A Model to Predict the Occurrence of Surviving Butternut Trees in the Southern Blue Ridge Mountains

van Manen, Frank; Clark, Joseph; Schlarbaum, Scott; Johnson, Kristine; Taylor, Glenn

ABSTRACT: Butternut (Juglans cinerea) is a hardwood species native to eastern North America and is valued for high-quality lumber and as a mast source for wildlife. Many butternut populations are currently being devastated by an exotic fungus (Sirococcus clavigignenti-juglandacearum), which causes multiple cankers that girdle the tree and cause mortality. The decimation of butternut populations has been severe in many areas, particularly in the southeastern United States. A breeding program involving resistant genotypes might be a feasible strategy to conserve and eventually reintroduce the species. However, such an effort is critically dependent on the ability to locate enough disease-resistant stock for evaluation and subsequent breeding. We developed a data-based model to predict probable occurrence of surviving butternut trees. Field surveyors initially mapped 134 sites with surviving butternut trees in Great Smoky Mountains National Park (GSMNP). These locations were entered into a geographic information system (GIS) and overlayed with digital map layers of eleven different geographic variables that described the topographical and biological conditions at each sampling point. We then calculated the Mahalanobis distance statistic, a multivariate measure of dissimilarity, for each 92.9- x 92.9-m pixel in a 31,235-km2 region encompassing the southern Blue Ridge Mountains to predict the relative likelihood of species presence. We then tested this model with independent presence/absence data from 130 test sites in this region as the dependent variable and Mahalanobis distance as the independent variable. We found that the Mahalanobis distance statistic showed a significant relationship with occurrence of butternut in the region. Although the field tests indicated that the species often was absent in areas predicted to be good butternut habitat, butternut decline due to the canker also may have contributed to that outcome. Biometric models such as the one we developed can be used by resource managers for collection purposes, restoration efforts, and to identify habitat associations across large areas.