Probabilistic Calculations of Groundwater Travel Time in Heterogeneous Three-Dimensional Porous Media

Abstract

ABSTRACTGroundwater travel time (GWTT) estimation at a potential high-level waste (HLW) repository is subject to various technical uncertainties. These uncertainties stem from model and data uncertainties and cannot be resolved with field tests because of the long time (> 1,000 yr) and large space (> 5,000 m) scales involved. Therefore, computational methods for demonstrating and determining compliance with the GWTT rule will be used. Stochastic theory based approaches constitute a natural framework for performing GWTT estimations under conditions of uncertain and/or limited data.This study employs the generation of spatially correlated hydraulic conductivity fields by the Nearest Neighbor Model (NNM). Repeated (Monte Carlo) realizations of the statistically equivalent random fields are obtained, and the saturated steady-state groundwater flow equation is solved. These results are then used to estimate GWTT along particular paths by releasing a large number of water particles at various starting points. By doing so, path variability is sampled through the realization ensemble space and also through the independent particle “flights” within a specific flow field realization. The uncertainty in predicted GWTT due to parameter variability is assessed for a data set characteristic of the saturated zone at Yucca Mountain at three levels of parameter heterogeneity. However, since several parameters and the boundary conditions of the problem have been arbitrarily assumed, direct conclusions regarding the proposed Yucca Mountain site cannot be drawn from this study.