Includes bibliographical references. / The mechanism by which trees die following a low intensity fire is poorly understood. Traditionally, cambial necrosis (Carbon starvation hypothesis) has been used to explain post-fire tree mortality, however, this does not explain why some trees die within days following a fire. To address this, the hydraulic death hypothesis argues that post-fire tree mortality is a result of a reduction in hydraulic conductance rather than the necrosis of cambium. There are a number of proposed mechanisms by which hydraulic failure can occur during a fire: firstly, plume-induced cavitation as a result of high vapor pressure deficit (VPD) in a fire-plume has been shown to reduce hydraulic conductance. Secondly, changing surface tension as water is heated has also been shown to increase the chance of cavitation. The final mechanism is a reduction in conductance as a result of direct xylem vessel deformation due to the visco-elastic properties of xylem walls (lignin). In order to determine the relative importance of each proposed mechanism, stems of Kiggelaria africana and Eucalyptus cladocalyx were exposed to 70 and 100„aC in two treatments designed to isolate the effect of each mechanism. An oven treatment was used as a surrogate for a fire-plume in order to demonstrate VPD-induced cavitation and a water bath treatment (transpiration inhibited) was used to demonstrate xylem deformation (along with microscopy). This was possible because post-treatment flushing was indicative of the initial cause of the reduction as cavitation is reversible while deformation is permanent. The data was then explored using a Hydraulic Death model we created based on a xylem conductance model from literature. The results showed that VPD-induced cavitation as well as deformation are able to reduce hydraulic conductance in trees exposed to fire, however, E. cladocalyx showed higher loss of conductance at 65„aC than K. africana and deformation was only seen to occur in water bath treatments and only in K. africana. Here we propose that a chain of events provides a mechanism for slowing the rate of heating in branches exposed to a fire-plume and that cavitation plays a protective role. Model exploration implied that vulnerability segmentation is responsible for preventing fire-plume induced runaway cavitation. This is in agreement with the ¡§safety valve hypothesis¡¨, however, rather than drought stress, it is a fire-plume which causes the cavitation. It was also found that E. cladocalyx was able to prevent deformation because of xylem vessel characteristics (thick vessel walls) and not bark properties. We propose that the necrosis of cambium and phloem leading to the inability to refill cavitated vessels is the actual cause of mortality in trees exposed to low intensity fires. The ability to refill is dependent on water availability and carbohydrate content, which is highly sensitive to drought. As resprouters store water and carbohydrates in lignotubers and stems, they are less sensitive to pre-fire conditions. However, the survival of cambium and phloem is essential to the refilling process and thus the mechanism for reducing heat transfer, bark properties as well as xylem characteristics work in combination to determine persistence after a fire.

Bird-wine grape damage is a globally acknowledged problem that has been the subject of considerable research in many wine growing regions. However, despite the Western Cape of South Africa being a major wine grape growing region, very little research has quantified bird damage in this region and very little is known about the extent of the problem. This research aimed to quantify bird damage to wine grapes in four grape growing regions of the Western Cape, through questionnaires. It also aimed to determine the factors that might explain the patterns of damage observed. 102 questionnaires were sent to the wine farms, and 52 were returned. Of the respondents, 71% reported bird grape damage. The amount of damage per wine farm ranged between 0% and 15% of the total rows of vines across a farm. When exploring the factors associated with whether a wine farm experienced bird damage, it was found that the border habitat of a farm was a significant factor. The composition of the border habitat greatly influenced the probability of bird damage. The PCA showed that tall trees, Fynbos and residential areas specifically increased the probability of bird damage. Within farms that experienced damage, it was found that in different vineyard blocks the areas closer to the edge of a vineyard and to trees were significantly more susceptible to bird damage. The most common reported damage-causing bird species were Red-Winged Starlings (Onycognathus morio) and White-Eyes (Zosterops lateralis). Damage was reported to occur mostly one month before harvest and at harvest time. This research provides a baseline for understanding the nature of bird damage in these regions and will aid future quantitative studies aimed at the management of the problem.

No description available.

