2015 GSA Annual Meeting in Baltimore, Maryland, USA (1-4 November 2015)

Paper No. 231-4
Presentation Time: 9:00 AM-6:30 PM

SOILS ON THE EDGE: EXPLORING THE LINKS BETWEEN INVASIVE PLANT COMMUNITY STRUCTURE AND EDAPHIC PROPERTIES AT FOREST EDGES IN WESTERN PENNSYLVANIA


COUGHENOUR, Christopher1, NORRIS, Tyler2 and SHANLEY, Rachel2, (1)Dept. of Energy & Earth Resources, University of Pittsburgh-Johnstown, 227 Krebs Hall, 450 Schoolhouse Rd, Johnstown, PA 15904, (2)Dept. of Energy & Earth Resources, University of Pittsburgh-Johnstown, 450 Schoolhouse Rd, Krebs Hall, Johnstown, PA 15904, tan42@pitt.edu

The forest edge is a transitional region that exhibits significant spatial gradients with regard to vegetation, microclimate, and soil properties. Edge regimes often exhibit high concentrations of invasive plants that grade into more native understory communities. Most forest edge studies have addressed general plant communities and microclimate characterization. Few studies have explored the relation of invasive plants to soil properties at forest edge at sub-hectare scale.

In this study, soil properties were collected over a 0.049 ha area at each of three sites in Johnstown, PA. Sampling began within two meters of the trunk line on the fieldward side of the forest. Physical soil data were collected on an 18 point grid, along with canopy closure and leaf litter depth. Invasive shrub locations and sizes were recorded at six sites. Soil data were used from the grid and interpolated for 32 randomized points at each site. Invasive spatial density was calculated at the same points.

Relative abundances of invasive species were significantly different between some sites. Pairwise Holm comparisons indicated site differences. At four sites 90% of invasives were within 6.5 m of trunk line. Invasive shrubs were spatially aggregated at all sites, indicated by species-blind density mapping and Ripley's L function. Dixon's function revealed two sites exhibited species segregation.

Canonical correlation analysis was used to probe the data. The first canonical variate pair at each of the sites exhibited strong canonical correlation. Matric potential was most correlated to the invasive density variate, with depth to Bt horizon also strongly correlated. No other soil properties were consistent in effects at all sites. Two sites with similar aspect, but differing closure times had the most dissimilar loadings. Prior studies of broad soil-plant interactions, generally not at forest edge, have often indicated little relation between vegetation and edaphic properties.