The CausalHeat Project: Land–climate feedbacks shaping ecosystem vulnerability to dry–hot extremes

Grasslands and forests play a central role in regulating terrestrial water, energy, and carbon (WEC) cycles. Vegetation–atmosphere interactions shape and respond to hydroclimatic extremes such as heatwaves, droughts and compound dry–hot events. Under climate change, the influence of these extremes on ecosystem functioning is intensifying, altering the extent to which vegetation modulates WEC fluxes and, in some regions, driving ecosystems toward turning points or functional role reversals. The CausalHeat project aims to characterize the causal dynamics underlying dry–hot extremes, their impacts on ecosystem resilience and vulnerability, and the subsequent ecosystem–climate feedbacks influencing the development and persistence of these events. Using entropy-based information-theoretic causality methods, we quantify the dominant drivers of dry–hot episodes and assess how these drivers propagate through ecohydrologic process networks to influence vegetation structure and function across biomes.

Observational evidence reveals a pronounced biome-dependent divergence in ecosystem responses — more resilient or vulnerable. Forested ecosystems and croplands exhibit strengthened ecohydrologic process coupling and increased network organization, consistent with adaptive reorganization under recurrent drought exposure. However, enhanced vapor pressure deficit (VPD) coupling to forest function and structure yields episodic shocks that can push systems into transiently vulnerable states. In contrast, grassland-, savanna-, shrubland-, and wetland-dominated ecosystems show progressive decoupling of ecohydrologic processes, indicative of potential declining resilience. Grassland ecosystems emerge as particularly sensitive to aridification, with vulnerability driven by the synergistic amplification of atmospheric water demand and declining soil moisture, rather than by the dominance of either factor alone. Together, these results highlight how hot extremes reorganize ecosystem process networks in biome-specific ways, with important implications for terrestrial WEC partitioning, ecosystem stability and ecosystem–atmosphere feedbacks. CausalHeat provides a framework for improving the prediction of dry–hot extremes and assessing ecosystem responses relevant to food and water security under climate change.