Classification and quality of groundwater supplies in the Lower Shire Valley , Malawi – Part 2 : Classification of borehole water supplies in Chikhwawa , Malawi

This paper compares data gathered from a study of the chemical and bacteriological quality of drinking-water from 28 rural borehole supplies in Chikhwawa, Malawi, with a tiered classification scheme (Class 0 being ideal through to Class III being unsuitable for drinking without prior treatment) developed by investigators from the Institute for Water Quality Studies, Department of Water Affairs and Forestry, South Africa. In general, the majority of borehole water supplies were classified as Class 0 or Class I supplies based upon the chemical analysis and bacteriological examination. However the classification of a borehole water supply was variable and depended upon the parameter, date of sampling and whether or not it was based on the mean or individual concentration. A number of boreholes were classified as II or III as they contained elevated levels of fluoride and nitrate suggesting that consumption over short or prolonged periods of time may lead to adverse or serious health effects, such as skeletal fluorosis in adults and methaemoglobinaemia in infants. Research is required to develop practicable, affordable and sustainable methods to enable villagers to treat Class II/III water supplies and improve the quality of their drinking-water to a class suitable for human consumption.


INTRODUCTION
Over the last decade an increasing number of studies have been undertaken on the bacteriological and chemical quality of borehole water supplies in Africa. he data accrued by investigators is more oten than not compared with the health-based guideline values recommended by the World Health Organization (WHO, 2008). When studies are undertaken in conjunction with European and American investigators, the data presented is oten compared with EU and USEPA drinking-water standards. However, it should be noted that these standards are based upon the quality of drinking-water supplied to the consumer's tap ater the source water has undergone some degree of conventional treatment. Where national standards exist, investigators have attempted to compare their indings with those standards, but these are invariably those recommended by the WHO. Little attempt has been made to publish drinking-water standards speciic to borehole water supplies in Africa.
To address this issue Kempster et al. (1997) undertook a review of South African drinking-water standards and WHO recommendations, and produced a scheme to facilitate the classiication of borehole water supplies in rural areas based on single parameters. he rationale behind such a scheme was awareness that 'the boundary between the no-efect level [of constituents] and the threshold for the initial appearance of undesirable efects is not a sharp one ' (Kempster et al., 1997 p. 163). In areas of the world under pressure from water resource scarcity, realistically, water quality, both source and treated, can luctuate on a temporary basis. he scheme was designed with the thought of giving 'a clearer picture of expected efects on the domestic user'. he scheme is based on a select number of aesthetic and health-based parameters that are commonly of concern in drinking-water. On this basis, Kempster et al. produced 4 classes of water quality in terms of suitability for use, ranging from the ideal (Class 0) through to that which is deemed unsuitable for use without further treatment (Class III).
Recently, we undertook a study of the chemical quality of borehole water supplies in Chikhwawa, Malawi (Grimason et al., 2013), and compared our indings with the maximum permissible standards laid down by the Malawi Bureau of Standards for borehole water supplies (MBS, 2005), and the WHO recommended health-based guideline values for drinking-water (WHO, 2008). In this paper we present a summary of the chemical and bacteriological data accrued from that study and compare our indings with the classiication scheme proposed by Kempster et al. (1997). We believe that this is the irst time that the proposed scheme has been utilised for the classiication of borehole supplies outside South Africa.

METHODOLOGY
Eighty-four groundwater samples were collected for chemical analyses from 28 boreholes (n=16 on east bank; n=12 on west http 574 bank) located in 25 remote, rural villages in the Chikhwawa District (n=3 per borehole), distributed along the east (n=15) and west (n=10) banks of the Shire River. Samples were collected every 2 months during the wet season (December, February and April) (Grimason et al., 2013). Samples for chemical analysis and bacteriological examination were processed in accordance with standard methods for the examination of water and wastewater (APHA, 1998). Bacteriological samples were collected in 500 mℓ sterile vessels ater laming the inside of the tap of the borehole pump and running of the water to waste for 30 s before obtaining a water sample. Vessels were sealed, labelled, stored in a portable ice chest (coolbox) at 4°C and transported to the Environmental Health Laboratory at the University of Malawi within 6 h. Samples were processed for faecal bacteria using a standard membrane iltration technique and subsequently with conirmatory media in accordance with APHA (1998).

