HP8 Field Biological Monitoring of Ecosystem Impairment
A105 & A106
1:20 PM - 5:20 PM, Thursday
() Derivation and application of stressor-specific tolerance values for diagnosing causes of biological impairment in streams.
Yuan, L1, Hawkins, C2, 1 US Environmental Protection Agency, Washington, DC, USA2 Utah State University, Logan, UT, USA
ABSTRACT- Biological assessments of stream assemblages have historically focused on identifying impaired water bodies by quantifying the degree to which their biota differ from an expected or reference condition. Several methods are available, and commonly used, for such assessments. However, methods for identifying the causes of impairment, which must be determined to guide remedial or restoration activities, are less well developed. The specific taxa collected from a stream should provide clues to the causes of biological impairment, because stream organisms differ in their sensitivities to different environmental stressors. We present an approach for identifying the likely causes of stream impairment that combines stressor-specific tolerance values with assessments of the degree to which taxonomic composition has been altered at a site. Empirical models can be used to predict the specific taxa that should occur at a site in the absence of stressors. The departure between the observed (O) and expected (E) lists of taxa for a site can be expressed as a ratio (O/Etaxa) that measures the degree of biological alteration, or impairment, at a site. Stressor-specific, diagnostic O/E indices can be derived by weighting the observed and predicted lists of taxa by their tolerances to different stressors (e.g., O/Esediment or O/Emetals). We use several data sets to show that weighted-averaging and generalized additive models can be used to derive stressor-specific tolerance values from field data and that the diagnostic O/E indices derived from these tolerance values respond in predictable ways to variation across sites in the type and magnitude of stress.
Key words: predictive models, tolerance value, diagnosis, field data