Skip to main content

Table 3 Selecting the best GLMM for log-transformed daily distance (n = 1842 travel days)

From: Adaptive drift and barrier-avoidance by a fly-forage migrant along a climate-driven flyway

Model (fixed effects)

df

AIC

∆AIC

AIC Weight

R2 full

R2 fixed

R2 random

~ (tailwind_track + abs (sidewind_track)) * travel_hrs * biome

27

758.48

0.00

1.00

0.71

0.70

0.00

~ (tailwind_track + abs (sidewind_track)) * biome + season

12

801.46

42.98

0.00

0.20

0.19

0.01

~ biome * travel_hrs

9

817.81

59.33

0.00

0.68

0.67

0.01

~ (tailwind_track + abs (sidewind_track)) * travel_hrs + biome

15

849.94

91.46

0.00

0.70

0.69

0.00

~ travel_hrs * abs (sidewind_track)

10

859.93

101.45

0.00

0.65

0.63

0.01

~ tailwind_track * travel_hrs

9

878.35

119.87

0.00

0.68

0.67

0.01

~ abs (sidewind_track) + travel_hrs

7

879.77

121.29

0.00

0.65

0.63

0.01

~ biome + abs (sidewind_track)

7

895.05

136.57

0.00

0.14

0.13

0.01

~ season + abs (sidewind_track)

6

908.30

149.82

0.00

0.03

0.02

0.01

~ season + biome + travel_hrs

7

909.63

151.15

0.00

0.67

0.67

0.00

~ season * tailwind_track

9

930.52

172.04

0.00

0.08

0.07

0.01

~ (tailwind_track + abs (sidewind_track)) * season

11

943.53

185.05

0.00

0.09

0.08

0.01

~ biome + tailwind_track + abs (sidewind_track)

8

992.51

234.03

0.00

0.18

0.18

0.01

~ tailwind_track + abs (sidewind_track)

7

1028.47

269.99

0.00

0.09

0.08

0.01

~ season + tailwind_track + biome

7

1045.48

287.00

0.00

0.18

0.17

0.01

~ season + tailwind_track

6

1055.16

296.68

0.00

0.08

0.07

0.01

~ season + abs (sidewind_track) + travel_hrs

7

1105.61

347.12

0.00

0.66

0.65

0.00

~ biome + travel_hrs

6

1110.26

351.78

0.00

0.67

0.66

0.01

~ abs (sidewind_track)

5

1125.04

366.56

0.00

0.02

0.01

0.01

~ (tailwind_track + abs (sidewind_track)) * travel_hrs * season

24

2575.59

1817.11

0.00

0.69

0.69

0.00

~ (tailwind_track + abs (sidewind_track)) * travel_hrs + season

15

2586.38

1827.90

0.00

0.69

0.68

0.00

~ (tailwind_track + abs (sidewind_track)) * biome * season

16

2587.98

1829.50

0.00

0.21

0.20

0.01

~ biome * tailwind_track

9

2611.55

1853.07

0.00

0.18

0.17

0.01

~ travel_hrs + tailwind_track + abs (sidewind_track)

8

2619.54

1861.06

0.00

0.68

0.68

0.01

~ (tailwind_track + abs (sidewind_track)) * biome

11

2619.68

1861.20

0.00

0.20

0.19

0.01

~ season * travel_hrs

9

2621.50

1863.02

0.00

0.65

0.65

0.00

~ (tailwind_track + abs (sidewind_track)) * travel_hrs

11

2685.52

1927.04

0.00

0.68

0.68

0.01

~ season * abs (sidewind_track)

9

2691.59

1933.11

0.00

0.03

0.02

0.01

~ tailwind_track * abs (sidewind_track)

10

2695.00

1936.52

0.00

0.09

0.08

0.01

~ biome + tailwind_track + travel_hrs

8

2697.32

1938.84

0.00

0.69

0.69

0.00

~ season * biome

8

2704.51

1946.03

0.00

0.15

0.14

0.01

~ biome + tailwind_track

7

2707.39

1948.91

0.00

0.18

0.17

0.01

~ season + travel_hrs

6

2808.74

2050.26

0.00

0.65

0.65

0.00

~ season + biome + abs (sidewind_track)

7

2809.29

2050.81

0.00

0.15

0.14

0.01

~ tailwind_track + travel_hrs

7

2810.58

2052.10

0.00

0.68

0.67

0.01

~ biome * abs (sidewind_track)

9

2810.77

2052.29

0.00

0.15

0.15

0.01

~ biome

5

2832.09

2073.61

0.00

0.14

0.13

0.01

~ travel_hrs

6

2834.03

2075.55

0.00

0.64

0.63

0.01

~ season + tailwind_track + travel_hrs

7

2835.69

2077.21

0.00

0.68

0.68

0.00

~ season + tailwind_track + abs (sidewind_track)

7

2928.90

2170.42

0.00

0.09

0.08

0.01

~ season + biome

6

2932.39

2173.91

0.00

0.14

0.14

0.01

~ tailwind_track

5

2950.18

2191.70

0.00

0.08

0.07

0.01

~ season

5

2951.18

2192.70

0.00

0.02

0.01

0.01

~ (1 | dev) + (1 | yr)

4

2963.73

2205.25

0.00

0.01

0.00

0.01

  1. We tested an exhaustive set of models including fixed effects of tailwind and sidewind along the realized travel direction, daily travel time, season, their additive effects, and interaction effects between wind variables and daily travel time, biome and season. We further allowed intercepts to vary randomly between individuals and between years. Models are ranked according to increasing ∆AIC values, with the best performing model on top