“…GWL modeling based on ML has the unique ability to find the likely relationships between GWL and controlling hydro-climatic-anthropogenic variables without constructing knowledge-driven conceptual or physically-based models. Therefore, researchers have studied the performance of ML methods for GWL modeling in India and Bangladesh (Nayak et al, 2006;Nury et al, 2017;Malakar et al, 2018;Mukherjee and Ramachandran, 2018;Bhanja et al, 2019b;Sun et al, 2019;Yadav et al, 2019 and the references therein) and other parts of the world (Coulibaly et al, 2001; 60 Feng et al, 2008;Sun, 2013;Nourani and Mousavi, 2016;Sun et al, 2016;Yoon et al, 2016;Barzegar et al, 2017;Ebrahimi and Rajaee, 2017;Wunsch et al, 2018;Chen et al, 2019;Lee et al, 2019 and the references therein). Most of these studies used popular methods like Artificial Neural Network (ANN), hybrid-ANN, Adaptive neuro-fuzzy inference system (ANFIS), Support Vector Machine (SVM) and few others using a wide range of frequency and temporal data on past GWLs, satellite observations derived groundwater storage (GWS), Normalized difference vegetation index (NDVI)), meteorological variables, river discharge, 65 variables on groundwater use and few dummy variables to simulate and/or predict GWLs.…”