Wednesday, June 6, 2007 - 11:30 AM
230

Development of biological indicators for New England lakes based on benthic macroinvertebrate assemblages

Yong Cao1, Charles P. Hawkins1, and Alan T. Herlihy2. (1) Western Center for Monitoring and Assessment of Freshwater Ecosystems, Department of Watershed Sciences, and Ecology Center, Utah State University, 5210 Old Main Hill, Logan, UT 84322-5210, (2) Department of Fisheries and Wildlife, Oregon State University, Corvallis, OR 97339

We used data collected from New England lakes (US EPA–EMAP) to explore the performance of two different types of biological indicators based on benthic macroinvertebrate assemblages: an O/E indicator and a multimetric index. The O/E indicator was derived from a Random Forest (RF) predictive model built from data collected at 24 reference lakes. This index was both precise (SD = 0.13) and responsive (86% of degraded sites rated as impaired). We also used RF to model the effects of natural environmental gradients on values of 44 candidate metrics, 9 of which varied with natural gradients (R2 = 0.10-0.37). Thirty metrics failed to differentiate stressed from reference lakes. Four metrics were rejected because of narrow ranges. Of the 10 remaining metrics, two unadjusted metrics (% of total individuals as oligochaetes and % of total individuals as Tanytarsini) and two adjusted metrics (Chironomidae taxa and Tanytarsini taxa) were selected based on their discriminatory power and independence. This index was less precise (CV = 0.21) than the O/E index and less responsive (77% of degraded sites rated as impaired). Our generally poor understanding of the tolerances of lake macroinvertebrates to stressors may currently constrain the utility of the multimetric approach in lakes.