Prevalence of multimodal species abundance distributions is linked to spatial and taxonomic breadth

Aim. Species abundance distributions (SADs) are a synthetic measure of biodiversity and community structure. Although typically described by unimodal logseries or lognormal distributions, empirical SADs can also exhibit multiple modes. However, we do not know how prevalent multimodality is, nor do we have an understanding of the factors leading to this pattern. Here we quantify the prevalence of multimodality in SADs across a wide range of taxa, habitats and spatial extents.Location. Global.Methods. We used the second-order Akaike information criterion for small sample sizes (AICc) and likelihood ratio tests (LRTs) to test whether models with more than one mode accurately describe the empirical abundance frequency distributions of the underlying communities. We analysed 117 empirical datasets from intensely sampled communities, including taxa ranging from birds, plants, fish and invertebrates, from terrestrial, marine and freshwater habitats.Results. We find evidence for multimodality in 14.5% of the SADs when using AICc and LRT. This is a conservative estimate, as AICc alone estimates a prevalence of multimodality of 22%. We additionally show that the pattern is more common in data encompassing broader spatial scales and greater taxonomic breadth, suggesting that multimodality increases with ecological heterogeneity.Main conclusions. We suggest that higher levels of ecological heterogeneity, underpinned by larger spatial extent and higher taxonomic breadth, can yield multimodal SADs. Our analysis shows that multimodality occurs with a prevalence that warrants its systematic consideration when assessing SAD shape and emphasizes the need for macroecological theories to include multimodality in the range of SADs they predict