Natural and Mathematical Sciences Departments

Permanent URI for this community

https://univendspace.univen.ac.za/handle/11602/1914

Browse

Now showing 1 - 20 of 651

Open Access
3D Cranial Morphometry, Sensory Ecology and Climate Change in African Rodents
(2019-05-15) Nengovhela, Aluwani; Taylor,, Peter; Braga, Jose
The order Rodentia is the most speciose group of mammals with muroids being the most diverse superfamily. Since they are represented in arboreal, semiaquatic, subterranean and terrestrial niches, rodents may exhibit morphological traits reflecting their adaptations to such diverse environments. This thesis focuses on the morphology of the endocranium, auditory bulla and cochlea in three tribes (Otomyini, Taterillini and Gerbillini) representing 10 species of African rodents, concentrating on their variability, function and adaptability, using micro-CT imaging and 3D shape comparative methods. Additionally, variations in cranial size were also studied in respective of global warming and climatic variables. Morphological changes/variations are a result of environmental change, therefore each chapter in this study details the effect of environmental change (in space and time) on different morphological traits i.e. general cranial size (chapter 2), cochlea and auditory bulla (chapter 3) and endocranial size and shape (chapter 4). With chapter 2 dealing specifically with climate change in its strict sense and the remaining two chapters looking at different environmental gradients. Chapter 2 tests the applicability of the “third universal response to warming” (i.e. declining body size) and the Resource Rule in two murid subfamilies, Murinae and Gerbillinae. The study shows that the third response is not as universal as only one species conformed to this response. Further, food availability (Resource Rule) was shown to be the more important factor correlated with body size variations in rodent species than Bergmann’s Rule. Chapter 3 looks at the auditory bulla and cochlea, the morphological traits that play a role in hearing capabilities of rodents. I found, with some exceptions, that bulla and cochlea modifications between species could be explained by environment, phylogeny and/or allometry. In addition, I concluded that true desert adapted laminate-toothed rats and gerbils use both bulla and associated cochlea hypertrophy. Chapter 4 shows larger brain size in Taterillini and two species of Otomyini, with life histories and environment being the most probably factors responsible for xiv this. Using a novel method of diffeomorphism (deformation models), there was more variation in endocranial morphology between the gerbils and laminate-toothed rats than within them with olfactory bulb, paraflocculi, and posterior ventral cortex showing the most variability. Overall, this thesis shows that variations in the morphological traits studied are strongly influenced by the environment and function.
Embargo
A Comparative Analysis of Machine Learning Models and Traditional Statistical Models for Continuous-Time Survival Analysis
(2026-05-19) Tshisikule, Ompha; Mulaudzi, T. B.; Bere, A.
Survival analysis is a statistical technique used to model time-to-event data, commonly applied in fields such as healthcare, engineering, and finance. Traditional approaches, including the Cox Proportional Hazards (CoxPH) model, have long been dominant due to their interpretability and theoretical foundation. However, recent advances in machine learning have shown promise in handling complex, high-dimensional datasets with nonlinear relationships. Despite this, there remains a gap in systematic comparative studies between traditional survival models and modern approaches such as regularized regression, ensemble methods, and deep learning architectures, particularly across diverse datasets with varying characteristics. This study conducts a comparative analysis of traditional, machine learning, and deep learning-based survival models, evaluating their predictive performance and computational efficiency for continuous-time survival data. The models considered include LASSO-regularized Cox regression, CoxPH, Random Survival Forest (RSF), and Long Short-Term Memory (LSTM) algorithms. Model performance was assessed using the concordance index (C-index), integrated Brier score (IBS), and Time-dependent Area Under the Curve (AUC) across three secondary datasets with different characteristics: a breast cancer dataset obtained from the SEER Program of the National Cancer Institute (2017 November update), the North Carolina Recidivism dataset (ICPSR 8987) obtained from ICPSR, and a heart failure clinical records dataset obtained from Kaggle. A rigorous statistical framework was employed, utilizing 100 iterations of stratified train-test splits to generate robust performance distributions. Distributional assumptions were systematically tested using Shapiro-Wilk and Levene’s tests to determine appropriate statistical tests, followed by omnibus tests (ANOVA, Welch’s ANOVA, or Kruskal-Wallis) and post-hoc pairwise comparisons with Bonferroni correction to control family-wise error rates. The analysis revealed that traditional survival models consistently outperformed deep learning-based approaches across all datasets. Random Survival Forest achieved the highest predictive accuracy, followed closely by CoxPH, with C-index values ranging from 0.66 to 0.73 and lower IBS scores indicating better calibration. In contrast, LSTM models performed poorly, often near random prediction levels (C-index 0.3–0.42), despite extensive optimization efforts including hyperparameter tuning, class balancing, and architectural modifications. Statistical testing confirmed that performance differences were highly significant across models and datasets (all p < 0.001), and post-hoc analyses demonstrated that RSF and CoxPH consistently outperformed LSTM for both discrimination and calibration metrics. These results suggest that traditional survival models remain the most reliable choice for moderate-sized datasets with censored observations and weak predictive signals, while LSTM networks are limited by dataset size, high censoring, and architectural mismatch with static survival data.
Embargo
A comparative evaluation of machine learning models for stock price prediction and uncertainity estimation
(2026-05-19) Nengovhela, Vhukhudo; Ravele, T.; Sigauke, C.; Ndogmo, J. C.
This study compares machine learning models for stock price prediction and uncertainty estimation using high-frequency one-minute stock data. The research looks at how different models perform across developed and emerging markets, which helps with model selection for practical financial forecasting. Four models were tested for point forecasting: Random Forest (RF), Gradient Boosting (GB), Multi-Layer Perceptron (MLP), and a hybrid stacking ensemble composed of multiple base learners. For uncertainty quantification, three interval prediction methods were used: Bootstrap Residuals, Quantile Regression Forests (QRF), and Conformalised Quantile Regression (CQR). The analysis used one-minute stock price data from Microsoft Corporation (MSFT) as a developed market example and Standard Bank Group (SBK.JO) as an emerging market example, covering the period from 3rd to 26th September 2025. The results show that GB performed best for point forecasts in both markets. For MSFT, GB had RMSE of 0.2875 and MAE of 0.1869, while for SBK.JO it achieved RMSE of 25.9248 and MAE of 14.3638. Statistical tests using the Diebold-Mariano and Giacomini-White frameworks confirmed that GB significantly outperformed the other models. For interval prediction, QRF gave sharper intervals in the relatively stable developed market, while CQR achieved better coverage in the more volatile emerging market. The Hybrid Stacking model showed some advantages in volatile conditions but didn’t consistently beat well-tuned individual models. These findings suggest that ensemble methods like GB are still very effective for financial forecasting, and that uncertainty quantification methods should be chosen based on market volatility. The study provides practical guidance for selecting forecasting methods depending on market conditions and data characteristics, which should help both researchers and practitioners working in financial risk management.
