Evaluation of Ambient NO 2 and SO 2 Concentrations at Lephalale, Polokwane, and Steelpoort Area in the Limpopo Province of South Africa

Sulfur dioxide (SO ₂) and nitrogen dioxide (NO ₂) are some of the air pollutants in industrial and urban areas. These are the most prevalent inorganic pollutants that are a serious risk to public health in populated areas. However, Lephalale, Polokwane and Steelpoort have not received enough attention to date. Therefore, this study investigated seasonal levels of NO ₂ and SO ₂ concentrations using passive sampling during summer, winter, autumn and spring of 2021. The sampling was done the first seven days of the month. The relationship of inorganic pollutants with meteorological factors was performed statistically using multiple linear regression model. Lephalale exhibited the highest NO ₂ concentration, 1.74 µg/m ³, in spring. Whereas Polokwane and Steelpoort peaked at 1.57 µg/m ³ and 0.84 µg/m ³ in winter, respectively. The concentrations of SO ₂ were higher in winter than in other seasons in all areas. The multiple linear regression models showed that NO ₂ and SO ₂ dispersion was influenced by meteorological parameters such as temperature, wind speed and relative humidity. In Polokwane and Steelpoort, NO ₂ and SO ₂ concentrations are inversely correlated to temperature and relative humidity. Similarly, NO ₂ concentrations are inversely correlated, while SO ₂ concentrations are directly correlated to both temperature and relative humidity in Lephalale. Wind speed had positive and inverse correlations to the concentrations of both air pollutants. The SO ₂ and NO ₂ concentrations did not exceed the annual average of 50 µg/m ³ and 94 µg/m ³ as set by National Ambient Air Quality Standards and World Health Organization. However, it is important to keep pollution concentrations of SO ₂ and NO ₂ to relatively safe for humans and the environment in the studied areas.

• Seasonal Monitoring
• SO ₂
• NO ₂
• Meteorological Parameters

1. Introduction

One environmental issue that affects people worldwide, in both developed and developing nations, is air pollution. It is linked to emissions from regional and local sources as well as atmospheric reactions with gaseous pollutants [1] [2] . Natural and anthropogenic sources contribute to the air pollutants in the ambient environment [3] . Burning biomass and engaging in industrial and automotive activity are examples of anthropogenic activity emissions [4] . About 40% of people use biomass fuels for cooking and heating, such as wood, charcoal, crop residues, and animal dung. These fuels release a complex mixture of pollutants, including gases and fine particles, that are harmful to human health [5] [6] .

The World Health Organization (WHO) states that air pollutants like nitrogen dioxide (NO₂) and sulphur dioxide (SO₂) pose a major risk to human health. Even though SO₂ and NO₂ are known to be toxic, they also play a role in particle formation because of intricate atmospheric photochemical reactions that involve ammonia from agricultural practices [7] [8] . Cardiovascular diseases have been reported as the prime cause of high mortality rates due to long-term exposure to NO₂. There are various detrimental diseases such as pulmonary edema, suffocation and laryngeal edema, that result from SO₂ gas inhalation. It has been reported that the changes related to NO₂ have led to hospital admissions, upper and lower respiratory illness, bronchitis and chronic cough [9] [10] .

Many people in South Africa, a developing nation plagued by poverty and inequality, heat their homes with biomass fuels. In addition, the energy sector in South Africa produces a lot of air pollution, which exacerbates local environmental problems and public health issues as well as global climate change. Hence, South Africa has air pollution legislation with international comparable pollution limits [11] [12] . The National Ambient Air Quality Standards (NAAQS) cover priority air pollutants such as SO₂ and NO₂ [13] . The annual ambient air quality guidelines and standards for SO₂ and NO₂ are 50 µg/m³ and 94 µg/m³, respectively [14] . This legislation primary objective is to reduce the pollution levels to concentrations relatively safe for humans and the environment. However, this aim is currently not being reached, resulting in a serious health problem due to the lack of capacity to implement the existing legislation by the government and provincial authorities. Currently, all municipalities are required to monitor air pollution. However, this route has negative implications since it is not cost-efficient and does not provide high-quality data with correct interpretations [13] [15] .

There are air quality monitoring networks all over South Africa, according to the 2018 Department of Forestry, Fisheries and Environmental (DFFE) state of air report (2005-2016). The report presented data from networks that met minimum requirements of 50 µg/m³ SO₂ concentrations, which included Western Cape, Gauteng, Limpopo and KwaZulu Natal [16] [17] . The Air Quality Act of 2004, Sections 15 and 16, prompted Limpopo Department of Economic Development, Environment and Tourism (LEDET) to launch an Air Quality Management Planning process in 2013. The primary goal was to supply the province with an implementable Air Quality Management Plan (AQMP) that complies with national standards and enhances the AQMPs already in place for District Municipalities [18] .

