Analysis of the Distribution of Some Potentially Harmful Elements (PHEs) in the Krugersdorp Game Reserve, Gauteng, South Africa

: The Mintails Mogale Gold (MMG) and the Rand Uranium (RU) are two large-scale mining consortiums active in re-mining old tailings dams and dumps in Krugersdorp and are a source of mine discharge feed into the Krugersdorp Game Reserve (KGR). This has resulted in a noticeable accumulation of potentially harmful elements (PHEs) over a number of years. Efforts were implemented to interpret the concentration levels of PHEs in soils of the study areas of which a total of 36 georeferenced soil samples were collected (in triplicate) from the MMG, RU and KGR, including samples from farmlands and waterways adjacent to the mining sites. Samples were then analysed by both inductively coupled plasma optical emission spectrometry (ICP-OES) and inductively coupled plasma mass spectrometry (ICP-MS) for 36 elements. From the 36 elements of this study, detailed evaluations of the occurrence of 12 selected elements were discussed. The geochemical landscape at the KGR is shown to be in flux. The major mediating influences on the behaviour of As, Co, Cu, Hg and Pb, as they enter the KGR largely in the form of acid mine drainage (AMD), are the geological substrate (mostly in carbonate form). Analysis of the soils showed high levels of contamination for As and Co in ppm. The mean maximum of As ranged from (5.00–170.30) with the highest level found in the Krugersdorp site. The mean maximum of Co ranged from (46.00– 102.30) with the highest level found in MMG. All of these values were well above the recommended maximum acceptable concentration (MAC) values, i.e., As (15–20) and Co (20–50). The mean maximum values for Pb (12.40–92.30); Cu (18.50–115.30) and Hg (12.40–92.30) content in surface soils of all four segments studied falls well within the MAC range for agricultural soils i.e. Cu (60– 150); Hg (0.5–5) and Pb (20–300).


Introduction
Krugersdorp Game Reserve (KGR) in the Gauteng Province of South Africa is one of the country's most important wildlife parks in terms of the revenue it generates from tourism and related activities. The Reserve is juxtaposed to two major mining outfits, Mintails Mogale Gold (MMG) and Rand Uranium (RU) at Krugersdorp, and smallholdings (SH). This makes the total environment an ideal setting for investigating the impact of mining and ore processing activities on the health of neighbouring wildlife communities, and other nearby ecosystem components.
Krugersdorp (26˚6'0"S; 27˚46'0"E), Figure 1, also referred to as Mogale City, is located in the West Rand of the Gauteng Province of South Africa. It has a population of close to half a million inhabitants. It is flecked with a number of abandoned gold (with Mn, Fe, asbestos and lime) mines and boasts an industry that contributes significantly to the economic activity of the Gauteng region. MMG and RU are two large-scale mining consortiums active in re-mining of old tailings dams and dumps in Krugersdorp. Although a comprehensive literature search revealed extensive published information on the hydrochemical environment and soil geochemistry at MMG, RU and SH, as well as wildlife health at KGR, information regarding any on soil geochemistry is scarce [1]. The main aim of this study was to provide a comprehensive data base of the occurrence and levels of potentially harmful elements (PHEs) in the study area, which is not available anywhere, rather than only focus on a few selected elements.
Heavy metals are a natural occurrence in nature all around the world or as a result of anthropogenic activities through emissions from the rapidly escalating industrial areas, mine tailings, discarding of high metal wastes and atmospheric deposition [1][2][3][4]. The metals commonly found in soil as a result of these activities include copper (Cu), lead (Pb), zinc (Zn), nickel (Ni), cobalt (Co), mercury (Hg) and cadmium (Cd). Some of these metals are essential in small doses for use in living organisms for their normal physiological functioning. However, large concentrations can tend to destabilize ecosystems due to bioaccumulation in organisms causing toxic effects through metabolic interference and mutagenesis [1,[5][6][7]. In addition, their accumulation in soils persists for a long period of time after their introduction, contributing to contamination of ecosystems [1,8,9]. Arsenic in particular, is quite toxic and readily absorbed by plants from groundwater and soil [10]. Mining of gold minerals, as seen in the mining consortiums in Krugersdorp is related to acid drainage problems that can cause long-term damage to waterways and biodiversity. Acid mine drainage (AMD) occurs when sulfide-bearing material are exposed to oxygen and water and typically ensues in iron sulfideaggregated rocks. Although this process occurs naturally, mining can stimulate AMD generation merely through increasing the measure of sulfides exposed. Furthermore, naturally-occurring bacteria can hasten AMD production by supporting the breakdown of sulfide minerals [11]. The obvious concern is the deterioration of water and soil quality from the impact of AMD from the West Rand Basin, which inspired the need to garner more detailed information on the distribution characteristics and fate of the heavy metals around the mining area at Krugersdorp and to establish their levels of concentration. The above considerations provided motivation for the present study, which is aimed at investigating and documenting the PHE-and other element additions from mine effluents (AMD). This study reports on the distribution and concentration of heavy metals due to RU and MMG mining activities in the Krugersdorp region and their migration to surrounding smallholdings and the KGR, focusing on cobalt (Co), mercury (Hg), aluminium (Al), zinc (Zn), copper (Cu), nickel (Ni), lead (Pb), cadmium (Cd) and arsenic (As). This information is vital in formulating intervention measures such as mitigation (AMD neutralisation), rehabilitation and soil amendment.

