Kouakou C. auteur aut IRD Paturel Jean-Emmanuel auteur aut IRD Satgé Frédéric auteur aut IRD Tramblay Yves auteur aut IRD Defrance Dimitri auteur aut IRD Rouché Nathalie auteur aut IRD text journalArticle eng text/pdf reformatted digital access Study region: Data-scarce basins located in West Africa and northern Central Africa. Study focus: Multiple studies have shown that global gridded precipitation datasets could provide an alternative to the lack of observed data in Sub-Saharan Africa. This work evaluated 15 precipitation datasets based on satellite rainfall (ARC v.2, CHIRP v.2, CHIRPS v.2, PERSIANN-CDR, MSWEP v2.2 and TAMSAT v3), reanalysis (ERA5, JRA-55 Adj, MERRA-2 PRECTOT, MERRA-2 PRECTOTCORR, WFDEI-CRU and WFDEI-GPCC) and ground measurements (CPC v.1, CRU TS v.4.00 and GPCC v.7), as well as a regional estimation method, based on spatial proximity, for the parameters of a simple monthly water balance model, GR2M. The regional simulations of the GR2M model were evaluated based on a Kling-Gupta Efficiency score in a split-sample spatiotemporal validation scheme. New hydrological insights for the region: The results show that among all the precipitation products, CHIRPS is the most effective for hydrological modeling in West and Central Africa at a monthly timestep. Also, among the top five products are WFDEI-CRU, CRU, WFDEI-GPCC and GPCC. Overall, regional hydrological modeling is more effective for basins smaller than 80,000 km2. The method of regionalization by spatial proximity causes an overall drop in the ability of the various precipitation products to reproduce discharge, most notably with WFDEI-GPCC and GPCC. CHIRPS remains the best product in terms of KGE2 values in regionalization. specialized Gridded precipitation estimates GR2M West Africa Central Africa Re-analysis Observed data AFRIQUE DE L'OUEST AFRIQUE CENTRALE 062 126 020 47 101409 [23 p.] 2023 2214-5818 https://www.documentation.ird.fr/hor/fdi:010090090 10.1016/j.ejrh.2023.101409 2214-5818 [F B010090090] https://www.documentation.ird.fr/hor/fdi:010090090 https://horizon.documentation.ird.fr/exl-doc/pleins_textes/2023-09/010090090.pdf IRD - Base Horizon / Pleins textes 2023-09-22 2025-02-24 fdi:010090090 fre