In many parts of the Province of Limpopo, the environment, human health, and life expectancy are seriously threatened by the poor quality of the air. The province is experiencing air quality issues due to high levels of particulate matter, volatile organic compounds, SO₂, and NO₂. These elevated concentrations are caused by a variety of sources, including mining activities, power generation, metallurgical activities, burning biomass, vehicle tailpipe emissions, and burning domestic fuel [19] . The province of Limpopo mines coal, iron, platinum, chromium, manganese, copper, gold, diamonds, lime, and asbestos, among other minerals and resources. In the mining sector, drilling, blasting, hauling, collection and transportation are done. When some of these functions are performed, they have significant environmental impact and public health effects. This is due to constant exposure to air pollutants such as SO₂ and NO₂ [20] [21] .

A network of monitoring stations is used to measure ambient concentrations throughout South Africa. The industrial, rural, and urban areas in and around Limpopo are home to the monitoring stations. Under the direction of the DFFE and the LEDET, there are more than 21 government-owned air quality monitoring stations in the province. Only a few uses continuous samplers, with most using passive samplers [22] [23] . There are several hot-spot areas in Limpopo, viz. Polokwane, Lephalale, Phalaborwa and Steelpoort. However, no air quality monitoring has been conducted on a small scale in Steelpoort, Lephalale and Polokwane to investigate if the mining industries, smelters and quarries are associated with the poor air quality to the surrounding environment. This study is part of the knowledge-building process by investigating compliance of NO₂ and SO₂ to the NAAQS in various areas in Limpopo within the residential area situated in the mining industries, urban and coal-powered industry. Also, to investigate the effects of meteorological parameters on NO₂ and SO₂ concentrations. The study focused on two pollutants namely SO₂ and NO₂ due to their environmental impact and public health effects. Therefore, this study was aimed to determine seasonal variations of NO₂ and SO₂ in Limpopo rural, urban and industrial areas using passive sampling from January 2021 to December 2021.

2. Materials and Methods

Sampling sites, shown in Figure 1 , were chosen because of mining activities,

Figure 1

Figure 1. Study areas showing (red dots) Lephalale, Polokwane and Steelpoort.

vehicle-dense areas and where the coal-powered station is located. Samples were collected at Lephalale (23.6665˚ South, 27.7448˚ East), Polokwane (23.8962˚ South, 29.4486˚ East) and Steelpoort (24.43˚48" South; 30.11˚59" East). Polokwane, formerly known as Pietersburg, is a city and capital of the Limpopo Province. It covers an area of about 106.8 km² with an altitude of 1310 m and a population of more than 130,000 people. Over the past ten years, the city has seen seasons that are noticeably warmer than average. It also has a dry climate, with July being the driest month and December or January (less frequently) being the wettest with annual rainfall of about 495 mm. Anglo American Smelter is situated about 19 km south of Polokwane city. Lephalale is a town found in the western part of the Limpopo province of South Africa and covers about 66.94 km² with has a population of about 19,000 people. This town experiences long, hot, and partly cloudy summers. In contrast, there are the short, chilly, clear, and dry winters. Throughout the year, the temperature typically varies from 7˚C to 32˚C. The town hosts two power stations, namely Matimba and Medupi, which use coal to produce electricity. Steelpoort is a mining area in Sekhukhune District Municipality in the Limpopo province. The elevation above sea level varies from 1500 to 2400 meters. The average yearly rainfall fluctuates between 630 and 1000 mm, mostly as summer thunderstorms. The settlement has an estimated population of approximately 1,105, 380 (122.09 per km²) households and covers 3.11 km² (Census 2011). Additionally, five villages, namely, Ga-Mahlokwane (3.8 km), Tukakgomo (3.8 km), Ga-Phasha (3.8 km), Ga-Mampuru (8.4 km), and Stocking (9.4 km) are located within a radius of ±10 km around it. Eight mines surround Steelpoort as well: Mototolo Platinum Mine, Two Rivers Platinum: Modikwa Platinum, Dwarsrivier Chrome, Tweefontein Chrome, Tubatse Ferrochrome, Lion Ferrochrome Smelter, and Marula Platinum (Pty) Ltd.

2.1. Data Collection

Sampling was performed approximately 10 km west of Polokwane CBD, 18 km east of Medupi power station in Lephalale and 6 km west of Steelpoort. This is where the provincial monitoring stations are housed. South African Weather Services (SAWS), which is overseen by the South African Air Quality Information System (SAAQIS), provided the temperature, relative humidity (%), and air pollution data for Steelpoort, Polokwane, and Lephalale from January to December 2021. Radiello passive samplers were deployed for 7 successive days to monitor SO₂ and NO₂. The samplers consist of a microporous polyethylene and polycarbonate cylindrical diffusive body coaxially containing a cylindrical adsorbing cartridge. A 530 ± 30 mg of 35 - 50 mesh activated carbon are held within the adsorbing cartridge, which is a 3 × 8 µm with 4.8 mm external diameter stainless-steel net cylinder mesh. The sampler was protected from bad weather and direct sunlight by the mountable polypropylene shelter on the pole. The samplers were positioned 1.5 meters above the ground at each sampling location. The NO₂/SO₂ sampler is made up of a blue, microporous high-density polyethylene diffusive body placed on a polycarbonate supporting plate and a microporous polyethylene chemiadsorbing cylindrical cartridge coated with 270 mg triethanolamine (TEA). While SO₂ is partially oxidized to sulphate on the cartridge as a sulphite ion, TEA is used to chemiadsorbed NO₂ as a nitrite ion. According to the supplier’s instruction (Ielpo et al., 2019), the sampling rate (Q_R) of NO₂ at 298 K (25˚C) is 0.141 ± 0.007 ng ppb⁻¹∙min⁻¹. The effect of temperature on the sampling rate of NO₂ is given by Equation (1) as:

