Predicting Groundwater Potential Areas Using Hybrid Artificial Intelligence Methods (Case Study: Birjand Plain)

Groundwater is one of the most valuable resources for communities, agriculture, and industry. In the present study, three new artificial intelligence models, including Modified Real AdaBoost (MRAB), Bagging model (BA), and Rotation Forest model (RF), have been developed by the Functional Tree Base Classifier (FT) model to predict groundwater potential in Birjand plain area. Therefore, for implementation, geo-hydrological data of 37 groundwater wells and ten factors of topography, hydrology, and geology were used. The performance of these models was evaluated using the area under the curve (AUC) and other statistical indicators. The results showed that although all the hybrid models developed in this study increased the prediction accuracy, MRAB-FT model (AUC = 0.742) has higher accuracy in predicting potential groundwater areas in Birjand plain. Accurate mapping of groundwater potential areas while maintaining a balance between consumption and operation will help feed the aquifer for optimal use of groundwater resources.