Includes bibliographical references. / Climate change continues to cause rising air and sea surface temperatures and changes in precipitation patterns across the globe. Many seabirds will be challenged by increasing temperatures because they must balance conflicting adaptations for dealing with cold environments when foraging and hot environments when nesting. Heat stressed seabirds often gular flutter for thermoregulation, a behaviour that is effective for dissipating heat but expensive in terms of evaporative water loss. This study examined gular fluttering behaviour of four species of southern African cormorants, crowned ( Microcarbo coronatus ), Cape ( Phalacrocorax capensis ), bank ( Phalacrocorax neglectus ), and white-breasted ( Phalacrocorax carbolucidus) cormorants. Results show that gular fluttering is influenced by temperature, body position and body size. Gular fluttering increases with temperature and larger cormorant species spend a greater proportion of time gular fluttering for a given temperature. Threshold temperatures for initiating gular fluttering are lower for large than for small cormorant species. Proportions of time spent gular fluttering are higher when birds are sitting than when crouching over the nest. Water loss shows the same pattern as gular fluttering, with the larger species estimated to lose a higher percentage of their daily water intake. Larger cormorant species can lose as much as 40% of their daily ingested water after eight hours of gular fluttering. These findings indicate that temperature increases from climate change will likely have serious direct impacts on nesting cormorant colonies in southern Africa. Gular fluttering could increase by as much as 25% by 2100 under current projected temperature increases, and increases in water loss could reach nearly 10%. Some species may shift their breeding dates to compensate for increasing temperatures, but if breeding activities are timed to coincide with peaks in their main prey specie s, this may result in poorer diets or increased competition from other species.

Seabird populations breeding within the Southern Ocean have experienced various threats over the past few decades. Albatrosses are particularly at risk due to several factors, inter alia, accidental bycatch on fishing gear, ingestion of polluted marine debris, invasive predatory species at breeding sites, as well as climate change-induced alterations to ocean circulation patterns. The recent decline in many albatross populations is mainly attributed to incidental fishing mortality, which decreases adult as well as juvenile survival rates and is thus detrimental to these long-lived, low fecundity birds. Recently, changes in fishing regulations to require the use of various mitigation measures have reduced the number of seabirds killed by fishing vessels. However fisheries may still impact seabirds either by direct competition for the same prey, or through ecosystem cascades arising from the removal of predatory fish and squid. Sub-Antarctic islands are important breeding grounds for many seabirds, including albatrosses. Monitoring of seabirds breeding on sub-Antarctic islands is important to detect changes in population dynamics to be able to implement timely conservation measures. Marion Island, the larger of the two Prince Edward Islands, some 2000 km southeast of South Africa, is a breeding site for four albatross species including the sooty (Phoebetria fusca) and light-mantled albatrosses (P. palpebrata). The Prince Edward and Crozet Islands are the only places where both Phoebetria albatrosses breed sympatrically in substantial numbers. Both archipelagos are 46°-47°S, at the southern and northern limit for sooty and light-mantled albatrosses, respectively. At-sea observations and diet studies suggest that sooty albatrosses forage mainly in sub-tropical waters to the north and light-mantled albatross in Antarctic or sub-Antarctic waters to the south. The sooty albatross is listed as Endangered due to recent global population declines whereas the light-mantled albatross is Near-threatened. The only comprehensive study of these species at the Prince Edward Islands was conducted during the late 1970s but annual estimates of breeding populations have been made from 1996 onwards. The previous analysis of these counts, up to 2008, suggested that the sooty albatross population on Marion Island decreased from 1996 to 2008, whereas numbers of light-mantled albatrosses increased over this period. Extending the count series to 2014, trends for both species were reversed, with sooty albatrosses recently increasing and light-mantled albatrosses decreasing. However, the timing of sooty albatross counts is in question as these were done towards the end of the incubation period when many nest failures have already occurred. Breeding success of both Phoebetria albatrosses was estimated during 2013/14 and 2014/15. The success of sooty albatrosses (51% overall) was higher than estimated at Marion Island in the 1970s (19%), but it was still lower than that of a neighbouring colony on Possession Island, Crozet archipelago (65%). The sooty albatross success was however skewed by a sub-colony with a very low 5 breeding success; excluding this sub-colony the breeding success is similar to that of the Crozets. Light-mantled albatross breeding success was the same as past estimates and lower than colonies at Macquarie and Possession Islands. Additional monitoring of a sub-sample of nests within the monitoring colonies was done to determine incubation and brood guard (light-mantled albatross only) shift lengths for both species. The shift lengths and distributions were not significantly different from previous data on Marion Island or other breeding sites.