Open Access
A metagenomic snapshot of selected wastewater treatment plants in Vhembe Region, Limpopo, South Africa: Investigating the resistome
(2024-09-06) Jacobs, Damien; Potgieter, N.; Traore, A. N.
Background: Water is crucial for human life. Rural communities often rely on natural water sources which may become polluted by wastewater due to various activities such as domestic use and agriculture. Antibiotic resistance genes (ARGs) may be transferred from wastewater to the environment and pose a global challenge they affect both human and animal-related sectors. Studying antibiotic resistance in wastewater treatment plants within Vhembe offers a representation of antibiotic resistance genes from entire communities. Knowledge of antibiotic resistance circulating in Vhembe has been sparsely studied. Metagenomics approaches allow for a broad overview of the resistome and the bacterial communities within environmental samples. Aim: To perform wastewater surveillance of antibiotic resistance genes and associated bacteria within Vhembe, Limpopo, South Africa, using a metagenomics approach. Method: A total of 32 sample duplicates were collected from the influents (n=18) and the effluents (n = 14) from nine wastewater treatment plants (WWTPs) around the Vhembe region, Limpopo, South Africa. One hundred milliliter was filtered using sterile cotton gauze and Wattman filter paper to remove debris and membrane filtered through 0.22um membrane filters to capture the bacteria within each sample. DNA was extracted directly from the 0.22 μm filters using a DNA miniprep kit. DNA was quantified using a spectrophotometer. Shotgun 18 metagenomic sequencing was performed on DNA extracts. Open-source bioinformatics pipelines were used to process and analyze raw sequence data, uncovering information of the bacterial community composition and associated ARGs in wastewater. Results: Site observations reveal animal and human activities within and near the sites. ARG analysis revealed an overall number 0f 220 ARGs detected across the WWTPs. Thirty-six genes were common to influent samples and 16 in within effluent samples, encoding predominantly against macrolides, sulfonamides and tetracyclines, beta-lactamases, and aminoglycosides. Some unique ARGS were detected at sites near South African borders. Bacterial Diversity showed the predominance of some genera, such as Arcobacter, Aeromonas, Pseudomonas and Acinetobacter. Pathogens were predominantly enteric and pulmonary, with some being linked to animals in past studies. A notable increase in some members of Mycobactericeae, among other bacteria, was noted in effluents.
Embargo
A Multi-level Model for a Vector-Borne Organ to Tissue life Cycle Dynamics
(2025-09-05) Mahada, Awelani Sydney; Netshikweta, R.; Garira, W.
Introduction: Malaria is among the World’s most lethal infectious disease. It is caused by a parasitic pathogen transmitted by the Anopheles mosquito, which inoculates sporozoites into the human host during a blood meal. The population dynamics of malaria are well-known for their complexity, stemming not only from the parasite’s lifecycle, which involves two hosts (humans and mosquitoes)but also from the intricate replication and transmission cycles across different levels of the infectious disease system organization. Like other infectious disease systems, malaria infections inherit multilevel and multiscale systems, which pose significant challenges to efforts aimed at eliminating and ultimately eradicating the infection in a malaria-endemic population. Methodology Mathematical modeling in the study of complex system has proven to be an invaluable tool for understanding and predicting the behaviour and dynamics of a complex system within the domain of complexity science. Thus, in this study, we propose a multiscale modelling framework that captures the dynamics of malaria across three organizational levels within infectious disease systems implicated in the spread of malaria in a community. We begin by formulating a mathematical model to describe the development and progression of malaria parasites within the liver and tissue(blood) stages of an infected human host. This is followed by the formulation of a multiscale model that integrates both the inside(i.e.,the organ-tissue level)host and the outside (i.e., the host level) host malaria dynamics. Results Mathematical analysis for both the malaria models presented in this study was carried out and proved that all the models are mathematically and epidemiologically well-posed. We also compute the basic reproduction number R0 for both models and use the R0 to determine the local and global stability of the disease-free equilibriumas well as the local stability of endemic equilibrium of both models, respectively. We demonstrate that if R0 < 1, then the diseasefree equilibrium pointy of both models is locally and globally asymptotically stable, respevctively. However, if R0 > 1 the endemic equilibrium point of both models is locally asymptotically stable. The numerical results for both the models have demonstrated that the goal of intervention during malaria infection should be to reduce the rates at which merozoites and gametocytes invade healthy liver tissue as well as the blood cells. Hence it is recommended that interventions during malaria infection be directed on reducing the pace at which merozoites infect healthy blood cells and the density of merozoites in circulation. Conclusion The study presents a method that incoporates the complexity of malaria pathogens which is significant not only for malaria treatment but also for other vector-borne disease system control treatment strategies.
