Hydro-climatic ecosystem risk mapping and spatial propagation analysis in the Upper Niger Basin, West Africa

Hydro-climatic variability threatens ecosystem stability and hydrological functioning across tropical river basins in West Africa. This study developed a remote-sensing and graph-based spatial-propagation framework to assess hydro-climatic ecosystem risk in the Upper Niger Basin, West Africa, during 2001–2024. CHIRPS rainfall, Landsat Collection 2 Level-2 land surface temperature, MODIS NDVI and evapotranspiration, SRTM terrain, and MERIT Hydro drainage-area data were processed in Google Earth Engine. Five normalised stress components (thermal, terrain-weighted runoff-response, moisture-balance, flow-connectivity, and hydrological stress) were combined into an equal-weighted Hydro-Climatic Ecosystem Risk Index (HCERI). NDVI and NDWI were retained as diagnostic indicators rather than direct HCERI components. The basin showed marked hydro-climatic heterogeneity. Rainfall was positively associated with NDVI (r = 0.919) and evapotranspiration (r = 0.965), whereas land surface temperature was negatively associated with NDVI (r = −0.879) and evapotranspiration (r = −0.878). Mann–Kendall tests identified no significant basin-wide trends except declining NDWI ( p = 0.0019). High-risk and Moderate-risk areas covered 56.59% and 38.93% of the basin, respectively. Under 1000 perturbed-weight simulations, equal-weight HCERI showed median Spearman correlation of 0.980, class agreement of 83.52%, Cohen’s kappa of 0.693, and High/Very High-risk overlap of 0.812; entropy weighting produced a distinct classification. Graph-based propagation using 3000 nodes and 13,686 edges preserved HCERI structure (R² = 0.999999; RMSE = 0.000090; MAE = 0.000065), reflecting propagation consistency rather than predictive validation. Temporal robustness was high across four six-year periods, and Gi* analysis confirmed spatial clustering. The framework supports adaptation, watershed management, and conservation.