Evaluating pesticide effects on freshwater invertebrate communities in alpine environment: a model ecosystem experiment

Pesticide loads in streams are potentially one of the most relevant stressors for macroinvertebrate communities. Nevertheless, real effects provoked at the community level are still largely unknown. Model ecosystems are frequently used as tools for the risk assessment of pesticides, especially for their regulation, however, they can be also applied to site-specific risk assessment in order to gain better understanding of the responses of aquatic ecosystems to chemical stress. In the present work, an experimental system was composed of 5 artificial streams that reproduced a mountain lotic environment under controlled conditions. This study was aimed to better understand, whether (and how) the biological community was influenced by pesticides pulse exposures. 5 mixture load events were simulated over the productive season (March–July 2010): biological community was regularly sampled and nominal concentrations of water were tested. The results were interpreted comparing the output of different metrics and statistical methodologies. The sensitivity of different metrics was analyzed considering single exposure events (maximum Toxic Units) as well as overall temporal trends. Results showed how some common taxonomic metrics (e.g. taxa richness, Shannon’s index, total abundance of organisms, and the Extended Biotic Index) were not suitable to identify the effects of pesticides at community level. On the contrary EPT%, SPEARpesticide and the Principal Response Curve methodology proved to be sensitive to this kind of stress, providing comparable results. Temporal trends of these metrics proved to be related to the concentration of chemicals. Remarkably, the first Principal Response Curve illustrates the trend followed by the most vulnerable species, while the second is more related to the trend of opportunistic species. A high potential risk for the invertebrate community was highlighted by a statistically significant decline of 40 points (comparison with the control) in both SPEARpesticide and EPT%.