Ecological Archives E094-059-A4

Peter C. le Roux, Jonathan Lenoir, Loïc Pellissier, Mary S. Wisz, Miska Luoto. 2013. Horizontal, but not vertical, biotic interactions affect fine-scale plant distribution patterns in a low-energy system. Ecology 94:671–682. http://dx.doi.org/10.1890/12-1482.1

Appendix D. Detailed Wilcoxon’s matched pairs test results testing for the improvement of species distribution models with the addition of herbivore and/or plant variables.

Table D1. Predictive performance (mean ± SE) for each of the four different models, as analyzed using generalized linear (GLM), generalized additive (GAM) and boosted regression tree (BRT) modeling. The significance of the improvement in predictive performance with the addition of plant or herbivore variables is indicated for each combination of parameters.

 

Method

Metrics

Baseline
(= M1)

Plant model
(= M2)

Sig. of improvement relative to M1

Herbivore model
(= M3)

Sig. of improvement relative to M1

Full model
(= M4)

Sig. of improvement relative to M2

Sig. of improvement relative to M3

Species occurrence models

 

 

 

 

 

 

 

 

GAM

AUC

0.66 ± 0.02

0.71 ± 0.02

0.010

0.66 ± 0.02

0.631

0.70 ± 0.02

0.922

0.060

TSS

0.24 ± 0.04

0.33 ± 0.04

0.023

0.24 ± 0.04

0.669

0.31 ± 0.04

0.842

0.099

GLM

AUC

0.67 ± 0.02

0.71 ± 0.02

0.022

0.66 ± 0.02

0.842

0.70 ± 0.02

0.981

0.026

TSS

0.23 ± 0.04

0.31 ± 0.04

0.036

0.22 ± 0.03

0.874

0.31 ± 0.04

0.854

0.010

BRT

AUC

0.66 ± 0.02

0.70 ± 0.02

0.002

0.66 ± 0.02

0.026

0.69 ± 0.02

0.474

0.001

TSS

0.23 ± 0.03

0.26 ± 0.03

0.070

0.23 ± 0.03

0.240

0.27 ± 0.03

0.240

0.011

Species cover models

GAM

Spearman corr.

0.15 ± 0.04

0.25 ± 0.03

< 0.001

0.15 ± 0.03

0.94

0.23 ± 0.03

1.00

0.010

GLM

Spearman corr.

0.15 ± 0.04

0.24 ± 0.03

< 0.001

0.15 ± 0.04

0.97

0.22 ± 0.03

1.00

< 0.001

 

BRT

Spearman corr.

0.17 ± 0.03

0.24 ± 0.03

< 0.001

0.18 ± 0.03

0.70

0.23 ± 0.03

0.83

< 0.001


[Back to E094-059]