COS 10-2
Drivers of plant species composition in an urban landscape: Which variables matter most?

Monday, August 10, 2015: 1:50 PM
324, Baltimore Convention Center
Bianca E. Lopez, Curriculum for the Environment and Ecology, University of North Carolina, Chapel Hill, NC
Peter S. White, Department of Biology, University of North Carolina at Chapel Hill, Chapel Hill, NC
Dean Urban, Environmental Sciences and Policy, Duke University, Durham, NC

Recently, the number of ecological studies in urban areas has been growing rapidly, as researchers recognize both the profound effects that urbanization can have on the environment, and the importance of maintaining ecosystem services like clean air and water in places where many people live. However, we still know little about the effects of habitat alteration, new species introductions, increased temperatures (the Urban Heat Island Effect), and altered hydrology (the Urban Stream Syndrome) on vegetation in and around cities. We used data on plant species composition from 50 riparian forest sites across the Research Triangle area of North Carolina (including the cities of Raleigh, Durham, and Chapel Hill) to determine how forests change with increasing urbanization. We also measured environmental variables such as canopy openness, stream channel depth, and temperature (using HOBO data loggers) in the field, and used these data to identify which environmental variables are most important for determining community diversity and structure in this complex urbanized landscape. This research has implications for restoration and management of urban areas in North Carolina and other areas of diffuse urban development, and for understanding how urbanization will affect ecosystems in the future.


Ordinations and Mantel’s tests showed that the largest differences in community composition across sites were related to minimum winter temperatures and the degree of variability between daytime and nighttime temperatures, which is lower in more urban areas where temperatures stay warm at night. Impervious surface cover surrounding sites, a metric commonly used to define the urbanization gradient, is also related to species composition, but less strongly—a partial Mantel’s test showed that when controlling for temperature, impervious surface is not related to species composition. We also found that, as expected, urban sites had a higher number of non-native, horticultural, and evergreen species and fewer grass species. Surprisingly, the prevalence of non-native and horticultural species was also more strongly correlated with temperature than urbanization per se, suggesting that either: 1) introduced species are favored by the warmer temperatures created by urbanization, or 2) that on-the-ground temperature measures may provide a more accurate measure of the degree of urbanization than coarse land cover data. Our results also suggest that to protect native vegetation, urban development plans should attempt to minimize the Urban Heat Island Effect by protecting larger forest patches.