Globally, ecosystems are under mounting pressure as biodiversity is lost at an ever increasing rate due to drivers such as habitat destruction and climate change. The systematic degradation of natural habitats witnessed today is often accompanied by a loss of ecosystem functioning and services which not only endangers the future of humankind but has consequences for all life on earth. To manage the ecological challenges facing us there is an urgent need to increase our understanding of how ecosystems function, the relation/ship between biodiversity and ecosystem functioning, and the effect habitat degradation can have on this relationship. Eastern Rûens Shale Renosterveld, located in the Overberg region of South Africa, is a vegetation type both critically endangered and poorly understood. Centuries of agricultural activity in the region has resulted in a landscape typified by fragments of pristine vegetation, in addition to communities in various states of degradation, embedded in an agricultural matrix. The current quantities of pristine vegetation are inadequate to meet conservation goals and conservation efforts are further challenged by the fact that little is known of the functional ecology of not only pristine fragments of Renosterveld, but the degraded communities in various stages of secondary succession. The overarching aim of this thesis is to better understand the relationship between biodiversity and ecosystem functioning across a degradation gradient in critically endangered Renosterveld vegetation within an agricultural landscape in South Africa. To achieve this, a variety of methods were employed using different lenses of analysis. In Chapter 2, I studied the components of biodiversity to assess the relationship between taxonomic and functional diversity indices and investigate the effect of habitat degradation. In Chapters 3 and 4, I used litter traps to investigate the effect of plant litter type, season and habitat degradation on litter decomposition rates and Springtail (Collembola) community dynamics, respectively. Finally, in Chapter 5, I constructed three high-resolution plant-pollinator networks from sites with distinct land-use histories and with different above-ground vegetation communities. The results revealed a complex association between different taxonomic and functional diversity indices, influenced by habitat degradation, with potential ecological and conservation implications. Particularly, the loss of functional redundancy in degraded sites is likely to reduce resilience to future environmental perturbations which may reduce ecosystem functions. Conversely, the similarities in both taxonomic and functional diversity indices between pristine and moderately degraded sites may be cautiously interpreted as the occurrence of successful passive restoration. Litter decomposition rates were shown to be variable with litter type and season revealed as important controlling factors. Although degradation did not appear to significantly affect iv decomposition rates, the initial nutrient content of litter appears to correlate with decomposition rate and it can be expected this ecosystem function will be accelerated where habitat degradation results in shifts in above-ground vegetation and subsequent litter input, specifically where the cover of non-native, nitrogen-rich annual species is increased. There was found to be a significant impact of litter type on community composition, and of sampling day on species richness, abundance and community composition, in Springtail communities. Despite the overall lack of effect of degradation on Springtail community dynamics, the abundance of the non-native Entomobrya multifasciata in degraded sites, and its absence from pristine sites, raises the intriguing possibility of its suitability as a bioindicator for habitat degradation. Comparisons to similar global studies revealed the Renosterveld networks to be highly functionally specialized. Assessing network dynamics across a degradation gradient showed the impact of above-ground vegetation structure on network properties with the more open and diverse vegetation structure and floral resources provided by the highly degraded site resulting in network indices more similar to that of the pristine site when compared to the relatively structurally uniform moderately degraded site. Although this thesis has enhanced our understanding of the functional ecology of Renosterveld, it has also highlighted knowledge gaps which still exist. Creating and collating a database of functional trait data can provide the building blocks for future ecological work. Furthermore, to truly gain a mechanistic understanding of the biodiversity-ecosystem function relationship one needs to utilise an integrated analysis which considers different facets of biodiversity, particularly functional diversity, across multiple trophic levels while simultaneously acknowledging the legacy effects that distinct land-use histories can impose at the community level.

