ResearchSpace

ResearchSpace is the institutional repository of the University of KwaZulu-Natal,
unlocking knowledge, empowering impact, and preserving UKZN's research legacy.


ReseachSpace Libguide

Communities in ResearchSpace

Select a community to browse its collections.

Now showing 1 - 5 of 11

Recent Submissions

Advancing the understanding of the hydrological dynamics of the Maputaland coastal region, north-eastern South Africa using remote sensing and In-situ observations.
(2025) Ramjeawon, Manish Roshenlal.; Demlie, Molla Bekele.; Toucher, Michele Lynn.
The Maputaland Coastal Plain in northern KwaZulu-Natal, South Africa, represents a groundwater-dependent region under significant hydrological pressure due to land-use change and climate variability. These pressures have resulted in declining groundwater recharge, wetland and surface water loss, and increased evaporation rates, threatening both ecological and socio-economic systems. This study investigates these challenges by addressing four key research areas: long-term land-use changes, groundwater storage variability, open water evaporation estimates, and water balance analysis. By using satellite remote sensing (GRACE, WaPOR and CHIRPS precipitation datasets), in-situ measurements, and bias correction techniques, this thesis provides a novel, integrated understanding of the hydro-ecological system in the MCP, highlighting the interactions and responses, particularly in relation to groundwater dynamics. The first component of this study analysed land-use changes in the northern half of the MCP from 1986 to 2019 which revealed significant transformations driven by afforestation, agriculture, and rural settlements. Using freely available Landsat satellite images processed and analysed via the Google Earth Engine platform with a random forest classification algorithm, the study identified the development of plantations, which more than doubled in extent. During the same period, surface water bodies and wetlands declined by 36.1% and 49.1%, respectively. These changes highlight the pressures of land-use transformation and climate variability on the MCP’s predominantly groundwater-driven system. The rapid expansion of commercial forestry plantations has contributed to significant reductions in groundwater recharge, amplifying the need for further investigation to quantify the hydrological impacts. The findings underscore the importance of balancing economic activities, such as forestry, with ecological conservation to sustain critical water resources in the region. Following this, groundwater storage changes in the Usutu-Mhlatuze Water Management Area, were analysed using GRACE satellite-derived terrestrial water storage data, GLDAS soil moisture data, and in-situ surface water level measurements. The results revealed significant groundwater storage losses between 2002 and 2020, with the primary aquifer losing 925 × 10⁶ m³ and the secondary aquifers losing 3614 × 10⁶ m³. The GRACE-derived groundwater storage anomalies demonstrated good agreement with limited in-situ observations from monitoring piezometers, achieving R² values of 0.79 and 0.74 for the primary and secondary aquifers, respectively. Seasonal patterns reveal that groundwater recharge predominantly occurs during the summer months, driven by precipitation, while periods of low rainfall intensify storage deficits. This highlights the region's vulnerability to climate variability, especially as rainfall patterns become increasingly variable. Despite the region's reliance on groundwater, the absence of comprehensive aquifer parameter data, such as storativity and specific yield, coupled with sparse and inconsistent monitoring networks, poses challenges for understanding storage dynamics. These findings underscore the urgent need for improved groundwater monitoring infrastructure to better understand and manage the region’s water resources under increasing climatic variability. By integrating GRACE satellite data with in-situ observations, this study bridged critical spatial and temporal gaps, offering a more detailed understanding of groundwater dynamics and emphasising the need for enhanced monitoring infrastructure to support sustainable water resource management. The third component focused on addressing the challenge of accurately estimating open water evaporation (Eo), a critical but under-researched hydrological component, this study developed a novel methodology particularly for Lake Sibayi and Kosi Bay. Satellite-based WaPOR reference evapotranspiration data, integrated with crop coefficients for open water was used to estimate Eo. A key innovation of this research was the incorporation of season-specific bias corrections tailored for each lake, validated using Penman-Monteith open water evaporation values derived from in-situ meteorological data. This approach significantly improved the accuracy of Eo estimates, with R² values increasing from 0.63 to 0.81 for Lake Sibayi and from 0.66 to 0.83 for Lake Kosi. These results underscore the importance of incorporating bias corrections in improving ET estimates and highlight the potential of remote sensing for water management in similar systems. Furthermore, the study provides critical insights into the influence of environmental factors, such as wind speed, temperature, and humidity, on open water evaporation in this region. This