CLASSIFICATION SYSTEM
he classiication scheme proposed by Kempster et al. (1997 p. 163) was used to classify rural borehole water supplies examined in this study (Table 1): • Class 0 'his is ideal water quality, suitable for lifetime use, with no adverse health efects on the user. his class is essentially the same as the target water quality range in the 2 nd edition of the South African Water Quality Guidelines for Domestic Use (DWA F, 19 96).' • Class I 'Water in this class is safe for lifetime use, but falls short of the ideal water quality in that there may be instances of adverse health efects, but these are usually mild, and overt health efects are almost sub-clinical and diicult to demonstrate. Water in Class I does not cause health efects under normal circumstances. Aesthetic efects may, however, be apparent.' • Class II 'Water in this class is deined as that where adverse health efects are unusual for limited short-term use. Adverse health efects may become more common particularly with prolonged use over many years, or with lifetime use. his class represents water suitable for short-term or emergency use only, but not necessarily suitable for continuous use over a lifetime.'

• Class III
'his water has constituents in a concentration range where serious health efects might be anticipated, particularly in infants or elderly people with short-term use, and even more so with longer term use. he water in this class is not suitable for use as drinking water without adequate treatment to shit the water into a lower and safer class.'

RESULTS AND DISCUSSION
A summary of the data gathered by Grimason et al. (2013) on a range of chemical constituents and bacteriological analysis of 28 borehole water supplies in Chikhwawa, Malawi, is presented in Table 2. Based upon this data, each borehole water supply was classiied according to the criteria proposed by Kempster et al. (1997) (Table 3). Tables 4 to 7 present a summary of the water classes determined for each borehole water supply based on the results of 3 samples. he physicochemical and bacteriological data gathered highlight the variable nature of classifying borehole water supplies, demonstrating the need for a tailored classiication scheme proposed by these investigators. For example, for the borehole supply at Matumula on the east bank the class for the iron parameter was Class 0 on 2 sampling occasions and Class II on a separate sampling occasion (Table 4). Despite this variation in quality, based on the mean concentration of the three samples taken, the borehole supply would be classiied as Class II (Table 3).

