| Summary: | Although watershed land use effects on in-stream fish habitat and fish-macrohabitat associations have been widely studied in the past, low-gradient, coastal Louisiana streams have been poorly described in the literature. In this thesis, I report the results of a two-year study exploring relationships among regional land use, in-stream physical habitat, and headwater stream fish assemblages.
In chapter two, I examined in-stream habitat variables such as depth, flow, and substrate combined with three-pass electrofishing depletion estimates at thirteen 100-m stream sites. I used a combination of principle component analysis and structural equation modeling to determine if trends were present in the habitat, fish composition, and species trait data. I found that a species-based structural equation model was a better predictor of relationships among fish, land use, and in-stream habitat variables, when compared to species grouped by functional traits. In addition, it appears that the amount of agricultural land may not have as detrimental effects on fishes in these coastal streams as has been reported in other aquatic systems and substrate type may be the most important manageable habitat parameter.
In chapter three, I measured various in-stream habitat variables (i.e. dominant substrate, depth, flow) for two dominant macrohabitats (pools vs glides). I collected fish from each macrohabitat via point-electrofishing and compared these data using canonical correlation analysis (CCorA) to determine if trends were present in macrohabitat, fish composition, and species trait data. I found that species- and functional trait-based CCorAs were able to determine correlations with various macrohabitats and variables within macrohabitats. Results indicate that species traits may be better measures of assemblage structure when macrohabitat-scale management, conservation, or restoration is the goal.
Although I expected land use to heavily influence in-stream habitat and fish assemblage composition, my research indicated that in-stream habitat was be more influential than land use in determining species composition of Louisiana headwater stream fish assemblages. I also expected species- and trait-based models to successfully predict fish-macrohabitat associations. The findings of this study confirmed my predictions that multiple techniques for assessing fish-macrohabitat associations exist, although the most appropriate method may depend on specific management, restoration, or conservation goals.
|
|---|
