Information flow causality reveals climate regime shifts during compound dry–hot events

The intensification in the frequency and severity of compound dry–hot events (CDHEs) threatens ecosystem stability worldwide. While land–atmosphere coupling has been recognized as a key driver of CDHE evolution, the causal dynamics between land surface and atmospheric variables under compound extreme conditions remain insufficiently understood. Since the variability of mass and energy transfers in the climate system can be explained as the function of information fluxes, this study applies a causal inference framework based on information flow (IF) to quantify the influences of soil water content (SM) and shortwave radiation (R) on near-surface air temperature (T) during CDHEs. To this end, simulation results from the fully coupled regional climate system model Terrestrial System Modeling Platform over the pan-European model domain region are utilised. Our results reveal that the study domain is predominantly energy-limited, characterized by strong R to T coupling, though water-limited regimes prevail in southwest Europe, showing a stronger SM to T coupling. E.g., the 2003 European CDHE was characterized by a widespread regime shift from energy- to water-limited conditions, which reversed, following the precipitation that ended the event. A significant increase in positive IF from SM to T is identified as a key contributor to the extreme nature of the 2003 CDHE, driven by the persistently dry antecedent conditions. These findings ratify the critical role of spatio-temporal land-atmosphere coupling in shaping CDHE trajectories from the perspective of information theory. Moreover, the IF framework is a powerful diagnostic for uncovering causal drivers of extremes, evaluating climate models, and enhancing the predictability of ecosystem responses under compound climate extremes.