This study aims to increase our perspective of the responses of Southern African climate to soil moisture forcings by drying or moistening the land surface using a regional climate model version 3, RegCM3. The sensitivity and response capabilities to soil moisture perturbations of the model are investigated. This includes identification of regions that may be influenced differently by antecedent soil moisture conditions as well as understand the implications of soil moisture conditions on frequency and intensity of rainfall. Exploratory analyses of soil moisture retention and comparison of climate model parameters with available observations or re-analysis data is done. The study then seeks out the large-scale atmospheric forcings under which the regional climate explicitly responds to perturbations in soil moisture using self-organising map technique. To investigate these underlying processes of atmosphere-soil moisture interactions a series of RegCM3 model experiments utilizing wet, dry and normal soil moisture conditions were designed. The experiments are based on changing the soil moisture field capacity in the RegCM3. The control simulations are run with RegCM3 nested in NCEP/NCAR reanalysis 2 data and using Emanuel convective scheme for the selected six summers (dry seasons - 1991/92, 1994/95 and 1997/98; wet seasons - 1995/96, 1996/97 and 1998/99). September to March simulations are performed with August as the spin-up month. The respective dry and wet soil moisture perturbation simulations are then initialised at field capacities of 25% (wilting) and 75% (saturation) within the land surface model, Bio-sphere Atmosphere Transfer Scheme. From the sensitivity studies, anomalously dry (wet) conditions have positive feedbacks with similar dry (wet) synoptic forcings of the regional climate. Anomalous dry forcing persists for longer and exacerbates the changes in the regional anticyclonic circulation especially during a drought or dry period. Soil moisture perturbations mostly affect the lower troposphere. Surface variables analysed especially surface temperature show strong responses to the soil moisture perturbations under all synoptic forcings but rainfall characteristics are strongly influenced by large-scale synoptic circulations. However, in some areas over southwestern parts of the region a weak feedback which can be either positive or negative depending on geographical and climatological setting has been detected.

This thesis explores the utility of the MM5 regional climate model iii product big a detailed wind climatology for South Africa in the context of wind power applications. Given the background of limited available computational resources a sophisticated temporal subsetting algorithm is developed and applied. The goal here is to allow a high resolution 18 by 18km simulation to be carried out over only 365 days of historical records (20 March 1996 until 19 March 1997), yet ensure that it is possible to construct a reasonably accurate climatology from such a short time period because it represents the most typical year available. Throughout the research the modelâs sensitivity towards different boundary layer and cwnulus parametrisations as well as observational and grid nudging is investigated in depth. New error metrics are specifically designed and applied based on 17 reliable lOni weather stations around the country with the specific purpose of evaluating model skill with respect to the wind patterns that are important for wind turbines. An optimal model configuration is established by using the Pleim Xiu land surface scheme coupled with the Atmospheric Convective Model for the planetary boundary layer, in addition to the Kain Fiitsch II cumulus parametrisation and observational wind nudging. The model is then integrated at 18km for the entire country and the results validated in detail. It is found that average wind speeds are within 0.2 + 0.7ms (68% confidence interval) and that the daily cycle, which constitutes the largest part of the observed variance, is predicted to within 18% + 20% of amplitude and lii of timing at coastal sites.

Includes abstract. / Includes bibliographical references (p. 213-247). / This thesis uses the Selous-Niassa wildlife corridor as a lens through which the process of constructing bioregions can be understood and the effects of that process on society properly evaluated. It specifically investigates the corridor as a cog in the creation of a bioregion in southeastern Tanzania, namely, the Selous-Niassa transfrontier conservation area. The study was motivated by claims that the creation of bioregions across international borders places the protection and conservation of biodiversity at the appropriate scale, and that bioregions of this type are beneficial for nature conservation and society. Though the study appreciates the ecological rationales for trans-border conservation, its focus is on the social side of the process. The main social claims for bioregions in general, and transfrontier conservation in particular, are that the establishment of cross-border protected areas, including transfrontier conservation areas, leads to the removal of colonial borders which disrupt ecological systems and local communities.

Includes abstract. / Includes bibliographical references. / Climate change is evident in the Western Cape province of South Africa, particularly in observed trends in average temperatures. Further increases are expected in the future, based on General Circulation Model (GCM) projections, as highlighted in the Inter-governmental Panel on Climate Change (IPCC) 4th (and previous) assessment reports. Whilst it is recognised that rises in temperature coupled with changes in rainfall will impact wheat yields (the province’s dominant field crop), little information exists to guide adaptation planning, especially on the potential range of climate change impacts on dryland winter wheat production. Furthermore the Western Cape is a highly diverse region with regard to geology, soils, topography, climatic influences and agricultural systems. Future climate change therefore, is likely to have different impacts in different zones of the province where wheat is produced. To address this heterogeneity, the APSIM crop model was applied to assess future climate impacts on wheat in 21 relatively homogeneous farming areas (RHFAs) across the province.

