As global warming intensifies, terrestrial ecosystems are increasingly confronting compound climate extremes. Predicting how these ecosystems will respond—whether they will survive or collapse—remains a major challenge in Earth system science.

A new study published recently in AGU Advances has uncovered a critical "personality switch" in ecosystem responses to such extremes. Led by Prof. DAN Li from the Institute of Atmospheric Physics of the Chinese Academy of Sciences, the research team found that the impact of climate extremes extends beyond immediate stressors; it is fundamentally governed by "hydrological memory"—the pre-existing soil moisture conditions.

The study focused on the low-latitude highlands of Southwest China, a biodiversity hotspot renowned for its sensitivity to climate fluctuations. The team conducted a "natural experiment" by comparing two distinct extreme events: the severe 2010 winter-spring drought and the record-breaking 2019 heatwave. Though both events occurred against similar large-scale atmospheric circulation backgrounds, their impacts on local ecosystems differed drastically.

"Ecosystems, much like humans, react to stress differently based on their prior state," explained PAN Wei, first author of the study.

The researchers documented that during the 2010 drought, the ecosystem experienced a systemic collapse in carbon uptake, measured as Gross Primary Production (GPP). In contrast, during the 2019 heatwave—despite even higher temperatures—the vegetation remained resilient and even thrived in some regions.

Why did the ecosystem collapse in 2010 but withstand the 2019 heatwave? The study identifies antecedent root-zone soil moisture as the "master switch" driving these divergent outcomes.

In 2010 (Water-Limited Paradigm): Soil moisture was already depleted prior to the drought. This lack of hydrological memory pushed the ecosystem to a physiological threshold, triggering widespread suppression of photosynthesis.

In 2019 (Energy-Driven Paradigm): The months leading up to the heatwave were characterized byrainfall, fostering abundant soil moisture. This accumulated hydrological memory acted as a buffer, enabling vegetation to access stored soil water. Consequently, high temperatures—rather than scorching the plants—enhanced their activity by providing additional energy for photosynthesis.

The findings highlight a phenomenon the researchers term a "state-dependent mechanistic shift": the ecosystem’s response pathway switches from water-limited to energy-driven, contingent entirely on its hydrological history.

This discovery challenges the static assumptions embedded in many current Earth System Models (ESMs). These models often fail to capture such nonlinear transitions, introducing uncertainties into predictions of future carbon sinks.

"Our results emphasize that geography and history are not merely backdrops; they are key players," said Prof. DAN. "Incorporating hydrological memory and state-dependent nonlinearity into climate models is crucial for reducing uncertainties in climate projections and accurately assessing ecosystem vulnerability in a warming world."