Can anybody suggest a robust way to decide the size of moving window to to assess landscape diversity?
I have a raster land use map with resolution 10 x 10m. Now I want to calculate diversity using land use map. I checked some papers suggesting very different numbers ranging from 100 to 500m.
Thank you
No numeric suggestion could be possibly meaningful without knowing the exact goals. The relevant scale is dictated by the scale of the process or ecological unit you attempt to model with your new moving-window diversity raster. If your target species has an average (assumed to be circular) home range size of e.g. 1 km^2, i would try including cells within approx. 500 m to 600 m distance to the centroid of the focal cell.
rainer
Ecologists and cia. have dealt with this for a long time. For practical purposes it can be said that the optimal plot size is the one that maximizes information content at the least cost. One approach is to increase the size of the plot several times and estimating a given measure of information content for each one. When you plot this against the plot values you will most likely see a logarithmic trend, which tells you that at the begging you gain a lot for each increase in plot, but after some point it satabilizeas.
The following paper should get you on track.
https://www.google.com/url?sa=t&source=web&rct=j&url=https://abcjournal.org/index.php/ABC/article/download/1568/1533&ved=2ahUKEwiW3PDUn8flAhUsIqYKHSNkDhsQFjASegQICxAB&usg=AOvVaw1qQMIun5gP97tXr32X1SJE
Hier a couple of more
http://www.sci.sdsu.edu/SERG/techniques/mfps.html
- Badges
- Science topic
- Similar topics
- Papers