How Strongly Are Mean and Extreme Precipitation Coupled?

Abstract

Changes in mean and extreme precipitation are among the most important consequences of climate change. Here, we examine the relationship between the mean and three different measures of extreme precipitation over Australia, from a regional climate model (RCM) ensemble. We show that model uncertainty in mean and extreme precipitation are tightly coupled for both the present-day climate and future changes. On the continental scale, strong correlations (0.79–0.99) are observed across models between the changes in mean and extreme precipitation. We also find that except in eastern coastal regions, precipitation statistics projected by RCM are highly predictable (up to 60%–70% of variance explained) from the mean precipitation of the global climate model (GCM) providing the boundary conditions. In coastal locations where RCMs are more accurate than GCMs for the present climate, they are less predictable, hence provide information—but this impact disappears for climate change, suggesting that improved present-day accuracy may not carry over to climate change.