 × 105 2.74 × 105 2.04 × 105 1.63 × 105 1.16 × 105 237U 1.37 × 102 1.80 × 102 2.02 × 102 2.15 × 102 2.30 × 102 240U 4.40 × 103 5.81 × 103 6.50 × 103 6.91 × 103 4.58 × 103 90Y 5.47 × 107 3.62 × 107 2.71 × 107 2.16 × 107 1.54 × 107 91mY 2.96 × 102 2.30 × 102 1.93 × 102 1.71 × 102 1.47 × 102 92Y 3.89 × 102 5.13 × 102 5.74 × 102 6.11 × 102 6.54 × 102 93Y 3.64 × 102 4.79 × 102 5.37 × 102 5.71 × 102 6.12 × 102 95Zr 6.55 × 104 8.64 × 104 9.67 × 104 1.03 × 103 1.10 × 103 97Zr/97mNb 3.66 × 102 4.83 × 102 5.41 × 102 5.75 × 102 6.16 × 102 with a uniform probability distribution from 0 to 2.5 (Thiessen Variation with time of the normalized activities of and Hoffman 2018); nuclide Z S = the water storage capacity of the plant (mm). Pröhl The variation with time of the normalized activities of (2009) proposed values for S of 0.2 mm for grass, cereals, radionuclide Z depositedonthegroundandonvegetation and corn (maize), and of 0.3 mm for all other crops; was calculated using the normalized values for the Trinity R = the total amount of rain during a single event (mm), and test provided by Hicks (1985) for a range of times during c = a unitless constant dependent on the type of plant the first year after the test. When needed, account was taken and ambient conditions (e.g., rainfall intensity and wind speed). of the environmental removal processes. For most radionu- Thiessen and Hoffman (2018) proposed a log-triangular clides, the variation with time was expressed with simple ex- probability distribution with a mode of 3 and lower and up- ponentials. More complex analytical expressions were used per bounds of 0.5 and 5, respectively. when the radionuclide Z had a radioactive precursor pro- duced by fission and radioactive equilibrium was not reached during the first few minutes after the shot. This was the case for 112Ag, 77As, 141Ce, 131I, 117mIn, 140La, 95Nb, 239Np, 149Pm, 105Rh, 125Sb, 99mTc, 90Y, 91mY, and 92Y. Normalized time-integrated concentrations in air and foodstuffs (ICpw/Apw) Normalized time-integrated concentrations in air were considered for two inhalation pathways: acute inhalation during the passage of the radioactive cloud and protracted inhalation from resuspension of deposited activity, for exam- ple during windstorm events, in the first year after the test. The ingestion pathways include the consumption of drinking water and of eight foodstuffs, namely leafy vegeta- bles, fruit vegetables, fruit and berries, cow’s milk, cow cheese, beef, mutton, and pork. The consumption of goats’ milk was also considered in the dietary survey; however, it was reported so infrequently that it could not be assumed or quantified on a population basis (Potischman et al. 2020). The origin of the contamination of all foodstuffs is the fallout deposition on plant leaves during the cloud pas- sage. The normalized time-integrated concentrations corre- spond to the food products consumed, considering the time delay between production and consumption, as well as the Fig. 2. Map of recorded precipitation, expressed in tenths of millimeters changes in concentrations due to the culinary preparation per day, on 16 July 1945 in all measuring stations of New Mexico (derived from the Global Historical Climatology Network–Daily Database; of the foodstuffs. In the absence of detailed information Menne et al. 2012). on local or regional patterns of foodstuffs other than cow’s www.health-physics.com 414 Health Physics October 2020, Volume 119, Number 4

−1 milk, those foodstuffs were assumed to have been produced vd is the average deposition velocity, expressed in m s ,for in the same precinct in which they were consumed. particles <10 mm. For Trinity, we found vd to be about 2 The inhalation pathways are discussed first. 10−3 ms−1, as estimated from Sehmel (1980) for particles Inhalation during the passage of the cloud of about 5 mm diameter (midpoint of the respirable range). The method used to estimate the time-integrated con- centrations in ground-level air during the passage of the Inhalation dose due to resuspended material The time-integrated concentrations in ground-level air, cloud is the same as in Simon et al. (1990). In the absence −3 of measured air concentrations, their values were derived using inBqdm , due to resuspension during the first year after the relationship proposed by Chamberlain and Chadwick the test, ICres(Z,L), were derived from measured values of (1953) between the dry deposition flux and the air concen- the resuspension factor Sf(t), which is the ratio of the air con- tration. The ratio of the two quantities, termed deposition centration and of the deposited activity (Anspaugh et al. velocity, is equal to the ratio of the deposition density and of 2002; Maxwell and Anspaugh 2011): the time-integrated concentrations in air. The dry deposition values have been reported to range over three orders of magni- Z365 R tude as they vary according to micrometeorological variables ICresðÞ¼Z; L Agd Z; L; ; t Sf ðÞt dt ð19Þ (e.g., atmospheric stability, temperature, and wind velocity), V TOA properties of the depositing material (e.g., particle size and −1 solubility), and surface variables (e.g., canopy growth and where the resuspension factor Sf (t), in m , is expressed as: structure) (Sehmel et al. 1980). Therefore, there are substan- tial difficulties in determining suitable deposition-velocity − − : − − : − S ðÞ¼t 10 5 e 0 07 t þ 710 9 e 0 002 t þ 10 9 ð20Þ values, as the values of many influencing parameters may f be unknown. Despite the difficulties associated with the de- where the time t after deposition at TOA is expressed in termination of the appropriate deposition velocity, the rela- days. tionship proposed by Chamberlain and Chadwick (1953) is For a radionuclide Z with no significant input of a pre- generallyusedtoestimatetheinhalation dose in the absence cursor after TOA and a radioactive decay constant, l in d−1, of measured air concentrations, and it was, for example, used the activity deposited on the ground varies as: in two studies related to fallout from tests detonated at the Nevada Test Site (NCI 1997; Simon et al. 1990). ; ; = ; The method proposed by Simon et al. (1990) also takes Agd ðÞZ L R V t into consideration the facts that the average particle size de- −l − ¼ A ðÞZ; L; R=V; TOA e ðÞt TOA ð21Þ creases as TOA increases and that the large particles are not gd respirable, i.e. cannot penetrate deep into the lung. Because and the time-integrated concentration in air from t = TOA to the Trinity test occurred in the summer before indoor air 1 t = 365 d is obtained as: conditioning was widely available, we assumed that the 2 buildings were naturally ventilated without filtration by open   R 10−5 −ðÞ0:07þl t1 −ðÞ0:07þl t2 windows, and so the air concentrations indoors were approx- ICresðÞ¼Z; L Agd Z; L; ; TOA e Þ−e V 0:07 þ l imately equal to those outdoors. In brief, the relationship be- −9  710 −ðÞ0:002þl t −ðÞ0:002þl t tween the deposited activity on the ground, Agd,andthe þ e 1 Þ−e 2 0:002 þ l time-integrated concentration of the respirable-sized particles −9  −3 10 − l − l in air, IC in Bq s m ,wasestimatedtobe(Simonetal. ðÞt1 − ðÞt2 : cd þ l e e ð22Þ 1990): For a radionuclide Z with a radioactive precursor produced A Z; L; R ; TOA −1 gd V by fission, dp, with a radioactive decay constant, ldp in d , ICcdðÞ¼Z; L xforðÞTOA ; ð18Þ vd the activity of the decay product deposited on the ground varies as: where: for is the fraction of the time-integrated air concentra- Agd ðÞdp; L; R=V; t tion on particles of respirable sizes, taken to be 10 mmor less. Simon et al. (1990) derived a function based on time ¼ Agd ðÞZ; L; R=V; TOA x ldp= ldp–l x expðÞ−l t –exp −ldp t Þ of arrival of fallout, TOA in hours, based on Nevada test þAgdðÞdp; L; R=V; TOA x exp −ldp t I ð23Þ data. For Trinity, however, we estimated for to be somewhat smaller and to vary from 0.08 to 0.2, depending on distance; and the time-integrated concentration in air from t1 = TOA and to t2 =365disobtainedas: www.health-physics.com c A. BOUVILLE ET AL. 415

ICres ðÞdp; L −5 ¼ Agd ðÞZ; L; R=V; TOA x ldp= ldp–l x 10 =ðÞ0:07 þ l x −5 ðÞÞexpðÞ−ðÞ0:07 þ l t1 – expðÞ−0:07 þ l t2 − 10 = 0:07 þ ldp x exp − 0:07 þ ldp t1 – exp −0:07 þ ldp t2 Þ −9 þ 7 10 =ðÞ0:002 þ l x expðÞ−ðÞ0:002 þ l t1 – expðÞ−0:002 þ l t2ÞÞ −9 − 7 10 =0:002 þ ldp x exp− 0:002 þ ldp t1 – exp −0:002 þ ldp t2ÞÞ −9=l −l – −l − −9=l þ 10 x expðÞt1 expðÞt2 ð 10 dp x exp −ldp t1 – exp −ldp t2 þAgd ðÞdp; L; R=V; TOA −5 x ½ 10 = 0:07 þ ldp x exp − 0:07 þ ldp t1 – exp −0:07 þ ldp t2 −9 þ 7x10 =0:002 þ ldp x exp − 0:002 þ ldp t1 – exp −0:002 þ ldp t2 Þ Ã −9 þ 10 =ldp x exp −ldp t1 – exp −ldp t2 gI: ð24Þ Ingestion of drinking water Drinking water from cisterns, wells, public network, rivers, and acequias (or “irrigation ditches”) could have been contaminated with radioactive fallout from the Trinity test. The following assumptions were made regarding the origin of the drinking water: • Tap water was derived from a river that flowed in 331 precincts (Fig. 3), including all precincts where main Fig. 3. Map of rivers in New Mexico (derived from US Census). towns and cities are located; • Cisterns and wells provided water for the other precincts precinct during the day of fallout transferred to the cistern with 50% of the consumption of drinking water in those the activity deposited on the roof. The volume of water precincts assumed to originate from cisterns, and the available immediately after rainfall at TOA was, therefore, other 50% from wells; and V +(R(L) S ). It was also assumed that rainfall that oc- • roof Acequias, i.e., irrigation ditches, that carried water, pri- curred during the days before and after the day of fallout did marily from rivers, to fields for irrigation, were rarely not remove any of the activity deposited on the roof; used as sources of drinking water for humans; hence, −1 le = l + ld,ind ,whereld represents the rate of de- they were not taken into consideration in the dose assess- crease of the concentration in the cistern water, assumed to ment. Contamination of water in acequias as a source for correspond to a half-time of 15 d, due to the sedimentation livestock was also not quantified for reasons that the sur- of the activity on the bottom of the cistern and on the replen- face area for collecting fallout would be extremely small ishment of the cistern with uncontaminated water. Because compared to lakes and rivers. of this relatively short half-time, which was assumed to be applicable to all radionuclides, the cistern water was only The contamination of drinking water from cisterns, contaminated during the first few months after the test; and wells, public network, and rivers was estimated as follows. TD = the time delay between rainfall at TOA and the Cisterns. The normalized time-integrated concentrations consumption of water, taken to be equal to 0.1 d. in drinking water from cisterns were obtained as: Wells. The water originating from wells was unlikely to −l : e TD ICcistðÞZ L Sroof e have been contaminated to a substantial degree by the fallout ¼ PRLðÞjA ð25Þ ; ; R ; gd l Agd Z L V TOA V þ RLðÞSroof e from Trinity. The time-integrated concentration of any radionu- clide from Trinity in well water was taken to be equal to zero. where Sroof = the area of the roof from which rainwater was Public water networks. In the absence of measure- collected, taken to be 10 m2; ments in New Mexico after the Trinity test, the estimation of V is the volume of water in the cistern at the time of the the normalized time-integrated concentrations in tap water test; V is assumed to correspond to an average monthly rain- is very uncertain. For that reason, we derived an empirical −2 fall of 30 mm (= L m ), so that V = Sroof x 30 = 300 L; relationship between the radionuclide concentration in tap P(R(L)|Agd) = 1 if rainfall occurred during the day of and rainwater. Measurements made in French Polynesia after fallout in precinct L; the nuclear tests carried out at Mururoa and Fangataufa P(R(L)|Agd) = 0 if rainfall did not occur during the day showed that the b activities in tap water were about 15 to of fallout. It was assumed that the rain R(L)thatfellinthe 20 times lower than in rainwater (Ministère 2006). In the www.health-physics.com 416 Health Physics October 2020, Volume 119, Number 4 absence of information for New Mexico or for specific weathering processes, assumed to correspond to half-times radionuclides, it was assumed that the time-integrated of environmental removal of 10 d for stable iodine, 13 d concentrations of any radionuclide from Trinity in tap water for stable strontium, 14 d for stable cesium, and 15 d for were 20 times lower than in rainwater. Rivers were assumed all other elements (Thiessen and Hoffman 2018). to be the main source of water for the public water supply. The calculation of the contamination of river water considered Ingestion of fruit vegetables (FV) the probability that rainfall (occurring during a 30-min Fruit vegetables, FV, were assumed to be ready for harvest storm every day) happened during the passage of the at the time of fallout deposition in 1945, and that they also radioactive cloud at TOA. Other assumptions made were: continued to ripen during the following 90 d after TOA. (1) the contamination of the river water was only due to Based on the results provided during the focus-group sessions wet deposition, and (2) the time-integrated concentration and the key-informant interviews, the typical fruit vegetable was obtained using a half-time of 15 d for the removal consumed was cooked squash (Potischman et al. 2020). due to environmental loss processes. The normalized time-integrated concentrations were calcu- Ingestion of leafy vegetables (lettuce, spinach, etc.) lated as: The activity intakes by man from leafy vegetables,LV, resulted from the direct deposition of radionuclides from ICFV ðÞZ; L TFðÞ Z; FV CFðÞ Z; FV the cloud at TOA. The leafy vegetables were assumed to ¼ ; ð27Þ A Z; L; R ; TOA Y ðÞFV l ðÞZ; FV be ready to be harvested at the time of fallout deposition veg V wet e in 1945. The normalized time-integrated concentrations, in where: Bqdkg−1 per Bq m−2, were expressed as: • the deposition A corresponds to the deposition on the IC ðÞZ; L CFðÞ Z; LV veg LV leaves and on the fruit of the plant; R ¼ ð26Þ Aveg Z; L; ; TOA YwetðÞLV leðÞZ; LV V • the values of le and of Ywet are the same as those selected for the leafy vegetables; where: • the transfer factor from deposition to fruit (TF)isas- −2 • the deposition, Aveg,inBqm , corresponds to the depo- sumedtobethesumofTF1 and TF2: (1) for fruit vege- sition on leafy vegetables; tables ready to be harvested immediately after TOA, TF1 • the culinary factor, CF(Z, LV), has 2 components. (1) corresponds to direct deposition on the fruit and is equal CF1(Z, LV), which is the decrease in concentration due to 1 since translocation does not have to be taken into ac- to the culinary preparation, consisting in washing and re- count; (2) for fruit vegetables that had not ripened yet at moving the outer leaves for the leafy vegetables; according TOA, the translocation factor TF2 from leaves to fruit to Simmonds and Linsley (1982): CF1(Z, LV) =0.2forSr, must be used. For this pathway, TF2 is expressed as the Cs, and Pu, this value was used for all radionuclides; and fraction of deposition on the leaves (per unit area of (2) CF2(Z, LV), which is the decrease in concentration ground) that is transferred to the fruit. Selected values between harvesting and consumption. The values of CF2 of the probability distribution of TF2 for the elements (Z, LV) were calculated accounting for the delay between considered in this study are presented in Table 11. In harvesting and consumption to be equal to 1 d for urban the calculations for the translocation of the specific ra- precincts and 0.1 d for rural precincts, corresponding to dionuclides considered in this study, the time taken for the minimum delay time recommended in Table 81 of the contamination deposited on the leaves to be transferred IAEA (2010) for urban precincts and to an even lower to the fruit was taken to be equal to 10 d. These translo- time for rural precincts. The rationale for selecting such cation factors, which depend on the plant characteristics, low values is that the delay times in 1945, when refriger- the stage of plant growth, and the characteristics of the ation was not yet widespread, were likely to have been deposition (wet or dry), have considerable uncertainty lower than those for more modern times as reported in (Thiessen and Hoffman 2018); and IAEA (2010). The values of CF(Z, LV) were obtained as • CF(Z, FV) is the product of CF1 and CF2. The values of the products of CF1(Z, LV) and CF2(Z, LV); CF1 were subjectively estimated to be 0.7 for all radionu- • Ywet(LV) is the standing crop biomass expressed in terms clides. The values of CF2(Z, FV) were calculated account- of wet weight [kg (wet weight) m−2]. Taking the water ing for the delay between harvesting and consumption to content of leafy vegetables to be 90% (IAEA 2010), the be equal to 2 d for urban precincts and 0.2 d for rural pre- ratio of dry to wet weight is 10, so that Ywet(LV) =3kg cincts, corresponding to the minimum delay time recom- (wet weight) m−2; mended in IAEA (2010) for urban precincts and to an −1 • le = l + ld where ld,ind , represents the rate of de- even lower time for rural precincts. The rationale for selecting crease of the concentrations in leafy vegetables due to such low values is that the delay times in 1945 were

www.health-physics.com c A. BOUVILLE ET AL. 417 likely to have been smaller than those for more modern with rivers and the dry lands areas in the other precincts, times reported in IAEA (2010). without rivers (Fig. 3). Under these assumptions, the radionuclide intake by dairy cows, ICOW, in Bq, due to Ingestion of fruit and berries (FB) direct deposition on pasture grass only occurred in Fruit and berries, FB,wereassumedtobereadyto 1945 during a few months after the explosion and was be harvested at the time of fallout deposition, TOA.Based calculated as: on the results provided from the focus groups and the key PGðÞ L ICOWðÞ¼ Z; L AvegðÞZ; L; R=V; TOA x ð28Þ informant interviews, the fruits consisted mainly of apples YwetðÞL leðÞZ and different types of berries (Potischman et al. 2020). The normalized time-integrated concentrations were cal- • Estimation of the transfer from feed to milk. The culated using equation 27 and appropriate values for its time-integrated concentrations in cow’s milk,ICCM in − parameters, the main difference being in the estimation Bq d L 1, are estimated using the transfer factor from an- of CF1. Reported values for the reduction in concentra- imal feed-to-milk, usually designated as Fm.Theyare tions due to washing berries are 0.76 for Cs and 0.64 for calculated as: Sr (Carini 1999). For the purposes of this study, a value ; ; : of 0.7 was selected for CF1 and was assumed to be the ICCM ðÞ¼Z L ICOWðÞ Z L xFmðÞelement xHLZðÞ ð29Þ same for all radionuclides. −1 Selected values for the central estimates of Fm,indL ,for Ingestion of fresh cow’smilk,CM the stable elements corresponding to the radionuclides In order to estimate the time-integrated concentrations under consideration are presented in Table 12. In order in fresh cow’s milk, the intake of radionuclides by the cow, to obtain the Fm value for a specific radionuclide, Z,the the transfer from feed to milk, and the origin and amount of Fm value for the corresponding element was multiplied milk consumed in the precinct need to be taken into consid- by a factor HL(Z) = Tr/(Tr +Tb), where Tr, in days, is the ra- eration as follows: dioactive half-life of radionuclide Z and Tb is the biological • Estimation of the intake of radionuclides by the cow. Milk half-time of the element in the cow, taken to be equal to 2 d cows are generally fed hay and grains but were assumed to (Thiessen and Hoffman 2018); and be highly dependent in New Mexico in 1945 on pasture • Estimation of the origin and amount of milk consumed in grass during the pasture season. Because of the importance the precinct. Because the number of milk cows was not of pasture grass to the radionuclide intake of dairy cows, sufficient in some of the counties in New Mexico to we assumed pasture grass to be the primary source of provide enough milk for the population of the county, contamination and that the other pathways of intake, such it was necessary to estimate the origin and amount of as inhalation and soil consumption, were negligible. milk imported from other counties. This was done The contamination of pasture grass was due to direct based on information on the milk production, use, deposition at TOA and decreased with time because of and consumption, which was estimated on a county radioactive decay and environmental removal processes, basis for the 1950s (NCI 1997). In that analysis, the with half-times of up to 15 d. The consumption of pasture state of New Mexico was divided into five relatively grass PG by cows in New Mexico, was taken to be homogeneous milk regions, MR, and the distribution different in irrigated areas, where there was a heavy of milk was first considered to occur within the same dependence on hay for most of the year, rather than fresh milk region (Fig. 4). As shown in Table 13, MR 351 pasture, and in dry land areas, where most of the cow’s wastheonlyregionwithanexcessofmilk.Theother intake was from pasture. The consumption of pasture milk regions received some milk from MR 351 and grass was assumed to be 2.5 kg d−1 (dry) (NCI 1997) some from other states (Arizona, Colorado, Kansas, and in irrigated areas and 5 kg d−1 (dry) in dry land areas. Texas). The estimated annual volumes of milk consumed The irrigated areas were assumed to be the precincts in each MR in the 1940s, VTOT in kL, are given in

Table 11. Element-dependent values of the translocation factor from leaves to fruit. Element Fe, Zr, Nb, Sb, Ce, Pr Co Sr, Ba Ru Sb Te, I, Cs Nd Pu, U

Distribution type Log-uniform Minimum (%) 0.1 0.3 0.01 0.01 0.1 0.5 0.1 0.0001 Maximum (%) 10 30 10 1 10 50 10 0.01 Average (%) 1 3 0.3 0.1 1 5 1 0.001

www.health-physics.com 418 Health Physics October 2020, Volume 119, Number 4 Table 14, along with their origin, distributed into three categories: local (from the same milk region, VL), imported from MR 351 in New Mexico (VNM), and imported from other states (VOS). The imported volumes of milk were distributed in the counties of the milk region MR according to their needs. Prior to that operation, a distribution within the MR was carried out in order to provide enough milk to the counties with small deficits to be in, or closer to, equilibrium. It was not necessary to transfer milk from one county to another in MR 352, 353, and 354 because all counties of those MRs were in deficit.

Estimation of the time-integrated concentrations in milk The time-integrated concentration of radionuclide Z in −1 the milk produced in precinct L, ICPCM(Z, L) in Bq d L , was expressed as:

ICPCM ðÞ¼Z; L ICOWðÞ Z; L xFmðÞZ : ð30Þ

Different procedures were used for the calculation of the time-integrated concentrations in consumed milk in the pre- Fig. 4. Map of milk regions (derived from NCI 1997). cinct according to whether: (1) there was excess milk in the county where the precinct is located, (2) there was a deficit average time-integrated concentration of radionuclide Z in of milk in the county but it was met with a supply from an- the milk produced in county CTY1 where L1 is located other county in the same milk region, or (3) the deficit of was equal to the mean of the values obtained for all pre- milk in the county was met with supplies from other milk cincts of the county: regions of New Mexico and regions from other states.

Precincts in category 1 (MC1): excess of milk in the ICPCM;AV ðÞ¼Z; CTY1 ½∑ICPCM ðÞZ; L1 =NL1ðÞð32Þ county. Assuming a delay of 2 d between production and where N is the number of precincts in the county. consumption, the time-integrated concentration, ICCM,MC1 −1 inBqdL , of radionuclide Z in the milk consumed in precinct Precincts in category 2 (MC2): deficit of milk in the L1 locatedinacountyCTY1 with excess milk was obtained as: county, met with supply from the same MR. The volumes −lðÞZ x 2 of milk transferred from one county to another were denoted ICCM;MC1ðÞ¼Z; L1 ICPCM ðZ; L1Þ x e : ð31Þ as VTR, Specifically, in MR 351, some of the excess milk In the absence of information on the production of cows’ from Quay county (1,092 kL) was transferred to Guadalupe milk at the level of the precinct, it was assumed that the (VTR = 181 kL), Mora (VTR = 58 kL), and San Miguel counties (VTR = 853 kL), so that the excess volume of milk Table 12. Central estimates of the transfer factor from feed-to-milk for in Quay county was reduced to 575 kL. In MR 355, excess −1 cows (d L ) for the stable elements corresponding to the radionuclides milk from Catron county (92 kL) was transferred to Socorro considered in the study. (VTR = 70 kL) and to Grant counties (VTR = 22 kL), and ex- 1 1 1 Element Fm (d L )ElementFm (d L ) Element Fm (d L ) cess milk from Sierra (163 kL) was transferred to Hidalgo Fe 3.7 × 105 Tc 0 Cs 4.9 × 103 (VTR = 118 kL) and to Luna counties (VTR =45kL).The Co 3.2 × 104 Ru 9.4 × 106 Ba 1.8 × 104 volumes transferred were sufficient to cover the consump- Cu 1.8 × 104 Rh 1.0 × 104 La 1.0 × 104 tion of milk in Guadalupe, Mora, San Miguel, Hidalgo, and As 1.4 × 104 Ag 0 Ce 1.5 × 105 Sierra counties, but served only to reduce the deficit in Grant Br 0 Pd 1.0 × 104 Pr 1.0 × 104 and Luna counties (Table 13). 4 4 Rb 0 Cd 2.6 × 10 Nd 1.0 × 10 The time-integrated concentration of radionuclide Z in 3 4 Sr 1.3 × 10 In 0 Pm 1.0 × 10 the milk produced, or more exactly, available for consumption Y1.0×104 Sn 1.0 × 104 Sm 1.0 × 104 6 5 3 in a precinct L2 of a county CTY2 with a deficit milk VTR that Zr 3.6 × 10 Sb 3.8 × 10 U 2.5 × 10 was provided by another county with excess milk CTY1 from Nb 4.1 × 107 Te 3.2 × 104 Np 1.0 × 104 the same milk region (either MR 351 or MR 355) has two com- Mo 1.2 × 103 I6.0×103 Pu 3.6 × 105 ponents: (1) the milk for fluid use, TMFU, available in CTY2; www.health-physics.com c A. BOUVILLE ET AL. 419 Table 13. Estimated annual production, distribution, and consumption of fresh cows’ milk in New Mexico (based on NCI 1997). County name Milk Production TMFUa Consumption EXCb EXC2c

region kL kL kL kL kL COLFAX 351 3,192 1,897 1,697 200 200 CURRY 351 6,400 3,805 2,991 814 814 DE BACA 351 957 569 357 212 212 GUADALUPE 351 852 506 687 181 0 HARDING 351 1,060 630 277 353 353 MORA 351 1,297 771 829 58 0 QUAY 351 5,243 3,117 1,450 1,667 575 ROOSEVELT 351 16,134 9,592 1,786 7,806 7,806 SAN MIGUEL 351 3,210 1,908 2,761 853 0 TORRANCE 351 3,258 1,936 806 1,130 1,130 UNION 351 4,970 2,955 746 2,209 2,209 BERNALILLO 352 6,082 5,590 21,374 15,784 15,784 MCKINLEY 353 134 123 3,459 3,336 3,336 RIO ARRIBA 353 1,886 1,733 2,700 967 967 SANDOVAL 353 612 563 2,710 2,147 2,147 SAN JUAN 353 3,594 3,303 3,632 329 329 SANTA FE 353 1,017 934 4,495 3,561 3,561 TAOS 353 2,028 1,205 1,821 616 616 VALENCIA 353 5,000 2,972 3,234 262 262 CHAVES 354 2,748 2,525 5,239 2,714 2,714 DONA ANA 354 3,101 2,849 5,280 2,431 2,431 EDDY 354 4,205 3,864 4,922 1,058 1,058 LEA 354 3,643 3,348 4,420 1,072 1,072 LINCOLN 354 1,258 748 826 78 78 OTERO 354 1,403 1,289 2,659 1,370 1,370 CATRON 355 745 443 351 92 0 GRANT 355 1,279 1,175 2,233 1,058 1,036 HIDALGO 355 728 433 551 118 0 LUNA 355 825 758 1,008 250 205 SIERRA 355 1,537 913 750 163 0 SOCORRO 355 1,102 1,012 1,082 70 0

aTMFU = total milk annually available for fluid use (kL). bEXC = excess or deficit of milk (kL), prior to distribution. cEXC2 = excess or deficit of milk (kL), after transfer from a county in the same milk region. and (2) the deficit, VTR(CTY1, CTY2), covered in its entirety, distribution of TMFU and of VTR at the precinct level, it with the exception of Grant and Luna counties, by the transfer was assumed that the ratio for CTY2 of TMFU(CTY2/ of milk from CTY1. In the absence of information on the [(TMFU(CTY2) + VTR(CTY1, CTY2)] had the same value for each precinct L2:

Table 14. Origin and amount of consumed milk in each milk region ICPCM;MC2ðÞZ; L2 of New Mexico (kL). ¼ ½ðÞICPCM ðÞZ; L2 x TMFUðÞ CTY2 Origin and annual amount of consumed milk (kL) þðÞÞICPCM ; AVðÞ Z; CTY1 x VTRðÞ CTY1; CTY2 =½ðTMFUðÞ CTY2 VL: from VNM: imported VOS: imported VTOT: Milk region MR same MR from MR 351 from other States total þ VTRðÞ CTY1; CTY2 :

351 14,599 0 0 14,599 ð33Þ 352 5,590 9,168 6,706 21,464 Assuming a delay of 3 d between availability and con- 353 10,833 1,834 9,119 21,786 sumption, the time-integrated concentration, IC in 354 14,623 1,523 7,197 23,343 CM,MC2 Bq d L−1, of radionuclide Z in the milk consumed in precinct 355 4,734 458 782 5,974 L2 located in a county CTY2 with deficit milk supplied by www.health-physics.com 420 Health Physics October 2020, Volume 119, Number 4 excess milk from another county, CTY1,fromthesamemilk concentration of radionuclide Z in the milk consumed in pre- region was obtained as: cinct L3 located in Grant or in Luna county was obtained as:

ICCM; MC3ðÞZ; L3 ; ; −lðÞZ x 3: hi ICCM;MC2ðÞ¼Z L2 ICPCM;MC2ðÞZ L2 xe ð34Þ −λðÞZ x 3 ¼f ICPCM; MC2ðÞZ; L3 xðÞ TMFUðÞþ CTY3 VTRðÞ CTY1; CTY 3 xe hi ; ; −λðÞZ x 4 = : For Grant and Luna counties, additional amounts of þ ICPCM; AV ðÞZ MR351 xVNMðÞ MR351 CTY 3 xe g CONSðÞ CTY 3 milk had to be imported from other areas in order to cover ð38Þ in its entirety the consumption of milk in those counties. Ingestion of cow cheese, CC The normalized time-integrated concentrations in soft cow Precincts in category 3: deficit of milk in the county, cheese were derived from the normalized time-integrated met with supply from milk region 351 of New Mexico concentrations in cow’s milk of local origin (same precinct), and from other states. All counties (and precincts) are in considering the values of the culinary factor, CF (Z, CC): category 3 in MR 352, 353, and 354, as well as Grant and IC ðÞZ; L ICP ðÞZ:L Luna counties in MR 355. CC ¼ CM CFðÞ Z; CC ð39Þ −1 AvegðÞZ; L ICvegðÞZ; L The time-integrated concentration, ICPCM,3 in Bq d L , of radionuclide Z in the milk produced, or more exactly, avail- where CF (Z, CC) has two components, such that CF = able for consumption in a precinct L3 of a county CTY3 with a CF CF : residual deficit of milk (see last column of Table 13) has three 1 2 • components: (1) the milk for fluid use, TMFU,inkL,available the culinary preparation, CF1, which involves both the in CTY3; (2) the volume of milk, VNM(MR351, CTY3), processing efficiency (weight of cheese per weight of transferred from MR 351; and (3) the volume of milk, milk) and the retention of the radionuclide in the cheese VOS, transferred from other states. As it was assumed that compared with milk. The value of CF1 is obtained as the the milk from other states was not contaminated by fallout ratio of the values for the retention and for the processing from Trinity, the third component was not considered. Also, efficiency. From limited information, CF1 is about 0.1/ in the absence of information on the distribution of TMFU 0.12 = 0.8 for most elements but could be as high as at the precinct level, it was assumed that the ratio for CTY3 0.7 / 0.12 = 5.8 for strontium. For the purposes of this of TMFU(CTY3)/CONS(CTY3) had the same value for each study, the proposed values of CF1 are 5 for isotopes of precinct L3: strontium and 1 for all other radionuclides; and • CF2, which represents the reduction in concentration due

ICPCM;MC3ðZ; L3Þ¼½ðICPCM ðZ; L3Þ x TMFUðÞ CTY3Þ to radioactive decay during the time delay, estimated to

þðÞICPCM ; AVðÞ Z; MR351 x VNMðÞ MR351; CTY3 =CONSðÞ CTY3 ð35Þ be 3 d, between the production of the soft cow cheese and its consumption. where the value of VNM for the milk region (Table 14) was apportioned to its counties CTY3 with milk deficits accord- Ingestion of beef (BF), mutton (MT), and pork (PK) ing to the volumes of milk consumed, CONS(CTY3),and −1 The normalized time-integrated concentrations in beef, the time-integrated concentration, ICP in Bq d L , −1 CM,AV mutton, and pork, in Bq d kg , were estimated in the same of radionuclide Z in the milk produced in MR 351 was ob- manner. Taking, for example, beef, the following equation tained as: was used: ; : ICPCM;AV ððÞZ MR351 ð36Þ ICBF ðÞZ L ÂÃ ; ; R ; Aveg Z L V TOA ¼ ∑ICPCM;AV ðÞZ; CTY4 x EXC2ðÞ CTY4 =∑EXC2ðÞ CTY4 PGBF ðÞL TrðÞZ ¼ Ff ðÞelement; BF CFBF ðÞZ ð40Þ where CTY4 is a county in MR 351 and EXC2 is its excess YwetðÞL leðÞZ TrðÞZ þ TbðÞBF of milk after distribution in the milk region (Table 13). Assuming a delay of 4 d between availability and con- where: sumption, the time-integrated concentration, ICCM,MC3 in Bq d • PGBF(L) is the pasture-grass intake of beef cattle, esti- − L 1, of radionuclide Z in the milk consumed in precinct L3 lo- mated to be 10 kg (dry mass) d−1; the corresponding cated in a county CTY3 with deficit milk supplied by excess values for sheep and swine were assessed to be 1.5 and milk from MR351 and from other states was obtained as: 0.7 kg (dry mass) d−1, respectively; • Ff (element, BF) is the feed-to-meat transfer coefficient −lðÞZ x 4 −1 ICCM;MC3ðÞ¼Z; L3 ICPCM;MC3ðÞZ; L3 xe : ð37Þ for beef cattle (d kg ) for the element. The selected values for Ff (element), as well as their probability distri- Exceptions are Grant and Luna counties, which are the butions, are presented in Table 15 for beef, mutton, and only counties with milk categories 2 and 3. The time-integrated pork (Thiessen and Hoffman 2018); www.health-physics.com c A. BOUVILLE ET AL. 421 • TrðÞZ is the expression which, when multiplied by • IMOT(Z, L) is the sum of the activity intakes, in Bq, by the TrðÞþZ TbðÞBF Ff (element, BF), provides the value of Ff (Z, BF), Z be- mother, from drinking water and from the consumption of ing a radioactive isotope of the element considered. The all foodstuffs; parameter Tb(BF) represents the biological half-time of the • FMM (element) is the intake-to-mothers’ milk transfer coef- −1 radionuclide in meat (muscle tissue). Avalue of 30 d was se- ficient (d L ) for the element. The selected values for FMM lected for all radionuclides and for beef, mutton, or pork; and (element) are presented in Table 16 (ICRP 2004; Simon • The culinary factor CFBF represents essentially the loss et al. 2010b); and of activity due to the delay between TOA and the time • TrðÞZ is the expression which, when multiplied TrðÞþZ TbðÞMM of slaughter. It was assumed that all slaughter of beef cat- by FMM (element), provides the value of FMM(Z),Zbeing tle, sheep, and swine occurred in the fall or winter; that a radioactive isotope of the element considered. In this is, at least 60 d after TOA.ThevaluesofCFBF were cal- equation, Tr(Z) is the radioactive half-life (d) of radionu- culated assuming radioactive decay during a period of 30 d clide Z. The parameter Tb(MM) represents the biological between TOA and slaughter and an additional 1 d between half-time of the radionuclide in mothers’ milk. Avalue of slaughter and consumption. 2 d was selected for all radionuclides.

Ingestion of mothers’ milk (MM) The time-integrated concentrations in mothers’ milk, −1 Normalized intakes by inhalation and ingestion (Q /IC ) ICMM inBqdL , were obtained by means of a different pw pw procedure: • The normalized intakes, Qpw/ICpw, consist of the ; ; TrðÞZ ICMM ðÞ¼Z L IMOTðÞ Z L FMM ðÞelement ð41Þ breathing rates for the inhalation pathways and of the TrðÞþZ TbðÞMM consumption rates of water and foodstuffs for the in- where gestion pathways.

Table 15. Selected log-triangular probability distributions for the feed-to-meat transfer coefficients, in d kg−1, for beef, mutton,a and porka (Thiessen and Hoffman 2018). Beef Mutton Pork Mode Min. Max. Mode Min. Max. Mode Min. Max.

Fe 1.4 102 4.7 103 4.2 102 3.0 103 4.0 104 3.0 102 Co 4.3 104 1.3 104 1.3 103 1.2 102 1.2 103 1.2 101 As 2.0 102 2.0 103 2.0 101 Sr 1.3 103 2 104 9.2 103 1.5 103 3.0 104 4.0 103 2.5 103 5.0 104 8.0 103 Y7.5 104 7.5 105 7.5 103 Zr 1.2 106 1.2 107 1.2 105 Nb 2.6 107 2.6 108 2.6 106 Mo 1.0 103 1.0 104 1.0 102 Ru 3.3 103 1.3 103 1.0 102 2.1 103 2.1 104 2.1 102 3.0 103 3.0 104 3.0 102 Rh 1.0 102 1.0 103 1.0 101 Cd 5.8. x.103 1.5 104 6.0 102 1.2 103 1.2 104 1.2 102 Sn 1.0 102 1.0 103 1.0 101 Sb 1.2 103 1.2 104 1.2 102 Te 7.0 103 7.0 104 7.0 102 I6.7 103 2.0 103 3.8 102 3.0 102 3.0 103 3.0 101 4.1 102 4.1 103 4.1 101 Cs 2.2 102 4.7 104 9.6 102 1.9 101 5.3 102 1.3 100 2.0 101 7.0 102 6.0 101 Ba 1.4 104 1.4 105 1.4 103 La 1.3 104 1.3 105 1.3 103 Ce 2.0 102 2.0 103 2.0 101 2.5 104 2.5 105 2.5 103 Pr 1.0 103 1.0 104 1.0 102 Nd 2.0 103 2.0 104 2.0 102 Pm 1.0 103 1.0 104 1.0 102 Sm 1.5 103 1.5 104 1.5 102 U3.9 104 1.3 104 1.2 103 4.4 102 4.4 103 4.4 101 Np 1.0 103 1.0 104 1.0 102 Pu 1.1 106 8.8 108 3.0 104 5.3 105 5.3 106 5.3 104

a Beef was used as a surrogate when specific transfer coefficients were not available.

www.health-physics.com 422 Health Physics October 2020, Volume 119, Number 4 Breathing rates total consumption rate of water was assumed to be pro- • The breathing rates BR as a function of age were taken portional to the urine excretion rate, also given in ICRP from ICRP Publication 66 (ICRP 1994). The selected Publication 89. The rounded values that were obtained values, which correspond to light exercise, are averaged for the consumption of drinking water vary with age but are independent of the data set (Table 17). over males and females and are assumed to be the same • ’ for Hispanics, Whites, Native Americans, and African The consumption rates of mother s milk, presented in Americans, depending on the age of the representative Table 17, were taken from ICRP Publication 95 3 −1 (ICRP 2004). Based on the information provided by individual under consideration: 0.19 m h for infants ’ under 1 y of age, 0.35 m3 h−1 for1-to2-y-oldinfants, Potischman et al. (2020), the consumption of mother s 0.57 m3 h−1 for 3- to 7-y-old children, 1.25 m3 h−1 for 8- milk was assumed to occur during the entire first year to 12-y-old children, 1.40 m3 h−1 for 13- to 17-y-old teen- of age. It was also assumed that no other foodstuff was agers, and 1.38 m3 h−1 for adults. consumed by infants <1 y of age. Consumption rates of drinking water and foodstuffs Doses per unit intake (Dpw/Qpw) • The consumption rates of drinking water, in L d−1,andof − • The absorbed doses per unit intake, also called dose con- foodstuffs, in kg d 1, considered in the dose assessment version factors, to lung, thyroid, active marrow, stomach are presented in Table 17. For all foodstuffs other than and colon, were estimated for the 63 radionuclides con- mother’s milk, the consumption rates CR for Hispanics, sidered in the dose assessment, for intake via inhalation Whites, and Native Americans, which vary according to and ingestion, and for all ICRP post-natal age groups, data set and to age group, were derived from information namely newborn, 1–2y,3–7y,8–12 y, 13–17 y, and adults. collected from the focus groups and interviews with key In addition, specific consideration was given to the absorbed informants that were conducted in various locations of doses received in utero per unit intake by the mother. New Mexico (Potischman et al. 2020). For reasons of • In order to estimate the conversion factors from intake to necessity, that information was collected for age groups absorbed dose delivered during the first year after intake, that are different from those used in the dose assessment. extensive use was made of ICRP publications (ICRP The values presented in Table 17 were interpolated from 1989, 1993, 1995a and b, 2001, 2004) in which 50-y the collected data using the assumption that the annual equivalent doses per unit intake are provided for the most consumption rates do not vary from year to year in the important radionuclides. The selected values for the class same age group. of solubility in lung and for the gastro-intestinal fractions • The consumption rates of drinking water and of mother’s were based on a literature review (Ibrahim et al. 2010). milk were derived from literature values. The total con- For all 54 radionuclides with physical half-lives shorter sumption rate of water (drinking water, other beverages, − than 3 mo, the doses per unit intake that are received dur- foodstuffs) was taken to be 2.2 L d 1 among adults, as ing the first year following the test are approximately recommended in ICRP Publication 89 (ICRP 2002). The equal to the 50-y dose conversion factors. However, for consumption rates of only drinking water were assumed the nine radionuclides with half-lives longer than 9 mo, to be equal to the total consumption of water minus the only a fraction of the 50-y doses was delivered during consumption of cow’s milk. The variation with age of the the first year following the test. For these radionuclides, Table 16. Selected values of the fraction transferred to the infant calculations were made of the first-year doses using spe- 1 in breast milk following maternal intake, in d L (ICRP 2004; cialized software consistent with the ICRP publications. Simon et al. 2010b). For 239Pu and 240Pu, which are a-emitters, the low-LET Element FMM Element FMM Element FMM and the high-LET components of the equivalent dose co- Fe 2.2 103 Tc 1.0 101 Cs 1.2 101 efficients are provided separately in the software that was Co 3.5 102 Ru 2.2 103 Ba 5.8 103 used. For the purposes of this study, the equivalent dose Cu 4.1 102 Rh 3.4 103 La 2.3 103 coefficients due to the high-LET component were divided 2 2 5 As 4.0 10 Ag 1.1 10 Ce 2.1 10 by 20 (which is the WR value for a particles) and added Br 8.5 102 Pd 2.8 104 Pr 2.3 105 to the equivalent dose coefficients due to the low-LET Rb 8.5 102 Cd 3.3 103 Nd 2.3 105 component (for which the WR valueisequalto1)inorder 2 3 5 Sr 6.1 10 In 1.2 10 Pm 2.3 10 to obtain the numerical values of the absorbed doses per 6 3 5 Y4.1 10 Sn 1.2 10 Sm 2.3 10 unit intake, expressed in mGy Bq−1. Zr 4.7 104 Sb 4.7 103 U9.7 104 • 4 3 5 With respect to the absorbed doses received in utero per Nb 2.8 10 Te 2.9 10 Np 4.1 10 unit intake by the mother, use was made of specialized Mo 1.2 102 I3.3 101 Pu 2.4 105 software,7 consistent with the results presented in ICRP

www.health-physics.com c A. BOUVILLE ET AL. 423 Table 17. Estimation of the consumption rate (kg d−1 or L d−1) according to age group and data set. Consumption rate (kg d1 or L d1) Foodstuff Age, y A B C D E F

Mothers’ milk 0 – 1 0.8 0.8 0.8 0.8 0.8 0.8 Drinking water 0 – 1 0.5 0.5 0.5 0.5 0.5 0.5 1 – 2 0.4 0.4 0.4 0.4 0.4 0.4 3 – 7 0.5 0.5 0.5 0.5 0.5 0.5 8 – 12 0.7 0.7 0.7 0.7 0.7 0.7 13 – 17 1.5 1.5 1.5 1.5 1.5 1.5 Adult 1.9 1.9 1.9 1.9 1.9 1.9 Leafy vegetables 1 – 2 0.052 0.013 0.013 0.031 0.045 0.039 3 – 7 0.094 0.021 0.021 0.066 0.061 0.068 8 – 12 0.16 0.036 0.036 0.095 0.091 0.12 13 – 17 0.23 0.086 0.086 0.13 0.15 0.21 Adult 0.26 0.134 0.14 0.16 0.19 0.28 Fruit vegetables 1 – 2 0.045 0.011 0.011 0.027 0.039 0.044 3 – 7 0.031 0.007 0.006 0.020 0.023 0.027 8 – 12 0.031 0.007 0.006 0.018 0.018 0.023 13 – 17 0.049 0.018 0.018 0.027 0.032 0.044 Adult 0.057 0.030 0.030 0.036 0.042 0.060 Fruit and berries 1 – 2 0.028 0 0.55 0.028 0.005 0.11 3 – 7 0.038 0 0.65 0.038 0.006 0.15 8 – 12 0.057 0 0.85 0.057 0.011 0.19 13 – 17 0.062 0 0.93 0.062 0.016 0.24 Adult 0.041 0 0.77 0.041 0.020 0.28 Cow’smilk 1– 2 0.27 0.25 0.12 0.54 0.118 0.28 3 – 7 0.48 0.38 0.15 0.53 0.189 0.33 8 – 12 0.65 0.50 0.18 0.72 0.142 0.49 13 – 17 0.65 0.39 0.19 0.84 0 0.53 Adult 0.59 0.18 0.18 0.61 0 0.39 Cow’scheese 1– 2 0.037 0.037 0 0 0 0 3 – 7 0.044 0.044 0.008 0 0 0.004 8 – 12 0.049 0.049 0.016 0 0 0.009 13 – 17 0.050 0.050 0.019 0 0 0.011 Adult 0.050 0.050 0.015 0 0 0.012 Beef 1 – 2 0.040 0.018 0.038 0.042 0.001 0.003 3 – 7 0.063 0.031 0.058 0.070 0.006 0.004 8 – 12 0.13 0.045 0.082 0.134 0.010 0.007 13 – 17 0.16 0.055 0.10 0.20 0.013 0.010 Adult 0.077 0.056 0.11 0.20 0.015 0.009 Mutton / Pork 1 – 2 0.004 0 0.003 0.022 0 0.011 3 – 7 0.005 0 0.005 0.035 0 0.019 8 – 12 0.006 0.021 0.006 0.069 0 0.025 13 – 17 0.008 0.053 0.008 0.14 0 0.025 Adult 0.008 0.053 0.009 0.20 0 0.023

Publication 88 (ICRP 2001). The time of intake was as- ; ; ; ; ; ; ̇ ; : Apw ðÞZ L TOA : ICpw ðÞZ L DpwðÞ¼Z L I m X ðÞ12 L ̇ ð7Þ sumed to occur during the fifteenth week of pregnancy, X ðÞ12; L ApwðÞZ; L; TOA ; ; ; ; Uncertainties : Qpw ðÞZ L I : DpwðÞZ I m : ICpw ðÞZ; L QpwðÞZ; L; I • The uncertainties in the estimated doses from internal ir- radiation were evaluated for each term of the right side of Uncertainties in the normalized deposition densities ˙ ˙ eqn (7), reproduced below, with the exception of X(12), [Apw/X (12, L)] which was covered in the section “Absorbed Doses From The important parameters influencing the estimates of External Irradiation:” the normalized deposition densities are: www.health-physics.com 424 Health Physics October 2020, Volume 119, Number 4 • b= ˙ −3 −2 X R;12, which is the ratio of the beta activity deposited distribution of ICres/Agd,inBqdm per Bq m , around on the groundV and of the exposure rate at H+12 h for a its best estimate was taken to be the same as that esti- given value of R/V. These ratios were derived from the find- mated in Maxwell and Anspaugh (2011) for Sf(t),thatis, ings of Hicks (1985) for R/V = 0.5 and Beck (2009) for uniform, U(0.1, 10); other values of R/V. The probability distributions around • Ingestion of drinking water. The uncertainty in the nor- the best estimate values were subjectively estimated to be malized time-integrated concentrations in drinking wa- −1 −2 log-normal with a GSD of 1.1; ter, ICwt/Agd,inBqdL per Bq m , depends on the • ðÞZ=b R; is the ratio of the activity of radionuclide Z and origin of drinking water (cistern or public network), on V 12 of the beta activity deposited on the ground for a given the characteristics of the cistern systems and of the rivers, value of R/V. These ratiosÂÃ were obtained from the same on the properties of the radionuclides, and on the validity b= ̇ sources as those used for X ðÞ12 R;12. The probability of the models used to calculate the water concentrations. distributions around the best estimateV values were sub- For any radionuclide, Z,andanyprecinct,L,theproba- jectively estimated to be log-normal with a GSD of 1.2; bility distribution of ICwt/Agd around its best estimate • Fgd(Z,TOA) is the function describing the variation with was subjectively assumed to be censored log-uniform time of the activity of radionuclide Z deposited on the ground, (0.1, 10); according to the laws of radioactive decay. Its uncertainty • Ingestion of leafy vegetables. The uncertainty in the nor- was taken to be negligible; malized time-integrated concentrations in leafy vegetables, IC /A ,inBqdkg−1 per Bq m−2, depends mainly on the • N50 ðÞL is the fraction of total beta activity that is on par- LV veg ticles <50 mmatTOA. Its deterministic value is used to uncertainty in the culinary preparation component of the assign the best estimate of R/V in the precinct under con- culinary factor. For any radionuclide, Z, and any precinct, sideration. Its probability distribution around its best L, the probability distribution of ICLV/Aveg around its best estimate was subjectively assumed to be triangular and to estimate was subjectively assumed to be log-normal with depend on the value of R/V:TRI(0.5,1,2)forR/V =3, a geometric standard deviation of 1.3; • TRI (0.6, 1, 1.5) for R/V =2,TRI(0.7,1,1.4)for Ingestion of fruit vegetables. The uncertainty in the nor- R/V = 1.5, TRI (0.8, 1, 1.2) for R/V = 1, and no uncertainty malized time-integrated concentrations in fruit vegetables, −1 −2 for R/V =0.5; ICFV/Aveg,inBqdkg per Bq m , depends mainly on the uncertainty in the transfer factor from leaves to fruit. • fdry is the fraction of the activity attached to particles <50 mm that is deposited and initially retained by vegetation For any radionuclide, Z,andanyprecinct,L,theproba- as a result from deposition via dry processes. Its proba- bility distribution of ICFV/Aveg, around its best estimate bility distribution around its best estimate was subjec- was subjectively assumed to be log-uniform with lower tively assumed to be triangular: TRI (0.5, 1, 2); and upper bounds equal to 0.1 and 10 times the deter- • f is the fraction of the activity attached to particles of ministic value, respectively; wet • all sizes that is deposited and initially retained by vegeta- Ingestion of fruit and berries. The uncertainty in the nor- tion as a result of deposition via wet processes, i.e., rain. malized time-integrated concentrations in fruit and berries, −1 −2 Given the fact that the occurrence of rain during the ICFB/Aveg,inBqdkg per Bq m , depends mainly on the passage of the radioactive cloud had a very small prob- uncertainty in the application of the model to different types of berries and of fruit. For any radionuclide, Z,andanypre- ability, the uncertainty in the value of fwet was not taken into consideration. cinct, L, the probability distribution of ICFB/Aveg around its best estimate was subjectively assumed to be log-normal Uncertainties in the normalized time-integrated with a geometric standard deviation of 1.3; concentrations (ICpw/Apw) • Ingestion of fresh cows’ milk. The uncertainty in the nor- All inhalation and ingestion pathways are considered malized time-integrated concentrations in cows’ milk, −1 −2 in turn. ICCM//Aveg,inBqdL per Bq m , depends essentially • Inhalation during the passage of the cloud. For any radio- on the transfer factor from feed to milk and on the origin nuclide, Z, deposited in any precinct, L, the probability of milk. Estimated values for any radionuclide, Z, are LN −3 −2 (2.0) for precincts in counties with category-1 milk distribution of ICcd/Agd,inBqsm per Bq m ,around its best estimate was subjectively assumed to be triangu- (Catron, Colfax, Curry, De Baca, Harding, Quay, Roose- lar: TRI (0.7, 1, 1.5); velt, Sierra, Torrance, Union); LN(3.0) for precincts in • Inhalation due to resuspended material. For any radionu- counties with category-2 milk (Guadalupe, Mora, San clide, Z, deposited in any precinct, L, the probability Miguel, Socorro, Grant, Hidalgo, Luna); and LN(4.0) for precincts in counties with category-3 milk (Bernalillo, Mckinley, Rio Arriba, Sandoval, San Juan, Santa Fe, Taos, 7 Personal communication, K. Eckerman and D. Melo; October 2017. Valencia, Chaves, Dona Ana, Eddy, Lea, Lincoln, Otero); www.health-physics.com c A. BOUVILLE ET AL. 425 • Ingestion of cow cheese. The uncertainty in the normalized SUMMARY time-integrated concentrations in cow cheese, ICCC//Aveg, − − The purpose of this document is to provide detailed in- in Bq d kg 1 per Bq m 2, depends on the uncertainties in formation on the methods, models, and parameter values ICP /A and in CF, as well as on the uncertainties in CM veg that were used in the assessment of the radiation doses that the transfer factor from feed to milk and in the origin of were received by New Mexico residents as the result of the milk. For any radionuclide, Z, and any precinct, L, the prob- detonation of the Trinity test (Simon et al. 2020). Three ability distribution of IC /A around its best estimate CC veg pathways of human exposure were included: (1) external was subjectively assumed to be log-normal with a geomet- irradiation, arising mainly from the radionuclides deposited ric standard deviation of 2.3; on the ground, and, for a small part, from radionuclides in • Ingestion of meat (beef, mutton, and pork). The uncer- the passing cloud, (2) inhalation of radionuclide-contaminated tainty in the normalized time-integrated concentrations −1 air during the passage of the radioactive cloud and, thereafter, in beef, mutton, and pork, ICBF/Aveg,inBqdkg per − of radionuclides transferred (resuspended) from soil to air, Bq m 2, depends mainly on the uncertainty in the transfer and (3) ingestion of contaminated water and foodstuffs. To coefficient from feed to meat, F . For any radionuclide, Z, f the extent possible, well established models and parameter and any precinct, L, the probability distribution of IC /A BF veg values were adopted for the calculation of the doses resulting around its best estimate was subjectively assumed to be from those three pathways. Sixty-three radionuclides and log-triangular with upper and lower bounds equal to 10 times five organs or tissues (thyroid, lung, active marrow, stomach, higher and lower than the best estimate, respectively; and and colon) were considered in the assessment. • Ingestion of mother’s milk. The uncertainty in the Each of the 721 precincts was classified according to time-integrated concentrations in mother’smilkde- ecozone (plains, mountains, or mixture of plains and moun- pends mainly on the uncertainties on the total intakes tains) and population density (urban or rural). The lifestyle by the mother and on the transfer of the radionuclides and dietary habits of representative individuals in each type from the intakes to breast milk. For any radionuclide, of precinct were obtained by means of focus-group sessions Z, and any precinct, L, the probability distribution of and interviews of key informants (Potischman et al. 2020). IC /A around its best estimate was subjectively as- MM pw Doses were assessed to representative individuals defined sumed to be equal to the uncertainty in cows’ milk of by the important characteristics of four ethnic groups (His- category 1, that is, log-normal with a geometric standard panics, Whites, Native Americans, and African Americans) deviation of 2.0. and seven age groups (in utero, newborn, 1–2y,3–7y, – – Uncertainties in breathing rates and consumption rates 8 12 y, 13 17 y, and adult). (Q/IC) Previous studies of fallout from nuclear weapons tests The probability distribution of the breathing rate, Q/IC have shown that as a result of the preponderance of short-lived in m3 d−1, was subjectively estimated to be distributed as radionuclides, most of the dose from external irradiation TRI(0.5, 1, 2) around the best estimate for any age and data Table 18. Estimated probability distribution of the uncertainty in the set, while the probability distribution of the consumption dose (mGy Bq−1) per unit intake for selected elements.a rate of any foodstuff was subjectively estimated to be log- Uncertainty distribution normal with a GSD of 1.3 around the best estimate, also for Element Reliability index around central estimates all ages and data sets. Ba Medium TRI (0.5, 1, 1.5) Uncertainties in doses per unit intake (D/Q) Ce Low TRI (0.3, 1, 1.5) The probability distribution of the dose per unit intake, Cs High TRI (0.8, 1, 1.5) D/Q in mGy Bq−1, for a given radionuclide varies, among I High TRI (0.8, 1, 1.5) other factors, according to the mode of exposure (inhalation La High TRI (0.8, 1, 1.5) or ingestion), the organ/tissue under consideration, the char- Mo Medium TRI (0.5, 1, 1.5) acteristics of the population group (age and sex), the level of Np Low TRI (0.3, 1, 1.5) complexity of the biokinetics and dosimetry relevant to the Rh Low TRI (0.3, 1, 1.5) considered radionuclide, and the quality of the underlying Ru Low TRI (0.3, 1, 1.5) Sr High TRI (0.8, 1, 1.5) information. In this analysis, the elements were classified Te Medium TRI (0.5, 1, 1.5) according to a subjective reliability index (low, medium, or U Low TRI (0.3, 1, 1.5) high) and all radionuclides of the element were assumed to Y Low TRI (0.3, 1, 1.5) have the same reliability index for all age groups (Table 18). Zr Medium TRI (0.5, 1, 1.5) The way in which the uncertainties in the individual pa- a The selected elements include the radionuclides that contributed to more than rameters were combined to evaluate the uncertainties in the 95% of the internal dose to the organs and tissues considered in the study. For dose estimates is provided in Simon et al. (2020). all other elements, the reliability index was judged to be low. www.health-physics.com 426 Health Physics October 2020, Volume 119, Number 4 is delivered during a few months following a nuclear test Carini F. Radionuclides in plants bearing fruit. An overview. J En- (Beck 2005) while the annual doses from internal irradiation viron Radioact 46:77–97; 1999. are much greater in the year following the test than in any Cederwall R, Peterson K. Meteorological modeling of arrival and deposition of fallout at intermediate distances downwind of the subsequent year (Simon et al. 2010b). Thus, the doses from Nevada Test Site. Health Phys 59:593–601; 1990. Trinity were calculated for only the first year following the Chamberlain AC. Interception and retention of radioactive aero- day of the test (16 July 1945) for all 721 precincts, all four sols by vegetation. Atmospheric Environment 4:57–78; 1970. ethnicities, and all seven age groups. Chamberlain AC, Chadwick RC. Deposition of airborne radioio- – The collection of models and parameter values presented dine vapor. Nucleonics 8:22 25; 1953. Church BW, Wheeler DL, Campbell CM, Nutley RV, Anspaugh here, enabling an assessment for multiple types of foods and LR. Overview of the Department of Energy’s Off-Site Radia- modes of intake, make the Trinity study one of the most de- tion Exposure Review Project. Health Phys 59:503–510; 1990. tailed assessments of exposure from nuclear testing fallout to Douglas RL. Levels and distribution of environmental plutonium have ever been conducted. There are, clearly, numerous uncer- around the TRINITY site. Las Vegas, NV: US Environmental Protection Agency, Office of Radiation Programs; Technical tainties in the results of an assessment of an event so long in Note ORP/LV 78–3; 1978. the past. However, the means for propagating the uncertainty Dunning GM. Radiation from fallout and their effects. In: Eighty- of estimated doses are provided and uncertainties of doses fifth Congress First Session on The Nature of Radioactive Fall- have been estimated in a companion publication (Simon out from Nuclear Weapons Tests. Hearings before the Special et al. 2020). The strategies and models presented here can Subcommittee on Radiation of the Joint Committee on Atomic Energy. Congress of the United States. Part I. Washington, DC: be adapted to assessments of other nuclear events where radio- United States Government Printing Office; 1957: 170–196. active fallout is a source of human exposure. Gordeev K, Vasilenko I, Lebedev A, Bouville A, Luckyanov N, Simon SL, Stepanov Y,Shinkarev S, Anspaugh L. Fallout from nuclear tests: dosimetry in Kazakhstan. Radiat Environ Biophys REFERENCES 41:61–67; 2002. Anspaugh LR, Simon SL, Gordeev KI, Likhtarev IA, Maxwell RM, Hansen WR, Rodgers JC. Radiological survey and evaluation of the Shinkarev SM. Movement of radionuclides in terrestrial ecosys- fallout area from the TRINITY test: Chupadera Mesa and White tems by physical processes. Health Phys 82:669–679; 2002. Sands Missile Range, New Mexico. Los Alamos, NM: Los Beck HL. Exposure rate conversion factors for radionuclides deposited Alamos National Laboratory; Report LA-10256- MS; 1985. on the ground. New York: US Department of Energy; Environmen- Henderson RW.Approximation of the decay of fission and activa- tal Measurements Laboratory Report EML-378; 1980. tion product mixtures. Los Alamos, NM: Los Alamos National Beck HL. External dose estimates from NTS fallout. Appendix E, Laboratory; Report LA-11968-MS; 1991. pages E-1 to E-57, in: Report on the feasibility of a study of the Hicks HG. Calculation of the concentration of any radionuclide health consequences to the American population from nuclear deposited on the ground by offsite fallout from a nuclear deto- weapons tests conducted by the United States and other nations. nation. Health Phys 42:585–600; 1982. Washington, DC: US DHHS; 2005. Available at http://www.cdc. Hicks HG. Results of calculations of external gamma radiation gov/nceh/radiation/fallout/default.htm. Accessed 6 August 2020. exposure rates from fallout and the related radionuclide Beck HL. Estimates of H+12 Exposure Rates and Associated Un- compositions—the TRINITY event. Livermore, CA: University of certainty Received by the Population of Selected Villages in California, Lawrence Livermore Laboratory; UCRL- 53705; 1985. the Vicinity of the Semipalatinsk Test Site in Kazakhstan at Hoffman FO, Thiessen KM, Frank ML, Blaylock BG. Quantifi- the Time of the Semipalatinsk Test Site Atmospheric Tests cation of the interception and initial retention of radioactive (1949-1962). Application of Joint Russian/American Method- contaminants deposited on pasture grass by simulated rain. ology for Fallout Dose Assessment to the Estimation of Internal Atmospheric Environ 26A:3313–3321; 1992. Doses and Associated Uncertainty. Final Report to NCI in Ful- Hoffman FO, Thiessen KM, Rael RM. Comparison of interception fillment of PSC # HHSN 261200800397P. March 15, 2009. and initial retention of wet-deposited contaminants leaves of differ- Beck HL, Bouville A, Moroz BE, Simon SL. Fallout deposition ent vegetation types. Atmospheric Environ 29:1771–1775; 1995. in the Marshall Islands from Bikini and Enewetak nuclear – Horton RE. Rainfall interception. Monthly Weather Rev 47: weapons tests. Health Phys 99:124 142; 2010. 608–623; 1919. Beck HL, Simon SL, Bouville A, Romanyukha A. Accounting for Ibrahim SA, Simon SL, Bouville A, Melo D, Beck HL. Alimentary unfissioned plutonium from the Trinity atomic bomb test. Health – tract absorption (f1 values) for radionuclides in local and regional Phys 119:504 516; 2020. fallout from nuclear tests. Health Phys 99:233–251; 2010. Bellamy MB, Dewji SA, Leggett RW, Hiller M, Veinot K, Manger International Atomic Energy Agency. Handbook of parameter RP, Eckerman KF, Ryman JC, Easterly CE, Hertel NE, Stewart values for the prediction of radionuclide transfer in terrestrial DJ. FGR 15. Federal Guidance Report No. 15: External Expo- and freshwater environments. Vienna: IAEA; Technical Report sure to Radionuclides in Air, Water, and Soil. Washington, Series No. 472; 2010. DC: U.S. Environmental Protection Agency, Office of Radiation and Indoor Air; Report EPA-402-R19-002; 2019. International Commission on Radiological Protection. Age-dependent Bouville A, Beck HL, Simon SL. Doses from external irradiation to doses to members of the public from intake of radionuclides: Marshall Islanders from Bikini and Enewetak nuclear weapons Part 1. Oxford: Pergamon Press; ICRP Publication 56, Annals tests. Health Phys 99:143–156; 2010. of the ICRP 20(2); 1989. Cahoon EK, Zhang R, Simon SL, Bouville A, Pfeiffer RM. International Commission on Radiological Protection. Age- Projected cancer risks to residents of New Mexico from dependent doses to members of the public from intake of ra- — exposure to Trinity radioactive fallout. Health Phys 119: dionuclides: part 2 Ingestion dose coefficients. Oxford: Pergamon – 478–493; 2020. Press; ICRP Publication 67, Annals of the ICRP 23(3 4); 1993.

