Prediction of Forest Type and Productivity Index on Disturbed Sites in Great Smoky Mountains National Park

A study was conducted in a portion of the permanent vegetation plots established by Uplands Field Research Laboratory in Great Smoky Mountains. The plots were located in the northwestern portion of the park in the vicinity of Cades Cove and Tremont on sites previously disturbed by chestnut blight, logging, fire, farming activities, and livestock grazing. Estimated plot productivity, and data on vegetation, soils, and topographic parameters were available from previous study of these plots. The purpose of the present study was to develop methods to predict forest cover type and plot productivity index score using site and disturbance history-related factors.

On a gradient ranging from generally xeric to generally mesic, eight forest cover types were delineated (1) Yellow Pines, (2) White Oak - Oak, (3) Chestnut Oak - Oak. (4) Mixed Sub-xeric Hardwoods, (5) White Pine - Hardwoods, (6) Chestnut Oak - Yellow- poplar, (7) Yellow-poplar, and (8) Mixed Mesic Hardwoods. Elevation, broad scale topographic factors, mean plot age, the degree of even-aged conditions within a plot, and the thickness of the organic and B soil horizons were the most important predictors of forest cover type.

Plot productivity index score was based on the volume increment of all trees in a plot over 30 cm dbh. A broad scale categorization of landform and total soil depth were the best predictors of plot productivity index score.

In general, broad scale topographic measurements derived from maps were better predictors of forest cover type and plot productivity index score than were categorizations of topography based on the investigator's perception of topography from within the plot itself. The effects of topography-related variables on forest cover type were more important at low elevation.

Disturbance history was found to be correlated to elevation. Topographic factors related to site moisture supply interacted with disturbance, particularly at low elevation. Thus, a given disturbance history did not always lead to the same forest cover type. Furthermore, a given forest cover type was found to be the result of more than one type of disturbance history scenario.