Evaluating QDM and CDF-t Bias Correction for CMIP6 Precipitation Projections and Hydroclimatic Extremes over West Africa

Earth System Models (ESMs) and General Circulation Models (GCMs) from the Coupled Model Intercomparison Project Phase 6 (CMIP6) still exhibit substantial biases over West Africa, particularly for daily precipitation extremes. This study evaluates raw CMIP6 simulations and simulations bias-corrected with the Quantile Delta Mapping (QDM) and CDF-transform (CDF-t) methods, for mean precipitation and two extreme indices: consecutive dry days (CDD) and total precipitation from very wet days (R95pTOT). Three observational datasets (CHIRPS, TAMSAT, MSWEP) are used over 1983-2012 for the West African Monsoon season (JJAS), with projections for 2071-2100 under SSP5-8.5. Raw simulations reproduce the main precipitation gradients but show considerable biases and inter-model spread, especially over the Guinean coast and Sahel. Both methods considerably improve performance, increasing spatial correlations and bringing normalized standard deviations closer to unity. CDF-t performs slightly better for mean precipitation and CDD, with comparable performance for R95pTOT; no single method is optimal for all indices. Projections indicate wetter conditions over the central and eastern Sahel, longer dry spells over the western Sahel, and a widespread intensification of extreme rainfall. Based on the Wilcoxon signed-rank test, this intensification is significant mainly at the ensemble level, while CDD changes show the weakest significance.