methodology offers a benchmark for future research and supports the development of more effective water resource management strategies. Finally, a water balance framework was developed for the W70 catchment to investigate changes in storage from July 2018 to December 2023. Characterised by groundwater-dependent hydrology, negligible surface runoff, and the presence of two major freshwater lakes, Lake Sibayi and Kosi Bay, the catchment offered a unique context for understanding hydrological dynamics. CHIRPS precipitation data, WaPOR-derived evapotranspiration estimates for vegetation (ET) and open water evaporation were used to calculate changes in storage as the residual of a simplified water balance equation. The analysis revealed persistent negative ΔS trends with a mean annual deficient of -0.24 mm.yr-1 across the study period. Sensitivity analysis revealed open water evaporation as the dominant driver of storage losses (sensitivity coefficient -10.43), exceeding the influence of precipitation (+2.19) and vegetative ET (-1.76). Additionally, a sensitivity analysis was performed to determine which climate parameter had the most impact on Eo. The analysis indicates that wind speed (𝑟 = 0.74) and net radiation (r = 0.57) have the strongest correlations with Lake Sibayi’s open water evaporation. Similarly, the sensitivity analysis for Kosi Bay reveals that wind speed (𝑟 = 0.72) and net radiation (r= 0.52) demonstrate the strongest correlations with Eo. This indicates that these parameters are the primary drivers of Eo. Statistical evaluation demonstrated strong correlations (r: 0.80–0.86) between remote sensing and in-situ data, though moderate Nash-Sutcliffe efficiencies (NSE: 0.45–0.73) underscored limitations in capturing short-term variability. This research underscores the necessity of including open water evaporation in water balance studies, particularly in groundwater-driven systems where it dominates hydrological losses. By integrating validated remote sensing datasets with sensitivity analyses, the study advances the understanding of changes in storage and provides a framework for similar data-scarce regions experiencing similar hydrological stress. In conclusion, this thesis offers a comprehensive and integrated approach to understanding the complex hydrological dynamics of the MCP, addressing critical gaps in knowledge related to land-use change, groundwater storage, open water evaporation, and storage changes. The findings emphasise the critical role of incorporating advanced remote sensing techniques, season-specific bias corrections, and sensitivity analyses to improve the accuracy of hydrological assessments in groundwater-dependent and data-scarce regions. By providing insights into the drivers of water loss and storage variability, this research not only enhances our understanding of the MCP’s hydrological system but also establishes a framework for sustainable water resource management. The methodologies and findings presented here serve as a benchmark for future studies, offering valuable tools and perspectives to address similar challenges in other regions.
Examining gender dynamics in the international institutional policies and power structures within Bretton Woods institutions.
(2025) Mfengu, Akho Ongezwa.; Hadebe, Sakhile.
Gender inequality remains one of the most serious global issues; it has been persistent and consistent. It has often been overlooked by power structures and influential institutions to the point where change, emancipation, and liberation are no longer seen as necessary, but are long overdue. Gender-focused policies and strategies have been developed to help narrow the gender gap, covering aspects from women's representation to their participation in decision-making. The study aims to establish how equity and justice are undermined using the Bretton Woods institutions as an example. Bretton Woods institutions have implemented gender mainstreaming strategies to promote economic and political interests, often ignoring the intersectionality of social identities like gender, race, culture, and power, which all influence outcomes across economic, political, and social spheres. Feminist and critical theories enable the study to challenge outdated traditional systems and examine how power structures shape and impact the effectiveness of these policies. Through qualitative document analysis, the study attempts to identify best practices while considering the experiences of marginalized groups. Findings indicate that the gender gap remains wide, as institutional strategies and policies tend to overlook specific country needs and prioritize the interests of power structures. Women’s participation in politics and access to positions of power remain slow and uneven, reflecting poor implementation and adoption. Gender equality is more than just closing the gender gap or creating a 50% men vs 50% women society. It goes as far as breaking traditional cycles rooted in culture and religion and dismantling entrenched patriarchal norms. Gender equality should not be viewed as a standalone issue but integrated into active economic and political strategies.
Identities, interests, regional integration, and economic development : the case of the African Continental Free Trade Area.
(2026) Sidaza, Sizwe.; Mtshali, Khondlo Phillip Thabo.
