Although mesopelagic fishes are an important component of marine food webs, the adaptive features used to facilitate niche partitioning among co-existing and presumably competing mesopelagic species is unclear. This study examined the trophic ecology of the two principal mesopelagic fishes off the west coast of South Africa (lanternfish Lampanyctodes hectoris and lightfish Maurolicus walvisensis) sampled during the spring 2014 and autumn 2015 cruises, using stable isotope and stomach content analyses. Stable isotope values were extracted from the white muscle tissue of fishes, but due to the high lipid content of both species, samples were processed in duplicate: δ13C was measured from lipid-extracted samples and δ15N from non-extracted samples. To validate the stable isotope results, stomach contents were examined and the relative importance of prey items was assessed using three measures: frequency occurrence (%F), numerical abundance (%N), and dietary carbon (%C). Both mesopelagic species occupied different isotopic niches that were separated by their δ15N values across a similar δ13C range. Furthermore, the relationship found between trophic position and standard length emphasizes the structuring effect of size within the assemblage, with the larger species (L. hectoris) occupying a higher trophic position than the smaller species (M. walvisensis). Although copepods dominated the diet of L. hectoris in terms of numerical abundance (42%), macro-zooplankton was by far the most important dietary component, with euphausiids contributing 53% of dietary carbon. Conversely, copepods - particularly Calanus sp. - were the most important component of the diet for M. walvisensis in terms of their occurrence (84%), numerical abundance (64%), and dietary carbon (67%). Though some dietary overlap exists between L. hectoris and M. walvisensis, the results of this study suggest resource partitioning within the mesopelagic assemblage, likely facilitated by differences in alimentary morphology (i.e. trophodynamically mediated), and possibly by differences in their respective foraging strategies. Similarly, ontogenetic shifts in trophic position were detected, which suggests that these adaptive features may also be used to mitigate intra-specific competition within populations. Furthermore, the trophic positions of both L. hectoris and M. walvisensis inferred from dietary and isotopic data signify that mesopelagic fishes (in the context of this study) are secondary and tertiary consumers in the marine ecosystem of the southern Benguela. Nevertheless, samples covering a larger area of the southern Benguela and multiple years would be needed for a more complete understanding of the trophic ecology of these two species.

Remotely sensed weekly MODIS data of chlorophyll α (Chl-α) concentration, sea surface temperature (SST) and satellite altimetry data of Absolute Dynamic Topography (ADT) and geostrophic velocities are used to examine the seasonal and interannual patterns in the Chl-α concentration and SST over three pilot study areas of the southwest Indian Ocean, namely Northeast Madagascar, Southeast Madagascar and Mascarene Islands. The weekly and monthly climatology and the weekly means of each variables are assessed using image displays and time series from 2003 to 2014. It is found that there is a seasonal cycle of phytoplankton blooms occurring twice a year across northeast and southeast Madagascar. The two blooms occur during the summer monsoon and during the winter monsoon, respectively. Unlike these two areas, the Mascarene Islands area has only one bloom during the summer monsoon. There is a negative correlation between SST and Chl-α concentration across all three areas; when SST is high, Chl-α concentration is low and vice versa. Also, the current patterns showed that the two Madagascar study areas, are more physically dynamic than the Mascarene Islands region. Unlike the Masacarene region, the Madagascar regions are more affected by the forcing of the South Equatorial Current that splits into the Southeast Madagascar Current and Northeast Madagascar Current, thus causing displacement of surface water. New outcomes of this study are that the north Indian Ocean (north of 100S) is not the only area that is affected by the summer and winter monsoons but the areas south of 100S may also be indirectly affected by the monsoons. Across Northeast Madgascar region, the summer monsoon bloom is well spread over the area while the winter monsoon bloom is mostly coastal. Across the Southest Madagascar region, the summer monsoon bloom spreads from east to west while, the winter monsoon bloom spreads from west to east. The Mascarene region is less productive with higher sea surface height and weaker eddies compared to the other areas and the mixed layer depth is greater across the Mascarene region, thus less nutrients are injected to the euphotic zone and the weaker eddies result in less mixing and consequently in weak Chl-α production. This study should improve our understanding of the seasonal and interannual variability of the SST and Chl-α and the dynamics of the ADT and geostrophic velocities in these regions for improved management of fishery resources using an ecosystem approach to fisheries.

