Figure 1. The interaction between warming and trait value, as a) the second axis of the PCA including all traits combined (describing plant size and water use strategy), as b) the second axis of a PCA with morphological traits only (describing plant size) and as c) the second axis of a PCA with ecophysiological traits only (describing water use strategy). Larger trait values correspond to larger size and/or more conservative water use strategy, as indicated by the arrow. Note that the axis in panel c) has been inverted to make the comparison between the panels easier. Lines are model predictions, points are raw data. The y-axes are changes in frequency (i.e. final/initial), in order to visualize the trends while taking into account the differences in initial frequency. The grey dashed line represents a relative frequency of 1, i.e. separating “losers” (below the line) from “winners” (above the line). For cover changes, see Supporting Information Figure S11a,c,d.
Figure 2. Species position in the trait space based on a) morphological traits only and b) ecophysiological traits only, c-d) indicating the traits governing the trait space. The first morphological axis describes leaf structure, and the second morphological axis describes plant size. The first ecophysiological axis describes plant gas exchange and the second ecophysiological axis describes integrated water use strategy. For both PCAs, the second axis correlates with species’ cover and frequency change to warming. In addition, the first morphological axis correlates with species cover change. Species are colored according to their position on the second axis, because for both PCAs, the second axis correlates with species’ cover and frequency change to warming; red colors indicate species less vulnerable to warming according to that trait axis. For the PCA including all traits combined, see Supporting Information Figure S10.
Figure 3. Community-weighted means for δ18O (correlating negatively with life-time stomatal conductance) and δ13C (correlating positively with life-time water use efficiency). Panels a) and c) show changes in community mean values when only changes in species frequency are taken into account (i.e. mean species trait values measured at +0°C). Panels b) and d) show changes in community-weighted values when both changes in species frequency and intra-specific trait changes are taken into account (i.e. site-specific trait values). See Supporting Information Figures S5-S6 for intra-specific changes in traits due to warming. The overall community-level effect of the transplantation was significantly higher life-time stomatal conductance (GS; lower δ18O; F(1,48) = 90.93, p < 0.001, panel b) and lower life-time water use efficiency (WUE; δ13C; F(1,48) = 8.75, p = 0.005m panel d) at the warmest site (1000 m, +5°C).
Figure 4. Relationship between species’ vulnerability to climate change based on morphological or ecophysiological traits only. Panel a) shows species ranked from most vulnerable to warming (1) to least vulnerable (16), according to their position on the second trait axis of the PCA with morphological or ecophysiological traits only (i.e. describing either size or water use strategy). The dashed line indicates 1:1 relationship, i.e. the further the species falls from the line, the more different are the two rankings for that species. Panel b) shows the correlation with the trait axes themselves.