I really need some help! I am doing a k-LoCoH analysis in OpenJump. The analysis is working, but I'm having trouble understanding the output. There is a "density" column in the output, and I'm not sure what units this is in, or what it indicates. I know by highlighting some of the polygons that the higher the number in the density column the more used the area...but I still would like to know what this number actually means before I continue with my results! Please help!!
The OpenJUMP HoRAE download http://www.geo.uzh.ch/~sstein/ojhorae/openjump171plus_moveantools_oct2014.zip comes with manual "horae_documentation_v1.2final.pdf". The description of the KDE method on page 17:
In general the KDE method will return a raster with each grid cell containing a “density” value, and not home range polygons. To derive the home range polygons a contouring algorithm needs to be applied. Assuming that density can be related to probability it can be said that a grid cell (i.e. location) with some density value reflects a probability that the animal can be found in that cell. If the value of a cell is zero, the animal is unlikely to be found there. Hence, if we want the area that has a 95% probability of encountering the animal within, we will calculate a contour line that encloses 95% of the density of all cells (volume).
I have no experience about KDE but it seems to me that the unit of density is related to probability but the scale is not fixed. The upper limit of scale may be different for individual source data set and for analysis made with different parameters. Density value of 0 means always that it is unlikely to meet animals in that cell but density value of 100 does not tell anything as is but you must make statistical analysis first to reveal where value 100 stands in the probability curve.
Be careful with my answer because I know nothing about home range analysis and KDE method. Fortunately it is easy to find good reading because the author, Stefan Steiniger, is listing eight scientific literature references about the KDE method tens more about HoRAE in general in the HoRAE manual.
It should be a simple count of the number of convex hulls/observation points that the cell is related to. So, if the count number is higher its more likely that the animal can be found in the cell.