Predicting forest cover types in Southwestern Virginia using topographic information

TR Number
Journal Title
Journal ISSN
Volume Title
Virginia Polytechnic Institute and State University

A computerized clustering algorithm, DIVIDE, was used to gain insight into the relationships between physical site factors and existing forest communities in southwestern Virginia. A pair of dichotomous keys was produced that "predicted" the forest type most likely to occur in an area based on topography. Maps of predicted forest types, using the Trayis (1982) and SAF (1980) vegetation classification systems, were produced for the entire study area. Accuracy levels of 57 to 78 percent were obtained. There were no significant differences in classification accuracy between Trayis and SAF forest type predictions (P > 0.25).

Herbaceous understory was sampled on the basis of cover, and mast production was estimated in each of the forest types. Forest types on sites had significantly greater amounts of forb and fern cover than those on drier sites. Production of grasses and leaves of woody plants was probably similarly affected, but differences were not significant. Estimates of acorn production were highest in old stands containing a high percentage of oaks. Oak stands on moist sites appeared to have higher estimated mast yields than those on dry sites, but differences were not significant.

Based on the results of these analyses, forest types were rated tor deer suitability using compatibility matrices. Deer habitat suitability maps vere produced for the entire study area based on these matrices. Differences between overall suitability values for the Travis and SAF systems suggest that selection of an appropriate forest type classification system is important for wildlife managers.