Distribution and Risk Assessment of Heavy Metals in Surface Water from Pristine Environments and Major Mining Areas in Ghana

mining sites were found to be slightly higher than those from the pristine sites. Conclusions. The concentrations of heavy metals in the Nyam, Subri, Bonsa and Birim Rivers from the mining sites and the Atiwa Range, Oda, Ankasa and Bosomkese Rivers from the pristine sites were found to be either below or within the USEPA and WHO’s recommended limits for surface water. The health risk assessment values for the hazard quotient for ingestion of water (HQing), dermal contact (HQderm) and chronic daily intake (CDI) indicated no adverse effects as a result of ingestion or dermal contact from the rivers. However, arsenic (As) in both the pristine and mining sites and chromium (Cr) in the pristine sites pose a carcinogenic threat to the local residents. Competing Interests. The authors declare no competing financial interests. J Health Pollution 9: 86–99 (2015)


Introduction
Ghana, like many countries in Africa, has a history of heavy metal pollution largely emanating from industrial effluent discharges and anthropogenic deposits from industrial activities due to prevailing winds. One of the biggest contributors to pollution in the Ghanaian environment is mineral mining. In Ghana, both local and foreign investor companies have equal rights to engage in mineral exploration. In 2000, a total of 224 local and foreign companies obtained mineral rights for gold exploration, as well as 600 registered small-scale miners. At the same time, the informal (illegal) sector locally referred to as "galamsey operators" dominated the small-scale gold sector with an estimated number of 200,000 people in 2004, up from a total of approximately 60,000 people in 1997. 1 In Ghana, the most common approach adopted by many mining companies is the use of cyanide, which makes it possible for mining companies to make large profits from low grade ores. 2 Lately, the use of cyanide in gold extraction has become unattractive as a result of several cyanide spillages which have caused significant damage to the environment. According to Amegbey and Adimado, the number of officially reported cyanide spillages between 1989 and 2003 in Tarkwa and Obuasi alone was 11 cases. 3 A separate investigative report issued by the Wassa Association of Communities Affected by Mining (WACAM) 4 indicated that cyanide spillages increased from 8 between 1989 and 2002 to about 13 cyanide spillages in 2006. This situation has led to the pollution of major water bodies such as streams and rivers, and over 52% of the population in the mining areas still lacks potable water. 5 Wassa Association of Communities Affected by Mining

World Health Organization
Zinc As gold is the main contributor to Ghana's economy, accounting for 38% of total stock and 95% of total mineral exports, 5 small-scale mining as well as improper regulation of large-scale mining pose enormous environmental challenges such as land degradation, subsidence due to gold mining and water pollution, all of which pose risks to human health. 6-8 Wastes from mining processes such as tailings can result in the influx of metals and toxic chemicals into the environment. Waste rocks are known to contain arsenic (As), mercury (Hg), cadmium (Cd), lead (Pb) and other toxic metals. 9,10 Eventually, the metals attain higher concentrations and accumulate in large quantities in crops and plants, and pose serious health hazards to humans and animals through bio-magnification. 11 There have been reports indicating increased concentrations of heavy metals and other pollutants in water bodies around mining communities in Ghana. Reports from studies have shown that from 1947 to 1992, mine effluents were discharged without restriction and treatment into water bodies, soil and air, thereby resulting in the degeneration of the environment. 12-14 Rivers are the main surface water resources for domestic, industrial and irrigation purposes, and their contamination as a result of large industrial wastewater discharges and seasonal run-offs from agricultural lands poses a growing threat to the environment. 15 In recent years, many rivers, streams and ponds in Ghana have had elevated inputs of heavy metals as a result of an increase in mining and other anthropogenic activities in and around water bodies, raising serious concern about the toxic effects on aquatic organisms and persistence in the environment. 16 Huge deposits of mine wastes as well as ore stockpiles and waste rocks are The aim of this study is therefore to determine the distribution and health risks of heavy metals in surface water from both pristine environments and major mining areas in Ghana.