$Q_K=Q_R{\left(K/298\right)}^{7.0}$ (1)

In Equation (1), Q_K represents the rate of sampling at temperature K (in Kelvin) within the experimental range between 263 and 313 K (−10˚C to 40˚C). The sampling rate at 298 K, the reference temperature, is denoted by Q_R. Equation (2) below is used to calculate the $C_{{\text{NO}}_{\text{2}}}$ concentration in parts per billion (ppb):

$C_{{\text{NO}}_{\text{2}}}=m_{{\text{NO}}_{\text{2}}}/Q_K^t$(2)

where $m_{{\text{NO}}_{\text{2}}}$ is the nitrite mass in ng found on the cartridge, t is the exposure time in minutes, and Q_K is the sampling rate at temperature K.

At 298 K (25˚C), the SO₂ sampling rate ( $Q_{{\text{SO}}_{\text{2}}}$) is 0.466 ± 0.022 ng∙ppb⁻¹∙min⁻¹. Using the masses of sulphite ( $m_{{\text{SO}}_{\text{3}}}$) and sulphate ( $m_{{\text{SO}}_{\text{4}}}$) on the cartridge, the uptake rate, and the exposure time t in minutes, the concentration $C_{{\text{SO}}_{\text{2}}}$ in parts per billion is computed using the following Equations (3) and (4):

$m_{{\text{SO}}_{\text{4}}}\left(\text{T}\right)=m_{{\text{SO}}_{\text{4}}}+1.2m_{{\text{SO}}_{\text{3}}}$(3)

$C_{{\text{SO}}_{\text{2}}}=m_{{\text{SO}}_{\text{2}}}\left(\text{T}\right)/0.466t$(4)

where $m_{{\text{SO}}_{\text{2}}}\left(\text{T}\right)$ is the mass of sulphate in ng found on the cartridge (wherein the original sulphate mass and mass of sulphite are converted to sulphate).

2.2. Analyses of SO 2 and NO 2

Following exposure, adsorption cartridges were put back into their original tubes and 5 mL of ultrapure water (having a conductivity of 0.055 mS∙cm^–1) was added for desorption. To guarantee total analyte extraction, the tubes were vigorously shaken for two minutes using a Vortex-style stirrer. After that, the solution with the cartridge was left for an additional one and half hours. The desorption liquid was then transferred to IC injection vials after the tubes had been shaken for 30 seconds. Using Dionex equipment, including the Dionex LC20 chromatography enclosure with ED40 electrochemical detector, the Dionex AG22 guard column (4 × 50 mm) and the Dionex AS22 analytical column (4 × 250 mm), SO₂ and NO₂ were measured by ion chromatography. The DX-120 ion chromatograph, equipped with an ASRS-ultra II suppressor and a Dionex AS40 automatic injector, was used to measure sulphate. The ion chromatographic system consisted of an AS12A separator column and an AG12A guard column with an isocratic eluent of Na₂CO₃/NaHCO₃ (2.1 mM - 0.8 mM).

2.3. Quality Control/Quality Assurance (QC/QA)

The analytical quality of the data was evaluated using calibration linearity, instrument status monitoring, and limit of detection (LOD) [24] . Each sampling site always had a minimum of one set of unexposed samplers during the sampling campaign. These samplers served as blank samples and were filled with adsorbent solution and blank filters. All these blank samples were sampled and returned to the labs for appropriate analysis. The average of the amounts in the blanks multiplied by three times the standard deviation was used to calculate the limits of detection, which were based on the blanks. The LODs for SO₂ 18.8 - 52.3 was ng/sample, while NO₂ was 6.4 - 40.8 ng/sample. All the concentrations from the passive sampling were reported here without any blank value deduction because the blank values were negligible. Regression analysis was used to determine the linearity of calibration standards, and the results ranged from 0.99 to 1.00 (r²) for both NO₂ and SO₂ as determined by spectrophotometry and IC. Standard-spiked samples were subjected to routine analysis to verify the instrument's functionality.

