Using GIS to analyse a severe forest blowdown in the Southern Rocky Mountains

Many studies have helped us understand where blowdowns occur, how forests regenerate and respond to large disturbances. However, few studies have analysed how physical and biological factors affect blowdowns or interact with one another to determine susceptibility to blowdown. These factors range from the natural physical setting to past disturbances, and human-related features, such as roads and timber harvest units. These factors were examined in this study. Patches of the blown-down forest were mapped for the 1997 Routt-Divide blowdown, which affected over 10 000 ha of subalpine forest in north-central Colorado. A systematic sampling scheme was used to extract information on predictors for sampling points inside and outside the blowdown. Data on predictors were used in Classification and Regression Trees (CART) and logistic regression to model blowdown occurrence and severity. Two models were applied, and both were able to predict with an overall success rate of 75% or more. These two models produced similar results, showing that the Routt-Divide blowdown was most influenced by factors pertaining to the physical setting: distance to the Continental Divide, wind exposure, elevation, aspect, and cover type. Surprisingly, soil permeability and water-holding capacity and distance to natural edges (e.g. forest-meadow) were not correlated significantly with the blowdown pattern.