Open Access
AB INITIO and DFT computational study of Myristinin A and A structurally- related molecule
(2019-09-20) Tshilande, Neani; Mammino, Liliana; Ghio, Caterina
The computational study of biologically active molecules is particularly important for drug development because it provides crucial information about the properties of a molecule, which determine its biological activities. The current work considers the results of a computational study of myristinin A and a structurally-related molecule (2-(4-hydroxyphenyl)-4-[2,4,6-trihydroxy-3-(9tetradecenoyl)phenyl]-3,4-dihydro-2H-benzopyran-7-ol, here denoted as DBPO). The two compounds pertain to the class of acylphloroglucinols. They were firstly isolated from Horsfieldia amygdaline, and they exhibit a variety of biological activities, including potent anti-inflammatory activity, potent DNA-damaging activity and DNA-polymerase ß inhibition. Their molecular structures differ only by the acyl chain. Both molecules have a bulky substituent meta to the acyl group consisting of a ring system (2-(4-hydroxyphenyl)-3,4-dihydro-2H-chromen-7-ol). The DBPO molecule can exist as cis and trans isomers in relation to the double bond present in the R chain, and both isomers are here investigated individually. The OHs ortho to the acyl group can form an intramolecular hydrogen bond (referred to as the first IHB) with the sp2 O atom of the acyl group. The phenol OHs neighbouring the substituent ring system can form O–Hπ interaction with the aromatic rings of the substituent, if suitable oriented. This study focuses on the identification of the stable conformers of these molecules (considering all the possible geometries obtainable by rotations about relevant single bonds), and of the factors stabilising the conformers. Full-optimisation calculations were performed in vacuo and also in three conveniently selected solvents. The results show that the dominant stabilising factors are the first IHB and the O–Hπ interactions. Other factors which have significant influence on conformational preferences are the orientation of the ring systems of the substituent, the orientation of the OHs on substituent, the mutual orientation of the OHs of the phloroglucinol moiety and also the orientation of the acyl chain. The results in solution are consistent with the findings of other acylphloroglucinols, for instance, the narrowing of the energy gaps and the increase of the dipole moment with the increase of solvent polarity.
Embargo
Activation of the p53 pathway in combination with photon irradiation for the treatment of neurological tumour cells
(2026-05-19) Maluleka, Musa; Nemangwele, Fhulufhelo; Fisher, Randall; Engelbrecht-Roberts, Monique
Medulloblastoma (MB) and glioblastoma (GB) are highly aggressive brain tumours for which treatment outcomes remain poor, particularly due to intrinsic and acquired resistance to radiotherapy. Molecular determinants, especially TP53 status, play a critical role in regulating tumour cell proliferation, cell-cycle control, and DNA damage response following irradiation. This study investigated the biological effects of the MDM2 inhibitor AMG232 in combination with photon irradiation in TP53- wild-type and TP53-mutant MB and GB cell lines, with the aim of assessing whether AMG232 enhances radiosensitivity in a TP53-dependent manner. Cell proliferation, cell-cycle distribution, and DNA damage were assessed using growth assays, flow cytometry, and H2AX foci analysis, respectively. The findings showed that TP53 status strongly influenced cellular responses to treatment. TP53-wild-type cell lines demonstrated clearer growth control following AMG232 treatment, consistent with activation of functional p53 signalling. In contrast, TP53- mutant cell lines showed slower growth, inconsistent cell-cycle regulation, and weaker responses to AMG232, indicating limited recovery of p53 function. Cell-cycle analysis revealed that AMG232 induced a stronger and more sustained G0/G1 arrest in TP53-wild-type cells, supporting activation of the canonical p53–p21 axis. TP53-mutant cells displayed only partial or transient G0/G1 accumulation, suggesting the involvement of p53-independent stress responses rather than effective checkpoint enforcement. H2AX foci analysis confirmed a dose-dependent induction of DNA DSBs following photon irradiation across all cell lines. AMG232 treatment was associated with increased persistence of H2AX foci, particularly in MB cell lines, indicating impaired or delayed DNA repair. Residual foci at later time points reflected the predominance of error-prone non-homologous end joining, especially in G0/G1-arrested cells. In GB cell lines, DNA repair efficiency remained limited irrespective of treatment, highlighting intrinsic radioresistance. This study demonstrates that AMG232 enhances radiosensitivity primarily by prolonging DNA damage signalling and reducing DNA repair capacity, with effects that are most pronounced in TP53-wild-type cell lines. These findings highlight the importance of TP53 status in determining the therapeutic efficacy of MDM2 inhibition combined with photon irradiation and support the potential for molecularly guided treatment strategies in aggressive brain tumours.
Open Access
Activity patterns, species composition and dietary analysis of bats on two macadamia farms in Levubu, South Africa
(2017-09-18) Mphethe, Vusani; Taylor, P. J.; Steyn, J. N.; Schoeman, M. C.
See the attached abstract below
Open Access
Adsorption of oxyanions of As, B, Cr, Mo and Se from coal fly ash leachates using A1/Fe modified bentonite clay
(2014-01-10) Masindi, Vhahangwele; Gitari, Wilson Mugera; Tutu, Hlanganani
South African coal-fired power stations rely on the use of low grade bituminous coal for electricity generation. During coal combustion processes, this type of coal generates large volumes of waste materials which include Fly ash, bottom ash, boilers slag, flue gas desulphurization and un-captured particles. From the array of generated residues, coal fly ash constitute 90% and of this only 5% is beneficially used with the rest being disposed-off to land as ash dumps or slurried to ash retention ponds. On disposal, coal fly ash leaches out toxic chemical species on contact with the aqueous media hence posing hazardous effects to the aquatic and terrestrial environment. Of prime concern are Oxyanionic species such as As, B, Cr, Mo and Se. This study aims to investigate the adsorption of As, B, Cr, Mo and Se from coal fly ash leachates on Al3+/Fe3+modified bentonite clay. Coal fly ash samples used in this study were collected from Tutuka, ESKOM power plant in Mpumalanga and Bentonite clay samples were collected from ECCA (pty) Ltd. pH, Electrical Conductivity (EC) and Total Dissolved Solids (TDS) of the samples were determined using a CRISON multimeter probe. Modifications of bentonite clay with Al3+ and Fe3+ cations were done in batch procedures and the parameters optimized included contact time, adsorbent dose and adsorbate concentration. Cation Exchange Capacity (CEC), Point of Zero Charge (pHpzc) and Water Holding Capacity (WHC) of raw and modified bentonite clay were determined using standard methods. Surface areas of raw and modified bentonite clays were determined by BET. Elemental composition of raw and modified bentonite clay was determined by XRF and Mineralogical composition of raw and modified bentonite clay was determined by XRD. AAS and GFAAS were used to monitor AI3+and Fe3+concentration in the reaction mixture during and after modification