temporal scenarios comparison using classes of Jenks natural breaks method

[Det.J]

Dear Maxenters

I was using the Jenks natural breaks method (in ArcGIS) to classify maxent output into 4 classes (unsuitable to highly) for my current and future scenarios. Knowing that these breaks differ for each scenario, would the comparison be valid between these suitability classes through time?

I found some papers using this method but they didn't explain if they kept these different class boundaries or if they calculated the average of each and then applied it to all scenarios (I saw one paper averaging it however that was for two 'present' scenarios that are different in the size of the study area). Any thoughts on this matter?