Water Resources Modeling in Data-Scarce Watersheds: Contribution of the SWAT Model and the SUFI2 Algorithm to the Study of the Thiokoye River Basin

Effective management of water resources in West African basins is increasingly hindered by sparse hydrometeorological datasets, high interannual rainfall variability, and nonlinear feedbacks among climate, land use, and hydrological processes. To address these constraints, this study implemented the Soil and Water Assessment Tool (SWAT) in conjunction with the SUFI-2 (Sequential Uncertainty Fitting v2) algorithm in the Thiokoye River Basin—an understudied sub-catchment of the Gambia River situated at the interface of Southern Sudanian and Guinean climatic zones. The model was calibrated (1979-1992) and validated (1998-2002) using multiple performance metrics (NSE, R2, KGE, PBIAS) and uncertainty indicators (p-factor, r-factor). Calibration yielded excellent results (NSE = 0.98, R2 = 0.98, p-factor = 0.90), with similarly robust validation scores (NSE = 0.82, R2 = 0.83, KGE = 0.81, PBIAS = –13.2%). Hydrological simulations showed strong seasonal partitioning, with blue water fluxes—surface runoff, percolation, and water yield—constrained to the May-October wet season, representing only 22% of annual rainfall. Conversely, green water availability remained low year-round (mean ET/PET = 0.19), suggesting persistent vegetative water stress. Sediment modeling indicated acute erosion risk, with over 96% of annual soil loss (15.5 t/ha) occurring during peak rainfall periods. Sensitivity analysis revealed CN2, ALPHA_BF, GW_DELAY, and CH_K2 as dominant controls on runoff and groundwater dynamics. These results underscore the utility of the SWAT-SUFI-2 framework for simulating hydrological and sediment processes in data-limited tropical basins and informing targeted strategies for watershed management and climate resilience.