Understanding the Zonal Variability in CMIP6 Projections of Sahelian Precipitation

The uncertainty in model projections of future precipitation across the Sahel has persisted across many generations of Earth System Models (ESMs), with some models predicting drying and others moistening across this region. These discrepancies in future projections pose a challenge for stakeholders and decision makers. Many projections of Sahel precipitation found in the ESMs participating in the sixth Phase of Coupled Model Intercomparison Project (CMIP6) show a zonal dipole in the sign of precipitation trend, with moistening across the Central and Eastern Sahel and drying projected for the Western Sahel. Previous studies have connected precipitation variability across the Sahel to changes in various factors, but particularly in sea surface temperatures (SSTs) and the warming of the land surface at higher latitudes. However, it remains unclear why some models produce the zonal dipole and others do not. Furthermore, for models that produce the dipole, why some transition from drying to moistening farther to the east than others remains an open question.

In this study, we seek to understand why some of these models produce this zonal contrast and others do not, along with to investigate and better understand the variability in the transition region for models that produce the dipole. To do this, we analyze the output from 43 state-of-the-arts CMIP6 models. Overall, twenty-nine (29) models produce a dipole pattern in precipitation change across Sahel from the 20^th to the 21^st century, while the remaining fourteen (14) models do not. Composite analysis shows that models that produce the dipole pattern simulate the shift of the Atlantic Intertropical Convergence Zone (ITCZ) farther to the south, while models that do not produce the dipole pattern shift the Atlantic ITCZ relatively to the north. An investigation of the relationship between the change in Sahel precipitation and SST based indices for the zonal and non-zonal models show that the strong relationship between these indices is driven by the non-zonal models. Finally, we propose an index that allows us to quantify where the east-west transition from drying to moistening occurs in each of the models. We found that regions, where these models transition from drying to moistening explains a large portion of the uncertainty in the total precipitation change across the Sahel. However, the proposed indices are not strongly correlated with most of the SST based indices. Thus, suggesting that fully understanding this zonal variability over the Sahel must go beyond the SST focused analysis of most previous studies.