The framing, adoption, and implementation of the African Continental Free Trade Area (AfCFTA) Agreement signify post-colonial Africa’s cherished commitment to advancing regional economic integration and development to build the Africa envisioned by successive generations of pan-Africanists. The Agreement Establishing the African Continental Free Trade Area aims to create the world’s largest free trade area, supercharge economic growth, foster structural economic transformation, and promote continental economic integration by creating a liberalised market for goods and services, boosting investment opportunities, and paving the way for a future continental customs union. The analysis of the AfCFTA vision has been deficient in examining the role of converging and diverging identities and interests in enabling or constraining the emergence and development of the AfCFTA. This study aimed to explore how the underlying rationale for the establishment of the African Continental Free Trade Area can be understood, interrogate the major players and evaluate their interests in its creation, and examine how the interests and identities of different players shape its development. The study used three theories: Nyerere’s theory of integration, intergovernmentalism, and social constructivism—a theoretical synthesis that embraces diverse perspectives and compensates for the limitations of each. A qualitative approach was adopted, and data was collected through semi-structured interviews. Expert and chain sampling were employed to identify information-rich participants with strong knowledge of the AfCFTA. Thirty-five participants were interviewed, including independent analysts and officials from the African Union’s AfCFTA Secretariat, the United Nations Economic Commission for Africa, the New Partnership for Africa’s Development, the United Nations Conference on Trade and Development, and the South African Institute of International Affairs. A thematic analysis tool was used to analyse the data and identify meaningful trends and patterns. The study found that the promotion of economic diversification, industrialisation, endogenous development, innovation and competitiveness, and the dismantling of colonial economic legacies underpin the establishment of the AfCFTA. It found that several players— including states, intergovernmental organisations, technical institutions, the private sector, the largest economies, and external players—each with different interests, have contributed to the creation and operation of the AfCFTA. It found that several ideational and ascriptive identities—including anticolonial, pan-African, African norms-related, hybrid liberalprotectionist, nationalistic, regional (geo-cultural), cultural, and extra-continental actor identities—shape the development and implementation of the AfCFTA. The study also found that the AU and the RECs have no structured framework for managing identity-related dynamics. Thus, it made several policy and future research recommendations to address the issues identified in the findings and ensure the long-term viability of the AfCFTA, including the need for the African Union and the Regional Economic Communities to develop a structured, institutionalised framework to cultivate a shared pan-African and free-market identity across the continent and, simultaneously, move toward independent, state-led financing to shield African agency from external dependency.
Gender, rurality and mentoring relationships in teaching practice contexts in Zimbabwean secondary schools.
(2022) Zinhumwe, Doreen Ruvimbo.; Moletsane, Relenbohile.
Mentoring relationships during teaching practice are a key medium in which student teachers develop effective professional learning and practice through guidance, supervision, critique and feedback from mentor teachers. While colleges and universities in Zimbabwe have been placing student teachers in a variety of geographic and cultural contexts, in recent years, there has been a decrease in the number of student teachers opting for teaching practice in rural areas. This mirrors the trend of qualified teachers shunning rural teaching placement. This is despite the government’s policy that upon qualification, student teachers have to serve in rural areas before they can permanently work in urban areas. The reasons for student teachers’ reluctance to be placed in rural schools include, among others, distance between the schools and urban centres, poor infrastructure, quality of mentoring and the choice of mentor in the schools. It is against this backdrop that the research examines the nature of mentoring relationships between student teachers and mentor teachers in rural schools, and how gender and rurality influence them. The study was conducted with 32 participants (18 student teachers who were on teaching practice and their 14 mentor teachers) in selected rural schools located in the Mashonaland East, West and Central provinces in Zimbabwe. The study was informed by two theoretical lenses, the ethic of care theory (Gilligan, 1982; Tronto, 1993) to explore the influence of gender on mentoring relationships and the generative theory of rurality (Balfour, Mitchell & Moletsane, 2008) to analyse the influence of rurality on relationships. The study was located within an interpretive paradigm, which suggests that human life experiences can be understood from the subjective construction of the