Sulphate, nitrate, faecal coliform and luoride parameters
With the exception of the borehole supply at Kanthema (Class III) on the west bank, every other supply fell within Class 0 (n=25; 89%) or I (n=2; 7%) for the sulphate parameter. Elevated levels of sulphate in water (400-600 mg/ℓ) may be associated with diarrhoea in sensitive, transient and non-adapted consumers, and can impart a bitter taste to the water (Kempster et al., 1997;WHO, 2008). Consumers of borehole water at Dyelatu School and Miagno on the west bank complained of a bitter taste associated with the water. Interestingly, these two supplies were the only ones to fall within Class I, with mean sulphate concentrations ranging from 219-252.3 mg/ℓ and 285.8-305.9 mg/ℓ, respectively. At levels > 600 mg/ℓ in water, such as the case with Kanthema (mean concentration 781.7 mg/ℓ), Kempster et al. (1997) state such water should not be used for drinking, especially by infants, due to the possibility of life-threatening diarrhoea in this sensitive, non-adapted group of individuals. No health-based guideline value has been established by the WHO, as the existing data do not identify a level of sulphate in drinking water that is likely to cause adverse health efects in humans. However, because of the gastro-intestinal efects that can result from drinking water with elevated sulphate levels, the WHO recommend that health authorities be notiied of sources of drinking-water containing levels greater than 500 mg/ℓ (WHO, 2008). In Malawi, the Malawi Bureau of Standards (MBS) derived maximum permissible level for sulphate in borehole water supplies is 800 mg/ℓ, although the grounds upon which this standard is based is not known. Of concern is the alarmingly high number of borehole supplies that were classiied as Class II (n=2; 7%) / III (n=11; 39%) for the nitrate parameter, given its known association with methaemoglobinaemia (blue baby syndrome) in bottle-fed infants (Kempster et al., 1997). he risk of methaemoglobinaemia in infants signiicantly increases with simultaneous exposure to nitrates and microbial contaminants in water and is rarely associated with nitrate in the absence of faecal contamination of drinking-water (WHO, 2008). his is of particular concern as 5 of the 13 borehole water supplies that were categorised as Class II (n=1) / III (n=4) for the nitrate parameter (Mfela School, Chinkole, Nedi, Kabudula I and Lukiuji), contained faecal coliforms at levels which placed them in Classes II (n=1) and III (n=4), respectively, for the faecal coliform bacteria parameter. Most supplies were classiied as 0 / I (n=19; 68%) with respect to the presence of faecal coliform bacteria. permissible level for the faecal coliform parameter in borehole water is 50 cfu/100 mℓ (Grimason et al., 2013), whilst most national standards set a standard of 0 cfu/100 mℓ in drinking water in accordance with the World Health Organization guideline value (WHO, 2011). he basis for the MBS MPL for faecal coliforms is unclear and clariication could not be provided by the MBS when consulted. With the exception of samples collected from boreholes located at Khumbulani (> 40 mg/ℓ), Wiliyamu I (> 40 mg/ℓ) and Migano (> 170 mg/ℓ), all other borehole samples complied with the maximum permitted level of 45 mg/ℓ nitrate set by the Malawi Bureau of Standards and the WHO health-based guideline value (50 mg/ℓ). Although no faecal coliform bacteria were detected in samples examined from these three borehole supplies, the reasons for the signiicantly high nitrate concentrations detected warrant further investigation by the health authorities. As a short-term measure, the WHO recommend that water should not be used for bottle-fed infants when nitrate levels are above 100 mg/ℓ; however, it may be used if medical authorities are increasingly vigilant when the nitrate concentration is between 50 and 100 mg/ℓ, provided that the water is known and is conirmed to be microbially safe (WHO, 2008).
Of equal concern is the number of borehole water supplies which were classed as II (n=10) and III (n= 10) with respect to the luoride parameter, along with the known association of high luoride levels with dental and skeletal luorosis (Kempster et al., 1997;WHO, 2008). Fluoride in water is considered to be an essential element for the development and protection of teeth and bones. he current WHO health-based GV (1.5 mg/ℓ), which is 4 times lower than the Malawi maximum permitted level, is considered to be a threshold where the beneit of resistance to tooth decay is weighed against the risk of developing dental luorosis. In this study evidence of dental luorosis was clearly visible in consumers of Class II and III borehole water supplies, especially supplies located on the west bank of the Shire River. Consumption of borehole water containing levels greater than 3.5 mg/ℓ over a prolonged period of time may cause skeletal problems (Kempster et al., 1997). herefore additional studies need to be undertaken to determine the skeletal health efects in people consuming water from Class III supplies such as those located at Kapufeni, Kabudula II and Lakiuji.

