Dose Estimation for Exposure to Radioactive Fallout from Nuclear Detonations.

ABSTRACT: In recent years, the prospects that a nuclear device might be detonated due to a regional or global political conflict, by violation of present nuclear weapons test ban agreements, or due to an act of terrorism, has increased. Thus, the need exists for a well conceptualized, well described, and internally consistent methodology for dose estimation that takes full advantage of the experience gained over the last 70 y in both measurement technology and dose assessment methodology. Here, the models, rationale, and data needed for a detailed state-of-the-art dose assessment for exposure to radioactive fallout from nuclear detonations discussed in five companion papers are summarized. These five papers present methods and data for estimating radionuclide deposition of fallout radionuclides, internal and external dose from the deposited fallout, and discussion of the uncertainties in the assessed doses. In addition, this paper includes a brief discussion of secondary issues related to assessments of radiation dose from fallout. The intention of this work is to provide a usable and consistent methodology for both prospective and retrospective assessments of exposure from radioactive fallout from a nuclear detonation.

INTRODUCTION

Since the Manhattan Project, when scientists began contemplating the test of the Trinity device in 1945, it has been recognized that radioactive debris in the form of particles of varying sizes, composition, and activity created by a nuclear detonation would eventually deposit on the earth. The phenomenon, termed fallout, was recognized from the onset of testing as a potential health hazard to man and animals. Similarly, it was recognized that the fallout from nuclear detonations could spread radioactive debris over sizeable geographic areas and potentially expose large populations at significant distances from the detonation site. Within the context of this paper and its companion publications, the term fallout refers primarily to the radioactive particulate debris from nuclear weapon detonations, though it also can include vapors of iodine and other volatile radionuclides. While the methods described herein are not specifically tailored for assessments of the doses from radioactive debris distributed by non-fission explosive dispersion devices, some applications to such scenarios are possible.

Over the 75 y since the Trinity test, hundreds of papers have been published on the basic mechanisms of the creation and dispersion of radioactive fallout and its movement and transport in the environment. Despite the many publications on specific physical and chemical attributes of radioactive fallout, there is little in the scientific literature to summarize the various approaches to dose estimation that have been used over the past decades to provide a complete methodology that can be used to assess radiation doses from exposures to radioactive fallout that might occur from events in the future, or for reassessments of past fallout exposures. Moreover, only a few investigations have published estimates of radiation doses received by specific populations downwind of the nuclear tests. While even fewer have included estimates of the resulting health risks, it is worthwhile to recognize that the endpoint of interest in studying exposures to fallout is to understand and reduce health risks from future events.

In recent years, threats of international terrorism, including the possible use of radioactive materials have come to the attention of the public and lawmakers alike. For example, in 2005, Public Law No. 108-276, i.e., the Project Bioshield Act of 2005 (https://www.congress.gov/bill/108th-congress/senate-bill/15) brought the potential threats of chemical, biological, and radiological terrorism to the attention of the public and Federal Government agencies. The law enhanced the capabilities of the Department of Health and Human Services and the Department of Homeland Security to prepare for future radiological threats. Following P.L. 108–276, the recognition of the importance of fallout from nuclear detonations as a threat to human health in the US was clearly noted in the Strategic Plan of the National Institutes of Health, National Institute of Allergy and Infectious Diseases (NIAID 2005).

The worst scenario would be the detonation of a nuclear explosive device which, in addition to causing enormous destruction from detonation and heat, would produce an intense burst of gamma radiation and large quantities of radioactive “fallout.”

One of the long-term goals of the NIAID program (NIAID 2005) has been to: Develop and validate methods to estimate radiation dose and future risk following exposure to radioactive material by various routes, including inhalation, ingestion, skin contact, or contamination of wounds.

Given the recognition of the potential scale of nuclear-related terrorism as well as the goals of the NIAID program on Radiation and Nuclear Countermeasures (https://www.niaid.nih.gov/research/radiation-nuclear-countermeasures-program), this paper and five others (Beck et al. 2022; Bouville et al. 2022; Anspaugh et al. 2022; Melo et al. 2022; Thiessen et al. 2022) present a well described method that remedies the absence of clear documentation of a dose-assessment method suitable for estimating doses from exposure to radioactive fallout at locations where exposure is entirely from fallout. The models presented can apply to both rural and urban locations, though some values of parameters would be strikingly different for the two environments.

The models are based on an extensive review of the literature on the physical and chemical attributes of fallout and mechanisms of fallout deposition and movement in the environment, and coalesce knowledge gained from years of experience in the fields of radiation dosimetry, fallout dose assessment, and dose reconstruction. We cite relevant and important publications that have used earlier incarnations of the methods presented here or portions of the methods. The methods presented are, of course, not entirely new but draw upon previous research and decades of accumulated experience. In this set of papers, we provide and discuss the model parameter values needed for estimating external and internal radiation absorbed doses and effective doses from exposure to fallout from nuclear detonations.

Herein, we specifically do not address the subject of measurement techniques or simulation and estimation of air dispersion of fallout, or of exposure from prompt gamma rays and neutrons released in the detonation itself. In the present work, determination or simulation of air dispersion is not considered as a requisite capability for estimating radiation doses from exposure to fallout. As discussed in the companion papers, the dose-calculation strategy begins with and relies on measurement data. The determination of the doses received by individuals and populations from exposure to deposited radioactive debris can be suitably addressed using estimates or measurements of either exposure rate above contaminated ground or of ground-radioactivity contamination at the site of exposure coupled with estimates of the transit time for the fallout to reach the measurement site. This work assumes that radiation-monitoring data over large geographic areas could be quickly available after nuclear detonations, regardless of whether those events were anticipated or not.

The primary purpose of this paper is to introduce the concepts presented in companion papers on the computational tools presented for estimating radionuclide deposition density (Beck et al. 2022), external dose (Bouville et al. 2022), internal exposure via ingestion and inhalation (Anspaugh et al. 2022), and intake to dose-conversion coefficients particular to fallout exposure (Melo et al. 2022), as well as to provide guidance on how best to use this methodology to assess doses from fallout to both individuals as well as populations. In addition, we introduce tabulations of relevant model parameter values (Thiessen et al. 2022) that are needed to model environmental transport processes of fallout and discuss other issues related to the assessment of doses and how they can be communicated. The suite of methods presented here is part of a proposed comprehensive schema for dose assessments for exposures to radioactive fallout from nuclear detonations.

RATIONALE, PURPOSES, AND PRIORITIES

As noted earlier, limited documentation exists in a single volume of a well conceptualized and well described dose assessment strategy for exposure to radioactive fallout that includes the more important pathways and radionuclides. Some publications are available that describe dose-assessment strategies from previous studies, including those of the Offsite Radiation Exposure and Review Project or ORERP (Church et al. 1990), University of Utah (Lloyd et al. 1990; Simon et al. 1990), a nationwide evaluation of thyroid dose due to Nevada fallout by the National Cancer Institute (1997), Beck et al. (2006), and a contemporary evaluation of the Trinity test (Bouville et al. 2020; Simon et al. 2020), though none are complete descriptions of dosimetric methods for all pathways, radionuclides, organs, and exposure scenarios with parameter values and dose coefficients. Hence, the rationale for this work is to provide and document a more complete method that may stand alone or supplement other available resources.

Some resources for fallout dose estimation today are already available as federal assets, e.g., the FRMAC (Federal Radiological Monitoring and Assessment Center) and its manuals. The FRMAC manuals are intended to address the early and intermediate phases of a radiological incident. However, the assessment methods for a nuclear weapon explosion resulting in a significant nuclear yield continue to be developed. Hence, the methods provided here can be viewed as a priority for national security reasons and health consequence aversion.

The more general purposes and priorities of dose estimation for nuclear events involving radioactive fallout include the needs: (1) to make assessments as part of a federal and/or state response to an unexpected nuclear detonation—in part to determine the necessity of protective actions including sheltering in place, evacuation or in some cases, medical triage; (2) to determine the potential effectiveness of proposed countermeasures in future fallout events and to support decision making to implement countermeasures, e.g., evacuation, food interdiction, stable iodine administration, or medical administration of chelating agents, etc.; and (3) to support health-risk evaluations and epidemiologic studies, usually months to years after the exposure, or even for lawsuits and compensation claims. Additional purposes of dose assessments include fulfilling the public’s and press’s “right to know” and for historical documentation providing information on the degree of exposure of nearby populations.

