Changes in extreme precipitation in East Africa and Mount Kenya based on high-resolution regional climate model simulations for the end of the 21st century

The climate in Africa is very diverse ranging from tropical rainforest to deserts. Also, East Africa is covered by different climate zones and is very dry compared to other tropical regions. This is owed to various large-scale drivers, such as the complex topography, large water bodies such as Lake Victoria and vicinity to the Indian Ocean. The southern part of East Africa is characterized by two rainy seasons, which are separated by dry periods. The long rains from March to May feature more continuous precipitation, while the short rains from October to November show high interannual variability with days of high precipitation intensities and drier intervals.

The CMIP5 and CMIP6 models project a general wetting of East Africa in the future, with a high model agreement. To obtain a better understanding of what this means for extreme precipitation and changes in the hydrological cycle we performed three different regional downscaling simulations using WRF: one for the present period from 1981–2010, and two for the end of the century (2071–2100). The latter two simulations are driven by, the RCP2.6 and the RCP8.5 scenarios, and the respective global forcing fields are based on CESM model runs. The regional model covers four different domains, whereby the first extends from the Sahara down to Madagascar with 27 km horizontal resolution, the second domain focuses on East Africa with 9 km resolution, the third domain at 3 km resolution zooms into the western part of Kenya, covering land with complex topography, and the last domain centers on Mount Kenya and surroundings at 1 km resolution.

Preliminary results show that the rainy seasons are difficult to capture by WRF, when driven by a global climate model. This might be related to the fact that some of the atmospheric circulation is misrepresented in the global model and cannot be corrected by the regional model dynamics. While the long rains are underestimated in the present compared to a downscaling of ERA5, the short rains show an overestimation. A sensitivity study with adjusted SSTs to overcome some of the circulation issues in the global climate model only weakly improves the results. The projections for the future show an increase in extreme precipitation days, but also in the extreme daily precipitation amounts compared to present extreme (p99) precipitation. While the rainy seasons are projected to be more intense, the dry seasons tend to become drier, leaving some months without precipitation at all. The results further suggest that the extreme precipitation events do not differ for the RCP2.6 and RCP8.5. Thus, extreme precipitation events in Kenya might be limited by an upper bound, but this is subject of ongoing research.