www.health-physics.com c A. BOUVILLE ET AL. 427 International Commission on Radiological Protection. Human respi- National Council on Radiation Protection and Measurements. ratory tract model for radiological protection. Oxford: Pergamon Deriving organ doses and their uncertainty for epidemiological Press; ICRP Publication 66, Annals of the ICRP 24(1–3); 1994. studies. NCRP Report No. 178; 2018. International Commission on Radiological Protection. Age-dependent Ng YC, Anspaugh LR, Cederwall RT. ORERP internal dose esti- doses to members of the public from intake of radionuclides: mates for individuals. Health Phys 59:693–713; 1990. part 4—inhalation dose coefficients. Oxford: Pergamon Press; Potischman N, Salazar SI, Scott MA, Naranjo M, Haozous E, ICRP Publication 71, Annals of the ICRP 25(3–4); 1995a. Bouville A, Simon SL. Methods and findings on diet and International Commission on Radiological Protection. Age-dependent lifestyle used to support estimation of radiation doses from doses to members of the public from intake of radionuclides: radioactive fallout from the Trinity nuclear test. Health Phys part 3—ingestion dose coefficients. Oxford: Pergamon Press; 119:390–399; 2020. ICRP Publication 69, Annals of the ICRP 25(1); 1995b. Pröhl G. Interception of dry and wet deposited radionuclides by International Commission on Radiological Protection. Corrected vegetation. J Environ Radioact 100:675–682; 2009. version, May 2002. Doses to the embryo and fetus from intakes Quinn VE. Analysis of Nuclear Test TRINITY radiological and of radionuclides by the mother. Oxford: Pergamon Press; Pub- meteorological data. NOAA, Weather Service Nuclear Support lication 88, Annals of the ICRP 31(1–3); 2001. Office, NVO 313; 1987. International Commission on Radiological Protection. Basic ana- Quinn VE. Analysis of meteorological and radiological data. tomical and physiological data for use in radiological protec- Health Phys 59:577–592; 1990. tion: reference values. Oxford: Pergamon Press; Publication Sehmel GA. Particle and gas deposition: a review. Atmospheric 89, Annals of the ICRP 32(3–4); 2002. Environ 14:981–1011; 1980. International Commission on Radiological Protection. Dose to in- Simmonds JR, Linsley GS. Parameters for modelling the intercep- fants from ingestion of radionuclides in mothers’ milk. Oxford: tion and retention of deposits from atmosphere by grain and Elsevier; Publication 95, Annals of the ICRP 34(3–4); 2004. leafy vegetables. Health Phys 43:679–691; 1982. International Commission on Radiological Protection. Conversion Simon SL, Bouville A, Land C, Beck HL. Radiation doses and coefficients for radiological protection quantities for external cancer risks in the Marshall Islands associated with exposure to radiation exposures. Oxford: Elsevier Publication 116, Ann. radioactive fallout from Bikini and Enewetak nuclear weapons ICRP 40(2–5); 2010. tests: summary. Health Phys 99:105–123; 2010a. Jacob P,Rosenbaum H, Petoussi N, Zankl M. Calculation of organ Simon SL, Bouville A, Melo D, Beck HL, Weinstock RM. Acute doses from environmental gamma rays using human phantoms and chronic intakes of fallout radionuclides by Marshallese and Monte-Carlo methods. Part II: radionuclides distributed in from nuclear weapons testing at Bikini and Enewetak and re- the air or deposited on the ground. Neuherberg: GSF-Institut lated internal radiation doses. Health Phys 99:157–200; 2010b. für Strahlenschutz; GSF-Bericht 12/90; 1990. Simon SL, Lloyd RD, Till JE, Hawthorne HA, Gren DC, Rallison Kinnersley RP, Scott LK. Aerial contamination of fruit through ML, Stevens W. Development of a method to estimate thyroid wet deposition and particulate dry deposition. J Environ Radioact dose from fallout radioiodine in a cohort study. Health Phys 59: – 22:191 213; 2001. 669–691; 1990. Larson KH, Neel JW, Hawthorn HA, Mork HM, Rowland RH, Simon SL, Bouville A, Beck HL, Melo DR. Estimated radiation Baurmash L, Lindberg RG, Olafson JH, Kowalewski BW.Dis- doses received by New Mexico residents from the 1945 Trinity tribution, characteristics, and biotic availability of fallout, Op- nuclear test. Health Phys 119:428–477; 2020. eration PLUMBBOB. Los Angeles, CA: University of California; Thiessen KM, Hoffman FO. Trinity study dose assessment WT-1488; 1966. parameters—final report to the National Cancer Institute. Oak Lindberg RG, Romney EM, Olafson JH, Larson KH. Factors Ridge, TN: Oak Ridge Center for Risk Analysis, Inc.; 2018. influencing the biological fate and persistence of radioactive fall- Thompson CB, McArthur ED, Hutchinson SW. Development out. Los Angeles, CA: University of California; WT-1177; 1959. of the Town Data Base: estimates of exposure rates and times Maxwell RM, Anspaugh LR. An improved model for prediction of fallout arrival near the Nevada Test Site. Las Vegas, NV: – of resuspension. Health Phys 101:722 730; 2011. US Department of Energy Nevada Operations Office; Report Menne MJ, Durre I, Vose RS, Gleason BE, Houston TG. An over- DOE/NV-374; 1994. — view of the Global Historical Climatology Network daily United Nations Scientific Committee on the Effects of Atomic Radi- – database. J Atmospheric Oceanic Technol 29:897 910; 2012. ation. UNSCEAR 1993 Report. Sources and effects of ionizing DOI: 10.1175/JTECH-D-11-00103.1. radiation. Annex B: Exposures from man-made sources of radia- Ministère de la Défense. La dimension radiologique des essais tion. New York: United Nations; 1993. nucléaires français en Polynésie. Paris: République Française; US Department of Health and Human Services. A report on the 2006 (in French). feasibility of a study on the health consequences to the American National Cancer Institute. Estimated exposures and thyroid doses population from nuclear weapons tests conducted by the United − received by the American people from iodine 131 in fallout States and other nations. Prepared for the U.S. Congress. following Nevada atmospheric nuclear bomb tests. Bethesda, Washington, DC: US DHHS; 2005. Available at http://www.cdc. MD: NCI; 1997. gov/nceh/radiation/fallout/default.htm. Accessed 6 August 2020. National Council on Radiation Protection and Measurements. Vandecasteele CM, Baker S, Förstel H, Musinsky M, Millan R, Uncertainties in the measurement and dosimetry of external Madoz-Escande C, Tormos J, Sauras T, Schulte E, Colle C. In- radiation. NCRP Report No. 158; 2007. terception, retention and translocation under greenhouse con- National Council on Radiation Protection and Measurements. ditions of radiocaesium and radiostrontium from a simulated Radiation dose reconstruction: principles and practices. NCRP accidental source. Sci Total Environ 278:199–214; 2001. Report No. 163; 2009a. National Council on Radiation Protection and Measurements. Un- certainties in internal radiation dose assessment. NCRP Report No. 164; 2009b. ■■

www.health-physics.com Paper

Estimated Radiation Doses Received by New Mexico Residents from the 1945 Trinity Nuclear Test

Steven L. Simon,1 André Bouville,2 Harold L. Beck,3 and Dunstana R. Melo4

human-consumed plant products including leafy vegetables, fruit Abstract—The National Cancer Institute study of projected health vegetables, fruits, and berries. Tables of best estimates of county risks to New Mexico residents from the 1945 Trinity nuclear test population-weighted average organ doses by ethnicity and age are provides best estimates of organ radiation absorbed doses presented. A discussion of our estimates of uncertainty is also received by representative persons according to ethnicity, age, provided to illustrate a lower and upper credible range on our and county. Doses to five organs/tissues at significant risk from best estimates of doses. Our findings indicate that only small exposure to radioactive fallout (i.e., active bone marrow, thyroid geographic areas immediately downwind to the northeast received gland, lungs, stomach, and colon) from the 63 most important exposures of any significance as judged by their magnitude radionuclides in fresh fallout from external and internal relative to natural radiation. The findings presented are the irradiation were estimated. The organ doses were estimated for most comprehensive and well-described estimates of doses received four resident ethnic groups in New Mexico (Whites, Hispanics, by populations of New Mexico from the Trinity nuclear test. Native Americans, and African Americans) in seven age groups Health Phys. 119(4):428–477; 2020 using: (1) assessment models described in a companion paper, Key words: dose reconstruction; exposure, radiation; fallout; (2) data on the spatial distribution and magnitude of radioactive health effects fallout derived from historical documents, and (3) data collected on diets and lifestyles in 1945 from interviews and focus groups conducted in 2015–2017 (described in a companion paper). The INTRODUCTION organ doses were found to vary widely across the state with the highest doses directly to the northeast of the detonation site and THE TRINITY nuclear test was unique in the annals of nuclear at locations close to the center of the Trinity fallout plume. science and socio-political and worldwide nuclear testing Spatial heterogeneity of fallout deposition was the largest cause of variation of doses across the state with lesser differences due history due to it being the first test of a nuclear fission to age and ethnicity, the latter because of differences in diets and weapon and, indeed, the first nuclear explosion in the his- lifestyles. The exposure pathways considered included both external tory of the world. The test took place at 5:29 a.m. on 16 irradiation from deposited fallout and internal irradiation via July 1945 about 56 km southeast of Socorro, New Mexico inhalation of airborne radionuclides in the debris cloud as well ′ ″ ′ ″ as resuspended ground activity and ingestion of contaminated (approximately 33°40 38 N, 106°28 31 W), on what was drinking water (derived both from rivers and rainwater cisterns) then the Alamogordo Bombing and Gunnery Range, now and foodstuffs including milk products, beef, mutton, and pork, part of White Sands Missile Range. Beyond being the first test, Trinity was a unique event in other ways. Unlike for later nuclear tests, because of the secrecy of the development of the atomic bomb, there was no public notice before the test and no prior evacuations 1Division of Cancer Epidemiology and Genetics, National Cancer of any nearby communities. In addition, the low detonation Institute, National Institutes of Health, Bethesda, MD; 2National Cancer Institute, National Institutes of Health, Bethesda, MD (retired); 3New height (30.5 m) and relatively light winds (Hawthorne York, NY, US DOE (retired); 4 MeloHill LLC, Rockville, MD. 1979) tended to create significant local fallout. The state For correspondence contact: Steven L. Simon, National Cancer Insti- of New Mexico was largely rural, though there were farms tute, 9609 Medical Center Drive, Room 7E558, MSC 9778, Bethesda, MD USA 20892-9778, or email at [email protected]. and ranches in all directions downwind from the White (Manuscript accepted 15 June 2020) Sands Gunnery range detonation site. The authors declare no conflicts of interest. 0017-9078/20/0 The Trinity detonation was the proof of principle of the Written work prepared by employees of the Federal Government as theory and mechanical design of the implosion concept de- part of their official duties is, under the U.S. Copyright Act, a "work of veloped by the Manhattan Project and used for the weapon the United States Government" for which copyright protection under Title 17 of the United States Code is not available. As such, copyright does not to be dropped on Nagasaki in August 1945. For those rea- extend to the contributions of employees of the Federal Government. sons, at least, Trinity has a definite place in history. Little DOI: 10.1097/HP.0000000000001328 is known, however, about any health consequences among 428 www.health-physics.com Estimated doses from Trinity nuclear test c S. L. SIMON ET AL. 429 the public as a result of the test. The purpose of this study is location. Each of these factors are further defined here, to present our assessment of the likely health consequences and uncertainty of dose is discussed in a later section. to residents of New Mexico exposed to radioactive fallout Five organs (or groups of organs) were considered in from the Trinity detonation. While in recent years there has the dosimetric analysis. Four were the same as in studies been considerable concern about the health consequences of health risk in the Marshall Islands following nuclear test- among regional New Mexico populations, the magnitude of ing, namely (1) colon, (2) active (red) bone marrow (RBM), these consequences, derived from well-buttressed scientific (3) stomach, and (4) thyroid gland (Land et al. 2010); lung inquiry and analysis, has been absent until now. The impacts, was added to the list of organs to be considered. The five or- in terms of both dose and health consequences from other gans are of generally high risk from exposure to radioactive nuclear test sites, including those in Nevada, Marshall fallout. It was considered beyond the feasibility as well as Islands, French Polynesia, and Kazakhstan (see for example, beyond any prevailing need to assess dose to every organ, Anspaugh and Church 1986; Stevens et al. 1990; Kerber such as eye lens and skin. Skin dose and particularly skin et al. 1993; Land et al. 2008, 2010, 2015; Drozdovitch burns, which had been reported on cattle in some locations et al. 2008; Gilbert et al. 2010), have been studied, leaving following Trinity, do not substantially contribute to the can- the first nuclear test, i.e., Trinity, to be the one of the few cer risk of the five organs (above) that are thought to be at remaining nuclear tests conducted near resident populations highest risk and that are the subject of the risk assessment to have never been studied in great detail. Because the for Trinity (Cahoon et al. 2020). magnitude of exposures received by local populations in The ethnic groups of interest (i.e., Whites, Hispanics, New Mexico, as well as the spatial pattern and local Native Americans,5 and African Americans) were defined heterogeneity, were heretofore not well known, the primary by the reported ethnicities in the US census of 1940 and goal of this study was to conduct a detailed analysis of the 1950. The New Mexico population in 1945 in terms of radiation exposure of residents of New Mexico from the numbers of persons, age distribution, and geography was Trinity fallout. This paper summarizes the findings from derived from US Census reports and is presented in later the dose assessment conducted for that purpose. sections of this report. The age groups used in this work are simplifications of OVERALL GOALS AND DEFINITIONS actual population age distributions but are intentionally con- sistent with age categories used by the International Com- As noted, the overall purpose of this paper is to report mission on Radiological Protection (ICRP) for deriving the findings of the Trinity dose assessment for New Mexico and reporting dosimetric factors necessary for our calcula- residents. For years, a major concern by New Mexico tions. The seven age categories considered were (1) in utero, populations has been the magnitude of doses received. But (2) 0–1 y of age, (3) 1–2yofage,(4)3–7yofage,(5)8–12 as is typical for exposures in the past, dose reconstruction yofage,(6)13–17 y of age, and (7) adults (18+ y). Dietary is challenging. Because organ doses, particularly internal data (Potischman et al. 2020), like census data, were not al- doses, cannot be assessed today by any physical or biological ways collected in the same age categories as used for defining assay, estimates are understandably dependent on models dose categories. In such cases, simple linear interpolation was and on the availability of data relevant to the modeling of used to recast the data into compatible age categories with exposures. The data necessary for the dose reconstruction, those of the ICRP dose coefficients. as well as exposure models used, are presented in Bouville The geographic extent of this work is the entire state of et al. (2020) and along with diet and lifestyle data New Mexico which, in 1945, included populations in 721 (Potischman et al. 2020) compose the necessary components voting precincts in 31 counties.6 Recognizing that fallout to understand the findings presented here. can be deposited at distances of many miles from a In addition to presenting the findings on the magnitude detonation site, particularly along the predominant wind of radiation doses received by regional New Mexico direction, we made the decision not to limit the dose populations from the Trinity detonation, we also report and assessment to only “close-in” counties. For this reason, we discuss the heterogeneity of those doses among regional have estimated doses for all the counties of New Mexico. populations resulting from differences in ethnicity, age, and This allowed us to provide long overdue estimates on location. Our interest in ethnicity is not because of any known genetically based differences in response to exposure but because recognized differences in diet and lifestyle 5The research findings in this paper do not explicitly apply to the people of the Navajo Nation. factors are the determinants of dose. 6While 31 counties existed in New Mexico in 1945, present-day maps In this analysis, dose refers to best estimates of radia- show 33. In 1981, Valencia was split into Cibola and Valencia, the former taking 80% of the original Valencia county land area but has tion absorbed dose (mGy) to specific organs of representa- only about 25% of the present-day population of both counties together. tive persons defined by ethnicity, age, and geographic Los Alamos county was created in 1949. www.health-physics.com 430 Health Physics October 2020, Volume 119, Number 4 exposures without preferential treatment or discrimination of any locations. Areas in other states adjacent to New Mexico (e.g., parts of Colorado, Texas, and Oklahoma) may have also received very low levels of exposure from Trinity, though only New Mexico was included in our study. The intent in this work should be understood to be different from an epidemiological follow-up study de- signed to quantify the risk per unit dose (e.g., in an analyt- ical dose-response study) or to provide information for administering medical countermeasures (e.g., immedi- ately after a radiation accident), or even one to assess individual-based probability of causation. In such cases, individual dose estimation is needed. Because much of the 1945 population is deceased and cannot be queried about their lifestyle in 1945 and because detailed individ- ual (rather than group) exposure-related information (e.g., diet and lifestyle) is impossible to obtain even among those still alive today, the requirements to estimate dose are unique in this work. The requirements of a risk projection are simply to quantify the average dose to each segment of the population with a unique exposure and risk profile and the number of persons exposed in each population segment. For the purposes of the Trinity risk projection, reported elsewhere in this issue by Cahoon et al. (2020), we present and summarize best estimates of dose by ethnicity, age, and county to five organs of representative persons of each ethnicity and age category. Clearly the deposition of fallout and the resulting exposure within any individual county close to the test site (Fig. 1) was very heterogeneous. For that reason, and because the US Census provided popula- tion data at the level of the voting precinct (i.e., on a much finer geographic scale than the size of counties), our dose assessment methods were applied to representative persons of each ethnicity in each of the 721 precincts in existence in 1945. While the exposure-rate information (Quinn 1987; Fig. 1. One-year integral air kerma (outdoor) from fallout deposited Cederwall and Petersen 1990) could be interpolated on a by the Trinity detonation (July 16, 1945 to July 15, 1946) at centroid geographic scale like that of the precinct sizes, other locations of 721 voting precincts and interpolated. Top panel: Air necessary data, such as specific types of foods available, kerma estimates at precinct centroids. Bottom panel: Interpolation map of air kerma. Star on each panel represents approximate location cannot be discerned reliably with the same high spatial of Trinity detonation. Gray rectangular area directly south of Trinity resolution. These restrictions suggest that county-level detonation site is the present-day White Sands Missile Range (in dose estimates by ethnicity and age are the finest spatial 1945 known as White Sands Proving Ground). discrimination that are appropriate for us to report today. populations, the expertise of the NCI dictated the emphasis For these reasons, we weighted our precinct-level dose on cancer risk. estimates by the precinct population size (according to age and ethnicity) to produce county-level ethnicity- and METHODS age-averaged dose estimates. Those estimates are presented in later sections. Dose reconstruction for a risk projection study only re- It is important to note that the risk of cancer is the only quires estimates of dose to representative persons in sub- health endpoint projected (i.e., estimated) in the National groups in which the dose and risk might be differentiated. Cancer Institute (NCI) study described in this issue. While The subgroups in this study that could possibly be distin- radiation can clearly play a role in the induction of guished, termed strata, were potentially based on ethnicity; non-cancer radiation effects (e.g., cataracts) in some sex; age; general geographic region in the state (north/ www.health-physics.com Estimated doses from Trinity nuclear test c S. L. SIMON ET AL. 431 south); environment type (also called ecozone), which in- undocumented exposure pathways may exist, particularly cluded plains, mountains, or plains/mountains; and popula- for populations that have been less well studied and reported tion density (urban and rural). As discussed in Potischman on, such as Native Americans, for example. However, et al. (2020), some combinations of attributes that might de- bounding assumptions can usually be made for most path- fine unique strata were found to be unnecessary or extraor- ways based on arguments of the physical amount of contam- dinarily difficult to define or to characterize. For this dose inated material that might be ingested. In this work, we have reconstruction, we rely on six defined data sets as presented assumed, based on our scientific understanding and years of in Potischman et al. (2020) that include the combinations of eth- experience conducting assessment of exposure to radioac- nicity (White, Hispanic, Native American, African American), tive fallout, that a group of relatively well-understood path- age, ecozone (plains and mountains), and population den- ways of exposure account for the largest proportion of the sity (rural/urban). dose. The important exposure pathways obviously fall un- The doses estimated in this work are those received der the categories of external and internal irradiation, and over the time of 1 y from the date of detonation; i.e., 16 internal dose includes components from both ingestion July 1945 through 15 July 1946. Pragmatic considerations and acute and long-term inhalation. dictated the decision to estimate doses only for the first full As described in Bouville et al. (2020), we accounted year after detonation as opposed to the lifetime dose. First, it for the pathways of exposure for the resident populations has been shown that more than 90% of the infinite-time ex- we believe would most significantly impact the population ternal dose from deposited fallout is received in the first cancer risk, and we applied the models using strata-specific year for fallout transit times up to about 30 h (Simon et al. data. While the individual pathway-specific models are de- 1995). It can also be shown that the annual dose from inter- scribed in the companion paper, for purposes of understanding nal irradiation is much greater in the year immediately fol- the doses reported here, we reiterate the exposure pathways lowing the test than in any subsequent year (Bouville et al. and food types quantitatively considered in this work: 2020). This is primarily a consequence of the short half-lives of the fallout radionuclides that deliver the greatest dose 1. External irradiation; (see Table 1 in Bouville et al. 2020). 2. Consumption of cows’ milk; Methods to estimate internal dose for longer periods 3. Consumption of mothers’ breast milk (specific to nurs- than 1 y are complicated by lack of information on changes ing infants); in diets and in bioavailability of the environmental contam- 4. Consumption of fresh cheese (from cows’ milk); ination over successive years, as well as the requirement to 5. Consumption of fruits and berries; change the biokinetic assumptions of an aging population. 6. Consumption of fruit vegetables; It is known, for example, that significant changes in diets 7. Consumption of leafy vegetables; in post-war years occurred as a result of widespread eco- 8. Consumption of beef (meat); nomic improvements, the introduction of home refrigerators, 9. Consumption of pork (meat); and greater transport and movement of regionally-produced 10. Consumption of mutton (meat); foods. These various social, environmental, and inter-individual 11. Consumption of river and cistern water; biological changes over time would add tremendous com- 12. Inhalation of fallout during the period of deposition; plexity to conducting a lifetime dose assessment. Given and the rapid decay of most of the radionuclides, the component 13. Inhalation of resuspended contaminated dust (entire of dose received with each passing year would contribute year). little to the lifetime dose and would not provide any signif- icant improvement to a dose or risk assessment. Consumption of goats’ milk might also be expected, The population number in each of the 31 New Mexico though as noted in Potischman et al. (2020), in interviews counties in 1945 varied considerably as did the mixture of for this study, the consumption of goats’ milk was reported ethnic groups. Table 1 in this paper provides data on the so infrequently that it could not be assumed to have been a numbers of persons of each ethnicity and age in the 31 commonly consumed food product in this population and, counties as derived and interpolated from the 1940 and furthermore, could not be quantified on a population basis. 1950 US Census reports (US Census 1940, 1950). Thedoseassessmentinthisworkconsistedofanum- As with any population, there can be exposure path- ber of sequential steps beginning with estimation of the ways that are exclusive to small groups of people but that ground deposition density (Bq m−2) for each of the 63 radio- are not well recognized and/or poorly understood. Such nuclides considered (see Table 1 of Bouville et al. 2020) pathways are typically expected to be quite minor in their using interpolated values of exposure-rate, fallout time-of- contribution to the total exposure even though their uncer- arrival, and the refractory to volatile ratio (R/V,see Bouville tainty can be substantial. Here we acknowledge that et al. 2020; Beck et al. 2020) at the centroid location of each www.health-physics.com 3 elhPyisOtbr22,Vlm 1,Nme 4 Number 119, Volume 2020, October Physics Health 432 Table 1. Population by county, age, and ethnicity (interpolated from 1940 and 1950 US Census). All cells are rounded whole numbers. White Hispanic County In-utero 0–1y 1–2y 3–7y 8–12 y 13–17 y Adult Total In-utero 0–1y 1–2y 3–7y 8–12 y 13–17 y Adult Total

Bernalillo 853 1,137 2,275 6,242 5,155 4,943 30,414 51,019 610 814 1,627 4,464 3,687 3,535 21,754 36,492 Catron 49 65 130 357 281 247 1,368 2,497 35 47 93 255 201 177 978 1,786 Chaves 294 392 783 2,150 1,802 1,770 11,181 18,372 210 280 560 1,538 1,289 1,266 7,997 13,141 Colfax 207 276 553 1,513 1,192 1,047 5,799 10,587 148 198 395 1,082 852 749 4,148 7,573 Curry 146 195 390 1,069 879 838 5,120 8,636 104 139 279 764 629 599 3,662 6,177 De Baca 43 57 115 314 247 217 1,204 2,197 31 41 82 225 177 155 861 1,572 Dona Ana 345 460 920 2,523 2,068 1,956 11,857 20,130 247 329 658 1,805 1,479 1,399 8,481 14,398 Eddy 326 434 868 2,380 1,946 1,833 11,059 18,846 233 310 621 1,703 1,392 1,311 7,910 13,480 Grant 241 321 642 1,756 1,383 1,216 6,733 12,291 172 229 459 1,256 989 870 4,816 8,792 Guadalupe 90 120 240 657 517 455 2,519 4,598 64 86 172 470 370 325 1,801 3,289 Harding 38 51 102 280 221 194 1,075 1,963 27 37 73 201 158 139 769 1,404 Hidalgo 57 76 153 419 330 290 1,605 2,929 41 55 109 299 236 207 1,148 2,095 Lea 256 341 682 1,869 1,529 1,442 8,706 14,824 183 244 488 1,337 1,094 1,031 6,227 10,603 Lincoln 91 121 242 661 521 458 2,536 4,629 65 86 173 473 373 328 1,814 3,311 www.health-physics.com Luna 88 117 233 639 503 443 2,450 4,473 63 83 167 457 360 317 1,753 3,199 McKinley 133 177 355 971 765 672 3,722 6,794 95 127 254 694 547 481 2,662 4,859 Mora 115 153 306 837 659 579 3,208 5,857 82 109 219 599 471 414 2,295 4,189 Otero 136 181 362 990 780 685 3,795 6,929 97 129 259 708 558 490 2,715 4,956 Quay 151 201 402 1,101 867 762 4,220 7,703 108 144 288 787 620 545 3,018 5,510 Rio Arriba 271 361 722 1,977 1,557 1,369 7,580 13,838 194 258 517 1,414 1,114 979 5,422 9,898 Roosevelt 180 240 480 1,315 1,036 911 5,043 9,206 129 172 344 941 741 651 3,607 6,585 Sandoval 125 166 332 909 716 630 3,486 6,363 89 119 238 650 512 450 2,493 4,551 San Juan 101 134 268 734 578 508 2,813 5,136 72 96 192 525 413 363 2,012 3,673 San Miguel 279 373 745 2,043 1,669 1,570 9,454 16,133 200 266 533 1,461 1,194 1,123 6,762 11,539 Santa Fe 321 429 857 2,353 1,962 1,911 11,968 19,801 230 306 613 1,683 1,403 1,367 8,560 14,163 Sierra 82 109 217 595 469 412 2,282 4,166 58 78 156 426 335 295 1,632 2,980 Socorro 196 262 523 1,433 1,129 992 5,493 10,028 140 187 374 1,025 807 710 3,929 7,173 Taos 199 265 531 1,453 1,145 1,006 5,570 10,169 142 190 380 1,039 819 720 3,984 7,274 Torrance 111 148 295 808 637 560 3,099 5,658 79 106 211 578 455 400 2,217 4,047 Union 95 127 253 693 546 480 2,657 4,851 68 91 181 496 391 343 1,901 3,470 Valencia 207 275 551 1,508 1,188 1,044 5,781 10,553 148 197 394 1,078 850 747 4,135 7,548 Native American African American

County In-utero 0–1y 1–2y 3–7y 8–12 y 13–17 y Adult Total In-utero 0–1y 1–2y 3–7y 8–12 y 13–17 y Adult Total

Bernalillo 43 57 113 283 198 182 1,055 1,931 11 14 28 82 68 72 715 990 Catron 0 0 0 1 0 0 2 4 00000 0 3 5 Chaves 0 1 1 3 2 3 16 27 81122645357562 777 Colfax 0 0 0 1 1 1 3 6 2351310995136 Curry 1 2 3 8 6 7 41 68 4 5 10 30 25 26 260 360 De Baca 0 0 0 0 0 0 1 2 00000 0 0 0 Dona Ana 2 2 5 12 9 9 52 91 12 16 31 84 65 62 655 925 Eddy 0 0 0 1 0 0 2 4 81122614950509 710 Grant 1 1 3 7 5 4 25 47 2351410997139 Guadalupe 0 0 0 0 0 0 0 0 00000 0 1 2

Harding 0 0 0 0 0 0 0 0 00000 0 1 2 test nuclear Trinity from doses Estimated Hidalgo 0 1 1 3 2 1 9 16 00121 111 16 Lea 1 1 2 5 4 4 25 42 10 13 27 77 63 66 659 915 Lincoln 0 0 0 0 0 0 0 1 00111 110 15 Luna 0 0 1 2 1 1 6 11 11375 550 72 www.health-physics.com McKinley 299 399 798 1,996 1,368 1,144 6,681 12,687 348211614151217 Mora 0 0 0 0 0 0 0 0 00000 0 1 1 Otero 21 28 55 138 95 79 462 877 348201513146210 Quay 0 0 0 1 0 0 2 4 12386 559 84 Rio Arriba 40 53 106 266 182 152 889 1,688 00000 0 2 3 Roosevelt 0 0 0 0 0 0 1 1 00000 0 1 2

Sandoval 81 108 216 541 371 310 1,810 3,437 00000 0 3 4 c .L S L. S. San Juan 207 276 552 1,381 947 792 4,623 8,778 00000 0 3 5

San Miguel 0 0 0 1 0 0 3 5 01132 225 35 AL ET IMON Santa Fe 18 24 48 120 85 83 476 854 347211718176244 Sierra 0 0 0 1 0 0 2 4 11265 444 63

Socorro 7 10 19 49 33 28 163 309 00000 0 3 4 . Taos 18 24 49 122 84 70 409 777 00000 0 1 2 Torrance 0 0 0 1 0 0 2 4 00000 0 0 0 Union 0 0 0 0 0 0 0 0 00132 218 26 Valencia 90 121 241 603 413 346 2,018 3,833 00011 1 6 9 433 434 Health Physics October 2020, Volume 119, Number 4 of the 721 voting precincts. For the purposes of our calcula- later section titled VALIDATION. As described in Bouville tions, deposition density was estimated at the approximate et al. (2020), the exposure rate at location L normalized to time of peak fallout arrival measured from the time of deto- 12 h post-detonation [i.e., ẊðÞ12; L ] as reported by nation (5:29 AM Mountain Time). Because the time interval Quinn (1987), is modified to account for the actual TOA from detonation to deposition is needed to properly account at each location of interest and the degree of fractionation for radioactive decay during the transit of debris from the of the fallout [see Appendix of Beck et al. (2020) for detonation site to the deposition site, fallout “time-of-arrival” further information]. (TOA, h) was a necessary parameter. In this study, the TOA From each of the 45,000+ deposition density estimates was estimated by interpolation of TOA data reported by in this work (721 precincts 63 radionuclides), N50,i.e., Quinn (1987) for the Trinity detonation. See Bouville et al. the fraction of fallout particles less than 50 mmindiameter, (2020) for further discussion on estimation of TOA and was estimated (see Bouville et al. 2020 for a discussion of exposure-rate at TOA. models) as an intermediate step to determine the contamina- Voting precincts were always relatively small areas and, tion of plants with radioactive particulates that are retained of course, dependent on population number and area. In on the leaf surfaces, a phenomenon that leads to each precinct, we assumed that estimated deposition density food-chain contamination. N50 was found to range from and TOA were relatively uniform and representative of the a minimum of 0.062 at locations very close to the detonation exposure conditions to the entire resident population of site where particles were predominantly much larger than 50 the precinct. mm and poorly retained on plant surfaces to unity at locations It is important to recognize that, similar to other dose where the fallout deposition occurred at times greater than reconstructions of fallout exposures in Nevada and Utah, about 14 h after the detonation and the particles were Marshall Islands, and Kazakhstan (Ng et al. 1990; Lloyd smaller than 50 mm. et al. 1990; Simon et al. 1990, 2002, 2006, 2010; NCI Ground deposition and N50 estimates for the 63 radio- 1997; Bouville et al. 2002, 2010; US DHHS 2005; nuclides were used as the basis for estimating contamination Gordeev et al. 2006a and b; Beck et al. 2006, 2010), the of both pasture plants eaten by grazing animals and local deposition density was derived from measurements human-edible plant foods [e.g., fruits and berries, fruit veg- (or interpolation of local measurements) of exposure-rate etables (such as tomatoes, peppers, squash, melons, etc.), or gamma spectrometry radionuclide concentrations in soil and leafy vegetables (such as spinach, greens, lettuce, and radioactivity measurements on gummed film (Bouville etc.)]. Time integrated concentrations (Bq d kg−1) of all ra- and Beck 2000) rather than from the use of atmospheric dis- dionuclides were calculated for all categories of pasture persion codes and models. There appears to be occasional grass and human-edible plants at all relevant locations by confusion among the public (TBDC 2017, p. 34) that at- considering biomass yield, weathering half-time, and other mospheric dispersion models, which are known to be related variables (Bouville et al. 2020). quite uncertain for complex radiation releases (e.g., nu- Intakes of radionuclides by animals in the food chain clear detonations), were the primary source of informa- were calculated for each plant-based food using conven- tion on fallout deposition in this and other fallout-related tional data on animal nutritional requirements appropriate dose reconstruction studies. While atmospheric dispersion to the mid-1940s. Intakes of radionuclides directly by model predictions are sometimes compared against ground humans were calculated for each plant and animal-based measurements as the basis for improving our understanding food as well as from drinking water and from inhalation at of dispersion processes (Cederwall and Peterson 1990; the time of fallout (termed “in-cloud” inhalation) and from Moroz et al. 2010), such models were only used for long-term inhalation of resuspended contaminated soil. estimating ground deposition in the Trinity study at Models of radionuclide intake by animals or humans are locations beyond the Quinn fallout deposition pattern. In similar for all sources of internal contamination in that each this study, ground deposition of fallout at all precincts in calculation requires the time-integrated concentration in the the Quinn pattern was estimated from interpolated values of source material and the rate of intake of that source material, actual ground-level exposure-rate measurements (Quinn a parameter that typically varies by ethnicity and age for 1987; also see Fig. 3 of Beck et al. 2020) taken within 21 d humans. Radionuclide intakes of plant-based foods were of the Trinity detonation. Beyond the Quinn pattern, near the calculated for all combinations of ethnicity/age/precinct as border of New Mexico and Colorado, the fallout pattern was the product of the time-integrated concentration (Bq−d supplemented with computer model calculations by kg−1) in the food product and the intake-rate (kg d−1)of Cederwall and Peterson (1990). At locations to the east, the food (Bouville et al. 2020). Intakes of radionuclides west, and south of the Quinn pattern, the absence of from animal products (cows’ milk, cow cheese, beef, mutton, significant fallout was confirmed by measurements from pork) were calculated as the products of time-integrated x-ray film badges (Hoffman 1945) discussed further in a concentrations in the food product, based on contamination www.health-physics.com Estimated doses from Trinity nuclear test c S. L. SIMON ET AL. 435 of pasture grass coupled with feed-milk or feed-meat factors estimating contamination of breast milk from reports radionuclide-specific transfer coefficients, and daily intakes of the International Commission on Radiological Protection rates by man, in a similar fashion to calculations for plant (ICRP). While the time-period of breastfeeding can vary by foods. Intakes of radionuclides from potable drinking water ethnic group and by family (Potischman et al. 2020), we as- sources, primarily cistern water, potentially contaminated sumed that, on average, breastfeeding continued for 12 mo, by rainfall events during the fallout cloud passage time and the intake rate of mothers’ milk for nursing infants was and from water derived from rivers, were calculated at 0.8 L d−1. the precinct level (Bouville et al. 2020). Well water was as- Estimates of intakes of radionuclides by ingestion and sumed not to be contaminated, and human consumption of inhalation were converted to organ dose using dose coeffi- water from open irrigation ditches, known as acequias, cients (mGy Bq−1) derived from publications of the ICRP, was assumed to have been too rare to account for on a as described in Bouville et al. (2020), with assignments of population-average basis. solubility class for lung and the gastrointestinal tract appro- Inhalation of radionuclides in air was accounted for (1) priate for regional fallout from nuclear testing (Ibrahim during the period when fallout was being deposited (i.e., et al. 2010). “in cloud” intake) and (2) during the entire first year from the resuspension of contaminated soil particles. FINDINGS AND DISCUSSION The calculation for the intake of radionuclides by inhala- tion during the period of fallout deposition accounts for In this section, we present graphical analyses illustrat- radioactive decay as well as for the approximate differ- ing (1) the relative magnitude of each exposure pathway ences in particle size with distance downwind, the latter by ethnicity and age, (2) the relative magnitude of estimated factor being important to determine whether the parti- doses for each of the five organs of interest by ethnicity, and cles are small enough to reach the deep lung after inha- (3) a ranking of counties in terms of population-weighted lation. “In cloud” inhalation only takes place after the dose (accounting for the ethnic and age distribution of the onset of deposition for a time roughly equal to (or county). In general, the analyses presented are intended to slightly less) than TOA, while inhalation from resuspen- illustrate differences in doses to the populations of New sion continues during the entire year. Mexico as well as the range and heterogeneity of dose In contrast to “in cloud” inhalation, the potential for the within each population group. We also present in a intake of soil by resuspension is known to be an ongoing series of tables the relative importance of individual phenomenon, though physical weathering of fallout parti- fallout radionuclides to the total dose at three locations of cles and their downward migration into the soil column re- increasing TOA. sults in a significant decrease in the availability of activity A discussion of our estimates of uncertainty is pro- for resuspension with the passage of time after deposition. vided to illustrate a lower and upper credible range on our Due to the downward migration of radionuclides in the soil best estimates of doses. Best estimates of doses were deter- column, and because only surface soil is available for re- mined directly from application of dose estimation formulae suspension, resuspension models (Maxwell and Anspaugh using diet input data, presented in companion papers 2011) predict the magnitude of the activity to be resus- (Bouville et al. 2020; Potischman et al. 2020). The deriva- pended in the second year after deposition to be less than tion of uncertainty, discussed elsewhere in this paper, could 1% of the activity available for resuspension in the first alternatively be used to estimate the mean value from the year. Moreover, the fraction available for inhalation con- resulting uncertainty distribution. For maximum transpar- tinues to decrease significantly in successive years. This ency, however, our reported dose estimates are derived di- phenomenon provides the primary rationale for limiting rectly from the application of the dose estimation formulae the calculation of dose from resuspension to the first year. and are closer to median rather than mean values of the un- See Bouville et al. (2020) for further detail about inhala- certainty distributions. The Appendix of this report presents tion models. tables of best estimates of county population-weighted aver- Another important pathway of exposure is by mothers’ age organ doses by ethnicity and age as a record for further milk and is only relevant, of course, for nursing infants. In research and archival purposes. this work, the radionuclide mixture and the possible degree External doses: magnitudes and spatial pattern of contamination of mothers’ breast milk was determined External doses to support the risk projection were esti- according to precinct location by considering the local mated to the whole-body as well as the five organs of inter- ground deposition density of each radionuclide, the intake est by county, ethnicity, and age. While the air kerma from of radionuclides according to the diet of adults of the ethnic deposited fallout is applied to all persons in each precinct, group under consideration, as well as from water and inha- the dose to each ethnic population in a precinct was mod- lation. All intakes used radionuclide-specific transfer ified to account for the shielding from common home www.health-physics.com 436 Health Physics October 2020, Volume 119, Number 4 construction materials and the reported time spent outdoors about 100 mGy for Whites, about 0.006 to 50 mGy for per day in summer months (Potischman et al. 2020; Bouville Native Americans, and 0.004 to about 100 mGy for et al. 2020). Hispanics and African Americans (all doses rounded to The maximum assigned exposure rate at 12 h post- two significant digits or less). detonation [i.e., Ẋ(12)] of 481 mR h−1 was for a single pre- The dose derived from external irradiation, in this cir- cinct in southern Torrance county, directly downwind of the cumstance, is approximately equal (ICRP 2010) for all organs Trinity detonation site.7 Sites directly on the periphery of the of the body since the energy of externally-received gamma fallout pattern were estimated to be at least 2,000-fold smaller rays from fallout is sufficient to completely penetrate the body. than the maximum, and locations well outside the pattern The variation of external doses with age was only about 30% were estimated to be up to 10,000-fold smaller. Relatively (ICRP 2010) with younger children receiving modestly greater high ẊðÞ12 contours from Trinity intersected numerous external doses because of smaller body sizes. External doses precincts in several counties, but because of the narrow- varied to a small degree between ethnic groups because of ness of the high exposure-rate contours (see Fig. 1 of modest differences in time spent outdoors and differences in Bouville et al. 2020), most of those counties had lower home occupancy and building shielding factors. area-averaged ẊðÞ12 values when each of the precinct’s areas with unique exposure rates, expressed as a fraction Comparison of doses by exposure pathway of the total county area, were used as weighting factors. Here we illustrate the importance of individual expo- Counties and precincts outside of the Quinn deposition pattern sure pathways to the total organ dose for adults. Data for − were conservatively assigned ẊðÞ12 values of 0.05 mR h 1 all data sets cannot be presented, nor is it necessary to do and TOA values from 12 to 36 h depending on the location so since the relationships between dose from individual ex- of the county. The spatial pattern of ẊðÞ12 strongly reflects posure pathways and food types for adults are reasonably the movement of the Trinity fallout cloud to the northeast similar to the relationships for other age groups, except for ’ from the detonation site as reported by Quinn (1987). 0-1 y of age, where only consumption of mothers milk is The time-integrated air kerma can be derived from assumed. This discussion focuses only on the relative im- X12̇ðÞand TOA as shown elsewhere (Simon et al. 1995; portance of individual pathways and food types to thyroid Bouville et al. 2020). Using such calculations, the spatial pat- dose since it is the organ that received, by far, the largest tern of the 1-y integral (outdoor) air kerma at locations across doses. Other organs would show a different ranking for the New Mexico can be derived. See the top panel of Fig. 1, importance of food types. which directly reflects the fallout deposition or X12̇ðÞ The ingestion doses within each ethnic group from pattern. As described earlier, Ẋ(12) was interpolated between each food type, external dose, inhalation, and resuspension, the isopleths at the locations of the centroid of each voting averaged over the entire population of New Mexico, can be precinct. Once interpolated, we assumed that exposure-rate compared in order to judge the importance of different routes of value to be relatively uniform across the precinct, consistent exposure within the ethnic group. However, the heterogeneity with our assumption that the populations were uniformly of doses among the precincts and counties results in different distributed in each precinct. The top panel of Fig. 1 shows comparisons depending on whether mean doses or median the precinct locations used to derive the interpolation map doses are compared. To minimize the effect of substantial in the lower panel. skewness in dose distributions, we compare median doses The outdoor 1-y integral air kerma (Fig. 1 lower panel) from each route of exposure. can be seen to be significantly elevated only in the localized Among adults, comparing the median dose for each ’ region directly northeast of the Trinity detonation site and food type (other than cows milk) to the median dose from ’ reached values up to 200-300 mGy (1-y integral air kerma) cows milk illustrates the relative importance of each path- over relatively small and sparsely populated areas. Outdoor way for each ethnic group as follows. Numbers in parenthe- ’ air kerma is an intermediate step to estimating external ses are the median dose relative to cows milk: dose received by persons living in the region since the air Whites - Cows’ milk (1.0) : Water (0.49) : Leafy vege- kerma need only to be modified by the time spent outdoors tables (0.40) : External Dose (0.095) : Fruit Vegetables : per day, a function of ethnicity and age, and the shielding (0.070) : Inhalation (0.061). All other routes of intake were provided by the type of residential or workplace construction less than 5% of the dose from cows’ milk; and by the body. One-year integral external doses, after Hispanics - Cows’ milk (1.0) : Water (0.49) : Leafy accounting for home shielding and time spent outdoors, vegetables (0.40) : External Dose (0.088) : Fruit Vegetables : varied among the 721 precincts from about 0.006 mGy to (0.070) : Inhalation (0.061). All other routes of intake were less than 5% of the dose from cows’ milk; ’ 7Traditional units are being used here to maintain consistency with the Native Americans - Cows milk (1.0) : Leafy vegetables historical data. (0.50) : Fruits and Berries (0.35) : Water (0.28) : External www.health-physics.com Estimated doses from Trinity nuclear test c S. L. SIMON ET AL. 437 (0.11) : Fruit Vegetables (0.087) : Inhalation (0.059). All substantially different for others, the differences primarily other routes of intake were less than 5% of the dose from reflecting differences in the internal dose component. Differ- cows’ milk; and ences in internal dose arise because of the differing chemi- African Americans - Cows’ milk (1.0) : Water (0.24) : cal and biokinetic characteristics of the radionuclides Leafy vegetables (0.24) : Fruits and Berries (0.059) : Exter- ingested in foods and water and inhaled in air. In general, nal (0.056). All other routes of intake were less than 5% of the thyroid gland received the highest internal dose, regard- the dose from cows’ milk. less of ethnicity or age. The larger doses resulted because of In some instances, the dose from drinking water was the predisposition of the thyroid gland to accumulate io- moderately high relative to cows’ milk. However, that occurred dine; an attribute unique to the thyroid. Other organs for only for the subset of persons resident at locations where river which dose was computed were, in general, smaller in mag- water had been contaminated by the temporal coincidence of nitude than the doses for the thyroid gland. rainfall and the passage of the fallout cloud. Because rainfall Fig. 3 presents a comparison of total organ doses to is episodic and infrequent in the desert environment, such a co- adults by ethnicity for the purposes of comparing organ incidence was, in general, not common. Models to estimate dose without the complication introduced by mixing persons this are discussed more in Bouville et al. (2020). of different ages. Only very minor differences were apparent Inhalation (“in-cloud”) dose and, to a much lower de- between ethnic groups in the ratio of thyroid dose to dose to gree, resuspension dose, were extremely small contributions other organs (colon, lung, RBM, and stomach). to thyroid dose, almost always much less than 5% of the Among adults, comparing the median dose to each or- dose from cows’ milk. More detail on the range of doses gan (other than thyroid) to the median dose to the thyroid il- from inhalation and resuspension are presented in Fig. 2a–d. lustrates the relative ranking of organs in terms of the dose Very small contributions to total dose were also contrib- received. Numbers in parentheses are the median dose rela- uted by animal meat products (beef, mutton, pork), primar- tive to that received by the thyroid gland: ily because of the additional limiting steps of transfer of Whites - Thyroid (1.0) : Colon (0.34) : Lung (0.083) : radionuclide activity at successive steps in the food chain Stomach (0.067) : RBM (0.065) and because consumption of animal meat in 1945 was a Hispanics - Thyroid (1.0) : Colon (0.34) : Lung (0.077) : luxury and consumption rates were reported to be low Stomach (0.069): RBM (0.065) (Potischman et al. 2020). The dose contributions from all Native Americans - Thyroid (1.0) : Colon (0.37) : Lung food types and exposure pathways including inhalation and (0.071) : Stomach (0.066): RBM (0.059) resuspension are provided in the panels of Fig. 2. African Americans - Thyroid (1.0) : Colon (0.36) : Lung (0.079) : Stomach (0.051) : RBM (0.051). Comparison of doses to organs of the body As can be seen here and in Fig. 3, the doses to the lung, The total organ dose, i.e., the sum of external plus in- stomach, and red bone marrow were similar in magnitude. ternal doses, can be similar for some organs and The moderately large relative dose to the colon arises

Fig. 2. Comparison of pathway and food type contributions to thyroid dose of adults by ethnicity: Whites (panel A), Native Americans (panel B), Hispanics (panel C), African Americans (panel D).

www.health-physics.com 438 Health Physics October 2020, Volume 119, Number 4

Fig. 3. Comparison of total organ doses to adults by ethnicity: Whites (panel A), Native Americans (panel B), Hispanics (panel C), African Amer- icans (panel D). Total organ dose is a sum over all exposure sources shown in Fig. 2. Note different y-axis scaling for each panel. because the chemical and physical form of the radionuclides factors. For external dose, the important factors are the following their vaporization and condensation as fallout TOA at the location of interest, the length of the time period particles is in the form of relatively insoluble particles over which dose is integrated, and, to a much smaller de- (Ibrahim et al. 2010). The colon is exposed during the final gree, the R/V ratio at that location. For internal dose, the steps of digestion of food material and its transfer to solid above factors are important as well as the relative magnitude waste in the lower part of the intestinal tract. of different components of the diet. The absolute values of doses to each organ received by The rankings provided in this section give an indication the different ethnic groups are illustrated in more detail in of the approximate relative importance of each radionuclide, Fig. 3 and can be attributed mainly to specific dietary differ- though the exact rank position of each nuclide among a ences. In general, higher doses, especially to the thyroid group of 20 or more should not be considered as precise in gland, would result from consumption of foods with greater all circumstances. The ranking is, of course, based on calcu- radioactivity content, e.g., fresh milk products. lations and depends on assumed parameter values in those calculations, e.g., TOA, transfer coefficients to individual Comparison of doses by age at time of exposure food products, and other variables. The effect of age alone (i.e., age in 1945 at time of ex- In Tables 2 and 3, we present analyses that illustrate our posure) on the dose received by the thyroid gland is illus- findings on the relative importance of individual radionu- trated in Fig. 4 for the ethnic groups. In the case of the clides to the 1-y integral external and internal doses at three dose to the thyroid gland, the average dose in each age different precincts covering a range of fallout TOAs: 3.1 h group relative to the average adult dose, regardless of eth- (Socorro County), 10.5 h (Bernalillo County), and 36.3 h nicity, is governed by the internal dose from consumption (Colfax County). See Fig. 1 of Bouville for county locations. of food products that contain substantial radioiodine and, The TOA is important because of the considerable decay of to a lesser degree, by external dose. Differences in internal the short-lived radionuclides during fallout transit. Two of doses by age are primarily due to age-dependence in dose the precinct locations, the closest being in Socorro County conversion factors and differences in dietary intake. and the furthest being in Colfax County, were rural/ Here, we present a comparison of thyroid dose by age, mountain locations, while the intermediate location in relative to the adult dose, averaged over the ethnicities: Bernalillo County was an urban/plains location. The environ- In-utero (~0.8) : 0–1 y (~2.0) : 1–2 y (4.5) : 3–7y(3.5): mental differences result in modest differences in availability 8–12 y (2.5) : 13–17 y (~2.0) : Adult (1.0). of foods and diet and, hence, the relative amounts of intake of Ranking of radionuclides to total dose different radionuclides, but more importantly, the change in Individual radionuclides contribute different fractions TOA among the locations demonstrates a change in the im- to the total body and organ dose depending on several portance of radionuclides according to their half-life. www.health-physics.com Estimated doses from Trinity nuclear test c S. L. SIMON ET AL. 439

Fig. 4. Comparison of total organ doses to adults by age and ethnicity: Whites (panel A), Native Americans (panel B), Hispanics (panel C), African Americans (panel D). Thyroid dose is a sum over all exposure sources shown in Fig. 2. Note that age groups are not equal sizes.

