Appendix
A

Representative Dose Reconstruction Studies

A NUMBER OF DOSE reconstruction studies of environmental releases of radioactive materials have been either completed or undertaken, and there are lessons to be learned from these efforts. The most important dose reconstructions have been associated with nuclear weapons testing (Maralinga, Pacific Test Site, Nevada Test Site, etc.); reactor accidents (Chernobyl, Three Mile Island, Windscale); routine releases from installations of the nuclear fuel cycle, especially during the early years of operation (Fernald, Hanford, Techa River); and careless disposal of industrial or medical radioactive sources (Goiania). The manner in which doses were reconstructed in some of these studies is discussed below.

NEVADA TEST SITE

The first modern dose reconstruction study done in the United States involved the Nevada Test Site (Voillequé and Gesell 1990). At this location, roughly 100 above-ground tests of nuclear weapons were conducted in the 1950s; tests also occurred in the early 1960s. The monitoring of the fallout (as determined by measurements of external gamma-exposure rate) from these tests was extensive within the close-in area, and calculations of the external gamma dose were made and tabulated for many communities (Anspaugh and Church 1986, Anspaugh and others 1990). Late in the 1970s, considerable controversy developed surrounding allegations that leukemias, and subsequently other cancers, had been caused



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 103
Appendix A Representative Dose Reconstruction Studies A NUMBER OF DOSE reconstruction studies of environmental releases of radioactive materials have been either completed or undertaken, and there are lessons to be learned from these efforts. The most important dose reconstructions have been associated with nuclear weapons testing (Maralinga, Pacific Test Site, Nevada Test Site, etc.); reactor accidents (Chernobyl, Three Mile Island, Windscale); routine releases from installations of the nuclear fuel cycle, especially during the early years of operation (Fernald, Hanford, Techa River); and careless disposal of industrial or medical radioactive sources (Goiania). The manner in which doses were reconstructed in some of these studies is discussed below. NEVADA TEST SITE The first modern dose reconstruction study done in the United States involved the Nevada Test Site (Voillequé and Gesell 1990). At this location, roughly 100 above-ground tests of nuclear weapons were conducted in the 1950s; tests also occurred in the early 1960s. The monitoring of the fallout (as determined by measurements of external gamma-exposure rate) from these tests was extensive within the close-in area, and calculations of the external gamma dose were made and tabulated for many communities (Anspaugh and Church 1986, Anspaugh and others 1990). Late in the 1970s, considerable controversy developed surrounding allegations that leukemias, and subsequently other cancers, had been caused

OCR for page 103
by the exposure, and a decision was made to carry out a dose reconstruction study to reevaluate the estimates of external exposure and dose and to attempt for the first time a complete assessment of exposure and dose from the ingestion and inhalation of radionuclides (Church and others 1990). The decision to perform this study was made in advance of any convincing demonstration that a dose reconstruction could actually be performed. An early debate ensued on the possible options of performing such a study with one option being a "model only" study using existing fallout and atmospheric transport models. The option eventually selected was to make maximum possible use of the existing historic data, which consisted of more than 100,000 measurements of the external gamma-exposure rate (Grossman and Thompson 1990) and more than 10,000 measurements of ground-level concentrations of radionuclides in air (Cederwall and others 1990). Beck and Krey (1983) had demonstrated that doses could be reasonably well reconstructed on the basis of contemporary measurement of 137Cs and 239+240Pu deposition densities in undisturbed soil, together with the determination of the 240Pu-to-239Pu ratio. The decision was made to enlarge the geographic coverage of the study to include the entire states of Arizona, Nevada, New Mexico, and Utah, as well as several counties of California, Colorado, Idaho, Oregon, and Wyoming (Church and others 1990). Two of the more difficult aspects of the study involved determining the relationship between the measured values of the external gamma-exposure rate and the deposition densities of the more than 100 radionuclides released by the tests and to then calculate the intake of different radionuclides by the exposed individuals and population. The first problem was solved with Beck's (1980) calculations of the external gamma-exposure rate per unit deposition density of a given radionuclide and by the development of normalized source terms (per mR hr-1 of exposure rate 12 hr after detonation) (Hicks 1982, 1990). (The mR represents a unit of exposure called milliroentgens.) The second problem was solved by the development of a dynamic, seasonally dependent food chain model with specific consideration of uncertainty (Whicker and Kirchner 1987, Whicker and others 1990). Example results (Henderson and Smale 1990, Ng and others 1990) of this study have been available for some time and intermediate results (Beck and Anspaugh 1990) have been used in epidemiologic studies (Kerber and others 1993, Stevens and others 1990) and to assess the thyroid doses of 131I received by the populations of the contiguous United States (Bouville and others 1990, Wachholz 1990). One important aspect of the Nevada Test Site study was the extensive use of contemporary measurements of long-lived materials in soil to confirm and extend the

