Extended Data Fig. 6: Principal component analysis of tree species functional traits across the six study regions.

Ordination diagram of the first two axes of the principal component analysis (PCA) of tree species trait values, for (a) all species (N = 1,207 species and 2,520 species-region combinations) in six neotropical forest regions (colored polygons), and (b) only species occurring in at least five plots in a given region (N = 484 species and 765 species-region combinations) which were included in our niche analysis, with colors representing their fate with forest loss: winners (green points) are those species whose distribution (that is abundance-weighted niche optimum) along forest loss gradients significantly deviate positively from random expectation, while losers (yellow points) are those whose abundance significantly decreases in communities embedded in more deforested landscapes, and neutral species (gray points) do not show significant response to forest loss. In (b), points representing winners and losers increase in size with species total abundance across plots, to highlight trait differences between dominant winner-loser species. Arrows indicate the direction and weighting of vectors representing the four continuous traits considered: WD = wood density, lnSM = logarithmic of seed mass, LMA = leaf mass per area, Hmax = maximum height.