The 22 radionuclides presented in Table 2 contribute, produced a significant amount of 237Uand239Np during the collectively, 95% or more of the 1-y integral external dose at detonation as a result of neutron capture by 238U (Beck et al. all three locations. More than 50% of the external dose, regard- 2020). The deposition density of 237Uand239Np, each nor- less of location, is contributed by 239Np, 140La, 95Nb, and 132I. malized to exposure rate at 12 h (i.e., mCi per m2 per mR per Similarly, Table 3a–e presents analyses that illustrate h at H+12), was reported by Hicks (1985) along with the es- the relative importance of individual radionuclides to the in- timates of normalized deposition factors for all other fission ternal dose (ingestion and inhalation + resuspension sepa- and activation products produced in the Trinity detonation rately) of the five organs studied. Analyses not shown (see Bouville 2020, this issue, for more detail). The indicate little difference in the relative ranking of radionu- site-specific deposition densities for 237U and 239Np, as clides by age or ethnicity at the same location. For that rea- for all radionuclides in this study, were estimated at each lo- son, we restrict the presentation on the relative importance cation by the product of the reported normalized deposition of individual radionuclides to adult Whites and Hispanics. factor for the specific radionuclide (Hicks 1985) and the ex- We present the ranking at the same three locations discussed posure rate at the location. above for external dose. For all organs, 90% or more of the ingestion dose, as well as the inhalation plus resuspension Comparison of county average doses dose, was contributed by 20 of the 63 radionuclides for A comparison of the population-weighted average dose which doses were estimated (Table 3a–e), though the rela- by county provides a relative ranking of the county-average tive radionuclide contributions depended on the organ, ex- exposures received in each of the 31 counties of New posure pathway, and TOA of the location where the exposure Mexico that existed in 1945. The relative position of each was assumed to have been received. Large contributors to county in the overall ranking is dependent both on the best colon, lung, red bone marrow, and stomach dose, both by in- estimate of dose to each ethnicity and age group in the gestion and inhalation (including resuspension), were 239Np, county but also on the relative numbers of persons of each 97Zr, 237U, 89Sr, 140Ba, 132Te, and 140La. As expected, the thy- ethnicity and age who were present in the county at the roid gland received nearly its entire dose from 131Iand133I time of exposure as estimated from the census. Because with smaller contributions from 132Te, 135I, and 131mTe. the diets and lifestyles of each ethnic group are different, The importance of two radionuclides, 237Uand239Np, the presence or absence of an ethnic group affects the total is worthy of mention as neither is a fission-product. In the population average dose. Counties do not have equal doses case of the Trinity device, part of the design was a heavy for all ethnic groups as ethnic-specific dose varies by uranium tamper/reflector around the plutonium core, which precinct, and the county-average dose depends on which www.health-physics.com 4 elhPyisOtbr22,Vlm 1,Nme 4 Number 119, Volume 2020, October Physics Health 440

Table 2. Radionuclides at three locations that contribute at least 95% of the first-year external dose: ranking by their fractional contribution to dose. Precinct 581, Socorro County TOA = 3.1 h Precinct 13, Bernalillo County TOA = 10.5 h Precinct 75, Colfax County TOA = 36.3 h Rank Radionuclide Fraction of annual dose Cumulative fraction Radio-nuclide Fraction of annual dose Cumu-lative fraction Radionuclide Fraction of annual dose Cumulative fraction

1 Np-239 0.160 0.16 I-132 0.194 0.19 La-140 0.240 0.24 2 La-140 0.128 0.29 La-140 0.189 0.38 I-132 0.207 0.45 3 Nb-95 0.117 0.41 Np-239 0.099 0.48 Nb-95 0.100 0.55 4 I-132 0.102 0.51 Nb-95 0.079 0.56 Np-239 0.098 0.64 5 I-135 0.057 0.56 I-133 0.061 0.62 Zr-95 0.048 0.69 6 Z-r95 0.057 0.62 I-135 0.054 0.68 Cs-137 0.044 0.74 Zr-97/ Ru-103/ 7 Nb-97m 0.049 0.67 Zr-95 0.038 0.71 Rh-103m 0.041 0.78 www.health-physics.com 8 La-142 0.045 0.71 Cs-137 0.033 0.75 I-131 0.036 0.81 Ru-103/ 9 I-133 0.037 0.75 Rh-103m 0.031 0.78 I-133 0.030 0.84 10 Sr-92 0.027 0.78 I-131 0.030 0.81 Ru-106 0.023 0.87 Zr-97/ 11 U-237 0.022 0.80 Nb-97m 0.025 0.83 Te-132 0.019 0.89 12 Ru-105 0.022 0.82 Te-132 0.018 0.85 U-237 0.018 0.90 13 Sr-91 0.019 0.84 Sr-91 0.017 0.87 Ba-140 0.017 0.92 Zr-97/ 14 Ce-143 0.016 0.86 Ru-106 0.017 0.88 Nb-97m 0.012 0.93 Ru-103/ 15 Rh-103m 0.016 0.87 U-237 0.015 0.90 Mo-99 0.009 0.94 16 I-131 0.014 0.89 Ru-105 0.014 0.91 Te-131m 0.008 0.95 17 Mo-99 0.014 0.90 Ba-140 0.014 0.93 Ce-143 0.007 0.96 18 Cs-137 0.010 0.91 Te-131m 0.011 0.94 Co-60 0.006 0.96 19 Sb-129 0.010 0.92 Ce-143 0.010 0.95 Ce-141 0.006 0.97 20 Te-132 0.009 0.93 Mo-99 0.009 0.96 Sb-127 0.006 0.98 21 Ba-140 0.009 0.94 Sb-129 0.006 0.96 Rh-105 0.005 0.98 22 Ru-106 0.008 0.95 Sb-127 0.005 0.97 I-135 0.005 0.99 Table 3a. Radionuclides at three locations that contribute at least 90% of the first-year internal dose to colon (Whites/Hispanics): ranking by their fractional contribution to dose. INGESTION Pct. 581, TOA= 3.1 h, Rural/Mountains Pct. 13, TOA= 10.5 h, Urban/Plains Pct. 75, TOA= 36.3 h, Rural/Mountains Data set B (see Bouville et al., this issue, this issue) Data Set D (see Bouville et al., this issue, this issue) Data set B (see Bouville et al., this issue) Cumulative fraction Cumulative fraction Cumulative fraction RANK Radio-nuclide Fraction of total dose of total dose Radio-nuclide Fraction of total dose of total dose Radio-nuclide Fraction of total dose of total dose

1 Np-239 0.29 0.29 U-237 0.55 0.55 Np-239 0.14 0.14 2 Te-132 0.07 0.37 Sr-89 0.18 0.73 Ba-140 0.14 0.28 3 Zr-97 0.07 0.44 Te-132 0.08 0.81 Te-132 0.13 0.41 4 Ba-140 0.07 0.51 Mo-99 0.06 0.87 La-140 0.10 0.50 5 U-237 0.06 0.57 Np-239 0.03 0.90 U-237 0.09 0.60 6 La-140 0.06 0.64 I-131 0.01 0.92 Ru-106 0.08 0.67

7 Ce-143 0.04 0.67 I-133 0.01 0.93 Ce-144/Pr-144 0.04 0.71 test nuclear Trinity from doses Estimated 8 Y-92 0.04 0.71 Ba-140 0.01 0.94 Sr-89 0.03 0.74 9 Mo-99 0.03 0.74 La-140 0.01 0.95 Mo-99 0.03 0.78 10 Y-93 0.03 0.77 Ru-106 0.01 0.96 Ru-103/Rh-103m 0.03 0.81 11 Ru-106 0.03 0.80 Nd-147 0.01 0.97 Ce-143 0.03 0.83 www.health-physics.com 12 Ce-144/Pr-144 0.02 0.83 Ru-103/Rh-103m 0.00 0.98 Zr-97 0.03 0.86 13 Sr-89 0.02 0.84 Zr-97 0.00 0.98 I-131 0.02 0.89 14 I-133 0.02 0.86 Ce-144/Pr-144 0.00 0.98 Pr-143 0.02 0.90 15 Nd-147 0.01 0.87 Pr-143 0.00 0.98 Nd-147 0.02 0.92 16 Ru-103/Rh-103m 0.01 0.89 Ce-143 0.00 0.99 I-133 0.01 0.93 17 Sr-91 0.01 0.90 Pm-149 0.00 0.99 Rh-105 0.01 0.94

18 Pm-149 0.01 0.91 Rh-105 0.00 0.99 Pm-149 0.01 0.96 c 19 I-131 0.01 0.92 Sb-127 0.00 0.99 Sb-127 0.01 0.97 S L. S. 20 Rh-105 0.01 0.93 Cs-137 0.00 0.99 Zr-95 0.01 0.97 MNE AL ET IMON Inhalation + Resuspension Pct. 581, TOA = 3.1 h, Rural/Mountains Pct. 13, TOA = 10.5 h, Urban/Plains Pct. 75, TOA = 36.3 h, Rural/Mountains

Data set B (see Bouville et al. 2020) Data Set D (see Bouville et al. 2020) Data set B (see Bouville et al. 2020) . Cumulative fraction Cumulative fraction Cumulative fraction Rank Radionuclide Fraction of total dose of total dose Radionuclide Fraction of total dose of total dose Radionuclide Fraction of total dose of total dose

1 Np-239 0.26 0.26 Np-239 0.28 0.28 Np-239 0.39 0.39 2 Zr-97 0.14 0.41 Zr-97 0.13 0.41 Te-132 0.13 0.52 3 Y-93 0.10 0.50 Te-132 0.09 0.50 Zr-97 0.08 0.61 4 Sr-92 0.07 0.58 Y-93 0.07 0.57 Ba-140 0.06 0.66 5 La-141 0.05 0.62 Sr-91 0.05 0.62 Mo-99 0.05 0.72 6 Ce-143 0.04 0.66 Ce-143 0.04 0.66 Ce-143 0.05 0.76 7 Mo-99 0.04 0.70 Mo-99 0.04 0.70 U-237 0.04 0.80 8 Y-92 0.04 0.74 Ba-140 0.04 0.74 Rh-105 0.03 0.83 9 Sr-91 0.03 0.78 Y-92 0.03 0.77 Y-93 0.02 0.86 441 Continued next page 442 Health Physics October 2020, Volume 119, Number 4 precincts each ethnic group resided in the county and the number of persons of that ethnic group present there. The four panels of Fig. 5 graphically present the rank- ing of the 31 counties in terms of population-weighted total

of total dose thyroid dose to adults for each of the four ethnicities. Dose

Cumulative fraction distributions for Hispanic and White populations were sim- ilar. For most counties, average doses to Native Americans were lower than for other ethnic groups except for Torrance County. As can be seen, population age-weighted doses for each group ranged from small fractions of a mGy up to about 50–60 mGy for all ethnic groups but Native Ameri- cans. In Torrance County, the average dose for Native Americans was estimated to be about 80 mGy, though the Data set B (see Bouville et al. 2020)

Pct. 75, TOA = 36.3 h, Rural/Mountains census population data indicates there were only four Native American adults present in Torrance County and, therefore, that average may not be representative and should be con- sidered with caution. The four counties (Torrance, Guadalupe, Lincoln, and San Miguel) had the greatest thyroid doses primarily be- cause of the greater depositions of fallout in each. While the Trinity test took place in Socorro county, the fallout pat- tern of Quinn (1987) illustrates that the deposition occurred of total dose Radionuclide Fraction of total dose to the northeast of the detonation site, primarily in other Cumulative fraction counties, resulting in average doses to Socorro county that were not extraordinarily large. Average doses to Native Americans reflected the counties and precincts that Native ouville et al. 2020) Americans resided in, which, for the most part, were not within the fallout pattern to any significant degree. Torrance

Inhalation + Resuspension County, relatively close to the Trinity detonation site, was an exception in that Native Americans lived in three precincts, in- Pct. 13, TOA = 10.5 h, Urban/Plains Data Set D (see B cluding one in which the largest thyroid doses were received. As an outcome of this analysis, we were able to esti- mate the population-weighted dose to residents of four counties suggested by TBDC (2017) to be at high radiation risk, and by inference, to have received higher radiation doses: Socorro, Lincoln, Otero, and Sierra. It should be noted that TBDC (2017) identified the health risk based on the findings from a health survey; however, health risk of total dose Radionuclide Fraction of total dose in the Tularosa Basin Downwinders Consortium (TBDC) Cumulative fraction report was not defined in conventional scientific terms as cases per 100,000 persons at risk but rather on the basis of the absolute number of reported cases—without reference to the size of the underlying population at risk. As the data on measurements of fallout exposure-rates (Quinn 1987) show a pattern of deposition that moved in a northwesterly direction from the Trinity test site in western Socorro

Data set B (see Bouville et al. 2020) county, it is reasonable to assume that the counties of Otero Pct. 581, TOA = 3.1 h, Rural/Mountains and Sierra would have received very low to negligible expo- sure and that the counties of Socorro, Lincoln, Torrance,

(Continued) Guadalupe, and San Miguel would likely have received the highest exposures. The data in Table 4 confirms that as- sumption. The few high dose precincts in Socorro (where 1011 Pr-14512 Te-13213 U-23714 Ru-10515 Ba-14016 La-142 0.0317 0.03 U-24018 Sb-129 0.0319 Pm-149 0.0220 Pm-151 0.02 Ce-144/Pr-144 0.01 0.81 0.84 0.01 0.01 0.87 0.01 0.01 0.89 0.01 0.91 Ru-105 Rh-105 0.92 0.93 U-237 La-141 0.94 0.02 0.94 0.96 Pr-145 0.02 0.95 Sr-92 0.02 Sb-129 0.02 Pm-149 Te-131m 0.02 Ru106 0.79 U-240 0.01 0.82 0.01 0.01 0.01 0.84 0.86 0.01 0.01 Sr-91 0.88 La-140 0.89 Ru-106 Pm-149 0.90 0.91 0.92 Sr-89 0.94 Te-131m 0.93 Sb-127 0.02 Ru-103/Rh-103m 0.02 Pr-143 Nd-147 0.01 0.01 Ce-144/Pr-144 0.01 0.01 0.01 0.87 0.01 0.89 0.01 0.01 0.90 0.01 0.91 0.95 0.92 0.93 0.94 0.96 0.95 0.97 Rank Radionuclide Fraction of total dose

Table 3a. Trinity was conducted) and Lincoln result in the relatively www.health-physics.com Table 3b. Radionuclides at three locations that contribute at least 90% of the first-year internal dose to lung (Whites/Hispanics): ranking by their fractional contribution to dose. Ingestion Pct. 581, TOA = 3.1 h, Rural/Mountains Pct. 13, TOA = 10.5 h, Urban/Plains Pct. 75, TOA = 36.3 h, Rural/Mountains Data set B (see Bouville et al. 2020) Data Set D (see Bouville et al. 2020) Data set B (see Bouville et al. 2020) Cumulative fraction Cumulative fraction Cumulative fraction Rank Radionuclide Fraction of total dose of total dose Radionuclide Fraction of total dose of total dose Radionuclide Fraction of total dose of total dose

1 Np-239 0.29 0.29 Sr-89 0.32 0.32 Cs-137 0.33 0.33 2 Te-132 0.07 0.37 Mo-99 0.21 0.54 Te-132 0.21 0.54 3 Zr-97 0.07 0.44 Te-132 0.18 0.71 I-131 0.10 0.64 4 Ba-140 0.07 0.51 Cs-137 0.11 0.82 Mo-99 0.09 0.73 5 U-237 0.06 0.57 I-131 0.08 0.90 Ba-140 0.09 0.82 6 La-140 0.06 0.64 U-237 0.03 0.93 Sr-89 0.05 0.86

7 Ce-143 0.04 0.67 I-133 0.03 0.97 Ru-106 0.03 0.90 test nuclear Trinity from doses Estimated 8 Y-92 0.04 0.71 Ba-140 0.01 0.98 La-140 0.03 0.92 9 Mo-99 0.03 0.74 Ru-106 0.01 0.98 I-133 0.02 0.94 10 Y-93 0.03 0.77 Sr-90 0.00 0.99 Ru-103/Rh-103m 0.01 0.96 11 Ru-106 0.03 0.80 La-140 0.00 0.99 Zr-95 0.01 0.97 www.health-physics.com 12 Ce-144/Pr-144 0.02 0.83 Ru-103/Rh-103m 0.00 0.99 Np-239 0.00 0.97 13 Sr-89 0.02 0.84 Te-131m 0.00 0.99 Te-131m 0.00 0.98 14 I-133 0.02 0.86 Np-239 0.00 1.00 U-237 0.00 0.98 15 Nd-147 0.01 0.87 Zr-95 0.00 1.00 Sb-127 0.00 0.99 16 Ru-103/Rh-103m 0.01 0.89 Sb-127 0.00 1.00 Zr-97 0.00 0.99 17 Sr-91 0.01 0.90 I-135 0.00 1.00 Rh-105 0.00 0.99

18 Pm-149 0.01 0.91 Zr-97 0.00 1.00 Nb-95 0.00 0.99 c 19 I-131 0.01 0.92 Sr-91 0.00 1.00 Sr-91 0.00 0.99 S L. S. 20 Rh-105 0.01 0.93 Nd-147 0.00 1.00 Ce-143 0.00 0.99 MNE AL ET IMON Inhalation + Resuspension Pct. 581, TOA = 3.1 h, Rural/Mountains Pct. 13, TOA = 10.5 h, Urban/Plains Pct. 75, TOA = 36.3 h, Rural/Mountains

Data set B (see Bouville et al. 2020) Data Set D (see Bouville et al. 2020) Data set B (see Bouville et al. 2020) . Cumulative fraction Cumulative fraction Cumulative fraction Rank Radionuclide Fraction of total dose of total dose Radionuclide Fraction of total dose of total dose Radionuclide Fraction of total dose of total dose

1 Np-239 0.39 0.39 Np-239 0.35 0.35 Np-239 0.39 0.39 2 U-237 0.08 0.47 Ba-140 0.09 0.44 Ba-140 0.11 0.49 3 Zr-97 0.05 0.52 Te-132 0.07 0.51 U-237 0.08 0.57 4 Ba-140 0.05 0.57 U-237 0.07 0.57 Te-132 0.07 0.64 5 Ce-144/Pr-144 0.04 0.62 Ru-106 0.06 0.63 Ru-106 0.07 0.71 6 Ce-143 0.04 0.65 Zr-97 0.04 0.67 Ce-144/Pr-144 0.04 0.74 7 Mo-99 0.03 0.69 Ce-144/Pr-144 0.03 0.70 Mo-99 0.03 0.77 8 Ru-106 0.03 0.71 Ce-143 0.03 0.73 Ru-103/Rh-103m 0.03 0.80 9 Y-93 0.03 0.74 Mo-99 0.03 0.76 Rh-105 0.03 0.83 443 Continued next page 444 Health Physics October 2020, Volume 119, Number 4 high ranks of those two counties (Fig. 5) for Whites and Hispanics. According to the census, few Native Americans or African Americans, however, lived in high-fallout depo- sition precincts in Lincoln county, resulting in the very

of total dose low position of Lincoln county for those ethnic groups. Or-

Cumulative fraction gans other than thyroid would have received even lower doses. In contrast, the counties of Otero and Sierra received very little fallout deposition and almost no radiation expo- sure at all, giving them near zero estimated doses. One further point about the location-specific dose esti- mates is important. It has been reported that there were a few dozens of ranches and farms within 64.4 km of the Trinity detonation site (LAHDRA 2009). In this analysis, we have Data set B (see Bouville et al. 2020) e Fraction of total dose

Pct. 75, TOA = 36.3 h, Rural/Mountains not attempted to estimate doses received by persons living at specific ranch locations because (1) average doses were estimated for all precincts, so in theory, their doses have been estimated—at least approximately; (2) those people were presumably included in the census and, therefore, in- cluded in the risk projection; and finally, (3) any misclassi- fication of dose for the persons living at these ranches and farms (because of their contamination) would not apprecia-

of total dose Radionuclid bly affect the risk projection for New Mexico because of the

Cumulative fraction few numbers of people residing at each ranch or farm.

ouville et al. 2020) Validation Validation in dose reconstruction is the process of using measurements of radiation dose, or measurements

Inhalation + Resuspension of quantities as close to dose as possible, to confirm model-based estimates of dose. Internal dose cannot be di- Pct. 13, TOA = 10.5 h, Urban/Plains Data Set D (see B rectly measured and, while it can be derived from bioassay measurements today, no such measurements were known to have been conducted among the public following Trinity. Measurements to validate external doses are less fraught with technical problems and thus are useful for purposes of validation or confirmation. In the case of the Trinity test in 1945, in addition to the post-shot monitoring data used by Quinn et al. (1987), there were film-badge data collected by of total dose Radionuclide Fraction of total dose the Los Alamos Scientific Laboratory (LASL) that can be Cumulative fraction used to examine the validity of our estimated external doses. That data is also useful to examine the validity of the pub- lished geographic footprint of the Quinn (1987)-based fall- out pattern. Hoffman (1945), in a summary of radiation monitor’s field notes, provided data on environmental exposure in Roentgens (R) derived from blackening of x-ray film-

Data set B (see Bouville et al. 2020) badges. The badges had been sent to numerous towns and Pct. 581, TOA = 3.1 h, Rural/Mountains communities across New Mexico before 16 July 1945 to be returned in the days afterward. Exposure was determined

(Continued) by densitometer readings of films and using calibration films exposed to a known radiation source. Little information is

1011 Te-13212 Sr-9213 La-14114 Sr-9115 Zr-9516 Ru-105 0.0317 Y-9218 0.03 Pr-145 0.0319 Nd-14720 La-142 0.02 Ru-103/Rh-103m 0.02 0.02 0.77 0.79 0.02 0.82 0.01 0.01 Ru-103/Rh-103m 0.01 0.83available 0.01 Rh-105 0.85 0.87 Sr-89 0.02 0.88 Sr-91 on 0.90 0.93 Ru-105 0.91 Y-93 how 0.92 Zr-95 0.02 Nd-147 the Ce-141 0.78 Y-92 0.02 film-badges La-141 0.02 0.02 0.02 Ce-143 0.81 0.01 0.01 0.83 were 0.01 0.01 0.85 0.01 instructed Sr-89 0.88 0.87 Zr-97 0.02 Ce-141 0.90 0.91 Nd-147 0.94 0.92 Zr-95 to 0.93 be Sb-127 Pr-143 0.02 Pm-149 La-140 0.85 0.02 Sr-91 0.02 0.01 0.01 0.01 0.88 0.01 0.01 0.00 0.90 0.92 0.01 0.94 0.93 0.96 0.95 0.97 0.98 0.97 Rank Radionuclide Fraction of total dose

Table 3b. deployed, though presumably, they were hung outdoors in www.health-physics.com Table 3c. Radionuclides at three locations that contribute at least 90% of the first-year internal dose to active (red) bone marrow (Whites/Hispanics): ranking by their fractional contribution to dose. Ingestion Pct. 581, TOA = 3.1 h, Rural/Mountains Pct. 13, TOA = 10.5 h, Urban/Plains Pct. 75, TOA = 36.3 h, Rural/Mountains Data set B (see Bouville et al. 2020) Data Set D (see Bouville et al. 2020) Data set B (see Bouville et al. 2020) Cumulative fraction Cumulative fraction Cumulative fraction Rank Radionuclide Fraction of total dose of total dose Radionuclide Fraction of total dose of total dose Radionuclide Fraction of total dose of total dose

1 Sr-89 0.23 0.23 Sr-89 0.78 0.78 Sr-89 0.29 0.29 2 Ba-140 0.21 0.44 Mo-99 0.05 0.83 Ba-140 0.24 0.53 3 Te-132 0.13 0.56 U-237 0.05 0.87 Te-132 0.13 0.66 4 Mo-99 0.08 0.64 Te-132 0.04 0.91 Cs-137 0.08 0.74 5 Np-239 0.05 0.69 Sr-90 0.04 0.95 Mo-99 0.05 0.79

6 La-140 0.05 0.75 Ba-140 0.01 0.97 La-140 0.05 0.84 test nuclear Trinity from doses Estimated 7 Cs-137 0.04 0.79 Cs-137 0.01 0.98 I-131 0.03 0.87 8 Zr-97 0.03 0.82 I-131 0.01 0.99 Ru-103/Rh-103m 0.02 0.89 9 Zr-95 0.02 0.84 I-133 0.00 0.99 Sr-90 0.02 0.91 10 I-131 0.02 0.86 La-140 0.00 0.99 Zr-95 0.02 0.92 www.health-physics.com 11 Sr-91 0.02 0.88 Np-239 0.00 0.99 Np-239 0.01 0.94 12 I-133 0.02 0.90 Ru-103/Rh-103m 0.00 1.00 U-237 0.01 0.95 13 U-237 0.02 0.91 Zr-95 0.00 1.00 Ru-106 0.01 0.96 14 Ru-103/Rh-103m 0.01 0.93 Ru-106 0.00 1.00 I-133 0.01 0.97 15 Sr-90 0.01 0.94 Te-131m 0.00 1.00 Sb-127 0.01 0.98 16 I-135 0.01 0.95 Sb-127 0.00 1.00 Zr-97 0.01 0.98

17 Ce-143 0.01 0.95 Nd-147 0.00 1.00 Te-131m 0.00 0.99 c

18 Te-131m 0.01 0.96 Zr-97 0.00 1.00 Ce-143 0.00 0.99 S L. S. 19 Ru-106 0.01 0.97 Sr-91 0.00 1.00 Ce-144/Pr-144 0.00 0.99 MNE AL ET IMON 20 Sb-127 0.01 0.97 Ce-143 0.00 1.00 Sr-91 0.00 0.99

Inhalation + Resuspension

Pct. 581, TOA = 3.1 h, Rural/Mountains Pct. 13, TOA = 10.5 h, Urban/Plains Pct. 75, TOA = 36.3 h, Rural/Mountains . Data set B (see Bouville et al. 2020) Data Set D (see Bouville et al. 2020) Data set B (see Bouville et al. 2020) Cumulative fraction Cumulative fraction Cumulative fraction Rank Radionuclide Fraction of total dose of total dose Radionuclide Fraction of total dose of total dose Radionuclide Fraction of total dose of total dose

1 Np-239 0.38 0.38 Np-239 0.31 0.31 Np-239 0.36 0.36 2 Zr-97 0.09 0.46 Te-132 0.17 0.48 Te-132 0.21 0.57 3 Te-132 0.08 0.54 Ba-140 0.11 0.59 Ba-140 0.14 0.71 4 Ba-140 0.07 0.61 I-135 0.06 0.64 Zr-95 0.04 0.75 5 Zr-95 0.05 0.66 Zr-97 0.06 0.70 Ru-103/Rh-103m 0.04 0.79 6 I-135 0.05 0.71 I-133 0.05 0.75 Zr-97 0.03 0.82 7 La-142 0.04 0.75 Zr-95 0.03 0.79 I-133 0.03 0.85 8 Sr-92 0.03 0.79 Sr-91 0.03 0.82 La-140 0.02 0.86 445 Continued next page 446 Health Physics October 2020, Volume 119, Number 4 reasonably open locations. Badges were returned to LASL by mail in the period from 17–23 July. More than 118 badges were deployed widely across the state. The exact number is difficult to determine because some were reported as

of total dose lost. In this analysis, we used data on 118.

Cumulative fraction Hoffman (1945) reports that approximately 82% (n = 97) of the 118 badges gave readings of “background” dose. Though “background” was not well defined, it can be presumed that those badges gave no evidence of exposure from Trinity fallout. The precise background value was not reported, though it can be assumed to be less than 0.1 R8 over the time-period when the badges were deployed, as some locations reported measurements of 0.1 R. About Data set B (see Bouville et al. 2020) e Fraction of total dose

Pct. 75, TOA = 36.3 h, Rural/Mountains 8% (n = 9) were very low, just above background; i.e., 0.10 to 0.13 R. Another 7% (n = 8) were also quite low, i.e., within 3 times background or 0.23–0.34 R. A single badge (sent to Pedernal, a present-day uninhabited town in Torrance County) gave a reading within 7 times background (0.68 R), i.e., a medium exposure level. Finally, 3% (n = 3) of the badges deployed directly northeast of the detonation site had readings significantly greater than all others, from

of total dose Radionuclid 3.3 to 8.2 R. A review of the estimated exposures from

Cumulative fraction our dose estimation calculations indicated that about 5.8% of the 721 precincts across the state received an accumulated exposure in the six days following Trinity of 3.3 R or greater, similar in magnitude to the 3% fraction of badges that reported exposures greater than 3.3 R. The geographic locations of the film-badges, colored

Inhalation + Resuspension by their approximate exposure, are presented in Fig. 6 along with the H+12 exposure isopleths derived from Quinn Pct. 13, TOA = 10.5 h, Urban/Plains Data Set D (see Bouville et al. 2020) (1987). The large number of “background” measurements, widely distributed throughout the state, confirms that only very low exposures were likely received by people resident outside of the presumed fallout pattern. Moreover, there were no readings of significance south of the detonation site, and the only high readings were in the very center of the isopleths where the H+12 exposure rate was thought to be several thousand times greater than the exposure rates on of total dose Radionuclide Fraction of total dose the periphery of the pattern. These findings provide a moder- Cumulative fraction ately high degree of confidence in the fallout pattern used as the basis for this dose reconstruction, and therefore, we con- clude that the pattern boundaries appear quite reasonable.

Uncertainty Clearly the estimation of exposures received more than 70 y ago is fraught with uncertainties. In this work, contem-

Data set B (see Bouville et al. 2020) porary interviews on diet and lifestyle data, benchmarked Pct. 581, TOA = 3.1 h, Rural/Mountains against historical reports and compendia on nutrition and di- etary habits, allowed estimates of doses to be made that are

(Continued) the best possible today. Variability of dietary data collected from the small groups interviewed, while recorded and 9 Sr-91 0.03 0.81 Ru-103/Rh-103m 0.03 0.85 I-131 0.01 0.88 101112 I-13313 Ru-103/Rh-103m Ru-1051415 Mo-9916 0.02 U-23717 0.02 Ce-143 0.0218 Sb-129 Ce-144/Pr-14419 0.0220 Te-131m 0.02 0.88 0.01 Y-92 0.01 0.84 0.86 I-131 0.01 0.01 0.89 Mo-99 0.91 0.01 Ru-105 Te-131m 0.92 0.94 0.00 0.93 U-237 0.95 0.01 I-131 0.02 0.01 Ce-143 Ru-106 0.95 Sb-129 0.96 0.01 I-132 0.89 0.01 0.01 0.01 Ce-144/Pr-144 Rh-105 0.87 0.88 0.01 Te-131m 0.01 0.90 0.01 Mo-99 U-237 0.01 0.92 0.93 0.94 0.93 Ru-106 0.01 0.95 Rh-105 Ce-143 I-132 0.96 0.01 0.01 0.96 Sr-91 Ce-144/Pr-144 Sb-127 0.01 0.92 I-135 0.01 0.01 0.01 0.91 0.89 0.01 0.01 0.93 0.00 0.01 0.96 0.95 0.94 0.97 0.96 0.98 0.98 Rank Radionuclide Fraction of total dose 8 ≅ Table 3c. 0.1 R = 100 mR 87.7 mrad to air = 0.88 mGy to air.

www.health-physics.com Table 3d. Radionuclides at three locations that contribute at least 90% of the first-year internal dose to stomach (Whites/Hispanics): ranking by their fractional contribution to dose. Ingestion Pct. 581, TOA = 3.1 h, Rural/Mountains Pct. 13, TOA = 10.5 h, Urban/Plains Pct. 75, TOA = 36.3 h, Rural/Mountains Data set B (see Bouville et al. 2020) Data Set D (see Bouville et al. 2020) Data set B (see Bouville et al. 2020) Cumulative fraction Cumulative fraction Cumulative fraction Rank Radionuclide Fraction of total dose of total dose Radionuclide Fraction of total dose of total dose Radionuclide Fraction of total dose of total dose

1 Y-92 0.19 0.19 U-237 0.42 0.42 La-140 0.12 0.12 2 Np-239 0.17 0.36 Sr-89 0.18 0.59 Np-239 0.12 0.24 3 Zr-97 0.06 0.42 Mo-99 0.11 0.71 Te-132 0.11 0.35 4 La-140 0.05 0.47 Te-132 0.08 0.79 Ba-140 0.07 0.42 5 Y-93 0.05 0.52 I-133 0.07 0.85 U-237 0.06 0.49 6 I-133 0.05 0.57 I-131 0.04 0.89 Mo-99 0.06 0.55

7 Te-132 0.04 0.61 Np-239 0.03 0.92 I-131 0.06 0.61 test nuclear Trinity from doses Estimated 8 Mo-99 0.04 0.65 Cs-137 0.01 0.93 Cs-137 0.05 0.66 9 U-237 0.03 0.68 La-140 0.01 0.94 I-133 0.05 0.71 10 La-141 0.03 0.71 Ba-140 0.01 0.95 Ru-106 0.05 0.76 11 I-135 0.03 0.74 Ru-106 0.01 0.96 Zr-97 0.03 0.79 www.health-physics.com 12 Sr-91 0.03 0.76 Y-92 0.01 0.96 Sr-89 0.03 0.82 13 Ba-140 0.03 0.79 Nd-147 0.01 0.97 Ce-143 0.03 0.85 14 Ce-143 0.03 0.81 Zr-97 0.00 0.97 Ru-103/Rh-103m 0.03 0.87 15 Pr-145 0.02 0.83 Ru-103/Rh-103m 0.00 0.98 Ce-144/Pr-144 0.01 0.89 16 I-131 0.02 0.85 Ce-143 0.00 0.98 Rh-105 0.01 0.90 17 Ru-105 0.02 0.86 Y-93 0.00 0.98 Pr-143 0.01 0.91

18 Ru-106 0.01 0.88 Pm-149 0.00 0.98 Nd-147 0.01 0.92 c 19 Cs-137 0.01 0.89 Rh-105 0.00 0.98 Y-93 0.01 0.93 S L. S. 20 Sr-89 0.01 0.90 Pr-143 0.00 0.99 Pm-149 0.01 0.94 MNE AL ET IMON Inhalation + Resuspension Pct. 581, TOA = 3.1 h, Rural/Mountains Pct. 13, TOA = 10.5 h, Urban/Plains Pct. 75, TOA = 36.3 h, Rural/Mountains

Data set B (see Bouville et al. 2020) Data Set D (see Bouville et al. 2020) Data set B (see Bouville et al. 2020) . Cumulative fraction Cumulative fraction Cumulative fraction Rank Radionuclide Fraction of total dose of total dose Radionuclide Fraction of total dose of total dose Radionuclide Fraction of total dose of total dose

1 La-142 0.12 0.12 Np-239 0.12 0.12 Np-239 0.28 0.28 2 La-141 0.11 0.23 Y-92 0.10 0.22 Te-132 0.12 0.39 3 Y-92 0.11 0.34 Zr-97 0.08 0.30 Zr-97 0.09 0.48 4 Np-239 0.09 0.42 Y-93 0.08 0.38 Ba-140 0.08 0.56 5 Y-93 0.08 0.50 La-141 0.07 0.45 I-133 0.05 0.60 6 Zr-97 0.07 0.57 Sr-91 0.07 0.52 Y-93 0.04 0.64 7 Sr-92 0.07 0.64 Ru-105 0.06 0.57 Mo-99 0.04 0.68 8 Pr-145 0.05 0.69 Te-132 0.05 0.62 Ce-143 0.04 0.72 9 Ba-139 0.04 0.73 I-135 0.05 0.67 Sr-91 0.03 0.75 447 Continued next page 448 Health Physics October 2020, Volume 119, Number 4 considered, does not in itself adequately capture the uncer- tainty on the mean data value, the reasons being the limita- tionsofthesampleofpersonsintermsofnumberofpersons and quality of memory recall. Hence, an uncertainty analysis

of total dose based only on statistical distributions of variability was not

Cumulative fraction reasonable. In this work, we conducted an uncertainty analy- sis using Monte Carlo methods using data-supported but judgement-based probability distributions that quantify our degree-of-belief in the mean values of the parameters used. Using the aforementioned strategy, we derived uncer- tainty factors that can be either applied to county-specific best estimate doses (i.e., those provided in the Appendix) or, as in this study, used to propagate uncertainty into the Data set B (see Bouville et al. 2020) e Fraction of total dose Pct. 75, TOA = 36.3 h, Rural/Mountains risk projection (Cahoon et al. 2020). Under conventional uncertainty analysis paradigms, variances or geometric standard deviations (GSDs) of log-normally distributed in- put parameters for dose models are used to derive “uncer- tainty distributions” that could be used to derive statistical confidence levels. In this work, however, we use the term “uncertainty factors” as we want to make the important dis- tinction that these estimates are not parameters of precise statistical distributions because of the large degree of sub- of total dose Radionuclid jectivity involved in their derivation. Cumulative fraction Because the total number of dose calculations we made for the 721 precincts, 63 radionuclides, 13 exposure path-

ouville et al. 2020) ways, 6 data sets, 7 age groups, and 5 organs was large (~124 million), we determined it was not feasible to conduct Monte Carlo simulations for every combination of parame- Inhalation + Resuspension ters. For that reason, we conducted the uncertainty analysis with a simplified strategy that we believe represented an ad- Pct. 13, TOA = 10.5 h, Urban/Plains Data Set D (see B equate cross-section of the combinations of exposure condi- tion such that the findings could be generalized. TOA, which is an important parameter in the calcula- tion of deposition densities of fallout radionuclides, varied acrossNewMexicofromabout1htoabout40h.Because of its importance, we chose precincts at three locations in the Quinn pattern (L1, L2, and L3) that we determined were representative of ranges of TOAs (L1 for close-in locations

of total dose Radionuclide Fraction of total dose with a TOA of about 3 h or less; L2 for mid-distance

Cumulative fraction locations with a TOA from 3 h to about 10.5 h, and L3 for a far-field locations with a TOA of about from 10.5 to about 36 h), plus two locations outside the Quinn fallout pattern − where we had estimated the X12̇ðÞto be 0.05 mR h 1 (L4 − was for TOA of 10–25 h and X12̇ðÞ=0.05mRh1 and L5 was for TOA of 25–40 h). At the locations chosen for simulation, we modeled the

Data set B (see Bouville et al. 2020) uncertainty of the external dose relative to the central best Pct. 581, TOA = 3.1 h, Rural/Mountains estimate by assigning probability density functions (PDFs) presented in Bouville et al. (2020) for the external dose

(Continued) model parameters normalized to the best estimate. Similarly, we modeled the uncertainty of the internal dose by assigning

101112 Ru-10513 Sr-9114 I-13515 Nd-14916 Ce-143 0.0417 Sb-129 0.0318 Rb-88 0.0219 Te-132 0.0220 0.02 Mo-99 0.01 I-133 0.77 Ba-140 0.01 0.80 0.01 0.82 0.01 0.84PDFs 0.85 I-133 0.01 0.01 0.87 Pr-145 (also 0.88 Ba-140 0.89 Sb-129from Ce-143 0.91 0.04 Sr-92 0.92 0.03 Bouville 0.93 Mo-99 0.03 Rh-105 0.02 0.02 Rb-88 et 0.02 Te-129al. 0.72 I-132 0.02 2020) 0.75 0.01 0.78 0.01 0.80 to 0.82 0.01 those Rh-105 0.01 0.84 I-132 0.86 radionuclides Ru-103/ Rh-103m 0.87 U-237 La-140 0.88 0.03 0.89 Zr-95 0.02 0.90 I-135 0.03 Te-131m 0.02 Ru-106 0.02 Ag-112 Pm-149 0.02 0.78 0.01 0.01 0.85 0.80 0.83 0.01 0.01 0.87 0.01 0.88 0.90 0.89 0.91 0.92 0.93 Rank Radionuclide Fraction of total dose

Table 3d. that accounted for at least 80% of the internal dose www.health-physics.com Table 3e. Radionuclides at three locations that contribute at least 90% of the first-year internal dose to thyroid (Whites/Hispanics): ranking by their fractional contribution to dose. Ingestion Pct. 581, TOA = 3.1 h, Rural/Mountains Pct. 13, TOA = 10.5 h, Urban/Plains Pct. 75, TOA = 36.3 h, Rural/Mountains Data set B (see Bouville et al. 2020) Data Set D (see Bouville et al. 2020) Data set B (see Bouville et al. 2020) Cumulative fraction Cumulative fraction Cumulative fraction Rank Radionuclide Fraction of total dose of total dose Radionuclide Fraction of total dose of total dose Radionuclide Fraction of total dose of total dose

1 I-131 0.71 0.71 I-131 0.80 0.80 I-131 0.86 0.86 2 I-133 0.20 0.90 I-133 0.14 0.94 I-133 0.08 0.93 3 Te-132 0.07 0.97 Te-132 0.06 1.00 Te-132 0.06 1.00 4 I-135 0.02 0.99 Te-131m 0.00 1.00 Te-131m 0.00 1.00 5 Te-131m 0.00 1.00 Sr-89 0.00 1.00 I-135 0.00 1.00 6 Cs-137 0.00 1.00 I-135 0.00 1.00 Cs-137 0.00 1.00 7 Te1-33m 0.00 1.00 Mo-99 0.00 1.00 Mo-99 0.00 1.00 siae oe rmTiiyncertest nuclear Trinity from doses Estimated 8 Mo-99 0.00 1.00 Cs-137 0.00 1.00 Ba-140 0.00 1.00 9 Ba-140 0.00 1.00 U-237 0.00 1.00 Sr-89 0.00 1.00 10 Tc-99m 0.00 1.00 Ba-140 0.00 1.00 Ru-106 0.00 1.00 11 I-132 0.00 1.00 Ru-106 0.00 1.00 I-132 0.00 1.00

www.health-physics.com 12 Sr-89 0.00 1.00 Sr-90 0.00 1.00 Tc-99m 0.00 1.00 13 Ru-106 0.00 1.00 I-132 0.00 1.00 Ru-103/Rh-103m 0.00 1.00 14 Sr-91 0.00 1.00 Ru-103/Rh-103m 0.00 1.00 Zr-95 0.00 1.00 15 Zr-95 0.00 1.00 Tc-99m 0.00 1.00 La-140 0.00 1.00 16 Ru-103/Rh-103m 0.00 1.00 Zr-95 0.00 1.00 Sb-127 0.00 1.00 17 La-140 0.00 1.00 Sb-127 0.00 1.00 U-237 0.00 1.00

18 Zr-97 0.00 1.00 La-140 0.00 1.00 Rh-105 0.00 1.00 c

19 Sb-127 0.00 1.00 Sr-91 0.00 1.00 Sr-91 0.00 1.00 S L. S. 20 Sr-92 0.00 1.00 Rh-105 0.00 1.00 Sr-90 0.00 1.00 MNE AL ET IMON Inhalation + Resuspension Pct. 581, TOA = 3.1 h, Rural/Mountains Pct. 13, TOA = 10.5 h, Urban/Plains Pct. 75, TOA = 36.3 h, Rural/Mountains

Data set B (see Bouville et al. 2020) Data Set D (see Bouville et al. 2020) Data set B (see Bouville et al. 2020) . Cumulative fraction Cumulative fraction Cumulative fraction of total Rank Radionuclide Fraction of total dose of total dose Radionuclide Fraction of total dose of total dose Radionuclide Fraction of total dose dose

1 I-133 0.46 0.46 I-133 0.50 0.50 I-131 0.55 0.55 2 I-131 0.26 0.72 I-131 0.32 0.82 I-133 0.37 0.92 3 I-135 0.22 0.94 I-135 0.14 0.96 Te-132 0.04 0.96 4 Te-133m 0.02 0.96 Te-132 0.03 0.99 I-135 0.02 0.98 5 Te-132 0.02 0.99 I-132 0.01 0.99 I-132 0.01 0.99 6 I-132 0.01 0.99 Te-131m 0.00 1.00 Te-131m 0.00 1.00 7 Te-131m 0.00 1.00 Ba-140 0.00 1.00 Ba-140 0.00 1.00 8 Np-239 0.00 1.00 Np-239 0.00 1.00 Np-239 0.00 1.00 9 Ba-140 0.00 1.00 Zr-97 0.00 1.00 Zr-95 0.00 1.00

Continued next page 449 450 Health Physics October 2020, Volume 119, Number 4 (sometimes as much as 96%). Similar to the calculations for external dose, we modeled the uncertainty of the internal dose by assigning PDFs to the internal dose model parameters normalized to the best estimate. As described

dose here, we conducted Monte Carlo simulations of doses for five pathways: external dose, ingestion dose from milk, ingestion dose from leafy vegetables, and doses from

Cumulative fraction of total inhalation and resuspension. This analysis was limited to adults and extrapolated to other ages. Our initial simulations confirmed that because the in- halation and resuspension doses are so small compared to the external and ingestion doses, their uncertainty contrib- utes very little to the overall dose uncertainty. For that rea-

Data set B (see Bouville et al. 2020) son, those pathways were not simulated further. The Pct. 75, TOA = 36.3 h, Rural/Mountains overall uncertainty was determined primarily by the most important contributors to the organ dose: external irradia- tion and ingestion of milk and leafy vegetables. Each Monte Carlo simulation for external and internal dose, run for 50,000 iterations (separately), produced a data set for each organ and location that closely fit a log-normal distribution. From each simulated dose data set, we derived a geometric standard deviation (GSD) by fitting a log-

of total dose Radionuclide Fraction of total dose normal type distribution and calculating the GSD from the

Cumulative fraction median, mean, and variance using standard statistical for- mulae. From the external and internal dose GSD values, the GSD of the total dose (external + internal) distribution was derived and the square of the GSD was assigned the term of “uncertainty factor.” The lower credible dose was Inhalation + Resuspension found by the best estimate of dose at each county divided by the uncertainty factor (i.e., best estimate of dose/GSD2), Pct. 13, TOA = 10.5 h, Urban/Plains

Data Set D (see Bouville et al. 2020) while the upper end credible dose was found by the best estimate of dose multiplied by the uncertainty factor (best estimate of dose GSD2). Based on well-known statistical properties for log-normal distributions, the range from the lower credible dose to the upper credible dose would en- compass 95% of the simulated values (i.e., from the 2.5% to 97.5%). Table 4 presents the uncertainty factors derived for ex-

of total dose Radionuclide Fraction of total dose ternal doses and internal doses at the precinct locations that

Cumulative fraction we generalized to other locations with similar TOAs. That is, the uncertainty factors were assigned to all locations with the same attributes of TOA as in the simulations: (1) TOA of 1–10 h, (2) TOA of 10–25 h, (3) TOA of 25–40 h, (4), TOA − of 10–25 h andẊðÞ12 =0.05mRh 1, and (5) TOA of 25–40 − handX12̇ðÞ=0.05mRh 1. The derived uncertainty factors, presented in Table 4, are in the range from 2.5 to 3.0 for all external dose esti- Data set B (see Bouville et al. 2020)

Pct. 581, TOA = 3.1 h, Rural/Mountains mates at all precinct locations. Uncertainty factors for inter- nal dose varied by organ and with TOA, with the largest

(Continued) uncertainty factors being for lung and thyroid at close-in lo- cations and for stomach and thyroid at distant locations (see Table 4 for all values). As an example, the uncertainty fac- 101112 Zr-9713 Zr-9514 La-14215 Sr-9216 Sr-91 0.0017 Ru-105 0.0018 0.00 U-23719 Mo-99 0.0020 Ce-143 0.00 0.00 Sb-129 1.00 Ru-106 0.00 1.00 0.00 1.00 0.00 1.00 0.00 Zr-95 1.00 0.00 1.00 Sr-91 Ru-105 1.00 Te-133m 1.00 1.00 Ru-106 0.00 Mo-99 1.00 0.00 0.00 1.00 U-237 0.00 Sb-129 Ce-143 0.00 Rh-105 0.00 1.00 Sr-92 0.00 1.00 0.00 1.00 0.00 1.00 0.00 1.00 Zr-97 1.00 0.00 La-140 Ru-106 1.00 U-237 1.00 1.00 Mo-99 Ce-143 0.00 1.00 0.00 0.00 Sr-91 1.00 Rh-105 0.00 Cs-137 0.00 Sb-127 0.00 Ce-141 0.00 0.00 1.00 0.00 1.00 1.00 0.00 1.00 0.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 Rank Radionuclide Fraction of total dose

Table 3e. tors for thyroid dose were estimated to be the largest www.health-physics.com Estimated doses from Trinity nuclear test c S. L. SIMON ET AL. 451

Fig. 5. Ordering (left to right) of counties by magnitude of population-weighted average doses to the thryoid gland of adults. Top panels (left to right): Whites, Hispanic. Bottom panels (left to right): Native Americans, African Americans. For discussion on Native Amerian doses in Torrance County, see section on “Comparison of county average doses.” Four counties with asterisks (*) are those identified as “high risk” (and presumably, “high dose” counties) by TBDC (2017).

(~12.5) for the precincts closest to the detonation site (TOA Context on the magnitude of estimated doses of <10 h), the smallest (~8.3) for precincts at TOA from 10 While the findings of cancer risks in the analysis of to 25 h, and intermediate (~10.6) for precincts with Cahoon et al. (2020) are the most important metric of health TOA>25 h. Uncertainties well outside the pattern were sim- impact from this study, it is useful to have an understanding ilar to the estimated values inside the pattern because the of the magnitude of exposures received in New Mexico uncertainty was dominated by factors other than the expo- from Trinity and other sources of nuclear testing fallout as sure rate and TOA. The uncertainty factors for the com- well as from natural sources. For this analysis, we provide bined external and internal doses varied depending on the individual county estimates of air kerma (integral over one magnitude of the external and internal doses relative to one an- year) and 137Cs deposition density (Bq m−2) from Trinity, other. Fig. 7 illustrates cumulative distributions of lower Nevada Test Site (NTS), and global fallout. Air kerma is credible doses, best estimates of dose, and upper credible used to normalize the comparison without the introduction doses for adults at each of the 721 precincts, derived by of building shielding and age-dependent dose factors. The the strategy described. findings are presented in Table 5 (including NTS and The modestly larger uncertainty for the thyroid gland global fallout data derived from US DHHS 2005). compared to other organs is a mathematical outcome of As can be seen in Table 5, the 1-y integral air kerma two conditions. First, the milk pathway model has a larger (the primary determinant of external dose to man) from number of uncertain parameters, e.g., feed-to-milk trans- Trinity was very heterogeneous across New Mexico, with fer coefficients (Bouville et al. 2020). Second, the thyroid county average values differing by almost 1,300-fold, from dose is nearly 100% contributed by only two or three ra- 0.018 to 23 mGy with a coefficient of variation (CV = dionuclides (see Table 3e), whereas doses to other organs standard deviation/mean) of 2.8, implying that the distribution have significant partial dose contributions by up to 20 ra- was highly positively skewed with a standard deviation dionuclides. In the Monte Carlo simulation, the uncer- equal to 2.8 times the mean value. In contrast, the CV was tainty distributions for each radionuclide must be summed much smaller for NTS fallout, about 0.5, and even more in proportion to their contribution to the total dose. A larger homogeneous for global weapons testing fallout with a CV number of components in the sum results in a narrower dis- of 0.3. The magnitudes of the average value of air kerma tribution of the sum distribution. among the counties for Trinity, NTS, and global fallout were

www.health-physics.com 452 Health Physics October 2020, Volume 119, Number 4 Table 4. Derived uncertainty factors (GSD2) for external and internal CVs for Trinity, NTS, and global fallout were 2.7, 0.52, dose for five representative locations (L1 = TOA of 3.1 h, L2 = TOA and 0.30, respectively. of 10.5 h, L3 = TOA of 36.3 h, L4 = TOA>10 h, and ẊðÞ12 of 0.05 (i.e., “outside the fallout pattern”) and L5 = TOA of 25–40 h The general finding from this analysis is that the and X12̇ðÞof 0.05 (i.e., “outside the fallout pattern”). statewide-average air kerma over the 31 counties from Trin- External Dose ity, NTS, and global fallout were similar (1.6, 2.5, 1.1 mGy, respectively), though because the heterogeneity was much Uncertainty factors Location TOA and ẊðÞ12 from simulation greater for Trinity, there were counties with much lower and higher values than the mean value, a situation which L1 TOA=1–10 h 3.0 did not occur to any significant extent for NTS and global L2 & L3 TOA>10 h 2.5 fallout. The statewide median values of air kerma for NTS L4 & L5 (outside X12̇ðÞ=0.05 and fallout pattern) TOA>10 h 2.8 and global fallout were 10 to 20 times greater than the me- 137 Internal Dose dian for Trinity. While the heterogeneity of Cs deposition COLON density estimates, as assessed by CVs, was similar to that of L1 TOA=1–10 h 8.9 air kerma, the relationships between mean values as well as L2 TOA=10–25 h 4.8 median values for Trinity, NTS, and global fallout were dif- L3 TOA=25–40 h 6.9 ferent than for air kerma (see Table 5). For example, the L4 TOA=10–25 h 4.8 statewide mean 137Cs deposition density for NTS and L5 TOA=25–40 h 6.9 global fallout were about 3-fold and 25-fold greater, respec- LUNG tively, than the statewide mean value for Trinity fallout, and – L1 TOA=1 10 h 10.7 the statewide median 137Cs deposition density for NTS and – L2 TOA=10 25 h 3.3 global fallout was about 28-fold and 244-fold, respectively, L3 TOA=25–40 h 4.7 – greater than the statewide median value for Trinity fallout. L4 TOA=10 25 h 3.3 137 – The different relationships for Cs deposition density L5 TOA=25 40 h 4.7 137 RBM and air kerma are not surprising given that Cs does not L1 TOA=1–10 h 8.9 provide a large contribution to the integral air kerma, partic- L2 TOA=10–25 h 5.9 ularly in the first few months when air kerma is greatest. L3 TOA=25–40 h 7.2 In summary, Trinity, as might be expected, resulted in a L4 TOA=10–25 h 5.9 very heterogeneous deposition pattern across New Mexico L5 TOA=25–40 h 7.2 where, in some locations, the air kerma exceeded the STOMACH maximum values from NTS and global fallout even though it L1 TOA=1–10 h 9.6 was similar on a statewide average basis. In contrast, global L2 TOA=10–25 h 4.8 fallout deposition of 137Cs was much more homogeneous L3 TOA=25–40 h 7.8 across the state and was much greater at most individual L4 TOA=10–25 h 4.8 locations, as well as for a statewide average, than from L5 TOA=25–40 h 7.8 either NTS or Trinity fallout. THYROID It is also interesting to compare the thyroid doses – L1 TOA=1 10 h 12.5 resulting from the three sources of fallout, because the doses L2 TOA=10–25 h 8.3 to the thyroid are greater than the doses to any other organ. L3 TOA=25–40 h 10.6 For that comparison, it is important to realize that the birth L4 TOA=10–25 h 8.3 L5 TOA=25–40 h 10.6 cohorts exposed to Trinity fallout and the peak years of NTS and global fallout were not the same. Most residents of New Mexico who were in childhood at the time of Trinity quite similar with mean values over the counties of 1.6, 2.5, were adults at the time of NTS or global atmospheric and 1.1 mGy. However, because of the wide variation for weapons testing. Because of the strong age-dependency of Trinity fallout, the median of the county values for Trinity the doses to the thyroid, the comparison of the thyroid doses falloutwasonly0.071mGycomparedto2.7and1.1mGy is easiest for the persons who were exposed as adults at for NTS and global fallout. As expected from a visual the time of Trinity and of NTS and global testing. Taking inspection of the Trinity fallout pattern, dose to air (kerma) as examples three counties that were exposed to high from fallout was very heterogeneous, with a much smaller (Guadalupe), moderate (San Miguel), and small amounts median value and a greater maximum value than from (Bernalillo) of Trinity fallout, one can see that the estimated fallout from NTS or global fallout. The variations of 137Cs adult thyroid dose to the residents of Guadalupe county ground deposition density from Trinity, NTS, and global (about 50 mGy) was substantially greater than the thyroid fallout were similar to the variations of air kerma. The doses from NTS fallout (30 mGy) and from global fallout www.health-physics.com Estimated doses from Trinity nuclear test c S. L. SIMON ET AL. 453 from Trinity (2 mGy) was smaller than both the dose from NTS fallout (10 mGy) and the dose from global fallout (3 mGy). Taking the entire state of New Mexico into consideration, the largest component of the dose to the thyroid gland was from fallout from NTS tests, though for nine counties, thyroid doses were dominated by exposure to fallout from Trinity. While comparison of Trinity exposures with NTS and global fallout are useful, it is also informative to compare Trinity exposures with other sources of radiation. Here we restrict our comparison to natural background radiation be- cause both Trinity and background radiation can be as- sumed to be an involuntary exposure. New Mexico, with an average elevation of 5,700 ft, is the fourth highest state9 on average and hence has a higher cosmic-ray exposure rate than most other states. As well, it has significant mineral deposits, typical of the Rocky Mountains, which leads to increased natural radiation exposure from the terrestrial environment, including both gamma radiation and radon. According to Brookins (1992), the terrestrial background and cosmic ray components of Fig. 6. Results of exposure measurements from Los Alamos Scien- natural radiation received by Albuquerque residents (as a tific Laboratory (Hoffman 1945) using x-ray film badges deployed representation of New Mexico residents) are about 0.6 and in communities across the state before the Trinity detonation. 0.8 mSv per year, respectively, corresponding to about 1.4 mGy in terms of annual whole-body dose. (9 mGy). In San Miguel county, the thyroid dose from Trinity As a comparison, the statewide average air kerma for (10 mGy) was smaller than the dose from NTS fallout (30 Trinity was about 1.4 mGy, similar in magnitude to the mGy) and approximately equal to the dose from global gamma-ray component of natural background radiation, fallout (10 mGy). In Bernalillo county, the thyroid dose though it varied considerably by county from 0.018 to 23

Fig. 7. Cumulative distributions of lower credible doses, best estimates of dose, and upper credible doses for adults at each of the 721 precincts.