HISTORY AND BACKGROUND

Nuclear testing and the early years

Before the first nuclear test explosion in 1945, physicists were aware of the potential problems of atmospheric fallout of fission products from a successful nuclear detonation and of dispersed plutonium should a nuclear explosion fail (Hacker 1987; Hoddeson et al. 1993). This led to, at least, rudimentary efforts to measure alpha, beta, and gamma activity in the near- and far-fields following the first nuclear test, Trinity, conducted on 16 July 1945 in New Mexico. By the time the Nevada Test Site (NTS) was established and the first nuclear test conducted there on 27 January 1951, it was routine that fallout was being measured and documented in Nevada, Utah, and parts of other nearby states, by field monitors using beta and gamma survey meters (Hawthorne 1979). More extensive monitoring evolved with time at continental test sites (Thompson et al. 1994) and Pacific test sites (Eisenbud 1997). Exposure-rate data were eventually used to estimate cumulative exposure and cumulative collective exposure for populations residing near to the NTS (Anspaugh and Church 1986; Anspaugh et al. 1990).

By the second series of tests in Nevada, which started on 22 October 1951, nationwide monitoring was implemented with the use of gummed-film collectors that were analyzed for beta activity (Beck et al. 1990; Eisenbud 1997). The nuclear tests conducted at the NTS were rather small in explosive yield compared to the larger tests being conducted by the U.S. on Pacific Atolls including the Marshall Islands and Johnston Atoll (Simon and Robison 1997) and by the former USSR at sites in Russia and Kazakhstan. The large tests led to the deposition of radioactive debris at much greater distances throughout the world in what is usually called “global fallout.” Global fallout, in turn, led to enhanced concerns about the possible health effects at great distances due to incorporation of long-lived radionuclides, particularly 90Sr and to a lesser extent 137Cs, into the human body. Congressional hearings were held (US Congress 1957, 1959) with a major emphasis on external dose and on internal dose from the ingestion of 90Sr and other long-lived radionuclides.

It was not until 1963 that the potential dosimetric and health impact of short-lived radionuclides released from nuclear tests was fully recognized, although there had been an early unsuccessful attempt to measure 131I in milk from a contaminated area in Utah (Johnson et al. 1953), and a US atomic energy scientist had indicated the possible importance of the dose to the thyroid from ingestion of fallout-derived 131I (Dunning 1958). Also, in 1957 there was a major accident involving a graphite fire at a reactor at Sellafield (formerly known as Windscale) in the UK with the release of about 7 × 1014 Bq of 131I. Due to resulting high levels of 131I in milk and other foods, milk and vegetables from an area of approximately 500 km2 were withdrawn from public consumption (UNSCEAR 1962). With knowledge of the problems caused by the release of 131I at Sellafield, Knapp (1963) analyzed a systematic set of data collected by the Public Health Service (PHS) of external gamma-exposure rates in the same location where cows’ milk was collected for measurement of 131I. These measurements followed the Small Boy nuclear test at the NTS on 14 July 1962. Small Boy was a slightly above-ground detonation with a yield reported as “low” (less than 20 kt) (US DOE 2000). The primary measurements were made at Alamo, NV (~90 km from the NTS) and Caliente, NV (~150 km). Knapp’s seminal contribution was an understanding that the level of 131I activity in milk could be correlated to the external gamma-exposure rate normalized to the same point in time following the detonation. The revelation that there could be such a correlation led to a speculative re-analysis of how large the radioiodine concentrations in milk and the related dose to infant thyroids might have been at other locations where there had been much higher levels of external gamma-exposure rate during 1953. These findings were discussed extensively during another Congressional hearing (US Congress 1963).

In the late 1970s and early 1980s, controversial studies were published that suggested an increase in leukemia and other cancers due to civilian and military exposure to low levels of radiation. Two of these studies were related to the NTS: (1) military participants at the Smoky test in 1957 (Caldwell et al. 1980, 1983), and (2) leukemia in Utah children (Lyon et al. 1979). Although these epidemiologic studies were not conclusive, they, combined with other factors [in particular, the filing of many claims of injury, two major lawsuits, and a provocative congressional report (US Congress 1980)], resulted in demands for a thorough evaluation of the radiation doses received by persons living downwind of the NTS.

In order to provide a thorough evaluation, doses from the ingestion of all short- and long-lived radionuclides had to be considered, in addition to external dose. There, for the first time, the relationship between 131I in environmental media and the normalized rate of external gamma-exposure was exploited for dosimetric purposes. The obvious extension of this process was to relate the presence of all short- and long-lived radionuclides in environmental media to the normalized external gamma-exposure rate. In order to perform that operation, two sets of information needed to be available: (1) results of calculations of external gamma-exposure rate for each radionuclide normalized to a unit of deposition density, and (2) a description of the time-dependent amount of each radionuclide created by a nuclear explosion. The first problem was solved by Beck (1980), who calculated the normalized exposure rates, and the second was solved by Hicks (1981, 1982), who created a time-dependent list of the activity of each fission and activation product for each individual test at the NTS using fission-product yields (Nethaway and Barton 1973) for each test coupled with an appropriately modified version of the ORIGEN computer code (RSIC 1979). Hicks coupled his calculations with those of Beck to create tabulations of the ground deposition density of 150 fission products and 27 activation products, normalized to the external gamma-exposure rate at 12 h post-detonation (H + 12). Hicks was able to include data on activation products because he had access to detailed radiochemical data on each test.

The factors in the tables published by Hicks thus provide the Rosetta Stone for calculating ground-deposition density (activity per unit area) at any time after a nuclear detonation and, in doing so, provided the means to estimate internal doses from radionuclides released in the detonation and subsequently transported in the environment. In addition, the decay rate for each test, usually approximated by the well-known t relationship for beta decay (Way and Wigner 1948), was improved and presented as a 10-term exponential relationship (Hicks 1982; Henderson 1991). With the estimates of the deposition density of each radionuclide, it is possible to estimate internal dose to humans through the use of environmental transport models and dose coefficients.

Assessments of fallout doses

The history of estimation of fallout-radiation doses spans the same time period as the history of nuclear detonations. Atmospheric nuclear testing began in July 1945 with the US Trinity test in New Mexico and continued until 1980 with the last atmospheric nuclear test by China. As is well known, the only detonations in the history of the world that were not part of nuclear testing programs were the Hiroshima and Nagasaki detonations in Japan in August 1945.

Many years were spent in reconstructing and improving the dose assessment for Japanese atomic bomb survivors (Auxier 1977; RERF 1986; Young and Kerr 2005; Kerr et al. 2013), though the methods developed have limited relevance to exposures from fallout because the bombings resulted primarily in prompt gamma-ray and neutron exposures with relatively little exposure to radioactive fallout (Sakata et al. 2014; Kerr et al. 2015).

One source of useful documentation on dose estimation for fallout exposures is from what was the first multi-institutional effort to reconstruct Nevada Test Site exposures: the DOE-sponsored Offsite Radiation Exposure and Review Project (ORERP). Much of the groundwork for estimating radionuclide deposition from exposure rate was developed in the 1980s (Beck 1980) and applied to the ORERP dose reconstruction, along with innovations in pathway exposure modeling (Whicker and Kirchner 1987). There have been many other publications from the ORERP, including Anspaugh et al. (1990), Church et al. (1990), Henderson and Smale (1990), Ng et al. (1990), Whicker et al. (1990, 1996), Kirchner et al. (1996).

Using techniques developed by the ORERP study, further assessments of fallout contamination, exposures, and risks were conducted in a variety of studies including the distribution of fallout from Nevada testing across the US (Simon et al. 2004) and leukemia and thyroid disease in cohorts in Utah (Stevens et al. 1990; Lyon et al. 2006) based on dose assessments by Simon et al. (1990), Lloyd et al. (1990), Till et al. (1995), and Simon et al. (2006a). In later years, dose assessments for nuclear testing conducted in the Marshall Islands (Simon et al. 2010), French Polynesia (Drozdovitch et al. 2008, 2021), Kazakhstan (Gordeev et al. 2006a, 2006b; Simon et al. 2006b; Land et al. 2015), and for the 1945 Trinity test in New Mexico (Beck et al. 2020, Bouville et al. 2020, Simon et al. 2020), were completed, all using the basic strategies of converting exposure rate to ground deposition of individual radionuclides for the purpose of computing internal doses, as developed by the ORERP.

More general assessments and reviews of assessments have also been conducted for populations residing near nuclear test sites (Simon and Bouville 2002) and for local and global fallout (Bouville et al. 2002; DHHS 2005).

