Evaluating CMIP6 simulations of historical Sahelian precipitation variability and potential links to future projections

Abstract

The Intergovernmental Panel on Climate Change has recognised the Sahel region in Africa as a hotspot for the effects of climate change due to its economic and environmental vulnerability to changes. Its climate is driven by the West African Monsoon (WAM) which occurs during the June to September season, accounting for ~75% of annual precipitation. Modelling the WAM has many associated uncertainties, predominantly linked to precipitation and its occurrence on the sub-grid-scale. Previous projections for the Sahel have covered both extremes of wetting and drying, complicating policymaking, and reinforcing the need for confidence in projections. This study has examined the output from the sixth phase of the Coupled Model Intercomparison Project for monsoon season precipitation over the Sahel. Models were initially evaluated by comparing historical simulations to observations. No model well represented the wet period of the 1950s and the dry period of the 1970s. However, atmosphere-only models using prescribed sea surface temperatures performed much better with considerably higher correlation coefficients at decadal time scales. This suggests that representing ocean processes (and their coupling) remains a major limitation on simulating decadal precipitation over the Sahel. Projection anomalies revealed an inter-model range of +2mm/day to -1mm/day for SSP1-2.6 which increased to +4mm/day to -2mm/day for SSP5-8.5. Following this, these projections were analysed in the context of historical model performance. There is no clear link between historical model performance and their respective projections, but historical model behaviour partially explains the differences in projections. However, this study highlights the range of potential impacts for the Sahel should global warming not be limited. Although, there is no guarantee the climate system will behave the same way as it did in the past, improving historical model performance, specifically ocean representation and its coupling, could add confidence to projections for the Sahel.

Date of Award	12 May 2022
Original language	English
Awarding Institution	University of Bristol
Supervisor	Paul J Valdes (Supervisor) & Rachel James (Supervisor)