Soils
Soils of Krugersdorp are mainly reddish brown to red and belong to the Hutton form (orthic topsoil on red apedal subsoil, usually on rock). The soils range in texture from clayey silt through silty clay to loam. During the field studies, sandy clay was also encountered at a few sites, but the sand content was always less than 10%. The soils are predominantly light textured (15%-25% clay), and vary in depth from less than 400 mm to over 1.2 m.
The KGR is primarily grassland, with shrub thickets, small forests and fynbos found in the valley and close to the small dams. The only information that refers to soil quality, and is relevant to the present study, is the description of agricultural potential of the proposed route of the Westgate substation, west of KGR [12].
Although Krugersdorp is designated as an agricultural hub in the Gauteng Province, the general potential for agricultural development remains relatively low. Based on soil factors (not climatic conditions), Paterson (2008) describes areas of moderate or high agricultural soils [13].

Sampling Technique and Sample Preparation
A comprehensive literature search was used to garner published information on the hydrochemical environment and soil geochemistry at MMG, RU and SH. Some data were found to exist on hydrochemistry and wildlife health [1], but hardly any on soil geochemistry. Therefore, in order to complement the existing hydrochemical data, a total of 36 soil samples were collected for analysis (in triplicate) from cultivable and uncultivable lands at the MMG and RU gold mine fields, and the KGR and adjacent SH to provide information on the soil geochemistry.
The map of the study area with sample sites is shown in Figure 2. At the mining sites (MMG and RU) soil samples were collected on a grid using pre-determined Global Positioning System (GPS) points. Adjacent to the mines and at the KGR, soil samples were taken at random intervals and at a distance of 15 m on either side of minor roads or footpaths (not shown) running approximately in the direction of rivers draining the two mining areas into the KGR (Figure 2). A distance of 15 m on either side of the paths was considered sufficient in minimizing anthropogenic influence on element fluxes in soils in the non-cultivable sectors of the land. This sample design was selected in order to determine concentration levels and migration characteristics of PHEs in pollution trails emanating from the mine sites and entering the Krugersdorp ecosystem, to provide a comprehensive data base of the occurrence and levels of PHEs in the study area.
The GPS coordinates as well as other geoenvironmental characteristics such as geology, topography, vegetation, pedology and hydrological characteristics were carefully recorded at each sampling site. The soil sampling protocol comprised of the construction in an undisturbed area, of a 1 m x 1 m pit to a depth of 40 cm using pick and shovel. A composite near-surface soil sample (30 cm below A0) was collected and stored in pre-labelled Kraft paper envelopes in accordance with The International Geological Correlation Programme (IGCP) 259 recommendations [13], modified for application in tropical conditions by [14].

Soil Sample Preparation and Analysis
Each soil sample was oven-dried (30 °C) for about 12 h, homogenized and then quartered in the Biochemistry Laboratories of the Mangosuthu University of Technology. Following disaggregation, the soil samples were sieved to <180 µm (80 mesh). The samples were then stored in plastic vials and shipped to the headquarters of ACME Laboratories in Vancouver, Canada, for analyses. ACME's sample preparation method comprised of the weighing out of a 30 g split followed by an Aqua Regia dissolution technique. Following soil sample digestion, a range of nutritional and PHEs were determined by inductively coupled plasma optical emission spectrometry (ICP-OES) and inductively coupled plasma mass spectrometry (ICP-MS) techniques, but only the PHEs of As, Co, Cu, Hg and Pb were selected for this study.

