Comparison of fish and macroinvertebrate bioassessments from South Carolina coastal plain streams

Abstract

Stream bioassessments are often based on a single taxonomic assemblage, such as fishes or macroinvertebrates, with the assumption that this assemblage is representative of other assemblages. However, ecological and physiological differences between taxonomic groups may cause different responses to disturbance and result in different assessment results. In this study, fish and macroinvertebrate bioassessments were conducted concurrently in South Carolina coastal plain streams and compared on the basis of precision, sensitivity, accuracy, and agreement. Fish and macroinvertebrate data were evaluated with previously developed multimetric indices including a modified Index of Biotic Integrity (IBI) based on electrofishing data and a benthic macroinvertebrate multimetric index (HDMI) based on data collected with Hester-Dendy artificial substrates. Benthic macroinvertebrates were also collected from natural substrates for comparative purposes. The IBI was more precise than the HDMI but the average difference between disturbed and reference sites was greater for the HDMI, resulting in equal sensitivity (i.e., ability to measure disturbance in relation to index variability). Regression of the HDMI on the IBI was significant (P{lt}0.0001) but moderate (R2 of 0.39). Agreement between indices was strong for highly disturbed sites but weak for slightly and moderately disturbed sites. Ordination of taxonomic data indicated that fish and macroinvertebrates responded differently to some disturbances regardless of whether macroinvertebates were collected from Hester-Dendy samplers or natural substrates. Disagreement between macroinvertebrate and fish assessments at moderately disturbed sites implies that biotic integrity cannot always be adequately evaluated from a single taxonomic group. Identification of disturbed sites was most accurate when HDMI and IBI results were combined. To improve the accuracy of stream bioassessments, future research should emphasize methods for cost-effectively sampling and integrating information from multiple taxonomic groups.