Survey of forest herb abundance at 55 sites in the French Broad River watershed, 2009.

We identified 14 native herbaceous species that represented a range of habitat generalists and specialists, and were common to the area. Sampling took place from May – August 2009. We selected sites by generating random points across 5 watersheds in ArcGIS and choosing a subset of 54 sites that were accessible, not recently logged, and not dominated by Eastern hemlock (Tsuga canadensis), rhododendron (Rhododendron spp.), or mountain laurel (Kalmia latifolia). A 20x20-m plot was established at each site containing 3 transects at 0-m, 10-m, and 20-m, perpendicular to the slope. Using a modified version of the Braun Blanquet (1932) scale, we recorded percent cover for each focal species within 1-m2 quadrats at 6 evenly spaced intervals along each transect (18 quadrats per plot). The entire plot was thoroughly searched and the presence of any focal species not recorded in a quadrat was noted and assigned to the lowest cover class (0-0.1%). Within each quadrat, we also measured percent total herb and shrub cover, and litter depth (an average of 4 measurements). Slope and terrain shape were determined using a clinometer. We extracted elevation and distance to stream for each plot using 30-m digital elevation models (DEMs) in ArcGIS. Distance to stream represents relative slope position and was obtained by calculating flow accumulation and the distance from plots to the nearest point of highest accumulation, using a threshold of 100 pixels. High values indicate positions upslope near ridges and low values indicate positions downslope close to streams or coves. Five soil cores (5 cm diameter, 15 cm depth) were obtained from each plot, composited, sieved (4 mm) and oven-dried. Soils were analyzed for total N, organic matter, P, K+, Ca++, Mg++, and pH by the University of Wisconsin-Madison Soil and Plant Analysis Laboratory using standard methods. Soil texture was determined in the laboratory using the hydrometer method.