Geophysical Data Integration and Conditional Uncertainty Analysis on Hydraulic Conductivity Estimation

Integration of various geophysical data is essential to better understand aquifer heterogeneity. However, data integration is challenging because there are different levels of support between primary and secondary data needed to be correlated in various ways. This study proposes a geostatistical method to integrate the hydraulic conductivity measurements and electrical resistivity data to better estimate the hydraulic conductivity (K) distribution. The K measurements are obtained from the pumping tests and represent the primary data (hard data). The borehole electrical resistivity data from electrical logs are regarded as the secondary data (soft data). The electrical resistivity data is used to infer hydraulic conductivity values through the Archie law and Kozeny-Carman equation. A pseudo cross-semivariogram is developed to cope with the resistivity data non-collocation. Uncertainty in the auto-semivariograms and pseudo cross-semivariogram is quantified. The methodology is demonstrated by a real-world case study where the hydraulic conductivity is estimated in the Upper Chicot aquifer of Southwestern Louisiana. The groundwater responses by the cokriging and cosimulation of hydraulic conductivity are compared using analysis of variance (ANOVA).