2.4. Statistical Analyses of Data

Statistical analyses were performed using IBM SPSS® to observe the relationships between the air pollutant concentrations and the meteorological factors, such as temperature (T), wind speed (WS) and relative humidity (RH). Multiple linear regression (MLR) analysis was conducted to investigate the relations between S_O2 and NO₂ concentrations using meteorological factors and to obtain mathematical expressions. In formulating the regression equations, SO₂ and NO₂ concentrations were taken as dependent variables and the meteorological factors as independent variables. Since there is more than one independent variable, the multiple linear regression analysis was performed. A general MLR equation can be expressed by Equation (5):

$y_i=C+{\beta }_1{x}_1+{\beta }_2{x}_2+{\beta }_3{x}_3$(5)

where y_i is the dependent variable (NO₂ or SO₂), C is the constant of regression, β is a regression coefficient and x₁, x₂, x₃ are independent variables, x₁ (temperature, T, ˚C), x₂ (wind speed, WS, m/s) and x₃ (relative humidity, RH, %).

3. Results

Figure 2 shows the SO₂ concentrations obtained in three different areas, Lephalale, Polokwane and Steelpoort.

Figure 2

Figure 2. Monthly concentrations of SO 2 , (a), and NO 2 , (b) from January to December 2021.

3.1. NO 2 and SO 2 Data

The results of concentrations of NO₂ and SO₂ during the period of study (January 2021 to December 2021 are shown in Figure 2 for Lephalale, Polokwane and Steelpoort. These set of data are for the four seasons which are defined as follows: winter (June, July, and August); spring (September, October, and November); and summer (December, January, and February); and autumn (March, April, and May). It is indicated in Figure 2(a) that the concentration of SO₂ is higher in winter than in other seasons. Even though the highest peaks occur in different month, viz. June in Steelpoort, July in Lephalale and August in Polokwane. The order of increasing concentration of air pollutant, SO₂, is Steelpoort > Lephalale > Polokwane. Maximum pollution levels of sulphur dioxide for Steelpoort, Lephalale and Polokwane are 1.9 µg/m³, 1.6 µg/m³ and 1.4, µg/m³, respectively. The expectation was that Lephalale would have high SO₂ concentrations due to. Medupi power station location and coal, which is the source of SO₂ [25] [26] . Low levels of SO₂ in Lephalale might be due to the sampling site, which is 18 km from the Medupi power station. Furthermore, the wind direction is predominantly north-easterly. The levels of SO₂ in Steelpoort are due to the closeness of the monitoring station to vehicular emission and domestic biomass. The highest SO₂ concentrations during winter in this study’s areas were due to increased fossil fuel, increased traffic volumes and meteorological conditions [27] . The wintertime inversion layer formation is another reason for the elevated SO₂ concentration. As a result, SO₂ is prevented from escaping into the atmosphere and the dispersion rate of emission is decreased. The SO₂ and NO₂ concentrations did not exceed the annual average of 50 µg/m³ and 94 µg/m³ as set by NAAQS and WHO.

Variations in NO₂ concentrations from January 2021 to December 2021 are shown in Figure 2(b) . Different maximum concentration values of NO₂ are attained in different monitoring stations and months. Lephalale, Polokwane and Steelpoort have highest NO₂ concentration of 1.74 µg/m³ in September, 1.57 µg/m³ in June and 0.84 µg/m³ in July, respectively. The annual non-exceedance of NO₂ for the two standards, i.e., NAAQS and WHO, was observed in all areas under this study. A slightly different seasonal trend was observed for NO₂ compared to SO₂. The higher concentration of NO₂ in Lephalale and Polokwane are due to traffic around the areas. Similar studies have reported direct vehicle emissions as a major source of NO₂ concentrations [28] . Polokwane was expected to have higher concentration because it is a capital city with many vehicles moving in and out of the city. However, the sampling site is about 10 km away from the city. NO₂ concentration in Steelpoort is due to domestic heating and heavy-load vehicles around the area.

3.2. Evaluation of the Effect of Meteorological Parameters on SO 2 and NO 2 Levels

Figures 3 - 5 illustrate the effect of atmospheric factors on concentrations of air pollutants under study.

The relationship between temperature, wind speed, and relative humidity and the concentrations of NO₂ and SO₂ in Steelpoort is depicted in Figure 5 .

The monthly concentrations of SO₂ and NO₂ differ in relation to the meteorological factors as shown in the figure. NO2 concentrations depend on relative humidity as decreases as relative humidity increases. A maximum value of 1.7 µg/m3 was achieved when the relative humidity of 39.21% is low. Moreover, the temperature and the wind speed have little effect on the influence of NO₂ concentration. The impact of atmospheric factors in Polokwane is shown in Figure 4 .

The two air pollutants’ respective concentrations are influenced by temperature. The concentrations of NO₂ and SO₂ are highest at low temperatures, or between 12 and 17˚C. Nonetheless, the temperature range of 18 to 20˚C is where the concentrations of these air pollutants are concentrated. Low SO₂ and NO₂ concentrations are found at slower wind speeds. Nevertheless, the dependence of concentration on wind speed is independent of the speed. When the wind speed

Figure 3

Figure 3. Scatter plots of (a) temperature, (b) wind speed, and (c) relative humidity on SO 2 and NO 2 concentrations in Lephalale.

