A quantitative analysis of species sorting across organisms and ecosystems

The estimation of the relative roles of environment, space, and their interactions in structuring community composition is one of the central topics of community ecology. I conducted a quantitative review to determine if the degree of species sorting (SS) by the abiotic environment varied predictably between organism types and ecosystems. SS was quantified as the relative fraction of community variation that is explained by environmental variables. I integrated data from 326 variation partition analyses in a generalized linear model, and found that a mean of 26.1% (minimum 0%, maximum 88%) of the variance in community composition was explained by environmental variables. I also found that organism body size and dispersal group were not related to the degree of SS. SS varied among trophic positions, being highest in autotrophs and omnivores and lowest in herbivores and decomposers. SS also varied among ecosystem types: it was lowest in lakes and highest in estuaries and marine environments. Studies using abundance data showed a higher degree of SS than studies based on presence-absence data. SS was lower when data sets were analyzed by spatial filters instead of by polynomials. These results suggest that although significant among-group differences emerged for trophic position and ecosystem type, variation in SS across body sizes or across different dispersal groups was relatively unpredictable. Nonetheless, these findings shed light on how the degree of SS varies across ecosystems or organism groups and may give important insights into the magnitude of environmental effects on biotic communities in a changing environment.