Ecological Archives E096-237-A3

Tom P. Bregman, Alexander C. Lees, Nathalie Seddon, Hannah E. A. Macgregor, Bianca Darski, Alexandre Aleixo, Michael B. Bonsall, and Joseph A. Tobias. 2015. Species interactions regulate the collapse of biodiversity and ecosystem function in tropical forest fragments. Ecology 96:26922704. http://dx.doi.org/10.1890/14-1731.1

Appendix C. Supplementary results with tables showing functional trait Principal Components and additional model results.

Table C1. Principal Component Analysis (PCA) for insectivores. Eigenvalues and the proportion of variance explained are presented. PC1 for locomotory and trophic traits were used in a second PCA to create an axis representing overall size. PC2 for locomotory and trophic traits represented variation of these syndromes independent of body size, and was used in further analysis.

PCA

PC

Proportion variance

PCA loadings

Locomotory

 

 

Tarsus length

Tail length

Wing chord

1

0.734

–0.529

–0.569

–0.630

2

0.198

0.779

–0.620

–0.095

Trophic

 

 

Beak length

Beak width

Beak depth

1

0.785

–0.526

–0.594

–0.608

2

0.162

0.845

–0.446

–0.669

Overall size

 

 

Trophic PC1

Locomotory PC1

1

0.858

0.707

0.707

 

Table C2. Principal Component Analysis (PCA) for frugivores. Eigenvalues and the proportion of variance explained are presented. PC1 for locomotory and trophic traits were used in a second PCA to create an axis representing overall body size. PC2 for locomotory and trophic traits represented variation of these syndromes independent of body size, and was used in further analysis.

PCA

PC

Proportion variance

PCA loadings

Locomotory

 

 

Tarsus length

Tail length

Wing chord

1

0.797

0.514

0.583

0.628

2

0.179

–0.821

0.546

0.165

Trophic

 

 

Beak length

Beak width

Beak depth

1

0.897

0.566

0.588

0.578

2

0.070

–0.802

0.229

0.776

Overall size

 

 

Trophic PC1

Locomotory PC1

1

0.804

–0.707

–0.707

 

Table C3. Phylogenetic signal of trait axes using Blomberg’s K. K values close to 1 indicate a Brownian motion model of trait evolution whilst a value of 0 indicates a random model of trait evolution.

Guild

Trait

K

P

Insectivores

Locomotory

0.59

<0.001

Trophic

0.54

<0.001

Dispersal

0.62

<0.001

Overall size

1.38

<0.001

Frugivores

Locomotory

1.11

<0.001

Trophic

0.53

<0.001

Dispersal

0.71

<0.001

Overall size

0.47

<0.001

 

Table C4. Average fixed effect estimates derived from the top linear models, <2AICc from the best fitting model. Standardized estimates for community assembly metrics were used as the response variable and environmental variables as the fixed effects. FS, Fragment size (ha); GF, Percentage of fragment covered in good quality forest; D, Distance of fragment away from nearest forest >1000 ha.

Guild

Trait syndrome

Metric

Variable

Estimate (SE)

Z value

P value

Insectivores

Phylogeny

MNTD

FS

–0.580 (0.143)

4.047

<0.001

Locomotory

SDNDr

GF

0.496 (0.381)

1.247

0.212

FS

0.516 (0.367)

1.345

0.179

D

–0.490 (0.380)

1.239

0.215

Variance

GF

–0.357 (0.315)

1.083

0.279

FS

–0.354 (0.315)

1.074

0.283

D

0.218 (0.320)

0.653

0.514

Dispersal

SDNDr

GF

0.564 (0.329)

1.637

0.102

FS

0.377 (0.339)

1.063

0.288

Variance

D

–0.952 (0.539)

1.700

0.089

FS

–0.859 (0.455)

1.827

0.067

D*FS

0.683 (0.621)

1.048

0.295

Foraging

SDNDr

D

–0.920 (0.555)

1.615

0.106

FS

–0.678 (0.402)

1.608

0.108

D*FS

0.68 (0.61)

1.059

0.290

Variance

D

–0.297 (0.432)

1.099

0.272

FS

–0.852 (0.393)

2.087

0.037

Overall size

SDNDr

FS

0.884 (0.335)

2.533

0.011

GF

0.384 (0.351)

1.042

0.297

Variance

FS

–0.297 (0.131)

2.271

0.031

Frugivores

Phylogeny

MNTD

D

0.673 (0.496)

1.296

0.195

FS

1.111 (0.517)

2.083

0.037

GF

–0.487 (0.398)

0.417

0.243

FS*GF

–2.013 (0.802)

2.392

0.017

Locomotory

SDNDr

D

–0.266 (0.239)

1.063

0.288

FS

0.255 (0.240)

