Thesis (Ph. D.)--University of Michigan, 1943. / Includes bibliographical references (leaves 184-198).

Thesis (M.S.)--University of North Carolina at Wilmington, 2005. / Includes bibliographical references (leaves: [70]-74)

Karenia brevis dinoflagellate blooms off the west coast of Florida can create devastating effects on marine communities when they release a neurotoxin known as a brevetoxin. These blooms, informally referred to as red tides, can cause massive fish kills, necessitate closures of shellfish fisheries, and can even leave lingering toxins that impact shelf communities long after the bloom has dissipated. As a result, much effort has been put into studying K. brevis bloom initiation and dynamics. However, how K. brevis blooms impact Florida's fisheries is not fully understood because the relationship between K. brevis cell counts and fish mortality is poorly described. To study this relationship and the ecosystem response to K. brevis blooms, Ecopath with Ecosim (EwE) modeling is used to force K. brevis bloom mortality on the shelf ecosystems by using a recently developed time series that indexes K. brevis bloom severity. This index dynamically drives K. brevis bloom mortality in EwE in a historical reconstruction scenario from 1980 to 2009. Three hypotheses on ecosystem response are explored using Gag grouper as a case study. We postulate a) that K. brevis blooms impose bottom-up and top-down effects on the food web, b) that episodic perturbations by these blooms shape the community structure and c) that fishing pressure exacerbates those effects. Results support the hypothesis that K. brevis blooms pose top-down food web pressures, which is seen by evidence of trophic cascading. Changes in community structure with bloom mortality are also evidenced by changes seen in biodiversity and richness. An exacerbation of those effects as a result of heavy fishing pressure is evident, however, is only seen during severe bloom events. Little to no changes were found in the mortality from K.brevis blooms during blooms of average severity, and less mortality was imposed on the system during blooms of particularly low severity. However, this may be an artefact of the mode of action of K. brevis in EwE. Investigation of bloom effects on Gag showed that natural mortality rates of Gag appear to be largely influenced by mortality incurred during K. brevis blooms relative to the low rate of predation on Gag. Moreover, consumption rates of Gag on its prey were found to increase under a realistic schedule of these blooms. This may be due to a combination of effects, including increased mortality on competitors (making more prey available for Gag) and a lowering of the mean age of the Gag stock, which increases population productivity.

A fish kill was observed in a historically freshwater region of the Everglades, coinciding with a spike in salinity. Sea level rise and altered hydrology allow freshwater wetland systems to be susceptible to saltwater intrusion. Most wetlands are characterized by highly organic soils with microbial communities starved of oxygen, seeking alternative electron acceptors. Seawater contains sulfate (SO42-), which is one of the alternative electron acceptors. Provided with sufficient quantities, sulfate-reducing bacteria (SRB) outcompete other anaerobes and produce toxic hydrogen sulfide. This study asked, what combination of soil organic matter (SOM) and SO42- are needed to produce sulfide concentrations potentially lethal to freshwater fish? Soil samples were collected from two freshwater wetlands with varying SOM content and incubated in microcosms at four SO42- concentrations (0, 0.75, 4.0, and 12.0 mM Na2SO4). Sulfide concentrations produced were compared to published data on lethal sulfide toxicity levels for aquatic life. This study demonstrated that high SOM soils (89.3 ± 0.2 % moisture) incubated with SO42- concentrations > 0.75 mM, rapidly produced toxic sulfide concentrations (> 5 ppm S2-) within 24 hrs.. While many freshwater species may tolerate a salinity of 1 ppt (0.75 mM SO42-), this SO42-concentration is sufficient to support SRB and lead to toxic sulfide production. ese results support the need to restore freshwater hydrology in wetland systems, such as the Everglades, thereby protecting against the rapid ecological effects of saltwater intrusion.

The Trinity River flows through the Dallas-Ft. Worth Metroplex in north central Texas where it receives effluents from numerous point sources including seven large regional wastewater treatment facilities. Historically, the Trinity River has been impacted by massive wastewater loadings which often constitute > 80% of the total river discharge during low flow periods. Normally, high mass loadings correspond to the summer months, compounding the effects of a naturally stressful period, characterized by high temperatures and low dissolved oxygen concentrations. Samples from 12 stations were collected quarterly over an 18 month period from the Trinity River and two tributaries. Water samples were analyzed for a variety of water quality variables, including metals, priority pollutants, pesticides, and general water quality parameters. Water samples were also tested for acute and subchronic effects with several test species. Fish were collected at each station and assemblages were characterized using traditional classification techniques and the Index of Biotic Integrity. In addition, sediment samples were assessed for toxic effects which could have adversely impacted fish recruitment and in situ biomonitoring experiments were performed. Quantitative habitat characterization analyses were performed to gain additional information that could possibly explains differences in fish assemblage structure related to habitat variability. Data were analyzed using regression, univariate, multivariate, and descriptive statistical techniques and new approaches for analyzing impact assessment data were discussed. Results indicated that the most substantial impacts on fish assemblages were confined to a segment of the river where a sequence of point sources, in close proximity to each other, were overloading the river's capacity to sufficiently dilute and/or detoxify the effluent. Data also indicated the presence of episodic toxicity from nonpoint sources. In addition, toxic effects in sediment samples and differences in habitat were detected and may have contributed to measured differences among fish assemblages in the Trinity River.