AbstractSpecies co-occurrences in local communities can arise independent or dependent on species’ niches. However, the role of niche-dependent processes has not been thoroughly deciphered when generalized to biogeographical scales, probably due to combined shortcomings of data and methodology. Here, we explored the influence of environmental filtering and limiting similarity, as well as biogeographical processes that relate to the assembly of species’ communities and co-occurrences. We modelled jointly the occurrences and co-occurrences of 1016 tropical tree species with abundance data from inventories of 574 localities in eastern South America. We estimated species co-occurrences as raw and residual associations with models that excluded and included the environmental effects on the species’ co-occurrences, respectively. Raw associations indicate co-occurrence of species, whereas residual associations indicate co-occurrence of species after accounting for shared responses to environment. Generally, the influence of environmental filtering exceeded that of limiting similarity in shaping species’ co-occurrences. The number of raw associations was generally higher than that of the residual associations due to the shared responses of tree species to the environmental covariates. Contrary to what was expected from assuming limiting similarity, phylogenetic relatedness or functional similarity did not limit tree co-occurrences. The proportions of positive and negative residual associations varied greatly across the study area, and we found a significant tendency of some biogeographical regions having higher proportions of negative associations between them, suggesting that large-scale biogeographical processes limit the establishment of trees and consequently their co-occurrences.
What drives interspecies graft union success? Exploring the role of phylogenetic relatedness and stem anatomy
