Okhombe is a ward in the Northern Drakensberg where community based restoration of degraded lands has been conducted for over a decade. In this important water supply region, payment for ecosystem services has been suggested as a means through which to conserve biodiversity, improve veld condition, provide income to local communities, and ensure water security. However, before such alternative market mechanisms can be considered, the effects of community based restoration must be properly quantified. The primary aim of this study was to determine some of the ecological impacts of community based restoration in these communal grasslands.

Eutrophication, one of the leading global water pollution challenges, can be addressed with nature-based solutions (NBSs). NBSs use or mimic natural processes (ecosystem services) to improve water quality, for example microorganisms, plants and filter feeders have already been shown to mitigate eutrophication. However, burrowing endobenthic organisms have not yet been considered as a NBS, despite their potential to greatly impact water quality. Therefore, the purpose of this study was to determine whether these endobenthic organisms could be effective as NBSs, more specifically whether the activities of the common sandprawn (Callichirus kraussi) can improve water quality in terms of pelagic nutrient, microalgal and suspended solids concentrations. C. kraussi is a benthic ecosystem engineer that, like other endobenthic crustaceans, affects water flow, sediment biogeochemistry and benthic-pelagic fluxes through bioturbation and bioirrigation. Their elaborate burrow systems may act as biological filters for pelagic microalgae, but they may also be a source of nutrient-rich excrement inputs into the water column. To determine the effects of C. kraussi on water quality, a laboratory mesocosm experiment and comparative in situ analysis were used in this study. In the mesocosm experiment, water quality at 3 increasing C. kraussi densities (0 [control], 100 and 200 individuals·m-2 ) were compared. At the end of the experiment, the water column NH4 + concentration in the high density C. kraussi treatment was 2.6 times greater compared to the controls. Despite this increase in the NH4 + concentration, the water column chl-a concentrations were 42.8% (50% density treatment) and 44.3% (100% density treatment) lower compared to the controls. Additionally, microalgal biomasses in the sediment were less concentrated on burrow mounds created by C. kraussi compared to the adjacent sediment between mounds and within the burrows. However, the overall chl-a concentrations in the sediment did not differ between treatments. This suggests that C. kraussi has the effect of breaking up microalgal mats on the sediment surface with their burrow mounds without changing the overall sediment surface chl-a concentration, thereby maintaining benthic-pelagic coupling. C. kraussi did not affect suspended solids concentrations in the water column over the 12-day study period. For the in situ component, water quality was compared at sites where C. kraussi was present and the adjacent sections where it was absent in the Zandvlei Estuary Nature Reserve. The findings from the mesocosm experiment were reflected in situ: Chl-a concentrations in C. kraussi habitats were reduced by 43.1% from upstream areas without C. kraussi, and again increased by 32.9% from C. kraussi habitats to downstream canalised areas without C. kraussi. Differences in suspended solids concentrations were also not explained by C. kraussi presence in situ. The conclusion from this study is that ecosystem engineering by C. kraussi has an overall positive effect on water quality in urban estuaries, principally by reducing chl-a concentrations in the water column and maintaining benthic-pelagic coupling. Therefore, C. kraussi should be conserved for its water purification ecosystem services. This can be achieved by maintaining soft sediments in estuaries, rather than constructing concrete canals in which C. kraussi cannot burrow. Regulation of C. kraussi extraction for fishing bait is also necessary to prevent the loss of their filtration function. Furthermore, this study supports the use of C. kraussi as a NBS to eutrophication and encourages the consideration of similar endobenthic organisms for this purpose. Endobenthic ecosystem engineering should not be underestimated in addressing the global challenge of eutrophication.

Includes bibliographical references.

Includes bibliographical references. / Seabird bycatch in commercial fisheries is one of the major factors causing decreases in many seabird populations. In trawl fisheries, high mortalities have been recorded as a result of seabirds being struck by trawl warps (the cables used to tow the trawl net). Tori (bird-scaring) lines have been used to decrease seabird mortality in some trawl fisheries by up to 90%. However, tori lines are not effective at reducing the number of birds that drift towards the trawl warps while feeding on factory discards alongside the vessel. The Albatross Task Force (ATF) helped to develop and test a new device, the Rory Line (RL), to be used in conjunction with tori lines, and designed to reduce warp strikes by placing a physical barrier between the scupper (where factory discards are released) and the danger zone (where the trawl warps enter the water) at the stern of the vessel. This study tests the efficacy of the RL at reducing the number of birds drifting into the danger zone and the number of birds being struck by the trawl warps.

