Assessing the groundwater recharge processes in intensively irrigated regions: An approach combining isotope hydrology and machine learning

• Isotope hydrology and Artificial Neural Network (ANN) to characterize groundwater recharge. • Precipitation, river, Irrigation Return Flow (IRF) and canal as the main sources of groundwater recharge. • Stable isotope modeling of evaporative enrichment in irrigated water. • Relationship between different hydrogeochemical variables. • Findings aid the policymakers in crafting sustainable water management strategies for irrigated regions. Agriculture is a major contributor to the global economy, accounting for approximately 70 % of the freshwater use, which cause significant stress on aquifers in intensively irrigated regions. This stress often leads to the decline in both the quantity and quality of groundwater resources. This study is focused on an intensively irrigated region of Northern India to investigate the sources and mechanism of groundwater recharge using a novel integrated approach combining isotope hydrology, Artificial Neural Network (ANN), and hydrogeochemical models. The study identifies several key sources of groundwater recharge, including natural precipitation, river infiltration, Irrigation Return Flow (IRF), and recharge from canals. Some groundwater samples exhibit mixing from various sources. Groundwater recharge from IRF is found to be isotopically enriched due to evaporation and characterized by high Cl − . Stable isotope modeling of evaporative enrichment in irrigated water helped to differentiate the IRF during various cultivation periods ( Kharif and Rabi ) and deduce the climatic conditions prevailed during the time of recharge. The model quantified that 29 % of the irrigated water is lost due to evaporation during the Kharif period and 20 % during the Rabi period, reflecting the seasonal variations in IRF contribution to the groundwater. The ANN model, trained with isotope hydrogeochemical data, effectively captures the complex interrelationships between various recharge sources, providing a robust framework for understanding the groundwater dynamics in the study area. A conceptual model was developed to visualize the spatial and temporal distribution of recharge sources, highlighting how seasonal irrigation practices influence the groundwater. The integration of isotope hydrology with ANN methodologies proved to be effective in elucidating the multiple sources and processes of groundwater recharge, offering insights into the sustainability of aquifer systems in intensively irrigated regions. These findings are critical for developing data-driven groundwater management strategies that can adapt to future challenges, including climate change, shifting land use patterns, and evolving agricultural demands. The results have significant implications for policymakers and water resource managers seeking to ensure sustainable groundwater use in water-scarce regions.