Understanding intermodel variability in future projections of a sahelian storm proxy and southern saharan warming

Abstract

Projected changes in the intensity of severe rain events over the North African Sahel-falling from large mesoscale convective systems-cannot be directly assessed from global climate models due to their inadequate resolution and parameterization of convection. Instead, the large-scale atmospheric drivers of these storms must be analyzed. Here we study changes in meridional lower-tropospheric temperature gradient across the Sahel (DTGrad), which affect storm development via zonal vertical wind shear and Saharan air layer characteristics. Projected changes in DTGrad vary substantially among models, adversely affecting planning decisions that need to be resilient to adverse risks, such as increased flooding. This study seeks to understand the causes of these projection uncertainties and finds three key drivers. The first is intermodel variability in remote warming, which has strongest impact on the eastern Sahel, decaying toward the west. Second, and most important, a warming-advection-circulation feedback in a narrow band along the southern Sahara varies in strength between models. Third, variations in southern Saharan evaporative anomalies weakly affect DTGrad, although for an outlier model these are sufficiently substantive to reduce warming here to below that of the global mean. Together these uncertain mechanisms lead to uncertain southern Saharan/northern Sahelian warming, causing the bulk of large intermodel variations in DTGrad. In the southern Sahel, a local negative feedback limits the contribution to uncertainties in DTGrad. This new knowledge of DTGrad projection uncertainties provides understanding that can be used, in combination with further research, to constrain projections of severe Sahelian storm activity.