The Effect of Stream Restoration on Turtle Species Assemblages in the Piedmont Region of North Carolina, USA

In response to the negative impact of urban and agricultural development on freshwater systems, stream restoration efforts often attempt to return degraded streams to a natural ecosystem structure and function. However, few attempts have been made to monitor the effectiveness of restored streams in supporting certain important groups of organisms found in native aquatic ecosystems, such as freshwater turtles. The purpose of this study was to compare six natural and six restored streams in the North Carolina Piedmont by quantifying habitat characteristics that might drive differences in turtle assemblages and by directly capturing turtles at each site. Stream habitat was characterized by water quality analyses, structural measurements of each stream, and floodplain vegetation surveys. Three baited hoop nets were set at each location for a total of 36 trap-nights at each stream, which were used to collect turtle population data from mid-May to late July 2009. In total 77 turtles were captured comprising eight species. At the natural sites, 24 turtles were captured representing five species (C. picta, C. serpentina, P. floridana, S. odoratus, and T. scripta scripta); 53 turtles were captured at the restored sites representing seven species (C. guttata, C. picta, C. serpentina, K. subrubrum, S. odoratus, T. scripta elegans and T. scripta scripta). Modified t-tests, based on randomized permutation tests, suggest that natural and restored sites differ in turtle abundance (p=0.13), but are not different in species richness (p>0.99) or gender ratios (p=0.80). A species community index suggests that natural and restored turtle assemblages overlap by 50%. Non-metric multidimensional scaling analyses of habitat characteristics indicate that natural and restored streams differ in channel structure, vegetation, and some water quality variables. Using Mantel’s test to compare turtle species composition with the most important variables separating natural and restored streams, canopy, slope, total phosphorus, chlorophyll A, and Juncus effuses abundance were found to be most strongly correlated with patterns in turtle assemblage composition. This is one of the first studies to address the possible impact of stream restoration on turtle assemblages, and the findings suggest that restored streams may be better habitat for turtles in the North Carolina Piedmont.