More than 200 million people worldwide are exposed to excessive fluoride in drinking water. According to the World Health Organization, the optimal concentration range of fluoride in drinking water is 0.5 to 1.5 mg/L. Above this range, populations may contract dental fluorosis or, in severe cases, crippling skeletal fluorosis. In the Gokwe area in NW Zimbabwe, where drinking water contains up to 11 mg/L fluoride, fluorosis prevalence has previously been estimated at 62%. This paper investigates the water quality of 126 water sources in Gokwe (58 pumped boreholes, 15 flowing artesian boreholes, 46 wells and 7 streams). The water chemistry, determined from high performance ion chromatography and field measurements, showed that the water source types exhibit significantly different (P < 0.05) concentrations of F, Cl, Na+, K+, Ca2+, Mg2+, temperatures, pH and conductivity values. Thirty-five (28%) of the 126 water sources (18 pumped boreholes, 15 artesian boreholes, one well and one stream) contained F > 1.5 mg/L, indicating that fluoride contamination in the area is a characteristic of deeper groundwater, possibly due to its interactions with the potentially fluoridic coaly and carbonaceous materials of the Lower Karoo Aquifer at depth. The plausibility of providing alternative low fluoride water, and defluoridation, should be investigated.

INTRODUCTION

Water is the principal source of fluoride for humans (World Health Organization 2011). In small doses (0.5–1.5 mg/L), fluoride helps prevent dental caries but it can be toxic at concentrations above 1.5 mg/L (US National Research Council 1993; Fawell et al. 2006; World Health Organization 2011). Human signs of chronic fluoride toxicity vary from dental fluorosis (staining, pitting and brittleness of teeth), through skeletal fluorosis (adverse changes in bone structure) to crippling skeletal fluorosis (crippling bone and joint deformities) as the level and period of exposure increase (US National Research Council 1993; Fawell et al. 2006). It is estimated that several tens of millions of people worldwide suffer from fluorosis, making fluorosis one of the biggest endemic health problems associated with natural geochemistry (Fawell et al. 2006; Fewtrell et al. 2006). In total, it is estimated that about 200 million people are exposed to excessive fluoride in drinking water worldwide (Edmunds & Smedley 2013).

Apart from water, other contributors to fluoride exposure in humans include food, beverages, dental products, fluoride supplements and infant formulas (World Health Organization 2011). In addition, atmospheric exposure to fluoride may occur due to indoor burning of high-fluoride coal and consumption of brick tea as in parts of rural China (World Health Organization 2011). Most fluoride in the environment can ultimately be attributed to the rocks of the Earth's crust and, in particular, to fluoride-bearing minerals such as apatite, fluorite, micas, amphiboles and topaz (Goldschmidt 1962; Koritnig 1978; Owen & Jones 1995; Edmunds & Smedley 2013). Intermediate igneous rocks contain an average of 400 ppm fluoride, whereas acid igneous rocks contain about 800 ppm F (Koritnig 1978) mainly in fluorapatite (Day 1963). Alkaline rocks contain an average of 1,000 ppm F and ultramafic rocks only about 20 ppm F (Koritnig 1978). In metamorphic rocks, fluoride is mainly restricted to skarns and metasomatic aureoles around granitic plutons. Although fluoride is the most abundant halogen in sedimentary rocks (Koritnig 1978), it is relatively rare except in areas of specific mineralisation (Edmunds 1995). Fluoride is generally more abundant in sediments of marine origin than those originating from non-marine environments (Koritnig 1978; Edmunds 1995). In coal, fluoride is mainly associated with mineral matter, rather than the organic component (Schultz et al. 1973), with fluorapatite being the dominant fluoride-bearing mineral (Bouska 1981).

