Skip to main content
. 2019 Apr 23;9:6443. doi: 10.1038/s41598-019-42827-1

Table 2.

Estimated effects of the variables in the top-ranked mixed effects models on community phylogenetic diversity (standardised effect size of the mean pairwise distance – SESmpd, and standardised effect size of the mean nearest taxon distance – SESmntd) in woody and herbaceous communities.

Community type Phylogenetic diversity Top-ranked model Variables Coefficient F-value p-value
Woody SESmpd latitude + latitude2 + size + vegetation Intercept 0.2437 2.9568 0.0868
Latitude 37.6293 <0.0001
Linear term 1.0241***
Quadratic term 0.5636***
Size 0.2065 3.1324 0.0781
Vegetation 2.6106 0.0758
Open −3.3202*
Semi-open –0.3985
SESmntd latitude + latitude2 + size Intercept −2.7331 538.3517 <0.0001
Latitude 13.9199 <0.0001
Linear term 0.5869***
Quadratic term 0.0522
Size 0.1664 3.2111 0.0745
Herbaceous SESmpd latitude + latitude2 + size + realm + vegetation Intercept 13.1456 4.3958 0.0365
Latitude 5.2857 0.0053
Linear term 0.5858
Quadratic term −1.4841**
Size −0.0714 0.9973 0.3184
Realm 2.8835 0.0541
Nearctic −14.3492*
Neotropical −12.4259*
Paleactic −13.1274
Vegetation 21.4457 <0.0001
Open −1.3713***
SESmntd latitude + latitude2 + size + realm + vegetation Intercept −1.3252 0.1169 0.7325
Latitude 3.4684 0.0318
Linear term 5.9833
Quadratic term 11.0210
Size −0.0380 0.4748 0.4910
Realm 2.8352 0.0569
Nearctic 1.9078
Neotropical 0.8520
Paleactic 0.5757
Vegetation 41.9209 <0.0001
Open −1.3868***

Variables were tested with ‘anova.lme’ function from nlme package72, using the marginal significance of each fixed effect variable coefficient when all other fixed effects variables are present in the model. Size = sampling unit size; vegetation = vegetation type (closed, open, semi-open); realm = biogeographic realm (Afrotropical, Australasian, Indo-Malayan, Nearctic, Neotropical, Palearctic)29. Significance of individual coefficients are flagged with stars: *p < 0.05, **p < 0.01, ***p < 0.001. Estimates for biogeographic realms refer to the deviation from Afrotropical, and for vegetation type to the deviation from closed. All parameter estimates of the top-ranked models are shown in Supplementary Tables S1 and S2. Study identification was included as a random effect variable in all models.