An Operational Model for Assessing Risk of Oriental Bittersweet, an Invasive Exotic Vine, in the Southern Appalachians

W. Henry McNab and David L. Loftis

USDA Forest Service Southern Research Station

Oriental bittersweet, a shade-tolerant, twining, rapidly-growing exotic vine, is a serious threat to hardwood silviculture in the eastern United States. Land managers need an economical and accurate method to assess the risk of occurrence of this species on the forested landscape. In a previous study of the distribution of bittersweet in a small study area, two variables were important predictors: (1) concave landforms were associated with its presence, and (2) the evergreen shrub Kalmia latifolia was associated with its absence. A logistic model using these variables to predict the presence of bittersweet performed well in the study area, achieving an accuracy of 87 percent when applied to a small sample of plots excluded from the analysis. Additional validation of the model was needed before it could be recommended to land managers as a method for assessing risk of bittersweet.

The study reported in this poster had two objectives: (1) validate the preliminary model for predicting the occurrence of bittersweet over a larger area and (2) devise an operational method for applying the model. Forty-three plots were installed in a uniform grid throughout the 6,000 ac watershed surrounding the 100 ac study area. Application of the model resulted in prediction accuracy of 79 percent; false positives were 19 percent, but more importantly, false negatives were only 2 percent. These results suggest the preliminary model is applicable over a broader area.

The topographic components of the model can be applied using a geographic information system (GIS) to determine land surface concavity, but a satisfactory method to predict occurrence of Kalmia is not available. Kalmia, however, is readily visible on dormant season, color infrared aerial photography; therefore the extent of this species should be accurately detected and mapped with image processing software.  We will develop an operational risk assessment model that utilizes classified imagery during the upcoming winter and test it during spring 2006. We are relatively confident that a satisfactory model utilizing a combination of GIS and aerial photography can developed for assessing risk of the occurrence of oriental bittersweet on forested landscapes in the southern Appalachians. 

corresponding author:

W. Henry McNab
USDA Forest Service
Southern Research Station
1577 Brevard Road, Asheville, NC 28806
828-667-5261x119
hmcnab@fs.fed.us