1.017

0.309

Variance

GF

0.060 (0.250)

0.241

0.811

Dispersal

SDNDr

D

–1.752 (0.844)

1.998

0.046

FS

0.577 (0.467)

1.178

0.239

GF

–0.900 (0.505)

1.692

0.091

D*FS

2.364 (0.755)

2.987

0.003

D*GF

2.252 (1.300)

1.644

0.100

Variance

D

0.816 (0.402)

1.950

0.051

GF

0.469 (0.431)

1.040

0.298

Foraging

SDNDr

D

0.879 (0.874)

0.963

0.336

FS

1.212 (0.509)

2.271

0.023

GF

–0.778 (0.527)

1.402

0.161

D*FS

–2.455 (0.878)

2.678

0.007

FS*GF

–2.729 (1.193)

2.172

0.030

Variance

FS

–1.032 (0.318)

3.113

0.002

GF

0.273 (0.345)

0.756

0.450

Overall

SDNDr

D

–0.531 (0.389)

1.316

0.188

GF

–0.494 (0.397)

1.201

0.230

Variance

D

–0.176 (0.340)

0.498

0.618

FS

0.048 (0.333)

0.346

0.890

GF

0.505 (0.296)

1.633

0.103

 

FigC1

Fig. C1. Structure of insectivore communities (n = 30) in fragments with different habitat characteristics. Regression lines are shown for significant relationships (derived from top model sets), and represent the fit of a model with individual habitat variables as the sole predictor of the structuring metrics. Final model outputs are included in Table C4.


 

FigC2

Fig. C2. Structure of frugivore communities (n = 30) in fragments with different habitat characteristics. Regression lines are shown for significant relationships (derived from top model sets), and represent the fit of a model with individual habitat variables as the sole predictor of the structuring metrics. Final model outputs are included in Table C4.


 

Table C5. Community structure in assemblages of avian insectivores surviving in tropical forest fragments (n = 30). Asterisks (*) indicate observed values for community structure metrics (MNTD, SDNDr, variance) significantly different from the null model in the direction providing evidence for competitive interactions (variance is predicted to be greater than expected; SDNDr is predicted to be lower than expected). The standardized effect size (SES) is provided along with the standard error (SE). Statistics (V) and P values are from Wilcoxon signed rank tests (two-tailed). Using an arbitrary 50:50 threshold, we ran analyses separately for 15 ‘small’ and 15 ‘large’ fragments (see Methods).

 

 

Small patch size

Large patch size

Metric

Trait syndrome

SES (mean + SE)

No. plots < expectation

V

P

SES (mean + SE)

No. plots < expectation

V

P

MNTD

Phylogeny

0.662 (0.938)

3

20

0.022*

–0.461 (0.663)

11

101

0.018

SDNDr

Locomotory

–0.562 (0.676)

14

119

<0.001

–0.122 (1.070)

10

92

0.037

Trophic

–0.240 (0.702)

11

108

0.002

–0.173 (0.950)

11

102

0.008

Dispersal

0.137 (1.007)

7

58

0.555

0.366 (0.834)

4

22

0.987

Overall

–0.785 (0.938)

13

115

<0.001

–0.241 (0.895)

9

92

0.037

Variance

Locomotory

0.112 (0.915)

9

77

0.359

–0.163 (0.794)

12

103

0.012

Trophic

0.270 (0.888)

8

69

0.639

–0.337 (0.814)

11

104

0.010

Dispersal

0.228 (0.924)

12

87

0.135

–0.331 (0.644)

14

119

<0.001

Overall

0.586 (0.610)

0

0

<0.001*

–0.041 (0.794)

5

26

0.055

 

Table C6. Community structure in assemblages of avian frugivores surviving in tropical forest fragments (n = 30). Asterisks (*) indicate observed values for community structure metrics (MNTD, SDNDr, variance) significantly different from the null model in the direction providing evidence for competitive interactions (variance is predicted to be greater than expected; SDNDr is predicted to be lower than expected). The standardized effect size (SES) is provided along with the standard error (SE). Statistics (V) and P-values are from Wilcoxon signed rank tests (two-tailed). Using an arbitrary 50:50 threshold, we ran analyses separately for 15 ‘small’ and 15 ‘large’ fragments (see Methods).

