Drought-induced forest mortality has major consequences for biodiversity and ecosystem services, and could result in a substantial positive carbon-climate feedback by releasing large amounts of carbon into the atmosphere within a relatively short period, which would cause further climate change. Critically, the physiological mechanism of two widespread patters, that (1) after a serious drought, trees are observed to grow normally for several years and then die and (2) large trees are more likely to die from drought than small trees, are not fully understood. We used a tree-level photosynthesis model dependent on leaf, functional xylem, soil water, atmospheric vapor pressure deficit, and atmospheric CO2, and optimality theory to understand the mechanisms governing the success and failure of tree recovery from drought-induced catastrophic xylem embolism. We simulated tree recovery from different degrees of xylem embolism for different size classes under optimal and unexpectedly dry conditions. We further examined how projected changes in atmospheric CO2 and climate impacted post-drought tree recovery. Tree model results were compared to observed drought mortality windows and global maps of productivity recovery times.
Results/Conclusions
We show that the dominant patterns associated with drought mortality can be explained by optimal carbon allocation in trees whose xylem has been damaged by drought. This occurs because severe drought irreparably reduces the ability of xylem to transport water, which requires stomatal closure that reduces photosynthesis, and results in negative carbon balance if the xylem damage is serious. We show that normal diameter growth during the post-drought period is an optimal strategy to potentially avoid mortality because of the critical need to build new xylem. Multi-year time lags before death are expected because trees with marginally negative post-drought carbon gain take years to starve to death while attempting to rebuild lost xylem. Big trees die at higher rates than small trees because the number of years of stem growth required to rebuild a healthy tree’s xylem and the amount of carbon lost to cambium and phloem respiration both increase with tree size. Forecasts of elevated future drought-kill conclude that increases in water use efficiency caused by CO2 fertilization will not be enough to compensate for increased future drought caused by elevated temperatures. However, because the optimal allocation hypothesis explains drought kill as carbon starvation induced by xylem damage, we show that CO2 fertilization should also reduce future drought kill because it increases photosynthesis, independent of its effect on in water use efficiency.