Building on the work of Martha Nussbaum and Amartya Sen, this research aims to enhance integration of human development and well-being in environmental assessment through developing a capabilities approach to the practice. The research emphasises the effectiveness and equity imperatives of public participation and highlights the inclusion of appropriate social considerations in environmental decision making. The participatory focus emphasises the potential for decision shaping by stakeholders and decision support for stakeholders to participate meaningfully in environmental assessment. The research develops an evaluative framework for public participation that better considers the capabilities of stakeholders. It explores the potential consilience of the capabilities approach and that of environmental assessment, with emphasis on the principles of justice in participatory decision making. A mixed methods approach explores, tests and evaluates a selection of five South African environmental assessment case studies using an applied capabilities framework. Four main methods are employed, a discourse analysis of environmental assessment reports, a conventional Q methodology, an adapted ranking Q methodology, and a survey using Likert scales. The research findings highlight the relationship between the stakeholder's capability considerations that relate to aspects of their 'ability', 'opportunity' and 'constraints' to participation. The research ranks an array of capabilities and provides insight into the types of capabilities stakeholders value highly when reflecting on their participation experience in environmental decision making. Reflecting on the emergent findings from the cases, the research contributes to the praxis of environmental assessment through theoretical development. The theoretical framework focuses on an individual's participation capabilities as well as a broader consideration of capabilities for practice to increase the realizable opportunities, or freedoms, to choose the kinds of environmental futures that can reasonably be considered as valuable and sustainable. Capability concepts of 'ceilings', 'thresholds' and 'capability sufficiency' are commended as supplementary to existing practice specifications of 'meaningful' participation. The research commends that the capabilities approach has potential to be included as a core part of the training for assessment practitioners. It also concludes that the field of environmental assessment provides a rich empirical context for the development of a more robust sustainability-orientated capabilities approach.

Includes abstract. Includes bibliographical references (leaves 178-191).

Drought and climate change pose a threat to southern African vegetation. This study examines the response of southern African vegetation to drought in both past and future climates. Multiyear and multi-simulation datasets from three dynamic global vegetation models (DGVMs), namely, Community Land Model version 4 (CLM4), Community Land Model version 4 with Variable Infiltration Capacity hydrology (CLM4VIC), and Organising Carbon and Hydrology in Dynamic Ecosystems designed by Laboratoire des Sciences du Climat et de l’Environnement (ORCHIDEE-LSCE). These three DGVMs and the Community Earth System Model (CESM) were analyzed for the study. The DGVM simulations were forced with the reanalysis climate dataset from the National Centers for Environmental Prediction (NCEP) and the Climatic Research Unit - NCEP (CRUNCEP). The simulated climate results were evaluated with observation datasets from the Climatic Research Unit (CRU), while the simulated vegetation index (i.e. Normalized Difference Vegetation Index, NDVI) were evaluated with NDVI data from the Global Inventory Modelling and Mapping Studies (GIMMS). Meteorological droughts were analyzed at different timescales (1- to 18-month timescales), using two drought indexes: the Standardized Precipitation Evapotranspiration Index (SPEI) and the Standardized Precipitation Index (SPI). The responses of vegetation to drought were quantified by means of Pearson Correlation Analysis. The DGVMs were applied to study the sensitivity of vegetation to fire, while the CESM was used to project impact of climate change on the characteristics of southern African vegetation in the future (up to the year 2100) under the 8.5 Representative Concentration Pathway (RCP8.5) scenario, focusing on impacts at 1.5oC and 2.0oC global warming levels (GWLs). Analysis of the observed data shows that the spatial distribution of vegetation across southern Africa is more influenced by the rainfall distribution than by the temperature distribution. The observed correlation between drought index and vegetation index is higher than 0.8 over southeastern part of the region at 3-month drought timescale, and there is no difference between the spatial distribution of the correlation between the SPEI and the vegetation index, and between the SPI and the vegetation index. The three DGVMs failed to capture the response of vegetation to drought; however, the CLM4 shows the best performance while ORCHIDEELSCE fared the worst of the three. The CLM4 simulation show that fire strongly influences growth of vegetation over the summer rainfall region but it has weak influence over vegetation in the western arid zone. The CESM strongly captures the spatial patterns of precipitation and the vegetation index across southern Africa, but it overestimates the magnitudes of the vegetation index across the region, except in Namibia and Angola. The CESM also underestimates the correlation between drought indexes with vegetation, and the timescales at which the vegetation respond to droughts. The CESM projects an increase in the drought intensity as a result of an increased temperature across southern African biomes. However the increase in drought intensity is more pronounced with the SPEI than with the SPI. CESM also projects a future decrease in the vegetation index (i.e. NDVI) in the region except in the dry savanna biome. The impacts of 1.5oC GWLs on the vegetation fluxes vary throughout southern Africa, and the magnitudes of changes in the vegetation fluxes are affected by a further increase in global warming over the region. While there is a good agreement among the CESM simulations on the projected changes in vegetation fluxes across the biomes, the uncertainty in the projections is higher with 1.5oC than with 2.0oC GWL. The results of the study can be applied to mitigate the impacts of climate variability and change on southern African vegetation. Specific mitigation efforts that could be applied to reduce the impacts of droughts and climate change are watershed management, improved vegetation management, impact monitoring, environmental awareness, and remote sensing tools.

