Table 4.
Model selection results for analysis of seasonal/monthly variation in bird-window collisions.
| Parameters | Ka | ΔAICcb | AIC weightc |
|---|---|---|---|
| Season * ResStatus | 7 | 0.0 | > 0.99 |
| Null | 2 | 105.8 | < 0.01 |
| Season | 4 | 108.2 | < 0.01 |
| Month | 3 | 108.5 | < 0.01 |
| Season + ResStatus | 5 | 108.9 | < 0.01 |
| ResStatus | 3 | 108.9 | < 0.01 |
| Month * ResStatus | 5 | 110.2 | < 0.01 |
| Month + ResStatus | 4 | 112.5 | < 0.01 |
Results of model selection comparison among candidate negative binomial models that included factors potentially affecting raw counts of total fatal collisions (including feather piles) found in morning, midday, and evening collision surveys in Stillwater, Oklahoma, USA in 2015–2016. Parameters included Month (4 = April, 5 = May, etc.), Season (spring, summer, fall) and ResStatus (residency status of individual birds: resident and migrant birds were included in analysis; individuals of unknown residency status and of = unidentified species were excluded from analysis).
aNumber of model parameters.
bDifference in AICc value between model and top model.
cAICc Weight—relative support for model.