Inverse relationship between taxonomic resolution and explanatory power of metacommunity models

Background/Question/Methods

Invertebrate communities in freshwater streams are impacted by a variety of environmental factors that are highly variable among taxa, and form the basis of environmental monitoring protocols. However, because of several practical limitations, these studies rely on a coarse taxonomic resolution. It is possible that this will group together species with potentially very different environmental preferences, thus masking the relationship between taxonomic composition and environmental variables. The alternative hypothesis is that closely related species share similar traits; therefore, refining the taxonomic resolution will not impact, or will even decrease, the strength of relationships between community composition and environmental variables. In addition, it is possible that increasing taxonomic resolution could expose spatially localized distribution patterns due to dispersal limitation, potentially independent from environmental heterogeneity.

To test these competing hypotheses, we characterized the metacommunity patterns of Trichoptera and Coleoptera in 21 streams in Algonquin Provincial Park, Ontario, Canada, as the metacommunity framework explicitly recognizes these interactions between both environmental filtering and dispersal processes. We determined community composition at the family, genus, species, and DNA barcode cluster levels. Twenty-two explanatory variables were characterised at each site then reduced through PCA, and their relative roles in structuring the metacommunity was analysed using variation partitioning.

Results/Conclusions

We found evidence for the second hypothesis. As taxonomic resolution increased, the proportion of community composition variability explained by the environment decreased by 22% for the Trichoptera, but did not change for the Coleoptera. The variation explained by spatial variables was 2% on average for both groups at all taxonomic levels. These results suggest ecological interchangeability of closely related Trichoptera (or neutral taxa) in this system given the environmental variables we measured. In addition, the Trichoptera are much more species rich compared to the Coleoptera (47 vs. 13, respectively), and this suggests that increasing the number of neutral taxa in a metacommunity increases random variation in their abundances and thus the amount of unexplained variation.  Our study illustrates the importance of using a metacommunity context in environmental monitoring, the need to establish the most efficient taxonomic resolution before starting routine monitoring, and the potentially counterintuitive importance of regional species richness.