| Title: | An Improved Method for Classifying Forest Fragmentation |
| Authors: | Jason Parent, James Hurd, University of Connecticut |
| Date/Time: | Wednesday, September 24 ~ 8:45 a.m. - 10:15 a.m. |
| Abstract: | Anthropogenic activities in the northeastern United States are causing forests in the region to become increasingly fragmented. Fragmentation not only reduces the quality of forest resources in terms of its ecological services, but also impacts its economic value. Quantifying the types of fragmentation in a forest landscape makes it possible to identify potentially threatened areas and help us plan new development to minimize additional impacts in the future. Several tools have been developed to classify the types of forest fragmentation from raster land cover datasets. In this presentation, we present a new tool, based on work done by Vogt et al. (2007), in which morphological image processing is used to classify spatial patterns at the pixel level. The procedure developed by Vogt et al. (2007) classifies forest pixels into one of four categories: core, perforated, edge, and patch. Core forests are interior forests with no fragmentation. Perforated and edge forest make up the interior or exterior edge of a forest patch respectively. Patch forest is a small fragment of forest surrounded by non-forest land cover. This new procedure is less prone to misclassification than previous pixel-based approaches to forest fragmentation. We present an ArcGIS model that produces a result functionally equivalent to that of the Vogt et al. (2007) procedure, but executes more efficiently. The model allows the user to specify the width of edge and perforated forests and can be easily applied to land cover maps of any spatial resolution. |
| Print this Page | |
| Close Window | |