BASIC CONCEPTS

In this section, we seek a common language and understanding of basic concepts of methods for dose estimation for exposure to radioactive fallout in order to facilitate an understanding of our companion papers as well as of the literature that we have drawn upon. While the term fallout, referring to the radioactive debris particles, and the process of their movement from the atmosphere to the ground is more or less universally understood, other concepts are briefly discussed here so as to ensure a consistent understanding: (1) temporal definitions, (2) empirical and theoretical assessment models, (3) time-dependent (dynamic) and time-integrated assessment models, and (4) dose units. The subject of dosimetric uncertainty, an important over-arching component to fallout dose assessments, is discussed throughout the five companion publications (Beck et al. 2022; Bouville et al. 2022; Anspaugh et al. 2022; Thiessen et al. 2022; Melo et al. 2022) and summarized here.

Temporal definitions

Assessments of radiation dose from exposure to nuclear testing fallout can have a temporal definition: retrospective or prospective. For example, an assessment may be conducted (1) for doses already received from a past event, i.e., a retrospective assessment; or (2) for doses that might be expected to be received from a recent event or for an event that has not yet taken place, i.e., a prospective assessment. While the models of the physics of fallout-transport processes (e.g., deposition, interception, retention, weathering, decay, etc.) may be chosen with little regard to whether the event has taken place yet, choice of parameter values and model modifications as discussed in the papers in this volume will depend on the temporal framework.

While the predictions of health risks can be made immediately after doses of exposed populations are calculated, observational or follow-up studies of health consequences generally require substantial passage of time after the exposure as there can be a long latency period (years) for some health outcomes, e.g., cancer. Epidemiological studies to quantify health consequences from past exposures to fallout require, by definition, retrospective assessments while risk projection, a type of epidemiological study that does not require quantitative records of actual health consequences, can be used to estimate expected health consequences either in the future or that may have occurred in the past. The dose-assessment methods described in this group of papers can generally be applied to either retrospective or prospective assessments. The main differences in application would be in the selection of model-parameter values and whether countermeasures need to be accounted for in the modeling. As discussed in the papers in this volume, retrospective studies may require additional or more detailed data than those needed for an initial quick assessment, such as soil-sample analyses, determination of device yield and fuel, etc., as well as specific information on the diets and human activities of the exposed populations.

Empirical and theoretical models

In the literature on dose assessment for radioactive fallout, as for other types of dose reconstructions (e.g., NCRP 2010), models used for dose estimation are often a composite of sub-models, which are smaller models that apply to a specific step in the dose reconstruction, e.g., movement of fallout from air to ground, from ground to biota, from biota to man, etc. The smaller model components are often empirically based, although in some cases, parts of the overall assessment model may be theoretically based. As is well known, empirical models are usually based on observations and measurements and are simplistic in the sense that they are usually based on statistical fitting of data sets using regression or interpolation/extrapolation techniques.

The weakness of empirical models is that their predictive capability is best suited for situations that are closely like those from which the data were collected. Empirical models are appealing, however, in that they can often be produced relatively quickly and easily and often closely reproduce the fitted data. Empirical models have been used widely, for example, to describe relationships of concentrations in different parts of the biosphere, usually under the assumption of steady state (i.e., equilibrium conditions). Many of the parameter values presented in this volume by Thiessen et al. (2022) are, in fact, based on simple empirical relationships or simple empirical models. This would include, for example, ratios between radionuclide concentrations in feed types and in animal tissues or animal products. In contrast, theoretical models, sometimes called mechanistic models, rely on a detailed understanding of physical or biological principles in order to simulate the steps in the processes. Theoretical and mechanistic models usually are complex in their construction and have demanding data requirements and computation times. An additional problem with theoretical models is that the underlying processes may be much more complicated than presumed by the modeler.

In our proposed methodology, the models are primarily empirically based, although some have theoretical bases as well. This would include, for example, the model for estimating refractory to volatile radionuclides and for determining the change in exposure rate with time after detonation (Beck et al. 2022). One example of a fallout dose-assessment code that included both empirical and mechanistic sub-models was PATHWAY (Whicker and Kirchner 1987; Whicker et al. 1990).

Time-dependent and time-integrated models

Models to estimate radiation doses to individuals or populations can be formulated with varying levels of detail depending on the particular purpose and interests to be served. As described in the previous section, the most detailed models, often called process-level or mechanistic models, attempt to account for the physics and biology of processes rather than just the outcomes of the processes (e.g., concentrations of radionuclides at specified steps of the food-chain or doses received). Process- or mechanistic-level fallout models can be formulated to produce time-dependent outcomes and are useful for improving our understanding of the processes of fallout transport and chemical behavior in the environment and in organisms. The PATHWAY model (Whicker and Kirchner 1987), which was created partly as a research tool and partly as a dose-assessment tool, had the capability to make time-dependent predictions. PATHWAY is discussed in more detail by Anspaugh et al. (2022). One advantage of time-dependent models is that model predictions for specific points in time and specific locations can be compared directly with measurement data (e.g., concentrations of radionuclides in defined parts of the biosphere or in man) obtained at the same points in time and locations. Process-level and time-dependent fallout models, however, usually require detailed sub-models, numerous parameters and parameter values, data that are often difficult to obtain, and are often uncertain to a high degree. Moreover, development or programming of software for solving sets of time-dependent differential equations is considerably more demanding than performing calculations based on closed-form solutions, making time-independent solutions quite attractive.

For the purpose of informing emergency-response decisions after a nuclear detonation, an assessment of fallout-related contamination of food products and related doses from ingestion of those foods can generally be estimated by time-integrated models, i.e., models that provide dosimetric quantities that represent the integral number of radioactive decays over a defined period of time. In Thiessen et al. (2022), a good example of time-integrated parameters is feed-to-milk transfer coefficients. Time-integrated parameters need to be used, however, with an understanding of their limitations and how they are to be applied. For example, some past usages have improperly used time-integrated parameters for predicting concentrations in a specific type of biota (e.g., Bq kg−1) at specific points in time. The correct usage is for estimating the integral decays in a specific type of biota (Bq d kg−1).

Dose units

Various units to quantify dose from fallout dose-assessment methods can be used depending on the purpose and intended use of the assessment. In this work, we present internal dose coefficients for both absorbed and effective dose and give guidance for estimating equivalent dose when needed. As explained, the dose coefficients we provide (Gy Bq−1 or Sv Bq−1) should allow the user to estimate the metric of dose most useful for the purpose at hand (Melo et al. 2022).

Absorbed dose (Gy; J kg−1) is often used to describe the dose received to individual organs and is particularly useful when the doses differ from one organ to another, e.g., as might be the case for external irradiation when there is partial body shielding or for internal exposure as a consequence of the biokinetic behavior of different radionuclides in the body.

The equivalent dose to a tissue is the product of the absorbed dose and a radiation-weighting factor that approximately reflects the relative biological effectiveness of the radiation type. A radiation-weighting factor of unity is recommended by ICRP for photons and electrons (ICRP 2007) and larger than unity for neutrons and alpha particles. Because radiation-weighting factors are unitless, the SI unit of equivalent dose is the same as that of absorbed dose (J kg−1); however, given that it is also called the sievert (Sv), it frequently introduces confusion with the effective dose, also reported in Sv.

While it is common to express regulatory quantities in terms of effective dose and common for medical professionals, writers and the press, and in some cases radiation-protection specialists, to use effective dose, it should be understood that the weighting factors used to derive effective dose are only approximations based on expert judgement and are subject to change over time with new information and updated evaluations. While effective dose can be used to estimate the excess health detriment due to cancer and hereditary effects (though not for specific organs, age groups, or populations; Menzel and Harrison 2012, Fisher and Fahey 2017), it should not be used for epidemiologic studies that are designed to quantify risk. Dose estimation for analytic and follow-up epidemiologic studies that seek to characterize the stochastic radiation risk must use organ absorbed dose as both equivalent and effective dose contain prejudgments on the risk.

METHODOLOGICAL APPROACH

In this section, we provide both generalizations about the overall proposed methods as well as discuss important details from each of the model components presented in the companion papers.

In fallout dose assessments, the level of detail, the level of realism sought, and the degree of individualization of doses can vary depending on the intended purpose and demands made on the assessors. At the time of dose computation, various approximations can be made considering the urgency and needs for either “quick and dirty” computations or for more detailed computations that attempt greater realism. In this work, for example, Bouville et al. (2022) discuss quick computations vs. more detailed dose assessments for external dose. Anspaugh et al. (2022) discuss two alternative methods for estimating internal doses, including the use of (1) pre-calculated and simple-to-use estimates of intake by man, by time of year, normalized to ground deposition, i.e., Bq intake per Bq m−2, as derived from detailed earlier published pathway analyses; and (2) analytical solutions for individual sub-models that allow the assessor to tailor the internal dose assessment more specifically to the situation of interest. Beck et al. (2022) also provide default model parameter estimates that allow quick and rough estimates to be made of deposition density that can be followed up at a later time by more precise estimates as more and better measurements become available.

