Analysis of raw water quality from Vaalkop dam, South Africa

The South African constitution states that everyone has a right to have access to an environment that is not harmful to their health or well-being. This includes a constant supply of clean and safe drinking water. In order to fulfil this constitutional right, it is imperative to carry out continuous monitoring of water at different stages of the water supply system so that real data can be obtained about the state of water quality in the environment. Management interventions can then be made to maintain or improve water quality.

Magalies Water is a state-owned entity providing a wide range of related water and sanitation services, and operates in four provinces that include Gauteng, Limpopo, North West and Mpumalanga. It obtains a significant fraction of its raw water from the Vaalkop Dam, a combined gravity and earth-fill type dam located in North West Province, South Africa. Its reservoir is located at the confluence of the Elands River and the Hex River, part of the Crocodile basin. The dam was established originally in 1972 and was renovated in 2008 in order to supply water for the platinum and associated metals mining operations in the area. The reservoir mainly provides water for irrigation purposes, municipal water supply and industrial uses.

Magalies Water recognises the critical role played by the monitoring of water quality and therefore carries out continuous measurement and collection of various water quality data. This data was handed over to the authors for analysis, which culminated in the production of this paper. The major water quality parameters that have been analysed include pH, conductivity and dissolved solids, turbidity and colour, total coliforms, E. coli, hardness, alkalinity, precipitation potential, chlorophyll and organic carbon.

The term pH (Jonnalagadda & Mgere 2001) describes the extent to which water is acidic or basic (alkaline). The pH scale ranges from 0 to 14. Water that is neutral has a pH of 7; water that is acidic has a pH less than 7 and water that is basic has a pH higher than 7. Very acidic or very basic water can be corrosive and can also interfere with water treatment process efficiency. pH is an important treated water quality parameter mainly due to the fact that (among other reasons), potable water should be non-corrosive.

The pH, as can be observed in Figure 1, fluctuates between 7.3 and 8.6, and is shown to exhibit a very small increase over time, which indicates slightly increasing alkalinity of the raw water. With a linear trend line whose slope is 0.3%, and without any anomalous deviations, this observation can be safely ignored. Moreover, the values fall within acceptable limits, thus in terms of pH, the water is healthy and there is no urgent need for any remedial action. However, an ever rising pH may indicate the presence of pollutants in the water, such as algae or an increase in carbonates. If this trend goes unabated, the pH in the future may rise above the recommended value of 9.5, which will compromise treatment efficiency.

Figure 2 shows that the conductivity of the water is high, surpassing the alarm level for approximately 50% of the period under consideration. It fluctuates between 42 and 89 mS/m. Although the results in Figure 2 generally imply a declining trend in conductivity, the very low coefficient of determination could suggest that the declining trend cannot be confirmed. What is certain, however, is that conductivity is not increasing over time. Moreover, while it is an indicator of the dissolved solid content of the water, as can be seen in Figures 3 and 4, all dissolved solids were found to have concentrations that were consistently and significantly lower than their recommended alarm levels. For the period under study (despite the missing data intervals): sodium and fluoride were less than 50% of the alarm level, sulphates were found to be less than 25% of the alarm level; potassium was less than 25% of the alarm level; the highest concentration of chlorides was 75% of its alarm level, and it occurred once. In addition, these concentrations were not increasing over time. Other dissolved solids are discussed as nutrients in a subsequent section, with similar results. This observation requires further verification.

Hardness is a measure of the amount of divalent salts, or positively charged ions, particularly calcium (Ca²⁺) and magnesium (Mg²⁺), in water. Total hardness is the sum of the concentrations of Ca²⁺ and Mg²⁺, expressed in ppm (mg/l) calcium carbonate. Calcium carbonate hardness is a general term that indicates the total amount of divalent salts present, although it does not specify which salts are causing water hardness. Hard water is generally not harmful to human health but wastes soap and forms scale precipitates upon heating (Kumar & Puri 2012; Sengupta 2013).

Alkalinity, on the other hand, is a measure of the capacity of water to neutralize acids. Alkaline compounds in the water such as bicarbonates, carbonates, hydroxides and phosphates lower the acidity of the water. Alkalinity, like hardness, is reported as milligrams per litre of calcium carbonate (mg/l CaCO₃), and owing to this fact, hardness and alkalinity are often confused. However, alkalinity measures negative ions (carbonates and bicarbonates) while hardness measures positive ions (calcium and magnesium), and sometimes these values can differ greatly. If limestone (calcium carbonate) is the cause of hardness and alkalinity, these values will be similar or identical. However, if sodium bicarbonate (NaHCO₃) is responsible for high alkalinity, it is possible for water to have high alkalinity and low hardness.

Figure 5 reveals declining trends for both hardness and alkalinity, whose values are already well below the alarm levels. The low and declining values for hardness can be explained by the low and reducing values of their causative agents (calcium and magnesium), as was observed in Figure 3.

Measured in milligrams per litre of calcium carbonate, this parameter indicates the level of corrosivity of the water (Randtke 2011). A negative potential indicates that the water is corrosive, i.e., can dissolve the metal of the pipeline. A positive potential indicates that calcium carbonate can be precipitated from the water and deposited into the pipeline. This action can have the advantage of protecting the pipeline through creating a lining, but can, if in excess, reduce the hydraulic capacity of the pipeline, hence the establishment of a limit.

The precipitation potential of Vaalkop water was mainly positive, with some scattered periods depicting negative values (Figure 6). Thus, the water is not corrosive, although it can be hard. However, the alarm level of the precipitation potential was never exceeded for the whole period under investigation. Furthermore, this is in agreement with the findings in the section on hardness and alkalinity, which shows that the water was not excessively hard. It is important to note that no trend could be accurately established due to gaps existing in the available data.

