With-in site variation of macroinvertebrate metrics and biocriteria from streams in Sublette County, Wyoming

B.D. Marshall^1,2.¹Center For Aquatic Studies, EcoAnalysts, Inc., Bozeman MT 59715, ²Montana State University, Department of Ecology, Bozeman MT 59717

I examined the within-site variability of biocriteria using data collected from streams in Sublette County, Wyoming. Samples were collected by the county’s conservation district using the probabilistic-composited-eight-Surber method commonly used for bioassessment in western states and RIVPACS. The method divides riffles into grids, and randomly selects eight locations for sampling. Samples were formed by combining all eight subsamples in the field. We used five replicates from each site—40 Surber samples (in all) from each site. Wyoming biocriteria rated one site as “Good,” “Fair,” or “Poor” depending on the samples used. Other sites had confidence intervals spanning two “condition-categories.” Power Analysis indicated that seven replicates were required to detect 10% change in condition-score, but the larger streams required more replication than smaller streams. Power analysis also indicated that three replicates could only detect a change of 10-20 taxa; 30-50% of the taxa must disappear, without being replaced, before a statistically significant change in richness is detected—even with a liberal alpha of 0.10. Composited samples do not necessarily homogenize the data. Moreover, they complicate site-specific monitoring by increasing field effort and obliterating correlations among biotic and abiotic variables. Many streams cannot support the collection if 40 Surbers.