Includes bibliographical references (leaves 56-67). / Knowledge of the vertical distribution of phytoplankton in the upper ocean is essential for accurate estimates of primary production. Satellite remote sensing has given scientists an unprecedented view of near-surface chlorophyll distribution and other surface conditions, including sea surface temperature and wind data, from regional to global scales but little information on the dynamics below the surface. As a result estimates of global production tend to use regional profile averages but these methods oversimplify the smaller scale dynamics, particularly in coastal regions where productivity is highly variable on time scales of weeks. A pilot study by computer science honours students in 2006 showed the viability of using a Dynamic Bayesian Network (DBN) in predicting a representative profile per pixel of a satellite map based on a database of time series satellite surface data. In this study, 5813 in situ profiles were obtained from the highly dynamic upwelling region around the southwestern coastline of southern Africa. The samples were collected between 1988 and 2006 between the coast and the continental slope. The region was divided into three sub-regions according to biophysical processes: the west Coast; the west Agulhas Bank; and the east Agulhas Bank. Of the 5813 profiles, 5557 were included in the sub-regions. Two consecutive processes were then applied to the profile database. First, the profiles were clustered using a k-means clustering program which produced 16 representative clusters.

Includes bibliographical references (leaves 48-54). / Ancient intertidal stonewall fish traps are found world-wide and those along the SouthAfrican south coast are the focus of my thesis. These fish traps, known locally as 'vywers', have recently enjoyed much media attention as interest increases in both South Africa's cultural heritage and its diminishing fish stocks. Two pioneering studies, by Goodwin (1946) and Avery (1975), provided the only documented knowledge of these vywers. My study aimed to locate, survey and document the main concentrations of vywers within a 300-km stretch along the south coast. A total of 43 sets of vywers was located by aerial surveys, 30 of which are only accessible through private land. This affords them some measure of protection, together with five located off reserves, but diminishes their educational and tourism value. Four sets of vywers were mapped in detail using aerial photogrammetry, a method that provided a rapid, extensive, accurate survey record in the form of geo-rectified ortho-images of these sites. The vywers are built in both exposed and sheltered environments, constructed from in situ rock material built into walls with either angular or curved shapes. These walls may occur singularly or in complexes of up to 25 traps. Vywers are prone to decimation by wave action and storm damage and so require maintenance to retain their characteristic form and associated cultural information. Tensions have, however, arisen between those who maintain and fish the vywers, and fisheries managers. Data from a questionnaire survey compared with records from the literature showed that species composition has not changed significantly in the last five decades. The fish are caught most frequently during new-moon spring-tides, especially in the winter months. The vywer fishery, currently active at only two sets of vywers, targets primarily mullet species (mostly Liza richardsonii) but infrequently enjoys 'bonanza' catches of over-exploited linefish stocks such as galjoen Dichistius capensis. It is these latter catches that concern managers, in addition to the noncompliance of fishers in terms of catch composition and size, permits and gear.

Bibliography: leaves 54-57. / The economy of Eritrea depends on rain-fed agriculture, which is highly sensitive to seasonal and inter-annual variations of rainfall. Adequate amount of rainfall is required every year for adequate agricultural production and drinking water. Therefore, the distribution of seasonal rainfall in space and time is critical to the country's economy. Drought has great influence on the socio-economic aspect and livelihood of the rural population. Drought occurs in Eritrea due to anomalous regional atmospheric circulation. This study focuses on the summer rainy season of June, July and August (JJA). It examines atmospheric and oceanographic mechanism patterns associated to summer rainfall anomalies. The investigation mainly concentrates on wet and dry spells that occurred over Eritrea during the rainy seasons (June to August) of 1950's-1990's. Rainfall data was available from the Climatic Research Unit (CRU) Hulme dataset for a period of 1950 -1996 and this together with rainfall data from the Africa Data Dissemination Service was used to determine the characteristics of wet and dry spells. Finally, NCEP-NCAR reanalysis climate data was used to analyse the atmospheric circulation associated with wet and dry years. It has been found that the pattern of the SST along the eastern Pacific and the Indian oceans seem to influence the occurrence of the anomaly rainfall years. Correlation analysis shows that Eritrean summer rainfall is more strongly correlated with the eastern Pacific and Indian Oceans SST. Thus, the drought records in Eritrea are mainly associated with EI Nino events.

