PREDICTION OF GROUNDWATER COLUMN IN BASEMENT COMPLEX TERRAIN USING ARTIFICIAL NEURAL NETWORK OVER IJEBU-JESA, SOUTHWESTERN NIGERIA

Abstract

The relationship between geoelectric parameters and groundwater column has not been fully understood quantitatively. This is worrisome in the basement complex terrain where these parameters are highly localized. In this study, an attempt was made to predict groundwater column using Artificial Neural Network (ANN) technique that can offer a nonparametric approach in modelling with higher reliability and precision. The ANN technique was applied to geoelectric parameters derived from the results of the interpretation of fifty one Vertical Electrical Sounding (VES) data acquired from the study area. The parameter values so obtained were input parameters used for the training and validation of the ANN. These parameters were the overburden thickness, aquifer thickness, overburden resistivity, aquifer resistivity and coefficient of anisotropy. The results obtained were used to develop groundwater column prediction model. The model was evaluated using two performance criteria namely; Mean Squared Error (MSE) and Regression coefficient (R). The results of the validation showed less mean square error (MSE) of 0.0014286 and the high regression coefficient (R) of 0.98731. This indicates that ANN can be used to predict groundwater column in a basement complex terrain with reasonably good accuracy. The ANN model achieved high predictive accuracy.