In understanding the behaviour of social-ecological systems, much focus has been placed on the role of institutions that govern how natural resources should be managed, and the biophysical processes affected by this management. Somewhat less attention has been given to the role played by the natural resource managers themselves. Novel insight into social-ecological systems can be gained from understanding why managers act as they do and how management actions become reinforced into (un)sustainable management regimes. In this thesis I applied a social-ecological systems perspective to the phenomenon of private land conservation. With increasing interest in the role that the private sector can play in global conservation efforts, a pertinent but largely unexplored question is whether private land conservation areas (PLCAs) can conserve biodiversity over sufficiently long time scales. This thesis contributes to social-ecological systems theory through an in-depth analysis of the multi-scale interactions among natural resources, managers and socioeconomic processes, which affect PLCA management practices and their sustainability. The potential ability for commercially operated PLCAs to generate the funds necessary for their maintenance makes them an attractive conservation strategy in an economically-orientated world. There are concerns, however, that (a) their long-term sustainability may be dependent on their ability to become and remain financially viable; and (b) they may be tempted to prioritize profit over biodiversity protection in their management practices, thereby jeopardizing their ecological sustainability. The objectives of this thesis were to investigate if, how and why commercial PLCA managers (a) meet their financial objectives and (b) adopt unsustainable ecological management practices. During 2014 and 2015 I interviewed the managers of 72 commercial PLCAs in the Western and Eastern Cape Provinces of South Africa, a region that supports a rapidly growing yet poorly understood PLCA industry. I applied theories from organizational ecology to understand patterns in the income-generating strategies adopted by PLCA managers. Distinct business models were evident, differentiated by incompatibilities in the biophysical (size, fauna) and socioeconomic (activities, affordability) characteristics of these conservation areas. PLCAs characterized by financial objectives but unprofitable business models suggest that these incompatibilities constrain the ability of managers to effectively adapt to their economic environment, a concept known as "structural inertia" in organizational ecology. Profitability was highest on PLCAs that supported megaherbivores and large predators, reflecting international tourist preferences for charismatic game. Managers' financial objectives influenced the strategies that they employed to manage these mammals. When managers used revenue-generation to inform their management decisions, they undertook management actions that enabled them to maximize and stabilize game populations. While this intensive management resulted in higher revenues, it corresponded in many cases with a lack of ecological monitoring and an increased risk of overstocking game. Regional policy guidelines for large predator management both mitigated and exacerbated the mismatch between financially-desirable and ecologically-sustainable management, depending on whether species-specific guidelines were ecologically appropriate or not. Simulations of a mechanistic PLCA model were used to test whether adopted management strategies influenced the observed constraints on business model adaptation. If the income-generating potential of an adopted business model was low, managers were unable to accumulate the capital necessary to overcome the biophysical and socioeconomic incompatibilities that separated business models, constraining their adaptive capacity. Intensively managed PLCAs were able to generate a more stable annual income, accumulate more capital and overcome constraints on adaptation faster than PLCAs managed according to ecological monitoring. This unique, large-sample assessment of the social-ecological mechanisms underlying PLCA sustainability emphasizes the significant role that managers can play in promoting resilient social-ecological systems. When financial viability is an important consideration, broad-scale socioeconomic factors can influence fine-scale management decisions. Through constraints on adaptation, and the presence or absence of corrective feedbacks between management actions and ecological monitoring, these management decisions can become reinforced into management regimes on a trajectory towards, or away from, sustainability. This study therefore provides a novel contribution to our understanding of how the interactions between managers and ecosystems influence the behaviour of social-ecological systems.

As the world's human population increases, so does the competition for natural resources between humans and wildlife. This competition may be intense for apex predators, such as raptors, which generally require large natural areas in order to maintain their populations. Anthropogenic development within territories can cause individuals to either abandon these sites, reduce their breeding productivity, or cause direct mortality to the territory holding birds. To mitigate such impacts, one method, employed as part of Environmental Impact Assessments (EIA), is the use buffer circles centred on nest sites. Within these buffers the most damaging forms of development are prohibited. This approach assumes that raptors use the space around their nest in a uniform way, but this assumption may not always be correct and few have evaluated the effectiveness of buffer circles at protecting a species' home range. This study uses tracking data to evaluate the effectiveness of buffer circles to cover the ranging movements of six southern African raptor species, throughout the year, as well as during their breeding and non-breeding season. My study revealed that buffer circles whose dimensions were based on the species' 95% Kernel Density Estimate (KDE) did relatively well at capturing the proportion of individual GPS fixes, but did less well at capturing the KDE area from tracked birds. For buffer circles to capture 95% of the home range polygons (95% KDE) they would generally need to be at least twice as large as those that were derived from the 95% KDE home range area, and for some species with very large home ranges (e.g. Lappet-faced Vultures) even buffer circles that were 3 times the size failed to cover 95% of the KDE polygons.

There are few studies that empirically test whether education can improve stakeholder support for a particular wildlife management/conservation intervention. Evidence-based decision making is critical for conservation actions, especially when people and wildlife are engaged in negative interactions to the detriment of both. Rising human and baboon numbers in the Cape Peninsula of South Africa is associated with increasing levels of damage to properties, both indirect and direct human-induced injury and mortality to baboons. A possible management intervention to reduce these negative interactions is a baboon-proof electric fence which has already been shown to be very effective in the suburb of Zwaanswyk on the Peninsula but surprisingly has low public acceptance. For this study, I produced a short survey that included an educational video on the welfare, conservation, and lifestyle benefits of a baboon-proof electric fence. Two versions of the same survey were created but with the educational video provided either before or after questions pertaining to the level of support for the fence. Using a randomised controlled trial method, I tested whether the level of support for the fence changed according to the order of viewing the video. I also used a pre- and post-test method to explore whether level of support for the fence changed for those who watched the video at the end and were given a chance to answer the question again about level of support for the fence. The results showed that watching the video before the survey increased the average marginal probability of supporting an electric fence by 15 percentage points and reduced the average marginal probability of not supporting the fence by 17 percentage points. There was also a significant difference in level of support for the fence in the pre- and post-test experiment, and my results show that women were more likely to adjust their response in favour of the fence. Implications of these results are significant as managing a peri-urban species is considerably more effective if there is collaboration, agreement, and support for management measures between and within key stakeholders. This study contributes to the emerging literature on the importance of education in managing human-wildlife conflict.