Quality Assurance Evaluation
Commercially available standards and previously analysed internal standards, as well as reagent blanks were included to check accuracy and precision. Triplicates were randomly incorporated into all analytical batches. The data were accepted if the relative standard deviation was <15% at five times the limit of detection for the triplicate samples. Triplicate samples comprised 16% of total samples. Standard solutions of all the elements were prepared by diluting their corresponding 1000 mg/L certified standard solutions. The blank reagents and standard reference materials (SRMs) of the metals were used to verify the accuracy and precision of digestion, while in view of data quality assurance, each sample batch was analysed in triplicate under standard opt-ionizing conditions within the confidence limit of 95% (see www.acmelab.com for further details).

Soil Fraction Analysed
The fine fraction of 180 µm (80 mesh) was considered suitable for the present study. A number of studies have determined that metals in the coarse-grained soils are mainly derived from parent materials, and metals in the fine fraction are mostly anthropogenic [15,16]. Other studies [17][18][19][20] have classified the fraction where enrichment of PHEs take place in (contaminated) soils as clay-and siltsized. The high specific surface area and charge of fine particles are the probable factors that account for the high contamination levels seen in the fine fraction [21]. The Aqua Regia technique used for decomposition of soil samples (in this study) is a partial dissolution procedure able to release metals associated with a recent pollution source, and could be released to the soil solution with input of water (non-silicate-bound forms); but would not dissolve metals associated within the solid phases, i.e., within the structure of the insoluble minerals [22]. As such, the method is deemed suitable for the assessment or prediction of crop plant metal content, plant growth and plant deficiency or toxicity [23,24]. The method could therefore be considered a suitable tool in the assessment of the long-term potential risk of heavy or toxic metals entering the biosphere.

Results and Discussion
As pointed out in the introduction, the main aim of this study was to provide a comprehensive data base of the occurrence and levels of PHEs in the study area, which is not available anywhere, rather than only focus on a few selected elements. These are presented in Tables 2-4.

Total Content of Selected Heavy Metals
Some soil pollution water sources and parameters are sighted in this section. February 2011 evidence of ADM in the Krugersdorp area involving the Tweelopie stream is shown in Figure 3. Previous (2006) KGR water hydrochemical parameters are outlined in Table 1 [25]. The general analytical results of the thirty-six (36) elements in Table 2 range from 9.22% (for Fe) to less than 0.1 ppm. From the 36 elements of this study, detailed evaluations of the occurrence of 12 selected elements are presented in Table 3. Of these, five heavy metals, namely As, Co, Cu, Hg and Pb, the data of which is presented in Table 4, from the respective study sites, were considered for a comparative analysis of their concentrations in the soil samples. The concentrations obtained in the study sites are compared with the reference values for background concentrations, Johnson [26] after Wedepohl [27].  In the KGR, surrounding smallholdings and RU, Mn had the highest mean concentration followed by Cr. These three study sites also had a similar trend for the four lowest heavy metal concentrations, Pb > As > Cd > Hg (Tables 3 and 4). The soil samples from the KGR had mean heavy metal concentrations of 93.41, 4280.33, 17.37, 64.73, 35.07, 88.56, 0.17, 0.12, 11.55, and 7.56 mg·kg −1 for Cr, Mn, Co, Ni, Cu, Zn, Cd, Hg, Pb, and As respectively. In the soils from MMG, the trend was Mn > Zn > As > Cr > Cu > Ni > Pb > Co > Cd > Hg (Table 3). The higher standard deviation reveals higher variations in heavy metal distributions from the point source of discharge to the adjacent areas.    * The mean concentrations were compared to reference values (mg.kg -1 ) for background concentrations after [27].
The sampling sites in the KGR (Table 3) showed a distinct migration pattern of heavy metal concentrations. Sampling sites KS8 and KS10 displayed higher mean concentrations than the other sampling sites in the KGR. This is interesting since KS8 and KS10 were located in close proximity to the TLS (Figure 1). The higher mean concentrations are to be expected since AMD first decanted into this waterway in September 2002 [29]. The sampling site KS1 in the KGR had predominantly the lowest mean heavy metal concentrations (Table 4). Again, this is to be expected, since this sampling site was the furthest from the point source of discharge (Figure 1). Table 4. As, Co, Cu, Hg and Pb (ppm) in agricultural soils and in living matter, compared with mean values in soils of the study localities. a [30]; b [31].

