Skip to main content

Table 4 Parameter estimates for the models of Bewick’s Swan behaviour and energetics with the lowest AICc values

From: Behavioural and energetic consequences of competition among three overwintering swan (Cygnus spp.) species

Variable

Effect type

Parameter

VIF

Mean

SE

Variance

SD

Aggression

Fixed

Intercept

– 2.09

0.83

  

A

1.94

– 0.25

0.18

  

C

1.00

5.5 × 10–4

7.4 × 10–4

  

A:C

1.94

1.4 × 10–2

3.3 × 10–3

 

Random

Observation ID

16.7

4.1

Foraging

Fixed

Intercept

− 0.91

0.19

  

A

2.00

0.40

0.05

  

B

1.21

8.00 × 10–4

4.00 × 10–4

  

C

1.21

5.00 × 10–4

1.00 × 10–4

  

A:C

1.99

– 1.40 × 10–3

8.00 × 10–4

 

Random

Observation ID

0.79

0.89

  

Residual

0.71

0.84

Vigilance

Fixed

Intercept

– 0.94

0.17

  

A

1.00

– 0.23

0.03

  

B

1.06

− 3.00 × 10–4

3.00 × 10–4

  

C

1.10

6.00 × 10–4

4.00 × 10–4

  

B:C

1.05

– 3.59 × 10–6

– 3.96 × 10–6

 

Random

Observation ID

0.61

0.78

  

Residual

0.48

0.69

Net energy gain

Fixed

Intercept

0.79

0.08

 

Random

Observation ID

0.26

0.51

  

Residual

0.21

0.46

  1. The variance and SD associated with the random effects, as well as the Variance Inflation Factor (VIF) associated with each parameter, are also indicated. Foraging and vigilance variables were logit-transformed response variable, whilst net energy gain was rescaled logit-transformed response variable. Model parameters: intercept (1), density of Bewick’s Swan sharing feeding habitat with focal individual (B), density of Mute and Whooper Swans sharing feeding habitat with focal individual (C), age class of focal individual (A), and observation identity (U). Adult parameter values refer to cygnets relative to adults