Skip to main content

Table 2 Linear mixed-effects regression (LMER) results showing best explanatory models of soil resistance (SR) in comparison with null models. Best models are shown in bold font, based on lowest Akaike (AIC) and/or Bayesian (BIC) Information Criterions. Null models include one or more grouping factors and a random factor only, single models include one main factor (explanatory variable) and random factor only, and best models include one or more grouping factor, the random factor, and one or more main factors for dataset group A (n = 252) derived from soils sampled over 3-week period in in subalpine meadows (n = 5) at Yosemite National Park, USA. An “X” denotes model factors included within a given model. Grouping factors included plant community type (PCT) and meadow gradient class (MGC), and the random factor is each of the five subalpine meadows (Mdw). Main factors included the following explanatory variables: bulk density (BD), gravimetric water content (GWC), and root mass areal density (RMAD)

From: Multi-scale drivers of soil resistance predict vulnerability of seasonally wet meadows to trampling by pack stock animals in the Sierra Nevada, USA

Models

Grouping factors

Random factor

Main factors

Model criterion

PCT

MGC

Mdw

BD (Mg/m3)

GWC (%)

RMAD (kg/m2)

AIC

BIC

Null model 1

X

 

X

   

227

243

Null model 2

 

X

X

   

244

254

Null model 3

X

X

X

   

225

243

Single model 1

  

X

X

  

211

221

Single model 2

  

X

 

X

 

210

220

Single model 3

  

X

  

X

246

256

Mixed model 1

X

X

X

X

 

X

209

232

Mixed model 2

X

X

X

X

  

211

232

Mixed model 3

X

X

X

 

X

X

210

233

Mixed model 4

X

X

X

 

X

 

210

230

Mixed model 5

X

X

X

  

X

224

244