meaning of the individuals involved and how they relate within their social contexts. Informed by this paradigm, the study adopted a qualitative approach, using such methods as focus group discussions, semi-structured and photo-elicitation interviews with student teachers and their mentor teachers to generate data. The findings suggest that three types of mentoring relationships existed in the rural secondary schools: cordial, dysfunctional and paradoxical caused by a combination of factors. Cordial relationships produced productive and mutual learning for the mentoring parties, whilst dysfunctional relationships were problematic and inhibited mutual learning, and paradoxical relationships were both productive and unproductive at different times. Whilst individual, school culture and socio-cultural factors accounted for these relationships, gender and rurality played a significant role in the formation and manifestation of these relationships. Same-gender relationships accounted for the majority of the cordial relationships whilst cross-gender mentoring partnerships were linked to dysfunctional and paradoxical relationships. The findings suggest that same-gender relationships opened more opportunities of interaction during and after working hours within and outside school premises, while cross-gender relationships were more restrictive. However, some trendsetters in cross-gender relationships had cordial relationships, but these were dependent upon caring attributes between the mentoring parties. Surprisingly, rurality, through its unique characteristics promoted agency and learning in the mentoring process and relationships. Due to communal lifestyles in rural communities, more opportunities for interaction and mutual understanding were available. The shortage of resources stimulated agency and innovation that saw the construction and accumulation of teaching and learning resources where there had been paucity. Both mentor teachers and student teachers reported a mindset shift regarding the benefits of rural placements in mentoring and professional nourishment. The findings suggest that student teachers should be exposed to a rural component of teaching practice during their training. In tandem, interventions that seek to develop skills and attributes that make mentoring relationships effective must be developed for student teachers in their teacher education curricula, and for school-based teachers through professional development programming.
Intention to adopt machine learning for credit risk assessment in a South African bank.
(2026) Thupayakgosi, Nicholus Olebile.; Price, Catherine Susan.
A bank functions as a credit institution engaging in transactions with individuals, businesses, and government entities; gathering funds for lending purposes; and engaging in supplementary investment activities. Its operations inherently depend on three key factors: the volume of outstanding debt, the probability of default which is a client not meeting their payment obligation, and the proportion of debt that may remain uncollected in case of default. The effective management of risks is critical to a bank's performance. Traditional credit risk assessment methods in South African banks face limitations in predictive accuracy and adaptability to complex borrower behaviours. While machine learning (ML) offers advanced analytical capabilities that could substantially enhance credit risk evaluation, its adoption in South African banking remains limited. Despite growing interest in ML technologies, little is known about the perceptions and readiness of technical teams responsible for implementing these systems in South Africa, where banking sector adoption remains limited. This study investigates the technological, organizational, and environmental factors influencing the intention to adopt ML for credit risk assessment in a South African bank, guided by an adapted Technology-Organization-Environment (TOE) framework. The results reveal that organizational factors emerged as the most critical barrier to ML adoption, surpassing both technological and environmental considerations. Key organizational barriers include limited top management support, absence of a clear strategic vision for ML integration, insufficient internal collaboration between departments, lack of skilled ML personnel, and absence of formal training programs. Technological challenges centred on data quality concerns, legacy system integration difficulties, infrastructure inadequacy, and the critical need for model explainability and transparency to satisfy regulatory requirements. Environmental factors include stringent regulatory compliance demands (POPIA and the National Credit Act), ethical considerations regarding algorithmic fairness, limited competitive pressure within the conservative banking sector, and concerns about customer trust and acceptance of automated credit decisions. The findings suggest that South African banks must prioritize organizational transformation including securing executive sponsorship, developing structured ML training programs, establishing cross-functional collaboration mechanisms, and building internal technical capacity before investing heavily in ML infrastructure. The research provides actionable insights for banking institutions seeking to implement ML responsibly and effectively in credit risk assessment.