while As, B, Cr, Mo, and Se were monitored using GFAAS, ICP- OES and JCP- MS. Adsorption of oxyanions of As, B, Cr, Mo and Se was carried out in batch experiments using synthetic solutions of the species. Effects of contact time, adsorbent dosage, adsorbate concentration and pH on the adsorption of As, B, Cr, Mo and Se were evaluated and optimized. The modification experiments revealed that loading of Al3+ onto bentonite clay interlayers is optimum at 60 minutes of contact time, 3 grams of adsorbent and 100 mg/L of adsorbate concentration and loading of Fe3+ onto bentonite clay matrices is optimum at 15 minutes of contact time, 2 grams of adsorbent and 100 mg/L of adsorbate. The CEC results showed that South African bentonite clay is characterized by high CEC of262 meq/l00g at pH 5.4 and 265.5 meq/l00g at pH 7.4 which decreased to 186.9 meq/l00g at pH 5.4 and 183.3 meq/l00g at pH 7.4 on introducing Al3+ onto bentonite clay interlayers and 195.5 meq/100g at pH 5.4 and 188.9 meq/100g at pH 7.4 on introducing Fe3+ onto bentonite clay interlayers. The results show that CEC is independent of pH. BET results indicated that the loading of Al3+ onto bentonite clay interlayers increased the surface area from 16 m2/g to 44.3 m2/g and 50 m2/g for Fe3+,external surface area from 11.1 m2/g to 33.1 m2/g for Al3+ and 37.4 m2/g for Fe3+,micro-pore area from 4.9 m2/g to 11.2 m2/g for AJ3+ and 12.5 m2/g for Fe3+. There was a slight difference in pHpzc of Al3+ modified bentonite clay (8.2) as compared to raw bentonite (8.0). For Fe3+ modified bentonite clay, the pHpzc was observed to increase from 8.0 for raw bentonite clay to 9.0. An increase in pHpzc shows that modification will favour adsorption of anions from aqueous media. WHC of bentonite clay was determined to be 42.7%. Elemental composition by XRF showed SiO2, AhO3, Fe2O3, CaO, Na2O, K2O and MgO as the major chemical species. XRF results also confirmed that bentonite clay is an aluminosilicate material. XRD disclosed that bentonite clay is dominated by montmorillonite as the major mineral phase. Characterization of fly ash leachates revealed that coal fly ash leachates contains 53 µg/L of As, 1730 µg/L of B, 271 µg/L of Cr, 165 µg/L of Mo and 170 µg/L of Se. Adsorption of As, B, Cr, Mo and Se onto Al3+ modified bentonite clay revealed that 30 minutes of contact time, 4 grams of adsorbent dosage, 10 mg/L of adsorbate concentration and pH 10 are the optimum conditions for removing those species from aqueous media. The adsorption affinity of those species obeyed the following order: B = Se > Mo = Cr > As. Adsorption of As, B, Cr, Mo and Se onto Fe3+ modified bentonite clay revealed that 30 minutes of contact time, 4 grams of adsorbent dosage, 10 mg/L of adsorbate concentration and pH 10 are the optimum conditions for removing those species from aqueous media. The adsorption affinity of those species obeyed the following order: B = Se > Mo > Cr = As. The adsorption capacities of modified bentonite clays were greater than the adsorption capacities of raw bentonite clay. The data fitted well to both Langmuir and Freundlich adsorption isotherm hence showing that the adsorption is energetically favourable. The unmodified clay showed poor adsorption for these chemical species. Both Al3+ and Fe3+ modified bentonite clay successfully removed As, B, Cr, Mo and Se from generated coal fly ash leachates to below DWAF water quality guidelines at the optimized conditions. This shows that AJ3+ or Fe3+ modified bentonite clay is an effective adsorbent for removal of oxyanionic species of As, B, Cr, Mo and Se from coal fly ash leachates and could be applied as a reactive barrier or liner in ash retention ponds.
Open Access
Alternative methods for solving nonlinear two-point boundary value problems
(2018-03-18) Ghomanjani, Fateme; Shateyi, Stanford
In this sequel, the numerical solution of nonlinear two-point boundary value problems (NTBVPs) for ordinary di erential equations (ODEs) is found by Bezier curve method (BCM) and orthonormal Bernstein polynomials (OBPs). OBPs will be constructed by Gram-Schmidt technique. Stated methods are more easier and applicable for linear and nonlinear problems. Some numerical examples are solved and they are stated the accurate findings.
Open Access
Alternative practices for optimising soil quality and crop protection for macadamia orchards, Limpopo Province, South Africa.
(2019-09-30) Steyn, Jakobus Nicolaas; Crafford, J. E.; Gliessman, S. R.; v. d. M. Louw, S.
The main aim of the research was to contribute means for converting conventional, high-input production systems to more sustainable ecological systems, thereby improving the sustainability of macadamia production and ultimately contributing to food security. This was achieved by a) investigating the potential use of cover crops and compost to enhance soil quality in macadamia orchards and b) investigating the potential use of use of cover crops and orchard heterogeneity to control stinkbug pests that target macadamia crops. Field experiments were conducted in three phases: phase one tested the potential of six cover crops for crop protection (as trap crops) and simultaneously for soil restoration or fertility enhancement purposes in macadamia orchards. Phase two repeated the trials of phase one (both soil restoration and trap crops) but with modifications to both categories. Soil restoration treatments were conducted with trees which were growing in what appeared to be healthy soils, and then repeated with trees in the same orchard where the topsoil had been degraded (totally removed) by agricultural operations. The third phase repeated the trap crop trials only, but this time on three different study areas (all commercial farms) with the single cover crop which performed the best as a trap crop during phase two. Trials were modified from the first to the last phase to overcome practical implementation problems encountered along the way and to adapt to local conditions experienced in the commercial macadamia farming systems which served as research sites. Diversity of natural orchard vegetation was enhanced in phase three to improve conditions for natural predators as part of the trap crop treatments in the last phase and cover crops were finally first composted and then returned to the root zones of the macadamia trees as part of the soil quality enhancement treatments in the second phase. The results from the trap crop trials shows a significant effect of trap crops combined with increased orchard diversity in reducing unsound kernel percentages caused by stinkbug pests and demonstrate that trap crops combined with an increase in orchard diversity could be utilized in macadamia orchards as a more sustainable alternative to inorganic pesticides against the stinkbug complex. The most notable changes in the soil that took place with soil quality enhancement treatments were the significant increases in soil phosphorous content and pH which resulted not in an improvement in soil quality in terms of these two indicators but revealed an important issue about the use of compost containing animal manure originating from dairies or feedlots. In summary however, it was clear that although not all the soil quality indicators that were employed to assess changes in the soil with compost treatments improved significantly, a holistic consideration of all indicators portrays an overall improvement which was particularly significant in the degraded soil plots where the topsoil had been removed by prior agricultural activities.
