I'm attempting to calculate the I and D niche overlap values for a couple predictions from MaxEnt. Do I need to transform them into binary maps with a threshold first or can these values be calculated from the continuous probability predictions that come from MaxEnt?