Figure 4

Figure 4. Dependency of air pollutants concentrations on meteorological parameters (a) Temperature, (b) Wind speed, and (c) Relative humidity in Polokwane.--Figure 4. Dependency of air pollutants concentrations on meteorological parameters (a) Temperature, (b) Wind speed, and (c) Relative humidity in Polokwane.

Figure 4

Figure 4. Dependency of air pollutants concentrations on meteorological parameters (a) Temperature, (b) Wind speed, and (c) Relative humidity in Polokwane.--Figure 4. Dependency of air pollutants concentrations on meteorological parameters (a) Temperature, (b) Wind speed, and (c) Relative humidity in Polokwane.

Figure 5

Figure 5. Variations of air pollutants as a function of meteorological factors in Steelpoort: wind speed (a), wind speed (b) and relative humidity (c).

is between 2.6 and 2.8 m/s, there are significant variations in the levels of SO₂ and NO₂. At a relative humidity of 59.52%, the concentration of SO₂ is high, while at 83.32%, the level of NO₂ is at its maximum.

The relationship between temperature, wind speed, and relative humidity and the concentrations of NO₂ and SO₂ in Steelpoort is depicted in Figure 5 .

As seen in the figure, relative humidity has little effect on concentrations of NO₂ and SO₂. At relative humidity between 55% and 60%, the SO₂ concentrations are high (1.2 to 1.8 µg/m³) compared to NO₂. In other relative humidity values, there is not much difference with respect to the concentrations of both pollutants. The effect of temperature depicts high levels of NO₂ and SO₂ at temperatures 15˚C and 26˚C, respectively, for the Steelpoort area. There was not much temperature effect on the air pollutants levels. Wind speed affects the concentration level of SO₂ when the wind speed was low, that is, between 1.1 to 1.3 m/s, the concentrations of SO₂ were high as shown in Figure 5 . Whereas, at speed between 1.4 m/s to 2.0 m/s, the concentrations range from 0.05 to 0.8 µg/m³.

3.3. Evaluation of the Effect of Meteorological Parameters on SO 2 and NO 2 Levels

In order to estimate how the NO₂ and SO₂ concentrations depend on meteorological factors a MLR was performed. The MLR coefficients of NO₂ and SO₂ concentrations with respect to meteorological parameters in different monitoring areas are illustrated in Table 1 . The null hypothesis, according to the table’s results, that there is no significant difference between any seasonal mean, is rejected. All p-values were less than 0.01.

The initial models, derived from Equation (5), for NO₂ are given by Equations

Table 1

Table 1. Summary of coefficients of air pollutants using MLR model.

	MLR	NO2 β (p-value)	SO2 β (p-value)	NO2 β (p-value)	SO2 β (p-value)	NO2 β (p-value)	SO2 β (p-value)
1	C	2.894 (0.003)	1.814 (0.028)	1.898 (0.058)	1.583 (0.111)	1.419 (0.021)	3.310 (0.003)
	T	−0.055 (0.208)	−0.114 (0.020)	−0.127 (0.027)	−0.018 (0.721)	−0.14 (0.620)	0.080 (0.096)
	WS	0.517 (0.293)	0.420 (0.373)	−0.094 (0.618)	0.080 (0.677)	−0.04 (0.379)	0.420 (0.004)
	RH	0.280 (0.076)	0.012 (0.387)	0.026 (0.153)	−0.016 (0.381)	−0.006 (0.428)	0.012 (0.009)
2	C	3.439 (<0.001)	2.316 (<0.001)	1.654 (0.040)	1.508 (0.098)	1.316 (0.014)	--
	T	−0.016 (0.458)	0.104 (0.718)	−0.0120 (0.022)	0.100 (0.568)	−0.245 (0.243)	--
	WS	--	-0.085 (0.005)	--	--	--	--
	RH	−0.040 (0.001)	--	0.023 (0.154)	0.023 (0.154)	0.023 (0.154)	--
3	C	3.256 (<0.001)	2.343 (<0.001)	2.292 (0.003)	1.758 (0.027)	0.944 (0.015)	--
	RH	−0.043 (<0.001)	−0.079 (<0.001)	--	--	−0.008 (0.165)	--
	T	--	--	−0.070 (0.048)	--	--	--

NB: C: Constant, T: Temperature, WS: Wind Speed, RH: Relative Humidity.

(6) to (8) for Lephalale, Polokwane and Steelpoort, respectively. Furthermore, models for SO₂ are shown in Equations (9) to (11).

