Hierarchical decomposition of trait patterns of macroinvertebrate communities in subarctic streams

Ecological research based on both species and their traits helps us understand the mechanisms structuring ecological communities. Our aim was to dismantle the effects of environmental variables measured at multiple spatial scales on the taxonomic and functional trait composition of benthic macroinvertebrate communities and to clarify the relationship between the environment and communities in high-latitude streams. Traits were combined into unique trait combinations (site-by-traits matrix), called hereafter the overall trait composition of macroinvertebrate communities, and then the matrix was decomposed into progressively smaller parts of individual traits (site-by-individual trait matrix). The effects of variables from different spatial scales on the variation in the overall trait matrix, decomposed trait matrices, and taxonomic data were analyzed using redundancy analysis and partial linear regression modeling. Our analyses indicated that: 1) the taxonomic composition of communities was more closely associated with factors measured at larger spatial scales, and the trait composition of communities was more closely associated with factors measured at smaller spatial scales, even within 1 drainage basin; 2) decomposing overall trait composition to its individual components of single traits revealed important patterns related to the potential causal factors; and 3) the abundances of organisms exhibiting different traits may be linked strongly to different environmental variables operating at different spatial scales. Our findings highlight the benefits of describing both the taxonomic and trait composition of communities when exploring the drivers of community composition. They also have direct applications in monitoring the vulnerability of high-latitude streams to future environmental changes.