Table 3. Ranking of models explaining the fix success rate (FSR) obtained by lightweight GPS collars in different suburban habitat types and sky availability classes during stationary collar tests (n = 36), Dunedin, New Zealand.
| Model Description | K | AICc | ΔAICc | wi | Model Likelihood |
| Sky Availability | 3 | 221.2 | 0.00 | 0.50 | 1.00 |
| Sky Availability+Vegetation complexity | 4 | 223.6 | 2.35 | 0.15 | 0.31 |
| Sky Availability+Distance to buildings | 4 | 223.6 | 2.36 | 0.15 | 0.31 |
| Null model | 2 | 224.7 | 3.50 | 0.09 | 0.17 |
| Sky availability+Vegetation complexity+Distance to buildings | 5 | 226.1 | 4.86 | 0.04 | 0.09 |
| Distance to buildings | 3 | 226.6 | 5.35 | 0.03 | 0.07 |
| Vegetation complexity | 3 | 227.0 | 5.74 | 0.03 | 0.06 |
| Vegetation complexity+Distance to buildings | 4 | 229.0 | 7.73 | 0.01 | 0.02 |
K = number of parameters; ΔAIC = change in AIC; wi = Akaike weight.
Models were ranked based on the Akaike’s second-order corrected Information Criterion (AICc).