I am currently developing a habitat model within ArcGIS Desktop for coyotes as a final project. One of the papers I read for my research did something similar, but they used a moving neighborhood analysis, setting the size of the neighborhood as the common home range for coyotes in the area.

I had no plans of doing this, my plan is to just use the finest resolution cell size as I can; I am curious though, what is the benefit of this?

I'm not really following how grouping up cells within a certain radius and adding there values together is any better than just evaluating each cell individually, could someone explain?

1 Answer 1


Using the finest resolution possible is not neccessarily the right choice when doing habitat modeling. I can't think of any papers to cite right now but consider this:

  • When creating a habitat model (HM) the resolution highly depends on the species you are looking at. A mouse has a very different moving and cognition range from an elephant. Thus, using a 1x1 m grid might be approptiate for a mouse HM, but for an elephant a resolution of 10x10 m or even 1000x1000 m might be just fine. Look for papers to get an appropriate cells size for your species.
  • When including moving windows analysis into your model, you integrate effects of neighbooring cells. So when there is a waterhole on cell a (e.g. raster cells with water) neighbooring to a cell b, the cell b will be more attrctive then cell c which has no neighbooring waterhole. And chosing the homerange of your modelled species makes sure that the effect is relevent for your modelled species. So doing an moving window before creating your actual model is quite a good idea.

EDIT: After a little research I found this paper which goes in to depth further. https://www.nature.com/articles/s41598-018-25437-1
One of there main points is that smaller cell sizes are important for land cover as an input variable when modelling habitat specialists because an accurate model should capture all the available terrain. While I'm by no means a specialist, I doubt this is true for coyotes.

  • awesome answer, thank you! The papers I am referencing did not state why they chose the cell sizes they did and I think it was likely the cell size that was most usable for all factors (i.e. whichever dataset had the coarsest cell size was chosen). Commented May 4, 2019 at 19:14
  • You are welcome @KarlJohnson. Please accept the answer using the little check to mark the question answered.
    – blabbath
    Commented May 5, 2019 at 7:22

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.