Water table prediction through causal reasoning modelling

José-Luis Molina; Jose-Luis García-Aróstegui

doi:10.1016/j.scitotenv.2023.161492

Water table prediction through causal reasoning modelling

Sci Total Environ. 2023 Apr 1:867:161492. doi: 10.1016/j.scitotenv.2023.161492. Epub 2023 Jan 10.

Authors

José-Luis Molina¹, Jose-Luis García-Aróstegui²

Affiliations

¹ IGA Research Group, Salamanca University, High Polytechnic School of Engineering Avila, Av. de los Hornos Caleros, 50, 05003 Ávila, Spain. Electronic address: jlmolina@usal.es.
² Geological Survey of Spain (IGME CSIC), Avenida Miguel de Cervantes, 45, 5A, Murcia 30009, Spain; University of Murcia, Institute for Water and Environment, Campus de Espinardo, 30010 Murcia, Spain.

PMID: 36639000
DOI: 10.1016/j.scitotenv.2023.161492

Abstract

This research is mainly aimed to analyze and model the relationship of the binomial Rainfall-Piezometry. In this sense, the inherent causality contained in temporal hourly Rainfall and Groundwater levels (piezometry) data records has been taken. This has been done through Bayesian Causal Reasoning (BCR) which is technique belonging to Artificial Intelligence (AI) based on Bayesian Theorem. The methodology comprises two main stages, first an analytical method from classic regression analysis, and second, a Bayesian Causal Modelling Translation (BCMT) that itself comprises several iterative steps. This research ultimately becomes a tool for aquifers management that comprises a bivariate function made of two variables Rainfall and Piezometry (Temporal Groundwater level evolution). This innovative methodology has been successfully applied in the Quaternary aquifer of the Campo de Cartagena groundwater body, which is an aquifer system that directly is connected to Mar Menor coastal lagoon (Murcia region, SE Spain).

Keywords: Aquifers; Bayesian Causal Modelling; Groundwater; Hydrodynamics; Uncertainty; Water management.