Modeling drought stress and tree mortality with dynamic global vegetation models

Background/Question/Methods

Increased drought severity due to climate change may cause extensive forest dieback, particularly along biome boundaries such as the ponderosa pine – grassland ecotone that we are studying at Wind Cave National Park (WICA) in the Black Hills of South Dakota, USA. Using the dynamic global vegetation model (DGVM) MC1, we projected future decreases in forest biomass at WICA due to increased fire frequency. However, MC1, like most DGVMs, simulates the direct effects of climate and CO₂ on productivity, but does not explicitly simulate drought-stress processes such as embolism, carbon starvation and insect-mediated death.

 Thus, we implemented a stress index as the ratio of net primary productivity to leaf area index (NPP/LAI), analogous to the vigor index developed by Waring and collaborators in the 1980s to assess the susceptibility of trees to forest pests and pathogens. We parameterized the mortality function based on our stress index to fit the observed ponderosa pine dieback along a dry ecotone in northern New Mexico following the 1950s drought. We then ran 100-year-long future simulations for WICA using statistically downscaled climate projections from the CSIRO Mk3, Hadley CM3 and MIROC 3.2 Medres general circulation models (GCMs), under the A2 emission scenario. 

Results/Conclusions

Applying the parameterized mortality function to WICA, we projected little stress-related mortality until near the end of the 21st century, when one or two dieback episodes occurred for the hottest and driest (MIROC) climate projection. The trees’ stress resistance was due in part to the simulated enhancement of NPP under high future CO₂ levels, based on results from long-term free air CO₂ enhancement (FACE) experiments. However, CO₂ effects are uncertain, as there are no FACE experiments with ponderosa pine. Varying the magnitude of the CO₂ enhancement effect in our model modulated our results, with substantially greater mortality projected when CO₂ effects on production and water use efficiency were turned off. Greater dieback of ponderosa pine is expected in the southwestern USA, where ensemble GCM results project substantially greater heat and drought stress than in the Northern Great Plains region that includes WICA.