K.8.0 ECOLOGICAL RISK UNCERTAINTY
This section provides a qualitative discussion of the uncertainties in the ERA. The ERA for this EIS used a screening level methodology to estimate potential radiological and chemical hazards to a suite of representative terrestrial receptors: the Great Basin pocket mouse, coyote, mule deer, red-tailed hawk, and loggerhead shrike (Volume Three, Section D.6). Pathways considered for the No Action alternative were food and water ingestion (all receptors except the mouse, which was assumed to obtain all water from metabolic sources), incidental soil ingestion (mouse and mule deer, coincident with consumption of vegetation), inhalation of routine releases (all), and direct external exposure (mouse, while in a burrow). Potential hazards to aquatic organisms were evaluated using the CRITRII program developed at the Pacific Northwest National Laboratory.
K.8.1 SOURCE TERMS
The source terms used for the ERA were the same as those used for the human health risk assessment. For the No Action alternative, the source terms were the inventory in the tanks for direct contact and food chain uptake, routine releases for air inhalation, and groundwater reaching the Columbia River in the future for water ingestion. Uncertainties in the source terms are described in Section K.3. The estimated ecological risks were directly proportional to the source contaminant concentrations, and uncertainty about these terms was considered relatively high. Therefore, the source terms were considered likely major contributors to the uncertainty in the ecological risk estimates. However, for the pathways involving direct exposure to stored wastes, radionuclide concentrations would have to have been overestimated by a factor of at least 10,000 for the "true" radiation doses to approach the 0.1 rad/day benchmark recommended by International Atomic Energy Agency (IAEA1992) for protection of terrestrial organisms, since most of the estimated doses were greater than 1,000 rad/day (Volume Three, Table D.6.4.1). Chemical concentrations would have to have been overestimated by a factor of 10 to 1,000 for the true HI to approach the benchmark value of 1.0, since most of the estimated HIs were between 10 and 1,000. If the inventories used for the existing risk estimates were underestimates, the corresponding true risks would be even greater than stated. Therefore, although the source terms were probably important contributors to the uncertainty in the absolute values of the risk estimates, it is not likely that better or more data would alter the conclusion that direct contact with the waste, either externally or through food chain uptake, would be very hazardous to ecological receptors.
Similar but converse arguments would apply to the estimates of inhalation and groundwater ingestion pathway risks to ecological receptors under the No Action alternative. Estimated radiation doses from these two pathways were very low compared with the two benchmarks used [0.1 rad/day for terrestrial organisms and 1.0 rad/day for aquatic organisms (IAEA 1992, NCRP 1991)] (Volume Three, Tables D.6.4.2, D.6.4.3, and D.6.4.6). The true radionuclide concentrations in air and water would have to be higher than those estimated by at least a factor of 10,000 for the maximum estimated doses to approach the lower of the two benchmarks of concern (0.1 rad/day). It is therefore unlikely that better estimates would alter the conclusion in the EIS that inhalation and groundwater ingestion would not be important sources of ecological risk under the No Action alternative. It is possible that more refined estimates of maximum future chemical concentrations in groundwater, assuming those estimates were 10-fold higher than the existing ones, could indicate potential chemical hazards to wildlife (Volume Three, Table D.6.4.5). This discussion would apply also to the estimated radiation doses resulting from inhalation of routine releases for the remediation alternatives. The true release terms would have to be higher than the existing estimates by several orders of magnitude for the estimated doses to approach the 0.1 rad/day benchmark of concern (Volume Three, Tables D.6.4.8 trough D.4.6.9). The exceptions would be the Ex Situ Intermediate Separation/Phased Implementation and the Ex Situ No Separation alternatives. The maximum radiation doses estimated for these two alternatives would exceed the 0.1 rad/day benchmark using the existing release estimates, and the real values would have to be 10- to 100-fold lower for the maximum radiation doses to fall below the level of concern. However, the minimum radiation doses for these two alternatives were approximately 100,000-fold lower than the maximum, supporting the claim that the existing dose values would be upper bound estimates.
K.8.2 ORGANISM VARIABLES
A number of species-specific variables contributed to the uncertainty in the ecological risk estimates: food, water, and soil ingestion rates; inhalation rates; body weights; home ranges; and the effective radius for absorption of energy from radioactive decay. Risk estimates for any given pathway are directly proportional to the associated contaminant intake rate. For example, a 10-fold difference in a food ingestion rate would produce a 10-fold difference in the estimated radiation dose or chemical HI. Therefore, the risk estimates would be expected to be sensitive to errors in the intake rates. However, as discussed in Section K.8.1, most of the risk estimates would need to be wrong by several orders of magnitude for the high values to fall below levels of concern or for the low values to exceed them. Errors of this size were considered unlikely for food ingestion rates, which were based on direct measures of various types (see sources for Volume Three, Table D.6.2.1). The exception would be the food ingestion rate for the loggerhead shrike, which was estimated from body weight using an empirical equation (Volume Three, Table D.6.2.1). Water and soil ingestion rates and inhalation rates were all estimated from body weight (water ingestion and inhalation) or dry matter intake (soil ingestion). Potential errors in these variables were likely greater than those for food ingestion, and the resulting risk estimates for the water and soil ingestion and the inhalation pathways could have a greater level of uncertainty than those for food ingestion. However, the food ingestion pathway has an additional source of uncertainty, biological transfer factors in the food chain (discussed in the following paragraphs), which the other pathways do not share.
