Impacts of Human Modification on Predator Behavior
We estimated how human modification directionally affects bobcat and
coyote resource selection behavior in the form of functional responses
in habitat selection (Hebblewhite and Merrill 2008, Moreau et al. 2012).
We used univariate regressions with the individual RSF coefficients for
five covariates (agriculture, exurban, and other landcover; distance to
water; and distance to road) as the response variables and the
proportion of human modification in each home range (Kennedy et al.
2019) as the explanatory variables. These regressions were separated
based on temporal period to discern temporal effects on these functional
responses. In addition, regressions of the means of coefficients for
each covariate during all temporal periods for each level of human
modification in home range (each individual) were performed to find if a
broad functional response was present regardless of temporal period.
Weighted regressions were used to account for uncertainty associated to
the RSF coefficients (Bastille-Rousseau et al. 2021). Intercept-only,
linear, and quadratic regressions were compared using
AICC to determine the top regression model (Burnham and
Anderson 2002). The “other” landcover category lacked biological
meaning, so was not included in the final results (Supporting
Information).