Table 2.
Snow cover, day length, temperature and NDVI values extracted at each location during spring (at one-day step length) and compared between days (from 10 days in the past to 10 days in the future), separately for each species.
| Species (n = number of locations) | Dep. var. | Pred. | Est. | SE | z/t value | p value | ΔAIC | AIC Weight | LL |
|---|---|---|---|---|---|---|---|---|---|
| Snowy owl n = 67572 | Snow cover | (Intercept) | 1.70 | 0.51 | 3.35 | <0.001 | 3461* | 1.0 | −30794 |
| Day | −0.10 | 0.00 | −56.70 | <0.001 | |||||
| Day length | (Intercept) | 15.44 | 0.37 | 41.64 | <0.001 | 446* | 1.0 | −181330 | |
| Day | 0.05 | 0.00 | 21.21 | <0.001 | |||||
| Temperature | (Intercept) | 0.82 | 1.23 | 0.67 | 0.51 | 4543* | 1.0 | −224904 | |
| Day | 0.30 | 0.00 | 68.57 | <0.001 | |||||
| NDVI | (Intercept) | −0.40 | 0.05 | −7.70 | <0.001 | 3420* | 1.0 | 2472 | |
| Day | 0.01 | 0.00 | 59.25 | <0.001 | |||||
| Rough-legged buzzard n = 133461 | Snow cover | (Intercept) | −1.04 | 0.15 | −7.00 | <0.001 | 8186* | 1.0 | −72895 |
| Day | −0.10 | 0.00 | −87.11 | <0.001 | |||||
| Daylength | (Intercept) | 15.44 | 0.23 | 67.46 | <0.001 | 501* | 1.0 | −366250 | |
| Day | 0.04 | 0.00 | 22.44 | <0.001 | |||||
| Temperature | (Intercept) | 7.07 | 0.44 | 15.93 | <0.001 | 6314* | 1.0 | −417478 | |
| Day | 0.20 | 0.00 | 80.43 | <0.001 | |||||
| NDVI | (Intercept) | −0.22 | 0.02 | −11.16 | <0.001 | 7321* | 1.0 | −42750 | |
| Day | 0.01 | 0.00 | 86.76 | <0.001 | |||||
| Peregrine falcon n = 6575 | Snow cover | (Intercept) | −3.90 | 1.08 | −3.61 | <0.001 | 134* | 1.0 | −899 |
| Day | −0.13 | 0.01 | −10.90 | <0.001 | |||||
| Day length | (Intercept) | 16.66 | 0.83 | 20.04 | <0.001 | 185* | 1.0 | −14843 | |
| Day | 0.06 | 0.00 | 13.76 | <0.001 | |||||
| Temperature | (Intercept) | 6.65 | 1.76 | 3.78 | 0.007 | 299* | 1.0 | −17929 | |
| Day | 0.13 | 0.01 | 17.56 | <0.001 | |||||
| NDVI | (Intercept) | 0.13 | 0.21 | 0.63 | 0.55 | 107* | 1.0 | −1079 | |
| Day | 0.01 | 0.00 | 10.46 | <0.001 |
Generalised linear mixed models (GLMM) with snow cover, day length, temperature and NDVI as dependent variables (each in a separate model), day as a predictor variable, and year nested in individual as a random effect. To evaluate the effect of the predictor, we compared a model with and without that predictor. Only the results of the full models are presented and those with ΔAIC > 2 are marked with *.