Integrated Inversion Algorithms to Analyse TDEM Data for Groundwater Resource Assessment in Volcanic Aquifers

Abstract

Summary Assessing groundwater resources in arid or semiarid environments often implies problems of previous data scarcity or logistical difficulties. In our work, we analyse one of such contexts, the volcanic island of Fogo, Cape Verde, to evaluate the presence of groundwater for solving drought problems of a local vineyard. We acquired a profile of Time Domain Electromagnetic (TDEM) soundings inside the target area, whereas we surveyed also external zones, near the few available wells, to correlate the acquisitions with stratigraphic information and water analyses. For the interpretation, we tailored three different inversion algorithms. First, we adopted a global search method, to obtain a 1D resistivity model without being trapped in possible local minima of the solution. We chose the stochastic Particle Swarm Optimization (PSO) algorithm, a computational-intelligence method based on the social dynamics of groups of animals. It was the starting point for the second inversion method, based on local search (1D inversion). In the end, a Spatially Constrained Inversion (SCI) interpretated the dataset with a pseudo-2D model. Our work provided a detailed characterization of the resistivity distribution in the subsurface of the vineyard, which outlined the presence of a probable thick suspended aquifer, which may contain exploitable groundwater resources.