The South African midwater trawl fishery targets adult horse mackerel (Trachurus capensis) on the Agulhas Bank. The bulk of the catch is taken by a single midwater trawler. The bycatch of this fishery is investigated by analysing observer records from 2004 to 2014 to report on bycatch composition, volumes and temporal and spatial patterns. Small fauna were analysed by extrapolation of sample weights and large fauna were counted. The midwater trawl fishery was estimated to have caught 25 415 t annually, with a bycatch rate of 6.9% of the total catch, by weight. Bycatch species with the highest average annual catches were chub mackerel (Scomber japonicus), redeye roundherring (Etrumeus whiteheadi), ribbonfish (Lepidopus caudatus) and hake (Merluccius spp.). Large fauna bycatch species included Cape fur seals (Arctocephalus pusillus), sharks (15 species), sunfish (Mola mola) and cetaceans (two species). Temporal analysis revealed significant differences between day-time and night-time CPUE for nine species, and significant differences in seasonal CPUE for 17 species. Spatial analysis of small fauna species composition and of large fauna species composition identified three groups in each. The 100% observer coverage is high compared to other South African fisheries, and, although the average sampling rate of 1.56% is low, an increase in the sampling rate might not be feasible owing to the size of catches. The bycatch rate in this fishery is low compared to other South African trawl fisheries but similar to global midwater trawl bycatch rates. There are species overlaps with various fisheries, namely the demersal trawl, small pelagic, linefishery, shark longline and squid fisheries, yet the total catch estimates from this fishery are generally small relative to catches taken in the target fisheries.

Includes bibliographical references (leaves 94-104). / Rainfall over Tanzania is highly variable. In recent decades the country has been devastated by floods and droughts. The Tanzanian population relies heavily on seasonal rainfall. Over the northern part of the country, the rainy season occurs in two phases, the short rains (October-December) and the long rains (March - May). Over the southern and the western areas, the rainy season occurs during November - May. This study aims at identifying factors that may playa role in the rainfall variability over the southern region of Tanzania. This study indicates that over the Indian Ocean, the sea surface temperatures (SSTs) over the northeast of Madagascar have a strong relationship with the rainfall in southern Tanzania during the OND rainfall, while the SSTs over the southeast Atlantic have a strong relationship during the JFM rainfall. It has also been revealed that the El Niño/ Southern Oscillation (ENSO) have influence on rainfall variability over the region. The warm/cold ENSO events would impact both a wet or dry year. Such contradiction has imposed a challenge to the forecasters of seasonal rainfall over southern Tanzania.

Includes abstract. / Includes bibliographical references (leaves 55-61). / An Age Structured Model (ASM) was develop in which dynamic and density-dependent cannibalism was included as a top-down control on a modeled population of M. paradoxus which used spawner biomass and maternal based recruitment. The ASM was used to evaluate the effects cannibalism had on age structure, recruitment and spawner biomass of the modeled population. The development of the model was described and evaluated with special emphasis on incorporating cannibalism and maternal based recruitment.

This study assessed a man-made salt-water pond (saltwater pond 1 SP1) as a potential site for the development of an oyster hatchery in Velddrif, South Africa. Over the study period it was observed that the site was not suitable for oyster larvae culture due to the following factors: temperature spikes during the summer months, low dissolved oxygen levels, high salinity levels, and the proliferation of the filamentous green algae (FGA), mainly the sp. Rhizoclonium implexum. Widespread patches of R. implexum were observed within SP1 and increased in biomass over the study period. Biomass patterns were not measured within this study, however microcosm experiments directed at nutrient depletion rates caused by FGA proliferation assessed the effect of the FGA on the system. Within microcosm experiments with and without FGA, nitrite within the system was significantly lower in the FGA inclusion treatment. Pond nutrient dynamics within the system indicated that widespread nutrient depletion occurred between the incoming water and the rest of the pond, and it was clear that the inflow station had significantly higher nutrient concentrations than all the other stations within SP1. Phytoplankton concentrations were extremely low and could be attributed to the FGA dominated state within SP1. Taken together with the fact the water levels within SP1 were not stable, the data suggested that SP1 was not an optimal source of seawater for either algal or larval oyster culture, and an alternative oyster hatchery site should be assessed.

Bibliography: leaves 73-78. / In this study Quality Index Method (QIM) schemes for raw headed and gutted (H&G) fish and skinless fillets as well as cooked fillets of kingklip (Genypterus capensis) were developed in a shelf life study. The QIM schemes for raw H&G and skinless fillets are based on six and seven quality attributes respectively, with a scoring system from 1 to 6, where a score of six is best quality (perfect). For the case of cooked fillets the scheme is based on six attributes, with a scoring system from I to 5, where a score of five is best quality. Sensory analysis of cooked fillets was carried out parallel to sensory analysis of raw H&G and skinless fillets as well as microbial count study until 18 days of storage. These were used to decide the maximum storage time in ice and to observe how the different quality attributes of cooked fish, especially odour and flavour, deteriorated with storage time in ice, as they are the best symptoms of fish spoilage. However, the microbial count study alone was continued until 22 days in ice in order to follow the bacterial spoilage pattern with storage time.