Shared environmental responses drive co-occurrence patterns in river bird communities

Abstract

Positive or negative patterns of co-occurrence might imply an influence of biotic interactions on community structure. However, species may co-occur simply because of shared environmental responses. Here, we apply two complementary modelling methodologies – a probabilistic model of significant pairwise associations and a hierarchical multivariate probit regression model – to 1) attribute co-occurrence patterns in 100 river bird communities to either shared environmental responses or to other ecological mechanisms such as interaction with heterospecifics, and 2) examine the strength of evidence for four alternative models of community structure. Species co-occurred more often than would be expected by random community assembly and the species composition of bird communities was highly structured. Co-occurrence patterns were primarily explained by shared environmental responses; species’ responses to the environmental variables were highly divergent, with both strong positive and negative environmental correlations occurring. We found limited evidence for behaviour-driven assemblage patterns in bird communities at a large spatial scale, although statistically significant positive associations amongst some species suggested the operation of facilitative mechanisms such as heterospecific attraction. This lends support to an environmental filtering model of community assembly as being the principle mechanism shaping river bird community structure. Consequently, species interactions may be reduced to an ancillary role in some avifaunal communities, meaning if shared environmental responses are not quantified studies of co-occurrence may overestimate the role of species interactions in shaping community structure.