 

 

Small patch size

Large patch size

Metric

Trait syndrome

SES (mean + SE)

No. plots < expectation

V

P

SES (mean + SE)

No. plots < expectation

V

P

MNTD

Phylogeny

–0.464 (0.972)

9

99

0.026

0.363 (0.829)

4

31

0.107

SDNDr

Locomotory

–0.381 (0.585)

14

119

<0.001

–0.114 (0.696)

14

117

<0.001

Trophic

–0.556 (0.889)

14

110

0.001

0.302 (1.515)

11

93

0.032

Dispersal

–0.482 (0.848)

11

109

0.002

0.177 (1.104)

8

58

0.555

Overall

–0.608 (0.899)

14

119

<0.001

–0.063 (0.707)

14

111

0.001

Variance

Locomotory

0.137 (0.547)

5

45

0.421

–0.226 (0.735)

10

90

0.095

Trophic

1.222 (0.886)

0

0

<0.001*

0.293 (0.794)

0

0

<0.001*

Dispersal

0.126 (0.828)

7

45

0.421

–0.344 (1.138)

9

70

0.600

Overall

–0.073 (0.673)

3

18

0.015*

0.015 (0.876)

3

17

0.012*

 

Table C7. Community structure in assemblages of avian insectivores surviving in tropical forest fragments (n = 30). Asterisks (*) indicate observed values for community structure metrics (SDNDr, variance) significantly different from the null model in the direction providing evidence for competitive interactions (variance is predicted to be greater than expected; SDNDr is predicted to be lower than expected). The standardized effect size (SES) is provided along with the standard error (SE). Statistics (V) and P values are from Wilcoxon signed rank tests (two-tailed). Using an arbitrary 50:50 threshold, we ran analyses separately for the 15 ‘least isolated’ fragments (i.e., closest to forests >1000 ha) and the 15 ‘most isolated’ fragments (i.e., nearest to forests >1000 ha).

 

 

Least isolated patches

Most isolated patches

Metric

Trait syndrome

SES (mean + SE)

No. plots < expectation

V

P

SES (mean + SE)

No. plots < expectation

V

P

MNTD

Phylogeny

0.057 (0.302)

8

119

0.806

0.144 (0.203)

6

113

1

SDNDr

Locomotory

–0.213 (0.667)

14

118

<0.001

–0.282 (0.646)

14

119

<0.001

Trophic

–0.158 (1.582)

13

111

0.001

–0.095 (0.987)

12

98

0.015

Dispersal

0.175 (1.187)

8

58

0.555

–0.480 (0.730)

11

106

0.003

Overall

–0.329 (0.748)

15

120

<0.001

–0.342 (0.952)

13

114

<0.001

Variance

Locomotory

–0.033 (0.776)

7

68

0.679

–0.056 (0.556)

8

66

0.762

Trophic

0.821 (1.012)

0

0

<0.001*

0.694 (0.881)

0

0

<0.001*

Dispersal

–0.071 (1.241)

7

46

0.454

–0.147 (0.746)

9

69

0.639

Overall

0.190 (0.804)

2

7

<0.001*

–0.249 (0.690)

4

27

0.064

 

Table C8. Community structure in assemblages of avian frugivores surviving in tropical forest fragments (n = 30). Asterisks (*) indicate observed values for community structure metrics (SDNDr, variance) significantly different from the null model in the direction providing evidence for competitive interactions (variance is predicted to be greater than expected; SDNDr is predicted to be lower than expected). The standardized effect size (SES) is provided along with the standard error (SE). Statistics (V) and P values are from Wilcoxon signed rank tests (two-tailed). Using an arbitrary 50:50 threshold, we ran analyses separately for the 15 ‘least isolated’ fragments (i.e., closest to forests >1000 ha) and the 15 ‘most isolated’ fragments (i.e., nearest to forests >1000 ha).

 

 

Least isolated patches

Most isolated patches

Metric

Trait syndrome

SES (mean + SE)

No. plots < expectation

V

Wilcoxen P

SES (mean + SE)

No. plots < expectation

V

Wilcoxen P

 

Phylogeny

0.089 (0.257)

5

114

0.964

–0.190 (0.254)

8

138

0.304

SDNDr

Locomotory

0.023 (1.246)

9

92

0.037

–0.464 (0.416)

15

120

<0.001

Trophic

–0.020 (0.886)

10

96

0.021

–0.392 (0.734)

12

112

0.001

Dispersal

0.661 (0.720)

2

4

1

–0.158 (0.928)

9

80

0.138

Overall

–0.262 (0.977)

9

88

0.060

–0.765 (0.865)

13

117

<0.001

Variance

Locomotory

–0.170 (1.011)

11

96

0.041

0.119 (0.663)

10

84

0.188

Trophic

–0.077 (1.031)

10

86

0.151

0.009 (0.763)

9

87

0.135

Dispersal

0.175 (1.084)

11

82

0.229

–0.278 (0.390)

15

120

<0.001

Overall

0.160 (0.974)

4

22

0.030*

0.385 (0.488)

1

1

<0.001*

 

Table C9. Community structure in assemblages of avian insectivores surviving in tropical forest fragments (n = 30). Asterisks (*) indicate observed values for community structure metrics (SDNDr, variance) significantly different from the null model in the direction providing evidence for competitive interactions (variance is predicted to be greater than expected; SDNDr is predicted to be lower than expected). The standardized effect size (SES) is provided along with the standard error (SE). Statistics (V) and P values are from Wilcoxon signed rank tests (two-tailed). Using an arbitrary 50:50 threshold, we ran analyses separately for 15 ‘high quality’ and 15 ‘low quality’ fragments (see Methods).

