资源环境科技发展态势分析平台

The rain gage networks in the African countries are notorious for their poor density and low frequency of observations. Rainfall products based on satellite estimates and/or ground observations have proven to be a viable alternative in the recent evolving campaigns to overcome this deficiency. The newly-borne country within the Nile Basin, South Sudan, has extremely few operating rain gage stations. Herein, evaluation of six long-term (1983 onward) rainfall products, i.e. the Global Precipitation Climatology Centre full data reanalysis version 7.0 (GPCC 7.0) and five Satellite-Rainfall Products (SRPs), is undertaken. Data from the only currently operating long-term stations (five) with reasonably up-to-date records are used to conduct point-to-pixel evaluation for the six products (from 1983 to 2010). The results of error and linear fit metrics rank GPCC 7.0 as the best performing product on monthly, maximum monthly, and annual scales, followed by Climate Hazards group Infrared Precipitation with Stations version 2.0 (CHIRPS v2.0). As regards the variability of annual rainfall, GPCC 7.0 outperforms the products whereas the Precipitation Estimation from Remotely Sensed Information Using Artificial Neural Networks Climate Data Record (PERSIANN-CDR) and CHIRPS v2.0 have the second-best performance. GPCC 7.0 and the Multi-Source Weighted-Ensemble Precipitation version 2.0 (MSWEP 2.0) show better agreement of variability of monthly rainfall with that of the station rainfall. The Africa Rainfall Climatology version 2.0 (ARC2) performs the best in capturing the variability in the maximum monthly rainfall followed by GPCC 7.0. In relation to capturing the median rainfall, a complex performance is evident across the stations with the following remarks: relatively good performance from MSWEP 2.0 on the annual scale followed by PERSIANN-CDR; GPCC 7.0 (PERSIANN-CDR) is mostly (relatively) operational for the median maximum monthly rainfall; GPCC 7.0 (CHIRSPS v2.0) is mostly (partly) suitable as an estimator for the monthly median. All the present SRPs unequivocally under-estimate the monthly peaks. Enhancing the rainfall estimation and observation network is key to improving the understanding and modeling of the hydrological processes and phenomena occurring in the basin in general and in the floodplains of the country in particular.