Skip to main content
. 2016 Apr 3;6(10):3092–3106. doi: 10.1002/ece3.2100

Table 2.

Multimodel inference from spatial‐error simultaneous autoregressive (SAR) models of phylogenetic diversity (PD), Net Relatedness Index (NRI), and phylogenetic endemism (PE) of angiosperms and gymnosperms in all grid cells covering continental North America north of the US–Mexican border

Model parameters SAR‐averaged coefficients Importance: sum Akaike weight OLS R 2
Angiosperms PD 0.551
ACC 0.083*** 0.993
WBL 0.049*** 1.000
EHET 0.129*** 1.000
Angiosperms NRI 0.390
TMIN −0.498*** 0.997
EHET 0.118*** 0.999
Angiosperms PE 0.779
MAT 0.097*** 0.605
WBL 0.097*** 1.000
PRS 0.033*** 0.821
EHET 0.139 1.000
Gymnosperms PD 0.215
TMIN −0.214** 0.810
WBL 0.264*** 1.000
EHET 0.328*** 1.000
Gymnosperms NRI 0.271
ACC 0.108*** 0.560
WBL −0.062*** 0.680
EHET 0.059*** 0.633
Gymnosperms PE 0.387
ACC 0.164*** 0.958
WBL 0.275*** 0.995
PRS 0.127*** 1.000
EHET 0.261*** 1.000

Only predictors with Akaike weights exceeding 0.5 are shown. Importance values and model‐averaged standardized parameters were reported. To describe explained variance, additional adjusted R 2 values were given for the ordinary least square (OLS) multiple regressions with the same formula of the lowest AIC SARerr models.

Predictors include: MAT, mean annual temperature; TMIN, temperature of coldest month; WBL, water balance; PRS, precipitation seasonality; TES, temperature seasonality; VT, temperature change velocity since LGM; ACC, tree accessibility; EHET, elevation heterogeneity.

The significances of predictors in the lowest AIC models were given in the following scales: *** < 0.001 < ** < 0.01 < * < 0.05.