Predicting drought-induced tree mortality in the western United States

Abstract:

Projected responses of forest ecosystems to warming and drying associated with 21^st century climate change vary widely from resiliency to widespread dieback. A major shortcoming of current vegetation models is the inability to account for mortality of overstory trees during extreme drought due to uncertainties in mechanisms and thresholds. In this talk, I discuss two modeling efforts to predict drought-induced tree mortality in the western United States. In the first, we identify a lethal drought threshold in the loss of vascular transport capacity from xylem cavitation, which provides insight into what initiates mortality, in Populus tremuloides in the southwestern United States. We then use the hydraulic-based threshold to produce a hindcast of a drought-induced forest dieback and compare predictions against three independent regional mortality datasets. The hydraulic threshold predicted major regional patterns of tree mortality with high accuracy based on field plots and mortality maps derived from Landsat imagery. Climate model simulations project increasing drought stress in this region that exceeds the observed mortality threshold in the high emissions scenario by the 2050s, likely triggering further widespread diebacks. In the second approach, we build a dynamic plant hydraulic model into a land-surface model and compare predictions against observed mortality patterns across multiple species. These methods provide powerful and tractable approaches for incorporating tree mortality into vegetation models to resolve uncertainty over the fate of forest ecosystems in a changing climate.