NO₂ = 2.894 − 0.055T + 0.517WS − 0.028RH (Lephalale)(6)

NO₂ = 1.898 − 0.127T - 0.094WS − 0.026RH (Polokwane)(7)

NO₂ = 1.419 − 0.014T - 0.204WS − 0.006RH (Steelpoort)(8)

SO₂ = 1.814 − 0.114T + 0.420WS + 0.012RH (Lephalale)(9)

SO₂ = 1.583 − 0.018T + 0.080WS − 0.016RH (Polokwane)(10)

SO₂ = 3.310 − 0.080T – 1.449WS − 0.038RH (Steelpoort)(11)

According to the statistical model Equation (6) through Equation (11), the NO₂ and SO₂ concentrations are inversely correlated to temperature and relative humidity apart from SO₂ in Lephalale, as shown in Equation (9). The pattern is consistent with the results obtained by Ebrahimi and Qaderi, 2021 [23] [29] . Also, this means that when temperature and relative humidity are increased by one ˚C and 1.0%, the concentration of NO₂ decreases by 0.006 µg/m³ to 0.028 µg/m³. Meanwhile, the concentrations of SO₂ decrease by 0.016 µg/m³ and 0.038 µg/m³ in Polokwane and Steelpoort. At the same time, the concentration of SO₂ increases by 0.012 µg/m³. It is worthwhile noting that wind speed has both positive Equations (6), (9) and (10), and inverse, Equations (7), (8) and (11) correlations to the concentrations of both air pollutants and this is due to the impact of vehicular traffic. The positive correlation of both NO₂ and SO₂ with wind speed in a low-income community [30] .

The backward elimination method in the MLR model was used to remove independent variables that are insignificant where alpha, α = 0.1. The final models are presented in Equations (12) to (17). This method begins by entering all terms specified on the stepwise list into the model. At each step, the least significant stepwise term is removed from the model until all the remaining stepwise terms have a statistically significant contribution to the model. After eliminating the non-significant variables, the final models indicate that relative humidity has an inverse correlation with the concentrations of NO₂ in all studied sites, which implies that lowering of the NO₂ concentration is followed by the increase of relative humidity and vice versa. However, the relationships between SO₂ and the independent variables differ. In Lephalale and Polokwane, there is an inverse correlation of temperature and relative humidity with SO₂ concentration, respectively. There is no statistical significance between the SO₂ concentrations and the meteorological parameters in Steelpoort. This demonstrates that meteorological conditions have a significant impact on how air pollutants are distributed.

According to the statistical model Equation (6) through Equation (11), the NO₂ and SO₂ concentrations are inversely correlated to temperature and relative humidity apart from SO₂ in Lephalale, as shown in Equation (9). The pattern is consistent with the results obtained by Ebrahimi and Qaderi, 2021 [23] [29] . Also, this means that when temperature and relative humidity are increased by one ˚C and 1.0%, the concentration of NO₂ decreases by 0.006 µg/m³ to 0.028 µg/m³. Meanwhile, the concentrations of SO₂ decrease by 0.016 µg/m³ and 0.038 µg/m³ in Polokwane and Steelpoort. At the same time, the concentration of SO₂ increases by 0.012 µg/m³. It is worthwhile noting that wind speed has both positive Equations (6), (9) and (10), and inverse, Equations (7), (8) and (11) correlations to the concentrations of both air pollutants and this is due to the impact of vehicular traffic. The positive correlation of both NO₂ and SO₂ with wind speed in a low-income community [30] .

The backward elimination method in the MLR model was used to remove independent variables that are insignificant where alpha, α = 0.1. The final models are presented in Equations (12) to (17). This method begins by entering all the terms specified on the stepwise list into the model. At each step, the least significant stepwise term is removed from the model until all the remaining stepwise terms have a statistically significant contribution to the model. After eliminating the non-significant variables, the final models indicate that relative humidity has an inverse correlation with the concentrations of NO₂ in all studied sites, which implies that lowering of the NO₂ concentration is followed by the increase of relative humidity and vice versa. However, the relationships between SO₂ and the independent variables differ. In Lephalale and Polokwane, there is an inverse correlation of temperature and relative humidity with SO₂ concentration, respectively. There is no statistical significance between the SO₂ concentrations and the meteorological parameters in Steelpoort. This demonstrates that meteorological conditions have a significant impact on how air pollutants are distributed.

NO₂ = 3.256 − 0.43RH (Lephalale)(12)

NO₂ = 2.292 − 0.70RH (Polokwane)(13)

NO₂ = 0.944 − 0.008RH (Steelpoort)(14)

SO₂ = 2.343 − 0.079T (Lephalale)(15)

SO₂ = 1.758 − 0.020RH (Polokwane)(16)

SO₂ = Null hypothesis (Steelpoort)(17)

4. Conclusion and Future Research

The findings of this study demonstrate that throughout the winter, SO₂ concentrations are high in all areas, viz., Steelpoort, Polokwane, and Lephalale. Higher SO₂ levels were found in Steelpoort, followed by Lephalale and lastly in Polokwane. However, this might be due to the distance between the Medupi power station and the sampling site. Lephalale had the highest levels of NO₂ as compared to Polokwane and Steelpoort. Air pollutants in Steelpoort come from various sources such as mining activities, domestic heating and vehicular emissions. The results of the multiple linear regression models demonstrated that meteorological factors like temperature, wind speed, and relative humidity have an impact on the dispersion of NO₂ and SO₂. However, the meteorological factors affect the air pollution levels differently. In Polokwane and Steelpoort, NO₂ and SO₂ concentrations are inversely correlated to temperature and relative humidity except for SO₂ in Lephalale. Wind speed has positive and inverse correlations to the concentrations of both air pollutants. The final model showed that relative humidity negatively influences air pollutants. Future research such as collecting daily data concentrations of SO₂ and NO₂; assessment of impact of climate change as well as the impact of these air pollutants on human health around the areas, needs to be studied as a preventative measure of air pollution.

