RESULTS
Predictors alone generally explained a very low proportion of the
variability in the LMMs (marginal R2 from 1% to 4%),
and only for BDMC of clonal species, the fixed effects accounted for
58% (Table 2). Most of the model variability was instead explained by
the random effect associated with species identity (conditional
R2 up to 93%; Table 2; Supplemental Material 2). As
of NMDS, clonal species (except for Koeleria distinguished by
consistently taller individuals and Carex by higher BDMC) showed
a greater overlap in the multifunctional space than non-clonal species,
which were characterized by distinct and largely species-specific
occupancy of the multifunctional space (Figure 2).
TABLE 2 Results of the LMMs for the clonal and non-clonal
species separately. Sign (arrow direction), significance (p-value: ** ≤
0.01; * ≤ 0.05; . ≤ 0.1) and strength (R2 values) of
single-trait predictor links are reported (marg = marginal
R2; cond = conditional R2). Only the
significant relationships are indicated – refer to Supplemental
Material 2 for detailed model summary statistics.