Determining the extent of freshwater contamination by agrichemicals is a major challenge. Biological indicators have been proposed as indirect measures of contaminants that can be used to reduce chemical monitoring costs by identifying pollution hotspots that warrant thorough chemical testing. Many general indicators are based on taxonomic properties of aquatic macroinvertebrate communities. However, it has been suggested that metrics based on traits, rather than taxa, can be used to develop more chemical-specific and efficient indicators. Here, we investigate whether macroinvertebrate family-level traits can be used as simple indicators of elevated levels of specific pollutants in farmland drainage ditches, to reveal areas in need of further chemical monitoring. We selected seven traits—body size, body armouring, feeding guild, habit, oxygen acquisition, dispersal mode, and voltinism—that we predicted would influence sensitivity to nitrate, to the two herbicides atrazine and glyphosate, and to neonicotinoid insecticides, and tested whether any trait-chemical relationships were strong enough to be reliable bioindicators. We collected macroinvertebrate samples and water samples for agrichemical analyses from 27 farmland ditches in Eastern Ontario, Canada. We indexed the sensitivity of each sampled macroinvertebrate family to the concentration of each agrichemical, using coefficients from multiple logistic regressions of family absence/presence on the concentrations of the four agrichemicals. We reduced the seven traits predicted to influence sensitivity to four—body armouring, oxygen acquisition, dispersal mode, and voltinism—after examining their inter-dependencies. We then tested for cross-family relationships between sensitivity to each chemical and the trait categories for each macroinvertebrate family. Two traits, oxygen acquisition and dispersal mode, were significantly associated with two agrichemical sensitivity coefficients: nitrate sensitivity was associated with mode of oxygen acquisition, with atmospheric breathers having a higher mean sensitivity coefficient than dissolved-oxygen breathers, and glyphosate sensitivity was related to dispersal mode, with passive dispersers having a higher mean sensitivity coefficient than active dispersers. However, inspection of these relationships revealed that the responses lacked enough consistency across families to be reliable, chemical-specific indicators. Instead, a taxa-level, post-hoc analysis indicated that further work should be conducted to determine whether there are individual taxa whose presence at a site is a strong indicator of a lack or low levels of certain contaminants. In particular, the presence of Corixidae appears to indicate low ditch nitrate levels. Overall, however, our results combined with previous work suggest that we are unlikely to find chemical-specific macroinvertebrate indicators that are more efficient than a rigorous chemical sampling scheme.

Additional Metadata
Keywords Agrochemicals, Aquatic invertebrates, Eutrophication, Nutrient enrichment, Pesticides, Water quality
Persistent URL dx.doi.org/10.1016/j.ecolind.2019.105965
Journal Ecological Indicators
Citation
Collins, S.J. (Sara J.), & Fahrig, L. (2020). Are macroinvertebrate traits reliable indicators of specific agrichemicals?. Ecological Indicators, 111. doi:10.1016/j.ecolind.2019.105965