Multivariate approaches optimize locations of groundwater pumping facilities for different hydrogeological scales

Abstract A number of optimization approaches regarding the design location of groundwater pumping facilities in heterogeneous porous media have elicited little discussion. However, the location of groundwater pumping facilities is an important factor because it affects water resource usage. This study applies two optimization approaches to estimate the best recharge zone and suitable locations of the pumping facilities in southwestern Taiwan for different hydrogeological scales. First, for the regional scale, this study employs numerical modelling, MODFLOW‐96, to simulate groundwater direction and the optimal recharge zone in the study area. Based on the model's calibration and verification results, this study preliminarily utilizes the simulated spatial direction of groundwater and compares the safe yield for each well group in order to determine the best recharge zone. Additionally, for the local scale, the micro‐hydrogeological characteristics are considered before determining the design locations of the pumping facilities. According to drawdown record data from six observation wells derived from pumping tests at the best recharge area, this study further utilizes the modified artificial neural network approach to improve the accuracy of the estimation parameters as well as to analyse the direction and anisotropy of the hydraulic conductivities of an equivalent homogeneous aquifer. The results suggested that the best locations for the pumping facilities are along the more permeable major direction. Copyright © 2011 John Wiley & Sons, Ltd.