How to compare SSP and RCP scenarios ?

[faramarz bozorgomid]

Hello MaxEet users
I want to predict the distribution of a species based on two scenarios, SSP and RCP, and then compare them.
The correlation test between bioclimatic layers showed that for the RCP scenario, layers ( 1, 2, 3, 4, 5, 12, 14, 15) and for SSP (1, 2, 3, 12, 14, 15) remained.  So, the rest of the layers are correlated and should be removed from modeling.
Now, to compare the outputs of the two scenarios, should the selected layers be the same? That is, when modeling, should I add layers 4 and 5 to the SSP scenario or not, and should I remove layers 4 and 5 from the RCPs?

[Bede-Fazekas Ákos]

Hello,
although I don't really know what do you mean "comparing" the two, but if you want to draw some conclusions based on the cell-wise comparison of the predicted suitabilities, then I suggest training the two models followin exactly the same parameterization and using the same predictors. If collinearity analysis let you remove some of the predictors from model no. 1 (i.e. based on the SSP), then you should remove these predictors from model no. 2 (i.e. RCP) as well.
HTH,
Ákos
_______
Ákos Bede-Fazekas
Centre for Ecological Research, Hungary

--
You received this message because you are subscribed to the Google Groups "Maxent" group.
To unsubscribe from this group and stop receiving emails from it, send an email to maxent+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/maxent/7d54b83c-3071-4d85-b13e-d2bf73b68ea6n%40googlegroups.com.

[Милош Поповић]

In my opinion, the layers should be the same if you would compare two scenarios. And it is generally better idea not to use derived varibles, but keep it simple. Note that Maxent could handle some level of correlation quite well, thus some people tend to use all the variables. Regularization could be used to pennalise correlated variables.

Best,

Miloš

--