Embargo
Ambient air pollution monitoring, modelling and human health risks in Thulamela Municipality, Vhembe District of Limpopo Province, South Africa
(2025-09-05) Enitan, Ibironke Titilayo; Edokpayi, J. N.; Piketh, S. J.
Ambient air pollution is a pressing global environmental concern, with pollutants from natural and anthropogenic sources contributing to deteriorating air quality. This research investigates the impact of sulfur dioxide (SO₂), nitrogen oxides (NOₓ), carbon monoxide (CO), and particulate matter (PM2.5 and PM10) on human health and the environment. PM2.5 in particular is a significant public health risk linked to respiratory and cardiovascular diseases. The study focuses on Thulamela Local Municipality (LM) in Limpopo Province, South Africa, a region with diverse semi-urban and rural environments, where air quality data is limited despite ongoing urbanisation and economic growth. Given the lack of comprehensive baseline data, this study aims to quantify, monitor, and model air pollution while assessing its potential health impacts. This research integrates the development of an emission inventory for criteria pollutants (SO₂, NOₓ, CO, PM2.5, and PM₁₀), real-time PM2.5 monitoring using Dylos DC1700 low-cost sensors (LCSs), PM2.5 dispersion modelling with AERMOD, health risk assessment (HHRA), and machine learning-based predictive modelling. By evaluating pollution sources, exposure risks, and predictive trends, the study provides critical insights to inform air quality management strategies in semi-urban settings. This study developed the first-ever 10-year vehicle emission inventory (VEI) for criteria pollutants (SO₂, NOₓ, CO, PM2.5, and PM₁₀) in Thulamela Local Municipality (LM), South Africa. In addition, it produced the first comprehensive bottom-up emission inventory for the year 2021, quantifying total SO₂, NOₓ, CO, PM2.5, and PM₁₀ emissions from two major anthropogenic sources: vehicular emissions and residential wood combustion (RWC). The results indicated annual emissions of 66.3, 1821.7, 21002.6, 2348.7, and 2649.3 tonnes, for SO₂, NOₓ, CO, PM2.5, and PM₁₀, respectively. RWC contributed over 96% of PM2.5 and PM₁₀ emissions, while NOₓ (84%) and SO₂ (60%) originated primarily from vehicular sources. The spatial analysis revealed that emissions were highest in the southwest and northwest, where population density and human activities were most concentrated. A one-year PM2.5 monitoring campaign (March 2021 – February 2022) was conducted at 14 sites, covering urban, suburban, and rural areas. The strong correlation (R² = 0.75) between Dylos and reference monitors (GRIMM EDM180) confirmed the reliability of LCS performance for PM2.5 measurements. The study found that annual PM2.5 concentrations exceeded both WHO guidelines and SA NAAQS, except at one site. Monthly trends showed higher PM2.5 levels from May to September, with diurnal peaks from 6–8 am and 6–8 pm, primarily influenced by traffic emissions and residential wood combustion. The bivariate polar plot (BVP) analysis identified pollution hotspots near traffic intersections, residential areas, and commercial zones. The study identified significant spatial and temporal variations in PM2.5 levels across different monitoring sites, with higher concentrations in suburban areas than rural ones due to vehicular emissions and wood-burning. VTV2 recorded the highest annual concentration (32.3 μg/m³), while RL4 had the lowest (19.1 μg/m³). The average seasonal PM2.5 levels, despite differences, consistently exceeded air quality guidelines, except in summer for specific locations. The results also showed that evening PM2.5 concentrations were generally higher than morning peaks, highlighting significant differences in pollution levels across the area. Using AERMOD, PM2.5 dispersion was modelled to assess spatial patterns and regulatory compliance. Results showed that pollution hotspots exceeded SA NAAQS (40 μg/m³ daily, 20 μg/m³ annually) and WHO standards (15 μg/m³ daily, 5 μg/m³ annually), with the highest daily PM2.5 concentration (49.2 μg/m³) in densely populated areas. Model validation comparing modelled and monitored PM2.5 showed systematic underestimation (R² = 0.176, negative fractional bias), indicating unaccounted sources such as biomass burning, unpaved roads, industrial activities, and agricultural emissions. At Lunungwi village (RL receptor group), the modelled hourly PM2.5 concentration (47.12 μg/m³) exceeded the observed value (32.27 μg/m³, FAC2 = 0.25), indicating localised overestimation. Daily peaks aligned with morning and evening rush hours. At the same time, monthly trends from May to September reflected seasonal variations and increased residential wood combustion, while a notable decline in July 2021 was due to COVID-19 lockdown restrictions, highlighting the impact of reduced human activity on air quality. Although AERMOD has limitations in accounting for some emission sources, it effectively identified pollution hotspots, offering valuable insights for policy interventions and air quality management. The study further assessed long-term exposure and health risks using the HHRA model, focusing on risk characterisation. The USEPA health risk assessment revealed that PM2.5 exposure posed both non-carcinogenic and carcinogenic risks, with infants and toddlers being the most vulnerable. Hazard Quotient (HQ) values exceeded safe limits (1.10 in winter) due to high PM2.5 levels in colder months. Excess Lifetime Cancer Risk (ELCR) values surpassed the acceptable threshold (10⁻⁴) for all age groups, with adults facing the highest long-term cancer risks (2.35 × 10⁻¹). These findings emphasise the urgent need for air pollution control measures, particularly in areas with high residential fuel combustion and vehicular emissions. To enhance PM2.5 predictions, machine learning models were applied using meteorological variables. Random Forest (RF) outperformed Multiple Linear Regression (MLR) and Polynomial Ridge Regression (PRR), with the highest R² (0.80) at VP4 for site-specific predictions and 0.604 for area-wide estimates. Seasonal performance showed better RF accuracy in winter (R² = 0.58, RMSE = 2.10 μg/m³) than in summer (R² = 0.45, RMSE = 4.72 μg/m³), reflecting the impact of seasonal meteorology on PM2.5 dispersion. In conclusion, this research provides a holistic approach to quantifying air pollution sources, exposure risks, and predictive modelling in Thulamela Municipality. The study highlights urgent public health risks and the need for stronger air quality regulations, improved monitoring networks, and targeted mitigation strategies by establishing a clear link between air quality and human health. The findings underscore the need to expand emission inventories, refine meteorological inputs, and integrate multi-pollutant assessments to develop a comprehensive air quality management strategy. To address identified challenges, the study proposes mitigation strategies, including eco-friendly phytoremediation techniques, promotion of cleaner cooking and transport technologies, and urban planning measures to minimise exposure in densely populated areas. By combining scientific research with practical applications, this work aims to support sustainable air quality management, improve public health, and contribute to a healthier environment.