Conservation units are a tool to guide policy such that conservation goals can be achieved. These units should ideally synthesise a wide array of data – genetic, ecological, demographic – to identify the appropriate scales at which conservation actions can then be directed. Despite being the most widespread of all felids in Africa, and facing numerous threats across its range, it has been proposed that the entirety of southern Africa be considered one conservation unit for leopards (Panthera pardus pardus). This proposal does not take into consideration the likelihood of existing population genetic structure across an increasingly fragmented landscape. Further complicating regional leopard conservation is the variability in conservation policies among the geopolitical leopard-range states. Within this single proposed conservation unit, the patchwork of different legislation does not support a unified policy for leopard conservation. Using a population genetic perspective, this study explores and tests the values and shortcomings of southern Africa as a single conservation unit, and explores the importance of leopard range states within the context of conservation units. Parallel investigations of leopard microsatellite genotype data within the framework of a genetic population study spanning eight countries across southern Africa were carried out. This study presents consilient evidence supporting the finding that southern Africa contains six clusters of unique genetic lineages, and as such does not constitute a single genetic unit. Furthermore, it is shown here that the spatial genetic structure that exists does not correlate with the separate geopolitical range states. Leopard range states within southern Africa instead capture varying levels of unique genetic structure and thus are not of equal value with respect to the conservation of genetic lineages. These findings have several implications for leopard conservation across the region. While the data presented here specifically consider a genetic element of conservation units, they do suggest shortcomings in adopting either the entirety of southern Africa as a single unit or separate geopolitical range states as conservation units. The variability in leopard conservation policy across southern Africa is unlikely to sufficiently protect their existing regional genetic structure. If conservation units are indeed a tool to guide conservation policy, then the southern Africa unit for leopards is potentially less effective than a smaller unit whose spatial scale more accurately captures the discrete variation in population genetic structure. Genetic diversity and population structure is an important component of conservation units and should not be neglected. Currently, however, an appropriate framework allowing for conservation policy to be informed at the necessary scale does not exist; although the establishment of Transfrontier Conservation Areas speaks to the growing acknowledgement that conservation needs to evolve beyond the historical confines of geopolitical range states. The evidence presented here further supports the need for a rethinking of existing policy structures.

Bibliography: leaf 195.

The colony of African Penguins at Boulders, Simon's Town on the southern Cape Peninsula, Western Cape formed in 1985, and is now a major ecotourism destination. The African Penguin is listed as Endangered due in part to competition with commercial fisheries. Improvements in the spatial management of these fisheries could assist with the conservation of the penguins, but would be costly. Various techniques were used to estimate the value of the colony, and to assess its impact on the economy of the southern Peninsula.

Leopards (Panthera pardus) are faced with increasing levels of anthropogenic pressure resulting in population declines across much of their historical range. While there is relatively limited knowledge of leopards occurring in African rain forests, their abundance and distribution is assumed to be impacted by a combination of several anthropogenic factors, most notably loss of prey and habitat conversion. In this study I used a long-term camera trap array that forms part of the Tropical Ecology, Assessment and Monitoring (TEAM) project - Terrestrial Vertebrate (Camera Trapping) Monitoring Protocol, to estimate the species richness of mammals, the relative abundance of leopard prey species and leopard habitat use in the Nouabalé-Ndoki National Park (NNNP) in the Republic of Congo. I investigated the impact of different environmental and anthropogenic factors on leopard occurrence at two camera trap arrays (a northern and southern cluster) within the NNNP using occupancy modelling. While there were no significant differences in mammalian species richness between the two clusters there was a higher relative abundance of the preferred prey species of leopards in the northern cluster. A total of 106 leopard photographic events were recorded across all camera traps and all survey years. The top occupancy model produced an average probability of site use (ψ) over all sites of 0.52 ± 0.14 (SE). The covariate specific β-coefficient estimate suggests that leopard occurrence and detectability were positively correlated with both the relative abundance of the blue duiker (Philantomba monticola) and the distance to the nearest river (β =0.062 ± SE 0.053 and 6.55 ± SE 10.84, respectively). Therefore there was no support for the prediction that the probability of leopard habitat use increases with a higher relative abundance of all potential prey species, nor was there support for the prediction that leopard habitat use would be higher further away from human settlements (β =3.42 ± SE 2.94). 2 Leopard habitat use was higher in the southern cluster which may be linked to the denser understory that would provide greater cover which is important for improved hunting success in leopards. Together, these results suggest that both the prey species and leopards appear to be thriving within the NNNP with limited evidence of anthropogenic impacts despite an increase in commercial logging and the itinerant bushmeat hunting in the peripheral area. It would be worth expanding the existing camera trap arrays to cover a greater spatial extent within NNNP and hence allow for a more detailed analysis of edge effects and to detect the potential impacts of anthropogenic activities which are predicted to increase in selected villages in the periphery of the park.