LANDTYPE-FOREST COMMUNITY RELATIONSHIPS - A CASE-STUDY ON THE MID-CUMBERLAND PLATEAU

Relationships between forest communities and landtypes (the most detai led level of a hierarchical land classification system) were determine d for the Prentice Cooper State Forest (PCSF), located on the southern tip of Walden Ridge, west of Chattanooga, Tennessee. Four extensive l andtypes within the Mullins Cove area of PCSF were sampled: 1) broad s andstone ridges - south aspect (LT-3), 2) north sandstone slopes (LT-5 ), 3) south sandstone slopes (LT-6), and 4) plateau escarpment and upp er sandstone slopes and benches - south aspect (LT-17). Rectangular, 0 .04-hectare plots specified sub-plots for sampling overstory, midstory , sapling/shrub, seedling/herb forest strata, and physical site charac teristics. Plots (139) were allocated by landtype using a random start with subsequent systematic location. Multivariate statistical techniq ues were used to 1) examine the distinctness of forest communities occ urring among landtypes (discriminant analysis), 2) describe the forest communities occurring within landtypes (cluster analysis), and 3) det ermine factors controlling the spatial distribution of forest communit ies on the landscape (factor analysis). Different relative importance values of species among communities along with different community com binations among landtypes resulted in distinct forest vegetation among landtypes. Chestnut oak (Quercus prinus L.), white oak (Quercus alba L.), and shortleaf pine (Pinus echinata Miller) communities occurred o n all four landtypes. Scarlet oak (Quercus coccinia Muenchh.) communit ies occurred on LT-5, LT-6, and LT-17. Black oak (Quercus velutina Lam .) communities occurred on LT-3 and LT-5. Yellow-poplar (Liriodendron tulipifera L.), northern red oak (Querons rubra L.), and eastern hemlo ck (Tsuga canadensis (L.) Carr.), communities occurred only on LT-17. Landscape scale factors that varied along an elevation gradient were d ominant in controlling spatial distribution of forest communities. Mic rosite factors were secondary controllers. Specific ecological factors could not be determined by factor analysis. Relatively distinct veget ation occurs among sampled landtypes on the PCSF. This study provides additional evidence that the land classification system divides the Mi d-Cumberland Plateau landscape into distinct ecological units.