Sampling and Analysis
Four water bodies were selected based on their proximity to the mining areas.  Research water solution and eventually rinsed with distilled water. Those samples meant for metal analysis were acidified to a pH of 2 using concentrated nitric acid before being transported to the laboratory in an ice-chest filled with ice chips. The samples were stored in a refrigerator at 4 o C upon arrival at the laboratory for further analysis. 24,25

Digestion and Analysis of Heavy Metals in Water
Digestion of the water samples involved sample measurement and reagent addition and digestion in a microwave. Then 5 ml of the water samples in three replicates were accurately weighed into TFM Teflon vessels of a microwave digester (Milestone ETHOS 900). Next, 6 ml of nitric acid (HNO3) (65%), 3 ml of hydrogen chloride (35%) and 0.25 ml of hydrogen peroxide (30%) were added to each of the vessels containing the samples. The vessels were swirled gently until thoroughly mixed and then fitted vertically into the microwave digester and digested for 21 minutes.
The Teflon bombs mounted on the microwave carousel were cooled in a water bath for 20 minutes to reduce high temperatures and pressure buildup within the vessels. The digestates were then transferred quantitatively into a volumetric flask and diluted to 20 ml using distilled water. A blank was also prepared in similar fashion, but without the analyte.
All the samples were analysed using a Varian AAS 240FS flame atomic absorption spectrometer coupled with an Atlas Copco LFx MED Superflow compressor at the Nuclear Chemistry and Environmental Research Center, Ghana Atomic Energy. To ensure the reliability of the analytical method during digestion and sample preparation, quality control (QC) and blank samples were digested along with each set of samples and subsequently analyzed for appropriate elements through the same procedure. Reference standards for this work were obtained from Fluka Analytical, Sigma-Aldrich Chemie GmbH, Switzerland.

Determination of Physico-chemical Characteristics
Physico-chemical measurements of pH, conductivity, temperature and turbidity were conducted to identify indications that would affect the results obtained from the study. The American Public Health Association (APHA, 1998) method for preparation and analysis of water samples was employed for the determination of physicochemical parameters in this study. The pH of the water samples was determined alongside the temperature using a pre-calibrated pH meter. Conductivity was measured using a pre-calibrated Hach SensION5 conductivity meter. The turbidity of the water samples was measured with a Hach turbidimeter. In addition, pH and temperature were measured on site.

Quality Control
Samples were analyzed in triplicates and after every 4 samples, a calibration standard was analyzed to check the response and efficiency of the analytical instrument alongside the blank which was used to constantly check for contamination.
Calibration curves were optimized by the use of quality control standards at every step of sample reading. All glassware were soaked in 10% HNO 3 overnight and thoroughly washed with distilled water and dried in an oven overnight at 50-60 o C. After oven-drying, the glassware were dried in a desiccator for about 20 minutes before use.

Data and Statistical Analysis
Data from the study were analyzed using IBM SPSS Statistics version 21 and the Excel Analysis ToolPak. Basic statistics such as mean and standard deviation were computed along with the multivariate statistics. Factor analysis (FA) and principal component analysis (PCA) were computed to identify significant principal components in the data as well as possible loadings and sources of the heavy metals. The PCA was carried out by the Promax normalized rotation method for the results. A scree plot was performed to determine how many important components were present in the data and to provide a visual explanation of the metal loading process which describes loading groups with a reduction in the original data. 26, 27

Human Health Risk Assessment
The human health risk assessment was used to characterize the nature and magnitude of possible human health risks and ecological receptors from heavy metal contaminants and other stressors that may be present in the environment. The health risk assessment for the heavy metals in surface water from both the mining areas and the pristine environments was estimated via ingestion and dermal contact based on the USEPA risk assessment method. 28 Details of the standards employed in the calculation of various parameters are shown above ( Table 2).
In the exposure assessment, the average daily dose (ADD) for the heavy metals level in surface water from the pristine and mining environments was calculated using the following slightly modified equations from USEPA protocol in 1989 and 2004.

