Global groundwater resources are under stress due to high extraction rates, pollution, landuse changes, and the impacts of climate change. In India, too, the decline in groundwater level is a growing issue for water supply. This is well investigated in northern India, but groundwater resources are also declining in parts of South India. While surface water historically supplied agriculture, it has been largely replaced by groundwater irrigation since the 1950s, resulting in the overexploitation of fragile aquifers. Population growth, urbanisation, and rising food demand are further intensifying water needs. Climate change adds uncertainty, with projections of higher temperatures, shifting rainfall patterns, and greater drought risks. The Sathaiyar basin, located in the South Indian state of Tamil Nadu, is marked by locally declining and rising groundwater levels, providing a unique case to study these dynamics. Half of the basin is covered by a cascade of irrigation tanks, which are small surface water bodies that store excess river water and rainwater. The primary objective of the present work is to assess the impact of i) natural groundwater recharge, as well as ii) indirect recharge from irrigation tanks on groundwater. The quantification of natural recharge was conducted using two approaches: the first involves simulating the infiltration of rainwater through the unsaturated zone, and the second evaluates the water level time series from an observation well. The local impact of tanks was assessed by modelling the saturated-unsaturated zone processes of two separate irrigation tanks. On the catchment scale, the effect of the tank cascade on groundwater was investigated using hydrochemical and stable isotope tracers. The natural groundwater recharge was simulated with three Hydrus 1D models. By calibrating the models with stable water isotopes from soil water at different depths, they provided reliable information on the quantity of recharge through different soils while considering the influence of varying vegetation. Using only the soil water content for model calibration would have resulted in a large range of recharge rates, which emphasises the importance of using multiple data sources to build a reliable model. At the site with the highest clay content of up to 80%, groundwater recharge is nearly negligible 0.01 cm from May to February, whereas at the site with only 20% clay, recharge amounts to 31 cm for the same period. The results are in the order of magnitude of the recharge estimated using the time series model Pastas, where the estimated annual recharge is up to 4.7 cm yr−1. Analysis of the data from rainfall stations in the catchment indicated a minor increasing trend for two stations, but no trend for the remaining locations. This suggests that natural recharge has only a small impact on groundwater level rise and is unlikely to increase substantially in the future. In contrast, the recharge from irrigation tanks is high: the spatio-temporal trend analysis of groundwater level time series shows that groundwater levels are mainly increasing below the tank cascade. Most likely, the increasing trends are caused by the iii infiltration capacity of the tanks. The results of the two tank models show a significant contribution of both tanks to groundwater recharge. As the tanks in the cascade are filled with water for most of the time throughout the year, recharge is mainly limited by low depth to groundwater and the resulting limited storage capacity of the aquifer. Model results suggest that regular desilting of the tank bed would increase infiltration rates. Another important parameter that strongly influences tank recharge efficiency is the hydraulic conductivity of the weathered hard rock. In the present study, it was constrained by calibration; future work could determine it directly using field experiments or tracer studies. The extent of the tanks’ impact was further investigated by using hydrochemical and isotope tracers 𝛿2𝐻 and 𝛿18𝑂 on the catchment scale. Samples from groundwater and tank water were taken before and after the monsoon season. The recharge effect of tanks could be demonstrated even in deep wells (up to 230m depth) by means of stable isotopes. An apparent difference was observed between the groundwater upstream of the tanks and the groundwater influenced by the tanks. Besides the benefit of enhanced groundwater quantity, a deterioration in its quality has been identified by hydrochemical analysis. Groundwater near tanks receiving sewage inputs showed evidence of anthropogenic pollutants, while agricultural activities additionally impacted the hydrochemical composition. In the present study, the influence of anthropogenic activities, such as sewage discharge and agrochemical application, could not be clearly distinguished from the natural processes that influence the chemical composition of groundwater. Employing tracers with higher specificity for human activities may enhance the resolution of future studies. The results of this work suggest that natural groundwater recharge is insufficient to balance current human extraction rates in the catchment during the study period. In contrast, the tank cascade provides sustained, indirect recharge and seems to be an effective tool for local water resources management. This benefit, however, is tempered by limited water quality, which emphasises the need to pair recharge management with upstream sanitation and water quality monitoring.
