PS 17-131
Urban tree crown attributes in the southeastern US: Differences in model form and driving factors by species and location
Urban forests improve the well-being of humans and animals, lessen air pollution and flood risk, reduce noise, and regulate temperatures. Proper care, monitoring, and management of these forests is essential to optimize their community benefit. The goal of this research is to develop predictive models of critical urban tree attributes for the southeastern US, such as crown characteristics and tree height. These models can greatly reduce the resources land managers need to conduct urban tree inventories. Through a newly created, standardized regional urban forestry database, we have high quality urban forest data from 10 college campuses and cities in the southeastern US (e.g., Auburn University, AL, Falls Church, VA, and Gainesville, FL), collected using the USDA Forest Service’s i-tree/Eco protocol. These data include basic forestry measurements and other tree and plot-level environmental variables such as ground covers, land use, and distances to adjacent structures. We will utilize this new urban forestry database to estimate models of crown characteristics and answer questions about the drivers of important urban tree attributes, such as: 1) Do factors that determine tree height and crown dimensions differ by species and region? 2) Do species growing in different locations differ in model form (i.e., nonlinearity)?
Results/Conclusions
The data from 10 sites across the southeastern US yielded ~15,000 tree records. Species-specific models could be estimated for 10 species (1 pine and 9 hardwoods) with >200 individuals across 3-9 sites, representing ~30% of all recorded trees. Genera-level models were estimated for Pinus spp., Quercus spp., and Ulmus spp. Unsurprisingly, mixed model analyses indicated that tree diameter at breast height (DBH) was the most important driver of tree height. Other significant drivers varied by locale and species, and there were differences even among species in the same genera. Tree height in most hardwood species (e.g., Quercus, Prunus, Acer, and Ulmus spp.) was significantly related to local tree density and crown light exposure (CLE). Pinus spp. heights were more affected by understory cover, and demonstrated a non-linear trend. Models of crown dimensions also showed species-specific differences and non-linearities. DBH and CLE were both significant drivers of crown area in most species, whereas ground cover had little relationship to crown area. This research is a first step in increasing our knowledge regarding urban tree attributes and how they differ among communities and species, providing the foundation for more refined estimates of urban tree and crown attributes and their community benefits.