Document Type

Thesis - University Access Only

Award Date


Degree Name

Master of Science (MS)

Department / School



This research was conducted to (1) determine an optimal suite of macroinvertebrate metrics for northern prairie streams, (2) examine standard EPA biological assessment methods and to test the effectiveness of those methods in identifying areas of degradation, and (3) determine which metrics are strongly correlated to stream physical habitat and chemical parameters. Invertebrate samples were taken from 18 sites on Bachelor and Brookfield Creeks in Moody County, South Dakota once per month from April through September 1998 and 1999. Data were applied to 51 candidate invertebrate metrics which were (1) randomly selected in sets often, (2) grouped as per USEPA (Plafkin et al., 1989) (3) optimized by minimizing reference site variability and maximizing site discriminatory power (Barbour et al., 1999), and (4) optimized using principal components analysis (PCA). A Spearman rank correlation analysis was performed to examine relationships among candidate invertebrate metrics, channel habitat, and chemical parameters. USEPA and randomly selected metric sets resulted in 60% of stream sites classified as slightly impaired and 40% of sites as moderately impaired. Optimized metrics resulted in 20% of stream sites classified as unimpaired, 47% of sites as slightly impaired, and 33% of sites as moderately impaired. PCA metrics resulted in 47% of sites classified as unimpaired and 53% of sites as slightly impaired. Three sites categorized as slightly impaired using the Plafkin set were considered non-impaired using the optimized set. All sites categorized as slightly impaired using the Plafkin set were considered nonimpaired using the PCA set. Random and USEPA (Plafkin et al., 1989) core sets provided similar site scorings with sites classified as either slightly or moderately impaired. In contrast, objectively chosen core metric sets (optimized and PCA sets) indicate that some sites were unimpaired. These metric set comparisons indicate that objective selection of core metrics is necessary to prevent type I errors from biomonitoring investigations. Measurements of specific conductance, ammonia-N, total dissolved phosphorus, percent pebble, clay, silt, and sand substrate, stream current velocity, and pool:riffle ratio displayed the strongest correlations with invertebrate metrics. Percent clinger mode of existence (a habitat utilization measure), abundance of Rheotanytarsus spp., percent Elmidae, percent gathering collectors (a functional group measure), and percent of invertebrates preferring depositional habitat were strongly correlated to physiochemical parameters. The optimization procedure is a well-designed, unbiased approach for selecting macroinvertebrate metrics to be included in an index of biological integrity. However, further refinement of the optimization procedure may be necessary to ensure the elimination of redundancy in the index. Also, the assumption of equal weights of metrics may not be optimal for differentiating macroinvertebrate communities.

Library of Congress Subject Headings

Water quality biological assessment -- South Dakota
Aquatic invertebrates -- South Dakota
Stream ecology



Number of Pages



South Dakota State University