Open Access
An analysis of exposure and slope gradient impact on tree layer profile of Maroelakop vegetation in Nylsvley Nature Reserve, Limpopo Province, South Africa
(2024-09-06) Mavhila, Tendani; Tshisikhawe, M. P.; Magwede, K.
Since the 1850s, plant ecologists have been eager to investigate challenges in vegetation communities that are related to slope. The aspect of the slope is crucial in terms of the diversity and distribution of plants. This research was conducted on Maroelakop Mountain within Nylsvley Nature Reserve located in Mookgopong, Limpopo Province. The investigation aimed to assess and contrast the vegetation composition on the north-facing slope with that of the south-facing slope. The specific goals included ascertaining the density and frequency of dominant tree species on each slope, calculating the importance value indices of trees on both slope gradients and analyzing the population structure of the slope gradient. Point Centred Quarter (PCQ) was used in data collection. Shannon-Weiner diversity index, Simpson diversity index, and Evenness index were used to measure plant species diversity. In this study, a total of 600 trees were discovered and identified by their botanical names. The data represented 21 plant species in 17 genera and 12 families. The tree layer of the north-facing slope was dominated by Diplorhynchus condylocarpon, whereas that on the south-facing slope was dominated by Burkea africana. Both tree species had high-importance value indices. Shannon diversity index of 2.20 was obtained for the North-facing downslope, whereas the South-facing downslope had 1.74. The average Simpson diversity index for the North-facing slope was 0.82 and South-facing slope 0.77. The average Evenness index for the North-facing slope was 0.93 and, on the South-facing slope 0.81. This study concluded that Diplorhynchus condylocarpon is the most prevalent species within Maroelakop Mountain in the Nylsvley Nature Reserve.
Open Access
An evaluation of flood control mechanisms to withstand and adapt to flooding in Vhembe District, Limpopo Province
(2024-09-06) Munzhedzi, Lily; Nethengwe, N. S.; Mugari, M. E.
The Vhembe district is currently grappling with a rising prevalence of flooding, a phenomenon attributed to a combination of natural elements, such as heavy rainfall, and human-induced activities like deforestation and inadequate infrastructure management. The resultant damages encompass infrastructure destruction, loss of life, crop and soil damage, and environmental harm, with implications extending to human health, housing, livelihoods, government resources, and significant economic consequences. Consequences vary based on the resources and development levels of different areas, emphasising the need to understand each community's unique aspects for the development of resilient flood control strategies. The aim of this study is to evaluate the flood control mechanisms to determine if they can withstand and adapt to flooding in the selected study areas of Musina, Duthuni, and Sane in the Vhembe District. The study pursues two objectives: investigating trends in extreme rainfall and its relationship to flood occurrences, and characterising the nature, effectiveness, and adaptability of flood control mechanisms in the Vhembe district area. Employing a mixed-methods approach, the research design integrates both quantitative and qualitative methods to gain comprehensive insights. The data collection process involved the administration of in-depth questionnaires to participants, coupled with observational analyses of flood control mechanisms. This multifaceted approach allowed for a qualitative assessment of the nature, effectiveness, and adaptability of these mechanisms. Additionally, daily rainfall data sourced from the NASA POWER website was utilised. This data facilitated the computation of an average seasonal precipitation table, the generation of monthly precipitation figures, and tables displaying the count of days when precipitation exceeded the 95th percentile. The findings reveal a recurring pattern of precipitation throughout the seasons, with the highest average precipitation consistently occurring from December to February, and January emerging as a critical month for flood risk. The analysis of flood events spanning from 2000 to 2020 establishes a connection between daily rainfall exceeding 20 mm and instances of flooding, particularly evident in Duthuni, where nearly annual flooding has been observed. The study identified diverse flood control mechanisms linked to the level of development and community traditions. Natural and traditional measures, such as furrows and vegetation-based flood control mechanisms, dominate the rural village of Sane. In contrast, the urbanised Musina prefers engineered solutions like culverts and concrete channels, while Duthuni integrates both approaches. Furrows are popular at the household level across all the study areas, although their effectiveness is considered to be lower in comparison to other household flood control mechanisms. The field observational analysis emphasises the construction of flood control mechanisms using reinforced materials and underscores the importance of regular maintenance. The study underscores the importance of region-specific flood control solutions tailored to local conditions and preferences, considering climate patterns and community traditions. By adopting this approach, the Vhembe district can enhance its readiness and adaptive capacity for flooding. The study's results provide a foundation for well-informed decision-making and policy development, emphasising the necessity for initiatives to align flood management with the distinct challenges and preferences of local communities.
Open Access
An Intelligent Surveillance System Using Deep Facial Expression Recognition
(2026-05-19) Mutshafa, Livhuwani; Moyo, B.
