Effects of riparian zone buffer widths on vegetation diversity in southern Appalachian headwater catchments

In mountainous areas such as the southern Appalachians USA, riparian zones are difficult to define. Vegetation is a commonly used riparian indicator and plays a key role in protecting water resources, but adequate knowledge of floristic responses to riparian disturbances is lacking. Our objective was to quantify changes in stand-level floristic diversity of riparian plant communities before (2004) and two, three, and seven years after shelterwood harvest using highlead cable-yarding and with differing no cut buffer widths of 0 m, 10 m, and 30 m distance from the stream edge. An unharvested reference stand was also studied for comparison. We examined: (1) differences among treatment sites using a mixed linear model with repeated measures; (2) multivariate relationships between ground-layer species composition and environmental variables (soil water content, light transmittance, tree basal area, shrub density, and distance from stream) using nonmetric multidimensional scaling; and (3) changes in species composition over time using a multi-response permutation procedure. We hypothesized that vegetation responses (i.e., changes in density, species composition, and diversity across the hillslope) will be greatest on harvest sites with an intermediate buffer width (10-m buffer) compared to more extreme (0-m buffer) and less extreme (30-m buffer and no-harvest reference) disturbance intensities. Harvesting initially reduced overstory density and basal area by 83% and 65%, respectively, in the 0-m buffer site; reduced by 50% and 74% in the 10-m buffer site; and reduced by 45% and 29% in the 30-m buffer site. Both the 0-m and 10-m buffer sites showed increased incident light variability across the hillslope after harvesting; whereas, there was no change in the 30-m and reference sites over time. We found significant changes in midstory and ground-layer vegetation in response to harvesting with the greatest responses on the 10-m buffer site, supporting our hypotheses that responses will be greatest on sites with intermediate disturbance. Ground-layer species composition differed significantly over time in the 0-m buffer and 10-m buffer sites (both P \u3c 0.0001), but did not change in the 30-m buffer and reference sites (both P \u3e 0.100). Average compositional dissimilarity increased after seven years, indicating greater within stand heterogeneity (species diversity) after harvesting. These vegetation recovery patterns provide useful information for evaluating management options in riparian zones in the southern Appalachians