ADDing = (Cx×Ir ×Ef×Ed)/ (Bwt ×At ×365)
Where, Cx is the concentration of toxicant metals in the drinking water (mg/L), Ir is the ingestion rate per unit time (L/day), Ed is the exposure duration (years), Ef is the exposure frequency (days/year), Bwt is the body weight of receptor (kg), and At is the averaging time (years) which is equal to the life expectancy of a resident Ghanaian. For the conversion factor from year to days, 365 was used. In addition, ADDing is the quantity of heavy metals ingested per kilogram of body weight. 29,31 In this study, surface water ingestion is assumed to be the main pathway for risk assessment because the rivers are potential sources of drinking water. However, dermal contact is another important pathway, because residents sometimes swim in these rivers and thus may come into contact with the toxic metals. 32 Average daily dose for dermal contact was calculated using the formula:

ADDderm = (Cx×Sa ×Pc×Et ×Ef ×Ed ×Cf )/(Bwt ×At ×365)
Where, Sa is the total skin surface area (cm 3 ), Cf is the volumetric conversion factor for water (1L/1000 cm3), Et is the exposure duration (h/day), Pc is the chemical-specific dermal permeability constant (cm/h), Ef is the exposure frequency (days/years), Ed is the exposure duration (years), Lt is the human lifetime (defined as 70 years) and Bwt is body weight.

HQ = ADD/RfD
Where HQ represents the hazard quotient via ingestion or dermal contact (no units) and RfD is the oral/ dermal reference dose (mg/L/day). 26 The chronic daily intake (CDI) of the metal was estimated using equation 4 below: :

CDI = C (DIing/Bwt)
Where C is the concentration of heavy metal in water, DI is the average daily intake rate (2 L) and Bwt is the body weight (70 kg).
Finally, the carcinogenic risks (CRs) of the metals were estimated to assess the probability of an individual developing cancer over a lifetime as a result of exposure to a potential carcinogen. The slope factor (SF) is a toxicity value that quantitatively defines the relationship between dose and response. Potential carcinogenic effect probabilities that an individual will develop cancer over a lifetime of exposure are estimated from projected intakes and slope factor. The range for carcinogenic risk acceptable by the USEPA is 1x10 -6 to 1x10 -4 .

Physico-chemical Variations in the Water Samples
The level of pH in the aquatic environment can affect chemical and biological processes. Even though humans may not be affected directly by pH, an elevated range of pH results in a bitter taste for drinking water. 30 The pH ranges for rivers in the study area are shown in Table 3. All rivers fell within the WHO range for potable water of 6.5 to 8.5. Temperatures ranged from 28.0 to 28.7º C with an average of 28.4º C. Conductivity of the water samples from the four mining sites ranged from 442 to 2890 µs/ cm, with a mean value of 1666 µs/cm and 228 to 938 µs/cm from pristine sites with a mean value of 583 µs/cm. Turbidity values of water samples from the mining areas in this study ranged from 0 to 92 nephelometric turbidity units (NTU) with a mean of 46 NTU, whiles samples from the pristine areas and 2.18 for components 1 and 2, respectively, as shown in Table 5. The FA/PCA on components 1 and 2 accounted for 47.60% and 31.14%, respectively, of the total variance.

Human Health Risk Evaluation
The evaluation of the heavy metals in the water samples from pristine and mining sites for possible adverse health effects associated with exposure to such chemicals revealed that the level of exposure through ingestion (ADDing) was observed in the order Cr > Cd > As from the pristine and mining sites. 28 The observed dermal exposure (ADDderm) was in the order Cr > Cd > As from both sites. In the present study, Zn, Fe, Co and Mn were not considered in the risk and hazard quotient estimation because their slope factors could not be obtained and Pb and Cu were below the detection limit.     46 Turbidity results from the Bonsa, Ankasa, Atiwa and Oda Rivers in this work can therefore be considered acceptable since they have values equal to or below the WHO's 5 NTU limit. However, turbidity values from the Subri, Birim, Nyam and Bosomkese Rivers were far above the WHO limit and therefore can be considered to be unfit for human consumption. Sample filtration was conducted for samples whose turbidity persisted after sample digestion.

