Comparing maxent outputs?

[Alex Rebelo]

Hi,

I would like to compare maxent's output (total relative occurrence rate) between species. 

I'm essentially looking for an alternative to distribution range-size, that is a proxy for the total suitable habitat.

I would like to know whether it makes sense to use maxent's raw output, or whether these need to be normalised or corrected before comparing between species.

Many thanks,

Alex

[Jamie M. Kass]

When comparing two different species predictions, or across scales, etc. you need normalized predictions. The logistic output in Maxent (see Merow et al. 2013 for details and caveats), makes an assumption about prevalence to achieve normalization, and thus you can compare logistic predictions in Maxent across species. However, see Elith et al. 2011 Box 2 for a description of how careful you need to be when assuming prevalence is equal across species.

Jamie Kass
PhD Candidate
City College, NYC