Radiation and temperature dominate the spatiotemporal variability in resilience of subtropical evergreen forests in China

Forest resilience is crucial to the mitigation of climate change, due to the enormous potential of forests to reduce atmospheric carbon dioxide concentrations and the possible conversion of forests from net carbon sinks into carbon sources following external disturbances. Subtropical forests are suffering the highest rates of forest change, but how they are evolving in response to climate change is little known. In this study, we estimated the spatial pattern and temporal trend of the resilience of subtropical evergreen forests in China by applying the lag-one autocorrelation (AC1) method to satellite kernel normalized difference vegetation index (kNDVI) data over the past two decades and identified the influential environmental factors that affect the ecosystem resilience by developing random forest (RF) regression models. The computed long-term AC1 based on kNDVI for the 2001-2020 period depicts considerable spatial variability in the resilience of the subtropical evergreen forests in China, with lower resilience at lower latitudes. The RF regression analysis suggests that the spatial variability in the forest resilience can be re-established by forest and climatic variables, and is largely affected by climate, with the three most influential variables being solar radiation (SR, %incMSE = 20.7 +/- 1.8%), vapor pressure deficit (VPD, %incMSE = 13.8 +/- 0.2%) and minimum temperature (T-min, %incMSE = 13.3 +/- 1.2%). Higher forest resilience is more likely to be located in areas with less radiation stress, adequate water availability, and less warming. Trend analysis shows a declining trend for the resilience of subtropical evergreen forests in China since the 2000s but an increasing forest resilience in the last decade, which is mainly dominated by temperature changes, including average and minimum temperatures. Considering the expected warming-dominated period in times of rapid climatic change, we suggest potential critical responses for subtropical forest productivity to the disturbances should be of greater concern in the future.