VARIABILITY, ACCURACY, AND TAXONOMIC COSTS OF RAPID ASSESSMENT APPROACHES IN BENTHIC MACROINVERTEBRATE BIOMONITORING

Twenty population and community measures (with structural measures inc luding richness, diversity indices, and biotic indices, and functional measures including functional feeding-group components) that have bee n proposed for use in benthic biomonitoring were examined in terms of within-habitat variability (using coefficients of variation), accuracy (i.e., indications of impact when impact occurred; no indications of impact when impact did not occur), and taxonomic effort required (e.g. , from simple discrimination among taxa to generic identification of t axa and determination of trophic status). Based on 54 samples from a u niform riffle of a California mountain stream, coefficients of variati on ranged from 7% for Simpson's diversity index to greater than 100% f or several measures. Richness measures, Margalef's diversity index, th e Family Biotic Index, and the ratio of individuals of the scraper fun ctional-feeding group to total number of individuals were accurate in detecting acid and thermal stress (i.e., statistically significant dif ferences were found) but did not indicate impact when no impact occurr ed. Functional measures (e.g., ratios of functional groups) require gr eater taxonomic effort than structural measures. Separation of natural variability from perturbation-induced variability is essential in the impact assessment process.