Concentrations and Site Variation of Heavy Metals
The distribution of heavy metals in the rivers from the selected sampling sites was generally low compared to that of national and international standards as shown in Table 4. The low levels of Cd suggest that the source may not necessarily be as a result of the mining activities, but Cd may occur naturally with zinc and sulfide ores. 47 The mean concentrations of the heavy metals in samples from both sites were below the WHO standard limit except for Fe, which showed varied metal levels with a few values higher than WHO and USEPA required limits. The high levels of Fe may result from natural geological sources or corroded iron metals carried in seasonal run-offs from illegal mining activities.
In Ghana, the presence of Hg in mining communities is linked with the use of Hg in gold mining. Average Hg levels were therefore expected to be high in water samples from the mining sites. Hg was detected in the Ankasa and Bosomkese Rivers, but the levels (0.002 mg/L) were below WHO recommended guideline limits. 39 The Hg presence in the pristine environment could be linked with illegal mining activities in the forest reserves or anthropogenic deposition of this metal.   Research concentrations of heavy metals in this study may be attributed to the fact that water levels increase during rainy seasons (wet season), resulting in a decrease in the concentration of metals. Most of the pristine sites are located in remote mountains and as a result, the concentrations of some of the measured metals were low.

Statistical Evaluation of the Heavy Metals
Results from the statistical analysis conducted to identify possible loading on components are shown in Table 5.
The results suggest that only the first two components have eigen-values greater than one (> 1.00), and together these explain over 79% of the total variability in the data. Also, Bartlett's test of sphericity gave a p-value of 0.000, which is below the recommended value of < 0.001, and therefore a good indication for a valid FA/PCA to be performed on the data. Looking at results from the pristine sites, the factor analysis resulted in 4 components with total %variance of 87.75% and eigenvalues of 2 The correlation analysis showed that a stronger correlation was observed at the mining sites than the pristine sites.
Correlations at the mining sites are shown in Table 5. The correlation results suggest that Zn, Mn, Fe and Cr may be coming from the same input sources, while As and Cd may be coming from different input sources. The correlation coefficients as well as the extracted component loading values of Cd and Co (r=0.425) and As and Co (r=0.653) indicate a closer association of Cd with As than with Co. The component plot for this phenomenon is shown in Figures 2 and 3. A similar trend of association of metals was observed in the pristine sites with a majority of the correlation coefficients falling below r=0.4, representing a weak correlation. The moderate correlation between Zn and Cd is not surprising since Cd occurs in the natural environment, typically in association with zinc ores (an impurity in zinc). Cobalt correlated positively and moderately with As (r=0.625) and Hg (r=0.545). There was strong and positive correlation between Hg and As (r=0.768) which could be due to the fact that along with mercury, arsenic is part of the toxic residue of gold mining. As, Hg and Co associations were confirmed by the metals' strong loading on component 1 in the FA/PCA results. However, most of the metals from the mining sites may result from anthropogenic deposition and the use of pesticides and fertilizers by farmers in the vicinity of the selected water bodies, which are eventually washed off into the rivers. 2

Analysis of Variance Estimation
Two-way analysis of variance  Table 6 and 7 for the pristine and mining sites, respectively. However, a significant difference was found between the concentrations of the metals, giving a p-value far less than 0.05 and an F-value greater than the F-crit (p-1.1×10-9, F=8.21, Fcri=1.95). These differences were confirmed by FA/PCA and the correlation results, where some of the metals had a stronger correlation than others and also had different loading on the components.

