Challenges and Solutions for Simulating Drought-Caused Tree Mortality in Earth System Models

Our prediction of future climate by Earth System Models (ESMs) greatly relies on how accurate we simulate the dynamics of terrestrial vegetation, which historically absorbed ~30% of anthropogenic fossil fuel carbon release. Unfortunately, recent studies have shown that current ESMs substantially underestimated the forest responses to drought. This underestimation is likely resulted from the fact that ESMs utilize carbon-based metrics of plant health to predict mortality, while our recent findings show that the universal signal of drought-caused tree mortality lies in the plant hydraulic function failure. To better predict vegetation dynamics under drought and vegetation feedbacks to climate, we developed a global model of plant hydraulic failure within a land model of DOE-Sponsored ESM, CLM(ED). The hydraulic model has been evaluated at our site level observations in New Mexico and will be evaluated against remote sensing estimates at regional to global scales. We expect that our hydraulic failure model will substantially improve the ESM performance in predicting global cycle response to drought.