OCR for page 103
original measurements. Furthermore, a convincing confirmation of the essential correctness of the external dose calculations was provided by contemporary solid-state dosimetry measurements of the thermoluminescence signals obtained from quartz contained in bricks in homes in the downwind areas (Haskell and others 1994). CHERNOBYL The April 1986 accident at Unit 4 of the Chernobyl nuclear power plant, in the Ukraine, about 30 km south of the border with Belarus, was the most severe in the history of the nuclear industry. The accident caused the death of 31 power plant employees and firemen from acute radiation exposures and burns. It brought about the early evacuation of 135,000 people and resulted in the contamination of vast regions of Belarus, Russia, and the Ukraine, as well as, to a lesser extent, many countries of northern Europe. About 330 petabecquerels1 (PBq) (8.91 x 106 Ci) of 131I, 35 PBq (9.45 x 105 Ci) of 134Cs, and 70 PBq (1.89 x 106 Ci) of 137Cs were released into the atmosphere over a period of 10 d. During that time, the winds blew in many directions, so that the radioactive materials released were transported over different regions. Most of the materials in the radioactive cloud were deposited on the ground, largely through precipitation, and this resulted in the contamination of milk and other foodstuffs. The short-lived 131I caused high thyroid exposures, especially among children, in the first few weeks after the accident. Radioiodine concentrations were measured in about 250,000 persons in Belarus, 150,000 persons in Ukraine, and 30,000 persons in Russia. Preliminary estimates indicate that the thyroid doses received by children range up to 30 Gy (3,000 rad) or more. A preliminary thyroid dose distribution among children in the most heavily contaminated districts of Belarus who were under the age of 7 years at the time of the accident is provided in Table A-1. The estimated arithmetic mean thyroid doses in subgroups of the same populations vary from 0.21 to 1.06 Gy or 21–106 rad (Table A-2). Efforts are being made to reconstruct the individual thyroid doses received by the most exposed populations of Belarus (Gavrilin and others 1992), Russia (Zvonov and Balonov 1993), and Ukraine (Likhtarev and others 1993), on the basis of the thyroid measurements and of personal interviews on dietary and lifestyle habits. Serious technical difficulties have been encountered when the thyroid measurements were carried out by inexperienced individuals using instrumentation that was not specifically designed for this type of measurement. In addition, the longer lived 134Cs, and, more important, 137Cs, deliver doses to the entire body and will be present in the environment for de-