9http://www.netstate.com/states/geography/mapcom/nm_mapscom.htm.

www.health-physics.com 454 Health Physics October 2020, Volume 119, Number 4 Table 5. Comparison of estimated integral air kerma in 1945 (mGy) and 137Cs deposition density (Bq m−2) by counties in New Mexico from three sources of nuclear testing fallout (TRINITY, Nevada Test Site, and global fallout). NTS and global fallout data taken from US DHHS (2005). TRINITY values are population weighted values derived from precinct estimates. All values rounded to two significant digits. 137Cs 137Cs 137Cs Integral air kerma (mGy) Integral air kerma (mGy) Integral air kerma (mGy) (Bq m−2) (Bq m−2) (Bq m−2) Trinity NTS Global fallout Trinity NTS Global fallout County (1945) (1951–1963) (1953–2000) (1945) (1951–1963) (1953–2000)

Bernalillo 0.14 4.0 1.4 12 380 2,800 Catron 0.039 1.7 0.91 3.3 150 1,800 Chaves 0.12 2.9 0.95 12 270 1,900 Colfax 0.61 2.8 1.6 73 260 3,400 Curry 0.022 2.4 1.3 2.4 230 2,800 De Baca 0.25 2.5 0.99 23 240 2,000 Dona Ana 0.019 0.66 0.65 2.0 55 1,200 Eddy 0.019 1.2 0.95 2.0 110 1,899 Grant 0.019 0.76 1.1 2.0 63 2,100 Guadalupe 10 3.0 1.0 810 280 2,000 Harding 0.065 2.6 1.5 7.1 250 3,300 Hidalgo 0.019 0.68 0.84 2.0 56 1,600 Lea 0.019 1.1 1.1 2.0 100 2,200 Lincoln 7.9 2.8 1.1 170 270 2,200 Luna 0.019 0.71 0.74 2.0 59 1,300 McKinley 0.023 5.6 1.0 2.3 530 2,100 Mora 0.60 2.8 1.6 64 260 3,500 Otero 0.019 0.76 1.1 2.0 67 2,200 Quay 0.056 2.7 1.6 5.5 250 3,300 Rio Arriba 0.089 2.8 1.3 9.9 260 2,600 Roosevelt 0.022 2.3 1.1 2.4 220 2,400 Sandoval 0.090 4.9 1.2 8.6 460 2,500 San Juan 0.018 3.5 0.76 2.2 330 1,500 San Miguel 1.2 2.9 1.4 110 270 3,000 Santa Fe 0.37 4.1 1.2 36 380 2,500 Sierra 0.022 1.1 0.86 2.0 93 1,600 Socorro 5.2 2.0 0.74 72 190 1,500 Taos 0.23 2.5 1.3 26 240 2,600 Torrance 23 4.0 0.97 880 380 1,900 Union 0.071 2.7 1.6 8.7 250 3,500 Valencia 0.12 4.4 0.59 10 410 1,200

Minimum = 0.018 0.66 0.59 2.0 55 1,200 Maximum = 23 5.6 1.6 880 530 3,500 Mean = 1.6 2.5 1.1 76 240 2,300 Median = 0.071 2.7 1.1 8.6 250 2,200 Coefficient of Variation = 2.8 0.51 0.27 2.7 0.52 0.30 mGy. This comparison suggests that the whole-body dose to natural radiation exposure were the much greater internal Trinity fallout as an external source of radiation, on average doses to the thyroid because of the importance of radioiodine for the whole state (~1.4 mGy), was about equal to the an- in fallout, whereas radioiodine does not contribute to the nual whole-body dose from the external component of nat- internal dose from natural background. ural background radiation. It can also be viewed that Trinity fallout resulted in an incremental increase (28% to 47%) in Tabular values for archival purpose the average total dose (Trinity + natural background) of Appendix Tables 1 through 4 and their sub-tables pro- New Mexico residents alive at that time. The main vide numerical values of population-weighted best esti- differences for Trinity fallout, however, compared to mates of organ doses by ethnicity and age for use in other www.health-physics.com Estimated doses from Trinity nuclear test c S. L. SIMON ET AL. 455 studies and for archival purposes. For example, Appendix for external exposure and internal exposure resulting from Table A1a–e provides doses for White ethnicity to colon, consumption of 11 different food types including mothers’ lung, active (red) bone marrow, stomach, and thyroid, re- breast milk, drinking water, in-cloud inhalation, and resus- spectively. Tables A2a–e, A3a–e, and A4a–e provide the pension over the first year following Trinity. doses to the same organs for Native Americans, Hispanics, Radiation doses were found, expectedly, to differ sig- and African Americans, respectively. The weighted value nificantly by location, age, and organ, and to a lesser degree for each county considers the fraction of the county’stotal by ethnicity. Doses received from Trinity by external irradi- population in each of its precincts and the radiation doses ation were not large except in very limited areas immedi- estimated for each precinct. ately downwind of the detonation site where they ranged It is important to recognize that the deposition and the up to 100 mGy. Organ doses, except to the thyroid gland dose estimates provided in the Appendix for counties well and for a relatively small fraction of the public, were also outside the central part of the fallout pattern appear to be rel- not large. For the thyroid gland, young age groups, e.g., atively homogeneous across the county. Therefore, individ- 1–2 y of age, received the largest doses, though few would ual precincts in those counties would likely have doses quite be considered high compared to annual doses from natural close to the population weighted value. In contrast, counties radiation and even less so compared to lifetime natural radi- near the central part of the fallout pattern, e.g., Socorro, Tor- ation. About 20% of that age cohort might have received rance, Lincoln, and Guadalupe, likely had precincts that had doses greater than 10 mGy extending up to (for a very few doses substantially lower as well as greater than the weighted persons) a few hundred mGy. Other organs and age groups average values. would have been less. Uncertainties were evaluated and uncertainty factors CONCLUSION (as previously defined) ranged from 2.5 to 3.0 for external dose and for internal dose from 3.3 to 10.7 for the lung (de- For the first time, organ doses received by representa- pending on location) and 8.3 to 12.5 for thyroid (also de- tive persons in four ethnic groups, all age groups, and all pending on location). Our analysis suggests that the counties of New Mexico have been estimated as a result “credibility range” of dose estimated from the best parame- of exposure to radioactive fallout from the 1945 Trinity ter estimates should capture the true average dose in each nuclear test. A high degree of spatial heterogeneity of precinct and county. Moreover, this analysis suggests that dose was estimated for New Mexico due to characteristics doses to individuals at either the very low or high end seem of the published Trinity fallout pattern. That pattern was a relatively unlikely since the mass of the uncertainty distribu- result of a wind dispersion to the northeast and was based tions is small at the extremes. For the most part, White and primarily on actual field measurements of exposure rate Hispanic populations received the highest exposures, while across central New Mexico in the first few weeks after Native Americans, except for the few in Torrance County, re- Trinity. We used the data on exposure rate and fallout ceived smaller doses than did other ethnic groups. The reason TOA to derive ground deposition densities of the 63 most for the generally smaller doses to Native Americans pueblos important radionuclides in fallout. We found the Trinity was due to their locations being outside the high contamina- fallout pattern exposure rates, the major available resource tion area of the fallout pattern. However, according to the US for estimating doses, to have overall good reliability based Census, there were Native Americans in New Mexico not on comparisons with historical film badges deployed resident in pueblos. For Native Americans living in other across New Mexico before the Trinity test. Fortunately, towns in New Mexico and not in the pueblos, their doses despite that radiation measurement instrumentation was would be expected to be similar in magnitude to the other not well developed in the mid-1940s, some aspects of the ethnic groups living in those same towns. Trinity test, such as the presumed fallout pattern (Quinn 1987; Despite that there was no public notice before the test Hoffman 1945), appear to be relatively well documented. and no evacuations and a low detonation height, our find- The crucial data needed for dose reconstruction were ings indicate that only small geographic areas immediately descriptions of diet and lifestyle from the mid-1940s for dif- downwind received exposures of significance as judged ferent ethnic groups, derived for this study from contemporary by their magnitude relative to naturally occurring background focus groups and interviews. While those data have clear and radiation. All locations other than the center line of the pat- obvious uncertainties due to limitations of memory recall, tern were found to have likely received doses from Trinity they were derived from persons alive and living in New at least 1,000-fold lower than those in the maximum ex- Mexico at the time of Trinity and for that reason are viewed posed locations. as relevant for the purposes of the dose reconstruction. These findings constitute the only comprehensive dose We used the fallout pattern, the data on diet and life- estimates for Trinity known to exist. Doses reconstructed in style, and a large suite of exposure models to estimate doses this study for representative persons of White, Hispanic, www.health-physics.com 456 Health Physics October 2020, Volume 119, Number 4 Native American, and African American ethnicity are being Gordeev K, Shinkarev S, Ilyin L, Bouville A, Hoshi M, Luckyanov used to project the excess cancer risk over their natural base- N, Simon SL. Retrospective dose assessment for the population living in areas of local fallout from the Semipalatinsk Nuclear line rate of occurrence (Cahoon et al. 2020). Test Site, part I: external exposure. J Radiat Res 47(Suppl): A129–A136; 2006a. Gordeev K, Shinkarev S, Ilyin L, Bouville A, Hoshi M, Luckyanov Acknowledgments—This research was supported primarily by the Intramural N, Simon SL. Retrospective dose assessment for the population Research Program of the NCI, NIH with additional support from the living in areas of local fallout from the Semipalatinsk Nuclear Intra-Agency Agreement between the National Institute of Allergy and Infectious Test Site, part II: internal exposure to thyroid. J Radiat Res Diseases and the National Cancer Institute, NIAID agreement #Y2-Al-5077 and – NCI agreement #Y3-CO-5117. The authors of this report are appreciative of 47(Suppl):A137 A141; 2006b. the expertise and contributions of authors and co-authors of companion papers Hoffman JG. Nuclear explosion 16 July 1945, part C: transcript in this issue: Nancy Potischman, Silvia I. Salazar, Marianne Naranjo, Mary of radiation monitor’s field notes. Film badge data on town Alice Scott, Emily Haozous, Lisa Cahoon, Ruth Pfeiffer, John Boice, Kathy monitoring. Los Alamos, NM: Los Alamos Scientific Labora- Thiessen, and F. Owen Hoffman. Furthermore, we appreciate the encourage- tory; Nuclear Testing Archive, Las Vegas, Nevada, Accession # ment and advice given by numerous experts, senior centers and organizations NV0059839; 1945. who were consulted in the State of New Mexico, including Las Mujeres Hawthorne H. Compilation of local fallout data from test deto- Hablan, the Tularosa Basin Downwinders Consortium, and Native American – Tribes. We are especially grateful of the time given and the commitment of nations 1945 1962. Extracted from DASA 1251. Defense residents of New Mexico who participated in interviews and focus groups Nuclear Agency, DNA 1251-1-EX Volume I -Continental and shared information on historical diet and lifestyles that enabled the U.S. Tests [online] 1979. Available from: http://www.dtra. radiation doses in this work to be estimated. mil/Portals/61/Documents/NTPR/4-Rad_Exp_Rpts/39_ DNA-1251-1-EX_Compilation_Fallout_Data_CONUS. pdf. Accessed 15 January 2017. REFERENCES Hicks HG. Results of calculations of external exposure rates from Anspaugh L, Church B. Historical estimates of external gamma fallout and the related radionuclide compositions—the Trinity exposure and collective external gamma exposure from testing event. Livermore, CA: Lawrence Livermore National Labora- at the Nevada Test Site. I. Test Series through Hardtack II, tory; Report UCRL-53705; 1985. 1958. Health Phys 51:35–51; 1986. Ibrahim SA, Simon SL, Bouville A, Melo D, Beck HL. Alimen- Beck HL, Anspaugh LR, Bouville A, Simon SL. Review of tary tract absorption (f1 values) for radionuclides in local methods of dose estimation for epidemiological studies of the and regional fallout from nuclear tests. Health Phys 99: radiological impact of NTS and Global Fallout. Radiat Res 233–251; 2010. – 166:209 218; 2006. ICRP. Conversion coefficients for radiological protection quanti- Beck HL, Bouville A, Moroz B, Simon SL. Fallout deposition in ties for external radiation exposures. Oxford: Pergamon Press; the Marshall Islands from Bikini and Enewetak nuclear ICRP Publication 116, Ann. ICRP 40(2–5); 2010. – weapons tests. Health Phys 99:124 142; 2010. KerberRA,TillJE,SimonSL,LyonJL,ThomasDC,Preston- Beck HL, Simon SL, Bouville A, Romanyukha A. Accounting for Martin S, Rallison ML, Lloyd RD, Stevens W. A cohort study unfissioned plutonium from the Trinity Atomic Bomb Test, – of thyroid disease in relation to fallout from nuclear testing. 2019. Health Phys 119:504 516; 2020. JAMA 270:2076–2082; 1993. Bouville A, Beck HL. The HASL gummed-film network and its LAHDRA. Los Alamos Historical Document Retrieval and Assessment use in the reconstruction of doses resulting from nuclear – Project [online] 2009. Draft Final Report. Available at http://www. weapons tests. Technol 7:355 362; 2000. lahdra.org/pubs/reports/Entire%20report/LAHDRA%20Draft% Bouville A, Beck HL, Simon SL. Estimation of doses from exter- 20Final%20Report_vJy23p.pdf. Accessed 15 January 2017. nal irradiation to Marshall Islanders from nuclear testing. Health Phys 99:143–156; 2010. Land CE, Zhumadilov Z, Gusev BI, Hartshorne MH, Wiest PW, Bouville A, Beck HL, Thiessen KM, Hoffman FO, Potischman N, Woodward PW, Crooks LA, Luckyanov NK, Fillmore CM, Simon SL. The methodology used to assess doses from the Carr Z, Abisheva G, Beck HL, Bouville A, Langer J, Weinstock first nuclear weapons test (Trinity) to the populations of R, Gordeev KI, Shinkarev S, Simon SL. Ultrasound-detected – thyroid nodule prevalence and radiation dose from fallout. New Mexico. Health Phys 119:400 427; 2020. – Bouville A, Simon SL, Miller CL, Beck HL, Anspaugh LR. Radiat Res 169:373 383; 2008. Estimates of doses from global fallout. Health Phys 82: Land CE, Bouville A, Apostoaei I, Simon SL. Projected lifetime – cancer risks from exposure to local radioactive fallout in the 690 705; 2002. – Brookins DG. Background radiation in the Albuquerque, New Marshall Islands. Health Phys 99:201 215; 2010. Mexico, U.S.A., area. Environ Geol Water Sci 19:11–15; 1992. Land CE, Kwon D, Hoffman FO, Moroz B, Drozdovitch V, Cahoon EK, Zhang R, Simon SL, Bouville A, Pfeiffer RM. Bouville A, Beck HL, Luckyanov N, Weinstock RM, Simon Projected cancer risks to residents of New Mexico from SL. Accounting for shared and unshared dosimetric uncer- exposure to Trinity radioactive fallout. Health Phys 119: tainties in the dose-response for ultrasound-detected thyroid 478–493; 2020. nodules following exposure to radioactive fallout. Radiat Res Cederwall R, Peterson K. Meteorological modeling of arrival and 183:159–173; 2015. deposition of fallout at intermediate distances downwind of the LloydR,GrenDC,SimonSL,WrennME,HawthorneHA, Nevada Test Site. Health Phys 59:593–601; 1990. Stevens W,Till J, Lotz T. Assigning individual exposures from Drozdovitch V, Bouville A, Doyon F, Brindel P, Cardis E, de Nevada Test Site fallout for Utah leukemia cases and controls. Vathaire F. Reconstruction of individual radiation doses for a Health Phys 59:723–737; 1990. case-control study of thyroid cancer in French Polynesia. Moroz B, Beck HL, Bouville A, Simon SL. Predictions of dispersion and Health Phys 94:418–433; 2008. deposition of fallout from nuclear testing using the NOAA-HYSPLIT Gilbert ES, Huang L, Bouville A, Berg CD, Ron E. Thyroid can- meteorological model. Health Physics 99(2): 252–269; 2010. cer rates and 131-I doses from Nevada atmospheric nuclear Maxwell RM, Anspaugh LR. An improved model for prediction bomb tests: an update. Radiat Res 173:659–664; 2010. of resuspension. Health Phys 101:722–730; 2011.

www.health-physics.com Estimated doses from Trinity nuclear test c S. L. SIMON ET AL. 457 National Cancer Institute. Estimated exposures and thyroid doses Simon SL, Till JE, Lloyd RD, Kerber RL, Thomas DC, Preston- received by the American people from iodine-131 in fallout Martin S, Lyon JL, Stevens W.The Utah leukemia case-control following Nevada atmospheric nuclear bomb tests. Bethesda, study: dosimetry methodology and results. Health Phys 68: MD: National Cancer Institute; 1997. Available at https:// 460–471; 1995. www.cancer.gov/about-cancer/causes-prevention/risk/ StevensW,ThomasDC,LyonJL,TillJE,KerberR,SimonSL, radiation/i131-report-and-appendix. Accessed 15 January 2017. Lloyd RD, Preston-Martin S. Leukemia in Utah and radioac- Ng YC, Anspaugh LR, Cederwall RT. ORERP dose internal esti- tive fallout from the Nevada Test Site—a case-control study. mates for individuals. Health Physics 59:693–713; 1990. JAMA 264:585–591; 1990. Potischman N, Salazar SI, Scott MA, Naranjo M, Haozous E, Tularosa Basin Downwinders Consortium. Unknowing, unwill- Bouville A, Simon SL. Methods and findings on diet and ing, and uncompensated—the effects of the Trinity test on lifestyle used to support estimation of radiation doses from New Mexicans and the potential benefits of Radiation Expo- radioactive fallout from the Trinity nuclear test. Health Phys sure Compensation Act (RECA) Amendments [online]. 2017. 119:390–399; 2020. Available at https://www.trinitydownwinders.com/health- Quinn VE. Analysis of nuclear test Trinity radiological and mete- impact-assessment. Accessed 1 March 2017. orological data. NOAA, Weather Service Nuclear Support Of- US Census. Available at https://1940census.archives.gov/ and fice, Las Vegas: NVO; NVO 313; 1987. https://www.census.gov/prod/www/decennial.html. Accessed Simon SL, Anspaugh LR, Hoffman FO, Scholl AE, Stone MB, 15 January 2015. Thomas BA, Lyon JL. Update of dosimetry for the Utah thy- US DHHS. A feasibility study of the health consequences to the roid cohort study. Radiat Res 165:208–222; 2006. American population from nuclear weapons tests conducted Simon SL, Bouville A, Melo D, Beck HL, Weinstock R. Acute by the United States and other nations. Washington, DC: and chronic intakes of fallout radionuclides by Marshallese Department of Health and Human Services; 2005. Available from nuclear weapons testing at Bikini and Enewetak and re- at http://www.cdc.gov/nceh/radiation/fallout/default.htm. lated internal radiation doses. Health Phys 99:157–200; 2010. Accessed 15 January 2017. Simon SL, Lloyd RD, Till JE, Hawthorne HA, Gren DC, Rallison M, Stevens W.Development of a method to estimate dose from fallout radioiodine to persons in a thyroid cohort study. Health Phys 59:669–691; 1990. ■■

www.health-physics.com 458 Health Physics October 2020, Volume 119, Number 4 APPENDIX Table 1A1. Estimates of precinct population-weighted average radiation absorbed doses (mGy) by county and age group for Whites/colon from all sources of internal and external exposure to fallout radionuclides (as discussed in the text). Doses rounded to two significant digits except those less than 0.01 mGy are rounded to one digit.

AGE GROUP (yrs)

COUNTY In-utero 0–11–23–78–12 13–17 Adult (18+) Bernalillo 8.2E-02 1.3E-01 4.3E+00 2.2E+00 1.8E+00 1.2E+00 7.7E-01 Catron 1.7E-02 9.6E-02 1.2E-01 7.7E-02 7.7E-02 7.1E-02 6.7E-02 Chaves 4.6E-02 6.3E-02 8.6E-01 5.0E-01 4.3E-01 3.0E-01 2.1E-01 Colfax 2.3E-01 4.1E-01 1.3E+00 9.4E-01 8.8E-01 7.2E-01 6.6E-01 Curry 9.0E-03 2.4E-02 7.6E-02 5.3E-02 4.9E-02 3.8E-02 3.0E-02 De Baca 9.8E-02 1.4E-01 6.9E-01 5.1E-01 4.7E-01 3.5E-01 2.8E-01 Dona Ana 1.1E-02 2.1E-02 6.5E-01 3.4E-01 2.7E-01 1.9E-01 1.2E-01 Eddy 1.1E-02 1.1E-02 6.5E-01 3.4E-01 2.7E-01 1.8E-01 1.1E-01 Grant 1.6E-02 9.1E-02 1.2E+00 8.0E-01 6.4E-01 3.6E-01 1.9E-01 Guadalupe 4.0E+00 5.6E+00 2.7E+01 2.0E+01 1.8E+01 1.4E+01 1.1E+01 Harding 2.7E-02 1.5E-01 2.6E-01 1.8E-01 1.7E-01 1.4E-01 1.2E-01 Hidalgo 8.6E-03 7.7E-02 9.6E-02 6.1E-02 6.2E-02 5.4E-02 4.9E-02 Lea 1.1E-02 1.1E-02 6.5E-01 3.4E-01 2.7E-01 1.9E-01 1.2E-01 Lincoln 2.9E+00 7.3E+00 1.1E+01 8.1E+00 8.3E+00 6.8E+00 5.6E+00 Luna 1.7E-02 1.1E-02 1.3E+00 6.6E-01 5.3E-01 3.6E-01 2.2E-01 McKinley 1.5E-02 8.1E-02 8.6E-01 4.5E-01 3.7E-01 2.6E-01 1.7E-01 Mora 2.4E-01 1.1E+00 1.7E+00 1.2E+00 1.2E+00 1.0E+00 9.6E-01 Otero 1.2E-02 6.5E-02 6.2E-01 3.3E-01 2.8E-01 1.9E-01 1.2E-01 Quay 2.2E-02 5.0E-02 1.7E-01 1.2E-01 1.1E-01 8.6E-02 7.1E-02 Rio Arriba 4.0E-02 2.1E-01 7.3E-01 4.7E-01 4.1E-01 2.8E-01 2.1E-01 Roosevelt 8.6E-03 1.4E-02 7.0E-02 5.0E-02 4.6E-02 3.4E-02 2.7E-02 Sandoval 4.4E-02 2.6E-01 7.7E-01 4.9E-01 4.3E-01 3.1E-01 2.3E-01 San Juan 1.2E-02 1.1E-02 8.1E-01 4.2E-01 3.4E-01 2.3E-01 1.4E-01 San Miguel 5.1E-01 2.9E+00 4.9E+00 3.2E+00 3.1E+00 2.7E+00 2.4E+00 Santa Fe 2.0E-01 1.5E+00 5.1E+00 3.2E+00 2.7E+00 2.0E+00 1.7E+00 Sierra 9.1E-03 1.4E-02 5.7E-02 4.0E-02 3.8E-02 3.2E-02 2.9E-02 Socorro 2.4E+00 2.5E+01 1.7E+01 1.0E+01 1.1E+01 1.1E+01 1.1E+01 Taos 9.5E-02 5.4E-01 9.5E-01 6.5E-01 5.9E-01 4.7E-01 4.1E-01 Torrance 8.8E+00 3.8E+01 3.8E+01 2.5E+01 2.6E+01 2.4E+01 2.3E+01 Union 2.8E-02 8.9E-02 2.3E-01 1.6E-01 1.5E-01 1.2E-01 1.0E-01 Valencia 6.3E-02 2.7E-01 1.2E+00 6.9E-01 6.1E-01 4.6E-01 3.4E-01

1E-notation used here to conserve space.

www.health-physics.com Estimated doses from Trinity nuclear test c S. L. SIMON ET AL. 459 Table 1B. Estimates of precinct population-weighted average radiation absorbed doses (mGy) by county and age group for Whites/lung from all sources of internal and external exposure to fallout radionuclides (as discussed in the text). Doses rounded to two significant digits except those less than 0.01 mGy are rounded to one digit.

AGE GROUP (yrs)

COUNTY In-utero 0–11–23–78–12 13–17 Adult (18+) Bernalillo 7.9E-02 7.4E-02 1.1E-01 9.2E-02 1.0E-01 9.5E-02 7.4E-02 Catron 1.7E-02 2.5E-02 2.8E-02 2.6E-02 3.0E-02 2.8E-02 2.3E-02 Chaves 4.4E-02 6.4E-02 7.6E-02 7.3E-02 9.2E-02 8.8E-02 7.1E-02 Colfax 2.2E-01 3.8E-01 4.5E-01 4.2E-01 5.2E-01 5.1E-01 4.2E-01 Curry 8.3E-03 1.3E-02 1.5E-02 1.5E-02 1.8E-02 1.7E-02 1.4E-02 De Baca 8.9E-02 1.3E-01 1.5E-01 1.5E-01 1.8E-01 1.7E-01 1.3E-01 Dona Ana 1.1E-02 1.1E-02 1.6E-02 1.4E-02 1.7E-02 1.6E-02 1.2E-02 Eddy 1.1E-02 1.1E-02 1.6E-02 1.4E-02 1.7E-02 1.6E-02 1.2E-02 Grant 1.5E-02 1.2E-02 2.0E-02 1.7E-02 1.8E-02 1.6E-02 1.3E-02 Guadalupe 3.6E+00 5.5E+00 6.1E+00 6.0E+00 7.5E+00 7.1E+00 5.6E+00 Harding 2.5E-02 3.9E-02 4.4E-02 4.3E-02 5.5E-02 5.2E-02 4.2E-02 Hidalgo 7.4E-03 1.1E-02 1.3E-02 1.2E-02 1.6E-02 1.5E-02 1.2E-02 Lea 1.1E-02 1.1E-02 1.6E-02 1.4E-02 1.7E-02 1.6E-02 1.2E-02 Lincoln 2.7E+00 4.1E+00 4.4E+00 4.4E+00 5.3E+00 5.0E+00 3.9E+00 Luna 1.7E-02 1.1E-02 2.0E-02 1.6E-02 1.8E-02 1.7E-02 1.3E-02 McKinley 1.4E-02 1.2E-02 1.8E-02 1.6E-02 1.8E-02 1.7E-02 1.3E-02 Mora 2.2E-01 3.4E-01 3.9E-01 3.7E-01 4.4E-01 4.3E-01 3.6E-01 Otero 1.1E-02 1.1E-02 1.6E-02 1.4E-02 1.7E-02 1.6E-02 1.2E-02 Quay 2.0E-02 3.1E-02 3.5E-02 3.4E-02 4.3E-02 4.0E-02 3.2E-02 Rio Arriba 3.7E-02 5.6E-02 6.7E-02 6.1E-02 7.3E-02 7.0E-02 5.9E-02 Roosevelt 8.0E-03 1.3E-02 1.5E-02 1.4E-02 1.8E-02 1.7E-02 1.4E-02 Sandoval 3.9E-02 5.4E-02 6.4E-02 5.7E-02 6.6E-02 6.3E-02 5.4E-02 San Juan 1.2E-02 1.1E-02 1.8E-02 1.5E-02 1.9E-02 1.7E-02 1.4E-02 San Miguel 4.5E-01 7.2E-01 8.5E-01 7.8E-01 9.3E-01 8.7E-01 7.1E-01 Santa Fe 1.4E-01 5.8E-01 6.2E-01 4.3E-01 4.5E-01 4.1E-01 3.5E-01 Sierra 8.4E-03 1.4E-02 2.3E-02 1.8E-02 2.0E-02 2.0E-02 1.8E-02 Socorro 2.4E+00 2.5E+01 1.6E+01 9.6E+00 1.1E+01 1.1E+01 1.1E+01 Taos 9.4E-02 4.7E-01 8.2E-01 5.9E-01 5.7E-01 4.8E-01 4.1E-01 Torrance 8.7E+00 3.6E+01 3.5E+01 2.5E+01 2.7E+01 2.7E+01 2.5E+01 Union 3.5E-02 4.8E-02 6.7E-02 6.4E-02 7.9E-02 7.8E-02 6.9E-02 Valencia 5.4E-02 7.6E-02 1.1E-01 9.7E-02 1.1E-01 9.4E-02 7.2E-02

www.health-physics.com 460 Health Physics October 2020, Volume 119, Number 4 Table 1C. Estimates of precinct population-weighted average radiation absorbed doses (mGy) by county and age group for Whites/RBM from all sources of internal and external exposure to fallout radionuclides (as discussed in the text). Doses rounded to two significant digits except those less than 0.01 mGy are rounded to one digit.

AGE GROUP (yrs)

COUNTY In-utero 0–11–23–78–12 13–17 Adult (18+) Bernalillo 8.8E-01 6.5E-02 3.2E-01 1.9E-01 2.1E-01 2.1E-01 9.9E-02 Catron 3.4E-02 3.4E-02 2.8E-02 2.4E-02 2.6E-02 2.5E-02 1.9E-02 Chaves 1.8E-01 4.9E-02 9.3E-02 7.1E-02 8.0E-02 7.7E-02 5.0E-02 Colfax 3.3E-01 2.8E-01 3.2E-01 2.8E-01 2.8E-01 2.7E-01 2.2E-01 Curry 1.5E-02 1.1E-02 1.2E-02 1.1E-02 1.2E-02 1.1E-02 8.1E-03 De Baca 1.4E-01 1.1E-01 1.2E-01 1.2E-01 1.3E-01 1.2E-01 9.0E-02 Dona Ana 1.3E-01 9.0E-03 4.7E-02 2.7E-02 3.1E-02 3.1E-02 1.4E-02 Eddy 1.3E-01 8.1E-03 4.8E-02 2.7E-02 3.1E-02 3.1E-02 1.4E-02 Grant 2.8E-01 2.6E-02 1.1E-01 6.9E-02 7.8E-02 6.8E-02 2.3E-02 Guadalupe 5.6E+00 4.3E+00 5.0E+00 4.7E+00 5.2E+00 4.8E+00 3.6E+00 Harding 4.7E-02 3.7E-02 3.5E-02 3.2E-02 3.6E-02 3.4E-02 2.5E-02 Hidalgo 1.6E-02 1.3E-02 1.1E-02 9.5E-03 1.1E-02 1.1E-02 7.6E-03 Lea 1.3E-01 8.1E-03 4.8E-02 2.7E-02 3.1E-02 3.1E-02 1.5E-02 Lincoln 3.2E+00 3.5E+00 3.6E+00 3.5E+00 3.8E+00 3.5E+00 2.7E+00 Luna 3.7E-01 8.1E-03 1.2E-01 5.9E-02 6.8E-02 7.2E-02 2.8E-02 McKinley 1.7E-01 1.4E-02 6.0E-02 3.4E-02 3.9E-02 4.0E-02 1.8E-02 Mora 3.8E-01 3.2E-01 3.1E-01 2.7E-01 2.8E-01 2.7E-01 2.2E-01 Otero 1.2E-01 1.3E-02 4.5E-02 2.6E-02 3.0E-02 3.0E-02 1.4E-02 Quay 3.3E-02 2.6E-02 2.8E-02 2.7E-02 2.9E-02 2.7E-02 2.0E-02 Rio Arriba 1.4E-01 5.7E-02 7.9E-02 6.0E-02 6.3E-02 5.9E-02 3.9E-02 Roosevelt 1.4E-02 9.7E-03 1.2E-02 1.1E-02 1.2E-02 1.1E-02 8.2E-03 Sandoval 1.5E-01 6.0E-02 8.2E-02 6.2E-02 6.5E-02 6.2E-02 4.1E-02 San Juan 1.7E-01 8.1E-03 5.8E-02 3.2E-02 3.6E-02 3.7E-02 1.6E-02 San Miguel 8.9E-01 6.7E-01 6.5E-01 5.8E-01 6.4E-01 6.2E-01 4.6E-01 Santa Fe 4.2E-01 2.2E-01 2.4E-01 1.9E-01 2.0E-01 2.0E-01 1.4E-01 Sierra 1.3E-02 9.9E-03 1.1E-02 1.0E-02 1.1E-02 1.0E-02 8.1E-03 Socorro 2.4E+00 3.0E+00 2.7E+00 2.3E+00 2.3E+00 2.2E+00 1.9E+00 Taos 2.0E-01 1.4E-01 1.5E-01 1.2E-01 1.2E-01 1.1E-01 9.1E-02 Torrance 1.0E+01 1.1E+01 1.2E+01 9.9E+00 9.3E+00 8.8E+00 8.1E+00 Union 7.5E-02 5.7E-02 6.7E-02 5.8E-02 5.8E-02 5.6E-02 5.0E-02 Valencia 2.4E-01 7.0E-02 1.1E-01 8.5E-02 9.6E-02 9.3E-02 5.8E-02

www.health-physics.com Estimated doses from Trinity nuclear test c S. L. SIMON ET AL. 461 Table 1D. Estimates of precinct population-weighted average radiation absorbed doses (mGy) by county and age group for Whites/stomach from all sources of internal and external exposure to fallout radionuclides (as discussed in the text). Doses rounded to two significant digits except those less than 0.01 mGy are rounded to one digit.

AGE GROUP (yrs)

COUNTY In-utero 0–11–23–78–12 13–17 Adult (18+) Bernalillo 8.1E-02 6.8E-02 3.0E-01 1.8E-01 1.6E-01 1.3E-01 9.2E-02 Catron 2.5E-02 4.2E-02 4.0E-02 3.3E-02 3.4E-02 3.1E-02 2.6E-02 Chaves 4.6E-02 5.0E-02 9.6E-02 7.5E-02 7.5E-02 6.6E-02 5.0E-02 Colfax 2.3E-01 2.8E-01 3.5E-01 3.0E-01 2.9E-01 2.7E-01 2.4E-01 Curry 8.8E-03 1.1E-02 1.4E-02 1.3E-02 1.3E-02 1.2E-02 8.9E-03 De Baca 9.8E-02 1.1E-01 1.6E-01 1.4E-01 1.4E-01 1.3E-01 1.0E-01 Dona Ana 1.1E-02 9.1E-03 4.4E-02 2.6E-02 2.3E-02 1.9E-02 1.3E-02 Eddy 1.1E-02 8.3E-03 4.4E-02 2.6E-02 2.3E-02 1.9E-02 1.3E-02 Grant 1.5E-02 1.5E-02 7.3E-02 4.9E-02 3.9E-02 2.8E-02 1.7E-02 Guadalupe 4.0E+00 4.5E+00 6.3E+00 5.6E+00 5.9E+00 5.2E+00 4.1E+00 Harding 2.7E-02 3.7E-02 4.4E-02 3.7E-02 3.9E-02 3.6E-02 2.9E-02 Hidalgo 8.4E-03 1.3E-02 1.4E-02 1.1E-02 1.2E-02 1.1E-02 9.2E-03 Lea 1.1E-02 8.3E-03 4.5E-02 2.6E-02 2.3E-02 1.9E-02 1.3E-02 Lincoln 2.9E+00 3.9E+00 4.2E+00 3.9E+00 4.2E+00 3.7E+00 2.9E+00 Luna 1.7E-02 8.3E-03 7.6E-02 4.2E-02 3.5E-02 2.9E-02 1.9E-02 McKinley 1.5E-02 1.5E-02 5.8E-02 3.4E-02 3.0E-02 2.6E-02 1.8E-02 Mora 2.3E-01 3.2E-01 3.6E-01 3.0E-01 3.0E-01 2.8E-01 2.5E-01 Otero 1.2E-02 1.3E-02 4.3E-02 2.6E-02 2.3E-02 1.9E-02 1.4E-02 Quay 2.2E-02 2.6E-02 3.4E-02 3.1E-02 3.2E-02 2.9E-02 2.3E-02 Rio Arriba 3.9E-02 5.4E-02 8.3E-02 6.3E-02 5.8E-02 5.1E-02 4.3E-02 Roosevelt 8.5E-03 1.0E-02 1.4E-02 1.2E-02 1.3E-02 1.1E-02 8.9E-03 Sandoval 4.3E-02 6.1E-02 8.9E-02 6.6E-02 6.2E-02 5.5E-02 4.7E-02 San Juan 1.2E-02 8.2E-03 5.3E-02 3.1E-02 2.6E-02 2.2E-02 1.5E-02 San Miguel 5.0E-01 7.1E-01 8.3E-01 6.9E-01 7.2E-01 6.6E-01 5.4E-01 Santa Fe 1.9E-01 2.3E-01 4.8E-01 3.3E-01 2.8E-01 2.5E-01 2.1E-01 Sierra 8.9E-03 1.0E-02 1.4E-02 1.2E-02 1.2E-02 1.1E-02 9.2E-03 Socorro 2.4E+00 4.5E+00 3.7E+00 2.8E+00 2.8E+00 2.8E+00 2.6E+00 Taos 9.4E-02 1.4E-01 1.6E-01 1.3E-01 1.2E-01 1.1E-01 1.0E-01 Torrance 8.9E+00 1.3E+01 1.4E+01 1.1E+01 1.0E+01 9.7E+00 9.1E+00 Union 6.2E-02 7.4E-02 9.6E-02 7.9E-02 7.4E-02 6.9E-02 6.8E-02 Valencia 6.2E-02 7.7E-02 1.3E-01 9.8E-02 9.7E-02 8.6E-02 6.7E-02

www.health-physics.com 462 Health Physics October 2020, Volume 119, Number 4 Table 1E. Estimates of precinct population-weighted average radiation absorbed doses (mGy) by county and age group for Whites/thyroid from all sources of internal and external exposure to fallout radionuclides (as discussed in the text). Doses rounded to two significant digits except those less than 0.01 mGy are rounded to one digit.

AGE GROUP (yrs)

COUNTY In-utero 0–11–23–78–12 13–17 Adult (18+) Bernalillo 1.9E+00 4.0E-01 1.5E+01 8.9E+00 6.0E+00 4.6E+00 2.2E+00 Catron 3.2E-01 8.4E-01 1.6E+00 1.1E+00 8.2E-01 6.8E-01 3.8E-01 Chaves 5.3E-01 1.1E-01 3.6E+00 2.2E+00 1.6E+00 1.2E+00 6.5E-01 Colfax 2.2E+00 2.6E+00 1.3E+01 9.4E+00 6.6E+00 4.9E+00 2.7E+00 Curry 1.3E-01 4.8E-02 8.3E-01 5.2E-01 3.6E-01 2.9E-01 1.5E-01 De Baca 1.1E+00 2.3E-01 6.9E+00 4.4E+00 3.1E+00 2.4E+00 1.3E+00 Dona Ana 2.9E-01 4.8E-02 2.4E+00 1.4E+00 9.4E-01 7.2E-01 3.5E-01 Eddy 2.9E-01 1.8E-02 2.4E+00 1.4E+00 9.4E-01 7.2E-01 3.5E-01 Grant 4.1E-01 6.4E-01 3.8E+00 2.8E+00 1.9E+00 1.2E+00 5.0E-01 Guadalupe 4.2E+01 8.5E+00 2.5E+02 1.6E+02 1.2E+02 9.1E+01 5.0E+01 Harding 4.1E-01 3.1E-01 2.4E+00 1.5E+00 1.1E+00 8.8E-01 5.0E-01 Hidalgo 1.4E-01 1.7E-01 7.5E-01 4.7E-01 3.5E-01 3.0E-01 1.8E-01 Lea 3.1E-01 1.8E-02 2.6E+00 1.5E+00 1.0E+00 7.7E-01 3.7E-01 Lincoln 1.1E+01 7.1E+00 5.5E+01 3.6E+01 2.7E+01 2.2E+01 1.3E+01 Luna 4.9E-01 1.8E-02 4.2E+00 2.4E+00 1.6E+00 1.3E+00 6.0E-01 McKinley 3.9E-01 1.8E-01 3.0E+00 1.7E+00 1.2E+00 9.4E-01 4.7E-01 Mora 2.6E+00 4.7E+00 1.2E+01 8.7E+00 6.4E+00 5.1E+00 3.1E+00 Otero 3.0E-01 2.0E-01 2.3E+00 1.4E+00 9.4E-01 7.2E-01 3.6E-01 Quay 2.5E-01 9.7E-02 1.5E+00 9.9E-01 7.0E-01 5.6E-01 3.1E-01 Rio Arriba 5.0E-01 1.0E+00 3.0E+00 2.2E+00 1.5E+00 1.1E+00 6.0E-01 Roosevelt 1.3E-01 2.5E-02 8.9E-01 5.5E-01 3.9E-01 3.0E-01 1.6E-01 Sandoval 5.2E-01 1.2E+00 2.8E+00 2.0E+00 1.4E+00 1.1E+00 6.1E-01 San Juan 3.5E-01 2.5E-02 3.0E+00 1.8E+00 1.2E+00 9.0E-01 4.3E-01 San Miguel 8.8E+00 7.2E+00 5.6E+01 3.5E+01 2.5E+01 2.0E+01 1.1E+01 Santa Fe 3.7E+00 1.1E+01 2.0E+01 1.4E+01 9.7E+00 7.8E+00 4.8E+00 Sierra 8.5E-02 6.7E-02 4.4E-01 3.1E-01 2.2E-01 1.7E-01 1.0E-01 Socorro 7.6E+00 2.5E+01 2.1E+01 1.4E+01 1.2E+01 1.2E+01 8.4E+00 Taos 9.3E-01 2.6E+00 4.6E+00 3.6E+00 2.6E+00 1.9E+00 1.1E+00 Torrance 4.2E+01 1.0E+02 2.1E+02 1.7E+02 1.2E+02 8.4E+01 4.8E+01 Union 5.3E-01 9.1E-01 2.9E+00 2.2E+00 1.5E+00 1.1E+00 6.5E-01 Valencia 9.1E-01 6.1E-01 5.7E+00 3.5E+00 2.5E+00 2.0E+00 1.1E+00

www.health-physics.com Estimated doses from Trinity nuclear test c S. L. SIMON ET AL. 463 Table 2A. Estimates of precinct population-weighted average radiation absorbed doses (mGy) by county and age group for Hispanics/Colon from all sources of internal and external exposure to fallout radionuclides (as discussed in the text). Doses rounded to two significant digits except those less than 0.01 mGy are rounded to one digit.

AGE GROUP (yrs)

COUNTY In-utero 0–11–23–78–12 13–17 Adult (18+) Bernalillo 7.6E-02 1.6E-01 3.5E+00 2.2E+00 1.8E+00 1.1E+00 7.6E-01 Catron 1.8E-02 9.9E-02 1.2E-01 8.1E-02 8.1E-02 7.5E-02 7.0E-02 Chaves 3.5E-02 4.3E-02 7.9E-01 5.0E-01 4.3E-01 2.9E-01 2.0E-01 Colfax 2.4E-01 4.6E-01 1.4E+00 1.0E+00 9.6E-01 7.9E-01 7.3E-01 Curry 8.3E-03 2.5E-02 8.2E-02 5.7E-02 5.3E-02 4.0E-02 3.1E-02 De Baca 1.1E-01 1.8E-01 7.9E-01 5.8E-01 5.4E-01 4.2E-01 3.5E-01 Dona Ana 1.1E-02 2.7E-02 5.0E-01 3.2E-01 2.7E-01 1.7E-01 1.2E-01 Eddy 1.0E-02 1.8E-02 4.8E-01 3.2E-01 2.6E-01 1.7E-01 1.1E-01 Grant 1.6E-02 1.1E-01 9.8E-01 7.8E-01 6.2E-01 3.3E-01 1.9E-01 Guadalupe 4.5E+00 7.4E+00 3.1E+01 2.2E+01 2.1E+01 1.6E+01 1.4E+01 Harding 3.0E-02 1.6E-01 2.9E-01 2.0E-01 1.9E-01 1.6E-01 1.4E-01 Hidalgo 9.5E-03 8.2E-02 1.0E-01 6.7E-02 6.8E-02 6.0E-02 5.3E-02 Lea 1.1E-02 1.8E-02 4.9E-01 3.2E-01 2.6E-01 1.7E-01 1.1E-01 Lincoln 3.2E+00 7.9E+00 1.2E+01 8.8E+00 8.6E+00 7.2E+00 6.3E+00 Luna 1.7E-02 4.6E-02 7.0E-01 6.1E-01 4.9E-01 2.9E-01 2.2E-01 McKinley 1.6E-02 1.0E-01 5.0E-01 4.1E-01 3.5E-01 2.2E-01 1.7E-01 Mora 2.5E-01 1.1E+00 1.9E+00 1.3E+00 1.3E+00 1.1E+00 1.0E+00 Otero 1.2E-02 7.9E-02 3.9E-01 3.1E-01 2.6E-01 1.6E-01 1.2E-01 Quay 2.5E-02 6.1E-02 1.9E-01 1.4E-01 1.3E-01 1.0E-01 8.6E-02 Rio Arriba 4.2E-02 2.2E-01 6.3E-01 4.7E-01 4.1E-01 2.8E-01 2.2E-01 Roosevelt 9.7E-03 1.9E-02 7.7E-02 5.7E-02 5.3E-02 4.0E-02 3.3E-02 Sandoval 4.6E-02 2.7E-01 6.6E-01 4.8E-01 4.3E-01 3.0E-01 2.4E-01 San Juan 1.3E-02 3.1E-02 4.5E-01 3.9E-01 3.2E-01 1.9E-01 1.4E-01 San Miguel 4.7E-01 2.9E+00 5.1E+00 3.4E+00 3.3E+00 2.8E+00 2.4E+00 Santa Fe 2.4E-01 1.5E+00 5.2E+00 3.2E+00 2.8E+00 2.1E+00 1.7E+00 Sierra 9.8E-03 1.6E-02 6.3E-02 4.4E-02 4.3E-02 3.6E-02 3.2E-02 Socorro 2.4E+00 2.5E+01 1.8E+01 1.0E+01 1.1E+01 1.1E+01 1.1E+01 Taos 9.6E-02 5.5E-01 9.3E-01 6.5E-01 6.0E-01 4.7E-01 4.1E-01 Torrance 8.8E+00 3.8E+01 3.8E+01 2.5E+01 2.6E+01 2.4E+01 2.3E+01 Union 3.0E-02 1.0E-01 2.5E-01 1.8E-01 1.7E-01 1.4E-01 1.2E-01 Valencia 7.0E-02 3.2E-01 9.2E-01 7.0E-01 6.2E-01 4.5E-01 3.7E-01

www.health-physics.com 464 Health Physics October 2020, Volume 119, Number 4 Table 2B. Estimates of precinct population-weighted average radiation absorbed doses (mGy) by county and age group for Hispanics/Lung from all sources of internal and external exposure to fallout radionuclides (as discussed in the text). Doses rounded to two significant digits except those less than 0.01 mGy are rounded to one digit.

AGE GROUP (yrs)

COUNTY In-utero 0–11–23–78–12 13–17 Adult (18+) Bernalillo 7.2E-02 6.0E-02 1.1E-01 9.3E-02 9.4E-02 7.6E-02 5.2E-02 Catron 1.8E-02 2.6E-02 3.1E-02 2.7E-02 2.9E-02 2.7E-02 2.4E-02 Chaves 3.3E-02 3.7E-02 7.7E-02 6.9E-02 7.1E-02 5.4E-02 3.1E-02 Colfax 2.3E-01 3.9E-01 4.7E-01 4.2E-01 5.1E-01 5.0E-01 4.3E-01 Curry 7.5E-03 1.1E-02 1.6E-02 1.4E-02 1.5E-02 1.3E-02 9.1E-03 De Baca 9.9E-02 1.4E-01 1.7E-01 1.5E-01 1.7E-01 1.6E-01 1.4E-01 Dona Ana 1.0E-02 8.9E-03 1.6E-02 1.4E-02 1.4E-02 1.2E-02 8.3E-03 Eddy 1.0E-02 9.0E-03 1.6E-02 1.4E-02 1.4E-02 1.2E-02 8.5E-03 Grant 1.5E-02 1.3E-02 1.9E-02 1.7E-02 1.8E-02 1.6E-02 1.3E-02 Guadalupe 4.0E+00 5.9E+00 7.0E+00 6.2E+00 7.0E+00 6.7E+00 5.9E+00 Harding 2.7E-02 4.2E-02 5.0E-02 4.4E-02 5.1E-02 5.0E-02 4.4E-02 Hidalgo 8.1E-03 1.2E-02 1.4E-02 1.3E-02 1.5E-02 1.4E-02 1.2E-02 Lea 1.0E-02 8.9E-03 1.6E-02 1.4E-02 1.4E-02 1.2E-02 8.5E-03 Lincoln 3.0E+00 4.4E+00 5.1E+00 4.5E+00 4.9E+00 4.7E+00 4.2E+00 Luna 1.8E-02 1.3E-02 1.8E-02 1.6E-02 1.7E-02 1.6E-02 1.4E-02 McKinley 1.5E-02 1.4E-02 1.8E-02 1.6E-02 1.7E-02 1.6E-02 1.4E-02 Mora 2.3E-01 3.5E-01 4.2E-01 3.7E-01 4.3E-01 4.2E-01 3.7E-01 Otero 1.2E-02 1.2E-02 1.6E-02 1.4E-02 1.6E-02 1.5E-02 1.3E-02 Quay 2.2E-02 3.4E-02 4.0E-02 3.5E-02 4.0E-02 3.9E-02 3.4E-02 Rio Arriba 3.8E-02 5.7E-02 6.9E-02 6.2E-02 7.1E-02 6.9E-02 6.0E-02 Roosevelt 8.9E-03 1.4E-02 1.7E-02 1.5E-02 1.7E-02 1.7E-02 1.5E-02 Sandoval 4.0E-02 5.5E-02 6.6E-02 5.8E-02 6.4E-02 6.2E-02 5.4E-02 San Juan 1.3E-02 1.3E-02 1.7E-02 1.5E-02 1.8E-02 1.7E-02 1.4E-02 San Miguel 4.0E-01 6.1E-01 9.2E-01 7.8E-01 8.2E-01 7.0E-01 5.1E-01 Santa Fe 1.8E-01 6.4E-01 6.8E-01 4.9E-01 5.2E-01 4.7E-01 3.9E-01 Sierra 8.9E-03 1.4E-02 2.4E-02 1.9E-02 2.0E-02 2.0E-02 1.9E-02 Socorro 2.4E+00 2.5E+01 1.7E+01 9.6E+00 1.1E+01 1.1E+01 1.1E+01 Taos 9.5E-02 4.7E-01 8.2E-01 5.9E-01 5.7E-01 4.8E-01 4.1E-01 Torrance 8.7E+00 3.6E+01 3.5E+01 2.5E+01 2.7E+01 2.7E+01 2.5E+01 Union 3.7E-02 5.0E-02 7.2E-02 6.4E-02 7.5E-02 7.5E-02 6.9E-02 Valencia 5.9E-02 8.2E-02 1.2E-01 9.9E-02 1.0E-01 8.9E-02 7.6E-02

www.health-physics.com Estimated doses from Trinity nuclear test c S. L. SIMON ET AL. 465 Table 2C. Estimates of precinct population-weighted average radiation absorbed doses (mGy) by county and age group for Hispanics/RBM from all sources of internal and external exposure to fallout radionuclides (as discussed in the text). Doses rounded to two significant digits except those less than 0.01 mGy are rounded to one digit.

AGE GROUP (yrs)

COUNTY In-utero 0–11–23–78–12 13–17 Adult (18+) Bernalillo 8.6E-01 9.3E-02 2.8E-01 1.9E-01 2.1E-01 2.0E-01 9.0E-02 Catron 3.6E-02 3.6E-02 3.1E-02 2.6E-02 2.5E-02 2.4E-02 2.1E-02 Chaves 1.7E-01 3.8E-02 1.1E-01 8.4E-02 8.8E-02 7.4E-02 3.6E-02 Colfax 3.6E-01 2.9E-01 3.4E-01 2.8E-01 2.7E-01 2.6E-01 2.3E-01 Curry 1.5E-02 1.0E-02 1.5E-02 1.3E-02 1.3E-02 1.1E-02 7.0E-03 De Baca 1.6E-01 1.3E-01 1.5E-01 1.3E-01 1.2E-01 1.2E-01 1.0E-01 Dona Ana 1.3E-01 1.5E-02 4.0E-02 2.8E-02 3.0E-02 2.9E-02 1.3E-02 Eddy 1.3E-01 1.5E-02 3.9E-02 2.7E-02 3.0E-02 2.8E-02 1.3E-02 Grant 2.8E-01 4.2E-02 8.7E-02 6.7E-02 7.6E-02 6.2E-02 2.3E-02 Guadalupe 6.4E+00 5.0E+00 6.0E+00 5.0E+00 4.8E+00 4.6E+00 4.0E+00 Harding 5.3E-02 4.2E-02 4.1E-02 3.4E-02 3.4E-02 3.4E-02 2.7E-02 Hidalgo 1.8E-02 1.5E-02 1.2E-02 1.0E-02 1.0E-02 1.0E-02 8.3E-03 Lea 1.3E-01 1.5E-02 4.0E-02 2.8E-02 3.0E-02 2.9E-02 1.3E-02 Lincoln 3.6E+00 3.9E+00 4.3E+00 3.6E+00 3.4E+00 3.2E+00 3.0E+00 Luna 3.6E-01 4.6E-02 6.5E-02 5.4E-02 6.2E-02 5.8E-02 2.8E-02 McKinley 1.7E-01 3.2E-02 3.9E-02 3.2E-02 3.5E-02 3.3E-02 1.9E-02 Mora 4.0E-01 3.4E-01 3.4E-01 2.8E-01 2.7E-01 2.7E-01 2.3E-01 Otero 1.2E-01 2.4E-02 3.1E-02 2.5E-02 2.8E-02 2.6E-02 1.4E-02 Quay 3.8E-02 2.9E-02 3.4E-02 2.8E-02 2.7E-02 2.7E-02 2.2E-02 Rio Arriba 1.4E-01 6.3E-02 7.5E-02 6.0E-02 6.1E-02 5.7E-02 4.0E-02 Roosevelt 1.6E-02 1.1E-02 1.4E-02 1.1E-02 1.1E-02 1.1E-02 9.0E-03 Sandoval 1.5E-01 6.8E-02 7.7E-02 6.2E-02 6.3E-02 5.9E-02 4.2E-02 San Juan 1.6E-01 2.5E-02 3.6E-02 3.0E-02 3.3E-02 3.1E-02 1.7E-02 San Miguel 8.8E-01 6.3E-01 7.9E-01 6.6E-01 6.7E-01 6.0E-01 3.9E-01 Santa Fe 4.5E-01 2.8E-01 3.0E-01 2.4E-01 2.6E-01 2.6E-01 1.7E-01 Sierra 1.4E-02 1.1E-02 1.3E-02 1.1E-02 1.0E-02 9.9E-03 8.7E-03 Socorro 2.5E+00 3.1E+00 2.8E+00 2.3E+00 2.2E+00 2.1E+00 2.0E+00 Taos 2.0E-01 1.4E-01 1.5E-01 1.2E-01 1.2E-01 1.1E-01 9.1E-02 Torrance 1.0E+01 1.1E+01 1.2E+01 9.9E+00 9.2E+00 8.8E+00 8.1E+00 Union 7.9E-02 6.0E-02 7.1E-02 5.8E-02 5.5E-02 5.5E-02 5.0E-02 Valencia 2.4E-01 9.5E-02 1.0E-01 8.6E-02 8.8E-02 8.5E-02 6.3E-02

www.health-physics.com 466 Health Physics October 2020, Volume 119, Number 4 Table 2D. Estimates of precinct population-weighted average radiation absorbed doses (mGy) by county and age group for Hispanics/Stomach from all sources of internal and external exposure to fallout radionuclides (as discussed in the text). Doses rounded to two significant digits except those less than 0.01 mGy are rounded to one digit.

