Adding multiple environmental variables to Maxent models

[Clarissa Crist]

Hello!

I was wondering if anyone knows if it is possible to produce Maxent models using multiple environmental variables. 

I am working on modeling plant species on Madagascar and have modeled using current bioclimate conditions on the island, but also soil types and topography. Is there a way I can produce a model using all of this data together. I have tried combining bioclimate and soil type variables in Maxent and it wouldn't work. I'm not sure if it's because it is processing too much data or the data types are different formats, so Maxent has a hard time registering both together. 

Thank you!

[mjb...@york.ac.uk]

Hello Clarissa,

Madagascar is my current working area for modelling, so I will be interested in hearing about your results when you have run your models.

As for your current problem: you can chuck in as many variables (within reason) into maxent but I wouldn't advise it and there are two immediate issues. Firstly be clear of what your question is. Are you looking to model plant species according to climate or abiotic conditions? There is likely to be a lot of correlation between variables that will produce a confused picture with you final results. Be selective about variables and run a correlation matrix beforehand. If you absolutely have to include multiple collinear variables, have a think about data reduction techniques such as PCA. But I would choose to simplify wherever possible. What are your current variables?

The second reason for not sticking climate in with soil types (etc) and modelling all in one go, is that as far as I am aware, the finest resolution climate data for Madagascar is approximately 1km, whereas topography, soil ph and so on are available at much finer scales. So, in order to get maxent to work, all the asc files (rasters) need to be at the same spatial scale. I suspect this is the reason your models don't run. Do you receive any error messages? A simple check is to open all of the asc files in notepad as text files. Each file has a header which states the number of rows and columns, map extent etc. These need to be the same on each asc file.

A final thought; I would advise having a think about the scale at which your subject organism experiences the environment.

best wishes

Michael