Climate change increasingly threatens water resources in semi-arid, rain-fed regions such as Nigeria's Benue River Basin. This study evaluates climate-induced variability in river discharge and groundwater recharge within the basin using the physically based ArcSWAT model to assess climate-driven variability in monthly river discharge from 1990 to 2024, integrating terrain, land cover, soils, and multi-source climate data, supported by Google Earth Engine remote sensing. Rigorous calibration (1990-2000) and validation (2001-2012) using the SUFI-2 algorithm confirmed strong model performance (R-2 = 0.99/0.85, NSE = 0.98/0.82), effectively capturing topography-driven runoff, soil-water interactions, and evapotranspiration dynamics. Results reveal pronounced seasonal contrasts, with peak discharges in August (145 m(3)/s) and extremely low flows (< 2 m(3)/s) during February-May, intensifying dry-season water stress. A marked January-July flow decline indicates shifts in atmospheric-hydrologic linkages. Spatial analysis shows greater discharge losses in upstream forested sub-basins than in downstream zones. Climate projections under RCP4.5 and RCP8.5 suggest mean annual streamflow reductions of 11.1% and 18.5%, with dry-season declines reaching 25%. Coupling CA-Markov land use simulations with CMIP6 ensemble projections enhanced ArcSWAT's forecasting accuracy under future scenarios. Combined land climate impacts led to up to 30% dry-season flow reduction and increased hydrological variability across sub-basins. As one of the few physically based long-term assessments in West Africa, the study underscores the compounded effects of land use and climate change on water resources. Urgent adaptive strategies such as aquifer recharge, climate-smart irrigation, and decentralized water storage are recommended. Future research should integrate groundwater and socio-economic water-use modeling to better inform resilient, sustainable basin-scale water management.