AGE GROUP (yrs)

COUNTY In-utero 0–11–23–78–12 13–17 Adult (18+) Bernalillo 7.5E-02 6.5E-02 2.7E-01 1.9E-01 1.6E-01 1.3E-01 8.3E-02 Catron 2.7E-02 4.4E-02 4.5E-02 3.5E-02 3.3E-02 3.1E-02 2.8E-02 Chaves 3.4E-02 3.5E-02 1.1E-01 8.7E-02 8.2E-02 6.3E-02 3.7E-02 Colfax 2.4E-01 3.0E-01 3.8E-01 3.1E-01 2.9E-01 2.7E-01 2.5E-01 Curry 8.1E-03 1.0E-02 1.7E-02 1.4E-02 1.4E-02 1.1E-02 7.8E-03 De Baca 1.1E-01 1.3E-01 1.9E-01 1.5E-01 1.4E-01 1.3E-01 1.1E-01 Dona Ana 1.0E-02 9.2E-03 3.8E-02 2.7E-02 2.3E-02 1.8E-02 1.2E-02 Eddy 1.0E-02 8.6E-03 3.7E-02 2.7E-02 2.3E-02 1.8E-02 1.2E-02 Grant 1.5E-02 1.7E-02 6.0E-02 4.7E-02 3.8E-02 2.6E-02 1.7E-02 Guadalupe 4.5E+00 5.2E+00 7.5E+00 6.0E+00 5.6E+00 5.1E+00 4.6E+00 Harding 3.0E-02 4.2E-02 5.0E-02 4.0E-02 3.7E-02 3.5E-02 3.2E-02 Hidalgo 9.2E-03 1.5E-02 1.6E-02 1.2E-02 1.1E-02 1.1E-02 1.0E-02 Lea 1.0E-02 8.6E-03 3.8E-02 2.7E-02 2.3E-02 1.8E-02 1.2E-02 Lincoln 3.3E+00 4.3E+00 5.0E+00 4.1E+00 3.8E+00 3.5E+00 3.3E+00 Luna 1.7E-02 1.4E-02 4.6E-02 3.9E-02 3.2E-02 2.4E-02 1.9E-02 McKinley 1.6E-02 1.9E-02 4.0E-02 3.3E-02 2.8E-02 2.3E-02 1.9E-02 Mora 2.5E-01 3.4E-01 4.0E-01 3.1E-01 2.9E-01 2.8E-01 2.6E-01 Otero 1.2E-02 1.5E-02 3.1E-02 2.5E-02 2.2E-02 1.8E-02 1.4E-02 Quay 2.4E-02 3.0E-02 4.1E-02 3.3E-02 3.0E-02 2.8E-02 2.5E-02 Rio Arriba 4.1E-02 5.7E-02 8.0E-02 6.3E-02 5.7E-02 5.0E-02 4.4E-02 Roosevelt 9.5E-03 1.2E-02 1.6E-02 1.3E-02 1.2E-02 1.1E-02 1.0E-02 Sandoval 4.5E-02 6.4E-02 8.6E-02 6.7E-02 6.0E-02 5.4E-02 4.8E-02 San Juan 1.3E-02 1.2E-02 3.4E-02 2.9E-02 2.4E-02 1.9E-02 1.5E-02 San Miguel 4.6E-01 6.6E-01 9.7E-01 7.7E-01 7.5E-01 6.5E-01 4.8E-01 Santa Fe 2.2E-01 2.9E-01 5.4E-01 3.8E-01 3.5E-01 3.1E-01 2.5E-01 Sierra 9.6E-03 1.1E-02 1.5E-02 1.2E-02 1.1E-02 1.1E-02 9.9E-03 Socorro 2.5E+00 4.6E+00 3.8E+00 2.9E+00 2.8E+00 2.8E+00 2.6E+00 Taos 9.5E-02 1.4E-01 1.6E-01 1.3E-01 1.2E-01 1.1E-01 1.0E-01 Torrance 8.9E+00 1.3E+01 1.4E+01 1.1E+01 1.0E+01 9.7E+00 9.2E+00 Union 6.2E-02 7.2E-02 9.7E-02 7.9E-02 7.2E-02 6.8E-02 6.7E-02 Valencia 6.9E-02 8.9E-02 1.3E-01 1.0E-01 9.3E-02 8.3E-02 7.3E-02

www.health-physics.com Estimated doses from Trinity nuclear test c S. L. SIMON ET AL. 467 Table 2E. Estimates of precinct population-weighted average radiation absorbed doses (mGy) by county and age group for Hispanics/Thyroid from all sources of internal and external exposure to fallout radionuclides (as discussed in the text). Doses rounded to two significant digits except those less than 0.01 mGy are rounded to one digit.

AGE GROUP (yrs)

COUNTY In-utero 0–11–23–78–12 13–17 Adult (18+) Bernalillo 1.9E+00 2.3E+00 1.2E+01 8.7E+00 5.8E+00 4.3E+00 2.2E+00 Catron 3.6E-01 1.2E+00 1.4E+00 1.2E+00 8.5E-01 6.8E-01 4.2E-01 Chaves 5.1E-01 3.6E-01 3.2E+00 2.2E+00 1.5E+00 1.2E+00 6.2E-01 Colfax 2.4E+00 5.7E+00 1.1E+01 9.6E+00 6.8E+00 4.9E+00 2.9E+00 Curry 1.4E-01 2.3E-01 7.1E-01 5.2E-01 3.7E-01 2.8E-01 1.6E-01 De Baca 1.3E+00 2.9E+00 5.3E+00 4.6E+00 3.3E+00 2.5E+00 1.5E+00 Dona Ana 2.9E-01 4.2E-01 1.8E+00 1.3E+00 9.1E-01 6.6E-01 3.5E-01 Eddy 2.9E-01 4.0E-01 1.8E+00 1.3E+00 9.1E-01 6.6E-01 3.5E-01 Grant 4.0E-01 1.2E+00 3.0E+00 2.7E+00 1.8E+00 1.1E+00 4.9E-01 Guadalupe 4.8E+01 1.1E+02 2.0E+02 1.7E+02 1.2E+02 9.4E+01 5.8E+01 Harding 4.5E-01 1.3E+00 1.8E+00 1.5E+00 1.1E+00 8.8E-01 5.6E-01 Hidalgo 1.6E-01 5.2E-01 5.9E-01 4.9E-01 3.6E-01 3.0E-01 1.9E-01 Lea 3.1E-01 4.3E-01 1.9E+00 1.4E+00 9.7E-01 7.0E-01 3.7E-01 Lincoln 1.3E+01 2.9E+01 4.5E+01 3.9E+01 2.9E+01 2.2E+01 1.5E+01 Luna 4.9E-01 1.3E+00 2.3E+00 2.2E+00 1.5E+00 1.0E+00 6.0E-01 McKinley 4.0E-01 1.1E+00 1.7E+00 1.6E+00 1.1E+00 8.0E-01 4.8E-01 Mora 2.8E+00 8.0E+00 1.1E+01 9.0E+00 6.7E+00 5.3E+00 3.4E+00 Otero 3.0E-01 8.2E-01 1.4E+00 1.3E+00 9.1E-01 6.3E-01 3.7E-01 Quay 2.9E-01 7.3E-01 1.2E+00 1.0E+00 7.5E-01 5.8E-01 3.6E-01 Rio Arriba 5.2E-01 1.5E+00 2.6E+00 2.2E+00 1.6E+00 1.1E+00 6.2E-01 Roosevelt 1.4E-01 3.6E-01 6.3E-01 5.7E-01 4.0E-01 2.9E-01 1.8E-01 Sandoval 5.4E-01 1.6E+00 2.4E+00 2.0E+00 1.4E+00 1.1E+00 6.4E-01 San Juan 3.5E-01 8.4E-01 1.7E+00 1.6E+00 1.1E+00 7.4E-01 4.3E-01 San Miguel 9.0E+00 1.6E+01 4.8E+01 3.5E+01 2.5E+01 2.0E+01 1.2E+01 Santa Fe 3.8E+00 1.1E+01 2.0E+01 1.4E+01 9.7E+00 7.9E+00 4.8E+00 Sierra 9.5E-02 2.1E-01 3.8E-01 3.2E-01 2.3E-01 1.8E-01 1.1E-01 Socorro 7.7E+00 2.6E+01 2.0E+01 1.4E+01 1.2E+01 1.2E+01 8.5E+00 Taos 9.5E-01 3.0E+00 4.5E+00 3.6E+00 2.6E+00 1.9E+00 1.1E+00 Torrance 4.2E+01 1.1E+02 2.0E+02 1.7E+02 1.2E+02 8.4E+01 4.8E+01 Union 5.7E-01 1.4E+00 2.6E+00 2.2E+00 1.6E+00 1.2E+00 6.9E-01 Valencia 9.9E-01 2.7E+00 4.0E+00 3.5E+00 2.5E+00 1.9E+00 1.2E+00

www.health-physics.com 468 Health Physics October 2020, Volume 119, Number 4 Table 3A. Estimates of precinct population-weighted average radiation absorbed doses (mGy) by county and age group for Native Americans/Colon from all sources of internal and external exposure to fallout radionuclides (as discussed in the text). Doses rounded to two significant digits except those less than 0.01 mGy are rounded to one digit.

AGE GROUP (yrs)

COUNTY In-utero 0–11–23–78–12 13–17 Adult (18+) Bernalillo 7.6E-02 1.4E-01 2.6E+00 1.6E+00 1.4E+00 9.1E-01 6.3E-01 Catron 1.9E-02 2.7E-02 1.3E-01 8.8E-02 7.6E-02 6.4E-02 6.1E-02 Chaves 5.3E-02 8.8E-02 7.7E-01 5.0E-01 4.4E-01 3.2E-01 2.8E-01 Colfax 3.4E-01 5.4E-01 2.0E+00 1.3E+00 1.2E+00 9.9E-01 9.6E-01 Curry 9.2E-03 1.6E-02 9.1E-02 6.3E-02 5.4E-02 4.3E-02 4.1E-02 De Baca 6.9E-02 1.0E-01 6.1E-01 4.2E-01 3.6E-01 3.0E-01 2.8E-01 Dona Ana 1.2E-02 2.4E-02 4.0E-01 2.5E-01 2.2E-01 1.4E-01 1.0E-01 Eddy 1.2E-02 2.4E-02 4.0E-01 2.5E-01 2.2E-01 1.4E-01 1.0E-01 Grant 1.4E-02 9.4E-02 3.9E-01 2.6E-01 1.9E-01 1.1E-01 8.6E-02 Guadalupe –––––– – Harding –––––– – Hidalgo 1.1E-02 9.8E-02 1.5E-01 9.2E-02 8.7E-02 7.9E-02 7.6E-02 Lea 1.2E-02 2.4E-02 4.0E-01 2.5E-01 2.2E-01 1.4E-01 1.0E-01 Lincoln 4.8E-02 1.2E+00 1.6E+00 1.1E+00 1.4E+00 8.3E-01 5.9E-01 Luna 1.6E-02 3.6E-02 7.4E-01 4.5E-01 3.9E-01 2.5E-01 1.7E-01 McKinley 1.7E-02 4.3E-02 5.2E-01 3.2E-01 2.8E-01 1.9E-01 1.4E-01 Mora –––––– – Otero 1.3E-02 8.4E-02 2.6E-01 1.8E-01 1.3E-01 7.3E-02 6.4E-02 Quay 3.3E-02 4.9E-02 3.0E-01 2.1E-01 1.8E-01 1.5E-01 1.4E-01 Rio Arriba 3.9E-02 6.5E-02 6.1E-01 4.0E-01 3.3E-01 2.3E-01 1.9E-01 Roosevelt 9.5E-03 1.6E-02 9.5E-02 6.5E-02 5.6E-02 4.5E-02 4.3E-02 Sandoval 4.7E-02 9.3E-02 7.5E-01 4.8E-01 4.2E-01 3.1E-01 2.5E-01 San Juan 1.2E-02 2.6E-02 4.8E-01 3.0E-01 2.6E-01 1.7E-01 1.2E-01 San Miguel 4.3E-01 3.0E+00 4.7E+00 3.0E+00 2.9E+00 3.0E+00 2.9E+00 Santa Fe 1.4E-01 9.4E-01 1.2E+00 7.5E-01 7.2E-01 6.7E-01 6.3E-01 Sierra 1.4E-02 3.7E-02 8.9E-02 5.7E-02 5.3E-02 4.7E-02 4.6E-02 Socorro 1.9E-01 7.5E-01 1.4E+00 9.8E-01 1.1E+00 8.3E-01 7.0E-01 Taos 1.4E-01 1.3E+00 2.2E+00 1.4E+00 1.3E+00 1.1E+00 1.0E+00 Torrance 3.0E+01 5.0E+01 8.7E+01 6.3E+01 6.1E+01 5.2E+01 4.8E+01 Union –––––– – Valencia 5.5E-02 1.1E-01 8.0E-01 5.2E-01 4.5E-01 3.3E-01 2.7E-01

www.health-physics.com Estimated doses from Trinity nuclear test c S. L. SIMON ET AL. 469 Table 3B. Estimates of precinct population-weighted average radiation absorbed doses (mGy) by county and age group for Native Americans/Lung from all sources of internal and external exposure to fallout radionuclides (as discussed in the text). Doses rounded to two significant digits except those less than 0.01 mGy are rounded to one digit. AGE GROUP (yrs)

COUNTY In-utero 0–11–23–78–12 13–17 Adult (18+) Bernalillo 6.4E-02 6.3E-02 8.1E-02 7.1E-02 7.8E-02 7.2E-02 6.2E-02 Catron 1.7E-02 2.4E-02 2.8E-02 2.6E-02 2.9E-02 2.8E-02 2.4E-02 Chaves 4.7E-02 6.5E-02 7.5E-02 6.9E-02 8.2E-02 8.0E-02 6.8E-02 Colfax 3.2E-01 5.1E-01 5.6E-01 5.3E-01 6.3E-01 6.1E-01 5.2E-01 Curry 8.1E-03 1.2E-02 1.4E-02 1.3E-02 1.5E-02 1.5E-02 1.3E-02 De Baca 5.5E-02 7.6E-02 8.5E-02 7.9E-02 9.1E-02 8.7E-02 7.5E-02 Dona Ana 1.1E-02 1.2E-02 1.6E-02 1.4E-02 1.6E-02 1.6E-02 1.3E-02 Eddy 1.1E-02 1.2E-02 1.6E-02 1.4E-02 1.6E-02 1.6E-02 1.3E-02 Grant 1.2E-02 1.6E-02 1.8E-02 1.7E-02 1.9E-02 1.8E-02 1.5E-02 Guadalupe –––––– – Harding –––––– – Hidalgo 9.1E-03 1.3E-02 1.4E-02 1.3E-02 1.6E-02 1.5E-02 1.3E-02 Lea 1.1E-02 1.2E-02 1.6E-02 1.4E-02 1.6E-02 1.6E-02 1.3E-02 Lincoln 1.1E-02 1.6E-02 1.8E-02 1.7E-02 1.9E-02 1.8E-02 1.6E-02 Luna 1.5E-02 1.3E-02 1.8E-02 1.5E-02 1.7E-02 1.6E-02 1.4E-02 McKinley 1.5E-02 1.6E-02 2.0E-02 1.8E-02 2.0E-02 1.9E-02 1.6E-02 Mora –––––– – Otero 1.1E-02 1.6E-02 1.8E-02 1.6E-02 1.9E-02 1.8E-02 1.5E-02 Quay 2.7E-02 3.8E-02 4.3E-02 4.0E-02 4.7E-02 4.5E-02 3.9E-02 Rio Arriba 3.5E-02 5.0E-02 5.7E-02 5.3E-02 6.3E-02 6.0E-02 5.2E-02 Roosevelt 8.2E-03 1.2E-02 1.4E-02 1.3E-02 1.5E-02 1.5E-02 1.3E-02 Sandoval 3.9E-02 5.0E-02 5.7E-02 5.2E-02 6.0E-02 5.7E-02 4.9E-02 San Juan 1.2E-02 1.3E-02 1.7E-02 1.5E-02 1.8E-02 1.7E-02 1.4E-02 San Miguel 3.4E-01 4.7E-01 5.1E-01 4.7E-01 5.6E-01 6.0E-01 5.1E-01 Santa Fe 1.2E-01 1.8E-01 1.9E-01 1.7E-01 2.0E-01 2.1E-01 1.8E-01 Sierra 1.2E-02 1.6E-02 1.8E-02 1.6E-02 1.8E-02 1.8E-02 1.5E-02 Socorro 1.5E-01 2.1E-01 2.2E-01 2.0E-01 2.2E-01 2.1E-01 1.8E-01 Taos 1.2E-01 1.7E-01 1.8E-01 1.7E-01 2.1E-01 2.0E-01 1.7E-01 Torrance 2.8E+01 4.2E+01 4.4E+01 4.1E+01 4.6E+01 4.4E+01 3.8E+01 Union –––––– – Valencia 4.3E-02 5.5E-02 6.1E-02 5.6E-02 6.3E-02 6.0E-02 5.2E-02

www.health-physics.com 470 Health Physics October 2020, Volume 119, Number 4 Table 3C. Estimates of precinct population-weighted average radiation absorbed doses (mGy) by county and age group for Native Americans/RBM from all sources of internal and external exposure to fallout radionuclides (as discussed in the text). Doses rounded to two significant digits except those less than 0.01 mGy are rounded to one digit.

AGE GROUP (yrs)

COUNTY In-utero 0–11–23–78–12 13–17 Adult (18+) Bernalillo 6.0E-01 1.1E-01 1.9E-01 1.3E-01 1.5E-01 1.5E-01 7.8E-02 Catron 2.8E-02 2.2E-02 2.5E-02 2.2E-02 2.3E-02 2.2E-02 1.8E-02 Chaves 1.6E-01 6.2E-02 8.1E-02 6.5E-02 7.0E-02 6.9E-02 5.0E-02 Colfax 4.5E-01 4.1E-01 4.4E-01 3.9E-01 3.9E-01 3.8E-01 3.2E-01 Curry 1.5E-02 9.9E-03 1.1E-02 1.0E-02 1.0E-02 1.0E-02 8.2E-03 De Baca 9.7E-02 6.6E-02 7.6E-02 6.6E-02 6.8E-02 6.6E-02 5.5E-02 Dona Ana 9.4E-02 1.8E-02 3.1E-02 2.1E-02 2.5E-02 2.4E-02 1.3E-02 Eddy 9.3E-02 1.8E-02 3.1E-02 2.1E-02 2.5E-02 2.4E-02 1.3E-02 Grant 7.3E-02 2.4E-02 3.9E-02 2.8E-02 2.8E-02 2.2E-02 1.4E-02 Guadalupe –––––– – Harding –––––– – Hidalgo 2.2E-02 1.7E-02 1.3E-02 1.1E-02 1.1E-02 1.2E-02 9.2E-03 Lea 9.4E-02 1.8E-02 3.1E-02 2.1E-02 2.5E-02 2.4E-02 1.3E-02 Lincoln 1.4E-02 1.4E-02 2.3E-02 2.0E-02 1.7E-02 1.3E-02 1.1E-02 Luna 2.4E-01 3.4E-02 6.6E-02 4.0E-02 5.0E-02 5.0E-02 2.2E-02 McKinley 1.2E-01 2.4E-02 4.0E-02 2.7E-02 3.2E-02 3.1E-02 1.7E-02 Mora –––––– – Otero 3.3E-02 1.9E-02 2.5E-02 2.0E-02 1.9E-02 1.6E-02 1.2E-02 Quay 4.8E-02 3.3E-02 3.7E-02 3.3E-02 3.4E-02 3.3E-02 2.7E-02 Rio Arriba 1.2E-01 4.7E-02 6.5E-02 5.2E-02 5.4E-02 5.1E-02 3.7E-02 Roosevelt 1.6E-02 1.0E-02 1.2E-02 1.0E-02 1.0E-02 1.0E-02 8.3E-03 Sandoval 1.5E-01 5.4E-02 7.2E-02 5.6E-02 6.1E-02 5.9E-02 4.1E-02 San Juan 1.1E-01 2.0E-02 3.6E-02 2.4E-02 2.8E-02 2.8E-02 1.4E-02 San Miguel 7.5E-01 5.7E-01 4.8E-01 4.1E-01 4.4E-01 5.1E-01 4.0E-01 Santa Fe 2.3E-01 2.1E-01 1.8E-01 1.5E-01 1.6E-01 1.7E-01 1.4E-01 Sierra 1.8E-02 1.6E-02 1.6E-02 1.4E-02 1.4E-02 1.4E-02 1.2E-02 Socorro 1.9E-01 1.9E-01 2.0E-01 1.8E-01 1.8E-01 1.7E-01 1.5E-01 Taos 3.7E-01 2.2E-01 1.9E-01 1.5E-01 1.6E-01 1.7E-01 1.2E-01 Torrance 3.2E+01 3.7E+01 3.7E+01 3.4E+01 3.4E+01 3.2E+01 2.8E+01 Union –––––– – Valencia 1.6E-01 6.0E-02 7.8E-02 6.2E-02 6.6E-02 6.4E-02 4.6E-02

www.health-physics.com Estimated doses from Trinity nuclear test c S. L. SIMON ET AL. 471 Table 3D. Estimates of precinct population-weighted average radiation absorbed doses (mGy) by county and age group for Native Americans/Stomach from all sources of internal and external exposure to fallout radionuclides (as discussed in the text). Doses rounded to two significant digits except those less than 0.01 mGy are rounded to one digit.

AGE GROUP (yrs)

COUNTY In-utero 0–11–23–78–12 13–17 Adult (18+) Bernalillo 7.5E-02 6.6E-02 2.1E-01 1.4E-01 1.3E-01 1.1E-01 8.2E-02 Catron 2.6E-02 2.3E-02 4.3E-02 3.4E-02 3.1E-02 2.9E-02 2.6E-02 Chaves 5.2E-02 5.6E-02 9.6E-02 7.5E-02 7.1E-02 6.4E-02 5.6E-02 Colfax 3.4E-01 4.1E-01 4.9E-01 4.2E-01 4.1E-01 3.8E-01 3.5E-01 Curry 9.0E-03 1.0E-02 1.5E-02 1.2E-02 1.1E-02 1.1E-02 9.5E-03 De Baca 6.7E-02 6.9E-02 1.1E-01 8.6E-02 8.1E-02 7.5E-02 6.9E-02 Dona Ana 1.2E-02 1.1E-02 3.2E-02 2.2E-02 2.0E-02 1.7E-02 1.3E-02 Eddy 1.2E-02 1.1E-02 3.2E-02 2.2E-02 2.0E-02 1.7E-02 1.3E-02 Grant 1.3E-02 1.9E-02 3.3E-02 2.4E-02 2.1E-02 1.7E-02 1.4E-02 Guadalupe –––––– – Harding –––––– – Hidalgo 1.1E-02 1.7E-02 1.8E-02 1.4E-02 1.3E-02 1.3E-02 1.2E-02 Lea 1.2E-02 1.1E-02 3.2E-02 2.2E-02 2.0E-02 1.7E-02 1.3E-02 Lincoln 1.2E-02 1.4E-02 2.4E-02 2.0E-02 1.6E-02 1.3E-02 1.2E-02 Luna 1.5E-02 1.3E-02 4.8E-02 3.2E-02 2.8E-02 2.3E-02 1.7E-02 McKinley 1.6E-02 1.6E-02 4.3E-02 3.0E-02 2.7E-02 2.3E-02 1.8E-02 Mora –––––– – Otero 1.2E-02 1.8E-02 2.7E-02 2.1E-02 1.8E-02 1.5E-02 1.3E-02 Quay 3.2E-02 3.4E-02 5.1E-02 4.2E-02 3.9E-02 3.6E-02 3.3E-02 Rio Arriba 3.8E-02 4.2E-02 7.3E-02 5.7E-02 5.3E-02 4.7E-02 4.1E-02 Roosevelt 9.2E-03 1.0E-02 1.5E-02 1.2E-02 1.2E-02 1.1E-02 9.8E-03 Sandoval 4.5E-02 4.7E-02 8.7E-02 6.6E-02 6.1E-02 5.5E-02 4.8E-02 San Juan 1.2E-02 1.1E-02 3.6E-02 2.4E-02 2.2E-02 1.8E-02 1.4E-02 San Miguel 4.1E-01 6.1E-01 6.8E-01 5.2E-01 5.2E-01 5.7E-01 5.1E-01 Santa Fe 1.4E-01 2.1E-01 2.1E-01 1.7E-01 1.7E-01 1.7E-01 1.6E-01 Sierra 1.3E-02 1.6E-02 2.0E-02 1.6E-02 1.6E-02 1.5E-02 1.4E-02 Socorro 1.9E-01 2.1E-01 2.7E-01 2.2E-01 2.2E-01 2.0E-01 1.9E-01 Taos 1.4E-01 2.1E-01 2.4E-01 1.8E-01 1.8E-01 1.7E-01 1.5E-01 Torrance 3.0E+01 3.8E+01 4.1E+01 3.6E+01 3.6E+01 3.3E+01 3.0E+01 Union –––––– – Valencia 5.4E-02 5.5E-02 1.0E-01 7.7E-02 7.1E-02 6.4E-02 5.6E-02

www.health-physics.com 472 Health Physics October 2020, Volume 119, Number 4 Table 3E. Estimates of precinct population-weighted average radiation absorbed doses (mGy) by county and age group for Native Americans/Thyroid from all sources of internal and external exposure to fallout radionuclides (as discussed in the text). Doses rounded to two significant digits except those less than 0.01 mGy are rounded to one digit.

AGE GROUP (yrs)

COUNTY In-utero 0–11–23–78–12 13–17 Adult (18+) Bernalillo 1.5E+00 3.5E+00 9.2E+00 6.3E+00 4.6E+00 3.4E+00 1.8E+00 Catron 2.6E-01 4.9E-01 1.0E+00 7.5E-01 5.2E-01 4.4E-01 3.1E-01 Chaves 7.1E-01 1.6E+00 3.1E+00 2.2E+00 1.6E+00 1.3E+00 8.6E-01 Colfax 2.2E+00 4.9E+00 1.0E+01 7.8E+00 5.0E+00 3.7E+00 2.7E+00 Curry 1.4E-01 3.2E-01 6.2E-01 4.5E-01 3.2E-01 2.6E-01 1.7E-01 De Baca 8.2E-01 1.7E+00 3.2E+00 2.3E+00 1.7E+00 1.4E+00 9.8E-01 Dona Ana 2.6E-01 5.9E-01 1.5E+00 1.0E+00 7.5E-01 5.6E-01 3.1E-01 Eddy 2.6E-01 5.9E-01 1.5E+00 1.0E+00 7.5E-01 5.7E-01 3.1E-01 Grant 1.8E-01 5.7E-01 1.2E+00 9.3E-01 5.6E-01 3.5E-01 2.1E-01 Guadalupe –––––– – Harding –––––– – Hidalgo 2.0E-01 6.3E-01 7.4E-01 5.1E-01 3.9E-01 3.5E-01 2.4E-01 Lea 2.7E-01 6.2E-01 1.6E+00 1.1E+00 7.9E-01 6.0E-01 3.2E-01 Lincoln 5.6E-02 1.1E-01 6.7E-01 5.7E-01 2.6E-01 8.3E-02 6.6E-02 Luna 3.9E-01 9.7E-01 2.4E+00 1.7E+00 1.2E+00 9.2E-01 4.7E-01 McKinley 3.3E-01 7.6E-01 1.8E+00 1.3E+00 9.2E-01 7.1E-01 4.0E-01 Mora –––––– – Otero 1.3E-01 4.3E-01 9.2E-01 7.0E-01 4.0E-01 2.4E-01 1.6E-01 Quay 4.1E-01 8.8E-01 1.6E+00 1.2E+00 8.3E-01 7.1E-01 4.9E-01 Rio Arriba 4.8E-01 1.1E+00 2.4E+00 1.8E+00 1.2E+00 8.9E-01 5.7E-01 Roosevelt 1.4E-01 3.3E-01 6.5E-01 4.6E-01 3.4E-01 2.7E-01 1.8E-01 Sandoval 6.2E-01 1.4E+00 2.8E+00 2.0E+00 1.4E+00 1.1E+00 7.3E-01 San Juan 3.0E-01 6.9E-01 1.8E+00 1.2E+00 9.0E-01 6.7E-01 3.6E-01 San Miguel 7.0E+00 2.1E+01 3.2E+01 2.2E+01 1.7E+01 1.6E+01 1.0E+01 Santa Fe 1.2E+00 4.1E+00 4.0E+00 2.7E+00 2.0E+00 2.0E+00 1.5E+00 Sierra 1.0E-01 2.5E-01 4.1E-01 3.0E-01 2.0E-01 1.7E-01 1.2E-01 Socorro 8.1E-01 1.6E+00 2.7E+00 1.9E+00 1.4E+00 1.3E+00 9.2E-01 Taos 2.5E+00 7.9E+00 9.3E+00 6.4E+00 4.9E+00 4.4E+00 3.0E+00 Torrance 7.7E+01 1.4E+02 3.5E+02 2.7E+02 1.6E+02 1.0E+02 8.4E+01 Union –––––– – Valencia 7.2E-01 1.6E+00 3.4E+00 2.4E+00 1.7E+00 1.4E+00 8.5E-01

www.health-physics.com Estimated doses from Trinity nuclear test c S. L. SIMON ET AL. 473 Table 4A. Estimates of precinct population-weighted average radiation absorbed doses (mGy) by county and age group for African Americans/Colon from all sources of internal and external exposure to fallout radionuclides (as discussed in the text). Doses rounded to two significant digits except those less than 0.01 mGy are rounded to one digit.

AGE GROUP (yrs)

COUNTY In-utero 0–11–23–78–12 13–17 Adult (18+) Bernalillo 2.0E-02 1.5E-01 1.5E-01 9.0E-02 9.3E-02 9.0E-02 8.7E-02 Catron 4.3E-02 5.9E-02 8.3E-01 4.8E-01 4.1E-01 2.9E-01 2.0E-01 Chaves 2.7E-01 4.9E-01 1.7E+00 1.3E+00 1.2E+00 9.4E-01 8.2E-01 Colfax 7.0E-03 1.1E-02 5.5E-02 3.9E-02 3.6E-02 2.7E-02 2.1E-02 Curry –––––– – De Baca 1.2E-02 4.0E-02 4.7E-01 3.3E-01 2.7E-01 1.7E-01 1.2E-01 Dona Ana 1.1E-02 1.6E-02 5.6E-01 3.3E-01 2.7E-01 1.8E-01 1.1E-01 Eddy 1.4E-02 1.1E-01 5.1E-01 4.1E-01 3.4E-01 2.0E-01 1.4E-01 Grant 2.6E+00 3.9E+00 2.0E+01 1.4E+01 1.4E+01 1.1E+01 9.6E+00 Guadalupe 2.5E-02 4.3E-02 1.9E-01 1.4E-01 1.3E-01 1.0E-01 8.4E-02 Harding 9.9E-03 9.7E-02 1.1E-01 7.3E-02 7.4E-02 6.6E-02 6.0E-02 Hidalgo 1.1E-02 1.4E-02 6.0E-01 3.3E-01 2.7E-01 1.8E-01 1.1E-01 Lea 2.9E-01 4.6E-01 1.4E+00 1.1E+00 9.7E-01 7.3E-01 6.2E-01 Lincoln 1.7E-02 4.6E-02 7.0E-01 6.1E-01 4.9E-01 2.9E-01 2.2E-01 Luna 1.6E-02 1.1E-01 5.0E-01 4.2E-01 3.5E-01 2.2E-01 1.7E-01 McKinley 3.2E-01 5.4E-01 2.5E+00 1.8E+00 1.7E+00 1.3E+00 1.1E+00 Mora 1.3E-02 8.4E-02 4.0E-01 3.3E-01 2.8E-01 1.8E-01 1.4E-01 Otero 2.7E-02 4.4E-02 1.9E-01 1.4E-01 1.3E-01 1.0E-01 8.5E-02 Quay 2.2E-02 4.1E-02 4.2E-01 3.3E-01 2.6E-01 1.4E-01 8.3E-02 Rio Arriba 8.1E-03 1.5E-02 6.4E-02 4.7E-02 4.4E-02 3.3E-02 2.7E-02 Roosevelt 4.5E-02 1.3E-01 6.7E-01 5.3E-01 4.6E-01 3.1E-01 2.4E-01 Sandoval 1.2E-02 3.0E-02 4.4E-01 3.8E-01 3.1E-01 1.9E-01 1.4E-01 San Juan 4.8E-01 4.4E+00 5.6E+00 3.5E+00 3.6E+00 3.3E+00 2.8E+00 San Miguel 2.1E-01 1.4E+00 5.8E+00 3.7E+00 3.2E+00 2.3E+00 1.7E+00 Santa Fe 1.0E-02 1.6E-02 8.1E-02 5.7E-02 5.4E-02 4.4E-02 3.7E-02 Sierra 1.4E-01 2.1E-01 4.4E-01 3.1E-01 3.1E-01 2.8E-01 2.7E-01 Socorro 5.5E-02 9.9E-02 5.4E-01 4.1E-01 3.4E-01 2.1E-01 1.6E-01 Taos –––––– – Torrance 3.3E-02 7.8E-02 2.7E-01 1.9E-01 1.8E-01 1.4E-01 1.2E-01 Union 4.4E-02 7.4E-02 6.4E-01 5.2E-01 4.4E-01 2.9E-01 2.2E-01 Valencia 2.0E-02 1.5E-01 1.5E-01 9.0E-02 9.3E-02 9.0E-02 8.7E-02

www.health-physics.com 474 Health Physics October 2020, Volume 119, Number 4 Table 4B. Estimates of precinct population-weighted average radiation absorbed doses (mGy) by county and age group for African Americans/Lung from all sources of internal and external exposure to fallout radionuclides (as discussed in the text). Doses rounded to two significant digits except those less than 0.01 mGy are rounded to one digit.

AGE GROUP (yrs)

COUNTY In-utero 0–11–23–78–12 13–17 Adult (18+) Bernalillo 7.6E-02 6.9E-02 1.0E-01 8.6E-02 9.9E-02 9.0E-02 7.1E-02 Catron 1.8E-02 2.7E-02 3.0E-02 2.6E-02 2.8E-02 2.7E-02 2.4E-02 Chaves 4.1E-02 5.9E-02 7.0E-02 6.8E-02 8.5E-02 8.0E-02 6.4E-02 Colfax 2.5E-01 4.3E-01 5.3E-01 4.7E-01 5.7E-01 5.5E-01 4.8E-01 Curry 6.7E-03 1.1E-02 1.2E-02 1.2E-02 1.6E-02 1.5E-02 1.2E-02 De Baca –––––– – Dona Ana 1.2E-02 1.2E-02 1.6E-02 1.4E-02 1.6E-02 1.5E-02 1.3E-02 Eddy 1.1E-02 1.1E-02 1.6E-02 1.4E-02 1.7E-02 1.6E-02 1.3E-02 Grant 1.3E-02 1.3E-02 1.6E-02 1.4E-02 1.6E-02 1.5E-02 1.3E-02 Guadalupe 2.2E+00 3.1E+00 3.7E+00 3.2E+00 3.5E+00 3.4E+00 3.0E+00 Harding 2.3E-02 3.7E-02 4.5E-02 4.0E-02 4.7E-02 4.5E-02 3.9E-02 Hidalgo 8.3E-03 1.2E-02 1.4E-02 1.3E-02 1.5E-02 1.4E-02 1.3E-02 Lea 1.1E-02 1.1E-02 1.6E-02 1.4E-02 1.7E-02 1.6E-02 1.3E-02 Lincoln 2.7E-01 3.9E-01 4.5E-01 4.0E-01 4.2E-01 4.1E-01 3.6E-01 Luna 1.8E-02 1.3E-02 1.8E-02 1.6E-02 1.7E-02 1.6E-02 1.4E-02 McKinley 1.5E-02 1.4E-02 1.8E-02 1.6E-02 1.7E-02 1.6E-02 1.4E-02 Mora 2.9E-01 4.7E-01 5.6E-01 5.0E-01 5.8E-01 5.7E-01 4.9E-01 Otero 1.2E-02 1.3E-02 1.6E-02 1.4E-02 1.6E-02 1.5E-02 1.3E-02 Quay 2.4E-02 3.7E-02 4.4E-02 3.9E-02 4.4E-02 4.2E-02 3.7E-02 Rio Arriba 2.2E-02 3.6E-02 4.5E-02 4.0E-02 4.8E-02 4.6E-02 4.0E-02 Roosevelt 7.5E-03 1.2E-02 1.4E-02 1.3E-02 1.5E-02 1.4E-02 1.2E-02 Sandoval 4.0E-02 5.1E-02 6.2E-02 5.5E-02 6.0E-02 5.7E-02 5.1E-02 San Juan 1.2E-02 1.2E-02 1.6E-02 1.4E-02 1.6E-02 1.5E-02 1.3E-02 San Miguel 4.0E-01 5.7E-01 6.5E-01 6.3E-01 7.7E-01 7.5E-01 5.9E-01 Santa Fe 1.4E-01 2.6E-01 4.8E-01 3.8E-01 4.0E-01 3.2E-01 2.3E-01 Sierra 8.9E-03 1.3E-02 1.5E-02 1.3E-02 1.5E-02 1.4E-02 1.3E-02 Socorro 1.4E-01 2.0E-01 3.6E-01 2.7E-01 2.9E-01 2.9E-01 2.8E-01 Taos 5.5E-02 1.0E-01 5.5E-01 4.3E-01 3.8E-01 2.6E-01 2.0E-01 Torrance –––––– – Union 4.0E-02 5.9E-02 7.9E-02 7.1E-02 8.5E-02 8.3E-02 7.1E-02 Valencia 3.9E-02 4.9E-02 9.3E-02 7.8E-02 7.6E-02 6.2E-02 5.0E-02

www.health-physics.com Estimated doses from Trinity nuclear test c S. L. SIMON ET AL. 475 Table 4C. Estimates of precinct population-weighted average radiation absorbed doses (mGy) by county and age group for African Americans/RBM from all sources of internal and external exposure to fallout radionuclides (as discussed in the text). Doses rounded to two significant digits except those less than 0.01 mGy are rounded to one digit.

AGE GROUP (yrs)

COUNTY In-utero 0–11–23–78–12 13–17 Adult (18+) Bernalillo 8.9E-01 6.4E-02 3.1E-01 1.8E-01 2.0E-01 2.1E-01 9.7E-02 Catron 3.4E-02 3.6E-02 2.7E-02 2.2E-02 2.2E-02 2.2E-02 1.9E-02 Chaves 1.8E-01 4.6E-02 8.9E-02 6.7E-02 7.5E-02 7.2E-02 4.5E-02 Colfax 4.2E-01 3.2E-01 3.8E-01 3.2E-01 3.1E-01 3.0E-01 2.6E-01 Curry 1.1E-02 8.2E-03 9.7E-03 9.0E-03 9.9E-03 9.2E-03 6.9E-03 De Baca –––––– – Dona Ana 1.3E-01 1.9E-02 3.7E-02 2.6E-02 2.9E-02 2.9E-02 1.5E-02 Eddy 1.3E-01 1.2E-02 4.2E-02 2.7E-02 3.0E-02 3.0E-02 1.5E-02 Grant 1.8E-01 3.2E-02 4.7E-02 3.7E-02 4.2E-02 3.7E-02 1.8E-02 Guadalupe 3.9E+00 2.7E+00 3.3E+00 2.7E+00 2.7E+00 2.6E+00 2.2E+00 Harding 3.9E-02 2.9E-02 3.5E-02 2.9E-02 2.9E-02 2.8E-02 2.3E-02 Hidalgo 1.9E-02 1.6E-02 1.3E-02 1.0E-02 1.0E-02 1.1E-02 8.5E-03 Lea 1.3E-01 1.0E-02 4.5E-02 2.7E-02 3.0E-02 3.1E-02 1.5E-02 Lincoln 4.2E-01 3.5E-01 4.1E-01 3.5E-01 3.3E-01 3.1E-01 2.7E-01 Luna 3.6E-01 4.6E-02 6.5E-02 5.4E-02 6.2E-02 5.8E-02 2.8E-02 McKinley 1.7E-01 3.2E-02 3.9E-02 3.2E-02 3.5E-02 3.3E-02 1.9E-02 Mora 5.0E-01 3.7E-01 4.5E-01 3.7E-01 3.6E-01 3.5E-01 3.0E-01 Otero 1.4E-01 2.6E-02 3.2E-02 2.6E-02 2.9E-02 2.7E-02 1.6E-02 Quay 4.0E-02 3.1E-02 3.7E-02 3.1E-02 3.0E-02 2.9E-02 2.5E-02 Rio Arriba 7.4E-02 3.0E-02 5.2E-02 4.1E-02 4.1E-02 3.5E-02 2.3E-02 Roosevelt 1.3E-02 9.5E-03 1.1E-02 9.6E-03 9.3E-03 9.0E-03 7.6E-03 Sandoval 1.9E-01 6.1E-02 7.7E-02 6.3E-02 6.5E-02 6.2E-02 4.4E-02 San Juan 1.6E-01 2.4E-02 3.5E-02 2.9E-02 3.3E-02 3.0E-02 1.6E-02 San Miguel 9.4E-01 7.1E-01 5.9E-01 5.2E-01 6.0E-01 6.1E-01 4.2E-01 Santa Fe 4.3E-01 2.3E-01 3.6E-01 3.0E-01 3.1E-01 2.7E-01 1.3E-01 Sierra 1.5E-02 1.1E-02 1.3E-02 1.1E-02 1.1E-02 1.1E-02 9.0E-03 Socorro 1.5E-01 1.7E-01 1.9E-01 1.6E-01 1.5E-01 1.4E-01 1.3E-01 Taos 1.2E-01 7.2E-02 9.9E-02 8.1E-02 7.8E-02 7.1E-02 5.6E-02 Torrance –––––– – Union 8.4E-02 6.1E-02 7.5E-02 6.2E-02 6.1E-02 5.9E-02 5.0E-02 Valencia 1.9E-01 5.7E-02 7.5E-02 6.3E-02 6.4E-02 6.0E-02 4.3E-02

www.health-physics.com 476 Health Physics October 2020, Volume 119, Number 4 Table 4D. Estimates of precinct population-weighted average radiation absorbed doses (mGy) by county and age group for African Americans/Stomach from all sources of internal and external exposure to fallout radionuclides (as discussed in the text). Doses rounded to two significant digits except those less than 0.01 mGy are rounded to one digit.

AGE GROUP (yrs)

COUNTY In-utero 0–11–23–78–12 13–17 Adult (18+) Bernalillo 7.6E-02 6.0E-02 2.9E-01 1.8E-01 1.6E-01 1.3E-01 8.9E-02 Catron 2.3E-02 4.1E-02 3.6E-02 2.7E-02 2.6E-02 2.6E-02 2.4E-02 Chaves 4.2E-02 4.6E-02 9.1E-02 7.0E-02 6.9E-02 6.1E-02 4.6E-02 Colfax 2.6E-01 3.3E-01 4.3E-01 3.5E-01 3.2E-01 3.0E-01 2.8E-01 Curry 7.0E-03 8.3E-03 1.1E-02 1.0E-02 1.0E-02 9.3E-03 7.3E-03 De Baca –––––– – Dona Ana 1.2E-02 1.2E-02 3.5E-02 2.6E-02 2.2E-02 1.8E-02 1.4E-02 Eddy 1.1E-02 9.2E-03 4.0E-02 2.6E-02 2.3E-02 1.9E-02 1.3E-02 Grant 1.3E-02 1.7E-02 3.6E-02 2.9E-02 2.4E-02 1.9E-02 1.5E-02 Guadalupe 2.5E+00 2.8E+00 4.2E+00 3.3E+00 3.1E+00 2.9E+00 2.6E+00 Harding 2.5E-02 3.0E-02 4.1E-02 3.3E-02 3.1E-02 2.8E-02 2.6E-02 Hidalgo 9.5E-03 1.6E-02 1.6E-02 1.2E-02 1.2E-02 1.2E-02 1.0E-02 Lea 1.1E-02 8.8E-03 4.2E-02 2.6E-02 2.3E-02 1.9E-02 1.3E-02 Lincoln 3.0E-01 3.6E-01 4.8E-01 3.9E-01 3.6E-01 3.3E-01 3.0E-01 Luna 1.7E-02 1.4E-02 4.6E-02 3.9E-02 3.2E-02 2.4E-02 1.9E-02 McKinley 1.6E-02 1.9E-02 4.0E-02 3.2E-02 2.8E-02 2.3E-02 1.9E-02 Mora 3.1E-01 3.7E-01 5.2E-01 4.2E-01 3.9E-01 3.6E-01 3.3E-01 Otero 1.3E-02 1.6E-02 3.2E-02 2.6E-02 2.3E-02 1.9E-02 1.5E-02 Quay 2.7E-02 3.2E-02 4.4E-02 3.6E-02 3.3E-02 3.1E-02 2.8E-02 Rio Arriba 2.2E-02 2.7E-02 5.0E-02 4.0E-02 3.4E-02 2.8E-02 2.3E-02 Roosevelt 7.9E-03 9.6E-03 1.3E-02 1.1E-02 1.0E-02 9.2E-03 8.3E-03 Sandoval 4.5E-02 5.1E-02 8.7E-02 7.0E-02 6.3E-02 5.5E-02 4.8E-02 San Juan 1.2E-02 1.1E-02 3.3E-02 2.8E-02 2.3E-02 1.8E-02 1.4E-02 San Miguel 4.6E-01 7.6E-01 7.9E-01 6.4E-01 6.8E-01 6.6E-01 5.3E-01 Santa Fe 1.9E-01 2.4E-01 6.4E-01 4.6E-01 4.1E-01 3.3E-01 2.1E-01 Sierra 1.0E-02 1.1E-02 1.7E-02 1.4E-02 1.3E-02 1.2E-02 1.1E-02 Socorro 1.4E-01 1.7E-01 2.1E-01 1.7E-01 1.6E-01 1.5E-01 1.4E-01 Taos 5.5E-02 6.9E-02 1.0E-01 8.1E-02 7.2E-02 6.3E-02 5.7E-02 Torrance –––––– – Union 5.9E-02 6.9E-02 9.7E-02 7.9E-02 7.4E-02 6.8E-02 6.1E-02 Valencia 4.4E-02 4.6E-02 8.6E-02 7.0E-02 6.2E-02 5.3E-02 4.6E-02

www.health-physics.com Estimated doses from Trinity nuclear test c S. L. SIMON ET AL. 477 Table 4E. Estimates of precinct population-weighted average radiation absorbed doses (mGy) by county and age group for African Americans/Thyroid from all sources of internal and external exposure to fallout radionuclides (as discussed in the text). Doses rounded to two significant digits except those less than 0.01 mGy are rounded to one digit.

AGE GROUP (yrs)

COUNTY In-utero 0–11–23–78–12 13–17 Adult (18+) Bernalillo 1.8E+00 5.0E-01 1.5E+01 8.8E+00 5.9E+00 4.5E+00 2.2E+00 Catron 2.6E-01 1.0E+00 1.1E+00 8.4E-01 6.2E-01 5.1E-01 3.1E-01 Chaves 5.1E-01 1.3E-01 3.4E+00 2.2E+00 1.5E+00 1.2E+00 6.1E-01 Colfax 3.2E+00 8.0E+00 1.5E+01 1.4E+01 9.6E+00 6.8E+00 4.0E+00 Curry 1.0E-01 2.7E-02 7.4E-01 4.6E-01 3.2E-01 2.5E-01 1.3E-01 De Baca –––––– – Dona Ana 3.0E-01 5.4E-01 1.7E+00 1.4E+00 9.2E-01 6.7E-01 3.7E-01 Eddy 2.9E-01 2.2E-01 2.1E+00 1.4E+00 9.2E-01 7.0E-01 3.5E-01 Grant 3.1E-01 9.3E-01 1.6E+00 1.5E+00 1.0E+00 6.9E-01 3.7E-01 Guadalupe 3.1E+01 7.0E+01 1.2E+02 1.0E+02 7.4E+01 5.8E+01 3.7E+01 Harding 2.8E-01 6.7E-01 1.2E+00 1.0E+00 7.3E-01 5.6E-01 3.4E-01 Hidalgo 1.7E-01 5.8E-01 6.2E-01 5.1E-01 3.8E-01 3.2E-01 2.1E-01 Lea 3.1E-01 1.4E-01 2.4E+00 1.5E+00 9.9E-01 7.6E-01 3.7E-01 Lincoln 1.6E+00 3.5E+00 6.7E+00 5.8E+00 4.2E+00 3.0E+00 1.9E+00 Luna 4.9E-01 1.3E+00 2.3E+00 2.2E+00 1.5E+00 1.0E+00 6.0E-01 McKinley 4.1E-01 1.1E+00 1.8E+00 1.6E+00 1.1E+00 8.1E-01 4.9E-01 Mora 3.4E+00 8.0E+00 1.4E+01 1.2E+01 8.6E+00 6.7E+00 4.2E+00 Otero 3.4E-01 9.2E-01 1.5E+00 1.4E+00 9.6E-01 6.9E-01 4.1E-01 Quay 2.9E-01 6.7E-01 1.2E+00 1.0E+00 7.4E-01 5.7E-01 3.5E-01 Rio Arriba 2.1E-01 4.8E-01 1.8E+00 1.5E+00 1.0E+00 5.9E-01 2.5E-01 Roosevelt 1.2E-01 3.0E-01 5.4E-01 4.9E-01 3.4E-01 2.5E-01 1.5E-01 Sandoval 5.8E-01 1.4E+00 2.5E+00 2.2E+00 1.5E+00 1.1E+00 6.9E-01 San Juan 3.5E-01 8.3E-01 1.6E+00 1.6E+00 1.1E+00 7.3E-01 4.3E-01 San Miguel 1.0E+01 1.2E+01 5.9E+01 3.7E+01 2.7E+01 2.3E+01 1.3E+01 Santa Fe 4.1E+00 1.1E+01 2.3E+01 1.6E+01 1.1E+01 8.7E+00 5.1E+00 Sierra 1.2E-01 2.6E-01 4.6E-01 3.8E-01 2.8E-01 2.2E-01 1.4E-01 Socorro 4.4E-01 8.2E-01 1.4E+00 1.2E+00 8.6E-01 7.0E-01 4.8E-01 Taos 4.3E-01 1.0E+00 3.0E+00 2.5E+00 1.7E+00 1.1E+00 5.2E-01 Torrance –––––– – Union 7.1E-01 1.7E+00 2.9E+00 2.6E+00 1.9E+00 1.4E+00 8.7E-01 Valencia 5.9E-01 1.3E+00 2.7E+00 2.4E+00 1.7E+00 1.2E+00 7.1E-01

www.health-physics.com Paper

Projected Cancer Risks to Residents of New Mexico from Exposure to Trinity Radioactive Fallout

Elizabeth K. Cahoon,1 Rui Zhang,1 Steven L. Simon,1 André Bouville,2 and Ruth M. Pfeiffer1

