Theses and Dissertations


Name: Black, Deirdre Darlene

Degree: MS

Chair: Elise R. Irwin

Resides: FAA Library

University: Auburn

Location: Auburn, Alabama

Date: 2005

Pages: 103

Keywords: acid mine drainage, biological, chemical, assessment, environmental impacts


Acid mine drainage (AMD) is responsible for serious environmental disturbance of aquatic systems in Alabama. A long-term data set (i.e., Clean Streams initiative project data; 1996-200I) that consisted of macroinvertebrate and water chemistry data from sites with varying levels of AMD impact was used to develop a rapid biological/chemical assessment technique (RBCAT) for characterization of AMD impacts in warmwater streams in Alabama. Data were collected monthly or quarterly from five sites associated with an AMD impacted stream and its receiving stream. Four AMD impact levels were identified associated with these sites (i.e., from high-impact AMD to low-impact pH neutral). Forty-six invertebrate metrics were calculated similar to the EPA's Rapid bioassessment techniques. Seventeen water chemistry variables were also calculated and analyzed. Data were analyzed for summer and winter. An analysis of variance was conducted to identify metrics and water chemistry variables that were statistically different among impact levels. Principal Components Analysis (PCA) was then conducted to reduce the number of metrics and/or water chemistry variables used to develop the RBCAT. Stepwise discriminant analysis (DA) was also performed to identify invertebrate metrics and water chemistry parameters that contributed to classification of sites by impact level.
Multiple invertebrate metrics differed among sites. Trichoptera and Megaloptera were collected from high-impact sites for both seasons; however, mean abundances were higher in winter than summer. Water chemistry variables also differed among sites but not seasons. Metrics values that were significantly lower at the high-impact site versus low-, moderate-, or recovery-impact sites included pH, hardness, sulfate, conductivity, unfiltered and filtered Ca, unfiltered and filtered Mg, and alkalinity. Filtered Mn, filtered Al, and unfiltered Al concentrations were significantly higher at the high-impact site than the other impact level sites. PCA was successful at differentiating impact level among sites used in the training data; however, PCA could not discriminate impact level at validation sites. DA was successful in differentiating impact level among sites in both the training data and validation data sets. Integrated approaches that include additional visual observations such water and sediment coloration and sediment chemistry as well as models with associated error rates are recommended.

Download: Download

203 Swingle Hall | Auburn, Alabama 36849 | (334) 844-4786 |
Website Feedback | Privacy | Copyright ©