Climate Simulation and Projection of Rainfall–Runoff Dynamics Using the GR4J Model in the Oti Sub-Basin: The Case of the Porga, Mandouri and Mango Outlets

Water resource management in the Sahelian-Sudanian transition zone faces growing uncertainty under climate change, yet hydrological projections remain scarce for the Oti-Pendjari basin (West Africa). This study develops an integrated modelling chain combining CMIP6 multi-model evaluation, bias correction, and GR4J hydrological modelling to project streamflow changes under SSP2-4.5 and SSP5-8.5 over 2021–2100. Eleven CMIP6 models were evaluated against ERA5 reanalysis data (1960–2014) using NSE, KGE, and MAE; the three best-performing models were bias-corrected using Linear Scaling, Variance Scaling, Quantile Mapping, and Quantile Delta Mapping. Linear Scaling proved most effective, with CMCC-ESM2 best reproducing observed precipitation (NSE and KGE up to 0.9), while the multi-model approach performed best for temperature. The GR4J model, calibrated at Porga, Mandouri, and Mango (KGE: 0.609–0.668), satisfactorily reproduces intermediate flows and flood dynamics, although structural limitations persist for low flows (KGE [1/Q]: −0.65 to −0.71). Projections reveal a marked divergence between scenarios: SSP2-4.5 yields September peak flow increases of +5.7% to +16.7%, whereas SSP5-8.5 leads to slight decreases of −1.1% to −3.6%, likely driven by increased potential evapotranspiration partially offsetting precipitation gains. These findings underscore the critical importance of scenario selection and model uncertainty in regional water resource planning.