Forest Fragmentation Analysis from Multiple Imaging Formats

Abstract In landscape ecology, forest fragmentation studies with emphasis on effects of scale on fragmentation patch metrics, is an important research area. With increasing availability of satellite data at multiple scales and varied resolutions, it has become important to understand effects of comparing fragmentation metrics acquired from coarse resolution images and those from finer resolution imagery. This is crucial because coarse resolution images such as Landsat imagery, are relatively easier to find because of their cheaper costs, availability and broad coverage, whereas finer resolution imagery is more expensive and therefore, spans only small areas. This paper examines effects of varied spatial resolutions on common fragmentation metrics using Landsat, Sentinel, National Agricultural Imagery Program (NAIP) and Unmanned Aerial Vehicle (UAV) imagery obtained in November, 2017 of the Whitethorne area near Blacksburg, Virginia. The images are analyzed using FRAGSTATS and ArcGIS software programs. The results show significant differences in fragmentation metrics despite simultaneous acquisition of all images in the same area. Discussion of results obtained in this study centers on the reasons for this disparity, and examines uses of imagery of different resolutions for forest fragmentation analysis.