Distribution of Soil Geochemical Data
Mean concentration ranges of Aqua Regia extractable As, Co, Cu, Hg and Pb fraction (pseudototal fraction) of soil samples collected for this study were compared with the range of maximum allowable concentrations (MAC) for agricultural soils worldwide, as given in [30]. Quite unexpectedly, mean PHE concentrations in soils of the KGR and SH are well within the range of MAC for agricultural soils worldwide. Table 4 gives the statistical parameters of surface soil PHE data, including standard deviations and location of mean contents within the range of MAC given for agricultural soils worldwide [30]. High standard deviations signify wider variations in PHE distribution from the point source of discharge of AMD to the adjacent areas. Of the four study segments (KGR, SH, MMG and RU), only MMG gives a mean value for As (130.07 ppm) falling above the MAC range for agricultural soils worldwide (15-20 ppm). Mean As values for the other three segments fall well within the MAC range (Table 4). Similarly, Co, KGR and SH gave mean values for surface soils (102.30 ppm) and (93.00 ppm) respectively, that is above the upper threshold of the MAC range for agricultural soils (20-50 ppm) given in [30]. The other two study segments gave mean values well within this range.
The mean Cu content in surface soils of all four segments studied falls well within the MAC range for agricultural soils (60-150 ppm) given in [30]. The highest Cu values were again found in samples from the MMG segment, but with a mean value (99.73) far removed from the upper threshold (150 ppm) of the MAC range. The mean Cu concentrations decrease in the order: MMG (99.73 ppm) > KGR (35.07 ppm) > SH (19.88 ppm) > RU (16.10 ppm) ( Table 4).
The mean Hg content in surface soils of all four segments studied falls well within the MAC range given for agricultural soils (0.5-5 ppm) in reference [30]. The highest Hg values were again found in samples from the MMG segment.
The mean Pb content in surface soils of all four segments studied falls well within the MAC range for Pb in agricultural soils (20-300 ppm) given in reference [30]. The highest Pb values were again found in samples from the MMG segment.

Sampling Sites
KS8 and KS10, in close proximity to the Tweelopiespruit (TLS) (Figure 1), displayed higher mean concentrations of PHEs in surface soil compared to the other sampling sites in the KGR. This is unsurprising, given that the TLS is the waterway where AMD first decanted in September 2002 [15]. The sampling site KS1 in the KGR gave the lowest PHE concentrations. Again, this is to be expected, since this sampling site was the furthest from the point source of AMD discharge (Figure 2). The relatively low concentrations of Cu, Hg and Pb in surface soils of the KGR, explained both on the basis of improved quality of influent mine water and carbonate-water interactions in the soil, is taken to indicate that the potential supply of these elements to herbage and crops would be limited.

Conclusion
Data from the hydrochemical literature as well as from current research on soil geochemistry indicate that large amounts of mine pollutants produced predominantly in the form of AMD have infiltrated soils and other environmental compartments of the KGR through ground-and surface waters, over several decades of gold mining. Analysis of the soils showed high levels of contamination for As and Co in ppm. The mean maximum of As ranged from (5.00-170.30) with the highest level found in the Krugersdorp site. The mean maximum of Co ranged from (46.00-102.30) with the highest level found in MMG. All of these values were well above the MAC recommended values, i.e., As (15)(16)(17)(18)(19)(20) and Co (20-50). The mean maximum values for Pb (12.40-92.30), Cu (18.50-115.30), and Hg (12.40-92.30) content in surface soils of all four segments studied falls well within the MAC range for agricultural soils, i.e., Cu (60-150), Hg (0. , and Pb (20-300). The major mediating influences on the behaviour of As, Co, Cu, Hg and Pb, as they enter the KGR largely in the form of acid mine drainage (AMD), are the geological substrate (mostly in carbonate form).
These pollutants can cause modifications to natural element cycling in the KGR-and adjacent soil environments. The fluxes of residual As, Co, Cu, Hg and Pb in soils from influent AMD that is available for plant uptake in KGR is considered to be determined largely by the nature of the underlying substrate (predominantly carbonate) and the level of treatment of decant from adjacent mining localities during any given season. The high As and Co contents found in most of the soils in our study areas provide a major motivation for proposing further investigation of the waters infiltrated by AMD contamination and comparing it to the soil values obtained to give a more comprehensive idea of the amount of PHE accumulation. Well-designed, quantitative research along these lines will enable the rehabilitation of metal-contaminated areas with appropriate indigenous species, allow identification of metal toxicities when they occur and allow for the effective regulation of metal emissions. This information is considered vital in the formulation of intervention measures such as mitigation, Acid mine drainage (AMD) neutralization, rehabilitation and soil amendment. Results from these kinds of studies would enable more informed decisions to be taken on site investigations, clean-up actions, and remedy selection. In accordance with the observations of DEAT [32], more research is required on the field impact of AMD on potentially receiving dolomitic (karst) environments and the re-activation of springs dried-up due to dewatering. Strategies should be evolved to utilize the storage potential of defunct/closed underground mine voids so as to optimally manage the generation of AMD and control its potential impact on the receiving surface and groundwater environments.