OCR for page 103
TABLE A–1 Preliminary Thyroid Dose Distribution Among Children in the Most Heavily Contaminated Districts of Belarus Who Were Less Than 7 Years Old at the Time of the Accident (Gavrilin and others 1992)   Nine Districts of Gomel Region, 32,420 Children Five Districts of Mogilev Region, 14,240 Children Fourteen Districts of Gomel and Mogilev 46,660 Children Thyroid Dose Range (Gy) Number % Number % Number % 0.00–0.03 15,128 46.66 9,637 67.68 24 ,675 53.08 0.03–0.75 8,951 27.61 2,975 20.88 11, 926 23.55 0.75–2.00 4,924 15.18 1,345 9.45 6,26 9 13.44 2.00–5.00 2,428 7.49 251 1.76 2,679 5.74 5.00–10.0 693 2.14 28 0.20 721 1.55 10.0–20.0 274 0.85 4 0.03 278 0.60 20.0–30.0 20 0.06     20 0.04 30.0–40.0 2 0.01     2 80.01 TABLE A–2 Estimated Arithmetic Mean Thyroid Doses to Children Under Age 7 in the Most Contaminated Districts of Belarus (Gavrilin and others 1992) Oblast (Province) Number of Districts Population Type Population Size Mean Thyroid Dose (Gy) Gomel 9 Rural 23,900 1.06     Urban 8,600 0.44 Mogilev 5 Rural 9,300 0.44     Urban 4,900 0.21 Gomel and Mogilev 14 Rural 33,200 0.88   Urban 13,500 0.36 cades to come. The resulting doses, which are less important than those delivered by 131I, also are being reconstructed for the populations living in contaminated areas. Doses from external irradiation are best determined from transport calculations based on measured cesium concentrations in soil and on thermoluminescent dosimeter (TLD) measurements, whereas the doses from internal irradiation are estimated from wholebody counting data or from 137Cs concentrations in milk associated with information on milk consumption rates. Attention is also being paid to 90Sr. The migration of this radionuclide from the soil through the terrestrial food chain could result in later years in doses from internal irradiation similar to those from 137Cs.

OCR for page 103
Finally, the resuspension into ground-level air of 241Am, 239Pu, and other long-lived transuranics is likely to lead to exposure to future generations, when 137Cs and 90Sr have decayed to negligible values. THREE MILE ISLAND On March 28, 1979, the No. 2 Unit of the Three Mile Island reactor in Middletown, Pennsylvania, had a loss-of-coolant accident that led to a partial core melt-down and the subsequent release of approximately 400 PBq (10 megacuries2 [MCi] of noble gases. The exact magnitude of this release is not known because the effluent monitors were insufficient to monitor the amount of the release. Despite the large release of noble gases, the amount of radioiodine released was estimated to be only 0.4–1 terabecquerel3 (TBq) (10–30 Ci). The iodine release was estimated from measurements performed on charcoal absorbers in the effluent line. Because the accident occurred in the early spring, cows were not yet in pastures and the amount of radioiodine transferred to milk was low, leading to maximum concentrations in milk of about 1.5 Bq L-1 (40 picocuries4 L-1). Several approaches were considered for estimating individual and population doses from the noble gases. The most useful was to rely on actual measurements of doses from external irradiation made by TLDs surrounding the site. The maximum individual dose was estimated to be less than 1 millisievert (mSv; 100 mrem) and the collective dose was 20–35 person-sievert (2,000–3,500 person-rem). The average dose to the 2 million people residing within 50 miles was 0.015 mSv (1.5 mrem). The Three Mile Island case illustrates the importance of using environmental measurement data for dose reconstruction when an accurate source them cannot be determined because of instrumental deficiencies or other causes. FERNALD The Feed Materials Production Center (FMPC) located at Fernald, Ohio, was, between 1951 and 1989, a government-owned, contractor-operated facility for producing uranium metal products used as feed materials in the production of nuclear weapons. Environmental releases of radioactive materials consisted mainly of uranium, thorium, and radon into the atmosphere. An environmental dose reconstruction is being conducted at Fernald. From preliminary information, it seems that the principal pathway is inhalation of uranium. The main difficulty so far encountered in the

