Ecological Archives E094-263-A1

Bert Hidding, Jean-Pierre Tremblay, Steeve D. Côté. 2013. A large herbivore triggers alternative successional trajectories in the boreal forest. Ecology 94:2852–2860. http://dx.doi.org/10.1890/12-2015.1

Appendix A. Statistical tables of the effects of immediate and delayed deer exclusion on woody regeneration on Anticosti Island, Quebec, Canada.

Table A1. Response of the tree community to the immediate and delayed exclusion of white-tailed deer on Anticosti Island on stem density (dbh > 1 cm) of all species and of Picea glauca in 2010. Models were fit assuming a Poisson distribution using the lme4 library in R. To correct for overdispersion, an observation level random effect was included. Likelihood ratio tests (LRT) compared models including the variable in question and models without them. A post hoc analysis was performed by merging two factor levels of treatment and contrasting it to the remaining factor, again using LRTs. Post hoc tests are designated by the codes of the only remaining factor level. LogLik denotes the log-likelihood. EXC = original exclosures, CTL = control plots, DEL = delayed exclosures. Significant p values indicated in bold font.

All species

df

logLik

χ2

p

Full model

8

-41.5

 

 

Treatment × Distance

6(2)

-41.97

0.9

0.627

Treatment

4(2)

-52.2

20.5

<0.001

Distance

5(1)

-42.28

0.6

0.427

EXC = DEL

4(1)

-46.22

7.9

0.005

DEL = CTL

4(1)

-44.96

5.4

0.021

EXC = CTL

4(1)

-52.31

20.0

<0.001

 

 

 

 

 

Picea glauca

df

logLik

χ2

p

Full model

8

-41.21

 

 

Treatment × Distance

6(2)

-41.44

0.4

0.801

Treatment

4(2)

-46.91

10.9

0.004

Distance

5(1)

-43.11

3.3

0.068

EXC = DEL

4(1)

-48.11

10.0

0.002

DEL = CTL

4(1)

-44.38

2.6

0.111

EXC = CTL

4(1)

-44.75

3.3

0.07

 

Table A2. Analysis of deviance on early and delayed exclusion effects on sapling and seedling densities and mean stem height of tree species over time since herbivore exclusion. Sapling and seedling densities were count variables and were modeled with a Poisson distribution, with forest edge distance as a covariate in lme4 in R. Mean stem height was modeled on transformed data (4√, log- and 2√ respectively) using a heterogeneous error structure in nlme in R. Likelihood ratio tests (LRT) compared models including the variable in question and models without them. Degrees of freedom used in the LRTs are indicated within brackets. Post hoc analyses were performed by merging two treatment factor levels followed by LRTs. EXC = original exclosures, CTL = control plots, DEL = delayed exclosures. Significant p values indicated in bold font.

 

Stem density tall regeneration

 

Seedling density

 

Stem height

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Picea glauca

df

logLik

χ2

p

 

df

logLik

χ2

p

 

df

logLik

χ2

p

Full model

9

-187.93

 

 

 

9

-157.18

 

 

 

11

101.53

 

 

distance

8(1)

-188.29

0.7

0.391

 

8(1)

-157.75

1.1

0.285

 

 

 

 

 

Treat. × Time

7(2)

-192.31

8.8

0.012

 

7(2)

-161.46

8.6

0.014

 

9(2)

98.42

6.22

0.045

Treatment

5(2)

-197.01

9.4

0.009

 

5(2)

-170.01

17.1

<0.001

 

7(2)

64.71

67.41

<0.001

Time

6(1)

-199.81

15.0

<0.001

 

6(1)

-192.43

61.9

<0.001

 

8(1)

43.36

110.12

<0.001

EXC = DEL

7(2)

-196.88

17.9

<0.001

 

7(2)

-169.4

24.4

<0.001

 

9(2)

66.72

69.61

<0.001

DEL = CTL

7(2)

-191.67

7.5

0.024

 

7(2)

-166.27

18.2

<0.001

 

9(2)

72.92

57.21

<0.001

EXC = CTL

7(2)

-190.32

4.8

0.092

 

7(2)

-158.65

2.9

0.23

 

9(2)

98.28

6.49

0.039

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

A. balsamea

df

logLik

χ2

p

 

df

logLik

χ2

p

 

df

logLik

χ2

p

Full model

9

-101.4

 

 

 

9

-202.53

 

 

 

11

-105.84

 

 

distance

8(1)

-102.32

1.9

0.173

 

8(1)

-202.55

0.0

0.848

 

 

 

 

 

Treat. × Time

7(2)

-113.26

23.7

<0.001

 

7(2)

-208.42

11.8

0.003

 

9(2)

-112.38

13.1

0.001

Treatment

5(2)

-124.82

23.1

<0.001

 

5(2)

-224.07

31.3

<0.001

 

7(2)

-117.98

11.2

0.004

Time

6(1)

-136.12

45.7

<0.001

 

6(1)

-237.47

58.1

<0.001

 

8(1)

-148.13

71.5

<0.001

EXC = DEL

7(2)

-108.99

15.2

<0.001

 

7(2)

-220.16

35.3

<0.001

 

9(2)

-105.86

0.1

0.976

DEL = CTL

7(2)

-108.84

14.9

<0.001

 

7(2)

-210.16

15.3

<0.001

 

9(2)

-114.46

17.2

<0.001

EXC = CTL

7(2)

-124.6

46.4

<0.001

 

7(2)

-213.5

21.9

<0.001

 

9(2)

-117.58

23.5

<0.001

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

B. papyrifera

df

logLik

χ2

p

 

df

logLik

χ2

p

 

df

logLik

χ2

p

Full model

9

-106.78

 

 

 

9

-160.7

 

 

 

11

23.98

 

 

distance

8(1)

-108.1

2.63

0.105

 

8(1)

-163.73

6.06

0.014

 

 

 

 

 

Treat. × Time

7(2)

-107.02

0.47

0.791

 

7(2)

-164.03

6.66

0.036

 

9(2)

-5.54

59.05

<0.001

Treatment

5(2)

-135.94

57.85

<0.001

 

5(2)

-165.8

3.55

0.169

 

7(2)

-24.04

37

<0.001

Time

6(1)

-128

41.96

<0.001

 

6(1)

-165.52

2.98

0.084

 

8(1)

-12.52

13.95

<0.001

EXC = DEL

7(2)

-107.66

1.75

0.418

 

7(2)

-162.66

3.93

0.14

 

9(2)

15.54

16.89

<0.001

DEL = CTL

7(2)

-126.59

39.61

<0.001

 

7(2)

-163.58

5.76

0.056

 

9(2)

1.41

45.15

<0.001

EXC = CTL

7(2)

-134.21

54.86

<0.001

 

7(2)

-163.74

6.08

0.048

 

9(2)

-15.11

78.19

<0.001


[Back to E094-263]