Resilience and regime shifts: Do novel communities impede ecological recovery in a historically metal‐contaminated stream?

Abstract

Abstract Novel communities that result from exposure to contaminants and other anthropogenic stressors often persist in ecosystems that have experienced regime shifts. Because these systems may not return to pre‐disturbance conditions after removal of a stressor, understanding the ecological consequences of regime shifts has important implications for how restoration success is defined. Long‐term observational studies can potentially identify regime shifts in disturbed ecosystems; however, experimental approaches may be necessary to demonstrate their ecological consequences. We report results of a long‐term (28 years) observational study and a series of stream mesocosm experiments that investigated a regime shift in a mining‐contaminated watershed. We tested the hypothesis that establishment of a novel, metal‐tolerant macroinvertebrate community in a previously contaminated stream impeded recovery of sensitive species, despite significant improvements in water quality over the past two decades. We observed significant improvements in abundance and species richness, but differences in community composition and trophic structure persisted downstream from a former source of metal contamination. These downstream communities were dominated by large, metal‐tolerant caddisflies that likely impeded colonization by metal‐sensitive groups (e.g. grazing mayflies). Mesocosm experiments conducted with reference and downstream communities demonstrated that novel communities retained their tolerance to metals, but were significantly more sensitive to other stressors. We suggest that the failure of downstream communities to recover represented a contaminant‐induced regime shift that resulted from the sustained dominance of metal‐tolerant species. Synthesis and applications. Restoration ecologists generally consider increased species diversity or abundance of recreationally important species as indicators of restoration success. However, few stream restoration projects have quantified ecological resilience, and none have included a significant experimental component. Our research is the first to experimentally demonstrate that novel communities in a stream recovering from historical metal pollution have lower resilience to other (e.g. non‐metal) stressors. These results have important implications for how restoration success is defined. In systems where the return to pre‐disturbance conditions is unlikely, a better understanding of the ecological resilience of novel communities may be critical for assessing restoration success. We suggest that resilience to novel stressors is an important indicator of restoration effectiveness that may be applicable in other aquatic ecosystems.