Nonlinear hydrothermal associations between coupled landscape ecological risk and resilience in a major grain-producing region of China.

In major grain-producing regions, hydrothermal processes are critical for regulating ecosystem stability. However, the long-term coupled dynamics between landscape ecological risk (LER) and vegetation-based ecological resilience (ER) remain inadequately quantified, particularly in intensively managed agro-wetland landscapes. Here, we developed a Coupled Risk and Resilience Integration (CRRI) framework that synergistically integrates LER, vegetation-based ER, hydrothermal balance, and interpretable machine learning. Using the Sanjiang Plain as a case study, we revealed that hydrothermal balance exerts contrasting nonlinear effects on LER and ER. Specifically, the ratio of evapotranspiration to potential evapotranspiration (ET/PET) showed opposing nonlinear associations with the two indicators: within an intermediate ET/PET range of approximately 0.50 to 0.60, vegetation-based ER peaked while LER remained comparatively low. By integrating this empirically derived threshold with Gaussian mixture modelling, we identified five distinct risk-resilience coupling states, each reflecting unique hydrological and landscape conditions. These findings establish a region-specific diagnostic framework for understanding risk-resilience coupling dynamics in the Sanjiang Plain and analogous agro-wetland landscapes.