Table 1.

Habitat configuration/quality and human relation predictors evaluated when modelling initial occupancy (ψ₁), colonisation (γ), extinction (ε) and detection (p) probability parameters of multiseason camera‐trap güiña (Leopardus guigna) surveys. Further details can be found in Appendix S1, S2 and Table S1

Parameter	Predictor	Abbreviation in models
	Habitat configuration
ψ1, ε, γ	Percentage of forest cover/habitata	Forest
ψ1, ε, γ	Percentage of shrub cover/marginal habitat	Shrub
ψ1, ε, γ	Number of forest patches	PatchNo
ψ1, ε, γ	Shape index forest patches	PatchShape
ψ1, ε, γ	Forest patch size areab	PatchAreaW
ψ1, ε, γ	Forest patch continuityb	Gyration
ψ1, ε, γ	Edge length of forest land cover class	Edge
ψ1, ε, γ	Landscape shape index of forestc	LSI
ψ1, ε, γ	Patch cohesionb	COH
	Human–predator relation
ψ1, ε	Land subdivision	Subdivision
ψ1, ε	Intent to kill (hypothetical scenario questions)	Intent
ψ1, ε	Predation	Predation
ψ1, ε	Frequency of predation	FQPredation
ψ1, ε, p	Frequency of encounterd	FQEncounter
ψ1, ε	Number of dogs	Dogs
	Habitat quality
p	Bamboo density (Chusquea spp.)	Bamboo
p	Density of understorey	Understorey
p	Sample Unit rotation block	Rotation
p	Intensity of livestock activity	Livestock
p	Intensity of logging activity	Logging
p	Water availability	Water

^aPools together all forest types: old growth, secondary growth and wetland forest.

^bPredictor excluded due to collinearity with percentage of forest cover (Pearson's │r│ > .7).

^cPredictor excluded due to collinearity with number of forest patches (Pearson's │r│>.7).

^dPredictor also fitted with detection probability.