Approximate Reasoning in Hydrogeological Modeling

The accurate determination of hydraulic conductivity is an important element of successful groundwater flow and transport modeling. However, the exhaustive measurement of this hydrogeological parameter is quite costly and, as a result, unrealistic. Alternatively, relationships between hydraulic conductivity and other hydrogeological variables less costly to measure have been used to estimate this crucial variable whenever needed. Until this point, however, the majority of these relationships have been assumed to be crisp and precise, contrary to what intuition dictates. The research presented herein addresses the imprecision inherent in hydraulic conductivity estimation, framing this process in a fuzzy logic framework. Because traditional hydrogeological practices are not suited to handle fuzzy data, various approaches to incorporating fuzzy data at different steps in the groundwater modeling process have been previously developed. Such approaches have been both redundant and contrary at times, including multiple approaches proposed for both fuzzy kriging and groundwater modeling. This research proposes a consistent rubric for the handling of fuzzy data throughout the entire groundwater modeling process. This entails the estimation of fuzzy data from alternative hydrogeological parameters, the sampling of realizations from fuzzy hydraulic conductivity data, including, most importantly, the appropriate aggregation of expert-provided fuzzy hydraulic conductivity estimates with traditionally-derived hydraulic conductivity measurements, and utilization of this information in the numerical simulation of groundwater flow and transport