Radionuclides

One of the more important issues is to determine the set of radionuclides for dose assessment. The radionuclides proposed here for consideration in future dose assessments of radioactive fallout are based on extensive experience in calculating the relative dose received for different tests at different locations and for various times-of-fallout arrival. For this work, we have derived a set of proposed radionuclides based on previous evaluations on the relative dose provided by each (Ng et al. 1990; DHHS 2005; Simon et al. 2010, 2020; Kraus and Foster 2014).

Table 1 specifies the 34 radionuclides we recommend for estimating internal dose from future fallout-related dose assessments. Based on our extensive experience in assessing fallout doses, these nuclides are believed to account for over 95% of the dose from both ingestion and inhalation. As discussed in Beck et al. (2022) and Bouville et al. (2022), all fission and activation-product nuclides contributing to external exposure are automatically included in the assessment of external dose. This list is somewhat shorter than the list of radionuclides considered in the DHHS report (2005) and the analysis of doses received in the Marshall Islands (Simon et al. 2010) and in New Mexico from the Trinity test (Simon et al. 2020). Possibly the radionuclide that is sometimes associated with nuclear detonations but is least common to other dose assessments is 239Np. Neptunium-239 played an important role in the Trinity nuclear test assessment (Simon et al. 2020) as it arises from neutron capture in 238U and can be significant for bomb designs such as that for Trinity, which included a heavy uranium tamper/reflector around a plutonium core, and for uranium-fueled detonations.

Table 1

Thirty-four radionuclides recommended for consideration in future assessments of dose from radioactive fallout; indented radionuclides are progeny of preceding (parent) nuclides. The doses from activities of progeny nuclides that arise in the body following intakes of the parents are accounted for by the parent dose coefficients; the doses from activities of progeny nuclides deposited on the ground are to be calculated separately.

Radionuclide	Half-life	Organs/tissues in adults receiving greatest dose via ingestion	Organs/tissues in adults receiving greatest dose via inhalation
89Sr	50.6 d	Bone surfaces, Colon, active marrow	Bone surfaces, Colon, active marrow, lung
90Sr	28.9 a	Bone surfaces, Colon, active marrow	Colon, bone surfaces, active marrow, lung
90Y	64.1 h	Stomach wall, colon	Stomach wall, colon, lung
91Sr	9.65 h	Colon	Colon, lung
91Y	58.5 d	Colon	Colon, lung
92Sr	2.66 h	Stomach wall, colon	Stomach wall, colon, lung
92Y	3.54 h	Colon	Colon, lung
93Y	10.2 h	Stomach wall, colon	Stomach wall, colon, lung
97Zr	16.7 h	Colon	Colon, lung
97Nb	72.1 min	Stomach wall, colon	Stomach wall, colon, lung
99Mo	66.0 h	Colon	Colon, lung
99mTc	6.0 h	Stomach wall, thyroid	Lung, colon
103Ru	39.2 d	Bladder wall, stomach wall, colon, ovaries, active marrow, uterus	Colon, lung
103mRh	56.1 min	Stomach wall, colon	Lung, stomach wall
105Ru	4.4 h	Colon, stomach wall	Lung, colon
105Rh	35.4 h	Colon	Lung, colon
106Ru	372 d	Stomach wall, colon	Colon, lung
132Te	3.20 d	Bone surfaces, colon, thyroid, ovaries	Bone surfaces, colon, thyroid
132I	2.30 h	Thyroid	Thyroid
131I	8.03 d	Thyroid	Thyroid
133I	20.8 h	Thyroid	Thyroid
135I	6.58 h	Thyroid	Thyroid
137Cs	30.1 a	All	All
140Ba	12.8 d	Bone surfaces, colon, ovaries	Colon, lung
140La	1.68 d	Colon	Colon, lung
141La	3.92 h	Stomach wall, colon	Stomach wall, colon, lung
142La	91.1 min	Stomach wall, colon	Stomach wall, colon, lung
143Ce	33.0 h	Colon	Colon, lung
143Pr	13.6 d	Colon	Colon, lung
144Ce	285 d	Colon	Colon, lung
144Pr	17.3 min	Stomach wall	Stomach wall, lung
145Pr	5.98 h	Stomach wall, colon	Stomach wall, colon, lung
239Np	2.36 d	Colon	Colon, lung
239Pu	24,100 a	Colon	Lung

Age groups, ethnicities, and sexes

Age groups are simple specifications of ranges in age that approximately match body sizes for which radiation-transport calculations have been completed in appropriately sized computational phantoms. In this work, we endorse the ICRP (2011) age groups of in utero, newborn (0–1), 1–2, 3–7, 8–12, 13–17, and 18+ y of age. In a dose assessment following a radioactive fallout event, all persons potentially exposed can be suitably assigned to one of the age categories. This assignment will determine both the external and internal dose coefficients selected.

While there can be differences in doses received by members of different ethnic groups because of differences in lifestyle and diet even in the same locations, in this work, no specific recommendations are made in our models for distinguishing ethnicities. In previous US assessments, fallout dose has most often been assessed for typical White populations, though occasionally also for Hispanic, African American, and Native American groups (e.g., Potischman et al. 2020; Simon et al. 2020). Depending on the location and event, consideration of other ethnicities, e.g., Asian, East Indian, Pacific Islander, etc., might also prove to be necessary in the future, even at US continental locations. However, in the modeling of both external and internal dose, differences in home construction that affect shielding and in dietary intakes that can affect internal dose can easily be incorporated by selection of relevant parameter values and, if necessary, minor modifications to dose models.

Modification of assessment parameter values for male and female is usually simpler than for age or ethnicity because the choice is simply a choice of two alternatives. Those assessment parameters that have sex specification in the presented methods include food-intake and ventilation rates (Anspaugh et al. 2022).

Spatial variations in parameters and attributes

Various attributes that describe a fallout-radiation field must be known or estimated from related data in order to properly use the presented methods for estimating doses, whether it be external dose, internal dose via inhalation, or internal dose via ingestion. In particular, the ground-level exposure rate at each site of interest at a known time and the time-of-fallout arrival [(TOA) measured relative to the time of detonation] are needed. Assessments conducted either for emergency decision making or for dose reconstruction for epidemiological studies will generally assume these quantities are homogeneous and unvarying within small areas around or near the measurement site, though the exact assumptions are at the discretion of the assessor. More sophisticated assessments can account for the variability in the exposure rate or deposition density over a defined area around the measurement site by treating the measured exposure rate as the central estimate of an estimated statistical distribution of exposure rates over the area of interest. It is also important to understand that a population living in the area around the measurement site may receive an internal dose from ingestion of radionuclides that were deposited at a distant site where the material consumed was originally produced. This is true when the distribution of food products is accounted for in the model calculations. Examples are NCI (1997) and Bouville et al. (2020).

Some parameters used in a dose assessment may differ significantly within small areas, generally as a consequence of varying human attributes reflecting different age and ethnic groups. For example, food types consumed, which influences the choices of modeling coefficients, can vary within small areas, though reliable decisions about parameter values require detailed knowledge of the potentially exposed populations.

The exposure assessment models presented in the companion papers are generally applicable to US and most western populations in temperate climates; however, model parameter values may need to be tailored to account for a range of environments, including desert and subtropical microclimates. For example, values of biomass and weathering rates can easily be modified to account for different exposure conditions and environment types (Beck et al. 2022; Bouville et al. 2022; Thiessen et al. 2022). While many values of parameters are presented in the companion papers, the specific choices for parameter values are normally left to the assessor and should be based on his or her knowledge about the exposure scenario, the population, and the environment type, including regional variations.

Exposure scenarios and exposure pathways

Exposure scenarios are largely defined by the age distributions, lifestyles, ethnicities, activities, and occupations of the exposed population, all of which help define the predominant exposure pathways. The assumed exposure scenarios should play an important role in choice of dose models and model-parameter values for the assessment.

The external dose pathway for exposure is clearly critical to consider in fallout dose assessments and, in fact, is the primary assessment made in most emergency-response strategies. In cases where there is little potential for ingestion of radioactive debris, external whole-body dose may be the most important calculation endpoint. (In general, external doses to the whole-body or to any organ are nearly equal, given that most nuclear weapons fallout has a mean gamma-ray radiation energy of several hundred keV.)

