Closing Africa’s water balance budget: Assessing the suitability of eight latest precipitation products

While numerous studies have evaluated precipitation products by comparing them with in-situ observations, the suitability of these products for closing the water balance budget in Africa remains largely unexplored. This study assesses the performance of eight latest precipitation products, namely MSWEP v2.8, CRU v4.07, CHIRPS v2.0, ERA5-Land, CPC, NOAH, IMERG v7, and MSWX v1, in closing the water balance budget over Africa. We begin by assessing the general consistency among these precipitation products using rain-gauge measurements and multiple linear regression analysis, followed by the generalized three-corner-hat (GTCH) method applied to estimate the relative uncertainties and signal-to-noise (SNR) ratios of the eight precipitation products, which help identify the potential causes of the water balance non-closure over Africa. To evaluate the suitability for closing the water balance budget, we compare the Gravity Recovery And Climate Experiment (GRACE) and its follow-on (GRACE-FO, hereafter GRACE) satellite missions’ derived total water storage changes (TWSC) with those of the water balance equation using each precipitation product. The influence of climate zones
and terrain on these evaluations is also examined. The results reveal significant discrepancies in precipitation trends among the eight precipitation products. The GTCH analysis indicates that MSWEP achieves the lowest uncertainties and highest SNR values in most cases, followed by MSWX and CHIPRS. Regarding closing the water balance budget, NOAH exhibits the highest consistency with GRACE-derived TWSC than other products, followed by IMERG, whereas CPC achieved the lowest suitability. The performance of these products appears to
be more influenced by climatic regimes than by terrain. All precipitation products indicate approximately 20% non-closure in the water balance budget over Africa, likely due to the underestimation of annual signals, as revealed by the principal component analysis (PCA). This study underscores that utilizing precipitation derived from the same data source as runoff and evapotranspiration demonstrates high consistency with GRACE-derived TWSC data. These findings suggest that precipitation estimates should merge multi-source datasets and consider the water balance budget, which can help reduce uncertainties and provide a more reliable foundation for hydrological studies in Africa.