Long - term patterns in vegetation - site relationships in a southern

ELLIO~, K. J., J. M. VOSE, W. T SWANK, AND P V. BOLSTAD. Long-term patterns in vegetation-site relationships in a southern Appalachian forest. J. Torrey Bot. Soc. 126320-334. 1999.-Wused permanent plot inventories from 1969-1973 and 1988-1993 to describe forest species distribution patterns of the Coweeta Hydrologic Laboratory, a 2,185 ha basin in western North Carolina, USA. We used canonical correspondence analysis to explore the vegetation-site patterns for the 1970s and 1990s inventories combined. Site variables were determined by direct measurements or calculated by digital geographical information system mapping methods. Site variables were percent slope, elevation, terrain shape, precipitation, modified azimuth, soil organic matter content, soil depth, soil clay content, depth of A-horizon, potential solar radiation, and mean temperature during the growing season. Fifty percent of the variation in the vegetation distribution was explained by the site variables used in the canonical correspondence analysis. Soil organic matter, terrain shape, and elevation were the variables most strongly related to vegetation distribution. Species associated with convex terrain (upper slopes and ridges), such as Pinus rigida, Quercus coccinea, and Quercus velutina, decreased in abundance from the 1970s to the 1990s; species associated with soils having high organic matter content and deep A-horizons, such as Liriodendron tulipifera, Rhododendron maximum, and Tsuga canadensis increased in abundance. Individual species responded differently to site gradients. For example, Acer rubrum, Quercus prinus, Oxydendrum arboreum, and Nyssa sylvatica were located in the center of the ordination space (i.e., their occurrence was not related to any of the site variables), which suggests that these species are habitat generalists.