The Worth of Stochastic Inversion for Identifying Connectivity by Means of a Long‐Lasting Large‐Scale Hydraulic Test: The Salar de Atacama Case Study

Abstract

AbstractUnderstanding groundwater flow involves characterizing the spatial variability of aquifer attributes and possible hydraulic connectivity structures. The latter are of crucial importance because high permeability channels may control groundwater flow and contaminant transport. In evaporitic aquifer systems, these preferential channels can consist of karst conduits, developed at different scales, as well as fault zones. These features condition the economic development of salt flats at the Central Andes. Hydraulic connectivity may affect exploitation efficiency by enhancing the inflow of less evaporated brine thus diminishing its mineral content (i.e., concentration of Li and K). In this context, we investigate if it is possible to use (tomographic) stochastic inversion (regularized pilot point method) in order to characterize the presence of connectivity structures in an evaporitic aquifer of great extension (some 1,500 km2) from head response measured at numerous observation points during a year long sequence of three hydraulic tests. Results show that, even though the solution is nonunique, the main preferential flow zones are identified. Numerous inversions yield similar fits to observed drawdowns with maximum errors of few centimeters. Preferential flow is identified not only by elongated high permeability regions, but also by a marked scale effect (model transmissivities are some 30 times larger than their local tests counterparts). The main high conductivity zones are consistent with independent information based on geophysics, isotopes, mixing ratios, piezometric data, and the expected dissolution processes.