Risk estimates for any given pathway would typically be inversely proportional to body weight, since body weight appears in the denominator of the equations for estimating intake of radionuclides or chemicals (Volume Three, Appendix D). However, as noted above, water ingestion rates and inhalation rates were estimated from empirical equations that are a function of body weight. This causes weight to appear in both the numerator and denominator of the intake equations, reducing its overall effect on the risk estimate. For example, a simple equation for the intake of a chemical by water ingestion is:
1) 
Where:
Ii = Intake rate of the ith contaminant, mg kg-1
day-1
Ci = Contaminant concentration in water, mg L-1
IR = Ingestion rate of water, L day-1
BW = Body weight, kg
FI = Fraction ingested from contaminated source, unitless
Here, the intake rate of a contaminant in water would be inversely proportional to the body weight, if the water ingestion rate remained constant. However, larger animals generally drink more water than smaller animals (on a per organism basis), so that as the denominator, body weight, increases, IR in the numerator also increases. In the TWRS EIS, this was explicitly the case, because IR was estimated from body weight using empirical equations. For example, the equation used to estimate water ingestion by the coyote and mule deer is:
2) 
Where:
IR = Ingestion rate of water, L day-1
BW = Body weight, kg
The equation used to estimate inhalation by the mouse, coyote, and mule deer is:
3) 
Where:
IR = Inhalation rate, m3 day-1
BW = Body weight, kg
The overall effect of using these equations is to reduce the potential effect of body weight variability on the water ingestion and inhalation risk estimates and on the uncertainty in the estimates. In addition, in parallel with the previous discussions of source terms and intake rates, the body weight values would need to be wrong by orders of magnitude for the high risk values to fall below levels of concern or for the low values to exceed them. Errors of this size were considered unlikely, because the body weights were based on reported measured values of real organisms, and the body weights of adult mammals and birds do not vary by orders of magnitude within species.
Risk estimates in this ERA were also inversely proportional to the estimated home range, except that home ranges equal to or less than 100 hectares (ha) (250 acres [ac]), the unit cell size used for the risk assessment (see Volume Three, Appendix D), all have the same effect on the risk estimate. That is, FI, the fraction of exposure to a contaminated source (see Equation 1 above), was set equal to the ratio of the cell size and the home range. For example, the coyote home range of 300 ha (750 ac) (Volume Three, Table D.6.2.1) results in an FI of (100/302) equals 0.33. Above a home range of 100 ha (250 ac), exposure would be directly proportional to home range size. Below this, all species have the limiting FI of 1.0. For example, the mouse and loggerhead shrike have home ranges of 0.09 and 10 ha (0.22 and 25 ac), respectively (Volume Three, Appendix D), but both have an FI of 1.0. The consequence is that potential variability or errors in the estimated home ranges would have different effects on different species risks. A 1,000-fold error in the mouse home range (i.e., the real value is 90 ha [220 ac]) would not affect the mouse risk estimates, because the FI would still equal 1.0. However, any error or variability in the home ranges for the coyote, mule deer, or hawk, with estimated home ranges of 300, 1,200, and 222 ha, (750, 3,000 and 540 ac) respectively, would have a proportional effect on the risk estimates. Errors of 10-fold in estimates of home range would not be unlikely, given that home range is defined as an area in which an animal sleeps and/or breeds and might not be the same as the area over which the animal forages for food. However, if the estimated home ranges were too low, more accurate (higher) values would decrease the risk, assuming that the resulting FI fell below the limiting value of 1.0.
If the home ranges used were too high, more accurate values would increase the estimated risk only for the coyote, mule deer, or hawk (which have home ranges greater than 100 ha [250 ac]). The worst case increase would be about 10-fold for the deer, with a home range of 1,200 ha (3,000 ac). Decreasing this value to 100 ha (250 ac) would increase the deer FI to 1.0 and increase estimated risks by a factor of 12 (1,200/100). This would affect conclusions about risk to deer only marginally. For example, the HI for direct contact with the waste in Cell 1WSS is 0.7, which is close to but not above the 1.0 benchmark of concern. The HI 0.7 multiplied by the estimated rise 12 is 8.4, which is above the benchmark. However, direct contact with the waste was already characterized as very hazardous to deer, and this conclusion would be unchanged. Therefore, although home ranges are not known with high confidence, uncertainty or variation in them would be unlikely to affect the conclusions of the ERA.