The presence of excessive fluoride in the sedimentary Lower Karoo Aquifer (LKA) in Gokwe, and in aquifers in other districts in Zimbabwe, was first noticed in routine water quality assessments. It was subsequently documented by Tobayiwa et al. (1991) and Mamuse (2003). Tobayiwa et al. (1991) estimated that fluorosis prevalence in an area, which falls within the current study area (Figure 1), was 62%. This paper explores the fluoride problem in the Gokwe area through hitherto unpublished statistical analyses of key results obtained by Mamuse (2003).
Figure 1

The study area. Subdivisions 1–3 refer to geological mapping as follows. 1 = mapped by Ait-Kaci Ahmed (in preparation), 2 = mapped by Leyshon (1969), and 3 = no published mapping.

Figure 1

The study area. Subdivisions 1–3 refer to geological mapping as follows. 1 = mapped by Ait-Kaci Ahmed (in preparation), 2 = mapped by Leyshon (1969), and 3 = no published mapping.

STUDY AREA

This study was carried out within an area measuring 6,000 km2 straddling Gokwe South and Gokwe North districts of Zimbabwe, some 300 km west of Harare, Zimbabwe (Figure 1). The area lies within the Sanyati-Sengwa Basin of the Zambezi Valley Geomorphic Province (Lister 1987) and experiences a warm to hot climate, with a mean annual temperature of 23 °C and an annual rainfall of 650–700 mm. Due to these semi-arid conditions, surface water supplies are limited and, consequently, the Gokwe area is heavily reliant upon groundwater. Aquifer materials and processes in many parts of the study area are such that groundwater is saline and/or contains excessive fluoride.

In Figure 1, the study area is divided into four segments, two of which (labelled 3) currently have no published geological map. The central portion of the study area (labelled 1) was mapped by Ait-Kaci Ahmed (in preparation) and the eastern part by Leyshon (1969). The geology of the area consists of rocks of the Karoo Supergroup (Lower and Upper Karoo groups) that were deposited in the Sengwa Sub-basin of the Mid-Zambezi Basin. The Upper Karoo is sub-horizontal and rests unconformably upon the Lower Karoo strata that dip and young to the northwest (Lepper 1992; Ait-Kaci Ahmed in preparation). Typical lithostratigraphic logs for the area, whose locations are indicated in Figure 2, are depicted in Figure 3 and show that the Upper Karoo is absent at the logged sites. The Lower Karoo consists of the Dwyka, Wankie and Madumabisa formations. The Wankie Formation is made up of the Lower Wankie Member, Middle Wankie Member and Upper Wankie Sandstone Member. Similarly, lower, middle and upper members of the Madumabisa Formation are recognised, but the Upper Madumabisa Member is absent in the study area. The Escarpment, Pebbly Arkose, Forest Sandstone and the Batoka Basalt formations constitute the Upper Karoo (Ait-Kaci Ahmed in preparation). Hydrogeologically, the Wankie Formation constitutes what is collectively known as the LKA, with the Madumabisa Formation being the upper confining layer (aquiclude). Mamuse (2003) suggested that carbonaceous materials (carbonaceous mudstone, carbonaceous shale and sub-bituminous to high-volatile bituminous coal) dominantly within the Lower Madumabisa Member and Middle Wankie Member likely constitute the source of fluoride contamination of water supplies in the area.
Figure 2

Geology of the central part (subdivision 1 on Figure 1) of the study area, after Ait-Kaci Ahmed (in preparation).

Figure 2

Geology of the central part (subdivision 1 on Figure 1) of the study area, after Ait-Kaci Ahmed (in preparation).

Figure 3

Lithostratigraphic section of the study area based on logs of boreholes at sites shown in Figure 2. Modified from Mamuse (2003).

Figure 3

Lithostratigraphic section of the study area based on logs of boreholes at sites shown in Figure 2. Modified from Mamuse (2003).

Leyshon (1969) mapped granitic and gneissic inliers, considered part of the Archaean Zimbabwe craton, in the eastern part of the study area (segment 2 in Figure 1) where they emerge through the cover of Karoo rocks.

