Skip to main content

Table 1 Binomial GLMM summary describing the probability of migration departure with respect to weather conditions experienced during the 3 days preceding departure and the departure day

From: Weather, sex and body condition affect post-fledging migration behaviour of the greater flamingo Phoenicopterus roseus

Probability of migration departure

Predictors

Estimate [95% CI]

SE

χ2

df

P

Intercept

 − 4.97 [− 7.49, − 2.44]

1.69

14.92

1

0.001

Crosswind

 − 0.27 [− 1.08, 0.53]

0.41

0.45

1

0.503

Wind speed

0.68 [− 018, 1.55]

0.44

2.38

1

0.120

Tailwind

1.41 [0.75, 2.07]

0.34

17.57

1

 < 0.0001

Air pressure

0.5 [− 0.54, 1.55]

0.53

0.89

1

0.345

Humidity

 − 0.51 [− 1.30, 0.28]

0.4

1.6

1

0.205

Cloud cover (medium)

 − 0.53 [− 1.85, 0.80]

0.68

0.6

1

0.438

Cloud cover (high)

 − 0.11 [− 0.57, 0.35]

0.23

0.22

1

0.634

Cloud cover (low)

 − 0.4 [− 1.64, 0.83]

0.63

0.41

1

0.523

Temperature

0.96 [− 0.28, 2.19]

0.63

2.31

1

0.128

Rain

0.19 [− 0.42, 0.79]

0.31

0.36

1

0.548

Colony (Margherita di Savoia)

3.09 [0.28, 5.91]

1.43

7.03

2

0.031

Colony (Molentargius)

0.57 [− 0.28, 3.42]

1.45

7.03

2

0.695

Year (2016)

 − 0.14 [− 2.24, 1.97]

1.07

0.017

2

0.898

Year (2017)

 − 0.17[− 3.40, 3.07]

1.65

0.017

2

0.919

Sex (Male)

 − 0.36 [− 1.54, 0.83]

0.6

0.347

1

0.555

Hours before sunset

4.37 [2.34, 6.41]

1.04

17.78

1

 < 0.0001

Hours before sunset2

 − 2.52 [− 3.89,  − 1.14]

0.7

12.87

1

0.0003

  1. Response variable was coded as “0” (n = 676) representing all locations 3 days prior to departure and as “1” (n = 30) the first location in flight. Predictors included sex, colony (reference level: Comacchio), year (reference level: 2015), and weather variables. Standardized estimates are shown alongside 95% confidence intervals, test statistics (Wald’s χ2, type III), degrees of freedom (df) and p-value (P). Predictors with significant p-values (α = 0.05) and with 95% confidence intervals not overlapping zero are shown in bold. Effects of biometrics and their interaction with weather variables were assessed preliminary and excluded as they were considered to have little influence