Magnesium, sodium and chloride parameters
All borehole water supplies fell within either Class 0 (n=16; 57%) or I (n=12; 43%) for the magnesium parameter. At levels below 70 mg/ℓ (Class 0 and I supplies) no undesirable health or aesthetic efects are discernible; however problems associated with soap lathering may be noticeable with Class I supplies. Elevated levels of magnesium in water (> 70 mg/ℓ) can impart a bitter taste to water supplies (Kempster et al., 1997). he highest magnesium levels recorded in this study were detected in samples from the borehole supply located at Migano (54.3 mg/ℓ), which consumers complained as having a bitter taste to it.
With respect to the sodium and chloride parameters, all samples on the east bank were classiied as Class 0. In contrast, the majority of borehole supplies on the west bank were classiied as II (n=2) or III (n=8) with respect to the sodium parameter, and two supplies were classiied as II for the chloride parameter. Both Class II chloride supplies (Nedi and Migano) were associated with supplies classiied as III for the sodium parameter. Health-based guidelines for magnesium, sodium and chloride in drinking water have not been established by the WHO as they are deemed to be naturally occurring chemicals which are usually found at concentrations which are not toxic to health (WHO, 2008). However, it is acknowledged that concentrations of chloride above 250 mg/ℓ can give rise to a detectable 'salty' taste in water. In this study, 3 villages had borehole water supplies that contained chloride values greater than 250 mg/ℓ (Nedi, Migano & Mtondeza); interestingly water consumers from all three villages complained of there being a distinct and mildly unpleasant salty taste to the water. he maximum permitted levels for magnesium, sodium and chloride recommended by the Malawi Bureau of Standards for borehole supplies are 200 mg/ℓ, 500 mg/ℓ and 750 mg/ℓ, respectively.

Iron, manganese and zinc parameters
he majority of borehole water supplies on both sides of the river fell within Class II (n=25; 89%) for the iron parameter. Two supplies on the east bank were classiied as Class I and one supply on the west bank as Class III (Migano). In general, no health efects are associated with the levels of iron in Class 0, I and II supplies (Kempster et al., 1997), except in sensitive individuals. However, Kempster et al. (1997) do state that negative health efects can occur in infants who consume water from Class III supplies. With regards to the manganese parameter two-thirds of borehole supplies were classed as 0 (n=15; 54%) or I (n=3), well below the WHO health-based guideline value (0.4 mg/ℓ). Eight supplies on the east bank and 2 supplies on the west bank were classiied as Class II; however, the concentrations detected were usually below the WHO guideline value, with slight exceedances detected at Bello, Chadula II and Mtuwawa (≤ 0.6 mg/ℓ).
All borehole supplies were classed as 0 with respect to the zinc parameter as they contained concentrations generally < 0.8 mg/ℓ, signiicantly lower than the upper limit for Class 0 supplies (3 mg/ℓ). At levels > 5 mg/ℓ zinc can impart a bitter taste to water and at levels > 10 mg/ℓ may be toxic (Kempster et al., 1997). However, the WHO reports that the concentrations of zinc found in groundwater are usually < 0.05 mg/ℓ (WHO, 2008), marginally lower than the levels detected in this study. Nevertheless, all borehole supplies complied with the Malawi Bureau of Standards maximum permitted levels for iron (3 mg/ℓ), manganese (1.5 mg/ℓ) and zinc (15 mg/ℓ), with the exception of the elevated iron levels detected at Migano (> 3.5 mg/ℓ). No WHO guideline values have been established for iron and zinc.
he main problems associated with iron and manganese in borehole water supplies tends to be aesthetic rather than healthrelated. Problems associated with unsightly loating 'black/ brown bits (particles)' and 'brown staining of white clothes (linen)' during washing were reported by women from diferent villages. his problem is probably brought about by the oxidation of soluble iron, manganese and/or copper to form insoluble precipitates brought about by the mechanical agitation processes involved during hand-washing of clothes (Grimason et al., 2013). Some women stated that they oten used a dishcloth to remove (ilter out) unsightly particles before storing water in clay pots to cool. Similar observations have been noted by Taulo et al. (2008) in villages located around the lakeshore in Malawi. In that study signiicantly high concentrations of faecal coliform (Escherichia coli) and Staphylococcus aureus bacteria were detected in rinsed water from dishcloths. herefore, the action taken to remove an aesthetic non-health related parameter could inadvertently result in the contamination of the water (MIWD, 2002). In light of the indings of this study we suggest that the MIWD consider revising their recommendation and require Government and non-Government providers of borehole water supplies to undertake an analysis of the parameters identiied by Kempster et al. (1997). his would enable the quality of the water to be classiied and determine whether any additional treatment is required to ensure that the water is both wholesome and it for human consumption. On too many occasions boreholes are sunk in Malawi without any consideration of the chemical and bacteriological quality of the supply. he recipient community, grateful for the provision of a borehole, may unknowingly consume water that is unit for human consumption. Nevertheless, it is recognised that water derived from borehole sources in rural Malawi is undoubtedly superior in quality to other protected (e.g. capped wells) and unprotected (shallow wells, rivers) drinking-water sources (Palamuleni, 2002;Pritchard et al., 2007;Mkandawire and Banda, 2010).
We recommend that the criteria for the classiication scheme developed by Kempster et al. (1997) should be updated in light of recent amendments to the South African Water Quality Guidelines for Domestic Use (DWAF, 1996) and WHO recommendations (WHO 2008). It is recommended that any new classiication scheme should be a default-based system whereby the overall class of a borehole defaults to the poorest class based on the health-related parameters. his would provide an incentive for the providers of boreholes to ensure that the water provided is wholesome, it for human consumption and does not constitute a signiicant medical risk to the consumer. Research is required to develop practicable, afordable and sustainable methods to enable villagers to treat Class II/III water supplies and improve the quality of their drinking-water to a class suitable for human consumption.
with a health-related contaminant. he appearance of unsightly discoloured borehole water resulted in some women reverting back to unprotected sources (Grimason et al., 2013).

