A Decision Support System for Identifying Optimal Well Placement

In East Africa, recurrent droughts over the past few decades have contributed to severe food crises, particularly in countries heavily dependent on agriculture. To enhance sustainable resilience to drought, utilizing groundwater through borehole drilling for crop irrigation presents a viable solution. While most well placement research has focused on developing optimization models to guide local farmers in selecting optimal borehole locations, it is crucial to study groundwater characteristics before creating such models in arid and semi-arid regions with varying groundwater depths and recharge potentials.This dissertation proposes an artificial intelligence-based decision support system encompassing groundwater level prediction, groundwater recharge estimation, and well placement optimization to identify suitable borehole sites for sustainable irrigation in data-limited and water-scarce regions. The system's analysis of groundwater potential and suggested optimal well placements aim to support borehole drilling decision-making processes. Focusing on a regional watershed in southern Ethiopia, where water and data availability are limited, this study addresses the substantial water needs and the necessity for cost-effective groundwater utilization strategies. First, a non-time series database of 75 boreholes was used to construct machine learning models, including multiple linear regression, multivariate adaptive regression splines, artificial neural networks, random forest regression, and gradient boosting regression, to predict the depth to the water table. Second, an observation-constrained land surface model was proposed to estimate groundwater recharge for the data-limited, water-scarce region in the Rift Valley basin of Ethiopia. Third, deterministic mixed integer programming and two-stage stochastic mixed integer programming problems were formulated and solved to identify optimal well placements that minimize total costs and satisfy water demand from a sustainable irrigation perspective.