An Evaluation of Taxonomic Data Quality in Fish Community Assessments

Accurate taxonomic identification is an important but often overlooked component of biological assessments. The evaluation of taxonomic data quality is gaining acceptance in laboratory-identified biological indicators, such as algae and benthic macroinvertebrates; however, field-based taxonomy, often associated with studies of fishes, presents unique challenges to quantifying taxonomic variability and thus has impeded its progress. To address this, voucher specimens (1 to 5 individuals per species) were retained from a random selection of 10% of sampling sites from fish community surveys in Maryland and Georgia.  Re-identification of the samples was completed by an independent quality control (QC) taxonomist, with results directly compared to those of the primary taxonomist and summarized as percent taxonomic disagreement (PTD).  Mean PTD of the QC samples was 7.8% (range: 0% - 17.6%) and was directly related to overall species richness (R²=0.38). Species comprising the families Petromyzontidae and Cyprinidae accounted for 71% of disagreements, followed by Centrarchidae (28%).  A detailed evaluation of the results for individual species and the identification of potential causes of disagreements prompted corrective actions, which decreased PTD to 6.8%. This procedure promotes the identification of difficult specimens, serves to increase the skills of the taxonomists, and helps identify recording or data entry errors.  Proper training and experience are critical elements for improving taxonomic consistency.  We recommend the adoption of standard procedures, taxonomic data quality standards, and routine and rigorous quality control evaluations to minimize the effects of taxonomic errors.