OCR for page 103
environmental pathway analysis appears to be that there are not enough local meteorologic data for 1951 to 1986. A recording meteorologic tower has been in regular operation at the site only since 1986. Data for earlier times (from 1951 to 1986) are available for regional airports, in Cincinnati and Dayton, but those sites are more than 50 km from Fernald. The options available to the dose reconstruction team were to use recent Fernald data for the earlier period, with a large degree of uncertainty applied to air concentration estimates, or to use a surrogate historic data set based on regional airport data. After examination of the meteorologic data available, it was concluded that the air dispersion predictions for the dose reconstruction study should be based on hourly wind and stability data from the Fernald meteorologic tower, with uncertainties based partly on the relationships between past and recent Cincinnati airport data. HANFORD The nuclear weapons facility at Hanford, Washington, released substantial quantities of radioactive materials into the atmosphere and the Columbia River from its plutonium production reactors and fuel reprocessing facilities. Two plutonium production reactors started operating at Hanford in December 1944. Two fuel reprocessing plants began extracting plutonium in the same month, and a third production reactor was added in 1945 (Cate and others 1990). The bulk of the releases of radioactive materials—131I discharged from the fuel-reprocessing plants—into the atmosphere occurred between 1944 and 1947. The amount of 131I released during that period is estimated at 685 kilocuries (kCi) (25 PBq) on the basis of the quantity and origin of reactor fuel reprocessed and on the time interval between removal from the reactors and reprocessing (Heeb 1992, TSP 1992, Robkin 1992). The production processes also resulted in the release of other radioactive materials to the atmosphere, to the Columbia River, and to groundwater. A dose reconstruction study began in 1988. In the first phase, scientists developed and tested methods for reconstructing the radiation doses to people who lived in the 10 Washington and Oregon counties closest to Hanford. To do this, they focused on the atmospheric releases of 131I from 1944 to 1947 and on the releases to the Columbia River from 1964 to 1966. Phase I of the study was completed in 1990 (TSP 1990). According to the preliminary results, the highest doses from 131I released to the atmosphere were to infants and young children drinking milk from cows pastured in north Franklin County. Thyroid doses for most individuals in this group of about 1,400 range from 0.15 to 6.5 Gy (15 + 650 rad). Individuals doses from radioactive materials released to the Columbia River during 1964 through 1966 are estimated to have been much lower than doses from

OCR for page 103
contaminated milk during the 1940s (TSP 1990). The final report on the study was published in 1994 (Shipler and Napier 1994). The main difficulty encountered in the part of the dose reconstruction study related to the atmospheric releases of 131I is that very few measurements of environmental radioactivity were made in the mid-1940s, so that the study is based largely on the use of models. In addition, important data about the commercial distribution of milk were not available and had to be obtained from interviews with people who remembered the activities of local dairies in the 1940s. An important characteristic of this dose reconstruction study is the extent to which the public is being kept informed on the progress of the work. The Technical Steering Panel (TSP) of the Hanford Environmental Dose Reconstruction Project, the group responsible for the study, consists of experts in various fields, as well as representatives of the Oregon and Washington state governments, of regional groups of Native Americans, and of members of the public. The TSP made all meetings open to the public and declared that any documents it received were available to the public and media. In addition, the TSP communicates to the public in quarterly newsletters, fact sheets, a video, and a toll-free telephone line. TECHA RIVER The Chelyabinsk-40 center, located near the town of Kyshtym in Russia, was the first Soviet nuclear installation dedicated to the production of plutonium for military purposes (UNSCEAR 1993). A uranium-graphite reactor with an open cooling-water system was commissioned in June 1948, and a fuel-reprocessing plant began operating in December 1948 (Nikipelov and others 1990). Liquid releases to the Techa River from 1949 to 1956 amounted to 2.7 MCi (100 PBq); 95% of this release was discharged between March 1950 and November 1951 (Kossenko 1991). The main constituents released were 89Sr (8.8%), 90Sr (11.6%), 137Cs (12.2%), rare-earth isotopes (26.8%), 95Zr-95Nb (13.6%), and ruthenium isotopes (25.9%). These large releases appear to have resulted primarily from a lack of waste treatment capability and from the storage of radioactive wastes in open, unlined earthen reservoirs (Trabalka and Auerbach 1990). A hydrologic isolation system, including a small reservoir called Lake Karachay, was built after 1952 to contain the low- and intermediate-level wastes (UNSCEAR 1993). The population along the Techa River was exposed to both external and internal irradiation. External irradiation was caused by gamma radiation from 137Cs, 106Ru, and 95Zr-95Nb in the flood plains, in vegetable gardens near houses, and inside houses. Internal irradiation mainly re-

