Table 2.
Models of annual variation in survival of Raso larks (2004–2017, n = 732).
| Model | Npar | Deviance | ΔAIC | Rel. Dev. |
|---|---|---|---|---|
| sex * year | 30 | 3478.3 | 0 | 1 |
| sex + year | 18 | 3507.5 | 4.72 | 0.816 |
| year | 17 | 3509.8 | 5.00 | 0.800 |
| sex + sex:PopSize + sex:Rain + sex:PopSize:Rain | 12 | 3532.6 | 17.6 | 0.657 |
| sex + sex: PopSize + sex:Rain + PopSize:Rain | 11 | 3536.0 | 19.7 | 0.631 |
| sex + PopSize + Rain + PopSize:Rain | 9 | 3553.5 | 32.5 | 0.525 |
| sex + Rain + sex:Rain | 8 | 3594.6 | 71.5 | 0.265 |
| sex + PopSize + sex:PopSize | 8 | 3611.7 | 93.3 | 0.128 |
| sex + Rain | 7 | 3600.4 | 75.3 | 0.229 |
| sex + PopSize | 7 | 3629.5 | 104 | 0.045 |
| sex | 6 | 3636.6 | 110 | 0 |
For each modelled set of covariates is given the number of parameters (Npar), the total deviance explained, AIC relative to the best model (ΔAIC), and the deviance relative to the full (ϕsex*year) and null (ϕsex) models (Rel. Dev.). Capitalised parameters are linear, those in lower-case factors (with 2, sex, or 13, year, levels). Models are listed in descending order of complexity and all include four re-encounter parameters (ptd*sex, see text).