The Berg River is a pivotal source of fresh water for domestic, industrial and agricultural use as well as for in stream ecology, therefore knowledge of what impacts this rivers water quality and assessing whether ecological resilience has been surpassed are of the utmost importance. Since diatom assemblages are inextricably linked to the chemical, physical and biological characteristics of their environment, they were chosen for this study to investigate long-term changes in water quality and availability and suggest potential drivers of such changes at an agricultural site (Rhenostervlei Farm) adjacent the Berg River in the Cape lowlands through fossil diatom analysis of a sediment core (RV3). Diatoms were extracted along the length of the RV3 core, chemically and physically treated to remove unwanted material and then were mounted on a slide to be counted. Twenty abundant diatoms were chosen as environmental indicators for the analysis. Their abundances were plotted against depth and age and changes in water quality and availability were inferred based on their autecological characteristics. The most prominent shift in the indicator diatoms at Rhenostervlei Farm as shown by the stratigraphic diagrams, CONISS analysis and the PCA occurred at the onset of the 20th century CE and was characterised by a shift from a saline, dry and nutrient-poor environment (ca. 1790-1890 CE) to a more turbid, nutrient-rich environment with increased freshwater influence that was prone to periodic flooding (ca. 1890-2011 CE). Through the chronological analysis and the interpretation of the historical record (climate and land-use) as well as other environmental proxies (sediment accumulation rate and macro-charcoal), the causes of the detected change in water availability and quality at the floodplain site were likely related to land-use change in the form of agricultural intensification at Rhenostervlei Farm and potentially in the Berg River catchment as a whole. This could have involved burning, clearance of natural vegetation, soil disturbance and fertilizer use - all of which contributed to increased surface runoff, erosion and nutrient and sediment loading into the site. Furthermore, water extraction and diversion in the 1950s could explain the decreased flooding signal (i.e. lower abundance of Aulacoseira granulata). Although no evidence of a catastrophic regime shift was identified, if land-use practices continue to intensify (e.g. increased fertilizer use) and future climate change interacts and influences the agricultural alterations to hydrological systems, we may expect increased vulnerability to global change and unexpected ecological outcomes such as regime shifts. In order to improve the interpretation of fossil diatom records in terms of changing water quality and availability, a study with multiple proxies should be undertaken to help infer environmental conditions in a complex environment that has many potential drivers, such as the Berg River.

Includes bibliographical references (p. 97-104).

In this thesis I determine how an emphasis on tourism may have both a top-down and bottom-up affect on the two most dominant tree species in the southern Kalahari Desert of central southern Africa. Both tree species are considered as keystone species especially along ephemeral rivers. My focus is on the impact of the introduction of an extralimital megaherbivore, Giraffa camelopardalis, (top-down) and anthropogenic water abstraction (bottom-up) on the reproductive output and vegetation structure of Acacia erioloba and Acacia haematoxylon in the Kgalagadi Transfrontier Park. I determine the impact of giraffe on canopy structure and reproductive potential of the two tree species by comparing three zones of giraffe density (high, medium and low) along the Auob River. I do this through a series of transects through the river to determine both giraffe and tree density and also photographed several trees to determine change in canopy width, percentage canopy death and change in the number of flowers and pods between the three zones and between species. My results show a significant negative impact of giraffe browse on canopy structure, specifically for A. haematoxylon. I did not however find any significant differences in recruitment between the different giraffe density zones. However, a noticeable decrease in numbers of flowers and pods in the giraffe browse height of trees (2 - 5 m) between zones suggest that giraffe will have a negative effect on the A. haematoxylon population but not the A. erioloba population in the future. I use stable hydrogen and oxygen isotope ratios to determine the water source for the two tree species. I do this relative to the stable isotope ratio of the water extracted from nearby boreholes. I also determine how close to physiological stress the trees are (using specific leaf area, leaf δ¹³C values, canopy death and midday xylem pressure potentials) both upstream and downstream of active boreholes used for two tourist camps. One camp, Nossob, in the Nossob River is abstracting at a high rate while the other camp, Urikaruus, in the Auob River is abstracting at a lower rate. My results for water isotope ratios show that both A. erioloba and A. haematoxylon are using deep groundwater. My results also show that in the dry season, A. erioloba in the Nossob are losing contact with groundwater and have to rely on some as yet unmeasured water source in the soil profile. This 'stress' is not reflected in canopy dieback as yet. The trees continue to transpire in the dry season suggesting that they are physiologically able to adapt to fluctuations in the water table of between 4.5 - 5.2 m between the wet and dry season by closing their stomata but are prone to drought-induced carbon starvation, specifically for trees in the Nossob. Overall, my results show that the impact of giraffe on vegetation structure is already evident for A. haematoxylon, indicative of future negative effects on reproduction. My research also shows that although current levels of water abstraction are as yet not affecting vegetation structure below the abstraction points, the trees are edging closer to thresholds of water stress that would be exacerbated with drought. These findings are discussed within the context of ecosystem health and management implications for the giraffe population and water use within the Park.