METHODS

Sampling and sample handling

This study is based on samples collected from 126 water sources in the Gokwe area, NW Zimbabwe between May and August 2002. Of these, 7 (5.6%) were collected from streams, 46 (36.5%) from dug wells, 58 (46.0%) from pumped boreholes and 15 (11.9%) from flowing artesian boreholes.

For each water source, two sub-samples were collected: (i) samples for well-head measurements, which were collected directly into clean plastic beakers, and (ii) samples for laboratory measurements, collected into clean high density polyethylene bottles. The laboratory samples were filtered through 0.45 μm membrane filters and refrigerated. On shipment to Curtin University, Australia for analysis, the sample bottles were packed into polystyrene insulating cartons and were refrigerated on arrival. Actual sample collection depended on type of water source. For pumped boreholes, pre-sampling water pumping was done to clean pipe fixtures and to ensure that aquifer water, not water standing in pipes, was sampled. Approximately three times the volume of borehole pipes was expelled this way, a routine which was not necessary for flowing artesian boreholes. Samples from open wells were collected manually with a clean plastic bailer. Flowing streams were sampled directly with clean plastic beakers for well-head measurements and filtration.

Sample analysis and statistical data analysis

Well-head measurements of water temperature, pH, conductivity and total dissolved solids (TDS) were performed using appropriate probes and read directly off a meter. Concentrations of F and those of major anions (Cl-, SO42−) and cations (Na+, K+, Ca2+ and Mg2+) were obtained using a Dionex® 1000 Ion Chromatograph in the laboratory. Anions were separated using a AS-12A separator column with AG-14A guard column and a (3.5 mM Na2CO3/ 1 mM NaHCO3) eluent at a flow-rate of 1 mL/min. Cations were separated using a CS-12A separator column and 30 mM methanosulphonic acid eluent at a flow rate of 1 mL/min. Sample volume for ion chromatographic analysis was 25 μL and ions were detected by conductivity, with a lower limit of determination of around 0.1 mg/L.

Results of chemical analyses of samples were subjected to statistical analyses using the software IBM® SPSS® Statistics (‘SPSS’) and, to a lesser extent, Microsoft® Excel® software. In order to proceed meaningfully with statistical analyses of the chemical compositions of the samples, it was necessary to determine if the chemistry of water from different source types was statistically different. Analysis of variance (ANOVA) was carried out to test the null hypothesis that water from streams, wells, pumped boreholes and artesian boreholes belonged to the same population.

RESULTS

Table 1 shows water chemistry results for 28 of the 126 water sources sampled in this study (seven most fluoride-contaminated sources for each of the four water source types). Fluoride contamination (F > 1.5) is dominantly associated with deeper water source types (flowing artesian boreholes and pumped boreholes), rather than the shallower water source types (wells and streams). Another observation is that all high-fluoride water sources (>1.5 mg/L F) in this study are restricted to neutral (pH 6.5–7.8), and moderately alkaline environments (pH 7.8–9) of Hounslow (1995). The association between alkaline conditions with high fluoride content has been explained in terms of fluoride desorption from kaolinite in reverse anion exchange that occurs in this environment (Hounslow 1995; Ming-Ho 2000).

Table 1

Water chemistry of the seven most fluoride-contaminated water sources of each of the four water source types sampled in this study

