ojaybee

Kernel Density Estimation in a confined area

Discussion created by ojaybee on Dec 26, 2011
Hi there,

I'm looking at homeranges and using the Geospatial Modelling Environment tool (previously Hawth's tool) to do some KDE work.  The population I'm studying lives in a confined area.  I'd like to use the shapefile of their habitat to confine the probablility surface.  So far I've found I can create the prob surface and then use the habitat shapefile to cut them out (method 1).  However, this doesn't make sense as there is 0 probability they will be found outside of their habitat.  I was wondering if anyone else had tackled this problem or if anyone knows how to use the border to confine the probabililty surface (method 2)?  Or indeed if this is the correct method.  I've looked at both Arc's KDE and GME's KDE functions but I can't seem to find anything to help.  Processing extents or masks tend to cut of the probability surface meaning it replicates method 1.

As always, any help would be appreciated.

And Happy Holidays!

Outcomes