90% UIs overlapped for all race/ethnicity groups for each cancer Abstract—The Trinity nuclear test, conducted in 1945, exposed grouping. Thus, most cancers that have occurred or will occur residents of New Mexico to varying degrees of radioactive fallout. among persons exposed to Trinity fallout are likely to be cancers Companion papers in this issue have detailed the results of a dose unrelated to exposures from the Trinity nuclear test. While these reconstruction that has estimated tissue-specific radiation absorbed ranges are based on the most detailed dose reconstruction to date doses to residents of New Mexico from internal and external exposure and rely largely on methods previously established through scientific to radioactive fallout in the first year following the Trinity test committee agreement, challenges inherent in the dose estimation, and when more than 90% of the lifetime dose was received. Estimated assumptionsrelieduponbothintheriskprojectionandincorporation radiation doses depended on geographic location, race/ethnicity, of uncertainty are important limitations in quantifying the range and age at the time of the test. Here, these doses were applied to of radiation-related excess cancer risk. sex- and organ-specific risk coefficients (without applying a dose Health Phys. 119(4):478–493; 2020 and dose rate effectiveness factor to extrapolate from a population Key words: cancer; fallout; nuclear weapons; radiation risk with high-dose/high-dose rates to those with low-dose/low-dose rates) and combined with baseline cancer rates and published life tables to estimate and project the range of radiation-related excess cancers among 581,489 potentially exposed residents of INTRODUCTION New Mexico. The total lifetime baseline number of all solid cancers [excluding thyroid and non-melanoma skin cancer (NMSC)] was THE TRINITY nuclear test was conducted on 16 July 1945 in estimated to be 183,000 from 1945 to 2034. Estimates of ranges south-central New Mexico as the culmination of the Manhattan of numbers of radiation-related excess cancers and corresponding Project, just 3 wk before the atomic bombings of Hiroshima attributable fractions from 1945 to 2034 incorporate various sources of uncertainty. We estimated 90% uncertainty intervals and Nagasaki, Japan. The device was similar to the Fat (UIs) of excess cancer cases to be 210 to 460 for all solid cancers Man-type plutonium implosion device used in the bombing (except thyroid cancer and NMSC), 80 to 530 for thyroid cancer, of Nagasaki. The test resulted in varying levels of radiation and up to 10 for leukemia (except chronic lymphocytic leukemia), dose from radioactive fallout to residents of New Mexico with corresponding attributable fractions ranging from 0.12% to 0.25%, 3.6% to 20%, and 0.02% to 0.31%, respectively. In the depending on geographic location of residence, age at the counties of Guadalupe, Lincoln, San Miguel, Socorro, and Torrance, time of the test, and race/ethnicity. Because the test occurred which received the greatest fallout deposition, the 90% UI for the over 70 y ago, it is not possible to retrospectively identify projected fraction of thyroid cancers attributable to radioactive the population exposed and collect statewide records of the fallout from the Trinity test was estimated to be from 17% to 58%. Attributable fractions for cancer types varied by race/ethnicity, but number and types of cancers that occurred in that population. For that reason, conducting an analytical epidemiological 1Division of Cancer Epidemiology & Genetics, National Cancer In- follow-up study of exposed individuals is not feasible. stitute, National Institutes of Health, Bethesda, MD; 2Retired (NCI/NIH). Recognizing, however, that ionizing radiation is an estab- For correspondence contact Elizabeth K. Cahoon, Radiation Epidemiol- ogy Branch, DCEG, National Cancer Institute, NIH, DHHS, 9609 Medical lished carcinogen, exposure of the New Mexico population Center Drive, Rm 7E452, MS 9778, Bethesda, MD 20892-9778, or email could lead to an increase in cancer incidence. Information at [email protected]; or Ruth M. Pfeiffer, Biostatistics Branch, that quantitatively relates radiation dose levels to cancer DCEG, National Cancer Institute, NIH, DHHS, 9609 Medical Center Dr., Rm 7E142, Bethesda, MD 20892-9778 or email at [email protected]. risk is required to provide an estimate (and/or projection) of (Manuscript accepted 15 June 2020) such an increase. Today, the best quantitative evaluation of The authors declare no conflicts of interest. 0017-9078/20/0 radiation-related cancer risk as a function of radiation dose Written work prepared by employees of the Federal Government as is based largely on epidemiological studies of the Japanese part of their official duties is, under the U.S. Copyright Act, a "work of atomic bomb survivors who were acutely exposed to low the United States Government" for which copyright protection under Title 17 of the United States Code is not available. As such, copyright does not to moderate/high external doses, in addition to some other extend to the contributions of employees of the Federal Government. radiation-exposed populations (Ron et al. 1995; NRC 2005; DOI: 10.1097/HP.0000000000001333 Preston et al. 2007). Cancer risk models developed using 478 www.health-physics.com Projected lifetime cancer risks c E. K. CAHOON ET AL. 479 datafromtheJapaneseatomicbombsurvivorscanbeusedto Table 1. Distribution of the New Mexico population by age and race/ estimate or project risks in other populations with different ethnicity in 1945. exposure scenarios; e.g., the residents of New Mexico alive Race/ethnicity at the time of the Trinity test. This extrapolation from one Native African population to another adds uncertainty in estimating the Age group (y) White Hispanic American American Total impact of radiation on cancer risks, but in the absence of a 0–4 38,816 27,764 5,544 467 72,591 large, well-defined and carefully followed cohort of Trinity 5–9 36,029 25,770 4,432 444 66,674 fallout-exposed individuals with individual dose estimates, 10–14 33,692 24,099 3,599 406 61,797 radiation risk projection based on models developed from 15–19 30,691 21,952 3,094 423 56,160 other exposed cohorts remains the only viable tool to estimate 20–24 28,079 20,084 3,367 524 52,053 the number of cancers that might have been caused by 25–29 28,169 20,148 2,952 755 52,024 fallout exposure to the New Mexico population (NRC 2005; 30–34 24,609 17,602 2,287 607 45,105 – Berrington de Gonzalez et al. 2012). 35 39 21,401 15,308 1,993 549 39,251 – Our objective is to estimate the range of radiation-related 40 44 17,424 12,463 1,533 494 31,915 45–49 15,335 10,969 1,149 407 27,859 excess cancers and corresponding attributable fractions from 50–54 12,139 8,682 1,166 265 22,252 exposure to fallout from the Trinity nuclear test among the 55–59 9,176 6,563 929 162 16,830 residents of New Mexico alive at the time of the test. We 60–64 6,757 4,833 862 163 12,614 use multiple databases to characterize the exposed New 65–69 5,537 3,960 520 127 10,143 Mexico population and baseline cancer rates and statistical 70–74 3,671 2,626 553 69 6,919 risk projection methods with reconstructed radiation doses 75 and older 3,828 2,738 693 42 7,301 (Simon et al. 2020) to estimate the potential magnitude and Total 315,352 225,561 34,673 5,903 581,489 proportion of radiation-related cancer risks. We consider uncertainties of the models and other factors used in the category was assumed to be Native Americans.3 The calculations and describe limitations in the uncertainty analysis. 1940 census combined counts for non-Hispanic whites This is the first assessment of cancer risks due to exposure and Hispanic whites for each county in New Mexico. to radioactive fallout from the Trinity nuclear test. However, a 5% sample of the 1940 census indicates that Hispanics (using the definition of having the Spanish mother MATERIALS AND METHODS tongue) represented 41.7% of the entire population of New Mexico at that time. We assumed the Hispanic population Based on methods and data described in companion to be 41.7% of every precinct and age group combination, papers (Bouville et al. 2020; Potischman et al. 2020; Simon rounded to the nearest person. The total population of et al. 2020), estimates of tissue-specific radiation absorbed New Mexico in 1945 was estimated to be 581,489 people doses from exposure to fallout from external and internal comprised of 315,352 Whites; 225,561 Hispanics; 34,673 sources were derived for residents of different precincts by Native Americans; and 5,903 African Americans. age at the time of the test and race/ethnicity for the first year following the Trinity test. In the current paper, total (internal Radiation dose and external combined) tissue-specific radiation doses are Tissue-specific absorbed doses (Table 2) from a range applied to baseline cancer rates, published life tables, and of radionuclides were estimated (Simon et al. 2020) for in- sex- and organ-specific risk coefficients to project the num- dividuals residing in the 721 precincts of New Mexico for ber of excess cancers among residents of New Mexico alive the primary organs at risk from exposure to radioactive at the time of the test. fallout: active bone marrow, thyroid, stomach, colon, and Study population lung, similar to analyses done for other populations exposed The population residing in New Mexico at the time to radioactive fallout (Land et al. 2010). For the cancer of the Trinity test on 16 July 1945 was estimated by linear risk projection of other organs, colon dose was used as a interpolation of the 1940 and 1950 United States census surrogate dose to project the excess of all solid cancers counts and is presented in Table 1 (USCB 2019). The total [excluding thyroid and non-melanoma skin cancer (NMSC)]. number of people reported in 5-y age groups by the census The dose reconstruction relied on fallout deposition estimates were apportioned equally to single year ages within that 5-y derived from ground-level exposure-rate measurements made group. For the wider age group of people 75 y and older, within 3 wk of the detonation, which were later confirmed by individuals were apportioned from ages 75 to 90 y based measurements from environmentally-placed x-ray film badges on 1939 US life tables for males and females. These censuses (Hoffman 1945), and on recall of diet and lifestyle from capture race/ethnicity of Whites, African Americans, and 3 “other races.” For New Mexico, the “other races” The research findings in this paper do not apply to the Navajo Nation. www.health-physics.com 480 Health Physics October 2020, Volume 119, Number 4 Table 2.4 Adapted from Simon et al. (2020). Mean radiation absorbed doses (mGy) to residents of New Mexico from Trinity radioactive fallout by county and age at exposure groupings. Abbreviations: ABM, active bone marrow. Doses rounded to two significant digits. Note: Weighted by population size in corresponding county and age grouping during first year at risk (1950 for solid cancers and 1947 for leukemia). Population Age group (y) ABM Thyroid Colon Stomach Lung

Totala <1 7.2E-01 1.1E+01 3.1E+00 8.7E-01 2.8E+00 1–2 8.2E-01 3.0E+01 4.9E+00 1.0E+00 2.4E+00 3–7 6.8E-01 2.2E+01 3.1E+00 7.9E-01 1.7E+00 8–12 6.7E-01 1.6E+01 3.0E+00 7.6E-01 1.8E+00 13–17 6.3E-01 1.2E+01 2.5E+00 6.8E-01 1.8E+00 Adult 4.8E-01 6.6E+00 2.1E+00 5.8E-01 1.6E+00 Selected countiesb <1 5.2E+00 6.6E+01 2.3E+01 6.5E+00 2.2E+01 1–2 5.4E+00 1.7E+02 2.6E+01 6.8E+00 1.9E+01 3–7 4.6E+00 1.3E+02 1.7E+01 5.5E+00 1.3E+01 8–12 4.5E+00 9.1E+01 1.7E+01 5.3E+00 1.4E+01 13–17 4.2E+00 6.8E+01 1.5E+01 4.9E+00 1.4E+01 Adult 3.5E+00 3.9E+01 1.3E+01 4.2E+00 1.2E+01 Other counties <1 9.7E-02 2.9E+00 3.2E-01 8.9E-02 1.3E-01 1–2 1.8E-01 1.1E+01 2.0E+00 1.9E-01 1.7E-01 3–7 1.3E-01 7.4E+00 1.2E+00 1.4E-01 1.4E-01 8–12 1.4E-01 5.2E+00 1.1E+00 1.3E-01 1.5E-01 13–17 1.4E-01 4.0E+00 7.6E-01 1.1E-01 1.4E-01 Adult 8.1E-02 2.2E+00 5.7E-01 8.7E-02 1.1E-01

aIncluding precincts with doses estimated to be greater than zero. bIncluding Guadalupe, Lincoln, San Miguel, Socorro, and Torrance. the mid-1940s in contemporary focus groups and personal Americans) in each precinct. Different dietary and lifestyle interviews from persons alive and residing in New Mexico patterns were assumed for each precinct based on collected at the time of Trinity. In the dose assessment, three exposure data from interviews and focus groups (Potischman et al. pathways were included: (1) external irradiation from the 2020). Table 2, adapted from Simon et al. (2020), presents radionuclides deposited on the ground, (2) inhalation of population-size weighted mean radiation absorbed doses by radionuclide-contaminated air during and after the passage of county and age at exposure groupings among residents the radioactive cloud, and (3) ingestion of contaminated water of New Mexico alive in 1945 with doses estimated to be and food. The ingestion pathway accounted for radioactive greater than zero. contamination of 13 types of food and water (Bouville et al. 2020; Simon et al. 2020). The calculations of contamination Baseline cancer rates of foods and air for the internal dose assessment accounted Baseline cancer rates are cancer incidence rates in a for 63 radionuclides in the fallout that are fission or activation population without known exposure to the factor of interest; products. Of the 63 considered, 54 have radioactive half-lives in this case, radiation from Trinity fallout. Baseline inci- of less than 3 mo, while only nine have radioactive half-lives dence rates for each specific cancer type vary between pop- longer than 9 mo. The risk analysis presented here uses ulations as well as over calendar time, across ages, birth estimates of intakes and doses received in the first year cohorts, sex, and race/ethnicity groups within a population. following the test. Any dose received in later years would To ensure stable estimates of baseline cancer rates by age at not only be much smaller than the component assessed but diagnosis, calendar year, sex, and race/ethnicity, we used would also be considerably more uncertain. Each derived cancer incidence rates reported by the Surveillance, Epide- dose estimate was considered to be the “best estimate” of miology, and End Results (SEER) cancer registry program radiation absorbed dose to specific organs for persons from 1973 to 2015, which includes between 9 and 18 representative of a specific age-at-exposure group (0 to <1 y, high-quality cancer registries across the United States. To 1–2y,3–7y,8–12 y, 13–17 y, and 18 y and older) and race/ derive baseline cancer incidence rates for 1945 to 1972, ethnicity (Whites, Hispanics, Native Americans, and African we extrapolated rates for all races combined from 1973 to 1987 from the nine SEER cancer registries (SEER9 2018). 4E-notation is used here due to space restrictions. We fit Poisson regression models that included 5-y age groups www.health-physics.com Projected lifetime cancer risks c E. K. CAHOON ET AL. 481 coded with dummy variables [combining 0 to 10-y-olds for cancer risk follows a step function so that it is equal to zero thyroid, leukemia—except chronic lymphocytic leukemia at time since exposure of less than 2 y for leukemia and less (CLL)], solid (except thyroid and NMSC) and 0 to 15-y- than 5 y for solid cancers. Unlike the BEIRVII report, we did olds for lung, colon, and stomach cancer), calendar year in not apply a dose and dose rate effectiveness factor (DDREF) single years fitted as a continuous variable, sex, and interac- to extrapolate from populations with high-dose/high-dose tion terms of calendar year by sex and age groups by sex. rates to those with low-dose/low-dose rates. A DDREF The SEER program includes cancer incidence rates for in- greater than 1, when applied to the excess relative risk, would dividuals of Hispanic ethnicity since 1992, so race ratios of reduce the number of estimated excess cases. Hispanics/Blacks/non-HispanicWhites/NativeAmericans compared to all races combined estimated using SEER Transfer of estimated excess relative risk to the exposed New Mexico population rates from 1992 to 2006 were applied to cancer incidence Selection of the weight given to the ERR and EAR rates for earlier periods. Poisson regression models using models can significantly impact the projected excess cancer SEER 18 data from 2000 to 2015 were used to extrapolate age-, risk when the baseline cancer rates differ between the sex-, and race/ethnicity-specific cancer incidence rates to years population from which the ERR or EAR was derived 2016 to 2034. Further details are provided in Appendix A. (e.g., Japanese atomic bomb survivors) and the population Models for estimating radiation-related cancer risk to which the risk is transferred (e.g., 1945 residents of New Excess radiation-related cancer risk can be computed Mexico). The BEIR VII approach uses the multiplicative either based on a multiplicative model with the excess rela- model (i.e., ERR) for thyroid cancer because mechanistic tive risk (ERR), an additive model using the excess absolute considerations suggest greater support for relative risk than risk (EAR), or a combination of the two models. We used for absolute risk transport (NRC 2005). For leukemia (except models based on the recommendations of the National CLL), stomach cancer, colon cancer, and for solid cancers Academy of Sciences BEIR VII report on the Biological (except thyroid and NMSC), the BEIR VII approach uses Effects of Ionizing Radiation (NRC 2005). BEIR VII a weighted average (on the logarithmic scale) of the ERR dose-response models were used for estimating the ERR and EAR models with weights of 0.7 on multiplicative and EAR per unit dose of radiation (Table 3). Most of transfer (ERR) and 0.3 on additive transfer (EAR). For the radiation dose-response coefficients in the BEIR VII lung cancer, the multiplicative model (ERR) weight is 0.3 report (denoted by b, g, h, d,andφ in the equation below) and 0.7 for the additive (EAR) model. We used a similar are based on analyses of data from the Life Span Study of approach, except that the weighted averages are on the Japanese atomic bomb survivors, which is considered the arithmetic, rather than the logarithmic, scale as used by gold standard in radiation risk assessment (Preston et al. previous studies to enable propagation of uncertainties 2007). The general form of the BEIR VII dose-response (Land et al. 2010; Berrington de Gonzalez et al. 2012). models for the ERR and EAR is Person-years The models recommended by BEIR VII provide esti- ERRðÞ D; s; e; a; t or EARðÞ D; s; e; t mates of the age-specific excess cancer risks. However, φ : ¼ bsDexpðÞge þ ha þ dt þ et ð1Þ projected lifetime cancer risk must reflect person-time at risk over the lifespan. Here, we use the term “person-years” Here, D is dose in Gy, e = (exposure age-30)/10 for ex- to represent the sum of years for which the study population posure age<30 and e = 0 for exposure age ≥ 30, a =loge is at risk of developing cancer. Person-years at risk reflect (attained age/60) and t =loge(time since exposure/25), the conditional probability of a person reaching a certain where time since exposure in years is attained by age minus age. United States (US) life tables provide 1-y survival data age at exposure. Age in years at exposure in this study is defined for persons alive at any given age during that calendar period as age at the time of the Trinity test. For example, based on and are published approximately every 10 y, based largely on these models (eqn 1) and coefficients (Table 3), the US census data (CDC 2019). We assumed that individuals in radiation-related risk of thyroid cancer is only modified by the exposed population could reach an age of up to 90 years age at exposure, with a stronger modification than other and estimated person-years from age in 1945 up until age cancers; thus, decreasing age at exposure substantially in- 90 y or year 2034 using published life tables by calendar creases thyroid cancer risk. The dose-response model for year, sex, and race/ethnicity, when available. For each corre- leukemia (excluding CLL) described in BEIR VII has a sponding sex, race/ethnicity, calendar year, and age, we linear-quadratic form for acute exposures, but the quadratic multiplied the corresponding baseline cancer rates by the term is typically dropped for protracted exposures, like those person-years for that age and summed over the different due to radioactive fallout (Land et al. 2010). We also used the ages to estimate lifetime cancer risk. This approach adjusts BEIR VII committee’s assumptions for cancer latency, that for competing age-specific mortality in estimating cumulative www.health-physics.com 482 Health Physics October 2020, Volume 119, Number 4 Table 3. Estimates (95% uncertainty limits) of preferred ERR and EAR model parameters for estimating site-specific cancer risk. Abbreviations: ERR, excess relative risk; EAR, excess absolute risk; PY, person-years. Model parameter Leukemiaa Thyroid Colon Stomach Lung All solidb

ERRd

bM 1.1 (0.10, 2.6) 0.53 (0.14, 2.0) 0.63 (0.37, 1.1) 0.21 (0.11, 0.40) 0.32 (0.15, 0.70) 0.33 (0.24, 0.47)

bF 1.2 (0.10, 2.9) 1.05 (0.28, 3.9) 0.43 (0.19, 0.96) 0.48 (0.31, 0.73) 1.40 (0.94, 2.1) 0.57 (0.44, 0.74) γ(e) -0.4 (-0.78, 0.0) -0.83c -0.3c -0.3c -0.3c -0.3 (-0.51, -0.10) h(a) 0 0 -1.4c -1.4c -1.4c -1.4 (-2.2, -0.7) d(t) -0.48 (-1.1, 0.20) 0 0 0 0 0 φ(e*t) 0.42 (0.0, 0.96) 0 0 0 0 0 EAR (per 104 PY)d

bM 1.62 (0.1, 3.6) 0 3.2 (1.8, 5.6) 4.9 (2.7, 8.9) 2.3 (1.1, 5.0) 22 (15, 30)

bF 0.93 (0.1, 2.0) 0 1.6 (0.8, 3.2) 4.9 (3.2, 7.3) 3.4 (2.3, 4.69) 28 (22,36) γ(e) 0.29 (0.0, 0.62) 0 -0.41c -0.41c -0.41c -0.41 (-0.59, -0.22) h(a) 0 0 2.8c 2.8c 5.2 (3.8, 6.6) 2.8 (2.15, 3.41) d(t)000000 φ(e*t) 0.56 (0.31, 0.85) 0 0 0 0 0 Multiplicative/additive transfer weights 0.7/0.3 1.0/0.0 0.7/0.3 0.7/0.3 0.3/0.7 0.7/0.3

a Leukemia excludes CLL. Because dose from fallout was considered to have been received at a low dose rate, the parameter for the dose-squared term in the BEIR VII model for leukemia was set equal to zero. b Except thyroid cancer and non-melanoma skin cancer. c Error assumed to be negligible, following BEIR VII (NRC 2006). d The form of the ERR and EAR models is bsDexp (ge+ha+dt+φet) where D is dose in Gy; (age at exposure-30)/10 for age at exposure<30 and e = 0 for age≥30; a = loge(attained age/60) and t = loge[(time since exposure)/25]. The sex-specific b, γ, h, d,andφ are uncertain param- eters where sex-specific b is assumed to have log-normal distributions for solid cancers and a four-parameter beta distribution for leukemia; γ, h, d,andφ are assumed to have normal uncertainty distributions when not constant. baseline and radiation-related excess risk. This approach time into person-years at risk, correspond to attributable allowed for survival probabilities to vary by sex, race/ethnicity, fraction measures based on incidence density in a closed and age and to change over calendar time. Further details cohort (Greenland and Robins 1988). are found in Appendix B. Uncertainty Calculation of excess and attributable fractions Our study accounts for many important but not all pos- The number of lifetime excess cancers for a given can- sible sources of uncertainty. We accounted for uncertainty in cer type associated with age at exposure, e; attained age, a; the baseline cancer rates, model parameters, model transfer latency period, l;transferweight,w; and baseline cancer rate weights, and radiation doses. For some sources, the uncer- B(a)atattainedagea is tainty magnitudes were based on empirical data, while for others the magnitudes of uncertainty were obtained more ∑90 − a¼eþlpyðÞ a ½ðw BaðÞERR þ ðÞ1 w EAR 2Þ subjectively from experts. The uncertainty of each component used in the calculation of the excess number of cancers was where py(a) is the person-years at risk of cancer during a described using probability distribution functions. Paramet- single year age interval adjusted by the life table probability ric bootstrap methods were used to propagate these numer- of survival to age a. The equation above is a simplified version ous sources of uncertainty. Possible sources of uncertainty for ease of exposition, since baseline cancer rates B(a)depend not accounted for include census-based population infor- on age, sex, race/ethnicity, and calendar year, and as described mation, national life tables, the extrapolation of model pa- previously, the ERR and EAR depend on age at exposure, rameters to low doses and low dose rates, and changes in attained age, sex, and tissue-specificdose.Inthepresentationof environmental and lifestyle factors that may have impacted the results, we rounded excess numbers to the nearest multiple the baseline cancer rates in early periods. of 10, and we do not present any numbers <10 after rounding. Since baseline cancer rates for the population of New Values for baseline and excess numbers of cancer cases Mexico were computed based on Poisson models for the were converted to estimates of attributable fractions, i.e., the periods 1945 to 1972 and 2016 to 2034, while we used projected proportion of cancers attributable to radiation dose, SEER rates more directly for the periods 1972 to 2015, by dividing the excess number of cancers by the total num- different approaches were implemented for accommodating ber, computed as the sum of baseline and excess cases. These uncertainty in the baseline rates depending on the calendar estimates, which incorporate latency, attained age, and calendar period. For years 1973 to 2015, we assumed the number www.health-physics.com Projected lifetime cancer risks c E. K. CAHOON ET AL. 483 of cancer cases for a specific cancer site was distributed the best estimate of dose within a race/ethnicity and age according to a Poisson distribution with the age- and group in a specified precinct to be log-normal (Simon et al. sex-specific counts and standard deviations provided by 2020). For the dose uncertainty component of the risk pro- the SEER program. Uncertainty for models extrapolating jection, we randomly drew a dose realization from a log-normal cancer risks from 1945 to 1972 and 2016 to 2034 used distribution that had as parameters the median dose (i.e., estimated parameters and covariance matrices describing the “best estimate”) and the variance computed from the correlations between parameters from the fitted Poisson geometric standard deviation (by age, voting precinct, and models in the baseline cancer risk models to generate rates race/ethnicity). To examine the impact of dose uncertainty directly from the Poisson models. See Appendix A for details. on our 90% uncertainty intervals (UIs) for the number of ex- The uncertainty in the ERR and EAR computations cess cancer cases, we compared our results for the total pop- (eqn 1) arises from two different sources: the uncertainty ulation (years 1945–2034) to results that did not incorporate in the parameters that are used and the uncertainty in the re- dose uncertainty (Appendix C). constructed dose. We accounted for uncertainty in model Parametric bootstraps/dose simulations based on the parameters used for ERR and EAR similarly to the BEIR assumptions described above generated 1,000 realizations VII report. The parameters in Table 3 were assumed to be of baseline and excess cases from which we calculated random variables, with bM and bF arising from log-normal distributions for solid cancers and from four-parameter beta 1,000 realizations of attributable fractions for each of 721 distributions for leukemia. Parameters γ, h, d,andφ were voting precincts, 4 race/ethnicities, 90 exposure ages, male assumed to have either normal distributions or were assumed and females, 90 calendar years of follow-up, and 6 cancer to be constant when no uncertainty limits were provided. The groups. The estimated numbers of cancers were summed to obtain totals for the entire New Mexico population and bM and bF parameters for solid cancers in Table 3 were as- sumed to be medians or geometric means (with 95% uncer- population sub-groups defined by selected counties/other tainty limits) and were converted to arithmetic means in the counties, and race/ethnicity categories. Using the 1,000 ERR and EAR calculations using the standard deviation cal- realizations of the projected number of excess cancers, culated from the uncertainty distribution. For example, for medians, means, and 90% uncertainty intervals (based on the th th colon cancer with bM distributed log-normally with upper 5 and 95 percentiles of the bootstrap distribution for the and lower uncertainty limits UL and LL, respectively, respective quantity), and corresponding attributable fractions the mean for the parameter bM was calculated as, were then generated for each cancer type and selected sub-population. By providing medians, means, and 90%  LnUL−LnLL 2 UIs, we intend to stress that our estimates should not be ðÞ21:96 : bM mean ¼ bM exp 2 interpreted as precise numbers but rather as ranges of possible excess cancer cases and attributable fractions.

Total tissue-specific doses used in eqn (1) were assumed RESULTS to be log-normally distributed (Simon et al. 2020). An important component of uncertainty relates to the Our estimates of excess cancer cases assume 581,489 transfer of excess relative risk (multiplicative) and absolute residents of New Mexico were alive at the time of the (additive) risks to the 1945 New Mexico population from Trinity test in July 1945 (Table 1). After accounting for a data derived largely from Japanese atomic bomb survivors. 2-y latency period for leukemia and a 5-y latency period This uncertainty arises from the lack of knowledge about for solid cancers, we calculated a total of 26.6 million which type of risk projection is accurate. The estimated excess person-years at risk of leukemia and 24.9 million person-years number of cases due to radiation exposure may be highly at risk of solid cancers from 1945 until 2034 or age 90 y, sensitive to the choice of weights, w, for the ERR and which ever came first (Table 4). EAR (eqn 2). We took a conservative approach, similar to The total lifetime baseline number of all solid cancers BEIR VII, in incorporating the uncertainty in the choice (excluding thyroid and NMSC) was estimated to be 183,000 of weights by assuming the weight follows a Bernoulli from 1945 to 2034, which is approximately 31% of the pop- distribution (NRC 2005). For example, for colon cancer, ulation alive and residing in New Mexico at the time of the when the multiplicative transport weight was taken to be Trinity test (Table 4). We estimated a radiation excess 90% 0.7, the weight for the additive transport was 1−w = 0.3, UI of 210 to 460 for solid cancers (except thyroid and and the variance of the weight was 0.7 0.3. In contrast, NMSC) corresponding to an attributable fraction between only multiplicative projections were used for thyroid cancer 0.12% and 0.25% of all solid cancer cases in this population. and consequently, no uncertainty was assigned to w. We estimated an excess 90% UI of 80 to 530 thyroid cancer Dose uncertainty was also accounted for. An assess- cases from 1945 to 2034, representing a range for the attributive ment of uncertainty found the uncertainty distribution on fraction of 3.6% to 20% in the total population. The 1945 www.health-physics.com 484 Health Physics October 2020, Volume 119, Number 4 to 2015 period includes most of the baseline and excess cancer, 36% for colon cancer, 40% for all solid cancers (ex- cancers, which represents 0 to 70 y since the test and cept thyroid cancer and NMSC), 49% for lung cancer, and about 96% of the estimated person-years. 50% for thyroid cancer. The sources of uncertainty for the Estimated numbers of baseline cancers, radiation-related remainders of the 90% UI length included baseline cancer rates, excess cancers, and the proportion of total cancer risk attribut- transport model weighting, and BEIR VII radiation-related able to Trinity fallout are shown for selected counties in cancer risk parameters. Table 4. Among the 30 counties in New Mexico at the time of the test, the counties of Guadalupe, Lincoln, San Miguel, DISCUSSION Socorro, and Torrance had the highest attributable risk ranges of all solid cancer (except thyroid and NMSC) and accounted Despite the widespread interest in quantifying the can- for over 70% of excess cancer cases. The 90% UIs of the cer burden from Trinity through an observational epidemio- total projected number of cancers in these counties were 150 logical study, it was not feasible to conduct a study of that to 330 for all solid cancers (except thyroid and type in part because of the absence of a tumor registry in NMSC), 60 to 370 for thyroid cancer, and up to 10 for New Mexico in the 30 y immediately following the nuclear leukemia (except CLL), with attributable fractions ranging test. Instead, the current study uses newly estimated from 0.65% to 1.4%, 17% to 58%, and 0.12% to 2.1%, radiation doses and existing epidemiologically-based radiation respectively. The county-specific attributable risk 90% UIs of risk data to estimate cancer risks to the radiation-exposed thyroid cancer are shown in Fig. 1. population of New Mexico. We applied estimates of tissue- Projected baseline and radiation-related cancers and specific radiation doses to the populations of 721 precincts proportion of total cancer risk attributable to radioactive of New Mexico by age and race/ethnicity to information from fallout are shown by race/ethnicity in Table 5. The number published, publicly available New Mexico census data, US life of mean baseline and excess cancer cases was highest tables, US baseline cancer rates, and radiation risk model among Whites, primarily reflecting the population size of parameters derived primarily from study of the Japanese Whites at the time of the test. Whites contributed up to atomic bomb survivors. 14.9 million person-years to the time at risk, and we esti- In this work, we determined the five counties of Guadalupe, mated an excess with an upper 90% uncertainty limit of Lincoln, San Miguel, Socorro, and Torrance likely accounted <10 for leukemia (except CLL), 90% UIs 50 to 330 for thy- for over two-thirds of excess cancers and reported the esti- roid cancer, and 130 to 310 for all solid cancers (except for mated excess and corresponding ranges for these counties thyroid and NMSC). These correspond to attributable combined (Table 4). Similarly, we reported the excess cancer fraction 90% UIs of 0.02% to 0.31%, 3.6% to 20%, and for the other 25 counties combined. The distribution of attrib- 0.11% to 0.26%, respectively. Hispanics contributed about utable fractions for thyroid cancer shown by county (Fig. 1) re- 10 million person-years, and we estimated an excess 90% flects the dose estimation (Simon et al. 2020). In this work and UIs of up to 10 for leukemia (except CLL), 30 to 210 for that of Simon et al. (2020), both the organ doses and the thyroid cancer, and 70 to 160 for all solid cancers (except resulting excess cancer cases were estimated at the voting pre- for thyroid and NMSC). These corresponded to attributable cinct level and summed at the county level. While precincts in fraction 90% UIs of 0.02% to 0.35%, 3.4% to 22%, and counties far outside the main fallout pattern (see Bouville et al. 0.13% to 0.29%, respectively. Native Americans contrib- 2020, Fig. 1), were relatively homogenous in the exposures uted approximately 1.5 million person-years, and we esti- they received, the exposure levels in precincts in counties mated an excess number of cases with an upper 90% within the main fallout deposition pattern were much more uncertainty limit of <10 thyroid cancers and <10 solid can- heterogenous (Simon et al. 2020). In counties within the cers (except for thyroid and NMSC), corresponding to at- fallout pattern, the precise boundaries of the fallout deposi- tributable fraction 90% UIs of 0.63% to 4.2% and 0.03% to tion pattern relative to the precinct boundaries locations 0.07%, respectively. African Americans contributed up to were difficult to assess. Consequently, precinct-level doses 232,000 person-years, and we estimated an excess number and cancer risks are not reported because such small divi- of cases with an upper 90% uncertainty limit of <10 thyroid sions are not considered reliable. cancers and <10 all solid cancers (except for thyroid and Among single cancer sites, in the total population, county NMSC) corresponding to attributable fraction 90% UIs of subgroups, and in all race/ethnicities represented in Tables 4 0.83% to 5.5%, and 0.02% to 0.05%, respectively. and 5, the fraction of cancer cases attributable to radiation We compared the length of the 90% UIs for our results exposure was highest for thyroid cancer. This reflects the for the total population (years 1945–2034) to results that did large effect of exposure to the radioactive isotope 131Iin not incorporate dose uncertainty. Not incorporating dose fallout from nuclear weapons tests. Iodine concentrates in the uncertainty substantially reduced the length of the 90% thyroid gland, which uses iodine to produce thyroid hor- UIs by 25% for leukemia (except CLL), 31% for stomach mones, resulting in exposures generally much greater than www.health-physics.com Table 4. Projected numbers of baseline cancers, excess radiation-related cancers, and proportion (in %) of total cancer risk attributable to radioactive fallout, by county group, cancer type, and time period (uncertainty distributions for excess cases represented by their means, medians, 5th and 95th percentiles). Abbreviations: PY, person-years. Results for person years, baseline, and excess presented for three significant digits or rounded to the 10th. Attributable risk presented for two significant digits or rounded to the hundredth. Projected lifetime cancers, 1945–2034 Estimated cancers, 1945–2015 Excess cases Attributable risk (%) Excess cases Attributable risk (%) Group/Cancer PY Baseline 5% Median Mean 95% 5% Median Mean 95% PY Baseline 5% Median Mean 95% 5% Median Mean 95%

Total a

Leukemia 26,600,000 2,900 <10 <10 <10 <10 0.02 0.09 0.12 0.31 25,500,000 2,580 <10 <10 <10 <10 0.02 0.1 0.13 0.34 risks cancer lifetime Projected Thyroid 24,900,000 2,170 80 210 240 530 3.6 8.7 9.7 20 23,800,000 1,880 60 150 180 380 3.1 7.5 8.4 17 Colon 24,900,000 16,300 20 30 30 40 0.11 0.17 0.17 0.25 23,800,000 14,600 20 20 20 40 0.1 0.15 0.16 0.24 Stomach 24,900,000 5,180 <10 <10 <10 20 0.02 0.03 0.1 0.33 23,800,000 4,650 <10 <10 <10 10 0.02 0.03 0.09 0.29 www.health-physics.com Lung 24,900,000 26,500 20 30 40 60 0.07 0.12 0.14 0.23 23,800,000 23,700 10 20 30 50 0.06 0.1 0.11 0.21 All Solidb 24,900,000 183,000 210 310 320 460 0.12 0.17 0.17 0.25 23,800,000 160,000 170 250 260 370 0.11 0.16 0.16 0.23 Selected countiesc Leukemiaa 3,210,000 360 <10 <10 <10 <10 0.12 0.62 0.82 2.1 3,070,000 320 <10 <10 <10 <10 0.13 0.67 0.88 2.3

Thyroid 3,010,000 270 60 150 170 370 17 35 36 58 2,870,000 230 40 110 130 270 15 32 33 54 c

Colon 3,010,000 1,970 10 20 20 30 0.66 0.97 1.0 1.5 2,870,000 1,760 10 20 20 30 0.59 0.89 0.94 1.4 C K. E. Stomach 3,010,000 620 <10 <10 <10 20 0.13 0.21 0.75 2.4 2,870,000 550 <10 <10 <10 10 0.12 0.20 0.67 2.1

Lung 3,010,000 3,230 20 30 30 60 0.56 0.95 1.0 1.8 2,870,000 2,880 10 20 30 50 0.43 0.73 0.87 1.6 AL ET AHOON All Solidb 3,010,000 22,400 150 210 220 330 0.65 0.94 0.98 1.4 2,870,000 19,600 120 180 180 270 0.61 0.89 0.92 1.3 Other counties a Leukemia 23,400,000 2,540 <10 <10 <10 <10 0 0.02 0.02 0.06 22,500,000 2,260 <10 <10 <10 <10 0 0.02 0.02 0.06 . Thyroid 21,900,000 1,900 20 60 70 150 1.2 3.0 3.6 7.4 21,000,000 1,650 20 40 50 110 1.1 2.6 3.0 6.3 Colon 21,900,000 14,400 <10 <10 <10 10 0.04 0.06 0.06 0.08 21,000,000 12,900 <10 <10 <10 10 0.03 0.05 0.05 0.08 Stomach 21,900,000 4,560 <10 <10 <10 <10 0 0 0.02 0.05 21,000,000 4,100 <10 <10 <10 <10 0 0 0.01 0.05 Lung 21,900,000 23,300 <10 <10 <10 <10 0.01 0.01 0.01 0.02 21,000,000 20,900 <10 <10 <10 <10 0 0.01 0.01 0.02 All Solidb 21,900,000 160,000 60 90 90 130 0.04 0.06 0.06 0.08 21,000,000 141,000 50 70 80 110 0.04 0.05 0.05 0.08

aExcludes CLL (chronic lymphocytic leukemia). bAll solid cancers except thyroid and non-melanoma skin cancer. cSelected counties include five counties with the highest attributable risk of all solid cancer: Guadalupe, Lincoln, San Miguel, Socorro, and Torrance. 485 486 Health Physics October 2020, Volume 119, Number 4

Fig. 1. Uncertainty intervals (5%, 95%) for proportion (in %) of thyroid cancer risk attributable to radioactive fallout from the Trinity nuclear test by county among New Mexico residents alive in 1945 (1945 to 2034). for other organs of the body. The radioactive isotope 131Iis population in that race/ethnicity category. Similarly, national indistinguishable by the thyroid gland from the non-radioactive life tables were used that for many years correspond to data version, making the thyroid gland especially vulnerable to this collected by the United States Census Bureau and may not form of radiation, particularly during childhood, as has been have accurately described the experience of all ethnic and reported in population-based studies of Chernobyl fallout racial groups in New Mexico. In addition, possible nonlinear (Brenner et al. 2011; Zablotska et al. 2011). The estimates changes, e.g., those caused by World War II, in the for attributable fraction of thyroid cancer were highest for His- population distribution during the period of time between panics and Whites, reflecting higher intakes of dairy products the two censuses were not accounted for. and locations of residence for these groups. However, the un- Cancer incidence data were not systematically col- certainty around these estimates was substantial, and 90% uncer- lected for residents of New Mexico for the entire time tainty intervals for attributable fractions of thyroid cancer do not period (Gibson and Jung 2002) in our study, so we relied support strong differences across races/ethnicities. on available data from high quality cancer registries in the There are several limitations that must be considered in United States SEER program that started in 1973 and have the interpretation of our excess and attributable fractions es- been updated until 2015 to obtain stable estimates of age-, sex-, timates. While the association between ionizing radiation and race/ethnicity-specific baseline cancer rates, projecting and cancer risk is considered one of the best quantified rates for periods 1945–1972 and 2016–2034, which were dose-response relationships for any environmental agent, outside the time period captured by SEER. We limited the the estimated number of excess cases is still uncertain and scope of projection to ages under 90 y and calendar year depends on numerous assumptions and input data. In this before 2034, for which projections of baseline cancer rates work, we sought to use the best available published data are more reliable. Projected baseline cancer rates prior to as input to our calculations and to make assumptions that 1973 did not account for changes in environmental and would not knowingly or purposefully bias the estimates. lifestyle risk factors. For example, the broader availability The study population and distribution of race/ethnicity of refrigeration in the 1940s, which reduced helicobacter was estimated based on US census data from 1940 and 1950 pylori prevalence, possibly results in an underestimation as that was the government-documented data available to of stomach cancer incidence in the early period (Luo et al. this study (USCB 2019). However, use of census data is ac- 2017). Modeled racial and ethnic patterns for cancer incidence knowledged to have possibly resulted in underestimates of based on data obtained from the SEER program do reflect the numbers of certain groups if they were less likely to partic- racial and ethnic patterns in cancer mortality that have been ipate in the census for either year. This could have been the reported for New Mexico from 1958 to 1982 (Becker et al. case, for example, for Native Americans who, through 1950, 1993). For example, similarly to Becker and colleagues, we were racially identified by a census taker rather than the indi- estimated a near two-fold increased risk of stomach cancer vidual interviewed (Jobe 2004). Hispanics were identified by among Hispanics compared to Whites. Overall, cancer those self-reporting the “Spanish mother tongue,” which may incidence for residents of New Mexico has been reported have underestimated the proportion of the New Mexico to be lower than the entire United States (NCI 2019), which

www.health-physics.com Table 5. Projected numbers of baseline cancers, excess radiation-related cancers, and proportion (in %) of total cancer risk attributable to radioactive fallout, by race/ethnicity, cancer type, and time period (uncertainty distributions for excess cases represented by their means, medians, 5th, and 95th percentiles). Abbreviations: PY, person-years. Results for person years, baseline, and excess presented for three significant digits or rounded to the 10th. Attributable risk presented for two significant digits or rounded to the hundredth. Projected lifetime cancers, 1945–2034 Estimated cancers, 1945–2015 Excess cases Attributable risk (%) Excess cases Attributable risk (%) Race/Cancer PY Baseline 5% Median Mean 95% 5% Median Mean 95% PY Baseline 5% Median Mean 95% 5% Median Mean 95%

Whites Leukemiaa 14,900,000 1,720 <10 <10 <10 <10 0.02 0.09 0.12 0.31 14,300,000 1,510 <10 <10 <10 <10 0.02 0.1 0.13 0.34 Thyroid 13,900,000 1,310 50 130 150 330 3.6 8.9 9.9 20 13,300,000 1,130 40 90 110 240 3.1 7.6 8.5 17 Colon 13,900,000 10,400 10 20 20 30 0.11 0.17 0.17 0.26 13,300,000 9,440 <10 10 20 20 0.1 0.15 0.16 0.25 Stomach 13,900,000 2,120 <10 <10 <10 10 0.02 0.03 0.15 0.49 13,300,000 1,910 <10 <10 <10 <10 0.02 0.03 0.13 0.43 rjce ieiecne risks cancer lifetime Projected Lung 13,900,000 19,200 10 20 20 50 0.06 0.11 0.13 0.24 13,300,000 17,100 <10 10 20 40 0.05 0.08 0.11 0.23 All Solidb 13,900,000 119,000 130 200 200 310 0.11 0.16 0.17 0.26 13,300,000 105,000 100 160 170 250 0.1 0.15 0.16 0.24 Hispanics a www.health-physics.com Leukemia 9,960,000 1,060 <10 <10 <10 <10 0.02 0.1 0.14 0.36 9,540,000 960 <10 <10 <10 <10 0.02 0.11 0.15 0.39 Thyroid 9,300,000 770 30 80 90 210 3.5 8.9 10 22 8,880,000 670 20 60 70 150 3.1 7.7 8.9 19 Colon 9,300,000 5,000 <10 10 10 20 0.13 0.2 0.2 0.3 8,880,000 4,310 <10 <10 <10 10 0.12 0.18 0.18 0.27 Stomach 9,300,000 2,680 <10 <10 <10 <10 0.02 0.03 0.09 0.26 8,880,000 2,390 0 0 0 <10 0.02 0.03 0.08 0.23 Lung 9,300,000 5,920 <10 10 10 20 0.12 0.18 0.2 0.31 8,880,000 5,310 <10 <10 <10 10 0.09 0.14 0.16 0.26 c b All Solid 9,300,000 55,500 70 110 110 160 0.13 0.19 0.2 0.29 8,880,000 48,000 60 90 90 130 0.12 0.18 0.18 0.26 C K. E. Native Americans a Leukemia 1,570,000 100 <10 <10 <10 <10 0 0.01 0.02 0.04 1,500,000 90 <10 <10 <10 <10 0 0.01 0.02 0.04 AL ET AHOON Thyroid 1,470,000 80 <10 <10 <10 <10 0.63 1.6 1.9 4.2 1,400,000 70 <10 <10 <10 <10 0.56 1.4 1.7 3.7 Colon 1,470,000 760 <10 <10 <10 <10 0.02 0.04 0.04 0.06 1,400,000 670 <10 <10 <10 <10 0.02 0.03 0.03 0.05

Stomach 1,470,000 310 <10 <10 <10 <10 0 0 0.01 0.03 1,400,000 280 <10 <10 <10 <10 0 0 0.01 0.03 . Lung 1,470,000 1,060 <10 <10 <10 <10 0 0.01 0.01 0.01 1,400,000 950 <10 <10 <10 <10 0 0 0 0.01 All Solidb 1,470,000 6,460 <10 <10 <10 <10 0.03 0.04 0.04 0.07 1,400,000 5,630 <10 <10 <10 <10 0.02 0.04 0.04 0.06 African Americans Leukemiaa 232,000 20 <10 <10 <10 <10 0 0.01 0.02 0.04 227,000 20 <10 <10 <10 <10 0 0.01 0.02 0.04 Thyroid 215,000 10 <10 <10 <10 <10 0.83 2.2 2.5 5.5 210,000 <10 <10 <10 <10 <10 0.74 1.9 2.2 4.7 Colon 215,000 210 <10 <10 <10 <10 0.02 0.03 0.03 0.06 210,000 200 <10 <10 <10 <10 0.02 0.03 0.03 0.05 Stomach 215,000 70 <10 <10 <10 <10 0 0 0.01 0.02 210,000 70 <10 <10 <10 <10 0 0 0.01 0.02 Lung 215,000 350 <10 <10 <10 <10 0 0 0 0.01 210,000 330 <10 <10 <10 <10 0 0 0 0.01 All Solidb 215,000 1,870 <10 <10 <10 <10 0.02 0.04 0.04 0.05 210,000 1,750 <10 <10 <10 <10 0.02 0.03 0.03 0.05

aExcludes CLL (chronic lymphocytic leukemia). bAll solid cancers except thyroid and non-melanoma skin cancer. 487 488 Health Physics October 2020, Volume 119, Number 4 would imply that the number of estimated baseline and excess on Radiological Protection (ICRP) has proposed a DDREF cancers may be overestimated under the multiplicative transfer of 2 (Wrixon 2008), the expert panel at the World Health model. However, inaccuracy in baseline cancer rates and Organization more recently did not apply a DDREF (i.e., all-cause mortality derived from US life tables apply similarly they set DDREF equal to 1) to its health risk assessment to estimates of excess and baseline cases, so that the projected following the nuclear accident in Fukushima (WHO 2013). proportion of cancer risk attributable to radioactive fallout, In addition, recently updated and expanded pooled analyses calculated as excess cases divided by the sum of baseline and of thyroid cancer, the organ most exposed to Trinity fallout, excess cases, should not be appreciably impacted. Thus, we supports linearity of the dose-response in the low dose range do not believe that these various limitations have substantially (<0.2 Gy) (Veiga et al. 2016; Lubin et al. 2017). biased our estimates of attributable fractions presented in Several other issues merit discussion. We did not ac- Tables 4 and 5 and Fig. 1. count for uncertainty in the life table- and census-based es- Uncertainty in dose estimation is an important limita- timates of person years at risk by age and race/ethnicity. It tion of this study. Dose estimates were based on the fallout should also be restated that the doses captured exposure pattern, information collected about individuals’ diet and from only the first year after the detonation, corresponding lifestyle, and exposure models used to estimate doses for to an estimated 90% of the lifetime dose. The effect of cap- external and internal exposure from consumption of food, turing 90% of the dose would suggest our risk projection water, in-cloud inhalation, and resuspension over the first might slightly underestimate total excess cases and attributable year following Trinity. Information on diet and lifestyle ob- risk, particularly for cancer outcomes related to longer-lived tained from focus groups have inherent limitations, including radionuclides. We also did not project risk for the in utero memory recall and difficulty in sampling groups representa- exposed population, which may also lead to a slight underesti- tive of all ages at the time of potential exposure. An analysis mation of total excess cancer cases. Although dose estimates of dose uncertainty, based on Monte Carlo simulations using were available for this group, robust long-term epidemiological subjectively and experience-derived probability density func- data are not available from which to obtain radiation-related tions resulted in geometric standard deviations on dose esti- risk parameters. mates ranging from 2.7 to 5.6 (Simon et al. 2020). When we compared our estimates of numbers of excess cancer cases CONCLUSION for the total population (years 1945–2034) to those that did not incorporate dose uncertainty, we found that 25% to 50% We provide estimates of the ranges of excess cancer cases of the length of the 90% uncertainty intervals could be attributed from exposure to Trinity fallout to residents of New Mexico to dose uncertainty. Thus, uncertainty in dose estimation had a alive in 1945. In this analysis, we accounted for uncertainty substantial impact on the total uncertainty around our estimates. of estimated doses, baseline cancer risks, model weights, Excess relative and absolute risk models rely on evidence and radiation-risk model parameters. There are several key from Japanese atomic bomb survivors, many of whom were conclusions from this analysis. Our 90% UIs suggest that exposed to moderate to high doses and high dose rates that as many as 1,000 or as few as 290 cancers have already resulted from near-instantaneous prompt gamma ray expo- occurred or are projected to occur in the future that would sure and very little protracted exposure from radioactive not have occurred in the absence of residential radiation fallout. In contrast, the residents of New Mexico were exposure from Trinity fallout. Most of the excess cancers are potentially exposed to lower doses and low dose rates over projected to have occurred or will occur among residents of a longer time since no New Mexico residents were close Guadalupe, Lincoln, San Miguel, Socorro, and Torrance enough to the detonation to be exposed to prompt gamma counties in 1945. Uncertainty in dose estimation had a rays (Simon et al. 2020). A DDREF of 1.5 was used in BEIR substantial impact in the total uncertainty around our VII to extrapolate from populations with high-dose/high-dose estimates. Finally, most cancers that have occurred or will rates to those with low-dose/low-dose rates, effectively reducing occur among the residents of New Mexico in 1945 are likely the number of estimated excess cancer cases. An evaluation to be cancers unrelated to exposures from Trinity fallout. of uncertainty in the BEIR VII report (Table 12-10, page 284), indicated that the main contribution to uncertainty for all solid Acknowledgments—This research was supported primarily by the Intramural Research Program of the National Cancer Institute with partial support from cancers (except thyroid and NMSC) was the DDREF. The the Intra-Agency agreement between the Radiation Nuclear Countermeasures current analysis differs from the BEIR VII report in that we Program of the National Institute of Allergy and Infectious Diseases with the did not apply either a DDREF or incorporate uncertainty National Cancer Institute, NIAID agreement #Y2-Al-5077 and NCI agreement #Y3-CO-511. The authors wish to acknowledge the contributions of other past around the values for the DDREF. There continues to be and present Trinity study team members including Lauren Houghton, Cheryl substantial debate among experts as to how to best extrapolate Deaguiar, Abigail Ukwuani, Kayla Myers, Jessica Lopez, Emily Haozous, Silvia Salazar, Nancy Potischman, and the Southwest Tribal Epidemiology to low doses (Ruhm et al. 2015; Rühm 2016; Shore et al. 2017; Center, particularly Kevin English. We are grateful to the many study partici- Tran and Little 2017). While the International Commission pants across the state of New Mexico for the insights, time, and commitment www.health-physics.com Projected lifetime cancer risks c E. K. CAHOON ET AL. 489 to providing the data needed for this work. We thank William Wheeler from Ron E, Lubin JH, Shore RE, Mabuchi K, Modan B, Pottern LM, IMS for computing help. This work utilized the computational resources of Schneider AB, Tucker MA, Boice JD Jr. Thyroid cancer after the NIH HPC Biowulf cluster (http://hpc.nih.gov). exposure to external radiation: a pooled analysis of seven stud- ies. Radiat Res 141:259–277; 1995. REFERENCES Rühm W. Dose-rate effects in radiation biology and radiation pro- tection. Washington, DC: ICRP; 2016. Becker TM, Wiggins CL, Elliott RA, Key CR, Samet JM. Racial Rühm W, Woloschak GE, Shore RE, Azizova TV,Grosche B, Niwa and ethnic patterns of mortality in New Mexico. Albuquerque: O, Akiba S, Ono T, Suzuki K, Iwasaki T, Ban N, Kai M, Clement University of New Mexico Press; 1993. CH, Bouffler S, Toma H, Hamada N. Dose and dose-rate effects Berrington de Gonzalez A, Iulian Apostoaei A, Veiga LH, Rajaraman of ionizing radiation: a discussion in the light of radiological pro- P,Thomas BA, Owen Hoffman F, Gilbert E, Land C. Radrat: a tection. Radiat Environ Biophys 54:379–401; 2015. radiation risk assessment tool for lifetime cancer risk projection. Rühm W, Azizova TV, Bouffler SD, Little MP, Shore RE, Walsh L, J Radiol Protect 32:205–222; 2012. Woloschak GE. Dose-rate effects in radiation biology and radia- Bouville A, Beck HL, Thiessen KM, Hoffman FO, Potischman tion protection. Ann ICRP Jun;45(1 suppl):262–279; 2016. doi: N, Simon SL. The methodology used to assess doses from the 10.1177/0146645316629336. first nuclear weapons test (Trinity) to the populations of New SEER 13: Surveillance, Epidemiology, and End Results (SEER) Mexico. Health Phys 119:400–427; 2020. Program (www.seer.cancer.gov) SEER*Stat Database: Inci- Brenner AV,Tronko MD, Hatch M, Bogdanova TI, Oliynik VA, Lubin dence - SEER 13 Regs Research Data, Nov 2017 Sub (1992- JH, Zablotska LB, Tereschenko VP, McConnell RJ, Zamotaeva 2015) - Linked To GA, O'Kane P, Bouville AC, Chaykovskaya LV, Greenebaum E, County Attributes - Total U.S., 1969-2016 Counties, National Paster IP, Shpak VM, Ron E. I-131 dose response for incident Cancer Institute, DCCPS, Surveillance Research Program, re- thyroid cancers in Ukraine related to the Chornobyl accident. leased April 2018, based on the November 2017 submission. Environmental Health Perspectives 119:933–939; 2011. Accessed 27 February 2019. CDC. United States lifetables [online]. Available at: https://www. SEER 18: Surveillance, Epidemiology, and End Results (SEER) cdc.gov/nchs/products/life_tables.htm. Accessed 26 April 2019. Program (www.seer.cancer.gov) SEER*Stat Database: Inci- Gibson C, Jung K. Historical census statistics on population totals dence - SEER 18 Regs Research Data + Hurricane Katrina Im- by race (1790 to 1990) and by hispanic origin (1970 to 1990) pacted Louisiana Cases, Nov 2017 Sub (2000-2015) - Linked To County Attributes - Bureau; 2002. Total U.S., 1969-2016 Counties, National Cancer Institute, DCCPS, Greenland S, Robins JM. Conceptual problems in the definition Surveillance Research Program, released April 2018, and interpretation of attributable fractions. Am J Epidemiol based on the November 2017 submission. Accessed 27 128:1185–1197; 1988. February 2019. Hoffman JG. Nuclear explosion 16 July 1945, part c: transcript of SEER 18: Surveillance, Epidemiology, and End Results (SEER) radiation monitor’s field notes. Film badge data on town monitoring. Program (www.seer.cancer.gov) SEER*Stat Database: Inci- Los Alamos, NM: Los Alamos Scientific Laboratory; Nuclear Test- dence - SEER 18 Regs Research Data + Hurricane Katrina Im- pacted Louisiana Cases, Nov 2017 Sub (2000-2015) Jobe MM. Native Americans and the U.S. Census: a brief histori- Rita Population Adjustment - Linked To County Attributes - cal survey. Boulder, CO: University Libraries Faculty & Staff Total U.S., 1969-2016 Counties, National Cancer Institute, DCCPS, Contributions 28; 2004. Surveillance Research Program, released April 2018, based on the November 2017 submission. Accessed 27 Land CE, Bouville A, Apostoaei I, Simon SL. Projected lifetime February 2019. cancer risks from exposure to regional radioactive fallout in Shore R, Walsh L, Azizova T, Ruhm W.Risk of solid cancer in low the Marshall Islands. Health Phys 99:201–215; 2010. dose-rate radiation epidemiological studies and the dose-rate ef- Lubin JH, Adams MJ, Shore R, Holmberg E, Schneider AB, fectiveness factor. Int J Radiat Biol 93:1064–1078; 2017. Hawkins MM, Robison LL, Inskip PD, Lundell M, Johansson Simon SL, Bouville A, Beck HL, Melo DR. Estimated radiation R, Kleinerman RA, de Vathaire F, Damber L, Sadetzki S, Tucker doses received by New Mexico residents from the 1945 Trinity M, Sakata R, Veiga LHS. Thyroid cancer following childhood nuclear test. Health Phys 119:428–477; 2020. low-dose radiation exposure: a pooled analysis of nine cohorts. – Tran V,Little MP.Dose and dose rate extrapolation factors for ma- J Clin Endocrinol Metab 102:2575 2583; 2017. lignant and non-malignant health endpoints after exposure to Luo G, Zhang Y, Guo P, Wang L, Huang Y, Li K. Global patterns gamma and neutron radiation. Radiat Environ Biophys 56: and trends in stomach cancer incidence: age, period and birth 299–328; 2017. – cohort analysis. Int J Cancer 141:1333 1344; 2017. USCB. Census of population and housing [online]. 2019. Available NCI. Quick profiles: New Mexico. [online] https://statecancerprofiles. at: https://www.census.gov/prod/www/decennial.html. Accessed 13 cancer.gov. Accessed 26 April 2019. 2019. May 2019. NRC. Health risks from exposure to low levels of ionizing radia- Veiga LH, Holmberg E, Anderson H, Pottern L, Sadetzki S, Adams tion: BEIR VII phase 2. Washington, DC: National Research MJ, Sakata R, Schneider AB, Inskip P, Bhatti P, Johansson R, Council (NRC) and Committee to Assess Health Risks from NetaG,ShoreR,deVathaireF,DamberL,KleinermanR, Exposure to Low Levels of Ionizing Radiation; 2005. Hawkins MM, Tucker M, Lundell M, Lubin JH. Thyroid cancer Potischman N, Salazar SI, Scott MA, Naranjo M, Hazous E, after childhood exposure to external radiation: an updated pooled Bouville A, Simon SL. Methods and findings on diet and life- analysis of 12 studies. Radiat Res 185:473–484; 2016. style used to support estimation of radiation doses from radio- WHO. Health risk assessment from the nuclear accident after the active fallout from the Trinity nuclear test. Health Phys 119: 2011 Great East Japan earthquake and tsunami based on a prelim- 390–399; 2020. inary dose estimation. Geneva: World Health Organization; 2013. Preston DL, Ron E, Tokuoka S, Funamoto S, Nishi N, Soda M, Wrixon AD. New recommendations from the international com- Mabuchi K, Kodama K. Solid cancer incidence in atomic mission on radiological protection—a review. Phys Med Biol bomb survivors: 1958-1998. Radiat Res 168:1–64; 2007. 53:R41–60; 2008.

www.health-physics.com 490 Health Physics October 2020, Volume 119, Number 4

Zablotska LB, Ron E, Rozhko AV, Hatch M, Polyanskaya ON, the reference group. We then tested if there were year by Brenner AV, Lubin J, Romanov GN, McConnell RJ, O'Kane race interactions in the Poisson models. There were some P, Evseenko VV, Drozdovitch VV, Luckyanov N, Minenko VF, Bouville A, Masyakin VB. Thyroid cancer risk in Belarus minor deviations from a constant race/ethnicity ratio by year among children and adolescents exposed to radioiodine after for the following sites/races/ethnicities: Hispanics for colon the Chornobyl accident. Br J Cancer 104:181–187; 2011. (men and women) and White women for stomach, which we did not accommodate further as they were not signifi- cant after controlling for multiple testing. The extrapolated APPENDIX A: BASELINE CANCER RATES rates from 1945 to 1972 were multiplied by the race/ethnicity Cancer incidence rates for the full period 1945 to 2034 ratio to obtain race/ethnicity and age-specific rates for each were estimated by age at diagnosis, calendar year, sex, and cancer site. The same race/ethnicity ratios were also applied race/ethnicity using data reported by the Surveillance, Epide- to cancer rates for all races combined from SEER 9, from miology, and End Results (SEER) cancer registry program 1972 to 1992, as not all race/ethnic groups were available in from 1973 to 2015. Since the SEER registries contained His- SEER during that period. For lung and all solid cancers, the panic ethnicity-specific data beginning in 1992, the period fol- race/ethnicity ratio estimates differed by sex; for all other sites, lowing 1992 was used to calculate race/ethnicity rate ratios they were the same for men and women. that were applied to earlier periods. All solid cancers included ICDO-III codes C00-C89 behavior 3, excluding leukemia, Years 1973 to 1991 lymphoma, myeloma, and other lympho-hematopoietic malig- The same race/ethnicity ratios used for the 1945 to nancies, plus brain and CNS tumors of benign, uncertain, or 1972 period were also applied to cancer rates for all unknown behavior (ICDO-III codes C70-C72, behavior 0,1) races/ethnicities combined from SEER 9, from 1973 to with in situ excluded. Site-specific codes were C34 for lung, 1992, as not all race/ethnic groups were available in SEER C18 for colon, C16 for stomach, and C73 for thyroid. For leu- during that time period. kemia, all malignant leukemia was included except CLL based Years 1992 to 2015 on the ICD-O-3/WHO 2008 definition (SEER 2019). Fig. A1 To use the most comprehensive data available in SEER, shows an example of the baseline cancer estimation by calen- incidence rates by age at diagnosis, calendar year, sex, and dar year for all cancer groups, by race/ethnicity, males and fe- race/ethnicity used SEER 13 from 1992–1999 and SEER males, for the 80–84-y age group. 18 from 2000–2015 (SEER13 2018, SEER18 2018).

