Are biological communities in naturally unproductive streams resistant to additional anthropogenic stressors?

Studies on the interactive responses to multiple simultaneously acting stressors have focused on individual or population-level responses in laboratory microcosms, while field-based studies on community-level responses are rare. We examined the influence of a natural (non-anthropogenic acidity) vs. human-induced stress (land drainage) and their interaction on species richness and spatial turnover (β diversity) of stream diatom, bryophyte, and benthic invertebrate communities. Our four stream categories were: circumneutral reference, circumneutral impacted, naturally acidic, and naturally acidic impacted streams. We expected the most sensitive species to be present only in the circumneutral reference streams. Therefore, species richness should be highest in these streams and lowest in the naturally acidic streams additionally stressed by forest drainage. Alternatively, communities in acidic streams may consist of the most tolerant taxa that are unaffected by further stressors, species richness in these streams remaining unaffected by drainage. We also expected spatial turnover to be highest in the circumneutral near-pristine streams and lowest in the drainage-impacted acidic streams. In all three taxonomic groups, α diversity was lower in the naturally acidic than in circumneutral streams. The additional impact of the anthropogenic stress on species richness varied between groups, having no effect on diatoms, antagonistic effect on bryophytes, and additive effect on invertebrates. We also found differences in how each stressor modified β diversity of each taxonomic group. For diatoms, β diversity showed an overall tendency to decrease with increasing stress level, while bryophyte β diversity responded mainly to forest drainage. Benthic invertebrate β diversity did not differ between treatments. Our results suggest that non-additive effects among stressors need special attention to improve the understanding and management of multifactor responses in streams. Our results also argue for the primacy of a multi-taxon approach to environmental impact detection, and for the inclusion of a wide array of ecological responses, particularly community turnover, in bioassessment programs to detect responses that may go unnoticed by conventional richness-based measures.