Arsenic, cadmium and ammonia parameters
Arsenic (MDL 15 µg/ℓ) and cadmium (MDL 1.5 µg/ℓ) were not detected in borehole samples above their minimum detection limits and therefore all borehole supplies fell into Class 0. An analysis of borehole water supplies for ammoniacal nitrogen was not undertaken as part of this study.

CONCLUSION
A review of the literature suggests that this is the irst time that any investigators outside South Africa have utilised the guideline criteria set out by Kempster et al. (1997) to classify rural borehole supplies in Africa. his may be due to the fact that most studies do not entail a comprehensive analysis of water samples for chemical constituents in order to facilitate such classiications. We found the criteria and range of classiications proposed by these investigators extremely useful to classify borehole water supplies as compared with the single guideline values recommended by the WHO for drinking-water (WHO, 2008;WHO, 2011) and maximum permitted levels imposed by the Malawi Bureau of Standards for borehole supplies (MBS, 2005). In addition, Kempster et al. (1997) also include a range of standards for parameters for which no WHO guideline values are currently proposed (e.g., iron, zinc, magnesium, electrical conductivity, total dissolved solids) which enabled the data accrued in this study to be compared and classiied.
In general, borehole water supplies fell mainly within Class 0 and I for the vast majority of water parameters. Borehole supplies which were classiied as II or II were primarily located on the west bank of the Shire River, mainly as a result of elevated levels of both health-related (e.g. nitrate, luoride, faecal coliforms) and non-health related parameters (e.g. iron, manganese and electrical conductivity).
he levels detected indicate that consumption over a short or prolonged period of time (depending upon the parameter) could have an adverse (Class II) or serious (Class III) health efect on the consumer, e.g., skeletal luorosis in adults and methaemoglobinaemia in infants. Of particular concern is the borehole located at Migano as it currently provides Class II water to consumers with respect to the levels of luoride, chloride and manganese detected and Class III water with respect to the signiicantly higher nitrate, sodium, iron and electrical conductivity levels detected. Kempster et al. (1997) recommend that Class II borehole water supplies are only 'suitable for short-term or emergency use only (and) not (for consumption) over a lifetime' and Class III supplies should not be consumed by sensitive groups (e.g., infants, elderly), even over a shortperiod of time, without adequate treatment. However, in reality most villagers will be unaware of the quality of their borehole water supply and will probably consume such water from the cradle to the grave. For many people who reside in rural areas in Malawi a borehole water supply is deemed to be safe if it is clear, odourless and palatable.
In Malawi, the Ministry of Irrigation and Water Development (MIWD) recommend that upon construction of a borehole, and before maintenance and control is oicially handed over to the community, a small number of water quality tests are undertaken (i.e. pH, total dissolved solids, sulphates, nitrates, luoride, chloride, electrical conductivity and iron)