Years1945to1972 Years 2016 to 2034 We used rates for all races/ethnicities combined from To obtain cancer rates for projections from 2016 to 1973 to 1987 from SEER 9 to extrapolate rates from 1945 to 2034 (inclusive), we used sex- and race/ethnicity-specific 1972 (Rühm et al. 2016). We modeled 5-y age-specific rates rates from SEER18, 2000 to 2015 (SEER18 2018). We fit for all cancers. As there were very few cancers observed in Poisson models including age in 5-y groups coded with the 0–4-y age groups, we combined that group with the – dummies, calendar year in single years (centered at 2007), 5 9-y age group and used 17 age groups in the modeling sex, and race/ethnicity and included interaction terms of for thyroid cancer, leukemia (except CLL), and all solid race/ethnicity with year, age, and sex for all sites. For all solid cancers (excluding thyroid and NMSC) combined. We fit cancers, we fit separate models for men and women that oth- Poisson regression models that included the age groups erwise had the same parametrization. Calendar year was fit coded with dummies, calendar year in single years treated with a linear trend (which seemed appropriate after some as a continuous variable fitted with a linear trend, sex and model checking). The coefficients from the Poisson models interaction terms of year by sex, and age groups by sex (17 de- were then used to project the rates. grees of freedom). For colon, lung and stomach cancer, we combined the 0–4, 5–9, and 10–14 y age groups into a single 15-y category and used 16 age categories. The terms in the APPENDIX B: LIFETABLES Poisson model were the same otherwise (interaction of sex by age group, year, interaction of sex by year). Availability of published US lifetables varied by calendar Race/ethnicity ratios using SEER rates for each site year and race/ethnicity. Separate male and female survival were computed using data from 1992 to 2006. We fit Poisson probabilities were available for each year and race/ethnicity in models to race/ethnicity specific counts and populations the table. We allowed for changes in survival probability across SEER 9 rates, adjusted for race/ethnicity [all race/ethnicities calendar years by updating the lifetables every decade, which combined (ALL), Black, Non-hispanic White, Native American, represented the period in which they became available. Race/ Hispanics], age (trend, single years), and calendar year (sin- ethnicity-specific lifetables were not available for non-Whites gle years, treated as a continuous variable) with “ALL” as until 1999. When a specific race/ethnicity was not available, www.health-physics.com Projected lifetime cancer risks c E. K. CAHOON ET AL. 491

Table B1. Selection of life tables by race/ethnicity for years Bootstrap for the baseline rates 1945 to 2034. The bootstrap computation for the baseline rates differed – Race/ethnicity by calendar time period. For the 1945 to 1972 and 2016 2034 African periods, we used a parametric bootstrap to account for the un- Lifetable American Native certainty in the predicted rates. We resampled coefficients for Ye a r s years White Hispanic (AA) American each of the Poisson models from a multivariate normal distri- 1945–1948 1939-41 White White White White bution that had as the mean the parameter estimates and the 1949–1958 1949-51 All races All races All races All races covariance matrix as estimated from the original model fit to 1959–1968 1959-61 White Non-white Non-white Non-white the SEER data. Similarly, we assumed that the estimates of 1969–1978 1969-71 White Non-white Non-white Non-white the race/ethnicity ratio parameters arose from a normal distri- 1979–1988 1979-81 White Non-white Non-white Non-white bution with known mean and covariance, and resampled 1989–1998 1989-91 White Non-white Non-white Non-white values from a multivariate normal distribution with that mean 1999–2008 1999-01 White All races AA All races and covariance matrix. These new estimates were used in the 2009–2015 2009 White Hispanic AA All races computation of the rates and the observed and excess numbers – 2016 2034 2015 White Hispanic AA All races of cancers. For the 1972 to 1991 period, we used SEER rates for we used lifetables for either non-Whites, all races, or what was the baseline rates, which we multiplied by a race/ethnicity available. For example, a lifetable based on data from 1939 ratio to obtain race/ethnicity specific rates. We thus ob- to 1941 among Whites was used for each race/ethnicity for tained bootstrap counts in cell I by resampling counts from the period 1945 to 1948. Table B1 summarizes the lifetables a Poisson distribution that has the mean the number of cases we used to estimate survival probability by calendar period in cell i reported in SEER, with the person years observed in and race/ethnicity. that cell. The race/ethnicity ratio uncertainty was again in- corporated by sampling the corresponding parameters as described for the 1945 to 1972 period. APPENDIX C: BOOTSTRAP VARIANCE For the 1992 to 2015 period, the age, sex, and race/ COMPUTATION ethnicity specific rates were directly obtained from SEER, We computed the variance of the following quantity, and we resampled counts from a Poisson distribution, as de- scribed for the 1972 to 1991 period, without any further race/ethnicity adjustment. ∑K * O ¼ i¼1personyearsðÞ i Bootstrap for ERR and EAR ½BiðÞ*ERRðÞ i *raceratio*w þ ðÞ1−w *EARðÞ i ; The uncertainty in the ERR and EAR computations ðC1Þ came from two different sources: the uncertainty in the param- eters that are used, and the uncertainty in the reconstructed where personyears(i) was the observed number of person-years dose. We used a parametric bootstrap to incorporate uncer- in a cell defined by precinct, sex, age, race, and calendar year tainty around parameter values in the ERR and EAR func- using a bootstrap procedure; w was the transfer weight; tions, given by eqn (1) (rewritten below for convenience), raceratio, the ratio of cancer incidence rates for different ERRðÞ D; s; e; a; t or EARðÞ D; s; e; t race/ethnicities; B(i) is the baseline cancer rates per cell; φ ; ERR(i) is the excess relative risk per cell; and EAR(i) is the ex- ¼ bsDexpðÞge þ ha þ dt þ et cess absolute risk per cell. We assumed that the person-years were fixed, as their variability is expected to be small com- where D = dose is in Gy, e = (exposure age-30)/10 for expo- pared to the variability in the baseline rates, the race/ethnicity sure age<30 and e =0forexposureage≥ 30, a =loge ratio (when incorporated) and the ERR and EAR functions. (attained age/60) and t =loge(time since exposure/25) where The variance computation for the baseline counts that time since exposure is attained age minus age at exposure. only used baseline rates, but not the EAR and ERR func- We used distributional assumptions based on recom- tions, proceeded along the same lines. Once we obtained mendations in the BEIRVII report (NRC 2005). For normally 1,000 bootstrap values of O, the 90% confidence interval (or log-normally) distributed parameters, we resampled values around the original estimate was calculated by taking the from a normal distribution that had as the true parameter 5th and 95th percentile of the bootstrap empirical distribution values the mean and variances given in the BEIR VII report. function as the lower and upper confidence limit, respec- We assumed that parameters bM and bF in the ERR and tively. Details for the individual component of eqn (C1) EAR models followed a log-normal distribution for every site follow next. except leukemia for which we sampled values from the normal www.health-physics.com 492 Health Physics October 2020, Volume 119, Number 4 distribution with the appropriate mean and variance and then The covariance matrices are: exponentiated them. For leukemia, the bM and bF parameters β β η γ were assumed to arise from a four-parameter beta distribution. CovERR(Log( M), Log( F), , )= We thus first applied the inverse normal density to them, and [0.031703 0.010922 0.011666 -0.021094 then the cumulative distribution function of a four-parameter 0.010922 0.017866 0.0094576 -0.011821 beta distribution. The parameters for the four–parameter beta 0.011666 0.0094576 0.010547 -0.023224 -0.021094 -0.011821 -0.023224 0.14091] distribution were obtained by solving a set of linear equations in the published confidence bounds and the mean, knowing CovEAR(Log(βM), Log(βF), η, γ)= that the lower interval bound was zero, to match up the inter- [0.029955 0.0094495 0.0093699 -0.0085991 vals given in BEIR VII (NRC 2005). 0.0094495 0.015269 0.0086235 -0.015065 For leukemia and all solid cancers, we additionally in- 0.0093699 0.0086235 0.009300 -0.020365 corporated the covariance between the parameter values. -0.0085991 -0.015065 -0.020365 0.10435] First, we drew all parameters for a multivariate normal dis- β β δ γ φ tribution and then exponentiated the first two components CovERR( M, F, , , )= for all solid cancers. Note that the covariances are not pre- [0.31306 0.24735 0.026282 0.026395 0.019503 0.24735 0.30365 0.020753 0.022953 0.015975 served after exponentiating. 0.026282 0.020753 0.11003 0.016966 0.057885 For all solid cancers, d = 0 and φ = 0 in eqn (1) for 0.026395 0.022953 0.016966 0.037307 0.013467 ERR. The remaining ERR parameters have means loge 0.019503 0.015975 0.057885 0.013467 0.069669] (bM)=−1.104; loge(bF)=−0.558; g = −0.3; h = −1.4.

For the EAR function for all solid cancers, d = 0 and φ = CovEAR(βM, βF, δ, γ, φ)= 0 in eqn (1) and the remaining ERR parameters have means [0.2568 0.22277 0 0.018272 -0.0029116 loge(bM)=−6.1; loge(bF)=−5.876; g = 2.779; h = 0.22277 0.24963 0 0.016169 -0.0021659 −0.4058. The covariance matrices for the parameters in 00000 the ERR and EAR functions were 0.018272 0.016169 0 0.021735 0.010381 -0.0029116 0.0021659 0 0.010381 0.015841]

CovERR(loge(bM), loge(bF) hg) 0.010922 0.011666 = [0.031703 Dose uncertainty −0.021094 To accommodate dose uncertainty in the estimates of the 0.010922 0.017866 0.0094576 numbers of excess cancers, we randomly drew a dose realiza- −0.011821 tion from a log-normal distribution that had as parameters the 0.011666 0.0094576 0.010547 mean dose and the variance computed from the 0.025th and −0.023224 0.975th quantile of the distribution for each voting precinct, − − − 0.021094 0.011821 0.023224 0.14091]; age group, and race/ethnicity. To examine the impact of dose CovEAR(loge(bM), loge(bF) hg) 0.0094495 0.0093699 = [0.029955 uncertainty on our 90% uncertainty intervals, we compared −0.0085991 results for the total population (years 1945 to 2034) that 0.0094495 0.015269 0.0086235 accounted for all sources of uncertainty to results that did not −0.015065 incorporate dose uncertainty. Results are presented in Table C1. 0.0093699 0.0086235 0.0093004 Table C1. Impact of dose uncertainty on 90% uncertainty interval for −0.020365 excess cancer cases. Thus, uncertainty in dose has a substantial impact −0.0085991 −0.015065 −0.020365 0.10435]. in the total population, 1945-2034. Length of 90% uncertainty interval For leukemias, h = 0 for the ERR and EAR functions. (5%, 95%) The remaining ERR parameters have means d = −0.4767, Without dose φ = 0.4211, loge(bM) = 0.05572; loge(bF)=0.1631;g = Cancer group With dose uncertainty uncertainty % change −0.4011 for ERR. For the EAR function the parameters Leukemia (except CLL) 8 6 25% have means loge(bM) = 0.485; loge(bF)=−0.0703; g = φ Thyroid 445 224 50% 0.2865 and = 0.557. Colon 22 14 36% To obtain realizations of a four-parameter beta distribu- Stomach 16 11 31% tion, we first compute X from applying a normal distribution Lung 41 21 49% with mean bM and variance given by the first diagonal term All Solid (except thyroid 246 147 40% in the ERR covariance matrix to bM andthenapplyingthein- and NMSC) verse cumulative distribution function of a four-parameter gamma distribution to the so-transformed values of X. ■■

www.health-physics.com Projected lifetime cancer risks c E. K. CAHOON ET AL. 493

Fig. A1. Projected (1945-1972 2016-2034) and observed (1973-2015) baseline cancer rates per 100,000 person-years by year for individuals aged 80-84 y smoothed using loess regression.

www.health-physics.com Paper

The Likelihood of Adverse Pregnancy Outcomes and Genetic Disease (Transgenerational Effects) from Exposure to Radioactive Fallout from the 1945 Trinity Atomic Bomb Test

John D. Boice, Jr.1,2

INTRODUCTION Abstract—The potential health consequences of the Trinity nuclear weapon test of 16 July 1945 at Alamogordo, New Mexico, are ON 16 July 1945, the first detonation of a nuclear device oc- challenging to assess. Population data are available for mortality curred at the Trinity site near Alamogordo, New Mexico (in but not for cancer incidence for New Mexico residents for the first 25 y after the test, and the estimates of radiation dose to the nearby Socorro County), on what is now part of White Sands population are lower than the cumulative dose received from Missile Range. Robert Oppenheimer, the director of Los ubiquitous natural background radiation. Despite the estimates Alamos National Laboratory at the time, was inspired by of low population exposures, it is believed by some that cancer “ ” rates in counties near the Trinity test site (located in Socorro the poetry of John Donne to assign the code name Trinity County) are elevated compared with other locations across the to the test. The test was of an implosion-design plutonium state. Further, there is a concern about adverse pregnancy outcomes device, informally nicknamed “The Gadget.” Among those and genetic diseases (transgenerational or heritable effects) related present at the test, 396 were commissioned officers or to population exposure to fallout radiation. The possibility of an intergenerational effect has long been a concern of exposed enlisted men in the US Army; they have been studied for populations, e.g., Japanese atomic bomb survivors, survivors of late effects as nuclear weapons test participants (Till et al. childhood and adolescent cancer, radiation workers, and 2014; Boice et al. 2019a). environmentally exposed groups. In this paper, the likelihood of discernible transgenerational effects is discounted because (1) in Population data are available for mortality for New all large-scale comprehensive studies of exposed populations, no Mexico residents (but not for cancer incidence) for the first heritable genetic effects have been demonstrated in children of 25 y after the test, and the estimates of radiation dose to the exposed parents; (2) the distribution of estimated doses from population (except that for the thyroid gland) are practically Trinity is much lower than in other studied populations where no transgenerational effects have been observed; and (3) there is no all <10 mGy (Simon et al. 2020), lower than the cumulative evidence of increased cancer rates among the scientific, military, external dose received from ubiquitous natural background and professional participants at the Trinity test and at other radiation of approximately 80 mGy over a 40 y period. Despite nuclear weapons tests who received much higher doses than New these estimates of low population exposures, concerns Mexico residents living downwind of the Trinity site. Health Phys. 119(4):494–503; 2020 have been raised by the citizens living in the vicinity of the Trinity test that the fallout radiation has caused Key words: epidemiology; fallout; health effects; nuclear weapons increased rates of cancer and transgenerational effects, i.e., genetic and adverse pregnancy outcomes (APOs) 1National Council on Radiation Protection and Measurements, (TBDC 2017). Bethesda, MD; 2Vanderbilt University Department of Medicine, Division The possibility of intergenerational effects has long of Epidemiology, Nashville, TN. The author declares no conflicts of interest. been a concern of exposed populations, e.g., Japanese For correspondence contact John D. Boice, Jr., National Council on atomic bomb survivors and other exposed groups. However, Radiation Protection and Measurements, 7910 Woodmont Avenue, Suite there is little to no convincing or consistent evidence among 400, Bethesda, MD 20814-3095, or Vanderbilt University Department of Medicine, Division of Epidemiology, 2525 West End Avenue, the offspring of environmentally exposed populations; Nashville, TN 37203-1738, or email at [email protected]. childhood, adolescent, and young adult cancer survivors; (Manuscript accepted 10 August 2019) 0017-9078/20/0 Japanese atomic bomb survivors; or radiation-exposed Written work prepared by employees of the Federal Government as workers for an excess of malformations, stillbirths, neonatal part of their official duties is, under the U.S. Copyright Act, a "work of deaths, cancer, cytogenetic syndromes, single-gene disorders, the United States Government" for which copyright protection under Title 17 of the United States Code is not available. As such, copyright does not or cytogenetic markers that would indicate an increase of her- extend to the contributions of employees of the Federal Government. itable genetic mutations in the exposed parents (UNSCEAR DOI: 10.1097/HP.0000000000001170 2001; COMARE 2004, 2016; Nakamura 2006; NA/NRC 494 www.health-physics.com Pregnancy outcomes and genetic disease following exposure c J.D. BOICE,JR. 495 2006; Fujiwara 2008; Winther and Olsen 2012; NCRP 2013; The section begins, however, with an overview of studies Brent 2015; Grant et al. 2015). of cancer risk among environmentally exposed populations Radiation clearly induces mutations in somatic cells of in New Mexico and among nuclear weapons test participants rodents and humans, and transgenerational (heritable) ef- present at the Trinity detonation. It is generally accepted that fects are established from experimental studies conducted a cancer risk in exposed populations is much more likely in the 1950s and 1960s of irradiated Drosophila and mice to be detected than a transgenerational (inheritable) risk, (UNSCEAR 2001; NA/NRC 2006; NCRP 2013). Thus, which has not been seen in the children of exposed parents the possibility of human germ-cell mutation following radi- (IOM 1995; NCRP 2013; Brent 2015; NA/NRC 2006). ation is recognized and considered by radiation protection committees (ICRP 2007; NCRP 2018b). However, the abil- Studies of cancer risk among environmentally exposed ity to establish an association between parental exposure populations and transgenerational effects in humans, if one exists, is in Studies of environmentally exposed populations and the future and would be related to advances in genetic tech- cancer risk in New Mexico. The studies described below are nologies (NCRP 2013; Brent 2015). It is noteworthy that the of New Mexico residents who lived near radiation facilities “mega-mice” studies involved nearly 7 million rodents, such as Los Alamos National Laboratory or the uranium which suggest the enormity of a comparable human investi- mill in Grants, New Mexico. The potential for exposure gation. Further, the lack of clear and convincing evidence was to any atmospheric release of radioactive material for transgenerational effects in human studies conducted during plant operation or to environmental contamination since the 1960s has reduced the level of concern of heritable or ingestion of any waste associated with uranium milling. effects (Fig. 1) (Hall 2009), and radiation protection com- Stebbings and Voelz (1981) examined both cancer mittees have reduced the genetic component assigned to mortality and incidence data from the New Mexico the radiation health detriment (ICRP 2007; NCRP 2018b). Tumor Registry for Los Alamos County, New Mexico, The experimental and human studies support the notion that where Los Alamos National Laboratory is located. They if transgenerational effects occur in humans, they are too found a suggestive excess mortality from leukemia, but small to be detected by epidemiologic study (IOM 1995). there was no parallel increase in leukemia incidence. There were suggestions of excesses in neoplasms of the METHODS reticuloendothelial system in the early years and of the colon and rectum; the latter were thought to be Human studies of the children of radiation-exposed explainable in terms of socioeconomic factors. There parents are discussed; specifically, studies of the offspring was no conclusive evidence of cancer risk among the of environmentally exposed populations; childhood, adoles- residents near these radiation facilities. cent, and young adult cancer survivors; atomic bomb survi- These observations of cancer risks among populations vors; and radiation-exposed workers. The studies sought to living near nuclear facilities in New Mexico are based on identify any excess of malformations, stillbirths, neonatal small numbers but are consistent with the much larger deaths, cancer, cytogenetic syndromes, single-gene disor- study conducted by the National Cancer Institute (NCI) of ders, or cytogenetic markers that would indicate an increase cancer risk among populations living near nuclear facilities of heritable genetic mutations in the exposed parents. throughout the United States (Jablon et al. 1991). A special scientific advisory committee of nongovernment scientists was established by NCI to provide guidance and oversight over the study. The committee concluded “that the survey produced no evidence that an excess occurrence of cancer had resulted from living near nuclear facilities. Further, that the measurements of radioactive releases from nuclear facilities indicate that the dose from routine operations is generally much below natural background radiation, and hence are unlikely to produce observable effects on the health of surrounding populations” (Jablon et al. 1990). Boice et al. (2010) examined both cancer incidence and mortality in populations living near uranium milling and mining operations in Grants, Cibola County, New Mexico, during 1950–2004. Lung cancer mortality and incidence Fig. 1. Schematic of how the level of concern about transgenerational (heritable) risks has decreased from the 1950s to the present as more in- were significantly increased among men but not women, formation from human studies has become available (Hall et al. 2009). and the excess was attributed to a previously reported risk www.health-physics.com 496 Health Physics October 2020, Volume 119, Number 4 of lung cancer among underground miners living in Grants observed rates with the general population and 95% and exposed to radon gas and its decay products (Boice et al. confidence intervals computed. Cox proportional hazards 2008). Stomach cancer mortality and incidence were both models were used to analyze leukemia and lung cancer significantly increased among women but not men. These dose response. Because only 3 leukemia deaths were due to excesses seem unlikely to be related to uranium milling leukemia other than chronic lymphocytic leukemia (CLL), and mining activities since the elevated risks were greatest a malignancy not considered to be increased following in the years before uranium mills and mines operated in radiation exposure (UNSCEAR 2008; Leuraud et al. 2015), Cibola County; furthermore, the stomach cancer rates the internal analyses could not be conducted of Trinity partic- decreased over time to normal levels. ipants but only of the entire cohort, which included 717 leu- Studies of nuclear weapons test participants at Trin- kemia deaths other than CLL and 8,027 lung cancer deaths. ity and other series, and cancer effects. To provide a Among the 396 Trinity participants, 319 (or 81%) had different look at the possibility that adverse pregnancy out- died, and the all cause of death SMR was 0.71 (95% confi- – comes or genetic disease might occur among the children dence interval [CI]: 0.63 0.79) (Table 2). Cancer mortality of parents exposed to fallout from the Trinity detonation, also was below expectations but not significantly so (SMR – the dose distributions and the mortality experience of the 0.95; 95% CI: 0.77 1.16). The dose distribution of Trinity 396 atomic veterans present at the Trinity shot is evaluated. participants was similar to that of all 113,806 participants This study of persons who were present at the Trinity detona- (Table 1). The mean dose to red bone marrow was 9 mGy tion in 1945 and who were followed through 2010 provides (maximum 35 mGy) and higher than the estimated red bone information on cancer risk at higher doses than were re- marrow doses received by New Mexico residents living ceived by residents living near the Trinity site. Those present near the Trinity site (Simon et al. 2020). No excess of at the Trinity test included Robert Oppenheimer, General leukemia, excluding CLL, or any other cancer was observed Leslie Groves, Hans Bethe, Enrico Fermi, Theodore Hall, among test participants at Trinity. The internal dose-response Louis Hempelmann, Hymer Friedell, Richard Feynman, and analyses for all 113,806 test participants did not show an Kenneth Bainbridge. The Trinity detonation was part of a increase for leukemia (excess relative risk [ERR] at 95% CI − − larger study of 113,806 nuclear weapons test participants con- for 100 mGy = 0.35 [ 1.05, 0.34], n = 717) or for lung − ducted within the Million Person Study of Low-Dose Health cancer (ERR at 95% CI for 100 mGy = 0.04 [ 0.11, 0.19], Effects (MPS) (Bouville et al. 2015; Boice et al. 2019a). Dose n = 8,027) (Boice et al. 2019b). estimates for the Trinity participants and all atomic veterans To place these analyses in perspective with regard to were determined for all participants at one of eight test series transgenerational effects, the Institute of Medicine (IOM (Till et al. 2014, 2018; Beck et al. 2017; Dauer et al. 2018; 1995) evaluated the likelihood that an epidemiologic study NCRP 2018a), and doses to red bone marrow are presented could detect an increase in heritable genetic effects among in Table 1. Extensive follow-up procedures located over 95% the children of atomic veterans (had there been an increase) of the cohort (Mumma et al. 2018) and identified a cause of and concluded that it was not possible. In the absence of any death for practically all known to have died. radiation effects, 15,000 newborn children with major birth Standardized mortality rates (SMRs) were computed defects would be expected to be diagnosed at birth among for all Trinity and other test series participants (Table 1 the estimated 500,000 offspring of 210,000 atomic veterans. – provides a listing of the eight test series) to compare An additional 3 5% of these children would be expected to be diagnosed with a major congenital anomaly in the first Table 1. Bone marrow dose distributions for atomic veterans at the 10 y of life. Thus, the study size would have to be enormous, a Trinity test and for all veterans in the Eight Series Study within the and controlling for confounding influences would be nearly Million Person Study.b impossible. Further, the gonadal doses to produce a possible Bone marrow dose (2-y lag) increase in transgenerational effects also would have to be Dose (mGy) Veterans at Trinity Total number of veterans very high. “Relatively high doses of radiation (greater than <2.5 225 (57%) 58,203 (51%) 2,000 mSv [200 rem]) would add only a small number of 2.5 to ≤5 27(7%) 19,092(17%) additional cases of genetic disorders to the large number 5to≤10 32 (8%) 17,050 (15%) that are expected to occur as a result of spontaneous muta- 10 to ≤25 53 (13%) 14,195 (12%) tions, most of which have existed in the population for many >25 59(15%) 5,266(5%) generations” (IOM 1995). Such high doses are not observed Total 396 113,806 among atomic veterans (if they had occurred, deterministic a The Eight Series Study included military aboveground weapon test partici- effects would have been evident) nor among the residents pants present at one of these eight test series (year): Upshot-Knothole (1953), near the Trinity site exposed to fallout radiation. Plumbbob (1957); Crossroads (1946); Greenhouse (1951); Castle (1954); Redwing (1956); Hardtack I (1958); Trinity (1945) (Till et al. 2014, In summary, nuclear weapons test participants received 2018); bBeck et al. 2017. the highest radiation doses of any population from nuclear test www.health-physics.com Pregnancy outcomes and genetic disease following exposure c J.D. BOICE,JR. 497 Table 2. Standardized mortality ratios (SMRs) and 95% confidence from a nuclear fuel reprocessing plant (Dean et al. 2000). intervals (CIs) among 396 atomic veterans at the Trinity site followed The radiation sources in the environment include thorium- from 1945 through 2010 with 18,884 person-y of follow-up.a containing monazite sands and effluents from nuclear facil- Trinity ities (NCRP 2013). No transgenerational effects have been Veterans at risk 396 demonstrated among people exposed to fallout from the Person-y of follow-up 18,884 Chernobyl reactor accident (WHO 2006). Cause of death (ICD9)a Observed SMR 95% CI Studies of Down syndrome and other genetic anoma- lies in populations living in high background radiation areas – c – All causes of death (001 999) 319 0.71 0.63 0.79 are mostly ecological and are limited because individual – – All malignant neoplasms (140 208) 97 0.95 0.77 1.16 doses and potential confounding influences are unknown. Colon (153) 9 0.99 0.45–1.88 An increased rate of Down syndrome among residents in Bronchus, trachea, and lung (162) 21 0.64c 0.40–0.98 areas of high background radiation in China was later attrib- Prostate (185) 16 1.3 0.74–2.11 Non-Hodgkin’s lymphoma (200, 202) 6 1.59 0.58–3.46 uted to increased maternal age at birth and to better case Leukemia and aleukemia (204–208) 5b 1.22 0.39–2.85 ascertainment in the high background radiation areas Smoking-related cancers (140–150, compared with the control areas (UNSCEAR 1993; Wei 33 0.67c 0.46–0.94 157, 161–162, 188–189) et al. 1990). Airborne radiation released from the Sellafield Non-smoking-related cancers 64 1.22 0.94–1.56 nuclear fuel reprocessing plant in England was claimed to Diabetes (250) 8 0.94 0.40–1.85 have caused a cluster of Down syndrome on the coast of Mental and behavioral disorders 6 0.8 0.29–1.74 Ireland but was later discounted (Dean et al. 2000). Studies (290-319) of 140,000 inhabitants residing in Kerala, India, in areas of Dementia and Alzheimer’s disease 10 0.87 0.42–1.60 (290.0-290.4, 331.0) high natural background radiation (15 to 25 mGy annual Diseases of the nervous system whole-body dose) reported increased rates of Down syn- 10 0.69 0.33–1.27 (320–389) drome (Kochupillai et al. 1976), which were not confirmed All heart disease (390–398, 404, 79 0.47c 0.37–0.59 in subsequent studies that used more reliable sources of in- 410–429) formation (Kesavan 1997). No correlation between increased Ischemic heart disease (410–414) 64 0.50c 0.39–0.64 c levels of natural background radiation and malformation, Cerebrovascular disease (430–438) 16 0.60 0.34–0.97 Nonmalignant respiratory disease stillbirth, or twinning was found in a comprehensive study 31 0.69c 0.47–0.98 (460–519) of over 40,000 newborn children and stillbirths in Kerala Bronchitis, emphysema, asthma (490–493) 9 0.66 0.30–1.25 (Jaikrishan et al. 1999). High natural background radiation Nephritis and nephrosis (580–589) 6 1 0.37–2.18 levels in Kerala also were not correlated with increases in All external causes of death (800–999) 10 0.43c 0.21–0.79 mental retardation, cleft lip, or cleft palate (Koya et al. Unknown causes of death 30 —— 2012). Clusters of Down syndrome in Germany were re- a SMRs shown only for causes of death with five or more cases. ICD9: Interna- ported just after the Chernobyl accident, but low-dose radi- tional Classification of Diseases, 9th Revision. ation was not considered a contributing cause (Little 1993; b 2 of the 5 leukemia deaths were due to chronic lymphocytic leukemia. Burkart et al. 1997). Offspring of residents in Kerala were c Denotes statistical significance at the p < 0.05 level, i.e., the 95% confidence limit does not contain 1.00. reported to have certain inherited genomic changes to mito- chondrial DNA (Forster et al. 2002) and to the Y chromosome detonations in the United States. These doses are still relatively (Premi et al. 2009), but results have not been replicated, and low but higher than those estimated for the populations living there is uncertainty as to the gonadal dose received by par- near or downwind of the Trinity site (Simon et al. 2020). No ents and the adequacy of the control groups. dose-related increases in cancer were observed among the Reproductive and hereditary effects have not been Trinity participants or among the large series of 113,806 demonstrated among people exposed to fallout from the atomic veterans. The low doses estimated for the smaller Chernobyl accident nor are any expected (Little 1993; number of New Mexico residents near or downwind of the WHO 2006). No effects on fertility, numbers of stillbirths, Trinity detonation site point to the implausibility that cancer or adverse pregnancy outcomes have been attributed to radi- or transgenerational (heritable) effects occurred in excess or ation, in large part because of the low radiation doses re- could be observed in any epidemiologic investigation. ceived. A modest but steady increase in reported congenital malformations in both contaminated and uncontaminated Studies of environmentally exposed populations areas of Belarus appeared related to improved reporting and and transgenerational effects. Transgenerational studies not to radiation exposure (WHO 2006). “An increased fre- have been conducted in areas of high natural background ra- quency of trisomy 21 in Berlin in January 1987, and in- diation in India (Jaikrishan et al. 1999) and China (Wei et al. creases in the frequency of neural tube defects in several 1990) and in areas in Ireland exposed to airborne releases small hospital-based series in Turkey, were not confirmed www.health-physics.com 498 Health Physics October 2020, Volume 119, Number 4 in larger and more representative series in Europe. No clear evidence of radiotherapy-related induction of persistent ge- changes in the prevalence at birth of anomalies which might nomic instability (Tawn et al. 2005). G2 chromosomal ra- be associated with the accident are apparent in Byelorussia diosensitivity evaluations were inconclusive but confirmed or the Ukraine, the republics with the highest exposure to that the radiosensitivity phenotype is heritable (Curwen fallout” (Little 1993). et al. 2010). Polymorphic variation in DNA repair genes These studies of environmentally exposed populations showed statistically significant genotype differences between provide little to no evidence for transgenerational effects. survivors and their partners for the APEX Asp148Glu site, but this initial observation was not confirmed in subsequent Studies of the offspring of childhood, adolescent, studies (Curwen et al. 2011; Wilding et al. 2007). No and young adult cancer survivors treated with radiation transgenerational effects of maternal exposure to cancer treat- and transgenerational effects. In 2006, the Committee to ment were seen in an evaluation of mutations in mitochon- Assess Health Risks from Exposure to Low Levels of Ioniz- drial DNA, but the size of the population was small (Guo ing Radiation (Biological Effects of Ionizing Radiation et al. 2012). [BEIR] VII committee) concluded that studies on the ge- Family blood studies also were used to examine pos- netic effects of radiotherapy on childhood cancer should sible radiation-induced germline minisatellite mutations. be encouraged (NA/NRC 2006). Subsequently, heritable Minisatellite mutations at hypervariable loci are tandemly disease among the children of cancer survivors treated with repeated regions of DNA which occur at a high frequency radiation in four countries was extensively evaluated in a throughout the genome. Some repeat DNA sequences ex- large-scale international collaboration, the genetic conse- hibit high frequencies of spontaneous germline mutations quences of cancer therapy study (Boice et al. 2003; NCRP to new allele lengths (up to 1,000 times more frequent than 2013). Over 35,800 children of 21,205 cancer survivors mutations in genes that code for proteins), and screening for were conceived after therapy had ended. The parents in length changes may indicate radiation-induced germline Denmark and Finland were cancer survivors diagnosed un- mutations using relatively small population samples. No der 35 y of age; the parents of US and Canadian cancer sur- convincing or consistent evidence has been found, however, vivors were under 20 y of age at cancer diagnosis. Estimates that radiation causes germline mutations based on changes of gonadal doses of radiation were based on original radiation- in minisatellite lengths among cancer survivor families therapy records and phantom reconstructions (Stovall et al. (mean parental gonadal dose ~500 mGy) or other exposed 2004). No associations between birth defects and gonadal populations (Tawn et al. 2011, 2015; Little et al. 2013; doses were found (Mulvihill et al. 2009; Green et al. 2009; NCRP 2013, 2015). Signorello et al. 2010; Winther et al. 2012). The mean testic- These and other studies (Byrne et al. 1998; Green et al. ular dose for men was 500 mGy, and the mean ovarian dose 2009) of the children of childhood, adolescent, and young for women was 1,200 mGy. High therapeutic doses to the adult cancer survivors treated with radiation provide little uterus of female cancer survivors was found to increase the to no evidence for transgenerational effects. rates of spontaneous abortions (miscarriages), preterm births, and stillbirths; these adverse pregnancy outcomes were at- Studies of the offspring of Japanese atomic bomb tributed to a somatic (deterministic) effect from a damaged survivors and transgenerational effects. The Japanese uterus and not to a genetic or heritable effect of the radiation atomic bomb survivor study was initially focused on evalu- exposure (Signorello et al. 2006; Winther et al. 2008). A ating and quantifying the risk of genetic disease associated small difference was reported for cytogenetic abnormalities with parental exposures received during the 1945 bombings (e.g., Down syndrome [relative risk, RR = 1.1] and Turner of Hiroshima and Nagasaki (Neel and Schull 1991). Nearly syndrome [RR = 1.3]) among the children of Danish cancer 80,000 children born to parents exposed to the atomic bombs survivors compared with the children of their siblings but were evaluated. The measures of possible transgenerational was not statistically meaningful (Winther et al. 2004). An effects included malformations, chromosomal abnormalities, altered sex ratio among the live-born children of cancer stillbirths, neonatal deaths, cancer, chromosomal transloca- survivors treated with radiation therapy was not observed tions, mutations in minisatellites, and multifactorial disease and provided no support for a possible transgenerational (Schull 2003; Fujiwara et al. 2008). Because experimental or germline effect (Winther et al. 2003). studies of Drosophila and rodents had established that radia- Molecular analyses of cancer family blood samples tion can cause heritable effects, it was surprising that there (blood taken from the irradiated cancer survivor, the spouse was no evidence for any significantly increased risk for or partner, and at least one child) have been used to study a any measure of genetic disease. Although there were no sta- number of mechanistic processes possibly related to trans- tistically significant findings, most of the measures of generational effects and cancer susceptibility. Analyses of transgenerational effects were in the direction of a positive unstable chromosome aberrations, however, provided no effect. The mean conjoined gonadal dose was of the order www.health-physics.com Pregnancy outcomes and genetic disease following exposure c J.D. BOICE,JR. 499 of 360 mGy, and it was estimated that the doubling dose Neel 1999b; Tawn 1995; UNSCEAR 1994; Wakeford 2000; (DD, the dose to a population that would produce the same COMARE 2004, 2016). Then a cohort study confirmed amount of genetic damage as occurs spontaneously each the statistical association between preconception radia- generation) is of the order of 2 Gy for acute exposures and tion of Sellafield workers and leukemia and lymphoma of the order of 4 Gy for chronic exposures, i.e., quite high (Dickinson and Parker 2002), but it was not an indepen- doses (Neel 1998, 1999a; Schull 2003). Over the years, there dent test of the hypothesis since it included the same cases has been a shift from concern over genetic effects (to future previously studied by Gardner et al. (1990). Dickinson generations) to concern about the individual and the subse- et al. (2003) had raised a number of important concerns quent development of cancer, a somatic effect (Fig. 1). about the original case-control study by Gardner et al. The Japanese atomic bomb survivor study of heritable (1990). An infectious agent associated with a high level effects is the most comprehensive of all human studies en- of population mixing was raised as a possible explanation gaged in examining the consequences of preconception irra- (Kinlen 1995, 2015; Sorahan et al. 2003). diation (Schull et al. 1981; Neel and Schull 1991; Satoh The possibility that an increase in minisatellite germline et al. 1996; Neel 1998; Izumi et al. 2003; Schull 2003; mutations following parental exposure could be related to NA/NRC 2006; Nakamura 2006; Fujiwara et al. 2008; leukemia was discounted when no increase in inherited Grant et al. 2015). A broad range of gonadal doses were ex- germline minisatellite mutations were found in children amined with respect to many indicators of genetic damage with leukemia (Davies et al. 2007) nor among workers at (NCRP 2013): (1) untoward pregnancy outcomes (i.e., still- the Sellafied nuclear fuel reprocessing plant (Tawn et al. born, neonatal death, major congenital malformation); (2) 2015). There was no convincing evidence that parental cancer in the offspring; (3) early death among offspring occupational exposure was related to increases in child- (Grant et al. 2015); (4) chromosomal aberrations; (5) fre- hood cancer in the children of US radiologic technolo- quency of sex-chromosome aneuploids, i.e., having one or gists (Johnson et al. 2008). more chromosomes above or below the normal number; Studies of workers at the Sellafield nuclear fuel (6) frequency of mutation-altering protein change or function reprocessing plant have reported a statistical association (electrophoretic mutations); (7) growth and development between paternal preconception exposure and stillbirth of the F1 offspring population; (8) inherited mutations in (Parker et al. 1999), which was questioned by Abrahamson minisatellite DNA (Kodeira et al. 2004); and (9) multifac- and Tawn (2001) at the time but also was not consistent with torial disease (Fujiwara et al. 2008). a larger study of workers in the UK nuclear industry (Doyle These comprehensive studies of the children of Japanese et al. 2000) or with the atomic bomb survivors study (Little atomic bomb survivors provide little to no evidence for 1999; Otake et al. 1990) or with studies of the children of transgenerational effects. The absence of detectable in- cancer survivors (Mulvihill et al. 2009; Signorello et al. creases in any measure of transgenerational effect is notable 2010; Winther et al. 2012). Maternal factors were not in light of similar findings in large-scale studies of the chil- considered (Boice et al. 2000). Further, an increase in dren of cancer survivors and to a lesser effect, in the studies minisatellite germline mutations following worker expo- of environmentally exposed and radiation-exposed workers. sures was not found in their children (Tawn et al. 2015). No association was found between preconception dose and Studies of the offspring of radiation-exposed workers congenital malformations among the children of workers in and transgenerational effects. Studies of the children of nu- the Canadian nuclear power industry (Green et al. 1997). A clear radiation workers and x-ray technologists are described study of preconception radiation among Hanford workers but are limited by small sample sizes, low gonadal doses, evaluated 12 major congenital anomalies, including Down minimal dosimetric information, or inadequate comparison syndrome (Sever et al. 1988a). There was no evidence for a groups (NCRP 2013). radiation association overall, except for neural tube defects, A cluster of leukemia and non-Hodgkin’s lymphoma which was based on only three cases and which was not con- in young people living in the village of Seascale, Cumbria, firmed by Doyle and colleagues in the United Kingdom UK, was reported in 1983 by a team of investigative televi- (2000). To further test the Gardner hypothesis, Sever et al. sion reporters (Black 1984). A subsequent case-control (1997) evaluated childhood cancers around three US De- study by Gardner et al. (1990) reported an association between partment of Energy nuclear facilities: Hanford Site; Idaho preconception irradiation and leukemia and non-Hodgkin’s National Engineering Laboratory; and the K-25, Y-12, and lymphoma in children of male workers at the Sellafield nu- X-10 plants at Oak Ridge National Laboratory. No statisti- clear fuel reprocessing plant adjacent to Seascale. Further cally meaningful associations were found between paternal studies failed to confirm that low doses to the testes re- exposure and childhood leukemia, leukemia plus non- ceived before conception is a cause of cancer (Doll et al. Hodgkin’s lymphoma, central nervous system, or all child- 1994; Kinlen 1993; Kinlen et al. 1993; Little et al. 1996; hood cancers. Studies of medical radiographers are in large www.health-physics.com 500 Health Physics October 2020, Volume 119, Number 4 part negative with respect to adverse inherited outcomes but the atomic bombs, cancer survivors treated with radiother- are hampered by inadequate dosimetry (Boice et al. 1992; apy who were able to have children, and the nuclear Roman et al. 1996). The low gonadal doses in most occupa- weapons test participants present at the Trinity and other tional studies preclude statistically powerful evaluations. tests. Given the absence of evidence for transgenerational effects among the ~80,000 children of Japanese atomic DISCUSSION bomb survivors and among the ~36,000 children of cancer survivors in the United States, Canada, Denmark, and Transgenerational studies in humans Finland, and the enormity of the dose needed to detect an ef- The possibility of transgenerational effects following fect had there been one (noting that the birth prevalence of radiation exposure has been a concern for over 70 y and has major congenital malformation is ~3%), it is not scientifi- been studied in atomic bomb survivors, survivors of childhood cally or biologically plausible that the low doses experienced and adolescent cancer, radiation workers, and environmentally from the Trinity fallout could result in transgenerational ef- exposed groups. No radiation-related transgenerational effects, fects in the children of exposed residents near the Trinity site. hereditary diseases, adverse pregnancy outcomes, germline minisatellite mutations, stillbirths, neonatal deaths, cancer, Acknowledgments—The research related to nuclear weapons test participants early death, chromosome aberrations, mitochondrial DNA was supported in part by contracts and grants from NCI (grant no. U01 CA137026) and the US Department of Energy (grant no. DE-SC0008944 changes, cytogenetic abnormalities, single-gene disorders, awarded to the National Council on Radiation Protection and Measurements), or any measure of transgenerational effect has been convinc- and a discovery grant from the Vanderbilt-Ingram Cancer Center (grant no. ingly or consistently demonstrated in any human population 404-3 57-9682). Publication costs for this paper were supported by Intramural Research Program of the NCI, National Institutes of Health, and the Intra- exposed to ionizing radiation before conception (UNSCEAR Agency Agreement between the National Institute of Allergy and Infectious 2001; NA/NRC 2006; NCRP 2013; Brent 2015). Diseases (NIAID) and NCI (NIAID agreement #Y2-Al-5077 and NCI agree- The average gonadal doses have been large, of the order ment #Y3-CO-5117). of 300 mGy among atomic bomb survivors, 500 mGy among male cancer survivors, and 1,200 mGy among female cancer REFERENCES survivors. These doses are much higher than those estimated Abrahamson S, Tawn EJ. Risk of stillbirth in offspring of men ex- posed to ionising radiation. J Radiol Protect 21:133–144; 2001. for participants at the Trinity test site or for the New Mexico Beck HL, Till JE, Grogan HA, Aanenson JW, Mohler HJ, Mohler residents exposed to fallout from the Trinity detonation. SS, Voillequé PG. Red bone marrow and male breast doses for The experimental and human studies support the notion a cohort of atomic veterans. Radiat Res 187:221–228; 2017. that if radiation causes transgenerational effects in human Black D. Investigation of the possible increased incidence of can- cer in West Cumbria. Report of the Independent Advisory populations, they are too small to detect by epidemiologic Group. London: HMSO; 1984. Available at https://archive. study (IOM 1995; NA/NRC 2006). Many birth anomalies org/details/op1278204-1001. Accessed 27 July 2019. occur (about 3%) from all causes, so that even high gonadal Boice JD Jr, Mandel JS, Doody MM, Yoder RC, McGowan R. doses of the order of 2,000 mGy would be expected to add A health survey of radiologic technologists. Cancer 69: 586–598; 1992. only a small number of additional cases of genetic disorders Boice JD Jr, Robison LL, Mertens A, Stovall M, Green DM, to the large number that are expected to occur as a result of Mulvihill JJ. Stillbirths and male irradiation. J Radiol Protect spontaneous mutations or other factors. Recognizing that it 20:321–322; 2000. would be close to impossible to control for the known, much Boice JD Jr, Tawn EJ, Winther JF, Donaldson SS, Green DM, Mertens AC, Mulvihill JJ, Olsen JH, Robison LL, Stovall M. less the unknown, confounding influences for birth anoma- Genetic effects of radiotherapy for childhood cancer. Health lies, the Institute of Medicine (1995) concluded that a valid Phys 85:65–80; 2003. epidemiologic study of atomic veterans on transgenerational Boice JD Jr, Cohen SS, Mumma MT, Chadda B, Blot WJ. A co- effects could not be conducted. hort study of uranium millers and miners of Grants, New Mexico, 1979–2005. J Radiol Protect 28:303–325; 2008. One of several possible explanations for the absence of Boice JD Jr, Mumma MT, Blot WJ. Cancer incidence and mortality radiation-related genetic effects in humans is a biological in populations living near uranium milling and mining filtering process; i.e., the germ cells that produce sperm or operations in Grants, New Mexico, 1950–2004. Radiat Res ova are so damaged by radiation that the body’snaturalpro- 174:624–636; 2010. Boice JD Jr, Cohen SS, Mumma MT, Ellis ED. The Million cesses filter out any defective embryo, leading to only a low Person Study, whence it came and why. Int J Radiat Biol [online]. chance of children being born with birth defects (Brent 2019a. Available at https://www.tandfonline.com/doi/abs/10. 1994; NCRP 2013). 1080/09553002.2019.1589015?journalCode=irab20. Accessed 28 September 2019. CONCLUSION Boice JD Jr, Ellis ED, Golden AP, Zablotska LB, Mumma MT, Cohen SS. Sex-specific lung cancer risk among radiation workers in the Million Person Study and among TB-fluoroscopy patients. The Trinity nuclear test of 16 July 1945 resulted in a Int J Radiat Biol [online]. 2019b. Available at https://www. much lower dose from radioactive fallout to the surrounding tandfonline.com/doi/abs/10.1080/09553002.2018.1547441? population than was experienced by Japanese survivors of af=R&journalCode=irab20. Accessed 28 September 2019. www.health-physics.com Pregnancy outcomes and genetic disease following exposure c J.D. BOICE,JR. 501 Bouville A, Toohey RE, Boice JD Jr, Beck HL, Dauer LT, risk of leukaemia/non-Hodgkin’s lymphoma in relation to Eckerman KF, Hagemeyer D, Leggett RW, Mumma M, paternal preconceptional irradiation. J Radiol Protect 23: Napier B, Pryor KH, Rosenstein M, Schauer DA, Sherbini S, 385–403; 2003. Stram DO, Thompson JL, Till JE, Yoder C, Zeitlin C. Dose Doll R, Evans HJ, Darby SC. Paternal exposure not to blame. Na- reconstruction for the million worker study: status and guide- ture 367:678–680; 1994. lines. Health Phys 108:206–220; 2015. Doyle P, Maconochie N, Roman E, Davies G, Smith PG, Beral V. Brent RL. Biological factors related to male mediated reproduc- Fetal death and congenital malformation in babies born to tive and developmental toxicity. In: Olshan AF, Mattison DR, nuclear industry employees: report from the nuclear industry eds. Male-mediated developmental toxicity. New York: Plenum family study. Lancet 356:1293–1299; 2000. Press; 1994: 209–242. Forster L, Forster P, Lutz-Bonengel S, Willkomm H, Brinkmann Brent RL. Protection of the gametes embryo/fetus from prenatal B. Natural radioactivity and human mitochondrial DNA muta- radiation exposure. Health Phys 108:242–274; 2015. tions. Proc Natl Acad Sci USA 99:13950–13954; 2002. Burkart W, Grosche B, Schoetzau A. Down syndrome clusters Fujiwara S, Suyama A, Cologne JB, Akahoshi M, Yamada M, in Germany after the Chernobyl accident. Radiat Res 147: Suzuki G, Koyama K, Takahashi N, Kasagi E, Grant EJ, Lagarde 321–328; 1997. E, Hsu WL, Furukawa K, Ohishi W, Tatsukawa Y, Neriishi K, Byrne J, Rasmussen SA, Steinhorn SC, Connelly RR, Myers MH, TakahashiI,AshizawaK,HidaA,ImaizumiM,NaganoJ, Lynch CF, Flannery J, Austin DF, Holmes FF, Holmes GE, Cullings HM, Katayama H, Ross NP, Kodama K, Shore RE. Strong LC, Mulvihill JJ. Genetic disease in offspring of long-term Prevalence of adult-onset multifactorial disease among offspring survivors of childhood and adolescent cancer. Am J Hum of atomic bomb survivors. Radiat Res 170:451–457; 2008. Genet 62:45–52; 1998. Gardner MJ, Snee MP,Hall AJ, Powell CA, Downes S, Terrell JD. Committee on Medical Aspects of Radiation in the Environment Results of case-control study of leukaemia and lymphoma (COMARE). Review of pregnancy outcomes following precon- among young people near Sellafield nuclear plant in West ceptional exposure to radiation, 8th report. Chilton, Didcot, Oxon, Cumbria. BMJ 300:423–434; 1990. UK: Health Protection Agency; 2004. Available at https:// Grant EJ, Furukawa K, Sakata R, Sugiyama H, Sadakane A, assets.publishing.service.gov.uk/government/uploads/ Takahashi I, Utada M, Shimizu Y, Ozasa K. Risk of death system/uploads/attachment_data/file/304586/ among children of atomic bomb survivors after 62 years of COMARE8thReport.pdf. Accessed 27 July 2019. follow-up: a cohort study. Lancet Oncol 16:1316–1323; 2015. Committee on Medical Aspects of Radiation in the Environment Green L, Dodds L, Miller A, Tomkins D, Jiehui L, Escobar M. (COMARE). Further consideration of the incidence of cancers Risk of congenital anomalies in children of parents occupa- around the nuclear installations at Sellafield and Dounreay, tionally exposed to low level ionising radiation. Occup Environ 17th report. Chilton, Didcot, Oxon, UK: Health Protection Agency; Med 54:629–635; 1997. 2016. Available at https://assets.publishing.service.gov.uk/ Green DM, Sklar CA, Boice JD Jr, Mulvihill JJ, Whitton JA, government/uploads/system/uploads/attachment_data/file/ Stovall M, Yasui Y. Ovarian failure and reproductive outcomes 554981/COMARE_17th_Report.pdf. Accessed 27 July 2019. after childhood cancer treatment: results from the childhood Curwen GB, Cadwell KK, Winther JF, Tawn JE, Rees GS, Olsen cancer survivor study. J Clin Oncol 27:2374–2381; 2009. JH, Rechnitzer C, Schroeder H, Guldberg P,Cordell HJ, Boice Guo Y, Cai Q, Samuels DC, YeF, Long J, Li CI, Winther JF,Tawn JD Jr. The heritability of G(2) chromosomal radiosensitivity EJ, Stovall M, Lahteenmaki P, Malia N, Levy S, Shaffer C, and its association with cancer in Danish cancer survivors Shyr Y, Shu XO, Boice JD Jr. The use of next generation and their offspring. Int J Radiat Biol 86:986–995; 2010. sequencing technology to study the effect of radiation Curwen GB, Murphy S, Tawn EJ, Winther JF, Boice JD Jr. A therapy on mitochondrial DNA mutation. Mutat Res 744: study of DNA damage recognition and repair gene polymor- 154–160; 2012. phisms in relation to cancer predisposition and G2 chromosomal Hall EJ. Radiation biology for pediatric radiologists. Pediatr Radiol radiosensitivity. Environ Mol Mutagen 52:72–76; 2011. 39(Suppl 1):S57–S64; 2009. Dauer LT, Bouville A, Toohey RE, Boice JD Jr, Beck HL, International Commission on Radiological Protection. The 2007 Eckerman KF, Hagemeyer D, Leggett RW, Mumma MT, Napier recommendations of the International Commission on Radio- B, Pryor KH, Rosenstein M, Schauer DA, Sherbini S, Stram DO, logical Protection. Oxford: Pergamon Press; ICRP Publication Thompson JL, Till JE, Yoder RC, Zeitlin C. Dosimetry and un- 103; 2007. certainty approaches for the million-worker study of radiation Institute of Medicine. Committee to study the feasibility of, and workers and veterans: overview of the recommendations in need for, epidemiologic studies of adverse reproductive out- NCRP Report No. 178. Review. Int J Rad Biol [online]. 2018. comes in the families of atomic veterans. Adverse reproductive Available at https://www.tandfonline.com/doi/abs/10.1080/ outcomes in families of atomic veterans: the feasibility of epi- 09553002.2018.1536299?journalCode=irab20. Accessed 28 demiologic studies. Washington, DC: National Academy Press; September 2019. 1995. Available at http://www.nap.edu/catalog/4992.html. Davies BG, Hussain A, Ring SM, Birch JM, Eden TO, Reeves M, Accessed 13 January 2019. Dubrova YE, Taylor GM. New germline mutations in the hy- Izumi S, Syuama A, Koyama K. Radiation-related mortality among pervariable minisatellite CEB1 in the parents of children with offspring of atomic bomb survivors: a half-century of follow-up. leukaemia. Br J Cancer 96:1265–1271; 2007. Int J Cancer 107:292–297; 2003. Dean G, Nevin NC, Mikkelsen M, Karadima G, Petersen MB, Jablon S, Hrubec Z, Boice JD Jr, Stone BJ. Cancer in populations O’Sullivan J. Investigation of a cluster of children with Down’s living near nuclear facilities. Washington, DC: US Government syndrome born to mothers who had attended a school in Printing Office; NIH Publ 90-874; 1990. Dundalk, Ireland. Occup Environ Med 57:793–804; 2000. Jablon S, Hrubec Z, Boice JD Jr. Cancer in populations living near Dickinson HO, Parker L. Leukaemia and non-Hodgkin’slym- nuclear facilities. A survey of mortality nationwide and inci- phoma in children of male Sellafield radiation workers. Int J dence in two states. JAMA 265:1403–1408; 1991. Cancer 99:436–443; 2002. Jaikrishan G, Andrews VJ, Thampi MV, Koya PK, Rajan VK, Dickinson HO, Hodgson JT, Parker L. Comparison of Health Chauhan PS. Genetic monitoring of the human population and Safety Executive and Cumbrian birth cohort studies of from high-level natural radiation areas of Kerala on the

