Sensitivity Analysis Techniques and Results for Performance Assessment at the Waste Isolation Pilot Plant

Abstract

The purposes of this report are to (1) review available techniques for the performance of uncertainty and sensitivity analyses for computer models, (2) discuss the approach to uncertainty and sensitivity analysis selected for use in performance assessment for the Waste Isolation Pilot Plant (WIPP), (3) present the results of uncertainty and sensitivity analyses performed in support of the 1990 preliminary performance assessment for the WIPP, and (4) discuss the US Environmental Protection Agency (EPA) release limits to the accessible environment (40 CFR 191) in the context of a formal representation for risk. The following approaches to uncertainty and sensitivity analysis are reviewed: differential analysis, Monte Carlo analysis, response surface methodology, and Fourier amplitude sensitivity test. Of these approaches, Monte Carlo analysis is felt to be the most widely applicable to the problems that arise in performance assessment for radioactive waste disposal. Monte Carlo analysis is based on performing multiple model evaluations with probabilistically selected model input, and then using the results of these evaluations to determine both the uncertainty in model predictions and the independent variables that give rise to this uncertainty. Overall, the releases were small, usually much smaller than the EPA release limits for the WIPP. Calculation of the EPA release limits involves a Monte Carlo calculation of a complementary cumulative distribution function (CCDF). The calculation of this CCDF is discussed and shown to fit into a formal representation for risk based on scenarios, scenario probabilities, and scenario consequences. The relationships between the Monte Carlo procedures used in the WIPP performance assessment and this formal representation for risk are developed. 164 refs., 79 figs., 18 tabs.