Kelp forests are amongst the most productive, diverse and dynamic ecosystems on earth and kelp are ecosystem engineers, which provide a structurally-complex habitat for many marine plants and animals. The Mediterranean mussel (Mytilus galloprovincialis) is an aggressive invader that has been in South Africa since the late 1970s, and does not normally occur in the subtidal zone, however in 2016 this mussel was first observed growing on the heads and stipes of kelp plants (Ecklonia maxima) in False Bay, South Africa. The overall aim of this project was to gain a better understanding of the ecological impacts of the invasive mussel colonising the canopy of kelp forests in False Bay. It was found that M. galloprovincialis were more likely to occur on the outer edge of kelp beds, where there is more water movement, and mussel masses are much more frequently found on kelp heads than on stipes. In addition, mussels infecting the kelp beds were likely younger than three years, indicating that colonisation has been a recent event. While the buoyancy of kelp plants was reduced by the invasion of mussels, it was not enough to sink most kelp plants. The mussel masses created habitat that increased species richness on both infected heads and stipes, and six species that are alien to the Western Cape were identified on the infected kelp plants. As detached kelps can raft vast distances, there is thus potential to spread not only alien species, but also native South African species to locations where they could become invasive. Lastly, a new species of amphipod that burrows into the primary blades of kelp heads was discovered during the course of the study. Further investigation is required to determine if the infection rate of kelp forests in False Bay is increasing and whether this phenomenon occurs in other locations.

Includes bibliographical references. / By protecting ecosystems from exploitation, no-take zones are considered the principal means by which marine species and their populations can be conserved for future generations. To be successful, no-take zones require continuous monitoring of the fish community to evaluate the response of marine ecosystems to anthropogenic impacts and environmental change. Obtaining an understanding of the patterns of species composition, abundance, and distribution, allows monitoring efforts to be focused, efficient, and properly interpreted. Baited remote underwater video (BRUV) was used to examine the effects of site, depth, andlevel of protection, on the diversity and relative abundance of temperate reef fish within the Table Mountain National Park (TMNP) Marine Protected Area (MPA). Four notake zones and adjacent exploited areas, subject to conventional management restrictions, were sampled monthly over a four-month period. A total of 36 species from three marine classes and 18 families was recorded. Species diversity (Shannon-Wiener index) was found to increase with sites closest to the mouth of the bay, whilst species abundance was found to increase with depth. Results indicated no consistent response to protection status among the sites at either the community or individual species level. However, the oldest no-take zone proclaimed for the purposes of reef conservation was found to harbour higher species diversity and a higher relative abundance of fish compared to its respective exploited area. Furthermore, the similar frequencies in which hottentot (Pachymetopon blochii) and roman (Chrysoblephus laticeps) were observed across the four study sites, suggests that these two commercially-important species are successfully recruiting inside and outside the no-take zones. These results indicate that physical factors, rather than protection status, within False Bay influence patterns of fish assemblage composition, abundance, and distribution. In future, and to improve comparability, assessments within the TMNP MPA should be designed to target similar locations and depth ranges within the bay. The success of no-take zones must be evaluated according to their individual design and management goals.

Includes abstract. / Includes bibliographical references (p. 99-110). / Citizen science is increasingly recognised as a useful tool for conducting scientific research and public outreach, producing multiple positive benefits for biodiversity conservation and the volunteers involved in such research.

