Sessile organisms interact for most of their life cycle only with their closest neighbors. The local biotic neighborhood should therefore determine the fate of individual plants, and the intraspecific variability in these local biotic neighborhoods should govern community assembly and dynamics.
We hypothesize that the stochastic geometry of biotic neighborhoods changes with species richness in fundamental ways: deterministic species interactions should lead in species poorer communities to a lower intraspecific variability compared to random locations in the forest whereas species rich communities should show high intraspecific variability in their biotic neighborhoods indistinguishable from that of random locations. To test this hypothesis we use data of large trees (dbh>10cm) in seven fully mapped forest plots with varying species richness (22‒340), ranging from temperate to tropical forests.
We define the biotic neighborhood B(r) of a focal tree as the mean pairwise dissimilarity of the focal tree to all other species within a given radius r (ranging from 1‒50m) and the intraspecific variability sd_rISARf(r) as the standard deviation of B(r) over all trees of given species. We repeated the analyses for different neighborhood radii and mean pairwise dissimilarities based on functional traits and on barcode molecular phylogenies.
Results/Conclusions
Null model results were very similar for functional and phylogenetic dissimilarity. The intraspecific phylogenetic biotic neighborhoods of two third of the 270 focal species tested were indistinguishable from that of random locations, and this proportion increased from 48% for temperate forests to 63% and 77% for subtropical and tropical forests, respectively. This increase supports the notion of stochastic dilution of neighborhood interactions that make species in more species rich (tropical) forests functionally more similar to each other.
Unexpectedly, departures from the null model were mostly positive for temperate and tropical forests (i.e., larger than expected intraspecific variability in the biotic neighborhood). However, subtropical forests showed the expected dominance of negative departures. Most of the significant effects occurred for small neighbourhoods (r < 12m), the occurrence of significant effects declined with increasing scale, and disappeared at 40-50m.
Our results show that large trees of a species are in general not surrounded by sets of phylogenetically or functionally consistent neighbors, but that they experience a wide range of biotic neighborhoods not much different from that of random locations in the forests. This result has important implications for coexistence theory as it outlines the strong stochasticity operating in relatively species rich forests at the individual level.