The effective energy absorbed from radionuclide decay has a direct proportional effect on the estimated radiation dose. This energy in turn depends on the effective radius assumed for the organism. The smaller the radius, the less energy is absorbed, although this varies across radionuclides due to their different physicochemical characteristics and the relative importance of alpha, beta, and gamma decay. The effective radii assumed for the organisms used in this ERA ranged from 1.4 cm (0.55 in.) (plants) to 30 cm (12 in.) (coyote and deer). An examination of Table D.6.3.7 shows that effective absorbed energies vary by approximately 10-fold across this range, with a number of isotopes showing no difference. Therefore, the effect of the assumed radius on the radiation estimates would be at most 10-fold if the entire dose were due to isotopes with this range of variation in effective absorbed energy. Such differences would not affect the primary conclusions of the ERA.
K.8.3 BIOLOGICAL TRANSFER FACTORS, HALF-LIVES, AND RETENTION FACTORS
The ecological risk estimates for food ingestion in the EIS were directly proportional to the soil-to-plant transfer factors (which determine how much of a contaminant moves into the food chain), the biological half-lives or turnover times of contaminants within organisms, and the fraction of the contaminant retained in the organism at each step of the food chain. Soil-to-plant transfer factors are likely to vary depending on local soil conditions and different plant species. This ERA used published default values that may or may not be applicable to the specific types and locations of plants consumed by the pocket mouse or mule deer. Therefore, these factors therefore probably contribute substantially to the uncertainty in the food ingestion risk estimates. However, as discussed for the source terms, the error or variability would need to be several orders of magnitude before it would affect the conclusions of the ERA. A similar concern exists for biological half-lives and retention factors for chemicals and radionuclides, although again, these would have to be wrong by several orders of magnitude before they by themselves would affect the conclusions of the ERA. The biological half-lives for radionuclides include the radiological half-lives in their calculation. These latter values are known with very high confidence and would not contribute significantly to the uncertainty in the risk estimates.
K.8.4 NO OBSERVED ADVERSE EFFECT LEVELS
The conclusions about potential effects of radiation on ecological receptors relied on single benchmarks of 0.1 rad/day for terrestrial organisms and 1.0 rad/day for aquatic organisms (Volume Three, Appendix D). These benchmarks are independent of specific radionuclides, reflect intense study, and have been widely reviewed. It is therefore unlikely that they contribute importantly to uncertainty in the conclusions of the ERA. However, the estimates of risks due to hazardous chemical exposure are based on the ratio of the estimated intake to the No Observed Adverse Effect Level (NOAEL); see (Volume Three, Appendix D). These values were derived largely from laboratory studies of species other than those of interest in this EIS. There are potential uncertainties in extrapolating from the species used in the laboratory studies to those in the EIS, from the dose ranges used in the studies to those estimated in the field, and from the general conditions in the laboratory to those in the field. All these factors may contribute significantly to the uncertainty in the HI estimates. The HI is inversely proportional to the NOAEL. An examination of Tables D.6.4.4 and D.6.4.5 suggests that errors of 10-to 100-fold in the NOAEL could reduce the high HI values in Table D.6.4.4, for direct contact with the stored waste, to low values below the 1.0 benchmark, if the real NOAELs were higher than those used in the ERA. Similar errors in the opposite direction could increase the low HI values in Table D.6.4.5, for future consumption of groundwater reaching the Columbia River, to values above the 1.0 benchmark. It is therefore possible that uncertainty in the NOAELs for hazardous chemicals could affect the ERA conclusions about chemical hazards associated with direct contact with stored waste or future consumption of groundwater at the maximum concentrations reaching the Columbia River. This would not affect conclusions about the presence or absence of radiological hazards or general conclusions about the need to prevent contact with the tank wastes by ecological receptors.
K.8.5 ECOLOGICAL RISK ASSESSMENT UNCERTAINTY CONCLUSIONS
Overall, the parameters of the equations used to estimate ecological risks would need to vary or be in error by several orders of magnitude to affect the conclusions of the ERA by themselves. Simultaneous variability in multiple parameters in the same direction could do so. For example, increasing IR and FI in Equation 1 above by 10-fold each would increase the estimated contaminant intake by a factor of 100. Such simultaneous variability is possible and would contribute to the overall uncertainty in the risk estimates. Nonetheless, because the ecological risk estimates in this EIS are so different for the various scenarios considered; very high for direct contact with stored wastes and very low for routine releases associated with either the No Action or various remedial alternatives, more detailed analysis would not be considered likely to alter those distinctions. Conversely, more detailed analysis would be unlikely to permit clear distinctions among the remedial alternatives based on potential radiological risks of routine releases, because these latter values are both low and similar to each other. The primary distinction in ecological risks thus remains between the No Action (assuming direct contact with the stored wastes at some future point) and remediation alternatives collectively.
|
NEWSLETTER
|
| Join the GlobalSecurity.org mailing list |
|
|
|
