Agricultural land use shapes dispersal in white-tailed deer (Odocoileus virginianus)

Table 1 Summary of statistical models evaluating Wisconsin white-tailed deer dispersal

Response	Hypothesis class	Covariates
Dispersal* (logistic regression)	Null	aKDE-based home range area
		Number of pre-dispersal fixes
		Longitude of capture location
		Year
	Full	Agricultural + average first poor location + proportion potential paths intersecting roads
		Agricultural + average first poor location + proportion potential paths intersecting roads + average proximity OR number proximate
		Agricultural + average first poor location + proportion potential paths intersecting roads + number proximate + body weight
Log-transformed dispersal distance (linear regression)	Null	aKDE-based home range area
		Number of pre-dispersal fixes
		Longitude of capture location
		Year
		Season^†
	Full	Proportion potential paths intersecting rivers and streams + proportion intersecting roads + seasonaverage proportion of steps falling in agricultural land OR seasonagricultural
		Proportion potential paths intersecting rivers and streams + proportion intersecting roads + seasonaverage proportion of steps falling in agricultural land + seasonaverage proximity OR season*number proximate
		Proportion potential paths intersecting rivers and streams + proportion intersecting roads + seasonaverage proportion of steps falling in agricultural land + seasonnumber proximate + body weight

Note: Models with “OR” indicate variables that were not included in models together, but were selected via AIC. *Dispersal logistic regressions were fit separately for spring and fall models. ^†Season was highly significant for dispersal distances and was therefore included in all dispersal distance models. Abbreviated covariates were: aKDE home range area = autocorrelated kernel density estimation home range area; agricultural = proportion of pre-dispersal range classified as planted or agricultural land use; average first poor location = the average first simulated dispersal step in “developed” or “water” land types; proportion intersecting roads = proportion simulated paths intersecting roads; average proximity = an individual’s average proximity score across potential associations; number proximate = the number of individuals proximate to the focal individual (per available individuals within 4 km). Season was spring or fall; Year was a categorical variable for the years 2017–2020

ISSN: 2051-3933