Table 3 Model selection results of apparent survival (φ) of adult barn owls breeding in western Switzerland in function of time (t: year), sex, dispersal status (status: immigrant vs. resident) and MC1R genotypes (MC1R: II/VI vs. VV)

From: Female-biased dispersal and non-random gene flow of MC1R variants do not result in a migration load in barn owls

Model

AICca

ΔAICcb

w c

K

Deviance

φ(Sex + status + t)

2191.23

0.00

0.34

37

785.17

φ(Sex + status + MC1R + t)

2192.63

1.40

0.17

38

784.46

φ(Sex + status + sex * status + t)

2193.34

2.11

0.12

38

785.17

φ(Sex + status + MC1R + sex * MC1R + t)

2193.80

2.57

0.09

39

783.51

φ(Sex + status + MC1R + MC1R * status + t)

2194.27

3.04

0.07

39

783.98

φ(Sex + status + MC1R + sex * status + t)

2194.74

3.51

0.06

39

784.46

φ(Sex + status + MC1R + sex * MC1R + MC1R * status + t)

2195.14

3.91

0.05

40

782.73

φ(Sex + status + MC1R + MC1R * status + sex * status + t)

2195.14

3.91

0.05

40

782.73

φ(Sex + status + MC1R + sex * MC1R + sex * status + t)

2195.92

4.69

0.03

40

783.51

φ(Sex + status + MC1R + sex * MC1R + sex * status + status * MC1R+ sex * status * MC1R + t)

2199.05

7.83

0.01

42

782.41

φ(Sex + t)

2199.40

8.17

0.01

36

795.45

φ(Sex + MC1R + t)

2200.56

9.33

0.00

37

794.50

φ(Sex + MC1R + sex * MC1R + t)

2201.60

10.37

0.00

38

793.43

φ(Status + t)

2205.87

14.64

0.00

36

801.91

φ(MC1R + status + t)

2207.69

16.46

0.00

37

801.62

φ(MC1R + status + MC1R * status + t)

2209.22

17.99

0.00

38

801.05

φ(MC1R + t)

2225.46

34.23

0.00

36

821.51

  1. Note that the model for recapture probability (ρ) was always time dependent (t) and is not included in the model notation. The most complex model for survival (i.e. (sex + status + MC1R + sex * MC1R + sex * status + status * MC1R+ sex * status * MC1R+ t)) corresponds exactly to model (g + t) from the first modeling step (see Table S1)
  2. K: Number of parameter estimated
  3. Deviance: model deviance
  4. aAIC value corrected for small sample sizes
  5. bDifference in a model’s AICc to the best-ranked model’s AICc
  6. cModel weight: probability of the model given the data