While many studies focus on identifying pollutants within an ecosystem or how they affect primary producers, few look at how pollutants move through trophic levels or their influence on animal demography. The aim of this study was to determine whether additional nitrogen, phosphorus and water, to a savanna would alter the vegetation quality enough to influence the feeding behaviour and diet of the ungulate populations both of which would alter the vegetation quantity. The study was conducted in the Kruger National Park. One site was supplied with additional nutrients and compared to three control sites. The nitrogen and phosphorus content of grass and tree leaves collected at the enriched site were higher than the leaves collected at the control site, indicating the additional nutrients are improving the vegetation quality. Feeding rates (determined from photos captured by camera traps) indicated a higher degree of herbivory at the enriched site. However, there was no difference in the δ¹³C value, nitrogen and phosphorus content in the ungulate dung collected amongst the study sites suggesting no change in the diet. The dominant grass was significantly shorter at the enriched site suggesting that increased grazing was diminishing grass biomass. Basic assessment of the trees indicated that the additional water at the enriched site seemed to be triggering an earlier start to the growing season for the trees. It was concluded that the additional nutrients have altered the vegetation structure enough to potentially influence animal demographics.

The priority of understanding marine systems increases as expanding pressures are exerted on them. In South Africa, efforts are underway to utilise the ocean to stimulate economic growth which will result in further pressures on marine systems. The aim is to mitigate pressures, largely through proposed offshore marine protected areas (MPAs). One of these proposed MPAs is situated off the Kei river mouth in the Eastern Cape of South Africa. The area offshore of the Kei river mouth is an endemism hotspot, and a transitional zone between the Subtropical East Coast and the Warm Temperate South Coast. The region waslargely unexplored because of itstreacherous sea conditions and because it was no longer an area of commercial fishing importance. A remotely operated vehicle (ROV) explored the regions’ fish and benthic habitats and investigated their associations. This study’s results provide quantitative information on fish and habitat diversity in the intermediate depths off the Kei river mouth. Observations of rare, commercially important, yet critically endangered charismatic reef fish species were made. These included red steenbras (Petrus Rupestris), seventy-four (Polysteganus undulosus), red stumpnose (Chrysoblephus gibbiceps) and dageraad (Chrysoblephus cristiceps). Habitat types, including rhodolith beds, sponges, and deep-water corals were documented. Maximum predicted fish diversity corresponded with mid-continental shelf, which is incorporated into the regions proposed MPA. The highest fish diversity was 10km within the shelf edge (which had a depth of approximately 100m) and was associated with the most structurally complex habitat biota: Fan Coral. Depth was a fundamental predictor associated with the presence and abundance of species distributions. The results support the location of the proposed MPA and are a step forward in identifying critical habitat to protect diversity and endangered species, and thus contribute to the regions spatial management and governance.

The Clanwilliam cedar, (Widdringtonia wallichii, formerly W. cedarbergensis) is a threatened conifer endemic to the fire-adapted fynbos vegetation of the Cederberg mountains, South Africa. Here its population size has drastically declined, and its conservation status subsequently escalated to critically endangered in 2013 by IUCN Red List of Plants. Studies have hypothesised that excessive exploitation for timber products, climate change and unfavourable fire regimes (frequency, intensity and season) have contributed to this species’ decline. This decline led to the overarching aim of the study to gain a better understanding of the effects of fire frequency and seasonality on the life history of Clanwilliam cedar. To characterise fire patterns in the Cederberg Wilderness Area, I used a latent class analysis on fire indices calculated from a fire history database. To explore the effects of fire seasonality on the cedar count numbers I used a negative binomial hurdle model using seasonal fire indices and environmental data. To examine the impact of fire frequency and seasonality on the life-history of the Clanwilliam cedar, I used a stochastic demographic model based on parameter values obtained from the literature. Findings from the latent class model indicated that the main axes of variation in fire frequency were the fire indices representing total fire frequency, summer fires, autumn fires in the last 30 years and fires in the last 30 years. Although these fire indices were able to distinguish relatively well between the three latent classes, it however was difficult to disentangle the relative importance of each fire index due to their strong covariation. This points to a more general pattern, suggesting that it is necessary to examine the entire fire frequency history and the seasonality pattern in order to understand the current state of the population of the Clanwilliam cedar. The linear count model revealed autumn fires as being positively associated, whereas mean annual precipitation and mean annual temperature and precipitation seasonality were negatively associated with the cedar numbers. The stochastic demographic model showed both summer and winter fires induce positive growth rates at firereturn intervals greater than 10 years, but winter fires always permitted a higher population growth rate. The sensitivity analysis of the stochastic population growth rate (log λs) to changes in the life-history parameters at fire-return intervals of 10 and 20 years showed that fire mortality was most important for a summer fire regime, and growth rates of adult trees were most important for a winter fire regime. The different methods used in this study provided different but complementary results, and thus insights from these various models could potentially contribute to the development of fire management strategies that reflect the complexities of fire frequency and seasonality on the population dynamics and thus persistence of the Clanwilliam cedar.