The internal dose pathway pertains to the exposure of organs and tissues from radioactive particles, gases, or debris entering the body by ingestion or via inhalation. Ingestion usually considers foods that can be grown at home and/or are commercially obtained and measured for radioactive contamination and whose intake rates can be determined over the age and ethnic distribution of the population. The foods most often included in assessment calculations are dairy products, greens and leafy vegetables, and meat, but sometimes also bread and eggs. Consumption of other types of foods, including fruit that can be peeled, nuts, seafood and aquatic organisms, honey, tea and coffee, as well as smoking and chewing tobacco, are generally not considered but can be at the discretion of the assessor. Other minor routes of internal exposure are also possible, e.g., inhalation by resuspension, skin absorption, consumption of drinking water (particularly for assessments based on typical contemporary municipal water supplies), soil ingestion (both intentional and inadvertent), wound penetration, and intakes from uncommon professions and activities, e.g., those that use earthen materials in pottery making. Based on our combined experience, these pathways are estimated to be usually too minor and often too variable among individuals to be included in a population dose assessment following a fallout event.

Inhalation dose is often of great concern. Historically, it has been a difficult pathway to model as it is known to depend on, among other things, the distribution of particle sizes at the locations of interest. As a rule, the contribution to the internal dose from inhalation during the passage of the radioactive cloud is much less than that from ingestion; however, in specific cases when ingestion of contaminated foods is prevented, inhalation might be the dominant internal dose pathway.

Internal doses are typically more difficult to assess than external doses, and the calculation models have many more parameters and possible variations than do external dose models. The relative importance of the external and internal exposure pathways will vary depending on the exposure scenario as well as the major purpose of the dose assessment. For example, assessments of the red bone-marrow (RBM) doses to “atomic veterans” exposed to fallout determined that internal doses were negligible and could be neglected compared to external exposure (Beck et al. 2017; Simon et al. 2019).

In determining the importance of each pathway, some consideration may be required about the ethnic groups exposed and the lifestyles of each, as well as the age distributions of each. Limited experience has been gained over the years in estimations of dose to ethnic groups other than typical North American citizens. Some examples are residents of the Marshall Islands, French Polynesia, and indigenous populations in the western US, Australia, and elsewhere. Special considerations may need to be made to account for differences in ethnic groups, though it is not always obvious prior to making dose calculations, whether the parameter values chosen for a specific ethnic group will result in significantly different doses than more general assumptions might make. It is possible that compensating differences in parameter values chosen can result in only small differences in estimated doses to different ethnic groups. The attributes of exposure that might deviate most from common assumptions are elements of the diet, housing type, and certain habits or customs, e.g., time spent indoors and outdoors and, possibly, soil ingestion.

STEPS IN FALLOUT DOSE ASSESSMENT: SUMMARY OF COMPANION PAPERS

The companion papers (Beck et al. 2022; Bouville et al. 2022; Anspaugh et al. 2022; Melo et al. 2022; Thiessen et al. 2022) present major steps and data necessary for the assessment of fallout doses. In this section, we summarize the important and primary elements described in detail in the companion publications. Along with the descriptions of the basic methodology, several of the companion papers also include appendices with extensive documentation that support the models and discuss uncertainties in greater detail. The three papers that present the models used for estimating doses provide specific examples of how to use the models.

Estimating ground-deposition density

The most important first step to determining internal doses that might be received either by inhalation or by food-chain transport is the estimation of ground deposition density (Bq m−2) of each radionuclide. Beck et al. (2022) summarize the implementation of the method for estimating ground-deposition density in 11 steps. The paper also discusses uncertainties of each of the major steps in order to allow an uncertainty analysis (error propagation) for internal doses from deposited fallout (Appendix C of Beck et al. 2022).

The estimation of ground-deposition density as presented in detail by Beck et al. (2022) uses measured post-detonation exposure rates, accompanied by reliable estimates of the fallout time-of-arrival or TOA (measured relative to the detonation time) at each site of interest. The method for fallout deposition density as a function of TOA is based on a joint US-Russian formulation that accounts for radionuclide fractionation by a semi-empirical formulation. Fractionation is the phenomenon that results in modification of the relative activity of each of the fission and activation products in fallout relative to the amounts produced in the explosion. The proposed methodology differs from that used for most of the earlier fallout dose assessments in that it accounts for the impact of fractionation on estimated doses (i.e., the effects due to varying particle sizes and radionuclide composition as a function of distance and TOA), on estimated deposition density, and on the fraction of activity deposited and retained on vegetation.

For calculations of contamination of plants involved in food-chain transport, it is important to estimate the fraction of the deposited radioactivity intercepted and retained on plant-leaf surfaces. A key assumption in the model is that only particles <50 μm in diameter are intercepted and initially retained on vegetation, thus making them available for food-chain transport. For this purpose, Beck et al. (2022) define the reduced time max where t is the estimated time for all particles >50 μm to be deposited. Beck et al. (2022) also estimate the fraction of the total deposited activity on small particles (<50 μm) (1) as a function of distance and (2) with respect to the site’s location relative to the central line of the downwind trajectory.

Furthermore, the model accounts for changes in fractionation as a function of TOA so that the doses at each site of interest reflect the nuclide composition of the fallout at that site. Close-in to a detonation site, the exposure rates can vary significantly over small distances as can the particle sizes and activity of the deposited particles. Earlier studies that failed to properly account for fractionation tended to overestimate the internal doses from fallout at sites close to the detonation site where the exposure rates were the highest.

Estimating external dose

As discussed by Bouville et al. (2022), the main source of external irradiation is that from gamma rays emitted from radionuclides deposited on the ground. However, the proposed methodology also accounts for the normally small external irradiation received by an individual from the descending particulate cloud. In that case, the in-cloud increment is accounted for by how the calculation of the dose is structured. The dependence of dose when indoors vs. outdoors is based on the assignment of location factors that vary due to differences in construction design and construction materials of residences, schools, and places of employment (Appendix E of Bouville et al.) as well as occupancy factors. Both are accounted for with multiplicative factors. A key difference in the methodology compared to most of the earlier studies (other than the ORERP study discussed earlier) is the use of a greatly improved model for the variation in exposure rates as a function of time. As illustrated in Bouville et al. (2022), the previous use of the t−1.2 approximation under some circumstances led to inaccurate estimates of integrated exposure. As discussed in Bouville et al. (2022), weathering of fallout (penetration of the activity to greater depth in the soil) can also affect the estimate of integrated exposure. Most early studies assumed the fallout was deposited on the surface and remained there indefinitely.

An implementation of the external dose models is provided by Bouville et al. (2022) in two forms that differ in the level of detail and precision offered: (1) an initial form for rough dose estimation soon after a nuclear detonation, and (2) an improved form for a later, more accurate, dose assessment. The later, more accurate dose assessment would presumably include the analysis of post-detonation measurements of radiation exposure and fallout deposition and information on the lifestyle of the exposed population. The choice of the initial or improved model will likely be related to the urgency of completing the assessment as well as to the level of detail and realism needed.

In these assessment strategies, the integration period can be specified, allowing the computation to estimate the external dose received over a specific interval, e.g., from time of deposition to time of evacuation or, as an alternative, over a lifetime. External dose calculations for fallout in a number of publications show that more than 90% of the lifetime external dose is received in the first year following deposition. For that reason, Bouville et al. (2022) emphasize the dose received in the first year after the detonation but also include the 70-y integral dose for the purpose of estimating lifetime exposure.

Secondary sources of external irradiation from deposited fallout are discussed in appendices in Bouville et al. (2022), and include: (1) the dose due to descending fallout during the passage of the radioactive cloud (Appendix A of Bouville et al. 2022), and (2) the dose from beta rays (Appendix B of Bouville et al. 2022). The dose from beta rays can be substantial for the skin and the lens of the eye but is almost always unimportant for all other radiosensitive organs and tissues of the body. Moreover, neither the skin nor the eyes are considered at high risk for cancer following radiation exposures (NAS 2006) or, specifically, exposure to fallout (Land et al. 2010). Estimation of dose to the skin may be a public concern when beta emitters come into prolonged contact with the skin, though an assessment of that scenario is particularly problematic to assess on a population basis. Reasons for such difficulties include that dose modeling for each person would require unique parameter values to specify the length of time of exposure, the fraction of the skin exposed, and possibly the location of the exposed skin on the body. Health-risk assessments and/or epidemiological studies involving exposure to fallout generally do not include skin dose (especially location-specific skin dose) or make projections of skin cancer because of the dosimetric problems noted but also because the particular parts of the skin irradiated would need to be identified.

Similar to Beck et al. (2022) who discuss uncertainties of individual model parameters for deposition, Bouville et al. (2022) discuss uncertainties of model parameters for external dose.

