PS 20-196 - Evaluating eastern U.S. forest response to climate change using a water balance approach

Background/Question/Methods

Climate change is expected to cause species-specific changes in tree growth and mortality, and ultimately lead to shifts in their distributions. At fine spatial scales, these changes could manifest along topographic gradients, which produce varied microclimates. Our objective in this study is to understand fine-scale vegetation-site relationships under contemporary climate, which would then enable us to model potential shifts under altered climate scenarios. Using a water balance approach, we modelled inter-annual variations in moisture demand and availability for each 3-meter pixel on 430 National Park Service plots in Maryland, West Virginia, Virginia, and the District of Columbia. Individual tree locations were linked to summer climatic moisture deficit for the driest year each park experienced 1998-2020; a “uniform year” across parks was also simulated using the most severe drought that any single park experienced. To assess whether species showed affinities for particular settings, we employed threshold indicator taxa analysis (TITAN) to detect changes in species distributions along a gradient of modeled climatic water deficit.

Results/Conclusions

Using threshold indicator taxa analysis we found a significant response of tree distribution and abundance to water deficit. Common species were not as restricted along the moisture-deficit gradient as uncommon species, but they did favor particular ends of the gradient. However, results were confounded by distinctive vegetation assemblages at particular parks, and the method of defining drought conditions. During the time frame of our analysis, this region has not experienced significant drought; coupled with the lack of topographic relief at many plots, the result is relatively low variability in moisture stress at most plots. For these mesic sites situated in a humid climate, other factors may exert stronger influence on growth and mortality. In our continuing research, we are using Bayesian General Additive Models to determine if there is an influence of water balance variables on tree growth rates, once we have accounted for the impacts of tree size, nitrogen and sulfur deposition, land use, and competition.