Master of Science (MS), Ohio University, 2004, Environmental Studies (Arts and Sciences)

One aspect of physical habitat that has been shown to influence stream macroinvertebrate communities both directly and indirectly is substrate quality. Part one of this study examines differences in surface substrate size distribution data derived from the Wolman pebble count, the zig-zag count, and the QHEI substrate metric at twenty-five stream sites in southeastern Ohio. Wolman counts were conducted in riffles only, and zig-zag and QHEI assessments were performed in multiple habitats throughout the entire 150-meter sampling reaches. The zig-zag and Wolman methods provide significantly different values for substrate size, diversity, and % fines (<6 mm) variables. The Wolman count records higher percentages of coarse gravels, lower % fines and a smaller range of grain sizes than the zig-zag method. In cluster analyses, sites pair together differently, depending upon which method – the QHEI, Wolman or zig-zag - is used to characterize the substrates. Part two investigates associations between reach-scale stream macroinvertebrate communities and substrate data, as quantified by the aforementioned habitat evaluation techniques. We had three predictions: 1) a general trend of increasing taxa diversity and abundance over a sequence of increasing particle sizes, up to bedrock; 2) higher biotic diversity at sites with greater substrate diversity; and 3) fewer EPT taxa and fewer % scrapers at sites with high % fines. Correlation analyses (Spearman coefficient) show that some variables, such as % Trichoptera, % scrapers and Cambaridae abundance, follow the expected patterns with substrate size variables, while others, such as % shredders and Tipulidae abundance, are positively correlated with finer-sized substrates, such as clay, silt and sand. Only two variables are significantly correlated with substrate diversity variables. Cambaridae are more abundant at sites with more diverse arrays of substrates, and Simuliidae are less abundant. Two variables are significantly corre (open full item for complete abstract)

Committee: Kelly Johnson (Advisor) Subjects: