This thesis is structured in the form of five main chapters, each written in the form of an independent paper and supported by a general conclusion and one previously published paper on the same general theme. The central theme is the dynamics, ecology and evolution of intraspecific and intraguild predation - specifically the determination of an interacting mechanism between the guild fish species Engraulis capensis and Sardinops ocellatus and its significance for future fisheries management.

Includes bibliographical references (leaves 210-273). / In South Africa, Ulva cultivation is of paramount importance to the marine aquaculture industry. Three local Ulva species (Ulva lactuca Linnaeus, Ulva rigida C. Agardh and Ulva capensis Areschoug) were selected for this research. The first two are currently cultivated on abalone farms for abalone feed and for use as bio-filters, and Ulva capensis was included because it is morphologically and biogeographically distinct from Ulva rigida in nature, despite the inability of molecular methods to separate them. Ulva rigida was collected at I & J farm and from nature at Kommetjie on the southwest of the Cape Peninsula, and U. lactuca was exclusively collected from I & J farm because it could not be found at sites where it had previously been recorded. Ulva capensis was exclusively collected from Kommetjie as this morphological species has not been recorded on abalone farms. The research also aimed to compare U. capensis with U. rigida on a variety of different measures, as molecular studies have suggested that they may represent a single polymorphic species.

Two morphs of Patiriella exigua occur in South Africa. These morphs occur on opposite sides of a temperature divide present at Cape Point. The green morph occurring on the West Coast (cold temperate) and the mottled morph occurring on the South Coast (warm temperate). In a previous study by Katherine Dunbar these two morphs were shown to be the same species. This project tests if there is a physiological difference between the two morphs of P. exigua. This was achieved by observing the activity coefficient of righting response of the two morphs of P. exigua under different temperatures and dissolved oxygen concentrations. The green morph was collected from Green Point (West Coast) and the mottled morph from Kalk Bay (South Coast). The results of a general linear model indicated that temperature and dissolved oxygen were significant factors determining righting time of P. exigua with P-values of less than 0.05. AT-test indicated that there was a significant difference between the righting times of the two morphs of P. exigua with the green morph turning at a faster rate than the mottled morph when placed under similar conditions.

Algal turfs play an important role in both tropical and temperate reef ecology, supporting a diverse array of macrofauna as well as being primary producers within both intertidal and subtidal ecosystems. Biodiversity and biomass patterns of the intertidal algal turfs at Sodwana Bay, South Africa have not yet been established so this study was undertaken to investigate these. It was hypothesised that both biodiversity and biomass would decrease with increasing height along the shore, as well as that sediment would have an effect on the community composition of the turfs. The community was sampled over a vertical range of 1.2 m and found to be dominated by the red alga Jania intermedia (24.21 % of total algal coverage) followed by Tolypiocladia glomerulata (18.05 %), Dictyota sp. (13.39 %), Jania adhaerens (12.32 %) and Sphacelaria tribuloides. (5.83 %). No significant vertical zonation patterns were found, though certain species showed vertical trends such as Asparagopsis taxiformis and Sphacelaria tribuloides, which occurred in greatest biomass lower down on the shore, while Tolypiocladia glomerulata occurred higher up on the shore. Sand was found to affect certain species more than others, with Tolypiocladia glomerulata, Jania adhaerens and Asparagopsis taxiformis occurring in greatest biomass when the amount of sediment found in the quadrat was low, while Sphacelaria tribuloides. showed the opposite pattern and Jania intermedia showed no distinct pattern. Sampling took place over 1.2 m vertical height along the shore at spring low tide therefore covering 60 % of the intertidal zone vertically (maximum tidal range in South Africa is 2 m). It is therefore possible that the turfs are the cause of a decline in zonation by remaining damp and decreasing desiccation gradients vertically along the shore.

Surface plankton samples were collected weekly over a 14-month period from the York River at Gloucester Point, Virginia. All dinoflagellates collected were identified to the species level when possible, and the relative abundance of each form was noted. The local dinoflagellate flora was examined with attention being given to environmental factors which might influence its composition. Special emphasis was placed on "red water" blooms. One hundred eighteen separable forms, representing 20 genera of Pyrrhophyta, were collected. Of these, 84 were identified to the species level with reasonable certainty. Five of the 118 belonged to the class Desmophyceae, the remainder to the class Dinophyceae. In general, thecate Dinophyceae dominated the dinoflagellate flora, except during warmer months when blooms of certain unarmored species occurred. Of the 84 identified species, 43 had not been recorded previously from Chesapeake Bay. Fifty-four species were considered major components of the dinoflagellate flora. Species occurrence exhibited definite seasonality. Five major floras were recognized: ubiquitous or "year round"; "winter" or cold water; "summer" or warm water; and "spring" and "fall", periods of temperature transition. Most species were found within moderately restricted temperature and salinity ranges. These species occurred in the plankton only when their requirements for both factors were satisfied. Temperature alone appeared to be limiting for many "summer" species, including most red water forms. Salinity alone had little correlation with dinoflagellate occurrence. Eight species were dominant in "red water" blooms during the sampling period. Major bloom components were usually accompanied by lesser numbers of other species. Blooms in the York River occurred during summer months when water temperature and salinity were high, although Peridinium triquetrum produced "red water" in the nearby James River in April. Temperature appeared most important in initiating blooms, with windless weather and stable water important in formation and maintenence of discolored water. The seasonal distribution of the 54 major species is discussed, and illustrations of each are provided.

The limited knowledge of the deep-sea is a paramount concern, affecting our ability to assess the overall health of the ocean’s ecosystem. Technology has made deep-sea fishing more accessible, but management plans cannot be implemented on ecosystems lacking fundamental information of the biology and species within them. Comprehensive ecological studies are needed to identify factors that may influence the distribution and abundance of the faunal groups that are becoming commercially relevant. This observational study was conducted over a 3-year period to provide an assessment of physical, environmental, and biological factors that drive benthic and benthopelagic community structure and function in Exuma Sound, the Bahamas. The use of deep-sea traps allowed the identification of the ecological community to the highest degree of taxonomy. A series of 115 deep-sea traps were sampled from 360 to 1480 meters deep from spring 2014 to spring 2017. During this study, two new species were discovered Booralana maxeyorum and Booralana sp. nov. Crustaceans dominated the catch (98%) with Teleosts (1.2%) and Elasmobranchs (0.1%) contributing the rest. Baited Remote Underwater Video Survey can also provide additional information about the ecosystem dynamics. The video surveys can help gather data on food availability and foraging behavior in the deep-sea. Baited video sampling will help describe the present conditions and distribution patterns in benthic communities of the deep-water of the Exuma Sound. The temperature might be a determining factor in the species that are present at different depths. A Total of 14 deep-sea BRUV surveys were conducted from depths of 651 to 1397 meters from fall 2013 to spring 2017. The BRUV recruited 22 different species within sample ranges of 651 to 1397 m, and 19 different families. Dominant taxa were Booralana sp., Bathynomus giganteus, Synaphobranchus sp., Simenchelys parasitica, Centrophorus sp., and Squalus cubenis. The deep-sea of the Gulf of Mexico are especially susceptible to anthropogenic influences due to location, abundant petrochemical resources, and easy access to fisheries. The magnitude and long term effects of these insults on the deep-sea are unknown. A more recent event, the Deepwater Horizon oil spill occurred on April 20th 2010 when methane gas erupted 1500 m below the surface northern Gulf of Mexico causing the most massive oil spill in US history. The benthic communities with acute damage have not been assessed for damage to the community structure and population size. A Total of 422 deep-sea sets were collected from depths of 184 to 2002 meters between April 2011 and April 2017. The total catch from the traps was 2898 individual animals, 1832 invertebrates from 27 species and 12 families. No benthic invertebrates were found in April 2011 and 2012 sampling immediately after the oil spill. There was a gradual increase in the subsequent years, with highest values in 2015. Then the CPUE tapered off in 2016 and 2017. The sampling was dominated by three species, Bathynomus giganteus (BGIG), Chaceon quinquedens (CQUI), Raymanninus shcmitti (RSCH). There is a preference to the geographic area which could be caused by access to nutrients, depth and sediment type. The more extended timescale measurements of impacts from the oil spill were evident to the most abundant invertebrate species. / A Thesis submitted to the Department of Earth, Ocean and Atmospheric Science in partial fulfillment of the requirements for the degree of Master of Science. / Spring Semester 2019. / April 16, 2019. / Baited Remote Underwater survey, Deep-sea, Exuma Sound, Gulf of Mexico, Trapping / Includes bibliographical references. / R. Dean Grubbs, Professor Directing Thesis; Jeffrey P. Chanton, Committee Member; Markus Huettel, Committee Member; Edward J. Brooks, Committee Member.

Polycyclic Aromatic Hydrocarbons (PAH) buried in sandy beaches of the northeastern Gulf of Mexico after the Deepwater Horizon accident posed a potential risk to environmental and human health. Therefore, concentrations and decay rates of the EPA’s 16 priority PAHs were determined in the supratidal area of the sandy beach at Pensacola Beach, Florida from 30 June 2010 through 16 June 2011. The results revealed no concentrations exceeding NOAA guidelines and decay to background concentrations within a year. The relatively rapid decay of the buried PAHs was facilitated by tidal pumping that maintained aerobic conditions within the beach sediment. Submerged-oil-mats (SOMs) that settled onto the seafloor in the northeast Gulf of Mexico were buried in the surface layers of the permeable inner shelf sands, which raised the question whether this embedded oil is preserved under anoxic conditions in the shelf sands despite their high permeability, and whether the sand layer covering the buried oil can effectively prevent release of potentially harmful petroleum hydrocarbons from the sediment. A set of laboratory flume experiments demonstrated that advective pore water flows, generated when bottom currents interact with the ripple topography of the sand bed, transports oxygen to and releases PAHs from the embedded oil to the water column. This process allows rapid aerobic decomposition of oil buried in the surface layers of permeable shelf sediment but also enhances the release of potentially harmful substances from this oil. Large phytoplankton blooms were associated with the Deepwater Horizon oil spill, and significant amounts of oil particles and algal cells settled in the form of marine snow onto northeastern Gulf of Mexico inner shelf and slope sediments. This raised the question how the metabolism of these sediments responded to this fossil and modern organic matter input and whether the combined input would affect the decomposition of the settled oil particles. Measurements of dissolved inorganic carbon, oxygen and nitrogen fluxes across the sediment-water interface at 5 to 20 m water depth and incubation of sediments retrieved from 310 m and 1000 m depth revealed how the seafloor at the different water depths responded to the experimental deposition of realistic amounts of phytoplankton, weathered oil particles, and a mixture of phytoplankton and oil particles. The permeable sand sediments and warmer temperatures of the inner shelf in general produced stronger absolute responses by the benthic microbial community, but the relative increases in sediment metabolism were higher in the muddy slope sediments. The results of this thesis research underline the role of oxygen, temperature and sediment composition for the decomposition of crude oil compounds that were deposited on shore, shelf and slope environments in the Gulf of Mexico. The findings emphasize the role of the transport mechanisms that facilitate aerobic microbial breakdown of the petroleum hydrocarbons. This research provides information to coastal managers and decision makers that can help when designing response plans to future oil spills in the Gulf of Mexico. / A Dissertation submitted to the Department of Earth, Ocean and Atmospheric Science in partial fulfillment of the requirements for the degree of Doctor of Philosophy. / Spring Semester 2019. / March 29, 2019. / Includes bibliographical references. / Markus Huettel, Professor Directing Dissertation; Bill Cooper, University Representative; Jeffrey P. Chanton, Committee Member; Eric Chassignet, Committee Member; Ian MacDonald, Committee Member.

Marshes are particularly important ecosystems, providing long-term soil carbon storage, flood protection and nutrient filtration. Nutrient filtering, specifically nitrate removal, is largely the result of belowground microbially mediated denitrification. Previous studies reveal that denitrification rates differ in Spartina alterniflora and Juncus roemerianus patches but determining how the associated microbial communities contribute to these differences is challenged by the inherent physicochemical variability in the belowground of plants at different elevations in the marsh. Here we had a unique opportunity to evaluate denitrification rates and the belowground microbial community in J. roemerianus and S. alterniflora collocated at the same elevation, thus experiencing the same inundation cycles, in a saltwater marsh. To determine denitrification rates sediment slurry incubations (15N-nitrate) were used. The microbial community structure was determined using “iTag” sequencing of 16S rRNA gene amplicons. Slurry experiments revealed that denitrification rates were consistently higher in J. roemerianus compared to S. alterniflora. Analysis of 16S rRNA exact amplicon sequence variants (ASVs) showed that the microbial communities were similar in both plant types, although oscillations in the abundance of some ASVs was observed. To link the rate and microbial community data, Random Forest Modeling (RFM) was used to determine if specific microbes could be accurate predictors of higher or lower denitrification rates. RFM identified ASVs classified as Deltaproteobacteria; Desulfobacteraceae and Chloroflexi; Anaerolineaceae as the most important predictors of denitrification rates. These microbial predictors were also identified as core members of the rhizosphere of both plants. The Desulfobacteraceae core member, indicates higher denitrification rates, while the Anaerolineaceae core member points towards lower rates of nitrate removal. Desulfobacteraceae are known sulfate reducers, however some have been shown to utilize both nitrate and sulfate to grow chemolithoautotrophically by coupling sulfide oxidation to dissimilatory nitrate reduction. In fact, this pathway was identified in metagenomic and metatranscriptomic datasets from one of the samples analyzed herein. Collectively, our data revealed that J. roemerianus promoted greater belowground nitrate removal compared to S. alterniflora which may result from different plant characteristics that lead to oscillations in the abundance and activity of core members of the microbial community that can serve as predictors of denitrification rates. Further, the data suggested that this reaction may be mediated by previously unsuspected sulfate reducing bacteria in our saltmarsh ecosystem. / A Thesis submitted to the Department of Earth, Ocean and Atmospheric Science in partial fulfillment of the requirements for the degree of Master of Science. / Spring Semester 2019. / March 28, 2019. / Includes bibliographical references. / Olivia Mason, Professor Directing Thesis; Angela Knapp, Committee Member; Sven Kranz, Committee Member; Brian Chadwick, Committee Member.

Bibliography: leaves 190-216. / A simple population dynamics model is constructed to simulate temporal variability in the biomass of a dominant copepod Calanoides carinatus (Copepoda: Calanoida) along the West Coast region of South Africa. C. carinatus is extensively preyed upon by the commercially important anchovy Engraulis capensis and variability in zooplankton production may serve as an useful predictor of variability in anchovy recruitment levels. The model developed here circumvents the need to include a large number of parameters because it uses satellite-derived estimates of chlorophyll a concentration and sea surface temperature as primary inputs. Abundance estimates necessary to initialise the model are readily obtainable from biannual research cruises. The model successfully simulates observed features of a copepod population's response to pulses of upwelling and results obtained are consistent with data from field studies. The model is robust with respect to most of its parameters because minor changes in their values result in predictable changes in model output. The effect on model predictions of errors in field estimates is quantified. The model showed greatest sensitivity to parameters which are difficult to determine empirically, such as predator-induced mortality rates. Gaps in our present understanding of the nature and scale of processes affecting copepod egg abundance, survival and viability in the Southern Benguela system, were identified as the dominant impediment to attempts to simulate copepod population dynamics in the region. The Southern Benguela system is patchy on a range of different space and time scales. The effect of fine-scale distributional heterogeneity on mesoscale patterns of copepod productivity was investigated by assuming that spatial patchiness affected the degree of overlap between zooplankton and phytoplankton populations. The effect of spatial patchiness is particularly prevalent under poor feeding conditions, and may result in predictions based on average feeding conditions underestimating zooplankton production by as much as 30% in some circumstances. Estimates of zooplankton production are sensitive to both the spatial arrangement and intensity of food patches in a heterogeneous environment. There is a need to isolate the essential mechanisms causing distributional heterogeneity and to quantify the effect of spatial patchiness on model predictions to permit the correct averaging of model results over broad horizontal areas. Because of the model's sensitivity to the predator-induced mortality rate, a temporally and spatially integrated system is used to quantify this parameter as a function of varying patterns of predator and prey abundance. Shoals of anchovy recruits are explicitly modelled feeding on patches of C. carinatus prey, and the fish's performance is quantified through temporal and spatial integration of periods and patches of prey abundance and shortage. Constant high fish densities dampen the spatial variability in copepod abundance, whereas a pulsed predation pressure permits locally depleted copepod populations a short respite in which to recover some growth, thereby allowing the persistence of a few good prey patches which offer favourable energy returns for foraging fish. The model suggested that at high densities of anchovy recruits, predicted growth rates are strongly density-dependent and predation rates may exceed copepod production rates. Absolute measures of prey availability are sometimes unable to predict anchovy feeding success as mechanisms permitting temporal and spatial segregation play a vital role in synchronizing the relationship between fish predation pressure and prey turnover rates. The model emulates observed variability in anchovy growth rates and analysis of the output indicates that the availability of high sustained abundances of food along the West Coast may be a critical "bottleneck" contributing to the strength of recruitment to the pelagic puseseine fishery in South African waters. Observed chlorophyll a concentration and sea surface temperature data in 1971 and 1972 were used as inputs into an annual version of the basic model, and model-predicted patterns of copepod biomass were compared with observed patterns of zooplankton biomass in the two years. The ability of the model to simulate major differences in the general features observed in the two years supports its use as a tool to describe net patterns of zooplankton productivity over large horizontal areas. The model · identified the need to quantify the role of major size-class groups, such as the microzooplankton and macrozooplankton, in mediating the flow of energy from phytoplankton to fish.

Bibliography: pages 114-116. / This study is one of three based on grab samples of macrobenthos obtained before and at different times after mining for diamonds off the coast of Namibia. The first study dealt with multivariate clustering analysis of the first samples before and after mining. The second study focused on recovery times after mining and this study is aimed at estimating the amount of stress encountered by benthic communities, for comparision with the descriptive multivariate approach. Two research areas, classified as 'northern' and 'southern' were investigated. Data were aggregated and analysed at the genus level. Graphical and statistical analyses were conducted on the data which was classified in three ways. First, on all unmined sites from the two research areas together to test for natural site-to-site variability. Secondly and thirdly, each research area (north and south) was analysed separately to test for differences between unmined and mined sites at each area. Stress levels in the community were assessed by Caswell's neutral model (the Vstatistic) and by interpretation of the value of the W-statistic (a summary statistic of the ABC curves). Correlation techniques were applied to assess if there was any relationship between the diversity indices (as indicators of the influence of disturbance on community structure) on the one hand, and the environmental indicators of disturbance (percentage gravel, sand, mud) on the other.