Coupled model simulations of extreme rainfall events over Africa coupled models

The ability of coupled climate models from the national multi-model ensemble (NMME) dataset to reproduce the basic state and interannual variability of precipitation in Sub-Saharan Africa and associated teleconnections is examined. The analysis is for the period 1982-2010 for most of the models, which corresponds to the NMME hindcast period, except for the CFS version 1 (CFSv1) which covers the period 1981-2009. The satellite based Global Precipitation Climatology Project (GPCP) data is used as proxy for observed rainfall and to validate the models. Sub-Sahara Africa lies across both the northern and southern hemispheres, and therefore its climate in most regions is dependent on the apparent seasonal march of the sun. We examine rainfall patterns during the four main seasons in DJF, MAM, JAS, and OND. Models are able to reproduce the northward migration of precipitation from winter when the area of maximum precipitation is located in southern Africa to the summer when it is in northern Sub-Saharan Africa and the later return to the south. Models also appropriately place precipitation over central Africa during the equinoxes in MAM and OND. However, there are considerable differences in the representation of the intensities and locations of the rainfall. Almost all the models have a dry bias in eastern southern Africa during the DJF season. Three of the models including the two versions of the NCEP CFS and the NASA models also have a systematic dry (wet) bias over the Sahel (Gulf of Guinea region) during the summer rainfall season, while the others show alternating wet and dry biases across West Africa. The models also exhibit a systematic dry bias in equatorial eastern Africa during MAM and OND. All models have spatially averaged values of standard deviation lower than that observed. Models are also able to reproduce to some extent the main features of the precipitation variability maximum, but again with deficiencies in the amplitudes and locations. The areas of highest variability are generally depicted, but there are significant differences among the models, and even between the two versions of the CFS. Teleconnections in the models are investigated by first conducting an EOF in the precipitation anomaly fields and then perform a regression of the first or second EOF time series onto the global SST. Focusing on JAS rainfall season, only the CFSv1 and the NASA models were able to depict the dipole pattern between the Sahel and the Gulf of Guinea rainfall. However, none of the models was able to reproduce the observed upward trend and high frequency rainfall episodes over the Sahel in the last decade. The relationship to SST is also examined. The observed influence of tropical north Atlantic SST on the Sahel rainfall is only partially represented even in the CFSv1, while the NASA model inconsistently emphasizes the role of the tropical South Atlantic. A majority of the models show a partial ENSO teleconnection combined with the tropical south Atlantic mode. However, observations indicate that the influence of ENSO on northern Sub-Saharan summer rainfall has been very weak over the past 30 years. Results for DJF, MAM, and OND are also presented.