Buffer vs. reprojected area: which area to use for the correlation test?

[J.wolk...@gmail.com]

Dear all, 

I'm, using buffers around presence points for specifying maxent background area and then using reprojection of models on the area of the chosen country. But could someone tell me which area I should use for correlation analysis? Should I use env. var. only from the area of buffers, reprojected country, or even different? 

Thank you,

J.W.

[Jamie M. Kass]

Do you mean a correlation test of the predictor variables used to build the model? In that case, the only extent that makes sense is the background extent (your buffers). The projected extent is not used to train the model. Please correct me if I misinterpreted your question.

Jamie Kass
PhD Candidate
City College of NY

[j.wolk...@gmail.com]

Yes, that is what I asked, thank you very much!

J.W.

Dne čtvrtek 29. března 2018 13:17:55 UTC+2 Jamie M. Kass napsal(a):