Estimating internal dose

Regardless of the specific dietary habits and food-production strategies, there is potential for significant intakes of fallout debris by ingestion and/or inhalation, particularly at early times after detonation. Thus, assessment strategies for estimating internal dose are needed, and the companion paper by Anspaugh et al. (2022) details our proposed methods for internal dose assessment.

Various approaches for estimation of internal dose from ingestion pathways have been used in previous assessments of nuclear weapons testing fallout. For example, for estimation of doses from Nevada Test Site fallout, a typical North American food-chain formed the basis for estimates of time-integrated intake per unit deposition for each radionuclide, fallout event, location, and age group (Whicker et al. 1996). The intake per unit deposition was then used together with the exposure rate (12 h after deposition) and the deposition of each radionuclide (per unit exposure rate) to estimate the total intake of each radionuclide at a given location. This method is one of the two main approaches described by Anspaugh et al. (2022).

Internal dose assessments are inherently more complex to make than external dose assessments because (1) the data on the decay characteristics of each radionuclide, including half-life, energies, and types of emissions, are needed; (2) the chemical attributes of each radionuclide and knowledge of the physical attributes, e.g., particle-size distributions, are needed, as well as (3) relevant data to estimate intakes, e.g., activity level for inhalation or dietary descriptions for contaminated foods. Data assumptions are available in Thiessen et al. (2022) and Ibrahim et al. (2010), the latter publication deriving assumptions from a wide range of literature, including the bioassay observations of Harris et al. (2010).

Two approaches for internal dose calculation are presented by Anspaugh et al. (2022): (1) a simple-to-implement method using results of calculations from methods that were used in the US Department of Energy’s (DOE) Off-Site Radiation Exposure Review Project (ORERP) (Church et al. 1990) and later used in a summary form for the US Department of Health and Human Services’ report to Congress on the feasibility of a study of the health consequences of nuclear weapons tests (Anspaugh 2005a,b, DHHS 2005); and (2) a more complex and flexible method based on consideration of individual food pathways using analytical solutions of time-dependent exposure assessment models.

The first method is a summarized version of the ORERP calculations and is presented in terms of Bq intake per Bq m−2 on the ground. This method is by far the simpler of the two methods to implement. Consequently, it is presented as an approach when a simple strategy or one that can be more rapidly applied, is appropriate. The ORERP method was developed with attention to distances of several hundred km from the site of detonation. For moderately close-in distances downwind, Anspaugh et al. (2022) used the traditional approach using a suite of sub-models for individual pathways and food types, with corrections for fractionation of radionuclides as specified by Beck et al. (2022).

In addition to the described methods that allow two different levels of complexity for internal dose assessment, two additional offerings are presented by Anspaugh et al., those being (1) for inhalation dose and (2) for exposure of the thyroid of the fetus. The former issue, inhalation dose, has often been ignored because of the inherent modeling difficulties and because many preliminary assessments have shown inhalation dose to be minor compared to ingestion dose. However, in cases where food interdiction has been employed, the role of inhalation dose is much greater on a relative scale. The latter issue, i.e., exposure of the fetus, is also sometimes ignored, presumably because the number of fetal exposures is normally small in a population compared to the other age groups.

Determination of inhalation dose is particularly problematic because of the difficulties in accounting for the presence of particles during the deposition process, many of which may be too large for deep penetration into the lung. In addition, fallout particle-size distributions change continually with increasing TOA and distance from the detonation site and from the center of the trajectory axis (Beck et al. 2022). It is, therefore, necessary to use dose coefficients that reliably account for the relevant particle sizes and the solubility of the radionuclides (Ibrahim et al. 2010). In addition, the historical data upon which to attempt to model in-cloud inhalation vary widely. The calculation strategy for inhalation dose follows some previous attempts at dose estimation (for example, Simon et al. 1990) in which the fraction of the fallout particles in a respirable size range (generally <20 μm diameter) is estimated. The calculations here are improved over most previous modeling attempts by the inclusion of dose coefficients for a range of particle sizes: 1 μm, 5 μm, 10 μm, and 20 μm with guidance provided for assessing dose from inhalation of particles 20–100 μm and >100 μm. Although some of the larger particles can be inhaled, they are not respirable and, in effect, lead to an ingestion dose that was usually not considered in previous fallout models. Here, dose from resuspension of contaminated soil is not explicitly presented because it has been shown in previous analyses to be an extremely small contributor to dose (Maxwell and Anspaugh 2011; Simon et al. 2020).

The issue of dose to the fetal thyroid from radioiodines has been considered previously by various organizations and authors, and the presentation by Anspaugh et al. (2022) adopts those efforts except for the critical radioiodine nuclides. As discussed in Anspaugh et al. (2022), conflicting assumptions on the biokinetics of radioiodine necessitated a re-examination of the dose coefficients. Anspaugh et al. (2022) refer to Melo et al. (2022) for a more complete discussion and derivation of those factors.

Similar to the other companion papers, Anspaugh et al. (2022) discuss uncertainties of model parameters for internal dose and also demonstrate simple error propagation for a lognormal multiplicative model.

Dose coefficients (DCs) for fallout dose assessments

A unique contribution to the suite of methods presented here for fallout dose assessments is the presentation of theory and calculated values of dose coefficients (DCs) tailored to the situation of radioactive fallout where particles of varying solubility would likely be encountered.

The 34 radionuclides discussed are our recommended priority list for fallout dose assessments (Table 1). These radionuclides are those that normally account for a preponderance of the organ doses that might be received by ingestion by persons of all ages (including via breast feeding for infants) and by inhalation following exposure to radioactive fallout. The presented dose coefficients for ingestion account for age and include modifications for variations in solubility with distance as discussed in a previous publication (Ibrahim et al. 2010), and those for inhalation similarly account for age, solubility, and particle sizes that would be relevant at various distances of exposure (Anspaugh et al. 2022). The proposed modifications peculiar to radioactive fallout account for systematic changes in solubility and particle sizes with distance from the site of detonation, termed here as the region of “local fallout” and the region “beyond local fallout.” A particular characterization of local fallout is that volatile nuclides such as radioiodines tend to be distributed primarily on the surface of fallout particles; they are more soluble than refractory nuclides that tend to be incorporated into the volume of the larger particles and thus are less soluble than generally assumed. Comparisons of the presented dose coefficients are made with values published by the International Commission on Radiological Protection (ICRP).

Melo et al. (2022) present DCs for 16 defined organs, all presented as Gy Bq−1 for absorbed dose and as Sv Bq−1 for effective dose. Values of f1 are specified for ingestion for each age group, and the choices of the values are discussed in the text. DCs are also presented for inhalation dose for particles of AMAD of 1, 5, 10, and 20 μm.

There are actually two sets of values for dose coefficients for inhalation and ingestion used in Anspaugh et al. (2022). The first is the traditional set of values that have been developed over many years by the ICRP and can conveniently be downloaded from the ICRP (2011) as “CD1.” Representative values of the ICRP dose coefficients are shown in Table 3 in Melo et al. (2022). The second set of values is that reported in the companion paper by Melo et al. (2022); this second set has been developed specially for the purpose of considering that particles formed in a nuclear explosion can be large, with radionuclides, particularly the more refractory ones, distributed throughout the volume, resulting in the radionuclides having reduced solubility. The two sets of dose coefficients are hereinafter referred to as the “ICRP” set and the “Reduced” set.

The guiding principles for the traditional method explored here (Anspaugh et al. 2022) are that the ICRP values should be used in all cases where the reduced TOA, called t, is more than 1.0. See the previous discussion on estimating ground deposition for the definition of t and Beck et al. (2022). In addition, the ICRP values should be used within the region where t = <1.0, if biological processing of radionuclides has occurred; for example, if the radionuclide reaches humans through the consumption of milk or meat or from the consumption of fruits and vegetables where the radionuclides have moved via soil-root absorption. The Reduced set of dose coefficients should be used within the region where t = <1.0 both for inhalation and for the consumption of any food that has been contaminated directly by fallout, for example the direct contamination of leafy vegetables. Historically, the ORERP used the ICRP values where available or values derived independently by Ng as presented in Kirchner et al. (1996). For the example ORERP calculations presented in Anspaugh et al. (2022), the ICRP dose coefficients were used because it is not possible to separate individual pathways.

As in the other companion papers, a discussion of uncertainties of the DCs is presented. Uncertainties for DCs are difficult to quantify, in part because of the complexities of estimating the uncertainty of complex biokinetic model calculations for which data are often sparse. Support for estimates of uncertainties for the DCs was derived from NCRP (2009).

