Ecological Archives E085-050-A2

Daniel Borcard, Pierre Legendre, Carol Avois-Jacquet, and Hanna Tuomisto. 2004. Dissecting the spatial structure of ecological data at multpile scales. Ecology 85:1826–1832.

Appendix B. A figure and two tables showing results of the analysis of the Amazonian fern data:

Table B1:. R2 and probabilities related to the spatial analysis of the fern Adiantum tomentosum at Nauta (Peru);
Table B2: R2 and probabilities related to the spatial analysis of the fern Adiantum tomentosum at Huanta (Peru);
Fig. B1: Square-root-transformed Adiantum tomentosum abundances and spatial PCNM model and submodels (ordinate) at Huanta (Peru).


Table B1. R2 and probabilities related to the spatial analysis of the fern Adiantum tomentosum at Nauta (Peru). First line: R2 of each spatial submodel. Second line: R2 of the regression of the submodel (fitted values) on a subset of backward-selected environmental variables. Third line: product of the two previous lines, i.e., variation of the fern data explained by the environmental variables at the scale considered. Other lines: P values of the regression coefficients of the environmental variables in the model considered.


A. tomentosum

Very broad

Broad

Medium

Fine


R2 of PCNM submodel on A. tomentosum

 

0.333

0.239

0.126

0.117

R2 of environment on submodel

 

0.347

0.334

0.157
 

R2 of environment on A. tomentosum

0.436

0.116

0.080

0.020
 

Elevation (m)

<0.0001

<0.0001

<0.0001
   

Thickness of soil organic horizon (cm)

0.0004

<0.0001
     

Drainage

0.0393
   

0.0007
 

Canopy height (m)

0.0011

<0.0001
     

Canopy coverage (%)

     

0.0024
 

Shrub coverage (%)

0.0621

0.0341
 

0.0001
 

Herb coverage (%)

 

0.0002
     

Trees 3–7.5 cm DBH

0.0083
 

0.0009
   

Lianas 3–7.5 cm diameter

     

0.0413
 

Lianas 8–15 cm diameter

     

0.0183
 

 

Table B2. R2 and probabilities related to the spatial analysis of the fern Adiantum tomentosum at Huanta (Peru). First line: R2 of each spatial submodel. Second line: R2 of the regression of the submodel (fitted values) on a subset of backward-selected environmental variables. Third line: product of the two previous lines, i.e., variation of the fern data explained by the environmental variables at the scale considered. Other lines: P values of the regression coefficients of the environmental variables in the model considered.


A. tomentosum

Very broad

Broad

Medium

Fine


R2 of PCNM submodel on A. tomentosum

 

0.324

0.047

0.170

0.152

R2 of environment on submodel

 

0.408

0.099

0.072

0.039

R2 of environment on A. tomentosum

0.227

0.132

0.005

0.012

0.006

Elevation (trend)

 

0.0107
     

Elevation (detrended)

<0.0001

<0.0001

<0.0001
   

Thickness of soil organic horizon (cm)

 

0.0087
   

0.0420

Drainage

0.0047
   

0.0002

0.0188

Canopy height (m)

0.0034
   

0.0006
 

Trees 3–7.5 cm DBH

 

0.0007
     

Trees 31.5–62.5 cm DBH

 

0.0215
     

Trees > 63 cm DBH

 

0.0384
     

Lianas 8–15 cm diameter

 

0.0017

0.0010
   

 


 

 

   FIG. B1. Square-root-transformed Adiantum tomentosum abundances and spatial PCNM model and submodels (ordinate) along the Huanta transect (abscissa, 260 quadrats). (a) A. tomentosum (thick line) and parsimonious PCNM model. (b) very broad (thick line) and broadscale models. (c) medium-scale model. (d) fine-scale model. The R2 are calculated with respect to the total variation in the transformed fern count data; they are thus additive.


[Back to E085-050]