 

 

High quality patches

Low quality patches

Metric

Trait syndrome

SES (mean + SE)

No. plots < expectation

V

P

SES (mean + SE)

No. plots < expectation

V

P

 

Phylogeny

–0.666 (0.244)

8

114

0.967

0.267 (0.264)

6

112

1

SDNDr

Locomotory

–0.238 (0.613)

14

119

<0.001

–0.258 (0.700)

14

117

<0.001

Trophic

–0.222 (1.017)

12

102

0.008

–0.032 (1.557)

13

108

0.002

Dispersal

–0.190 (1.116)

11

91

0.042

–0.115 (0.960)

8

69

0.319

Overall

–0.127 (0.929)

13

111

0.001

–0.545 (0.714)

15

120

<0.001

Variance

Locomotory

–0.174 (0.577)

10

81

0.252

0.086 (0.736)

5

51

0.639

Trophic

0.783 (0.859)

0

0

<0.001*

0.732 (1.034)

0

0

<0.001*

Dispersal

–0.230 (0.598)

10

77

0.359

0.012 (1.307)

6

42

0.330

Overall

–0.350 (0.681)

5

38

0.229

0.292 (0.735)

1

1

<0.001*

 

Table C10. Community structure in assemblages of avian frugivores surviving in tropical forest fragments (n = 30). Asterisks (*) indicate observed values for community structure metrics (SDNDr, variance) significantly different from the null model in the direction providing evidence for competitive interactions (variance is predicted to be greater than expected; SDNDr is predicted to be lower than expected). The standardized effect size (SES) is provided along with the standard error (SE). Statistics (V) and P values are from Wilcoxon signed rank tests (two-tailed). Using an arbitrary 50:50 threshold, we ran analyses separately for 15 ‘high quality’ and 15 ‘low quality’ fragments (see Methods).

 

 

High quality patches

Low quality patches

Metric

Trait syndrome

SES (mean + SE)

No. plots < expectation

V

P

SES (mean + SE)

No. plots < expectation

V

P

 

Phylogeny

0.060 (0.259)

5

120

0.775

–0.160 (0.254)

8

139

0.285

SDNDr

Locomotory

–0.431 (0.540)

14

119

<0.001

–0.010 (1.211)

10

94

0.028

Trophic

–0.240 (0.748)

11

108

0.002

–0.172 (0.914)

11

102

0.008

Dispersal

–0.112 (0.987)

9

72

0.262

0.615 (0.693)

2

13

0.998

Overall

–0.755 (0.959)

13

114

<0.001

–0.271 (0.891)

9

91

0.042

Variance

Locomotory

0.137 (0.576)

10

81

0.252

–0.188 (1.058)

11

96

0.041

Trophic

0.126 (0.625)

9

84

0.188

–0.193 (1.097)

10

84

0.188

Dispersal

–0.092 (0.820)

14

105

0.008

–0.011 (0.872)

12

96

0.041

Overall

0.450 (0.650)

2

17

<0.001*

0.095 (0.851)

3

17

0.012*

 

Table C11. Variance Inflation Factor (VIF) for each of the continuous predictor variables used in the GLMs. A VIF threshold of >4 (or >10) is used as a rule of thumb for excessive multicollinearity in regression analyses (O’Brien 2007).


Variable

Variance Inflation Factor (VIF)

Fragment size

2.033

Habitat quality (% of good forest)

1.515

Distance to forest

2.296

 

 

 

 

Table C12. Mantel test for GLMs assessing spatial correlation in functional trait syndromes across our study site.

Guild

Metric

Trait syndrome

Monte-Carlo test observed value

P value

Insectivore

SDNDr

Locomotory

0.072

0.245

Trophic

–0.083

0.796

Dispersal

–0.112

0.892

Overall

–0.068

0.696

Variance

Locomotory

0.074

0.222

Trophic

–0.076

0.771

Dispersal

–0.124

0.845

Overall

0.028

0.367

Frugivore

SDNDr

Locomotory

–0.029

0.635

Trophic

0.013

0.402

Dispersal

0.192

0.077

Overall

0.011

0.386

Variance

Locomotory

–0.026

0.577

Trophic

–0.011

0.524

Dispersal

0.103

0.162

Overall

0.013

0.542

Literature cited

O’Brien, R.M. 2007, A caution regarding rules of thumb for Variance Inflation Factors, Quality and Quantity 41:673–690.


[Back to E096-237]