Rapid assessment of rivers using macroinvertebrates: the role of experience, and comparisons with quantitative methods

We assessed rapid biological assessment (RBA) of macroinvertebrate communities in comparison with quantitative sampling at 18 pairs of river sites in south-eastern Australia. One member of each pair served as a reference site and the other was affected by mild to moderate human disturbance from a point or diffuse source (fish farm effluents, small municipal sewage discharges, a dam, agriculture and grazing). Samples were taken from riffles, stream edges and rocks in pools, mostly using hand nets (RBA) and Hess samplers (quantitative). Macroinvertebrates in RBA samples were always subsampled by live-picking by eye on site for 30 min. Comparisons were made between novice operators (university undergraduates) and experienced river biologists in the application of RBA, including sample identification. Quantitative samples were collected only by experienced river biologists, preserved and picked under stereomicro scopes in the laboratory, either in their entirety or after mechanical subsampling. Specimens were identified to family level for both methods. Novices recorded slightly fewer families than experts and misidentified some specimens, and expert data for the same site and habitat were on average more consistent than novice data. Nevertheless, Procrustes analysis of ordinations showed that differences between RBA and quantitative data did not depend on the type of RBA operator. And regardless of the operator, RBA data were better than quantitative data at discriminating between reference and disturbed sites for all three habitat types. We conclude that this type of RBA is sensitive and cost-effective, and we recommend improvements to inter-operator consistency.