Using causal networks to constrain regional drought projections

Severe droughts in the Cape Town region (CTR) are projected to become more frequent in the coming decades, posing significant societal challenges. However, while climate models consistently predict a precipitation decline for the CTR until the end of this century, these projections carry substantial uncertainties, with decreases ranging from almost zero to as much as -50%.

In this study, we employ causal networks to evaluate climate models based on their ability to accurately represent the large-scale dynamical processes that drive precipitation in the CTR. While previous research has identified links between precipitation in the CTR  and various large-scale drivers, such as the eddy-driven jet and sea surface temperatures in the South Atlantic, the interactions between these drivers remain poorly understood and the relative contributions of individual drivers to precipitation in the CTR remain unexplored.

Following causal inference theory, the causal relationships among the large-scale drivers of precipitation in the CTR are quantified in reanalysis data, pinpointing the main precipitation drivers, their interactions and relative contributions to precipitation and drought events. The resulting causal network is then applied to constrain precipitation projection. The study’s insights into the links between planetary-scale circulation patterns and regional processes could enhance our understanding of extreme and compound events, with potential implications for drought management.