Table 3. Model results showing the effect of setting a camera on a feature (game trail or log) based on paired comparisons with nearby random locations.
| Species | Trail Feature (ΔAICc) |
Log Feature (ΔAICc) |
|---|---|---|
| Odocoileus virginianus |
+ (20.44) |
- (37.77)§ |
| Didelphis virginiana |
+ (8.09) |
- (2.17) |
| Tamias striatus* |
+ (7.56) |
NA |
| Sciurus carolinensis |
+ (2.28)☩ |
+ (49.04)¶ |
| Sylvilagus floridanus* |
+ (1.95) |
NA |
| Procyon lotor | No effect (2.21) |
+ (3.82) |
| Ursus americanus | No effect (3.66) |
No effect (4.11) |
| Mouse (Peromyscus sp.)✝ | NA |
+ (37.78) |
| Sciurus niger✝ | NA |
+ (23.24) |
| Urocyon cinereoargenteus | NA | No effect (3.69) |
The data columns show the effect of feature presence on detection; the direction of this difference (positive [+] or negative [–]), and the ΔAICc value between the global detection model (all covariates) with and without feature type. An NA value indicates that there were insufficient detections for a given species in the camera pairs for a certain feature to allow investigation. All detection models were run with the top model for local occupancy (θ) of the species. The null model for theta (θ (.)) was the top model for almost all species; three exceptions are noted in the table.
* Trail data only sufficient to support basic models. Covariates on θ and p not investigated.
✝ Log data only sufficient to support basic models. Covariates on θ and p not investigated.
§The model used for calculating detection differences was ψ(.) θ (CovAll)
☩The model used for calculating detection differences was ψ(.) θ (CovAll)
¶The model used for calculating detection differences was ψ(.) θ (global)