Parameter values for internal dosimetry modeling

The companion paper by Thiessen et al. (2022) presents a lengthy discussion and a compendium of numerical values of the common parameters needed for internal dose assessments for exposures to radioactive fallout. Both point values and probability distributions are proposed. Parameters include those needed to assess the interception and initial retention of radionuclides by vegetation, translocation of deposited radionuclides to edible plant parts, root uptake by plants, transfer of radionuclides from vegetation into milk and meat, transfer of radionuclides into non-agricultural plants and wildlife, and transfer from food and drinking water to mother's milk (human breast milk). Also discussed are the weathering half-life for contamination on plant surfaces, biological half-lives of radionuclides in organisms, food processing (culinary factors), and contamination of drinking water. As appropriate, and as information exists, parameter values or distributions are specific for elements, chemical forms, plant types, or other relevant characteristics.

Uncertainty

The ability to estimate dosimetric uncertainty has improved over the decades as least as much as any single component of the dose-calculation strategy. In particular, the understanding of random and systematic errors and the possibility to conduct Monte Carlo calculations to simulate numerous alternative outcomes have enabled quantitative uncertainty analysis to develop as an accepted (and expected) tool. There are numerous discussions of the Monte Carlo method (for example, Morgan and Henrion 1990) for purposes of uncertainty analysis. Examples of application of the Monte Carlo method in fallout dose assessments include Breshears et al. (1989), Whicker et al. (1990), Simon et al. (1990), Kirchner et al. (1996), Simon et al. (2010), and Simon et al. (2020), as well as a real-time application in the National Cancer Institute’s fallout dose calculator (https://radiationcalculators.cancer.gov/fallout/).

Uncertainty of estimated fallout doses needs to be interpreted with care as uncertainty may refer to one of several quantities, e.g., the estimated organ dose for an individual or the average dose for persons of certain attributes. The use of uncertainty estimates in communication with the public can be a particularly difficult challenge as understanding the meaning of confidence or credibility intervals requires a minimum level of numerical literacy.

Typical levels of uncertainty in retrospective fallout calculations can be summarized either as confidence limits or as multiplicative factors on the mean, median, or best estimate of dose that produce lower and upper credibility levels. For example, the uncertainty may be expressed as a geometric standard deviation (GSD) of the distribution of possible values. In this case, the GSD expresses the width of the lognormal distribution relative to the geometric mean (GM) with the range representing alternative possible values of what is believed to be the true dose. The 95% range of possibilities extends roughly from GM/GSD2 at the low end to GM x GSD2 at the upper end. For external dose, GSDs of 1.3 to 2.0 are typical, while significantly greater values of 2.5 to 5.0 are typical for internal dose. This finding has been found in a variety of studies and publications related to fallout exposures (Whicker et al. 1990; Kerber et al. 1993; NCI 1997; Simon et al. 2006a, 2020). See eqn (4) of Anspaugh et al. (2022) for an example of analytical propagation of GSD values.

In each of the companion papers (Beck et al. 2022; Bouville et al. 2022; Anspaugh et al. 2022; Melo et al. 2022; Thiessen et al. 2022), representative uncertainties are presented for the model parameter values allowing dose assessors of future fallout events to propagate uncertainties in model calculations. Some of the numerical representations of uncertainties in Thiessen et al. (2022) are more complex than in other companion papers. These distributions represent state-of-knowledge uncertainty distributions and are often mathematically more complex than simple lognormal distributions, which are typically used to characterize the distributions of measurements. In addition to lognormal, the types of distributions include log-uniform, triangular, and log-triangular. In these cases, useful quantiles, e.g., mean and modes, are provided so that the distributions can be easily and properly characterized in Monte Carlo (MC) error propagation codes, as simple error propagation for a multiplicative model with lognormal distributions is not viable. Typically, MC codes can sample from the appropriate distribution type as defined by the specified quantiles.

Verification/validation/confirmation/corroboration

As discussed in NCI (1997, Chapter 10), model verification demonstrates the extent to which results calculated by computerized dosimetry programs agree with results hand-calculated using the equations and the parameter values, while model corroboration, sometimes termed validation, refers to a comparison of available measurements of concentrations in man, in animals, and in the environment with the results obtained from the algorithms and models. Verification of the computer program developed for an assessment is often not given the same importance as choice of the algorithm and parameter values, though it is vitally important to the quality-control and quality-assurance process.

For purposes of confirmation, corroboration, or validation, the computed value is compared against available measurements for confirming that the choice of algorithm and parameter values produces reasonable and unbiased results. The process described by these terms, under limited conditions, can lead to the conclusion of validity of the dose algorithm and selected parameter values, but only for the range of conditions represented by the samples chosen for comparison.

For fallout dosimetry, few exercises in confirmation, corroboration, and/or validation have been completed, and these are primarily limited to external dose where measurements of radiation-induced signals in fallout irradiated materials are compared against dose calculations or to internal dose to a limited number of tissues.

Primarily, the measurement techniques of EPR (electron paramagnetic resonance) and OSL (optically stimulated luminescence) have been used to make measurements in a variety of materials including bricks and tooth enamel from man and animals, to confirm calculations of external dose. A few examples of confirmation exercises using EPR after fallout irradiation are Ivannikov et al. (2006) and Sholom et al. (2007).

Other means of model verification and corroboration are comparisons of concentration of radionuclides measured in tissue, blood, urine, or feces with model calculations. As described in NCI (1997), modeled values of 131I concentrations in urine were compared with measurements of 131I in urine collected from military men in 1955. Similarly, 131I was measured both in animal thyroids and in cows’ milk from many different locations. Such measurements, even when viewed as satisfactory, can only be used to confirm the part of the dose calculation to which they pertain. The value of confirmation measurements is underscored by the fact that internal dose cannot be observed directly.

In more recent studies, measurements of chromosome aberration rates have been reported in military men exposed to radiation from nuclear testing as a means of corroborating dose-calculation strategies (McKenna et al. 2019; Simon et al. 2019). Studies of both EPR and chromosome aberrations have been used to validate external and internal doses to Russian persons living downstream of the Mayak Production Association (Degteva et al. 2015). Measurement of chromosome aberration rates could, in theory, reflect both external exposure and internal exposure, at least to the degree that circulating blood lymphocytes are exposed to internal radiation. Other types of biodosimetry are under development, and their continual improvements will likely open additional opportunities in the future for corroborating doses (Sproull and Camphausen 2016).

In addition, a type of model confirmation can be achieved by comparing calculations with environmental measurements, either instantaneous measurements (e.g., the concentration in biota at a specified time) or integral measurements (e.g., readings of film badges or other types of integrating dosimeters exposed to fallout radiation fields). For example, Simon et al. (2020) compared model-based external exposure calculations with historical exposure data derived from film-badges deployed in various environmental settings prior to the 1945 Trinity nuclear test.

Results of comparisons of model-based calculations and measurements reported to date have suggested reasonably good agreement, at least, on average. It should be recognized though that all exercises in corroboration, confirmation, or validation of models by single samples is particularly fraught with uncertainty and effects of random variations. Agreement needs to be judged using the mathematical average from several appropriate samples. Beyond the statistical limitations of single samples, comparisons of model predictions with measurements of any type have further limitations due to uncertainties of the measurements themselves. For example, measurements of exposure rate and concentrations have well known instrument and laboratory/technique-associated uncertainties. Applicability of the collected samples to the assessment endpoint also introduces uncertainty. Biological assays, e.g., chromosome aberration rates, have considerable uncertainties related to inter-individual variation, individual and population baseline rates, and laboratory technique. Corroboration and confirmation exercises clearly have value but need to be interpreted cautiously in light of their various uncertainties.

Because the sophistication of external dose models has increased only modestly over the years of fallout dosimetry, it is difficult to definitively state that a high degree of improvement has been achieved with recent modeling efforts. However, as newer models (Beck et al. 2022) are based both on theoretical concepts (e.g., fractionation) and statistical fitting of actual data sets (e.g., measurements of fallout decay-rates, interception fractions), the presented models provide a degree of confidence as good as or superior to any past calculation strategies. As noted, internal dose calculations are more difficult to confirm but are similarly based at their most fundamental level on estimates of radionuclide concentrations in the environment derived from the same relationships between exposure rate and radionuclide-decay patterns as used in the external dose calculations.

OTHER ISSUES

Countermeasures

In contrast to literature that discusses medical radiation countermeasures (i.e., chemical agents for reducing radiation injury and risk of radiation-related disease, typically administered after the exposure is received; for example, Obrador et al. 2020), here we refer to countermeasures as physical procedures that prevent or reduce possible exposure (NRC 2004). Some possible countermeasure procedures are sheltering, evacuation, food interdiction, environmental remediation, prophylactic administration of stable and competing nuclides, and chelation therapy. While the models discussed in the companion papers assume no countermeasures or remediation that could prevent or reduce possible exposure, here we discuss how our models might be adapted to account for such practices.