Colour in water may be caused by the presence of minerals such as iron and manganese or by substances of vegetable origin such as algae and weeds. Colour tests indicate the efficiency of the water treatment system. Potable water should be colourless and non-turbid. Colour levels were found to vary between 13 and 70 mg/l Pt, and generally fell below the alarm level (Figure 7). The linear trend line depicts a very small increase in colour over time. With a slope of 5%, this variation can be safely neglected. However, an interesting observation is the sharp unprecedented increase in colour during the month of December 2010. It is not immediately clear what triggered this sharp increase.

Turbidity is a measure of water clarity or the extent to which colloidal substances in water decrease the passage of light through the water (Juntunen et al. 2012). Colloidal substances are chemically reactive suspended substances that cannot be removed by filtration and may include clay, silt, algae, plankton, microbes, and other substances. Turbidity levels were found to vary between 1 and 23 NTU (Figure 8), and generally falling below the alarm level. The linear trend line depicts a very slight increase in turbidity over time (6% slope), which can be safely neglected. This could be due to developments within the catchment. As in the case of colour, an interesting observation is the sharp unprecedented increase in turbidity in the month of December 2010.

Coliforms are used as indicators of the presence of pathogens (Onda et al. 2012; Reddy & Lee 2012; Bartram et al. 2014). Total coliforms can exist in several media such as the waste of animals and humans, in soil, and wood. Studying Figure 9, it can be inferred that prior to 2010, the total coliform count was well below the alarm level (at under 40% of the alarm level). The count then started increasing, eventually surpassing the alarm level in December 2011. The highest count of total coliforms recorded for the period under consideration was 34,824/100 ml, or approximately 35 times the alarm level in December 2011.

The subsequent continuous rise in the coliform count starting December 2010 should be verified. During the same month, there is also a noticeable increase in colour (Figure 7), turbidity (Figure 8), and conductivity (Figure 2). Explanation of this unprecedented behaviour requires verification of the events that may have led to this result. Although the graph shows a declining total coliform count in the second half of 2012, more data is required to confirm this trend.

E. coli is an indicator of exclusive contamination of raw water by human waste. From Figure 10, it can be observed that the E. coli count was above the alarm level between April 2007 and May 2008, from April 2008 to June 2008, and from October 2008 until the end of the evaluation period. The highest E. coli count was 700.75/100 ml, or 70 times the alarm level, in April 2012. Although the graph shows a declining E. coli count in the second half of 2012, more data is required to confirm this trend. The subsequent continuous rise in the coliform count starting December 2010 should be verified.

It is unacceptable to find E. coli in raw water at such an alarming rate. This suggests that perhaps, management and disposal of waste water in areas that comprise the catchment of Vaalkop dam are not appropriate. Disinfection of the treated water should be thorough in order to destroy all pathogens. Effort should be made to ensure adequate disposal of waste water.

This parameter measures the amount of organic matter in the water (Kaire Toming et al. 2016). Too much organic matter will attract undesirable microorganisms that will grow in pipelines. The dissolved organic carbon content (Figure 11) of the water is below the recommended alarm level. Establishment of the trend is not possible due to extensive missing data.

Chlorophyll is a pigment in plants that enables photosynthesis to take place. The concentration of chlorophyll is a convenient indicator of the amount of algae in the water. Too much algae can cause undesirable tastes and odours, clog filters, consume chlorine and can lead to the production of undesirable disinfection by-products. All of this can increase the cost of treating water for domestic use. Chlorophyll values were consistently below 42% of their recommended values for the period July 2007 – June 2012 (Figure 12), implying a low algal concentration in the water. These findings can be explained by the low levels of feeder nutrients discussed in the following section. These findings, though, are not conclusive due to several gaps in the available data.

This section discusses ammonium, nitrates, nitrites and phosphates as nutrients, due to their importance in supporting algal growth (Thi Minh Hanh et al. 2011; Yan et al. 2015; Liu & Chan 2016). With respect to their impact on human health, the concentrations shown in Figure 13 do not pose any threat. As dissolved inorganic solids, they generally fall below the recommended alarm levels for human health for the entire six years of the study period, and only exceeded the alarm levels in few and scattered cases. For example, in the case of ammonium, the alarm level was only exceeded in February 2011, and nitrates only exceeded their alarm level in May 2011. An accurate establishment of their trend was complicated by missing data.

• i.

  In general terms, the pH, colour, turbidity, dissolved solids, hardness, alkalinity, precipitation potential, chlorophyll and organic carbon were found to be within acceptable limits.

• ii.

  The conductivity of the water is high, surpassing the alarm level for approximately 50% of the period under consideration, even though the causative dissolved solids are appreciably below their alarm levels.

• iii.

  There was a continuous rise in the coliform count starting December 2010. During the same month, there was also a noticeable increase in colour, turbidity, and conductivity. Sharp increases were also noted in the total coliform count in December 2011; and in the E. coli count in April 2012.

• i.

  A further investigation, requiring more data, is required to explain the contradiction between high conductivity and low dissolved solid concentration.

• ii.

  It is important to establish the root cause of increased turbidity, colour, conductivity and coliform counts in December 2010 and thereafter, in order to explain the unprecedented increases in their values.

• iii.

  It is critical to determine the events that took place in December 2011 in order to explain the unprecedented spike in the total coliform count, and in April 2012, in order to explain the unprecedented spike in the E. coli count.