Water balance dynamics under climate variability and lulc changes—a catchment scale assessment

The aim of this study was to develop a catchment scale water balance model capable of simulating water balance components (WBCs), with a specific focus on data-sparse environments. Using the Soil and Water Assessment Tool (SWAT), an open-source semi-distributed hydrological model, the WBCs dynamics simulated by integrating decadal land-use and land-cover (LULC) changes, daily meteorological inputs, and time-invariant soil and topography data sets. The model simulates the spatial and temporal distribution of key WBCs, including actual evapotranspiration (AE), surface runoff, lateral flow, percolation, and soil water content (SW) at monthly and annual scale. The model framework was applied to the Chittar catchment in Tamil Nadu, India, covering a 20 years historical period (January 2001–December 2020) and a 30 years forecast period (January 2021–December 2050). The model was calibrated and validated with river gauge discharge data using the SWAT-Calibration Uncertainty Procedures (SWAT-CUP) database. The R-square values were 0.87 for calibration and 0.93 for validation. To complement the single-variable calibration (i.e. river gauge discharge at the outlet) and assess model consistency, monthly simulated AE across three representative sub-catchments was cross-validated against gridded AE estimates using Global Land Evaporation Amsterdam Model (GLEAM) v3.6a. This validation yielded R-square values ranging from 0.78 to 0.82 across the sub-catchments. This dual validation strategy enhanced model robustness by ensuring spatial consistency in flow and ET dynamics, addressing equifinality concerns inherent in data-sparse environments. The LULC dynamics were incorporated using decadal historical maps (2000, 2010, 2020) derived from Landsat satellite imagery, while future scenarios (2030, 2040, 2050) were projected using the Cellular Automata (CA)-Markov model. Daily meteorological data for the historical were obtained from the Cheranmadevi observatory. The future daily data were reconstructed using Centro Euro-Mediterraneo sui Cambiamenti Climatici—Climate Model (CMCC-CM) project data based on the Representative Common Pathways version 8.5 (RCP 8.5) emission scenario from 2022 to 2050 after removing bias using observatory data. The WBCs were simulated by forcing historical and future LULC and meteorological conditions at monthly and annual scale. Meteorological extremes recurred every 10–15 years across the catchment. During December 2005 and November 2030, over 60% of monthly rainfall was converted to surface runoff, reflecting heightened flood risks. The AE exhibited strong seasonal variability and a 3–4 years cyclic pattern, with projections suggesting marginally increased contributions in future periods due to warming and LULC changes. Approximately 70% of annual rainfall partitioned into runoff and AE, indicating chronic water losses that amplify drought and flood vulnerabilities. Rainfall patterns showed a rapid response in surface runoff, percolation, and lateral flow, while SW and AE exhibited lagged hydrological responses, adding complexity to water resource planning. These insights aid policymakers, researchers, and practitioners in strategizing climate adaptation and sustainable management of water and land resources under shifting climatic and anthropogenic pressures.