Vegetation Dynamics of the Forest Edge -- Trends in Space and Successional Time

Abstract

1 Forest herbs, shrubs, and tree seedlings were censused at five distances from human-generated edges at 14 sites in south-eastern Pennsylvania and northern Delaware, USA. Forest edges varied in orientation and in the degree of canopy closure, allowing comparisons of vegetation by aspect and successional stage. Soil depth and pH were measured at vegetation sampling points at a subset of sites; microclimate data were available from a parallel study. 2 Edge-related pattern was observed in overall species composition, and in distributions of 15 individual species. At recently created forest edges, species abundances were correlated with steep climatic gradients. Although most edge-orientated species were clustered $\leq 5$ m from the edge, some species consistently reached peak densities at distances up to 40 m, the greatest distance surveyed. 3 Edge-related pattern was most frequently observed at recently created edge sites, but persistent pattern was also observed at older edges, including those closed for 55 years by succession. Species distributions at older edges showed fewer instances of edge-related pattern, modal distances were farther from the edge, and the strength of residual pattern was lower than at recent sites. 4 Edge-related pattern was often observed at north-facing sites. North-facing edges were not generally distinguishable from other recently exposed edges on the basis of edge effect strength (Gini coefficient) or mean distance of individual species from the edge. 5 These observations suggest a cycle of pattern imposition and relaxation as the edge ages, and are consistent with the great clonal longevity observed in many forest species. A Piedmont forest reserve designed to protect interior habitat from vegetational changes at forest edges should include points further than 92 m from the nearest edge, and should be comprised of forest more than 100 years old.