Both physical–chemical and biological indicators are used to assess the ecological health of ecosystems. The functional trait composition of communities has the potential to predict the impact of environmental change on ecosystems, yet, we miss evidence that information on pollution can be used to forecast the functional responses of freshwater communities. We used chemical and ecological data collected in the Nestore river basin (Central Italy) to test whether (i) there are generalities in the response of macroinvertebrate functional diversity to river pollution; (ii) water pollution filters certain functional traits and (iii) traits do better than chemical analyses at revealing ecological stress in rivers. Due to interactions with hydrology, chemical variables showed higher seasonal variation at impacted sites, so occasional water samples represented snapshots. Functional traits did well at detecting water pollution by various chemicals, and there was a clear shift of trait combinations from unimpacted to impacted sites. Anthropogenic disturbances homogenized benthic communities in terms of taxonomic and functional composition, both in space (impacted sites) and time (seasons). Traits related to feeding habits (mainly reliance on fine particulate organic matter) and tolerance to organic pollution were prevalent at impacted reaches, and functional trait compositions were more stable than water chemistry over seasons because they integrate environmental conditions over longer time periods. Functional traits can play a relevant role in the bioassessment of polluted river ecosystems, even under low to moderate disturbance. Trait‐based assessment can be complemented by chemical analyses, notably at highly polluted sites where biological diversity drops to its lowest levels. Copyright © 2016 John Wiley & Sons, Ltd.

In applied river ecology it is implicitly assumed that community structure is a surrogate for functional properties. To test this hypothesis, 26 sites were sampled for macroinvertebrates in a Central Italian stream system impacted by 13 trout farms (1 upstream and 1 downstream site per facility). We used a fuzzy-coding approach of biological traits combined with species distribution analysis. Trout farm locally impacted the structure of macroinvertebrate communities, but serial disturbances at the stream system scale did not affect the overall regional diversity pattern.Impacted sites had species with large body size, which fed upon microorganisms and fine detritus, and which showed asexual reproduction. Least impact sites hosted small-sized invertebrates showing sexual reproduction, and feeding upon plant material and/or algae.

No abstract