Table 2 Final set of covariates (DA-hydrologic index; ALT-altitude, BIOC.SO-summer ombrothermic index; BIOC.TH-thermicity index) and covariance structures for the best models retained for each response variable at the population (Ae, number of effective alleles; Fis, inbreeding; uHe, expected unbiased heterozygosity; Ho, observed heterozygosity) and at the individual level (HL, homozygosity level).

From: A spatial stream-network approach assists in managing the remnant genetic diversity of riparian forests

Dependent variable

Fixed effects parameters

Covariance parameters

Level

Y

Environmental Factor

Estimate

Std Error

p-value

Name

varcomp

range(m)

Population

Ae

DA

0.4768

0.1733

0.0105

Spherical.TU

0.4790

8287.01

  

ALT

0.0010

0.0008

0.2138

LinearSill.TD

0.2377

154617.24

  

R2 = 0.2337

   

Nugget

0.0496

 

Population

Fis

BIOC.TH

0.0002

0.0001

0.0341

Mariah.TU

0.1304

14505.31

  

DA

−0.0086

0.0056

0.1336

LinearSil.TD

0.5663

116814.96

      

Exponential.EUC

0.0016

1855642.4

  

R2 = 0.1956

   

Nugget

0.106

 

Population

uHe

DA

0.0152

0.0111

0.1810

Exponential.TU

0.2016

209.89

      

Exponential.TD

0.0001

88792.04

      

Gaussian.EUC

0.6454

17820.05

  

R2 = 0.0629

   

Nugget

0.0901

 

Population

Ho

ALT

0.0002

0.0001

0.0061

LinearSill.TD

0.7025

7257.30

  

DA

0.0280

0.0175

0.1211

Exponential.EUC

0.0000

3594.72

  

R2 = 0.2970

   

Nugget

0.0004

 

Individual

HL

BIOC.SO

−0.0023

0.0008

0.0039

Exponential.TU

0.7537

0.00

      

LinearSill.TD

0.1416

11.87

  

R2 = 0.0137

   

Nugget

0.0908

 
  1. The R2 values indicate the percentage of variation explained by environmental factors, while varcomp indicates the percentage of variation explained by each covariance structure within the final model mixture, and the nugget (i.e. the unexplained variation) that accounts for the variability that occurs at a scale finer than the closest measurements, as well as measurement error. The range represents the distance after which the spatial autocorrelation becomes zero.