Surveillance systems are critical tools for maintaining security, enhancing public safety, and safeguarding assets in diverse settings, from public spaces to private facilities. Despite their importance, these systems often face challenges that require human oversight. Recent studies have explored deep learning techniques to address such challenges, primarily focusing on face recognition and anomaly detection in static images. This study proposes a deep learning approach for detecting and interpreting facial expressions in dynamic images to enhance surveillance applications. The methodology involved a comprehensive literature review, dataset preprocessing, development of deep learning models, and rigorous model evaluation. A fine-tuned MobileNetV2 and a hybrid MobileNetV2–LSTM models were designed to capture both spatial and temporal features of facial expressions. The models were trained on benchmark datasets, including the Amsterdam Dynamic Facial Expression Set (ADFES) and the Chinese Face Dataset with Dynamic Expressions, and evaluated using accuracy, precision, recall, and F1-score metrics. Results demonstrated that the MobileNetV2–LSTM model significantly outperformed the standard MobileNetV2, achieving 95% accuracy, 95% precision, 95% recall, and 95% F1-score, highlighting the advantages of temporal modeling. The models maintained high computational efficiency, achieving 43.09 frames per second and a per-frame inference time of 0.0232 seconds, indicating strong real-time feasibility. This study contributes to intelligent surveillance by providing a highly reliable facial expression recognition framework for dynamic scenarios, with future work focusing on real-time deployment, expanded datasets with diverse ethnicities, and enhanced robustness under challenging surveillance conditions.
Open Access
Analysis of a boundary value problem for a system on non-homogeneous ordinary differential equations (ODE), with variable coefficients
(2015-01-16) Makhabane, Paul Suunyboy; Hlomuka, V. J.; Garira, W.
In this study we present a condition for the existence and uniqueness of the solution y(x) for a system of nonhomogeneous linear first order Ordinary Differential Equations (ODE). The existence and uniqueness of the solution of y(x) was confirmed through the Picard Lindelof Theorem. We then study the stability of matrix A(x) using its spectrum, moreover, A(x) is symmetric. This is a pre-condition for the application of Lefschetz direct stability method. We then modify the given Lefschetz system (Meyer, 1964) to suit the problem at hand. The direct method requires the construction of a suitable Lyapunov function; not easy for a time-independent (non-dynamic) problem. For a time-dependent problem the energy thereof becomes a suitable candidate for a Lyapunov function. For a non-dynamic problem it is harder to construct a Lyapunov function as there are no rules for that purpose. In our study we modified the Lefschetz system for the direct stability method and applied it to confirm the Lefschetz stability criterion using the modified systems of linear first order ODEs with variable coefficients. The Lefschetz method afforded us the construction of a credible Lyapunov function which enabled us to confirm the stability of the null solution to our problem. From our modified Lefschetz direct stability system, we solved the Makhabane / Hlomuka equation (5) for B(x) (7) which we later confirmed as both symmetric and positive definite.
Open Access
Analysis of antibiotics and PFAS as potential emerging pollutants in the water bodies and their accumulation in food samples
(2021-04) Mpela, Ntshani Gershom; Tavengwa, N. T.
Antibiotics and perfluoroalkyl substances (PFAS) are groups of anthropogenic substances that are found in industries and consumer products, i.e., antibiotics are used in healthcare facilities, pharmaceutical companies, and agriculture, whereas PFAS are used in industries and households. These compounds find their way into the environment through emission, landfill disposal, agricultural run-offs, irrigations, leaching and wastewater spillage. Their residues are widely detected in almost every environmental compartment. Their detection in drinking water and food is one of the global concerns. Their effect in living organisms have been widely documented ranging from allergic reaction to carcinogenic effect, and in serious situations they can lead to death. In this project, the background review on the behaviour of antibiotics and PFAS in different environments are discussed in chapter two. This has been done by looking into their physico-chemical properties, distribution pathways, accumulation and toxicity in various environmental compartments through available literature. After understanding their behaviour in the environment two review papers, namely Paper I and Paper II, were written. Paper I was a book chapter which focused on the application of miniaturised liquid phase extraction techniques used for extraction of antibiotics in environmental water samples. In this book chapter, the use of liquid-liquid extraction techniques was discussed and the evolution of liquid-liquid extraction technique into miniaturised extraction technique was reviewed. Paper II was a critical review which focuses on the application of dispersive liquid-liquid micro-extraction of PFAS in different environmental samples. In this paper, the principles of dispersive liquid-liquid micro-extraction was explained. Later, its application for pre-concentration of PFAS was reviewed and the future trends were discussed. In chapter four, three lab-based papers, namely Paper III, Paper IV and Paper V, were written. In Paper III, for the first time a dispersive liquid-liquid microextraction technique was developed and applied for extraction of multi-class antibiotics in macadamia nuts prior to UHPLC-qToF-MS analysis. In paper IV, a novel deep eutectic solvent was synthesised and characterised using FTIR spectroscopy. For the first time the synthesised deep eutectic solvent was applied for extraction of multiclass antibiotics in green beans prior to UHPLC-qToF-MS analysis. In Paper V, for the first a dispersive liquid-liquid micro-extraction technique was developed and applied for extraction of PFAS in food contact materials prior to UHPLC-qToF-MS analysis. Owing to their ease of operation, efficiency and sensitivity, these techniques can also be applied for preconcentration of antibiotics and PFAS in different food, environmental and biological samples.
Open Access
Analysis of drought incidence, gendered vulnerability and adaptation in Chivi South, Zimbabwe
(2016) Chineka, Jestina; Musyoki, A.; Kori, E.; Chikore, M. R. H.
Climate change has brought about a number of global environmental challenges and the worst, climatic disasters such as floods and droughts. In Zimbabwe, particularly in the semi-arid Chivi District, droughts have become more frequent. At the household level, there are a multiplicity of vulnerability and coping mechanisms to this scourge, which have gender dimensions. This study analysed the vulnerability of the Chivi South community to drought and its adaptation with specific consideration to the gender dynamics. The specific objectives for this dissertation were to establish the characteristics and extent of drought occurrences in Chivi District in the last 30 years, assess Chivi South‟s vulnerability to drought, evaluate levels of gender vulnerability to drought, and analyse gendered adaptation to drought. Subsequently, a strategy for drought adaptation in the rural areas of Zimbabwe was drawn. The research was based on the mixed methods approach, as it employed both qualitative and quantitative approaches. Data collection methods included a review of official documents, key informant interviews with community leaders, Non-Governmental Organisations and government officials working in the area, household questionnaires and focus group discussions. The Standardized Precipitation Index was used to determine drought severity. Data was analysed using the SPSS 22.0 software. Capabilities such as Chi-square and cross tabulation were used to effectively analyse data. The Household Vulnerability Index was employed to infer vulnerability and adaptation of the community to drought. Research findings were illustrated using charts, graphs, tables and photographs. The study established the occurrence of droughts in Chivi, with a high prevalence of low magnitude droughts. A decrease in crop food production closely related to drought patterns was noted. Drought vulnerability cuts across the whole gender spectrum, with a sizable number of female headed households being severely exposed. However females adapt better to the effects of drought than males. A holistic approach which seeks to integrate both men and women in decision-making and to improve the community‟s adaptation to drought and other disasters was proposed
Embargo
Analysis of gully erosion as a hydro-geomorphological hazard in Thulamela Municipality, Limpopo Province, South Africa
(2025-09-05) Makuya, Asakundwi; Kori. E.; Maponya, M. G.
