Phreatic levels hybrid model from different environmental variables

Abstract

Groundwater systems are complex and heterogeneous. The levels are affected by a combination of natural and anthropogenic factors. These factors, added to the inherent natural characteristics (such as permeability, porosity and grain size) represent challenges for hydrological modeling. This study aimed to verify the influence of soil physical-hydric and management, topography and vegetation variables on the Bauru Aquifer System (SAB) groundwater levels oscillations. These aspects summed up to 21 possible predictive variables of groundwater levels. Information were collected at the watersheds of the Santa Barbara Ecological Station/SP-Brazil. The most relevant variables for the multiple regression model were chosen using principal components analysis, which determined those variables with greater variability. The results indicated a robust fit to the data by the model and good predictive capacity for new observations. This information can help decision-making on water use and support water resource management instruments, as tools for water resource planning.