Includes abstract. / Includes bibliographical references. / Since the 15th century, human activity has altered and degraded nearly half of the available land of the Cape Peninsula, South Africa; this has resulted in significant restriction and fragmentation of the historic geographic range of the peninsula's Chacma baboon (Papio ursinus) population.

Recent studies suggest that a significant proportion of the plastic pollution that enters the sea is disappearing from the surface, despite being less dense than seawater. Observations across size classes show that it is particularly smaller plastics, <5 mm in length, that are susceptible to removal. The dynamics and transport mechanisms that determine the pathways taken by floating marine plastic debris are poorly understood and the processes causing this disappearance of plastic are unknown. A spatial gradient in the size composition of floating litter has also recently been observed by visual at-sea surveys in the South Atlantic Ocean, where smaller plastic litter items are found in greater relative abundance closer to the coast becoming less frequent with increasing distance out to sea. Conversely, larger, more buoyant plastic items were found to be proportionally more abundant at greater distances away from the coastal source. Both the observations of missing microplastic and the apparent spatial gradient evident in the size composition of dispersing litter suggest that size selective mechanisms are removing smaller fragments of plastics from the surface. The nature of these and the whereabouts and ultimate fate of these smaller plastic fragments is unknown. Two studies were conducted. The first was an empirical investigation to confirm how the size and buoyancy of litter items are influenced by dispersal distances from a point source. Beach littler samples were collected from beaches at increasing distances from a major pollution source: Cape Town in the Western Cape province of South Africa. The size and buoyancy compositions of litter at each distance interval were compared. Mean size and buoyancy increased significantly with increased distance from Cape Town. Mean item volume rose from 5.1 ml to 604 ml. Over 90% of the items recovered closer to Cape Town were in the two lowest buoyancy categories, in contrast to the furthest sampling site, where only 20% of the litter recovered occupied these categories and 55% occupied higher bin ranges. The findings from the beach litter samples were comparable to those of the recent at-sea surveys in the same region and confirmed the spatial gradient in size composition shown by that study. The second study was an experimental investigation into marine biofouling as a possible explanation for the size-selective sinking of smaller plastics at sea. . The study was conducted in situ at the False Bay Yacht Club in Simon's Town in the Western Cape. Samples of highdensity and low-density polyethylene plastic of varying thickness were cut into squares of three different sizes, tethered to exposure rails and submerged approximately 10 cm below the surface for a 12-week study period. A subset was removed bi-weekly, their buoyancy observed and change in dry weight measured. My results showed that both fragment size and material thickness were significant determinants of fouling mass accumulated. Exposure times required for sinking varied from 17 days for the smallest thinnest samples to 66 days for the largest, thickest ones. All sample sizes sank within the study period. Sample volume was a close correlate of time to sinking. Refined versions of these estimates could be scaled by factors such as environmental conditions and proximity to litter inputs, and potentially included into numerical models of floating litter abundance and distribution. Both studies showed that high volume debris items persist longer at the ocean surface.