Sheltering, sometimes referred to as sheltering-in-place, simply refers to the purposeful act of staying indoors to achieve the greatest possible shielding from external radiation originating from descending and deposited fallout. The estimation of external dose can be easily modified to account for sheltering-in-place by modification of the time spent indoors and the location factors. See Appendix E of Bouville et al. (2022) for various values that could be customized for other conditions as needed.

Accounting for evacuation in the external dose models would necessitate, for example, truncation of the integration period over which dose is accumulated (eqn 1 of Bouville et al. 2022). Some consideration needs to be given to the possibility that only a reduced exposure-rate, rather than 0, might be received at the location to which evacuation is assumed.

The concept of food interdiction, i.e., removing contaminated foodstuffs from public availability, is preferably implemented immediately after contamination is realized but before consumption begins. In that case, modifications of internal dose models would be a simple step to zero-out the rate of intake of the contaminated foods. Here, similar to the case for external dose, consideration needs to be given to the possible (but probably lower) contamination of foods that replace those interdicted. In contrast, if food interdiction is assumed to occur sometime after public consumption has begun, modifications to assessment models need to be specific for the date at which the food availability is interrupted.

Environmental remediation is also possible during the exposure period as an attempt to reduce ongoing exposure from ground contamination (Cox et al. 2005). For example, urban surfaces, e.g., streets and houses, can be washed to remove surficial contamination. For modeling the effects of such activities on external dose received, the changes to the dose model would be similar to the case of evacuating to a different location with a presumably reduced exposure rate. The initial date of the remediation (relative to the date of deposition) would be needed, coupled with a post-remediation exposure rate for the remainder of the exposure period. External dose models can be modified to account for the effectiveness of various countermeasures (Thiessen et al. 2009).

Environmental remediation of crop and pastureland can also be implemented by physical removal of contaminated soil, deep ploughing of contaminated soil so as to reduce the average surface concentration, and chemical soil amendments to prevent uptake of radionuclides into crops. Activities to reduce possible internal doses by reducing uptake of radioactivity into plants is generally a long-term initiative and would not likely be part of a short-term emergency response. For that reason, it is unlikely that internal dose estimates for the months soon after a detonation would need to reflect environmental remediation, though, if necessary, the internal dose model can be adapted to reflect the periods of time over which uptake occurs and the degree of uptake, both of which could be amended to reflect remediation.

Prevention of internal dose to the thyroid gland from internally deposited radioiodines is possible through the prophylactic administration of stable iodine (NRC 2004). In the case of administered stable iodine, the uptake of radioactive iodine is reduced or prevented. As discussed in the literature, the timing of the administration of stable iodine, usually in the form of potassium iodide, is critical to its success for preventing internal dose. Modification of the internal dose calculations to account for prophylactic administration of stable iodine would be to significantly reduce the uptake of iodine by the thyroid gland. It has been reported that 95% of the thyroid dose can be averted by proper administration of stable iodine (Verger et al. 2001).

For a limited number of other radionuclides that might result in internal exposure from fallout, some internal dose can be averted by chelation therapy, i.e., administration of compounds to increase the rate of excretion of the radionuclide from the body and, thereby, reduce the absorbed radiation. REMM and NCRP (2008) present a summary of radionuclides for which chelation therapy can be attempted and the treatment compounds to be used. The modification of dose-assessment calculations for chelation therapy, however, would involve modification of the biokinetic models used for calculations of the dose coefficients (Melo et al. 2022), a subject beyond this discussion. Melo et al. (2022) show that DCs cannot be easily adjusted based on the f factor alone.

Stakeholder-related issues

Groups or individuals who are either the subject of dose and/or risk assessments, or groups of persons who have any vested interest in the outcome of assessments, might be considered as stakeholders. It is useful to recognize that the numerical results of dose and risk assessments can be highly controversial among stakeholders, and engagement of those persons and obtaining their support, sometimes as an outcome of participation in the assessment, can be useful to ensure wide acceptance of the assessment results. The methods provided here, while technical, are intended to be made available through open publication for public and professional review and acceptance. The transparency afforded by publication and open access is intended to increase public acceptance of dose-assessment results, although experience suggests that this presumed advantage is not always realized.

Communication issues

One of the important but often overlooked issues related to assessing doses and health risks from fallout-related exposures is communicating the findings on the dose and risk assessments to the public. The challenges and means of communicating dose and risks from fallout exposures have been discussed by numerous authors (Hoffman et al. 2002; Caffrey 2021). As explained by Hendee (1991), the risks of radiation exposures are best communicated by someone who is knowledgeable, is recognized as a health expert, is trustworthy, and has no conflict of interest but a stake in the community—such as a community physician or, by extension, representatives of a public health protection agency. Such a spokesperson must respond first to perceptions of risks and underlying emotions because use of only facts often fails to counteract fears.

In the US, both the CDC and the NIH have significant experience in communicating with the public about fallout exposures. See for example, the National Cancer Institute’s (NCI) Fallout Dose Calculator for Nevada Test Site fallout (https://radiationcalculators.cancer.gov/fallout/) and other documents prepared by the NCI as part of a communication campaign about the risks of exposure to 131I (https://www.cancer.gov/about-cancer/causes-prevention/risk/radiation/i-131). The NCI documents above were developed from a comprehensive two-volume report (NCI 1997 and https://www.cancer.gov/about-cancer/causes-prevention/risk/radiation/i131-report-and-appendix) on doses to the American people from 131I in Nevada Test Site fallout, prepared in response to Congressional requests for information. In later years, in cooperation with the NCI, the CDC assisted in the development of a technical report that included an assessment of doses from global fallout to US residents (https://www.cdc.gov/nceh/radiation/fallout/default.htm). These various assessments were precursors to fallout exposure studies conducted later for populations in the Marshall Islands, Kazakhstan, and New Mexico.

Preparation for the future

The value of a comprehensive fallout dose-assessment methodology as presented here is primarily for preparation for the future. Part of an effective preparation for future fallout events is to ensure the capability to obtain the needed data for both external and internal dose assessments. Estimation of external dose requires measurements of environmental exposure rates and TOA, though estimates of exposure rate can also be derived from deposition density of individual radionuclides and TOA. The same measurements are similarly needed for the estimation of internal dose. Depending on the size of the area potentially contaminated with fallout, the capability for obtaining measurements may need to be considerable.

As described earlier, estimation of internal dose is more complex than for external dose because of the necessity to quantitatively estimate radionuclide intakes among the population. In general, estimates of radionuclide intakes (or radionuclide-intake rates) from a particular nuclear event may require post-detonation measurements of radionuclide concentrations in foods or at least the capability to predict food concentrations using data on soil contamination. In addition, data from a survey of lifestyle and dietary practices can be useful to estimate dietary intake of radionuclides among the age and ethnic groups resident. Two examples of retrospectively obtaining data on diet and lifestyle are Schwerin et al. (2010) and Potischman et al. (2020) though details on methods of obtaining such data are beyond the scope of this paper and its companion reports.

SUMMARY

This publication and its companion papers provide a well conceptualized, well described, and internally consistent suite of methods for assessing radiation doses to persons or population groups that may have been exposed in the past or may be exposed presently or in the future to radioactive fallout from a small to moderate-yield nuclear fission detonation. For these purposes, we have extensively reviewed literature about the physical and chemical attributes of fallout, mechanisms of fallout deposition and movement in the environment, and coalesced knowledge gained from years of experience in the fields of radiation dosimetry, fallout dose assessment, and dose reconstruction.

Here, it was our intent to provide a practical set of methods that could be employed by dose assessment experts in the context of (1) federal, state or other expert group responses to a nuclear detonation, including considerations of physical countermeasures; (2) health research, e.g., for epidemiology and/or retrospective risk assessment; or (3) documenting exposures and providing data for compensation, historical purposes, etc. The methods are based on the most inclusive set of knowledge, experiences, and research in fallout measurements and fallout dose assessment to date. Some of the assessment models and sub-models have been validated, albeit to a limited degree, but should be considered as viable and suitable for dose-assessment purposes without reference to a specific identity, age, sex, or ethnicity of a particular population, except to the degree that diet and lifestyle for a population need to be considered in the choice of parameter values. While some parameter values might need to be tailored to a specific age or ethnic group of interest, the methods presented are general enough to account for most known fallout exposure conditions and situations.