Acknowledgements

The authors acknowledge the Limpopo Department of Economic Development, Environment and Tourism, South African Department Forestry, fisheries and Environment, for availing meteorological data through SAAQIS, University of Limpopo and Mintek for availing facilities and material resources to conduct the study.

References

1. Alolayan, M.A., Almutairi, A., Aladwani, S.M. and Alkhamees, S. (2023) Investigating Major Sources of Air Pollution and Improving Spatiotemporal Forecast Accuracy Using Supervised Machine Learning and a Proxy. Journal of Engineering Research, 11, 87-93. >https://doi.org/10.1016/j.jer.2023.100126

2. Bikis, A. (2023) Urban Air Pollution and Greenness in Relation to Public Health. Journal of Environmental and Public Health, 2023, Article 8516622. >https://doi.org/10.1155/2023/8516622

3. Duffney, P.F., Stanek, L.W. and Brown, J.S. (2024) Air Pollution: Sources, Regulation, and Health Effects. In: Wexler, P., Ed., Encyclopedia of Toxicology, Academic Press, 215-228. >https://doi.org/10.1016/b978-0-12-824315-2.00754-5

4. Garland, R., Naidoo, M., Sibiya, B. and Oosthuizen, R. (2017) Air Quality Indicators from the Environmental Performance Index: Potential Use and Limitations in South Africa. Clean Air Journal, 27, 33-41. >https://doi.org/10.17159/2410-972x/2017/v27n1a8

5. Bonjour, S., Adair-Rohani, H., Wolf, J., Bruce, N.G., Mehta, S., Prüss-Ustün, A., et al. (2013) Solid Fuel Use for Household Cooking: Country and Regional Estimates for 1980-2010. Environmental Health Perspectives, 121, 784-790. >https://doi.org/10.1289/ehp.1205987

6. Röllin, H. (2017) Evidence for Health Effects of Early Life Exposure to Indoor Air Pollutants: What We Know and What Can Be Done. Clean Air Journal, 27, 2-3. >https://doi.org/10.17159/2410-972x/2017/v27n1a1

7. Agarwal, R. and Aggarwal, S.G. (2023) A Year-Round Study of Ambient Gaseous Pollutants, Their Atmospheric Chemistry and Role in Secondary Particle Formation at an Urban Site in Delhi. Atmospheric Environment, 295, Article 119557. >https://doi.org/10.1016/j.atmosenv.2022.119557

8. WHO (2019) Air Pollution. World Health Organization.

9. Çapraz, Ö., Deniz, A. and Doğan, N. (2017) Effects of Air Pollution on Respiratory Hospital Admissions in İstanbul, Türkiye, 2013 to 2015. Chemosphere, 181, 544-550. >https://doi.org/10.1016/j.chemosphere.2017.04.105

10. Hossain, M.S., Frey, H.C., Louie, P.K.K. and Lau, A.K.H. (2021) Combined Effects of Increased O ₃ and Reduced NO ₂ Concentrations on Short-Term Air Pollution Health Risks in Hong Kong. Environmental Pollution, 270, Article 116280. >https://doi.org/10.1016/j.envpol.2020.116280

11. Pretorius, I. (2015) The Impact of the South African Energy Crisis on Emissions. WIT Transactions on Ecology and the Environment, 198, 255-264. >https://doi.org/10.2495/air150211

12. Shezi, B. and Wright, C.Y. (2018) Household Air Pollution Exposure and Respiratory Health Outcomes: A Narrative Review Update of the South African Epidemiological Evidence. Clean Air Journal, 28, 43-56. >https://doi.org/10.17159/2410-972x/2018/v28n1a11

13. Okello, N.O., Okello, T.W. and Zunckel, M. (2020) Changes in Health Risk Associated with Air Pollution and Policy Response Effectiveness, Richards Bay, South Africa. Clean Air Journal, 30, 1-10. >https://doi.org/10.17159/caj/2020/30/1.8012

14. Joss, M.K., Eeftens, M., Gintowt, E., Kappeler, R. and Kunzli, N. (2017) Time to Harmonize National Ambient Air Quality Standards. International Journal of Public Health, 62, 453-462. >https://doi.org/10.1007/s00038-017-0952-y

15. Pienaar, J.J., Beukes, J.P., Van Zyl, P.G., Lehmann, C.M.B. and Aherne, J. (2015) Passive Diffusion Sampling Devices for Monitoring Ambient Air Concentrations. Comprehensive Analytical Chemistry, 70, 13-52. >https://doi.org/10.1016/bs.coac.2015.09.002

16. Department of Environmental Affairs (2016) Draft Strategy to Adress Air Pollution in Dense Low-Income Sttlements. Staatskoerant, No. 40088, 1008-1023.

