Ecological species groups of South Carolina's Jocassee Gorges, southern Appalachian Mountains

ABELLA, S.R. AND V.B. SHELBURNE (Department of Forest Resources, Clemson University, Clemson, SC 29634-0331). Ecological species groups of South Carolina's Jocassee Gorges, southern Appalachian Mountains. J. Torrey But. Soc. 131: 220-231. 2004-Ecological species groups, consisting of assemblages of co-occurring plant species exhibiting similar environmental affinities, were developed for ground-flora and tree strata in late-successional forests on a 13,000 ha southern Appalachian landscape. We distinguished 11 ground-flora groups that included 50 species and six tree groups comprised of 19 species. Ground-flora groups ranged from a xeric Vaccinium group (including Vaccinium pallidum, Euphorbia corollata. and Piptochaetium avenaceum) to a mesic Rhododendron group (typified by Rhododendron maximum, Mitchella repens, and Hexastylis heterophylla). Tree groups ranged from a Quercus coccinea group to a Tsuga canadensis group. Consistent with previous research, species groups exhibited a range of amplitudes from widely distributed Smilax and Vitis groups to a Sanguinaria group restricted to one ecosystem type. A given species group occupied a variety of different combinations of measured environmental variables, which apparently interacted to produce repeating environmental complexes across the landscape favorable for specific species groups. We also tested two multivariate methods for quantifying associations among species groups, and found that Mantel tests using traditional distance measures were inappropriate because of the double-zero problem of species absences, whereas canonical correlation modeled species group associations consistent with species distributions among sites. This study is among the first to develop ecological species groups in the southern United States, and the species group approach was useful for explaining vegetation-environment relationships, identifying groups of ground-flora and tree species that varied together across the landscape, and for determining the environmental gradients most strongly associated with species distributions.