www.health-physics.com 502 Health Physics October 2020, Volume 119, Number 4 southwest coast of India. I. Prevalence of congenital at http://www.ashg.org/2009meeting/abstracts/fulltext/ malformations in newborns. Radiat Res 152(6 Suppl): f21783.htm. Accessed 28 September 2019. S149–S153; 1999. Nakamura N. Genetic effects of radiation in atomic-bomb survi- Johnson KJ, Alexander BH, Doody MM, Sigurdson AJ, Linet MS, vors and their children: past, present and future. J Radiat Res Spector LG, Hoffbeck RW, Simon SL, Weinstock RM, Ross JA. 47(Suppl. B):B67–B73; 2006. Childhood cancer in the offspring born in 1921–1984 to US ra- National Academies/National Research Council. Health risks diologic technologists. Br J Cancer 99:545–550; 2008. from exposure to low levels of ionizing radiation: BEIR VII Kesavan PC. Indian research on high levels of natural radiation: Phase 2. Washington, DC: National Academies Press; 2006. pertinent observations for further studies. In: Wei L, Sugahara National Council on Radiation Protection and Measurements. T, Tao Z, eds. High levels of natural radiation, 1996: radiation Preconception and prenatal radiation exposure: health ef- dose and health effects. Amsterdam: Elsevier; 1997: 112–117. fects and protective guidance. Bethesda, MD: NCRP; Re- Kinlen LJ. Can paternal preconceptional radiation account for the port 174; 2013. increase of leukaemia and non-Hodgkin’s lymphoma in Seascale? National Council on Radiation Protection and Measurements. Health BMJ 306:1718–1721; 1993. effects of low doses of radiation: perspectives on integrating Kinlen LJ. Epidemiological evidence for an infective basis in radiation biology and epidemiology. Bethesda, MD: NCRP; childhood leukaemia. Br J Cancer 71:1–5; 1995. Commentary 24; 2015. Kinlen L. Population mixing and childhood leukaemia. Eur J National Council on Radiation Protection and Measurements. Epidemiol 2015 30:1331; 2015. Deriving organ doses and their uncertainty for epidemiologic Kinlen LJ, Clarke K, Balkwill A. Paternal preconceptional radia- studies (with a focus on the one million US workers and vet- tion exposure in the nuclear industry and leukaemia and non- erans study of low-dose radiation health effects). Bethesda, Hodgkin’s lymphoma in young people in Scotland. BMJ MD: NCRP; Report 178; 2018a. 306:1153–1158; 1993. National Council on Radiation Protection and Measurements. Kochupillai N, Verma IC, Grewal MS, Ramalingaswami V.Down’s Management of exposure to ionizing radiation: radiation pro- syndrome and related abnormalities in an area of high back- tection guidance for the United States (2018). Bethesda, MD: ground radiation in coastal Kerala. Nature 262:60–61; 1976. NCRP; Report 180; 2018b. Kodeira M, Izumi S, Takahashi N, Nakamura N. No evidence of Neel JV.Genetic studies at the Atomic Bomb Casualty Commission- radiation effect on mutation rates at hypervariable minisatellite Radiation Effects Research Foundation: 1946–1997. Proc Natl loci in the germ cells of atomic bomb survivors. Radiat Res Acad Sci USA 95:5432–5436; 1998. 162:350–356; 2004. Neel JV. Changing perspectives on the genetic doubling dose of Koya PK, Chougaonkar MP, Predeep P, Jojo PJ, Cheriyan VD, ionizing radiation for humans, mice, and Drosophila. Teratol Mayya YS, Seshadri M. Effect of low and chronic radiation ex- 59:216–221; 1999a. posure: a case-control study of mental retardation and cleft lip/ Neel JV. Two recent radiation-related genetic false alarms: leuke- palate in the monazite-bearing coastal areas of southern Kerala. mia in West Cumbria, England, and minisatellite mutations in Radiat Res 177:109–116; 2012. Belarus. Teratol 59:302–306; 1999b. Leuraud K, Richardson DB, Cardis E, Daniels RD, Gillies M, Neel JV, Schull WJ. The children of atomic bomb survivors. O’Hagan JA, Hamra GB, Haylock R, Laurier D, Moissonnier Washington, DC: National Academy Press; 1991. M, Schubauer-Berigan MK, Thierry-Chef I, Kesminiene A. Otake M, Schull WJ, Neel JV: Congenital malformations, still- Ionising radiation and risk of death from leukaemia and lym- births, and early mortality among the children of atomic bomb phoma in radiation-monitored workers (INWORKS): an inter- survivors: a reanalysis. Radiat Res 122:1–11; 1990. national cohort study. Lancet Haematol 2:e276–281; 2015. Parker L, Pearce MS, Dickinson HO, Aitkin M, Craft AW. Stillbirths Little J. The Chernobyl accident, congenital anomalies and among offspring of male radiation workers at Sellafield nu- other reproductive outcomes. Paediatr Perinat Epidemiol clear reprocessing plant. Lancet 354(9188):1407–1414; 1999. 7:121–151;1993. Premi S, Srivastava J, Chandy SP,Ali S. Unique signatures of nat- Little MP. A comparison of the risk of stillbirth associated with ural background radiation on human Y chromosomes from paternal pre-conception irradiation in the Sellafield work- Kerala, India. PLOS ONE 4:e4541; 2009. force with that of stillbirth and untoward pregnancy outcome Roman E, Doyle P, Ansell P, Bull D, Beral V. Health of children among Japanese atomic bomb survivors. J Radiol Protect 19: born to medical radiographers. Occup Environ Med 53: 361–373; 1999. 73–77; 1996. Little MP,Wakeford R, Charles MW,Andersson M. A comparison Satoh C, Takahashi N, Asakawa J, Kodaira M, Kuick R, Hanash of the risks of leukaemia and non-Hodgkin’s lymphoma in the SM, Neel JV.Genetic analysis of children of atomic bomb sur- first generation offspring of the Danish Thorotrast patients with vivors. Environ Health Perspect 104(Suppl 3):511–519; 1996. those observed in other studies of parental pre-conception irradi- Schull WJ. The children of atomic bomb survivors: a synopsis. J ation. J Radiol Protect 16:25–36; 1996. Radiol Protect 23:369–384; 2003. Little MP,Goodhead DT, Bridges BA, Bouffler SD. Evidence rel- Schull WJ, Otake M, Neel JV. Genetic effects of the atomic evant to untargeted and transgenerational effects in the off- bombs: a reappraisal. Science 213:1220–1227; 1981. spring of irradiated parents. Mutat Res 753:50–67; 2013. Sever LE, Gilbert ES, Tucker K, Greaves JA, Greaves C, Buchanan Mumma MT, Cohen SS, Sirko JL, Ellis ED, Boice JD Jr. J. Epidemiologic evaluation of childhood leukemia and paternal Obtaining vital status and cause of death on a million persons. exposure to ionizing radiation. Springfield, VA: National Tech- Int J Radiat Biol [online]. 2018. Available at https://www.ncbi. nical Information Service; U50/CCU012545-01; 1997. Avail- nlm.nih.gov/pubmed/30412007. Accessed 9 November 2018. able at https://www.cdc.gov/niosh/oerp/pdfs/2001-133g7.pdf. Mulvihill JJ, Boice JD, Winther JF, Malila N, Olsen JH, Accessed 27 July 2019. Lähteenmäki PM, Madanat-Harjuoja LM, Sankila R, Stovall Sever LE, Hessol NA, Gilbert ES, McIntyre JM. The prevalence M, Tawn EJ, the GCCT Study Group. Germ cell mutations: at birth of congenital malformations in communities near the genetic disease in offspring of cancer survivors. American Hanford site. Am J Epidemiol 127:243–254; 1988a. Society of Human Genetics Meeting, Honolulu, HI, 20–24 Sever LE, Gilbert ES, Hessol NA, McIntyre JM. A case-control October 2009, Platform Abstract 111 [online]. 2009. Available study of congenital malformations and occupational exposure

www.health-physics.com Pregnancy outcomes and genetic disease following exposure c J.D. BOICE,JR. 503 to low-level ionizing radiation. Am J Epidemiol 127:226–242; Tularosa Basin Downwinders Consortium. Unknowing, unwill- 1988b. ing, and uncompensated: the effects of the Trinity test on Signorello LB, Cohen SS, Bossetti C, Stovall M, Kasper CE, New Mexicans and the potential benefits of Radiation Expo- Weathers RE, Whitton JA, Green DM, Donaldson SS, Mertens sure Compensation Act (RECA) amendments [online]. 2017. AC, Robison LL, Boice JD Jr. Female survivors of childhood Available at https://www.kob.com/kobtvimages/repository/ cancer: preterm birth and low birth weight among their chil- cs/files/TBDC%20HIA%20final%20report%202-8-2017. dren. J Natl Cancer Inst 98:1453–1461; 2006. pdf. Accessed 15 March 2019. Signorello LB, Mulvihill JJ, Green DM, Munro HM, Stovall M, United Nations Scientific Committee on the Effects of Atomic Ra- Weathers RE, Mertens AC, Whitton JA, Robison LL, Boice diation. Sources and effects of ionizing radiation. New York: JD Jr. Stillbirth and neonatal death in relation to radiation ex- United Nations; 1993. posure before conception: a retrospective cohort study. Lancet United Nations Scientific Committee on the Effects of Atomic Ra- 376:624–630; 2010. diation. Sources and effects of ionizing radiation. New York: Signorello LB, Mulvihill JJ, Green DM, Munro HM, Stovall M, United Nations; 1994. Tawn EJ, Weathers RE, Mertens AC, Whitton JA, Robison United Nations Scientific Committee on the Effects of Atomic Ra- LL, Boice JD Jr. Congenital anomalies in the children of can- diation. Report to the General Assembly, with scientific an- cer survivors: a report from the childhood cancer survivor nexes. Hereditary effects of radiation, UNSCEAR 2001. New study. J Clin Oncol 30:239–245; 2012. York: United Nations; 2001. Simon SL, Bouville A, Beck H, Melo D. Estimated radiation United Nations Scientific Committee on the Effects of Atomic Ra- doses received from the 1945 Trinity nuclear test. Health Phys diation. Effects of ionizing radiation, UNSCEAR 2006. Scien- 119:428–477; 2020. tific annex A: epidemiological studies of radiation and cancer; Sorahan T, Haylock RG, Muirhead CR, Bunch KJ, Kinlen LJ, Vol I. New York: United Nations; 2008. Little MP, Draper GJ, Kendall GM, Lancashire RJ, English Wakeford R. Study of childhood cancer and paternal preconceptional MA. Cancer in the offspring of radiation workers: an investi- irradiation at USA nuclear facilities. J Radiol Protect 20: gation of employment timing and a reanalysis using updated 331–332; 2000. dose information. Br J Cancer 89:1215–1220; 2003. Wei L, Zha Y, Tao Z, He W,Chen D, Yuan Y. Epidemiological in- Stebbings JH Jr, Voelz GL. Morbidity and mortality in Los vestigation of radiological effects in high background radiation Alamos County, New Mexico. I. Methodological issues and areas of Yangjiang, China. J Radiat Res 31:119–136; 1990. preliminary results. Environ Res 25:86–105; 1981. Wilding CS, Curwen GB, Tawn EJ, Sheng X, Winther JF, Stovall M, Donaldson SS, Weathers RE, Robison LL, Mertens Chakraborty R, Boice JD Jr. Influence of polymorphisms at AC, Winther JF, Olsen JH, Boice JD Jr. Genetic effects of ra- loci encoding DNA repair proteins on G2 chromosomal radio- diotherapy for childhood cancer: gonadal dose reconstruction. sensitivity. Environ Mol Mutagen 48:48–57; 2007. Int J Radiat Oncol Biol Phys 60:542–552; 2004. Winther JF, Olsen JH. Does cancer treatment in childhood in- Tawn EJ. Leukaemia and Sellafield: is there a heritable link? J duce transgenerational genetic damage? J Clin Oncol 30: Med Genet 32:251–256; 1995. 225–226; 2012. Tawn EJ, Curwen GB, Rees GS, Jonas P. Germline minisatellite Winther JF, Boice JD Jr, Thomsen BL, Schull WJ, Stovall M, Olsen mutations in workers occupationally exposed to radiation at JH. Sex ratio among offspring of childhood cancer survivors the Sellafield nuclear facility. J Radiat Protect 35:21–36; 2015. treated with radiotherapy. Br J Cancer 88:382–387; 2003. Tawn EJ, Rees GS, Leith C, Winther JF, Curwen GB, Stovall M, Winther JF, Boice JD Jr, Mulvihill JJ, Stovall M, Frederiksen K, Olsen JH, Rechnitzer C, Schroeder H, Guldberg P, Boice JD Tawn EJ, Olsen JH. Chromosomal abnormalities among offspring Jr. Germline minisatellite mutations in survivors of childhood of childhood cancer survivors in Denmark: population-based and young adult cancer treated with radiation. Int J Radiat Biol study. Am J Hum Genet 74:1282–1285; 2004. 87:330–340; 2011. Winther JF, Boice JD Jr, Svendsen AL, Frederiksen K, Stovall M, Tawn EJ, Whitehouse CA, Winther JF, Curwen GB, Stovall M, Olsen JH. Spontaneous abortion after radiotherapy for child- Olsen JH, Guldberg P, Rechnitzer C, Schrøder H, Boice JD hood cancer: a population-based cohort study. J Clin Oncol Jr. Chromosome analysis in childhood cancer survivors and 26:4340–4346; 2008. their offspring—no evidence for radiotherapy-induced persis- Winther JF, Olsen JH, Wu H, Shyr Y, Mulvihill JJ, Stovall M, Nielsen tent genomic instability. Mutat Res 583:198–206; 2005. A, Schmiegelow M, Boice JD Jr. Genetic disease in the children Till JE, Beck HL, Aanenson JW, Grogan HA, Mohler HJ, Mohler of Danish survivors of childhood and adolescent cancer. J Clin SS, Voillequé PG. Military participants at U.S. atmospheric nu- Oncol 30:27–33; 2012. clear weapons testing—methodology for estimating dose and World Health Organization. Health effects of the Chernobyl acci- uncertainty. Radiat Res 181:471–484; 2014. dent: n overview [online]. 2006. Available at https://www.who. Till JE, Beck HL, Aanenson JW,Grogan HA, Mohler HJ, Mohler SS, int/ionizing_radiation/chernobyl/backgrounder/en/. Accessed Voillequé PG. Dosimetry associated with veterans who partic- 1 August 2019. ipated in nuclear weapons testing. Int J Radiat Biol [online]. 2018. Available at https://www.tandfonline.com/doi/abs/10. 1080/09553002.2018.1551639?journalCode=irab20.Accessed 28 September 2019. ■■

www.health-physics.com Paper

Accounting for Unfissioned Plutonium from the Trinity Atomic Bomb Test

Harold L. Beck,1 Steven L. Simon,2 André Bouville,3 and Anna Romanyukha4

INTRODUCTION Abstract—The Trinity test device contained about 6 kg of pluto- nium as its fission source, resulting in a fission yield of 21 kT. THE TRINITY test device was reported to have contained 239 However, only about 15% of the Pu actually underwent fission. about 6 kg (US DOE 2002) of plutonium as its major fission The remaining unfissioned plutonium eventually was vaporized in the fireball and after cooling, was deposited downwind from the source. However, based on the reported fission yield of 21 kT test site along with the various fission and activation products pro- (US DOE 2000), and the fact that about 1/3 of the yield was duced in the explosion. Using data from radiochemical analyses of from fission of 235U in the thick natural uranium tamper sur- soil samples collected postshot (most many years later), supple- rounding the plutonium core, only about 15% of the 239Pu mented by model estimates of plutonium deposition density esti- 5 mated from reported exposure rates at 12 h postshot, we have actually underwent fission. The remaining unfissioned plu- estimated the total activity and geographical distribution of the tonium was instantaneously vaporized in the fireball and af- deposition density of this unfissioned plutonium in New Mexico. ter cooling, was deposited downwind from the test site along A majority (about 80%) of the unfissioned plutonium was with the various fission and activation products produced in deposited within the state of New Mexico, most in a relatively small area about 30–100 km downwind (the Chupadera Mesa the explosion. In this paper, we estimate the deposition den- area). For most of the state, the deposition density was a small sity of this unfissioned plutonium at various distances down- fraction of the subsequent deposition density of 239+240Pu from wind from the test site as well as the total cumulative activity – Nevada Test Site tests (1951 1958) and later from global fallout deposited within the state of New Mexico. The amount of from the large US and Russian thermonuclear tests (1952–1962). The fraction of the total unfissioned 239Pu that was deposited in plutonium contamination of the New Mexico environs has New Mexico from Trinity was greater than the fraction of fission understandably become an issue of concern to some residents products deposited. Due to plutonium being highly refractory, a of the state, particularly those residing at locations near the 239 greater fraction of the Pu was incorporated into large particles White Sands test site (TBDC 2017). that fell out closer to the test site as opposed to more volatile fission products (such as 137Cs and 131I) that tend to deposit on the surface of The total unfissioned plutonium can be estimated from smaller particles that travel farther before depositing. The plutonium the reported amount of plutonium in the device, the estimated deposited as a result of the Trinity test was unlikely to have resulted ratio of 90Sr to 137Cs in the deposited fallout, and the reported in significant health risks to the downwind population. – explosive yield. According to Glasstone and Dolan (1977), Health Phys. 119(4):504 516; 2020 23 239 235 239 1.45 10 fissions of either Pu or U results in a Key words: Pu; fallout; nuclear weapons; plutonium yield of 1 kT. If all the reported yield of 21 kT were from plutonium fission, 1.197 kg of 239Pu would have fissioned − ([21 kT 1.45 1023 fissions kT 1]/[2.52 1024 plutonium atoms kg−1]), leaving 6 − 1.2 = 4.8 kg of Pu unfissioned. 1US DOE Environmental Measurements Laboratory (retired); 2Divison 3 However, because ~1/3 of the fissions were actually from of Cancer Epidemiology, National Cancer Institute; National Cancer Institute 235 (retired); 4Centre for Medical Radiation Physics, School of Physics, University fission of U, the amount of unfissioned plutonium was of Wollongong, Wollongong, NSW, Australia. actually ~5.2 kg. The authors declare no conflicts of interest. Although the unfissioned 239Pu from nuclear tests has For correspondence contact Steven L. Simon, Division of Cancer Epidemiology and Genetics, National Cancer Institute, National Institutes generally been of less concern to knowledgeable experts of Health, 9609 Medical Center Drive, Bethesda, MD 20892–9778, or in regard to the risk of health effects (compared to the risks email at [email protected]. (Manuscript accepted 25 June 2019) from deposited fission products), there is a perception by 0017-9078/20/0 Written work prepared by employees of the Federal Government as 5 90 137 part of their official duties is, under the U.S. Copyright Act, a "work of Because the ratio of the fission yields for Sr to Cs differs considerably for 239Pu and 235U fission (3.21 vs. 1.14), the fraction of the total fissions from the United States Government" for which copyright protection under Title 235 90 137 17 of the United States Code is not available. As such, copyright does not U can be estimated from the ratio of Sr and Cs deposition densities of extend to the contributions of employees of the Federal Government. Trinity fallout (2.04) estimated by Hicks (1981) (Beck 2001a). Although the actual fraction of the Trinity yield from 235U has not been officially re- DOI: 10.1097/HP.0000000000001146 ported, our estimate is in good agreement with unofficial estimates.

504 www.health-physics.com Accounting for unfissioned plutonium c H.L. BECK ET AL. 505 the general public that plutonium is especially hazardous. It penetrated much deeper. Based on limited data, Hansen is, therefore, worthwhile to attempt to document exactly and Rodgers (1985) found that plutonium inventory was how much and where this plutonium was deposited and to uniform in the top 2.5 cm of soil and decreased exponen- discuss the possible impact on the residents of New tially with further depth. Based on the cores collected at Mexico—past, present, and future. 17 sites by Hansen and Rodgers (1985), about 1/2 the pluto- To estimate the amount of plutonium deposited in var- nium collected down to a depth of 15 cm was in the top ious areas of New Mexico, we relied on data from analyses of 2.5 cm in the mid-1970s when most of the soil data were postshot soil samples for 239+240Pu activity supplemented by collected (Fig. 2). Even the samples collected closer to the model-based estimates of 239+240Pu activity in soil derived time of the test (Olafson et al. 1957) are suspect because from postshot exposure-rate monitoring data. The use of a the sample depth was sometimes only 2.5 cm or even less, model to supplement the soil analyses was necessary and rain and natural processes would have caused some of because the available soil sample data were limited to areas the plutonium to move below the sampling depth. In their relatively close to the test site and to the fallout pattern axis, 1972 study, Hakonson and Johnson (1973) observed that as well as, in general, being very imprecise. the plutonium concentration penetrated to a soil depth of Using the combination of soil measurements and esti- at least 30 cm. Of 12 profile samples taken by the US Envi- mates from models, we estimated plutonium deposition den- ronmental Protection Agency (EPA) (Douglas 1978), 6 had sity at ~1,000 sites covering the entire state of New Mexico. detectable levels of plutonium down to at least 10–15 cm. We then interpolated those data to obtain estimates on a Nyhan et al. (1976) noted that their soil samples reflected 2km 2 km grid allowing for a fairly precise numerical “considerable downward migration of plutonium into the soil” integration of activity with increasing distance from ground with time. McArthur and Miller (1989) collected soil samples zero (GZ). The specifics of the methodology and results are at a number of sites in New Mexico in 1982 in conjunction detailed in the following sections of this paper. with the US Department of Energy (DOE) Offsite Radiation Exposure Review Program (ORERP). Those samples were METHODS obtained under a more rigorous sampling strategy, and soil preparation and chemical analysis procedures were conducted Estimation of deposition density from soil sample data with more strict quality assurance than many of the earlier A number of investigators and institutions collected 239+240 soil samples downwind from GZ for plutonium measure- Trinity sampling programs. Although the Pu at the ment. Table 1 lists the various sampling programs and the sites sampled was mostly from Nevada Test Site (NTS) approximate number of unique sites sampled. Unfortunately, and global fallout rather than from Trinity, on average some of these samples were obtained in a manner that made 60% was found below 10 cm. However, while these – their use in estimating deposition density unreliable, leading samples were taken 5 10 y later than the Trinity samples, us to determine they were not suitable for use in estimating the NTS and global fallout was deposited from 6 to 17 y the total plutonium deposited. Most of the other samples re- later than the Trinity fallout. quired corrections to account for insufficient depth of sam- The amount of soil collected at each site varied as well. pling and/or unspecified soil density. Fig. 1 shows the For some of the sites sampled, only three 12.8-cm-diameter cores were obtained (Hansen and Rodgers 1985), although location of all the sampled sites. 2 Most of the soil samples were taken only to a depth of for the samples collected by Douglas (1978), ~1,000 cm 5 cm and some only as deep as 2.5 cm. For samples taken (10 cores) were sampled at most sites. Thus, even for the many years after the test, the plutonium is known to have Douglas samples, an individual sample may not be a true representation of the mean deposition density over the sur- Table 1. Summary of soil samplings. rounding area, particularly considering that the samples in Date of # Unique # Profiles (total those surveys were usually not obtained in ideal terrain ≥ b Reference sampling sites depth 15 cm) where the deposited activity might have been expected to Olafson et al. (1957) 1948 17 0 have deposited uniformly over the general area and would Hakonson et al. (1973) 1972 8 8 have been expected to have remained in place from the time Douglas (1978) 1973, 1974 88 11 of deposition to the time of sampling. It is well known (US Nyhan et al. (1976) 1976 4 2 EPA 1980) that in sparsely vegetated environs windblown Hansen and Rodgers (1985) 1977 54 17 fallout in surface soil tends to preferentially collect at the a McArthur and Miller (1989) 1982 14 14 base of desert brush, and sampling in open areas may thus aMost of these samples were remote from the areas impacted by Trinity fallout result in an underestimate of the true deposition density. and served as control samples. Some analyses were reported only as activity per mass bMost of the profile samples were taken in 5 cm increments; of the 38 profile samples, excluding the McArthur and Miller samples, 20 reached to only of soil rather than per unit area. In that case, the specific 15 cm while only 13 extended to ≥25 cm. bulk density of the soil is required to estimate deposition www.health-physics.com 506 Health Physics October 2020, Volume 119, Number 4

Fig. 1. Soil sampling locations. density (Bq m−2). To correct the samples where only activity the reported sample depth to attempt to account for the plu- per mass was reported, we assumed an average bulk density tonium likely not collected (Table 2). These correction factors of 1.2 g cm−3 based on the mean density reported by Douglas are based on the observed depth distributions at sites where (1978) for the samples they collected. profiles were obtained, the McArthur and Miller (1989) soil To address the problems with the limited depth sam- data, and our best judgement. Because the number of complete pled at most of the sites, we applied corrections based on profiles were relatively few, as shown in Table 1, and the depth

Fig. 2. Fraction of plutonium activity in top 5 cm, based on 17 profile samples 0–15 cm taken in 1977. LADB refers to sample number in the Los Alamos database, reproduced in Appendix A of Hansen and Rodgers (1985).

www.health-physics.com Accounting for unfissioned plutonium c H.L. BECK ET AL. 507 profiles varied from profile to profile, these correction factors 1015 Bq kg−1, respectively), and the ratio of 137Cs activity in may not be strictly accurate for any specific site, but on aver- deposited fallout normalized to E12 for Trinity (29 Bq m−2; age, they should provide more reasonable estimates of the total Hicks 1981). plutonium deposited over the entire area than the estimates However, corrections are required to account for frac- based on the uncorrected data. tionation, i.e., the phenomenon whereby refractory radionu- As shown in Fig. 1, the available soil sample data covers clides such as plutonium tend to be incorporated into larger only a limited geographical area relatively close to the test particles that fall out more rapidly, and thus closer, to the test site and to the fallout pattern axis (center line of the trajec- site than volatile nuclides such as 137Cs or 131I. The volatile tory). Some of these data, in our judgement, were anomalous nuclides condense later and attach preferentially to smaller or not suitable for estimating deposition density due to insuf- particles that travel farther distances before depositing ficient sample depth. The remaining soil data, notwithstand- (Hicks 1981; Freiling 1961). Thus, the ratio of refractory ing the uncertainty due to lack of representativeness and nuclides to E12 is greater than the ratio for unfractionated particularly to measurement imprecision, are far too few to fallout at locations close-in to the detonation site and along interpolate and extrapolate for estimating the cumulative de- the trace6 axis than for fallout characteristic of deposition at position over the entire state. Thus, it was necessary to use an greater distances and less than the unfractionated value at alternate methodology to extend the fallout pattern out to far- distances far removed from the detonation site or far off ther distances as well as to fill in the areas with a limited the axis. number of sites with good measurement data. Using the calculations of Hicks (1981), we can calcu- late the activity ratio R/V (of a refractory radionuclide R 239+240 such as plutonium to a volatile radionuclide V such as Estimation of Pu from the reported exposure − rate pattern 137Cs) for E12 = 1.0 mR h 1 and for different values of The model used to estimate plutonium deposition in R/V.7 Using the joint US-Russian deposition density model, this study is based on a joint US-Russian semiempirical we can estimate R/V as a function of fallout time-of-arrival model for estimating the deposition density of fission prod- (TOA) and distance from the fallout pattern axis to correct ucts at sites relatively close to a test site where fractionation the value of Pu/E12 and thereby, account for fractionation affects the relative deposition of refractory vs. volatile radio- at each specific location. Table 3 gives our estimates of nuclides as a function of distance from GZ. The details of Pu/E12 vs. R/V8 and illustrates that the correction re- this model are summarized in the Appendix. quired can be quite significant, particularly for sites close In brief, we estimated the plutonium deposition density to GZ, i.e., with low TOA values and at locations close to at a large number of sites throughout New Mexico from the the trace axis. reported exposure rates at 12 h postshot (E12) from Quinn The required estimates of E12 and TOA were obtained (1987) using conversion ratios of 239+240Pu activity to E12 by interpolation of the published fallout pattern (Fig. 3) that (Bq m−2 mR h−1; abbreviated as Pu/E12), corrected for was constructed from an analysis of all postshot monitoring fractionation as described in the Appendix. data supplemented by meteorological data (Quinn 1987). Pu/E12 for unfractionated Trinity fallout was estimated Unfortunately, because the acquisition of monitoring data to be 2.7 Bq m−2 (mR h−1)−1 from the fission yields for was limited to locations with roads, and the precision and 137Cs (6.58% for 239Pu and 3.22% for 235U; England accuracy varied depending on the accuracy of the particular and Ryder 1994), the number of fissions (21 kT instruments used, the required corrections for radioactive − 1.45 1023 fissions kT 1), the estimated amount of decay and consequently, the published fallout pattern itself unfissioned plutonium (5.2 kg), the estimated fraction is somewhat uncertain. Furthermore, the contours provided of the yield due to 235U fission (32%), the specific are fairly broad, i.e., the spatial resolution is poor, particu- activities of 239Pu and 137Cs (2.27 1012 and 3.21 larly close-in to GZ where the E12 changes rapidly with dis- tance. This made precise interpolation of both E12 and Table 2. Corrections applied for insufficient sample depth. TOA difficult. Sample depth (cm) Correction factor

<2.5 Not used 6 7We use the term trace to refer to the pattern of fallout downwind from GZ. 2.5 3.0 Conventional units are used for E12 to be consistent with historical liter-

8ature on deposition of fallout from nuclear testing. 52.0The Pu/E12 ratios in Table 3 include a small correction to account for 10 1.6 240Pu in the original fuel. According to Douglas (1978), ~2.3% of the Trinity 240 240 15 1.4 plutonium mass deposited was Pu. This amount of Pu is consistent with the fraction of 240Pu used in early nuclear tests (Hicks 1990). Thus, based on 22.5 1.1 therelativehalf-livesof240Pu (6,600 y) to 239Pu (24,000 y), ~8% of the depos- 240 241 30 1.0 ited activity was from Pu. The fraction of Pu in early devices was only about 0.05% (Hicks 1990).

www.health-physics.com 508 Health Physics October 2020, Volume 119, Number 4 Table 3. Estimated Pu/E12 vs. R/V. which can be compared with the limited coverage of the soil R/V Pu/E12 (Bq m−2 [mR h−1]−1) sites shown in Fig. 1. Although the model estimates are uncertain, primarily 0.5 1.7 as a consequence of the uncertainty in the estimates of 1.0 2.7 1.5 3.3 E12 and TOA but also due to the uncertainty in the esti- 2.0 3.7 mated R/V and the exact amount of unfissioned pluto- 3.0 4.3 nium, a comparison of model calculated vs. measured 4.0 5.1 deposition density (Fig. 5) exhibits a fairly strong correla- tion (r2 = 0.7), particularly if we restrict the comparison to calculated-to-measured (C/M) ratios within a factor of 5-fold. Nevertheless, we estimated E12 and TOA at all the We have assumed that ratios outside a factor of 5-fold sample sites as well as at the centroids of all the 1945 voting indicate either that the soil data is anomalous (most likely, precincts for which Simon et al. (2020) calculated organ for the reasons discussed earlier regarding the various sam- doses from fallout radionuclides and activation products. pling and analysis uncertainties) or the model calculations We then applied the R/V corrected Pu/E12 conversion are in error (considered less likely, except very close to the based on the estimated TOA and E12 and the calculated dis- trace axis and to GZ where large corrections for R/V were tance from the trace axis to estimate a plutonium deposition required). The average ratio of C/M was found to be 1.18 density. The locations of these sites are shown on Fig. 4 (standard error [SE] = 0.10) suggesting that calculated ratios

Fig. 3. Fallout contours (E12, TOA) (Quinn 1987).

www.health-physics.com Accounting for unfissioned plutonium c H.L. BECK ET AL. 509

Fig. 4. Locations where deposition density was calculated from E12, TOA. for data deemed reliable were generally within about 25% of of New Mexico as well as to fill in the gaps at close-in the measured ratios. Furthermore, the C/M ratio varies little distances. with R/V, suggesting the estimates of Pu/E12 vs. R/V are reasonably accurate. Considering, as discussed earlier, that Estimation of plutonium deposition density using a the soil sampling carried out many years after the event combination of soil data and model estimates would tend to underestimate the deposition, a C/M ratio Because of the limited number of soil samples and the greater than unity is not unexpected. spatial gaps in geographic coverage (Fig. 1), it was not pos- However, as can be seen from Fig. 5, there is a slight sible to obtain a credible estimate of total deposition from trend toward increasing C/M as the measured plutonium de- the soil data alone. However, the reasonably strong correla- creases. For example, the ratio of C/M for measured plutonium tion between the soil data and the model calculations allowed <1,000 Bq m−2 is about 1.4, while C/M for >1,000 Bq m−2 is us to calculate the deposition density at sites without mea- about 0.9. Thus, either the model overestimates the pluto- surement data and estimate the total deposition in New nium at lower activities or underestimates the plutonium at Mexico two different ways. First, we used a combination of high activities, the measured plutonium tends to be progres- thesoildatadeemedcredible(C/M=0.2–5) supplemented sively more underestimated at lower activities, or more by model calculations at additional sites. This method likely, some combination of the above. However, consider- minimizes any potential error due to having to estimate ing the multiple sources of uncertainty discussed in the pre- R/V because the soil data are used for most of the sites ceding paragraphs, the correlation between the measured where the R/V values are highest. Second, we used only the and calculated plutonium is quite satisfactory, and we be- model-calculated deposition density under the assumption lieve it demonstrates that the model estimates can be reli- that this might provide better overall precision without ably used to extend the deposition density estimates to all affecting the estimated geographical precision. To allow a www.health-physics.com 510 Health Physics October 2020, Volume 119, Number 4

Fig. 5. Calculated vs. measured deposition density (Bq m−2): (a) only sites with calculated-to-measured ratios between 0.2 and 5; (b) all soil- sampling sites. precise numerical integration vs. distance, we used kriging in within a distance of <400 km downwind. Given that a line both cases to estimate (interpolate) the deposition density on from Trinity along the trace intersects the eastern border a2km 2kmgrid. of New Mexico at about 350 km, most (if not all) of the plutonium deposited within 400 km was deposited within RESULTS the state of New Mexico. Our estimate of the total unfissioned plutonium depos- Fig. 6 is a plot of the interpolated deposition density es- ited in New Mexico is considered somewhat uncertain, 239+240 timates of Pu activity throughout New Mexico. Fig. 6 while the fraction deposited within a range of distances is based on using only calculated (model) data because we (the pattern of fallout) is believed to be more precise. found the interpolation plot using a combination of model Because R/V is only significantly >1 over a small area and soil data to be almost identical. However, as expected, close to GZ and within a few tens of kilometers from the the data values using model data alone appeared to be axis of the fallout trace, any uncertainty in R/V is slightly more precise as indicated by less scatter. This is expected to have had only a relatively small effect on the not unexpected given that individual soil estimates are not estimated distribution of plutonium vs. distance and the very representative of the local area compared to the estimated fraction of unfissioned plutonium deposited in New Mexico. E12, as any site can have been unknowingly disturbed or eroded. This conclusion is supported by the overall relatively good Table 4 shows the calculated total deposition (total plu- correlation shown in Fig. 5 between calculated and measured tonium activity) as a function of distance from GZ for both methods along with the fraction of unfissioned plutonium 9 9 The measured and calculated deposition densities are in Bq m2 of (mass or activity). As indicated by both methods, about 239+240Pu. The mass of plutonium deposited per unit area is directly pro- 80% of the unfissioned plutonium was likely deposited portional to the activity of 239+240Pu deposited. www.health-physics.com Accounting for unfissioned plutonium c H.L. BECK ET AL. 511

Fig. 6. Calculated deposition density of 239+240Pu. plutonium deposition density and the fact that the C/M ratio In addition, 239Np was produced in the blast by neutron did not vary significantly with R/V. activation of the 238U tamper surrounding the plutonium The reasons for the significant uncertainty in total de- core. Based on the 239Np activity per mR h−1 estimated position are several. First, the exact amount of plutonium for Trinity by Hicks (1981), as much as an additional used in the Trinity device is not known precisely. US 0.8kgof239Pu might have been produced from the beta DOE (2002) reported the fuel to have been “about” 6kg decay of this 239Np. This extra source of plutonium is not of plutonium, which is the value used in our model. In included in our primary calculations because it does not comparison, the Defense Special Weapons Agency (DSWA represent the unfissioned remainder of the device’score. 1997) estimated 6–7 kg was used in the first Soviet test, Including this source of plutonium would have increased reputed to have been “very similar” to Trinity in its our estimate of the deposition density at each location and construction. In that test, only about 15% of the 239Pu was the total deposition in New Mexico, and would have reported to have fissioned. Other reported estimates range increased the observed C/M ratio, by about 10%. from ~5.5 kg to 6.5 kg. Thus, because only a small amount The amount of plutonium deposited very close to the of plutonium actually fissioned, if there was 10% more or test site is also very uncertain. Based on sparse available less plutonium in the Trinity device than the nominal 6 kg data, we crudely estimated ~5% of the total unfissioned plu- that we used for our calculations, the amount of tonium to have been deposited within 10 km from GZ. The unfissioned plutonium would have been ~10% higher or available soil activity data at distances <10 km, i.e., on the lower. The yield of the device is also uncertain to ±2 kT Trinity test site, ranged widely, from a few very high values (Young and Kerr 2005), which could have resulted in to no alpha activity. Unfortunately, we could not use our about 7% more or less unfissioned plutonium having model to estimate the plutonium on-site because the E12 been produced. measured near GZ was mostly from activation of the soil

www.health-physics.com 512 Health Physics October 2020, Volume 119, Number 4 Table 4. Fraction of unfissioned plutonium deposited vs. distance and Measurements (NCRP) and the International Commission a. Using soil data supplemented by model estimates of deposition density. on Radiological Protection (ICRP). Distance (km) Unfissioned 239Pu deposited (%) Cumulative (%) The activity levels in New Mexico soils that we have estimated can be compared to these US EPA recommended – 0 10 5.0 5.0 action levels. The US EPA action level for 239Pu is 10–30 3.9 8.9 − 7,400 Bq m 2 in the top 1 cm of soil, which is considered 30–100 53 62 the depth from which soil particles containing plutonium 100–200 19 81 200–300 1.4 82 could be resuspended into the air. Even with the uncertainties 300–500 0.4 83 discussed earlier, for areas off the site of the Trinity >500 17 100 detonation, the plutonium soil activity was well below these action levels, even for the Chupadera Mesa region and even b. Using model estimates of plutonium deposition density only. shortly after deposition. Note that our estimated plutonium Distance (km) Unfissioned 239Pu deposited (%) Cumulative (%) deposition densities include the small contribution from 240 0–10 5 5.0 Pu. The subsequent penetration of activity to deeper levels 10–30 3.6 8.6 in the soil was such that the activity in the top 1 cm is now 30–100 52 61 far below these action levels at all locations. In fact, based on 100–200 17 78 our estimates of the depth distribution in 1976 (Table 2), the 200–300 1.2 80 current levels in the top 1 cm are a factor of 3–4 times lower 300–500 0.4 80 than the levels shown in Fig. 6. >500 20 100 Measurements of plutonium activity in air carried out by the EPA in 1973–1974 in one of the highest soil activity areas were a factor of 25 less than the EPA recommended rather than from deposited fission products, and the avail- action level for airborne activity (Douglas 1978). Undoubt- able soil analysis data was too unreliable to include in the in- edly, the potential inhalation hazard was greatest shortly af- terpolations of deposition density vs. distance. ter the test when the activity was closer to the surface, and Finally, any systematic bias in the reported E12 would re- there was a possibility of inhaling descending fallout by sult in a corresponding bias in the total plutonium deposited. those living in the fallout areas. However, the hazard during fallout would be from inhalation of small respirable particles Implications for health risk (Simon et al. 2020). Because of the refractory nature of Two possible modes of intake of plutonium from Trin- plutonium, most of the plutonium in the descending fallout ity are most important with respect to potential health ef- would have been on nonrespirable large particles so the fects: inhalation of descending fallout and inhalation at potential hazard was limited. As discussed, over time the later times from the resuspension of activity on the surface plutonium penetrated deeper into the soil, thus greatly of the ground. Due to plutonium’s low transfer from gut reducing the chance in the years after Trinity of inhaling to blood, plutonium is not considered a significant hazard airborne plutonium particles, and thereby, little risk would from ingestion relative to inhalation (ICRP 1993, 1995; be expected to persons living today or in past years even in Burley 1990). the highest fallout areas. The concern is that inhalation of plutonium could lead Analyses for the Trinity dose reconstruction (Simon et al. to an increased risk of lung, bone, or liver cancer (Burley 2020) indicate that there are other radionuclides in Trinity 1990). The radiation dose would be a consequence of inhal- fallout that contribute more effectively to the exposure of the ing respirable particles of descending fallout or of contami- lung, especially in the first year when doses are the highest. nated soil that was resuspended by wind, human activity, or Those radionuclides include 239Np, 140Ba, and 237U. other activity. Previously, the US EPA (1977), using conser- vative models to relate surface soil activity to surface air ac- SUMMARY AND CONCLUSION tivity and inhalation of those airborne particles containing plutonium to organ doses, recommended action levels for Based on the soil data and our model calculations, about plutonium surface soil activity that would insure that activity 80% of the plutonium from the Trinity detonation was depos- in surface soil below these levels would result in minimal risk ited in New Mexico, primarily in a region from 50–150 km from either inhalation or ingestion. Those action levels are from GZ in a northeast direction. While the geographical based on the level below which doses (particularly lung doses) distribution of the deposited plutonium is believed to be would be well under the suggested annual radiation dose limits fairly accurate, the estimates of the total unfissioned to the public recommended by national and international orga- plutonium and the total deposited in New Mexico are nizations such as the National Council on Radiation Protection acknowledged to be uncertain for the reasons discussed www.health-physics.com Accounting for unfissioned plutonium c H.L. BECK ET AL. 513 earlier. The plutonium fallout pattern is generally consistent accounted for the unfissioned plutonium produced by the with the exposure rate pattern, though there is more Trinity test. While the residual plutonium in New Mexico from plutonium along the trace axis and close-in than fission the Trinity test is unquestionably a man-made contaminant and product activity due to fractionation. is understandably a source of apprehension to many, it does The estimated fraction of plutonium deposited within a not technically differ from the larger amount of plutonium few hundred kilometers from GZ is consistent with the ex- deposited in New Mexico from global and NTS fallout. pected higher deposition of refractory radionuclides close- Furthermore, most of this plutonium has now penetrated in due to fractionation. Hicks (1982) estimated that about to depths well below the ground surface due to natural 1/2 the refractories, on average, deposit locally, while weathering processes. Freiling (1962) suggested that, for the most refractory nu- clides, this fraction would be even greater. Based on the Acknowledgments—Publication costs for this paper were supported by Intramu- Joint US-Russian model (see Appendix), we interpret ral Research Program of the National Cancer Institute (NCI), National Institutes of Health (NIH), and the intra-agency agreement between the National Institute close-in to be the distance in which all particles >50 mmdi- of Allergy and Infectious Diseases (NIAID) and the National Cancer Institute, NIAID agreement Y2-Al-5077 and NCI agreement Y3-CO-5117. ameter will deposit. The time at which this occurs, Tmax,is based on the maximum height of the stabilized debris cloud from Hawthorne (1979) and the average gravitational set- REFERENCES tling velocity (see Appendix) of 50 mm particles. For Trin- Beck HL, Bennett BG. Historical overview of atmospheric nuclear ity, T is 14.6 h. Based on our interpolated deposition weapons testing and estimates of fallout in the continental max – densities, about 65–70% of the total plutonium was deposited United States. Health Phys 82:591 608; 2002. Beck HL. Appendix D. External radiation exposure to the population in less than 14.6 h, in reasonably good agreement with the of the continental United States from Nevada weapons tests and Hicks and Freiling fractionation estimates considering the un- estimates of deposition density of radionuclides that could signif- certainties discussed earlier. Although about 80% of the icantly contribute to internal radiation exposure via ingestion. In: unfissioned plutonium was deposited within New Mexico, A feasibility study of the health consequences to the American population from nuclear weapons tests conducted by the United the fraction of more volatile radionuclides deposited in States and other nations. Washington, DC: Department of Health New Mexico would be expected to have been much less and Human Services, Centers for Disease Control and Pre- than 80% as a result of the volatile radionuclides being vention, and the National Cancer Institute; 2001a. Available concentrated on the smaller particles that travel longer at http://www.cdc.gov/nceh/radiation/fallout/default. htm. Accessed 19 September 2019. distances before depositing. As for the unfissioned plutonium Beck HL. Appendix F. External radiation exposure to the popula- and fission products deposited outside New Mexico, we note tion of the continental United States from high yield weapons that fallout from Trinity was detected as far away as Indiana tests conducted by the United States, UK and USSR between where it caused fogging of film produced in a Kodak film 1952 and 1963. In: A feasibility study of the health conse- quences to the American population from nuclear weapons tests factory (Webb 1949). conducted by the United States and other nations. Washington, Based on soil data obtained far from the pattern axis DC: Department of Health and Human Services, Centers for (Douglas 1978; Hansen and Rodgers 1985), as well as data Disease Control and Prevention and the National Cancer Insti- reported by Beck (2001a and b), the mean deposition den- tute; 2001b. Available at http://www.cdc.gov/nceh/radiation/ fallout/default.htm.Accessed. sity in New Mexico from NTS and global fallout was − Burley G. Transuranium elements. Vol.2: technical basis for remedial about30Bqm 2. So, as can be seen from Fig. 6a, only a actions. Washington, DC: US EPA; EPA 210/1-90-016; 1990. relatively small geographical area in New Mexico has Defense Special Weapons Agency. Estimate of the radioactively Trinity plutonium activity levels appreciably higher than contaminated region due to atmospheric nuclear detonation, fi- nal report to Defense Special Weapons Agency from Russian the plutonium deposited by either NTS or global fallout Academy of Sciences. Washington, DC: DSWA; DSWA 01- and, for most of the state, the levels are at most a small 97-C-0007; 1997. fraction of the NTS and global fallout levels (Beck and Douglas RL. Levels and distribution of environmental plutonium Bennett 2002). In fact, due to the low annual precipitation around the Trinity site. Las Vegas, NV: US EPA, Office of Ra- diation Programs; technical note ORP/LV-78-3; 1978. in New Mexico, global plus NTS plutonium deposition England TR, Ryder BF. Evaluation and compilation of fission density levels there are far less than the levels in most product yields. Los Alamos, NM: Los Alamos National Labo- of the continental United States that range from ratory; LA-UR-94-3106 (ENDF-349); 1994. – −2 Freiling EC. Radionuclide fractionation in bomb debris. Science 30 150Bqm (Beck and Bennett 2002). The areas – −2 133:1991 1998; 1961. shown in Fig. 6a as 0–10 Bq m are areas where the Freiling EC. Fractionation in surface bursts. In: Radioactive fallout Trinity plutonium, if any, is indistinguishable from global from nuclear weapons tests; conference proceeding. Washington, and NTS fallout. DC: US Atomic Energy Commission; TID-7632; 1962: 25–46. Basedonourreviewandanalysisofallthehistorical Freiling EC, Crocker GR, Adams CE. Nuclear debris formation. In: Klement AW, ed. Radioactive fallout from nuclear weapons monitoring data on plutonium in New Mexico from the tests; proceeding of 2nd conference. Washington, DC: US Atomic detonation of the Trinity test, we believe we have reasonably Energy Commission; 1965: 1–41. www.health-physics.com 514 Health Physics October 2020, Volume 119, Number 4 Glasstone S, Dolan PJ. The effects of nuclear weapons. Washington, Olafson JH, Nishita H, Larson KH. The distribution of plutonium DC: US Department of Defense and US Energy Research and in the soils of central and northeastern New Mexico as a result Development Administration; 1977. of the atomic bomb test of July 16, 1945. Los Angeles, CA: Hakonson TE, Johnson LJ. Distribution of environmental plutonium University of California at Los Angeles; Report 12; 1957. in the Trinity site ecosystem after 27 years. In: Proceedings, 3rd In- Quinn VE. Analysis of nuclear test Trinity radiological and mete- ternational Radiation Protection Association Meeting. Springfield, orological data. Las Vegas, NV: National Oceanic and Atmo- VA: Washington, DC: US Atomic Energy Commission; spheric Administration for the US DOE, Nevada Operations CONF-730907; 1973. Office; NVO-313; 1987. Hansen WR, Rodgers JC. Radiological survey and evaluation of Simon SL, Bouville A, Beck HL, Melo DR. Estimated radiation the fallout area from the Trinity test: Chupadera Mesa and White doses received by New Mexico residents from the 1945 Trinity Sands Missile Range, New Mexico. Los Alamos, NM: Los nuclear test. Health Phys 119:428–477; 2020. Alamos National Laboratory; LA-10256-MS/UC-11; 1985. Tularosa Basin Downwinders Consortium. Unknowing, unwilling, Hawthorne H. Compilation of local fallout data from test detonations and uncompensated: the effects of the Trinity test on New 1945–1962. Vol. I: continental US tests. Santa Barbara, CA: Gen- Mexicans and the potential benefits of the Radiation Exposure eral Electric Company-TEMPO; DNA-1251-1-EX; 1979. Compensation Act (RECA) amendments [online]. 2017. Hicks HG. Additional calculations of radionuclide production fol- Available at https://docs.wixstatic.com/ugd/2b2028_ lowing nuclear explosions and Pu isotopic ratios for Nevada 8e221b260de7468bbcb67cbddc498dbe.pdf. Accessed 1 Test Site events. Health Phys 59:515–524; 1990. March 2019. Hicks HG. Calculation of the concentration of any radionuclide US Department of Energy. United States nuclear tests July 1945 deposited on the ground by off-site fallout from a nuclear det- through September 1992. Las Vegas, NM: US DOE Nevada onation. Health Phys 42:585–600; 1982. Operations Office; DOE/NV-209-REV 15; 2000. Hicks HG. Results of calculations of external gamma radiation ex- US Department of Energy. Restricted data declassification decisions posure rates from fallout and the related radionuclide compositions, 1946 to the present. Washington, DC: US DOE Office of Health, parts 1–8. Livermore, CA: University of California, Lawrence Safety and Security Office of Classification; RDD-8; 2002. Livermore Laboratory; UCRL-53152; 1981. US Environmental Protection Agency. Proposed guidance on International Commission on Radiological Protection. Age- dose limits for persons exposed to transuranium elements dependent doses to members of the public from intake of ra- in the general environment. Washington DC: US EPA; EPA 520/ dionuclides: part 2—ingestion dose coefficients. Oxford: 4–77-016; 1977. ICRP; Publication 67; 1993. US Environmental Protection Agency. Soil profiles of mounds on International Commission on Radiological Protection. Age-dependent plutonium-contaminated areas of the Nevada Test Range Com- doses to members of the public from intake of radionuclides: plex. Las Vegas, NV: US EPA, Environmental Systems Labo- part 4—inhalation dose coefficients. Oxford: ICRP; Publica- ratory; EMSL-LV-0539-33; 1980. tion 71; 1995. Webb JH. The fogging of photographic film by radioactive con- Land CE, Kwon D, Hoffman I, Moroz B, Drozdovitch V,Bouville taminants in cardboard packaging materials. Physical Rev 76: A, Beck H, Luckyanov N, Weinstock R, Simon SL. Account- 375–380; 1949. ing for shared and unshared dosimetric uncertainties in the Young RW, Kerr GD. Reassessment of the atomic bomb dosimetry dose response for ultrasound-detected thyroid nodules after ex- for Hiroshima and Nagasaki—Dosimetry System 2002 (DS02). posure to radioactive fallout. Radiat Res 183:159–173; 2015. Hiroshima: Radiation Effects Research Foundation; 2005. McArthur RD, Miller FL. Offsite radiation exposure review pro- Available at www.rerf.jp/shared/ds02/index.html. Accessed ject phase-II soils program. Las Vegas, NV: US DOE Nevada 19 September 2019. Operations Office/Desert Research Institute Water Resources Center; DOE/NV/10384-23RE; 1989. Nyhan JW,Miera FR, Neher RE. Distribution of plutonium in Trinity soils after 28 years. J Environ Quality 5.4:431–437; 1976. ■■

www.health-physics.com Accounting for unfissioned plutonium c H.L. BECK ET AL. 515 APPENDIX maximum height of the debris cloud, all particles >50 mm will have deposited, and all activity subsequently deposited will be on particles <50 mm. Model for estimating Pu/E12 vs. R/V (fractionation) N50, the fraction of particles less than 50 mm, is di- The ratio of Pu/E12 for unfractionated fallout, as de- rectly related to the average R/V, i.e., to the average ratio − − − scribed earlier, was calculated to be 2.7 Bq m 2 (mR h 1) 1 of refractory nuclide activity to volatile nuclide activity. Be- but to vary with the degree of fractionation (Table 3). R/V cause plutonium is a highly refractory element, an estimate of is known to be >1 close to the Trinity test site where most the relationship between R/V and N50 allows us to estimate of the activity that was deposited would be on relatively large R/V and thus, the ratio of plutonium activity to the activity particles, compared to R/V < 1.0 at distances where most of of nuclide Z (Pu/Z), where Z is a volatile nuclide such as the activity would be on small particles. Thus, we need to es- 137Cs. Using the Pu/137Cs ratio calculated from the amount timate R/V at each location to estimate the Pu/E12 ratio at of fuel in the device and the reported test yield and 137Cs/ that location to estimate the plutonium deposition density at E12 (Bq m−2 [mR h−1]−1) vs. R/V based on Hicks (1981), that site. Using a semiempirical model based on the work we then can calculate Pu/E12 (Bq m−2 [mR h−1]−1)vs.R/V. 10 11 of Gordeev as modified by Beck, in conjunction with Using regression fits to actual measurements of N50 the calculated relative radionuclide activities in Trinity fallout for a number of NTS and Russian tests (unpublished),11 vs. E12 from Hicks (1981), we can accomplish this using the N50 is estimated by: measured E12 and fallout TOA reported by Quinn (1987). The predictive model for Pu/E12 is based on estimat- 3 N50a ¼ ðÞ1−a EXP½−ðÞd tr ðÞonthe axis ing the fraction of radioactive fission products on particles m <50 m in diameter. This fraction allows one to estimate − : the fraction of activity that would be deposited and retained N50 ¼ N50a 1 3SQRTðÞN50a ðA1Þ on vegetation and available for transfer from animals such lnðÞ El2=El2max ðÞoff axis at same TOA ; as cows to milk consumed by humans (Land et al. 2015). −1 The predictive model reflects the fact that refractory nu- where tr=TOA/Tmax, Tmax = CT/0.73, 0.73 km h is the clides condense from the nuclear debris plasma earlier than gravitational settling velocity of 50 mm particles, CT is the volatile nuclides and thus, tend to be incorporated in larger height of the cloud top in km, and E12max is the exposure particles (Freiling 1961, 1962; Hicks 1982). As the debris rate at H + 12 h on the axis at a given TOA. cloud cools, the more volatile nuclides condense and tend Based on the regression fits to measured particle activ- to deposit on the surface of smaller particles. Due to gravi- ity, the parameters a and d have been shown to vary slightly tational settling, the larger particles deposit earlier after the among nuclear tests depending on the height of burst (a) and detonation than smaller particles. Hence, the earlier the the amount of wind shear (d). Based on these fits, we esti- TOA, the larger the proportion of large particles deposited mated 1 − a =0.98,d = 1.8 for Trinity based on the reported as well as a greater proportion of total activity deposited that height of the burst and the observed width of the fallout pat- is on large particles and conversely, the smaller the fraction tern. As indicated by eqn (A1), N50 at a given TOA increases on small particles. This phenomenon is termed fractionation as one moves away from the fallout pattern axis. and describes the phenomenon where the ratio of refractory The relationship between N50 and R/V was inferred to volatile radionuclides (R/V) in the deposited fallout dif- from observations of 137Cs/E12 at the NTS and Russian nu- fers from the ratio of refractory to volatile fission products clear testing test sites at various TOAs supplemented by pub- in the debris cloud. This implies that R/V is higher at loca- lished observations of the fraction of activity on small tions close to the detonation site and lower at more distant particles (Beck 2009; Hicks 1982; Freiling 1961, 1962; locations. At some time after detonation, depending on the Freiling et al. 1965). It is predicated on the model assumption that N50, by definition, approaches 1.0 as the TOA ap- 10 Gordeev KI. Radiation exposure to the population of the Semipalatinsk proaches Tmax, i.e., that all particles less than 50 mmhave region from Semipalatinsk weapons tests. Part I. Experimental and theo- retical investigation of the processes of radioactive contamination of grass resulting from local fallout from nuclear explosions and justification of the Table A1. Estimated relationship between N50 and R/V. concept of “biologically active fraction” of fallout. Report to the National N50 R/V 11Cancer Institute, Bethesda, MD; 1999. Unpublished. Beck HL. Estimates of H+12 exposure rates and associated uncertainty received by the population of selected villages in the vicinity of the Sem- >0.83 0.5 ipalatinsk Test Site in Kazakhstan at the time of the Semipalatinsk Test Site 0.43–0.83 1.0 – atmospheric tests (1949 1962). Application of joint Russian/American 0.23 to <0.43 1.5 methodology for fallout dose assessment to the estimation of internal doses and associated uncertainty. Final report to NCI in fulfillment of 0.09 to <0.23 2.0 PSC #HHSN261200800397P; 15 March 2009; unpublished. A peer-reviewed <0.09 3.0 publication is being prepared for journal submission.

www.health-physics.com 516 Health Physics October 2020, Volume 119, Number 4 been deposited when t = Tmax. Based on Hicks’ observations Although the semiempirical model for N50 described of R/V for fallout from NTS tests, R/Vasymptotes to a value here has been shown to adequately reproduce measured 11 of 0.5 as N50 approaches 1.0. We have estimated an approx- N50 forbothNTSandRussiantests, the basis for the imate, though simplistically modeled, relationship between estimated relationship between N50 and R/V is less rigor- N50 andR/VshowninTableA1. ous and thus somewhat more uncertain than the esti- For each location where plutonium deposition was mates of N50. calculated, we adjusted the unfractionated Pu/E12 value by the estimated R/V at that site, using Table 3, eqn (A1), and the above relationship between N50 and R/V, to estimate from the interpolated E12 corrected pluto- nium deposition density. ■■

www.health-physics.com