Which Environmental Layers Do I need for MaxEnt
1,801 views
Skip to first unread message
sebastian...@gmail.com
May 19, 2017, 6:08:52 AM5/19/17
to Maxent
Hi there,
I am new in MaxEnt, and would like to basically ask some questions:
1) One sheet of CV file, with more than one species: Is it possible to do one cvs file wiht more than one species? Or even is it possible to only use the genus of a species, without exact species classifiaction? I ask because this would increase the amount of occurance data I have.
2) Environmental Layers: Which environmental layers do I basically need to performa a good modeling. I thought I use DEMs too, but then I read that the bioclim data is already based on a DEM, is this correct? I would like to use a raster with vegetation zones in Bolivia, my study site, too, or is it possible to only clip all environmental data with a shape for Bolivia?
I think this is it, at the beginning.
Thanks for you respone
Best,
Sebastian
I am new in MaxEnt, and would like to basically ask some questions:
1) One sheet of CV file, with more than one species: Is it possible to do one cvs file wiht more than one species? Or even is it possible to only use the genus of a species, without exact species classifiaction? I ask because this would increase the amount of occurance data I have.
2) Environmental Layers: Which environmental layers do I basically need to performa a good modeling. I thought I use DEMs too, but then I read that the bioclim data is already based on a DEM, is this correct? I would like to use a raster with vegetation zones in Bolivia, my study site, too, or is it possible to only clip all environmental data with a shape for Bolivia?
I think this is it, at the beginning.
Thanks for you respone
Best,
Sebastian
Jamie M. Kass
May 27, 2017, 8:06:39 AM5/27/17
to Maxent
1) Each CSV should be of a single species (as far as I know). If you want to run Maxent many times on many species, you should use R to access Maxent to do some looping.
2) This entirely depends on your species and study. What do you think affects the distribution of your study species? Using just elevation is usually a bad idea because it is not a proximal variable for explaining occurrence. In other words, it's not directly related, but instead correlated with temperature or precipitation, which are actually driving the distribution. Use more proximal variables, but if you want to include elevation too because you think it gives unique information, by all means toss it in. Maxent is pretty good at handling some colinearity between variables, as it does internal variable selection via L1 regularization (see the Phillips et al. 2006 or 2008 paper).
Bioclim variables (i.e. bio1-19 from Worldclim) are interpolated surfaces made from weather station data, and use elevation to correct the interpolation. You can certainly use a categorical variable like veg zones. Not sure what you're asking about clipping variables -- yes, it's possible using a GIS software or with R, etc.
Jamie Kass
PhD Candidate
City College of NY
2) This entirely depends on your species and study. What do you think affects the distribution of your study species? Using just elevation is usually a bad idea because it is not a proximal variable for explaining occurrence. In other words, it's not directly related, but instead correlated with temperature or precipitation, which are actually driving the distribution. Use more proximal variables, but if you want to include elevation too because you think it gives unique information, by all means toss it in. Maxent is pretty good at handling some colinearity between variables, as it does internal variable selection via L1 regularization (see the Phillips et al. 2006 or 2008 paper).
Bioclim variables (i.e. bio1-19 from Worldclim) are interpolated surfaces made from weather station data, and use elevation to correct the interpolation. You can certainly use a categorical variable like veg zones. Not sure what you're asking about clipping variables -- yes, it's possible using a GIS software or with R, etc.
Jamie Kass
PhD Candidate
City College of NY
Dimitris Poursanidis
May 27, 2017, 8:10:17 AM5/27/17
to max...@googlegroups.com
On what Jamie sais about the CSV, you can have as much species as you want in one CSV (i have use 100 in one)
On the data perspective - the which are ecological relevant have ti be used. Climatic (WORLDCLIM or CHELSEA or other sources) or topographic related (DSM and 1st/2nd order parameters), LULC data and other related to the case study.
>->->->->->->->->->->->->->->->-
Dimitris Poursanidis
Ph.D. Marine Ecology
The latest papers !!!
--
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+unsubscribe@googlegroups.com.
To post to this group, send email to max...@googlegroups.com.
Visit this group at https://groups.google.com/group/maxent.
For more options, visit https://groups.google.com/d/optout.
Jamie M. Kass
May 27, 2017, 9:56:38 PM5/27/17
to Maxent
Thanks for the clarification, Dmitris. Never tried running the GUI with multiple species, but glad to know it works. Keep in mind, however, that if you do it this way, I think they all use the same background extent. This is probably not a great idea if your species are located far apart from each other and/or have different areas they are able to disperse to.
-Jamie
-Jamie
Dimitris Poursanidis
May 28, 2017, 3:08:47 AM5/28/17
to max...@googlegroups.com
I know that and yes i use it in similar bg extend species
>->->->->->->->->->->->->->->->-
Dimitris Poursanidis
Ph.D. Marine Ecology
The latest papers !!!
Reply all
Reply to author
Forward
0 new messages