Landscape metrics from rasters

[JamesDoyle1]

Hello everyone,

I am working on a habitat suitability model and need some help creating environmental variable layers. Does anyone know of some way to create a continuous map of various landscape metrics from a National Land Cover Dataset raster in Arc? I am trying to analyze some of the same metrics fragstats has, but the fragstats output is not very friendly for the modeling software I am using (Maxent). For instance, I would like to create a distance-to-edge map that calculates the euclidean distance from any given cell in the raster to a cell of a different class (in this case land cover, e.g. distance to forest). I haven't been able to find a tool that calculates the distance to another class within the same raster, only distance to a shapefile. I would also like to calculate the edge length or density within a neighborhood (e.g. 10 cell x 10 cell window), and some sort of shape complexity index. Fragstats can give me all of these, but I need to do it based on my land cover raster so I will have a continuous background map of the whole study area that I can use in Maxent. I'm pretty new to ArcGIS so I may be missing some obvious solutions to this.

Any help or suggestions are much appreciated!

Jimmy

[JeffreyEvans]

I would recommend reading the FRAGSTATS help more thoroughly. If you go to the Fragstats > Set Run Parameters menu there is an entire dialog on specifying an input raster, defining window size and outputting specified metrics as rasters. You will need to write a class properties file (.fdc) first but otherwise it is very straight forward. Certainly, much easier than writing your own landscape metric models in Python. MaxEnt requires ASCII grid input which FRAGSTATS can directly output.

[JamesDoyle1]

Thanks Jeffrey. I think the moving window analysis in Fragstats gives me what I need for a background layer in Maxent. I'm starting a run now so fingers crossed...