Human Health Risk Evaluation
In addition, HQing and HQderm for Cr, Hg and Cd from the mining and the pristine sites were below 1.0, indicating their minimum hazard effect on the local residents who utilize the rivers. However, the HQing for As and Cd values of 0.76 and 0.71, respectively, were a bit near unity, which indicates the potential of these metals posing a hazard threat should their levels continue to accumulate. From the pristine and mining sites, the average levels of HQing for Cr, Cd, and As were found to be 0.64 and 0.55, respectively, as shown in Table 8, while the observed values for HQderm were 0.17 and 0.11, respectively. According to USEPA risk assessment guidelines, when the value of the hazard quotient is greater than 1.0, the probability of adverse health effects due to exposure is high. 28, 45 In the current study, the hazard quotients were below 1.0, which suggests that communities which depend on the rivers for drinking or swimming may not be at high risk of illnesses associated with high levels of consumption of contaminants.
The chronic daily intake (CDI) values for the metals from the pristine and mining sites are shown in Table 8. The low CDI indices indicate that mining activities and agricultural practices like fertilization and run-off are not impacting the heavy metals load in the river bodies in the present study and also do not affect the water quality of the rivers.
Carcinogenic risk through ingestion of heavy metals (CRing) for Cr, Cd and As were estimated to be 2.57×10-6, 4.68×10-7 and 1.52×10-4, respectively, from the pristine sites, and 1.50×10-6, 5.15×10-7 and 1.52×10-4, respectively, from the mining sites. The CRderm from the pristine and mining sites ranged from 10-6 to 10-8, which is below the remedial target goal of 10-6. The USEPA has estimated the CR from projected intakes and slope factor. The range for carcinogenic risk acceptable by the USEPA 50 is 1×10-6 to 1×10-4. The CRing results from this study for Cr and Cd were within the USEPA's acceptable limits, with most of the values falling within the upper boundary. The CRing for As was above the remedial goal target of 1×10-6 in both the pristine and mining sites as well as Cr in the pristine sites, therefore raising carcinogenic concerns for the local residents around the catchment areas. 51 Inorganic arsenic and chromium (VI) are known human carcinogens. High levels of arsenic can cause cancer of the skin, lungs, liver and bladder. Lower intakes of As may cause nausea and vomiting, abnormal heart rhythm, and damage to blood vessels. Chromium (VI) compounds are toxins and known human carcinogens, and breathing elevated levels can cause irritation to the lining of the nose and nose ulcers. 52 The results also show that the carcinogenic risks were found to be higher than the non-carcinogenic risks to the residents through ingestion of water from water bodies around the mining areas.

Conclusions
The concentrations of heavy metals in the Nyam, Subri, Bonsa and Birim Rivers from the mining sites and the Atiwa Range, Oda, Ankasa and Bosomkese Rivers from the pristine sites were found to be either below or within the USEPA and WHO's recommended limits for surface water. The health risk assessment values for HQing, HQderm and CDI, were found to be below 1.0 in the pristine and mining sites, indicating no adverse effects as a result of ingestion or dermal contact of the rivers. However, CRing and CRderm values for As in both the pristine and mining sites and Cr in the pristine site were above the remedial goal target of 1×10-6, and therefore pose a carcinogenic threat to the local residents.
Inorganic arsenic is a known carcinogen and can cause cancer of the skin, lungs, liver and bladder. At low concentrations, As exposure can cause nausea and vomiting, decreased production of red and white blood cells, and damage to blood vessels. 53 Multivariate statistical analysis (PCA/FA) has confirmed both geogenic and anthropogenic sources of metal introduction into the rivers. This study is the first of its kind undertaken to compare metal pollution levels in areas believed to be highly polluted (mining areas) with areas believed to be out of metal pollution range (pristine locations). Even though further studies are needed, the current study nevertheless provides preliminary information on heavy metal levels in rivers from mining areas and pristine environments in Ghana which can be used for future water pollution monitoring.