African wild dogs (Lycaon pictus) are one of the most endangered carnivores in southern Africa. Direct persecution, prey decline and habitat loss and fragmentation all contributed to a rapid decline in this species’ population size and distribution during the 20th century. Following a thorough population viability analysis in the late 1990s the decision was taken to manage the South African population as a metapopulation. This involved the reintroduction of packs to small, fenced protected areas and the subsequent transfer of individuals or small groups between reserves to avoid inbreeding. A key component of successful metapopulation management is post-release monitoring to provide data on the determinants of reintroduction success and failure, particularly when establishing new populations. This study aimed to provide information on the post-release behaviour and movements of a pack of eight African wild dogs introduced to the Northern Tuli Game Reserve in eastern Botswana in February 2017. Two individuals from the introduced pack were fitted with GPS collars. A total of 933 GPS locations were recorded between February 2017 and October 2017. Movement data was used to analyse home range, habitat resistance and resource utilisation by this pack across a range of spatial and temporal scales. Monthly 95% kernel density estimations revealed a mean home range of 330.02 km2 . A reduction in home range size to 37% of the average monthly 95% kernel density estimations revealed that the pack commenced denning in May 2017. However, this denning attempt failed, as shown by the home range size increasing only a month after it initially contracted, which is less than the expected contraction period required to produce a successful litter. Habitat resistance analysis revealed that the pack readily crossed fences but not rivers, with the Limpopo river serving as a very hard barrier that consistently deflected pack movement parallel to its course. Resource utilisation functions showed a preference for sites far from riverine areas, with low elevation and rough and rocky terrain. I propose that this may reflect a predator avoidance pattern, with lions (Panthera leo) in particular preferring riverine habitat with a less rough terrain in this area. The persistence of this pack in the landscape nine months post-release indicates that this reintroduction has been a partial success. Large perennial rivers provide important barriers to the movement of this pack and may thus be important for mitigating local human-wild dog conflict. Fences, however, were readily traversed and therefore communities outside of fenced areas are likely to experience conflict with this pack and any future packs re-introduced to this area.

The aim of this study was to evaluate the potential costs and benefits of increasing conservation in the Sabaki River Mouth Important Bird Area. This was achieved by assessing the vulnerability of waterbirds to disturbance from human activity in the estuary, the current use and importance of the area to local livelihoods, and the value (current and potential) of tourism. Waterbird densities and levels of human activity were quantified from 20 September to 25 November 2010 in the intertidal area of the Sabaki River Mouth Important Bird Area on the central Kenyan coast. Household surveys were conducted in the adjacent Sabaki Village from 15 October to 24 November 2010 (N = 190). The current and potential value of ecotourism was investigated by recording visitation rates and interviewing visitors to ascertain their preferences and willingness to pay an entry fee from 5 October to 4 November 2010. Three types of response variables were collected at six sites to characterize relative responses of waterbirds to simulated human disturbances. These were 1) changes in bird density within a 40 m radius of a stationary disturbance (D40); 2) minimum distance of birds from the source of a stationary disturbance (¡Ü 40 m); and 3) the time for 90% of original bird abundance to recover following a mobile human disturbance which caused all birds to flee the immediate vicinity. Disturbance response metrics were estimated from these variables by calculating normalized mean residuals from regressions of density (D40) and minimum approach distance against expected densities measured in the absence of disturbance, for 14 waterbird species.

In this thesis I determine the effect of population decline on reproduction for two critically endangered African cedars. Widdringtonia whytei (Mulanje cedar) endemic to Mt Mulanje, in Malawi and Widdringtonia cedarbergensis (Clanwilliam cedar) endemic to the Cedarberg Mountains in the Western Cape Province of South Africa. Populations of both species have declined dramatically in the recent past and both show evidence of insufficient recruitment. Recent research has shown that a reduction in population density may limit pollen dispersal, reduce seed viability and increase self-pollination rates, causing an inbreeding depression, resulting in less fit offspring. Based on this research, I hypothesise that pollen transfer in Widdringtonia is distance-dependant and therefore populations with greater distances to the nearest adult neighbour will have lower seed viability, due to pollination failure. I also hypothesise, that trees which receive less outcross pollen due to distance-dependant pollination will have higher rates of self-pollination. I further hypothesise that seed viability may also increase with increased soil nutrients and more amenable climate differences linked to changes in altitude. To test these hypotheses, I determine the extent to which reduced population density has resulted in a decline in viable seed using germination experiments, followed by cut tests and tetrazolium chloride tests. For W. cedarbergensis seedling survival was determined in a greenhouse and self-pollination was assessed with ISSR markers, using DNA extracted from parent and offspring. My results show that seed viability for both W. cedarbergensis and W. whytei is not significantly correlated with distance to nearest-neighbour, altitude or soil nitrogen, carbon or phosphorus. These results suggest that current population densities are not effecting the reproduction either of W. whytei or W. cedarbergensis. My results for population genetics show relatively low levels of genetic variation in W. cedarbergensis typical of endangered and endemic species. The genetic differentiation between populations is low, suggesting that pollen flow between populations is adequate and populations are not genetically isolated. I conclude that there is no evidence that population decline is causing any noticeable limitations on pollen transfer and reproduction in Widdringtonia.