Assessing the performance of volunteers in monitoring streams

1. Citizens are concerned about the quality of water resources and many participate in monitoring activities, though doubts remain about the quality of the data volunteers collect. We trained volunteers to collect benthic macroinvertebrates using professional protocols. Of the seven stream sites sampled by volunteer crews, six sites were also sampled by professional crews.
2. In the laboratory, volunteers used morphological features to identify as many different taxa as possible within the major insect orders; their identification was approximately to family. Volunteers calculated five metrics: total taxon richness, richness of three key groups (Ephemeroptera, Plecoptera and Trichoptera), and percentage dominance of the three most abundant taxa. All metrics were strongly correlated with (a) the percentage of urbanized area in the catchment and (b) the metrics derived from a more complete taxonomic identification by a professional scientist. Taxon richness metrics declined with urban development, while percent dominance increased.
3. An overall summary multimetric index was used to compare the field and laboratory procedures of volunteers and professionals. Using an ANOVA model, we detected no significant difference between field samples collected by volunteers and professionals. The variance of index values associated with differences between crews was zero. The ability of the index to detect significant differences among sites (statistical power) improved by only 13% for assessments based on professional laboratory identification instead of volunteer laboratory identification.
4. Citizen volunteers, when properly trained, can collect reliable data and make stream assessments that are comparable to those made by professionals. Data collected by volunteers can supplement information used by government agencies to manage and protect rivers and streams.