Predictor Variables
We identified 30 predictor variables that were either previously
linked to moose ecology or of interest to WLFN (Table 1). These
variables fell into three broad categories: land cover, human features
(or footprint), and burn area. Various land cover and human feature
categories were combined based on their similar impacts (e.g., oil
wells, gas wells, and other types of wells were all binned under
“well”) or structures/functions (e.g., bogs and wetlands). We z-scaled
all predictors (mean=0, s.d. = 1) before checking for spatial
autocorrelation ( Zuur et al., 2010). We then calculated the
area covered by each predictor within different buffer sizes surrounding
camera locations from 250 to 5000m radii, sensu Fisher et al. (2011)