What Can We Know About Recent Past Precipitation Over Africa? Daily Characteristics of African Precipitation From a Large Ensemble of Observational Products for Model Evaluation. Issue 10 (1st October 2021)

• Record Type: Journal Article
• Authors: Dosio, Alessandro; Pinto, Izidine; Lennard, Christopher; Sylla, Mouhamadou Bamba; Jack, Christopher; Nikulin, Grigory
• Is Part Of: Earth and space science. Volume 8:Issue 10(2021)
• Page Start: n/a
• Page End: n/a
• Publication Date: 2021-10-01
• Subjects: observations -- Africa -- precipitation daily characteristics -- CORDEX; Space sciences -- Periodicals; Geophysics -- Periodicals; 500.5
• DOI: 10.1029/2020EA001466
• Languages: English
• ISSNs: 2333-5084

Abstract

Abstract: We assess the daily characteristics of recent past precipitation over Africa by means of a large ensemble of observational products, including reanalysis, gauge‐based, and satellite‐based products. The spatial distribution of seasonal mean precipitation varies considerably amongst products especially over areas where gauge networks are sparse. Large uncertainties in the annual precipitation cycle are visible in particular over the Ethiopian Highlands, the eastern Sahel, the coasts of the Gulf of Guinea, and the Horn of Africa. Interannual variability shows large differences especially amongst reanalysis data sets whereas satellite and gauge‐based products usually show more consistent results. Plain Language Summary: Southern Africa and the Atlas region show the largest agreement amongst different products for most of the precipitation characteristics, including annual time series and trends, spatial and temporal correlations, and indices of mean and extreme events. Over other regions, discrepancies amongst different products are larger, especially for indices based on short time scales, in particular, over the regions affected by the monsoon (the Sahel, the coasts of the Gulf of Guinea, and Central Africa). Observed precipitation characteristics have been compared to the results of state‐of‐the‐art regional climate models (RCMs). Given the impossibility to select a single "best" observation data set for a realistic representation of all precipitation characteristics, our results show that comparing model results to a very limited set of observations is not only pointless, but it can be even misleading. When similarly large ensembles of model results and observations are compared, our findings show that RCMs are able to reproduce daily precipitation characteristics as well as the observations. Key Points: Precipitation over Africa is assessed by means of a large ensemble of reanalysis, gauge‐based, and satellite‐based products Southern Africa and the Atlas region show the largest agreement amongst different products for most of the precipitation characteristics Similarly large ensembles of regional models are able to reproduce daily precipitation characteristics as well as the observations