III. WATER PATHWAYS
The authors relied heavily on modeling to estimate the source term. Modeling was necessary because the environmental monitoring data were lacking or because the concentrations of the radionuclides in the environmental media were below the limits of detectability. Radioactivity releases due to reactor operation, including cleaning of the reactor pipes with abrasive diatomaceous earth, and the nearly 2,000 fuel-element failures were included in the model. Most of the radionuclides released were due to neutron activation of trace impurities in the water. The authors should have added a comment to the effect that fall and spring turnover of the lakes created by dams upstream from the Hanford Site caused changes in the mineral content of the cooling water which led to differences over short periods of time in the radionuclides induced by the neutron irradiation. This was a potential source of uncertainty that was not addressed.
The total amount of radionuclides released and the river radionuclide burden for the years 1944-1971 are presented graphically. Because of various factors (e.g., physical half-life), there was no relationship given between the amount of radionuclides released and radiation dose.
Transport of radionuclides by the Columbia River was modeled with a modified version of the computer code WUS-CHARIMA. The historical data required for estimating a sediment transport were unavailable; therefore, this parameter was not used. In the report’s Figure 3.5, the concentrations of chronium-51 predicted by the model were compared with active measurements; the committee has not reviewed the document that discusses the agreement between estimates and measured data in more detail (Walters et al., 1994). Concentrations were highest in the winter months, when river flows were lowest, but it is not clear that seasonal variations in the water consumption by members of the exposed population were taken into consideration.
PNWD-2227 HEDR summarizes the technical approach used to estimate radiation doses received by three classes of representative individuals who might have used the Columbia River as a source of drinking water, or food or for recreational or occupational purposes in 1944-1992: maximally exposed individuals, typically exposed individuals, and occupationally exposed individuals. The report briefly explains the approaches used to estimate the radioactivity released to the river, the development of the parameters used to model the uptake and movement of radioactive materials in aquatic systems (such as the Columbia River), and the method of calculating the Columbia River’s transport of radioactive materials. Dose estimates were calculated for the three representative types in 12 segments of the Columbia River from the Hanford Site to the mouth of the river; they include ingestion of Willapa Bay shellfish and salmon or steelhead caught in the river. Doses were calculated for five radionuclides that together contributed over 94% of the total dose: sodium-24, phosphorus-32, zinc-65, arsenic-76, and neptunium-239 (see page v, PNWD-2227 HEDR). Six additional radionuclides were included in source-term estimates because they were needed for river-transport validation or were of particular interest to the TSP (page 3.4). Doses are presented as the effective dose equivalent (sum of the committed effective dose equivalent from internal deposition of radionuclides in the body and the effective dose equivalent from external radiation) and dose equivalent (determined by multiplying the absorbed dose in rads by a quality factor) for the red bone marrow and lower large intestine.
One of the committee’s concerns is that important details are missing, such as the physical half-lives of the five most important radionuclides, the average time taken by water to cross each segment of the Columbia River, most of the model validation results, most of the uncertainty estimates, and the distribution of the population along the Columbia River. In addition, there is a lack of consistency between the selection of periods in the figures and the discussion of the periods in the results. It would be preferable to discuss the three periods—1944-1949, 1950-1971, and 1972-1992—separately in each section. The reason for treating the three differently should be explained more clearly.
The committee is concerned that the authors of Tables 3.1 and 3.2 (pages 3.15 and 3.16 in PNWD-2227 HEDR) have differentiated the uncertainty in the bioconcentration factor into components (e.g., waterfowl, predator, warm season; omnivore, warm season; etc.) but not one of the largest uncertainty factors, namely, the “ingestion dose conversion factor.” At a minimum, the committee would have expected at least three components: variations among persons in uptake, in retention, and (within a given age and sex group) in intake of water, fish, and waterfowl.
The last component, which might well be the largest of the three, applies only to intake assessments for “representative” persons and not to real persons. Hence, if the authors’ uncertainty analyses were based on data on real persons, the size of the “ingestion dose conversion factor” relative to other factors might be substantially different from what they have presented. The reader is not told what components went into the index of uncertainty for the “ingestion dose conversion factor.” For example, how much variability did they impute to the uptake and retention components and where did this information come from?
The largest contributor to the uncertainty in the intake of radionuclides might be the unreliability in personal reporting of the intake of water, fish, or waterfowl. Did the authors consider this? If so, how? The committee notes that this applies only to intake assessments for “representative” persons and not to real persons.
In general, the presentation of the uncertainty analysis suffers from the failure to present a table of the specific sources of uncertainty and the sizes and shapes of the assumed distributions for these sources. Until some type of documentation is given, the uncertainty analyses have to be regarded as questionable. Given the large uncertainties that exist in estimating the dose from water pathways, the committee recommends that the lack of precision in these doses be recognized explicitly.
Ingestion dose conversion factors and drinking-water transmission factors for the five radionuclides included in the model are based on published data (see Table 3.3, page 3.23, PNWD-2227 HEDR). It would have been useful to indicate the ranges of possible choices for these factors and where in the ranges the values used lie. It is unclear why intake of drinking water by a typical representative individual is 40% lower than that by the maximum representative individual and occupational representative individual.
Finally, given the relative magnitude of doses associated with this pathway, the committee is concerned that too much effort has been expended on estimating doses acknowledged to be very small. This pathway needs to be considered, but a decision could be made to limit consideration to screening calculations, as was done for the earlier years (1949-1950). With all the uncertainties in the estimation of dose and the small doses associated with this pathway, such detailed and elaborate calculations seem unnecessary and project a sense of reliability that is unwarranted. Furthermore, it is unlikely that this pathway will be used in the epidemiologic study of thyroid disease.