After reclassifying a raster land-cover map into foreground (forest and wetland) and background (all other classes), MSPA uses a series of image processing routines to identify hubs, links (corridors), and other features that are relevant to green infrastructure assessments (Vogt et al., 2007). The green infrastructure elements identified by MSPA include core, islet, bridge, loop, branch, edge, and perforation (Soille and Vogt, 2009)(Table 2). In the terminology of green infrastructure, core is equivalent to hub, and bridge is equivalent to link (corridor). MSPA processing starts by identifying core, which is based on the connectivity rule used to define neighbors and the value used to define edge width (Soille and Vogt, 2009). Connectivity can be set to either four (cardinal directions only) or eight neighbors. Edge width affects the minimum size of core and the number of pixels classified as core (Fig. 1). Increasing edge width increases the minimum size of core, thereby reducing the number of pixels classified as core. The ‘loss’ of core that results from increasing edge width results in gains for all other classes, not just edge (Table 3). Increasing edge width can change core to islet if the area of core is small, and core to bridge if the area of core is narrow (seeFig. 1). We used eight-neighbor connectivity and edge width values of one (1), two (2), and four (4) for this analysis. The physical distance (width) of edge translates to 30 m, 60 m, and 120 m for values one (1) two (2) and four (4), respectively, as a result of the native 30 m pixel size of the Landsat TM imagery usedto produce the NLCD (Homer et al., 2007; Fry et al., 2009). Edge width can be set to any integer multiple of the pixel resolution (http://forest.jrc.ec.europa.eu/biodiversity/GUIDOS/).
đang được dịch, vui lòng đợi..