step three.2. Timing beneficial and you can incident locations

I checked out four housing thickness kinds aggregated at a three hundred m scale; 0 domiciles each ha (wildland), 0 to at least one.47 (outlying and exurban), 1.47 so you can 10 (suburban), and you can higher than 10 home each ha (urban). Hr of the GPS collar towns is discretized into the five bins: night (–), morning (–), go out (–), and you will evening (–) schedules. For each and every individual and you will period of time, we summed just how many GPS collar location findings inside the for every houses thickness class. To standardize the latest proportional use of the property densities across date (as time periods differ during the course), i upcoming split the ensuing worth by the final amount from towns and cities (for the individual) one decrease for the each time bin classification.

Presenting aforementioned, i utilized the inverse logit of projected cougar event regression equations, excluding the brand new intercept, in order to aesthetically define the newest relative likelihood of an excellent cougar incident and constrain the costs anywhere between no and one, then multiplied for each and every epidermis of the a facial skin representing relative cougar environment possibilities

To help you design the risk of cougar situations, we operating an effective have fun with?access framework complement having fun with a good logistic regression (Manly, McDonald, Thomas, McDonald, & Erickson, 2007).