Drought-related forest mortality is a widespread phenomena and is expected to increase with continued climate warming. Recent research has focused on mechanisms that underlie drought-related tree mortality and has identified both endogenous (carbon starvation and hydraulic failure) and exogenous factors (insect and disease) that lead to death. In this paper, we show that a mechanistic model of carbon starvation can be accurately estimate spatial differences in drought stress mortality for a Ponderosa Pine site in New Mexico. Estimates of net primary productivity and mortality from a coupled hydrologic and ecosystem carbon cycling model (RHESSys) are compared with observed measurements for the 3 plots that cross an elevational gradient. We assess how well the model represents differences among these plots and in particular the greater response to moisture stress in the lower elevation plot that led to a stand level mortality event in the early 2000s. We then use this model implementation to investigate the relative importance of different environmental controls on productivity and mortality across the elevational gradient and how this might change under a warmer climate.
Results/Conclusions
Results show that for a fairly wide set of parameter assumptions, the model accurately estimates spatial differences and inter-annual variation in productivity in both drought and non-drought periods. The model also accurately captures spatial differences in Ponderosa Pine mortality during the early 2000s drought. We show that if the 2000s drought had been 2C warmer, mortality would likely have extended to the mid-high elevation sites. Our results demonstrate the potential for a coupled hydrology-ecosystem mechanistic model that includes a simple model of carbohydrate storage to predict stand to watershed scale mortality patterns related to carbon starvation. While drought stress related mortality in many cases may be due to other exogeneous factors (insects/disease), reduced productivity and carbohydrate storage reserves likely to contribute to susceptibility and thus the model could be used to estimate this risk and its variation across space and climate scenarios.