Gully erosion, a severe form of environmental process, poses significant global environmental and socio-economic challenges, worsened by deforestation, overgrazing and climate change. This study examines gully erosion as a hydro-geomorphological hazard in Thulamela Municipality, Limpopo Province, South Africa. The aim of this research was to investigate the occurrence and intensity of gully erosion in Thulamela Municipality. The study's first objective was to map gully distribution in Thulamela Municipality. The second objective was to map the severity of gullying. The third and last objective was to map factors contributing to gully erosion formation. This study investigated the three objectives across four temporal benchmarks (2008, 2013, 2018 and 2023) using GIS and Remote Sensing techniques. Distribution in gully occurrence was tracked through interpolation. Sampled gullies were overlaid to assess their size and shape transformation over time to determine the severity of gully erosion. Land use/cover changes were classified to track how they influenced gully development over the 15-year period. Topography was analysed based on elevation, terrain characteristics, flow accumulation and flow direction to determine areas that are more prone to gully development. Rainfall was assessed based on its intensity and distribution over the period of the study to determine how it contributes to gully erosion. Soil moisture was analysed to determine susceptibility to gully erosion. The results revealed a significant and exponential increase in gully development influenced by all the analysed factors. The findings highlight that Thulamela Municipality contains areas highly susceptible to gully erosion, presenting severe threats to land stability, safe habitation and sustainable development.
Open Access
Analysis of land use and land cover change and its impact on soil erosion in Nzhelele Valley, Limpopo Province, South Africa
(2022-07-15) Mavhuru, Blessing; Nethengwe, N. S.; Odhiambo, B. D.O.; Chikoore, H.
Globally, the rate of land use and land cover changes has affeceted the magnitude of soil eroion. Strategies to combat soil erosion can give assistance as solutions to provide food security in many agricultural areas. Although some changes are caused by natural factors, anthropogenic factors and an increase in population are major drivers of soil erosion. This study analyses land use and land cover change and its impact on soil erosion in Nzhelele Valley. To attain this goal, the objectives of this study are to classify land use and land cover change using GIS from 2005 to 2019 in Nzhelele Valley; assess the human influence of soil management strategies on soil erosion; determine the impact of soil fertility of different land use on soil erosion, and model soil erosion on different land-use areas of Nzhelele Valley. To select farmers, study used a stratified random sampling technique by dividing groups based on their villages. A systematic sampling technique was used to select 392 farmers. To collect 78 soil samples from Nzhelele Valley, the study used the Sampling Design Tool of ArcGIS 10.The study used supervised classification in Geographic Information System (GIS) to classify land use and land cover types. Semi-structured questionnaires were used to solicit data on human influence on soil management strategies on soil erosion. The study draws a systematic sample using Slovin’s formula to determine the size of the sample from the Nzhelele Valley, and with a target population size of 657 farmers from the village, 398 farmers were selected for the study. The study also measured soil organic matter, soil pH, phosphorus, and nitrogen from collected soil samples to determine the impact of soil fertility on soil erosion. To collect soil samples from Nzhelele Valley, the study used the Sampling Design Tool of ArcGIS 10. The tool assisted in selecting the soil sample points within the study area and a total of 78 soil samples were collected for the study. The Soil and Water Assessment Tool (ARCSWAT) with an interface of ArcView Geographic Information System software modelled soil erosion within Nzhelele Valley. In terms of LULCC, results show that from 2005 up to 2019, forest increased significantly by 5%, agricultural land cover significant decrease by 11% from 36% to 25% land use/cover, bare land built-up land increase by 3% from 14% to 17%, as well as grassland increase by 3% from 19% to 22%. The results also show a significant correlation (p < 0.05) between land use management strategies and economic support systems. The physical characteristics (slope and terrace) significantly correlate (p< 0.05) with soil management strategies that limit soil erosion. The results illustrate significant variation (p ≤ 0.0002) in nitrogen among different land-use classes. Significant variation was observed (p ≤ 0.0001) in soil pH, phosphorus, and organic matter among the different land-use areas. Changes in LULC are more likely to have an effect on soil erosion in the grassland and bare land/built-up areas where infiltration is very limited. The recorded decrease in agricultural land use and an increase in forest cover are likely to reduce the impact of soil erosion. Using SWAT to model soil erosion, 26-sub basin and 301 hydrological response units were delineated. The statistical elevation of the watershed obtained a minimum elevation of 384 m, maximum elevation of 1680 m with a minimum elevation of 808.81 m. Sub basins 13, 23, and 25 demonstrated high erosion-prone classes with average sediment yield of 15.3%, 11.26%, and 11.5% respectively. The lowest sediment yield (2.1 t/ha-1yr-1) in the study area was observed in sub-basin 3. Overall synthesis of the findings illustrates that research into land-use changes and type of land use activities factors were key aspects in addressing soil erosion challenges. These findings vibrate strongly on the need to develop a systematic land management system that can solve unplanned land cover and land-use change. Based on the results, fostering better management investments through collaborative land management in Nzhelele Valley might improve the dissemination of information. Farmer to farmer training can improve the skills of the individual, which might allow smooth dissemination of information vital for land management. The problem of soil nutrient deficiency can be a central issue put forward to the local government, which needs to be addressed with specific sectors of the government to create a better environment for agriculture and food security especially in the marginalized community of South Africa. Sediment output based on the land use classes is potentially influenced by cropland with the study. This allows for further solutions on the type of land use activities to be narrowed to understand the levels of soil erosion based on the activities. The study findings from this area can be used for comparison to other areas with the same characteristics to gain insight into possible solutions to the challenges of soil erosion.