17. Department of Environmental Affairs (2018/2019) Annual Report. >https://www.gov.za/sites/default/files/gcis_document/201911/environmental-affairs-1819-anreport.pdf

18. Bird, T., Liebenberg-Enslin, H., Von Gruenewaldt, R., Modisamongwe, D., Thivhafu-ni, P. and Mphahlele, T. (2011) Developing an Air Quality Management Plan: Lessons from Limpopo. Limpopo Department of Economic Development, Environment and Tourism (LEDET).

19. Maciejczyk, P., Chen, L. and Thurston, G. (2021) The Role of Fossil Fuel Combustion Metals in PM _2.5 Air Pollution Health Associations. Atmosphere, 12, Article 1086. >https://doi.org/10.3390/atmos12091086

20. Dou, S., Xu, D. and Keenan, R.J. (2023) Effect of Income, Industry Structure and Environmental Regulation on the Ecological Impacts of Mining: An Analysis for Guangxi Province in China. Journal of Cleaner Production, 400, Article 136654. >https://doi.org/10.1016/j.jclepro.2023.136654

21. Quarm, J.A., Anning, A.K., Fei-Baffoe, B., Siaw, V.F. and Amuah, E.E.Y. (2022) Perception of the Environmental, Socio-Economic and Health Impacts of Artisanal Gold Mining in the Amansie West District, Ghana. Environmental Challenges, 9, Article 100653. >https://doi.org/10.1016/j.envc.2022.100653

22. Arowosegbe, O.O., Röösli, M., Künzli, N., Saucy, A., Adebayo-Ojo, T.C., Schwartz, J., et al. (2022) Ensemble Averaging Using Remote Sensing Data to Model Spatiotemporal PM ₁₀ Concentrations in Sparsely Monitored South Africa. Environmental Pollution, 310, Article 119883. >https://doi.org/10.1016/j.envpol.2022.119883

23. Swartz, J.-S., Van Zyl, P.G., Beukes, J.P., Labuschagne, C., Brunke, E.-G., Portafaix, T., et al. (2020) Twenty-One Years of Passive Sampling Monitoring of SO ₂, NO ₂ and O ₃ at the Cape Point GAW Station, South Africa. Atmospheric Environment, 222, Article 117128. >https://doi.org/10.1016/j.atmosenv.2019.117128

24. He, J., Xu, H., Balasubramanian, R., Chan, C.Y. and Wang, C. (2014) Comparison of NO ₂ and SO ₂ Measurements Using Different Passive Samplers in Tropical Environment. Aerosol and Air Quality Research, 14, 355-363. >https://doi.org/10.4209/aaqr.2013.02.0055

25. Nakaishi, T. (2021) Developing Effective CO ₂ and SO ₂ Mitigation Strategy Based on Marginal Abatement Costs of Coal-Fired Power Plants in China. Applied Energy, 294, Article 116978. >https://doi.org/10.1016/j.apenergy.2021.116978

26. Ai, H., Zhou, Z., Li, K. and Kang, Z. (2021) Impacts of the Desulfurization Price Subsidy Policy on SO ₂ Reduction: Evidence from China’s Coal-Fired Power Plants. Energy Policy, 157, Article 112477. >https://doi.org/10.1016/j.enpol.2021.112477

27. Emami, F., Masiol, M. and Hopke, P.K. (2018) Air Pollution at Rochester, NY: Long-Term Trends and Multivariate Analysis of Upwind SO ₂ Source Impacts. Science of the Total Environment, 612, 1506-1515. >https://doi.org/10.1016/j.scitotenv.2017.09.026

28. Schatke, M., Meier, F., Schröder, B. and Weber, S. (2022) Impact of the 2020 COVID-19 Lockdown on NO ₂ and PM ₁₀ Concentrations in Berlin, Germany. Atmospheric Environment, 290, Article 119372. >https://doi.org/10.1016/j.atmosenv.2022.119372

29. Ebrahimi, M. and Qaderi, F. (2021) Determination of the Most Effective Control Methods of SO ₂ Pollution in Tehran Based on Adaptive Neuro-Fuzzy Inference System. Chemosphere, 263, Article 128002. >https://doi.org/10.1016/j.chemosphere.2020.128002

30. Adesina, J.A., Piketh, S.J., Burger, R.P. and Mkhatshwa, G. (2022) Assessment of Criteria Pollutants Contributions from Coal-Fired Plants and Domestic Solid Fuel Combustion at the South African Industrial Highveld. Cleaner Engineering and Technology, 6, Article 100358. >https://doi.org/10.1016/j.clet.2021.100358