Probability Encoding of Hydrologic Parameters for Basalt. Elicitation of Expert Opinions from a Panel of Three Basalt Waste Isolation Project Staff Hydrologists

摘要

The present study implemented a probability encoding method to estimate the probability distributions of selected hydrologic variables for the Cohassett basalt flow top and flow interior, and the anisotropy ratio of the interior of the Cohassett basalt flow beneath the Hanford Site. Site-speciic data for these hydrologic parameters are currently inadequate for the purpose of preliminary assessment of candidate repository performance. However, this information is required to complete preliminary performance assessment studies. Rockwell chose a probability encoding method developed by SRI International to generate credible and auditable estimates of the probability distributions of effective porosity and hydraulic conductivity anisotropy. The results indicate significant differences of opinion among the experts. This was especially true of the values of the effective porosity of the Cohassett basalt flow interior for which estimates differ by more than five orders of magnitude. The experts are in greater agreement about the values of effective porosity of the Cohassett basalt flow top; their estimates for this variable are generally within one to two orders of magnitiude of each other. For anisotropy ratio, the expert estimates are generally within two or three orders of magnitude of each other. Based on this study, the Rockwell hydrologists estimate the effective porosity of the Cohassett basalt flow top to be generally higher than do the independent experts. For the effective porosity of the Cohassett basalt flow top, the estimates of the Rockwell hydrologists indicate a smaller uncertainty than do the estimates of the independent experts. On the other hand, for the effective porosity and anisotropy ratio of the Cohassett basalt flow interior, the estimates of the Rockwell hydrologists indicate a larger uncertainty than do the estimates of the independent experts. (ERA citation 10:040310)