Biotic and abiotic influences on wind disturbance in forests of NW Pennsylvania, USA

Forests in northwestern Pennsylvania experienced a severe windstorm in July 2003. The storm damaged some forests and left others in its path intact. This varied impact raised the question of whether biotic and abiotic stand characteristics influenced storm damage. To answer this question we investigated data on windthrow severity, vegetation characteristics, and physiographic variables provided by the three largest landowners affected by the storm. These local forest managers provided stand level data on 258,000 ha, about half of the storm swath, of which about 5000 ha (2%) experienced moderate or severe blowdown. The study includes over 1002 disturbance patches, 60% of which were under 3 ha. We used classification tree analysis, a non-parametric method of statistical inquiry, to identify the variables that were most useful in predicting storm damage. Our model used biotic and abiotic site factors to correctly predict affected and unaffected stands 89% of the time (kappa value 0.63). The most predictive biotic variables were stand structure and stand age. Predictive abiotic variables included mean elevation, the range of elevations across the stand, and topographic position relative to neighboring stands. Results show that windthrow was more likely in older stands, stands at the highest elevations, and in flatter stands at lower elevations. Except for red maple stands on wet sites, which were disproportionately affected, forest type was not a useful predictor of the storm's impact.