Please have look at our ASF article for a condensed review of spatial auto-correlation and multi-collinearity for maxent input data, our empirical findings and relevant references to some of our earlier papers.

As you are mentioning environmental predictor variables, you may be interested in multi-collinearity between them.

Spatial auto-correlation may an issue when using incidental presence data point locations.

Hope it helps

Hi Malik, as suggested by Professor Gils above, you can use PCA to check collinearity between the variables which will invariably deliver a few non-correlated variables suitable for your maxent model building.

Cheers!

Hi Malik, you may have moved on already but you can use SDM ToolBox in ArcMap to evaluate the spatial correlation between your variables of interest. The output is a very helpful correlation matrix - and thus you can identify and remove highly correlated variables (one from the pair) from the model. Cheers!

Spatial Autocorrelation and Moran's I in GIS - GIS Geography

gisgeography.com › GIS Analysis

Jul 15, 2014

Dear Shakil Malik

the software SPSS can be taken into consideration to examine the correlation among the different variables,whose regression analysis can support the significant coefficient to make sure that which one shoule be discard, namely keep which one cariable.

Hope it helps.