Table 3.
Movement decisions of arctic raptors according to environmental predictors (snow cover, wind support, and crosswind) at one-day step length.
| Species (n = number of locations) | Pred. | Coef. | SE | z value | p value | ΔAIC | AIC Weight | LL |
|---|---|---|---|---|---|---|---|---|
| Snowy owl n = 41316 | Snow cover | −0.01 | 0.04 | −0.13 | 0.90 | 2 | 0.3 | −35026 |
| Wind support | −0.05 | 0.00 | −10.99 | <0.001 | 118* | 1.0 | ||
| Crosswind | −0.00 | 0.01 | −0.52 | 0.61 | 2 | 0.3 | ||
| Rough-legged buzzard n = 72479 | Snow cover | −0.05 | 0.03 | −1.88 | 0.06 | 2 | 0.7 | −72409 |
| Wind support | 0.01 | 0.00 | 2.92 | 0.004 | 6* | 1.0 | ||
| Crosswind | 0.06 | 0.00 | 17.02 | <0.001 | 282* | 1.0 | ||
| Peregrine falcon n = 4147 | Snow cover | 0.07 | 0.25 | 0.29 | 0.77 | 2 | 0.3 | −2753 |
| Wind support | −0.06 | 0.02 | −3.53 | <0.001 | 10* | 1.0 | ||
| Crosswind | 0.03 | 0.02 | 1.78 | 0.08 | 1 | 0.6 |
We performed mixed conditional logistic regression models with movement choice (chosen vs alternative locations) as a dependent variable, snow cover, wind support, and crosswind as predictors and stratum nested in individual as a random effect. We performed the models separately by species. To evaluate the effect of the predictor, we compared models with and without that predictor. Only the results of the full models are presented and those with lower AIC and ΔAIC > 2 are marked with *.