OCR for page 103
sulted from consumption of water and of local foodstuffs contaminated with 89Sr and 90Sr. Average cumulative effective doses are estimated to have been as high as 1.4 Sv in the village of Metlino, 7 km downstream from the point of discharge. The evacuation of the village started in 1953; from 1955 to 1960, inhabitants of another 19 settlements were moved away from the river. Altogether, 7,500 persons were relocated (Kossenko 1991). More recently, Kossenko and Degteva (1994) have estimated that the average effective dose to their Tartar-Bashkir population of 6,123 persons is 0.37 Sv (37 rem); that to the Russian population of 20,563 is 0.13 Sv (13 rem). The person-years weighted mean doses to the bone marrow and soft tissue have been estimated to be 0.37 and 0.14 Gy, respectively (37 and 14 rad). Scientists from the Ural Research Center for Radiation Medicine are following up 28,000 persons who lived in 38 villages along the banks of the Techa River in 1949. External doses have been reconstructed on the basis of gamma-dose rate measurements in the early 1950s along the river bank, on the shore within a few hundred meters of the river, in specified areas of villages, and inside houses. A survey was also made of the lifestyle habits of the populations from the riverside villages (Kossenko and others 1992). Internal irradiation doses have been reconstructed from measurements of the surface-beta activity of teeth, done from 1960 to 1976, and from whole-body counter measurements begun in 1974 (Kossenko and others 1992). GOIANIA In 1987, a 137Cs teletherapy (radiation therapy) source in Goiana, Brazil, was broken by a scrap metal collector who dispersed parts of the 50.9 TBq (1,375 Ci) source (cesium chloride powder) in his house and garden and to other properties in this city of 1.3 million inhabitants. The accident has been described by several papers in a special issue of Health Physics (1991). External gamma irradiation was the main cause of radiation exposure. Acute radiation sickness led to the death of 4 persons whose doses were reconstructed largely on the basis of hematologic observations (Brandão-Mello and others 1991). Dose reconstruction for other persons was begun in 1987 (more than 100,000 were considered to be potentially affected) and used whole-body counting where possible, dose rate measurements in front of highly contaminated persons for which whole-body counter measurements were not possible because of too high a count rate, and dicentric chromosome biologic dosimetry (Ramalho and others 1988). External exposure fields were reconstructed with hand-held dose rate meters (several weeks later and partially after the decontamination most

OCR for page 103
urgently needed) and TLD measurements were made of wall materials. Time-and-space information was assembled on the whereabouts of individuals during the time of the accident. Internal doses of 137Cs were reconstructed by assessment of 137Cs concentrations in air and in food. The following pitfalls were encountered in the dose reconstruction: The measured dose rates were influenced by 137Cs in the environment as well as on the skin and clothes of the persons studied. Estimating organ doses from the measured values is difficult. The external skin contamination of the persons also influenced the interpretation of whole-body counting measurements. Often, the count rate was too high for these sensitive instruments. Whole-body counting cannot indicate the external exposures that contributed most of the doses. There is no calibration curve for the induction of dicentric chromosomes in human lymphocytes by combined internal and external 137Cs irradiation at low doses, and sometimes for partial-body irradiation (a clear research need); in addition, the number of people potentially affected was prohibitively high (more than 100,000 individuals were at risk). The shielding by structures in the urban environment made retrospective determination of external gamma fields difficult and uncertain. The highly heterogeneous nature of the contamination and of its resulting radiation field required an accuracy in the retrospective time-and-space records that could not be met in most cases. The early environmental dispersion of 137Cs resulted from human actions and from resuspension that could not be reconstructed with sufficient detail. Detailed measurements could not be carried out until several weeks after the accident, when inhalation and ingestion played a negligible part in the total exposure and uptake. In summary, the reconstruction of the doses of highly affected people was carried out by clinical evaluation, dose reconstruction by biologic dosimetry was severely hampered by the lack of calibration curves for combined external and internal 137Cs irradiation at a low dose rate, and case-specific exposure pathway analysis models (necessarily unvalidated) based on measured dose rates and 137Cs concentrations weeks later were only useful for approximative retrospective dose estimates for reference persons. NOTES 1.   A petabecquerel (PBq) is 1015 Bq. 2.   A megacurie (MCi) is 106 Ci. 3.   A terabecquerel (TBq) is 1012 Bq. 4.   A picocurie (pCi) is 10-12 Ci.