Inter-comparison and validation against in-situ measurements of satellite estimates of incoming solar radiation for Central Africa: From the annual means to the diurnal cycles

• Published in: Atmospheric research Vol. 287; p. 106711
• Main Authors: Ouhechou, Amine, Philippon, Nathalie, Morel, Béatrice, Trentmann, Jörg, Graillet, Alexandre, Mariscal, Armand, Nouvellon, Yann
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 15.05.2023; Elsevier
• Subjects: Atmospheric and Oceanic Physics; Central Africa; Diurnal cycle; Physics; Satellite estimates; Solar radiation; Diurnal cycle; Satellite estimates; Solar radiation; Central Africa
• ISSN: 0169-8095; 1873-2895
• DOI: 10.1016/j.atmosres.2023.106711

This study pictures for the first time incoming solar radiation mean evolution in Central Africa, intercomparing 8 gridded products (namely CERES-EBAF, CERES-SYN1deg, TPDC, CMSAF SARAH-2, CMSAF CLARA-A2, CAMS-JADE satellite products, as well as ERA5 reanalysis and WorldClim 2 interpolated measurements) and station-based estimations (FAOCLIM 2) or measurements. At the mean annual scale, all products picture low levels of global horizontal irradiance (GHI) to the west (SW Cameroon to SW Republic of Congo) and higher levels towards the north and south margins of the region. However, GHI levels in the CMSAF products are much higher than in CERES and TPDC. The mean annual cycles of GHI extracted for 6 sub-regions are bimodal, with two maxima during the two rainy seasons (March–May and September–November) and two minima during the two dry seasons (December–February and June–August). These seasonal cycles are well reproduced by most products except their amplitude which is dampened in TPDC. At the daily and sub-daily time-scales, products were compared with in-situ measurements from ten meteorological stations located in the western part of Central Africa. The products' performance is assessed through scores as bias and RMSE but also by considering the diurnal cycles' shape, amplitude and frequency of occurrence along the annual cycle. The products properly reproduce the shape of the four types of diurnal cycles with nonetheless noticeable differences in the cycle's frequencies of occurrence. •We evaluated eight solar radiation products for Central Africa against in-situ data.•Most products overestimate solar radiation, especially SARAH-2.•Satellite estimates perform better than ERA5 on interannual and sub-daily timescales.•Products accuracy was also assessed considering the diurnal cycle approach.