Table 2. Carnivore species detected during the camera trap survey in Mole National Park, Ghana, and estimated mean occurrence (ψ) and detection (p) probabilities and covariate effects on occurrence.

Common namea	Prop. sites	ψ (SD)	p (SD)	Covariate effects indicatedb
Spotted hyena	0.442	0.544 (0.050)	0.173 (0.039)	small prey(+), riverine(+), edge(−), hunting(−), seasonal NDVI(+)
Leopard	0.299	0.526 (0.077)	0.140 (0.038)	small prey(+), riverine(+), patrol(−), hunting(+)
White-tailed mongoose	0.259	0.292 (0.039)	0.119 (0.031)	small prey(+), riverine(−), seasonal NDVI(−), patrol(−)
Large-spotted genet	0.246	0.263 (0.037)	0.146 (0.041)	small prey(+), edge(+), hunting(+), seasonal NDVI(+)
African civet	0.098	0.189 (0.062)	0.123 (0.047)	small prey(+)
Caracal	0.054	0.096 (0.045)	0.100 (0.047)	riverine(−), small prey(+)
Marsh mongoose	0.049	0.095 (0.053)	0.124 (0.060)	small prey(+)
Gambian mongoose	0.018	0.075 (0.073)	0.094 (0.053)	small prey(+)
Side-striped jackal	0.013	0.072 (0.089)	0.087 (0.054)	small prey(+)

The proportion of 224 sampling sites at which carnivore species were detected reflects observation data, whereas ψ and p are model-averaged estimates from the multi-species hierarchical mixture model (means and standard deviations from posterior probability distributions for species-specific parameters). Site covariates of occurrence are shown for cases where the posterior probability distribution from the full model for the corresponding species-specific coefficient indicated a potential effect (i.e., posterior mass not concentrated at 0; distributions are given in Appendix S3).

^aScientific names in Table S1.

^bDirection of effect indicated as either positive (+) or negative (−) association of species occurrence probability with the particular covariate. For the different prey biomass covariates, only the strongest effect is indicated.