The premise of the thesis is that the developmental use of urban land should be ethical, fair and promote social justice. Using multiple research approaches and mixed methods this thesis examines the urban land question in South Africa and particularly Cape Town where land distribution and ownership are inequitable. The thesis unpacks land redistribution, land tenure reform and land restitution within this context. It is argued that the South African Constitution commits government to protect the property rights of those who currently own property and at the same time redistribute land to those who have been dispossessed and explores what this tension means for urban land redistribution and reform using the freedoms approach developed by Amartya Sen as a conceptual framework and as alternative to the neo-classical model. The main findings of the thesis can be summarized as follows. a) The thesis demonstrates that there is no logical reason why the freedoms approach cannot be extended to include urban land. b) The entitlements and endowments that urban land could bestow on the urban poor are shaped by how the State invests in land through the instruments of land planning and land use management which call for a significant role for the State as custodian of public land to not only make explicit the land asset register under public ownership but also instill trust in the poorer sectors of urban society. c) A two track system of land planning and land use management may be more appropriate in the post apartheid South African city, one stream for market driven land and one for targeted public land programmes that directly address urban poverty provided that the State is able to make strong connections between the philosophical and the technical aspects of land and land use management systems. d) As a two track system is suggested the land use management system requires to be reframed. To facilitate land redistribution and reform in urban areas of South Africa therefore, the thesis suggests that a deliberative and systemic planning approach needs to be adopted that is intervention focused. Only when the State assumes a more critical interventionist role in public land programmes would it be possible to obtain social justice and the principles of the good city in the South African urban context. e) Gaining access to and control over land resources beyond the market is possible but limiting for the majority of the urban poor when land and housing debates are conflated. This conflation results in other land debates being silenced yet these have the potential to offer alternatives to the neo-classical model of land and land use management as well as promote a wider role for public land than just housing.

Includes abstract. / Includes bibliographical references. / Like many of the world’s subtropical regions, the Western Cape of South Africa is highly sensitive to oscillations in the earth’s climate system triggered by major tectonic changes, local variations in orbital forcing, better known as Milankovitch cycles, and its position at the interface between temperate and tropical circulation systems. Regrettably, a dearth of reliable and continuous palaeoenvironmental records means that relatively little is known about how regional environments have been impacted over centennial to multi-millennial timescales. Palaeosols constitute an important stratigraphic marker for past environments and may provide useful validation of pedogenic and other earth system process models. However, the characterisation and analysis of palaeosols has been a largely neglected source of information in the Western Cape, South Africa. This thesis aims to improve the understanding of the environments and climate dynamics using evidences from palaeosols. Various palaeosol-based proxies including geochemistry, mineralogical, macro- and micromorphological, δ13C and δ18O isotope, and selected physico-chemical properties of palaeosols were studied and records obtained for four sites in the Western Cape. A number of chemical weathering indices and geochemical climofunctions were used to calculate weathering intensities, pedogenesis, palaeotemperature and palaeoprecipitation for the various locations using geochemical and stable isotope data. Results of this study indicate that palaeosol-based proxies have the potential to provide snapshots into the palaeoenvironments and palaeoclimate of Western Cape and may complement previous studies done with other proxies such as pollens and diatoms. For example, the interpretation of the pedofeatures (calcareousness, vertic, gleyic, illuviation) - from the micromorphology - of thepalaeosols suggests cyclic patterns of erosion and deposition that correlate with climate changes of the past. Podzolization and laterization are the principal pedogenic processes responsible for the red palaeosol formation at the Cape Peninsula, while calcification and salinization are accountable for the Quaternary palaeosols. Inferences of gleization and lessivage are only evident in the mid-Miocene palaeosol at Langebaanweg. At LBW, pedogenesis was more advanced in the Mid-Miocene and Early Pliocene layers signifying a more humid and warmer climate with more stable landscape. Two major regional climate cycles were evident at the Cape Peninsula: relatively warm and humid subtropical climate which gave rise to the pedogenically modified buried red palaeosol and dry semi-arid Mediterranean climate under which the soils overlying the stone line is currently forming, as seen from the poor horizonation and translocation of materials. Clay mineral assemblages suggest the Quaternary palaeoclimate of the Western Cape has been predominantly characterised by low precipitation and active coastal erosion, which jointly accounted for poor soil profile development. This is in agreement with the early reports obtained from marine records of the African continent (e.g. deMenocal, 2005). A cross plot of the δ13C and δ18O of the carbonate palaeosols indicates they all formed under strong marine influence and C3 plants have been dominant since late Quaternary.