  Concentrations in mg/L
 
Other Measurements
 
Sample ID F Cl SO42− Na+ K+ Mg2+ Ca2+ TDS1 T (°C) pH EC2 (mS cm−1
Flowing Artesian Boreholes 
w005 11.0 32.2 298.4 318.4 3.1 2.1 4.3 663 29.3 8.2 1.3 
w043 7.2 74.7 116.4 359.7 5.8 1.5 5.4 1,140 34.1 7.4 2.3 
w097 6.7 96.4 1,199.6 2,089.3 3.8 2.4 7.5 1,243 32 8.1 2.5 
w029 6.1 28.4 50.6 388.6 4.2 2.7 0.5 790 28.5 8.0 1.6 
w044 5.7 289.9 6.1 639.9 7.7 3.8 15.9 2,680 35.1 7.1 5.4 
w041 4.7 14.1 138.3 152.1 2.2 0.8 1.9 590 36.2 8.3 1.2 
w059 4.4 607.7 2.4 478.4 5.2 3.0 12.1 3,210 31.8 7.0 6.4 
Pumped Boreholes 
w126 9.8 118.9 236.2 520.5 7.4 11.4 7.1 1,155 26.7 7.4 2.3 
w017 8.3 22.2 73.2 299.3 3.5 2.5 5.6 802 26.7 8.2 1.6 
w006 8.1 29.3 102.5 311.2 3.1 2.1 4.0 688 29.1 8.3 1.5 
w004 7.6 102.2 298.4 39.8 0.3 0.4 0.1 748 26.8 8.2 1.5 
w145 7.6 39.6 244.2 168.4 4.9 1.1 2.5 884 29.7 8.4 1.8 
w151 6.7 96.4 1,199.6 40.6 2.2 29.2 38.2 447 27.3 6.6 0.9 
w096 6.0 111.2 0.2 240.2 4.5 3.0 4.4 1,297 21.6 8.6 2.6 
Wells 
w117 5.9 167.1 147.6 276.1 2.7 27.1 3.6 896 26.4 6.9 1.8 
w149 1.4 12.1 1,001.0 319.8 1.2 122.1 71.4 1,319 20.6 6.9 2.6 
w116 1.1 13.5 15.6 19.6 1.6 20.9 13.5 260 25.8 6.8 0.5 
w056 0.8 315.8 8.8 794.1 10.8 11.6 27.7 3,480 23.1 7.2 6.9 
w035 0.7 4.0 632.5 61.3 19.7 32.3 121.5 770 19.1 7.5 1.5 
w221 0.7 9.2 2.2 12.0 5.3 1.9 12.1 82 21.4 7.4 0.2 
w095 0.6 3.8 294.9 69.0 12.3 29.5 20.4 464 25.2 7.0 0.9 
Streams 
w255 2.9 2.8 69.7 11.8 2.8 24.1 48.4 284 21.5 7.3 0.6 
w100 0.5 8.2 504.8 288.2 17.3 41.3 8.4 720 23.8 7.5 1.4 
w092 0.4 6.4 118.0 89.1 5.8 33.4 12.3 546 19.5 8.3 1.1 
w088 0.4 11.4 272.2 99.2 20.2 62.9 71.5 816 22.9 7.9 1.6 
w053 0.2 3.9 16.2 69.6 3.8 16.8 23.1 310 25.8 7.2 0.6 
w089 0.2 0.1 14.6 29.9 4.2 20.1 10.5 264 23.3 8.0 0.5 
w064 0.1 59.9 0.5 75.2 15.4 4.0 6.1 229 21.6 7.5 0.4 
  Concentrations in mg/L
 
Other Measurements
 
Sample ID F Cl SO42− Na+ K+ Mg2+ Ca2+ TDS1 T (°C) pH EC2 (mS cm−1
Flowing Artesian Boreholes 
w005 11.0 32.2 298.4 318.4 3.1 2.1 4.3 663 29.3 8.2 1.3 
w043 7.2 74.7 116.4 359.7 5.8 1.5 5.4 1,140 34.1 7.4 2.3 
w097 6.7 96.4 1,199.6 2,089.3 3.8 2.4 7.5 1,243 32 8.1 2.5 
w029 6.1 28.4 50.6 388.6 4.2 2.7 0.5 790 28.5 8.0 1.6 
w044 5.7 289.9 6.1 639.9 7.7 3.8 15.9 2,680 35.1 7.1 5.4 
w041 4.7 14.1 138.3 152.1 2.2 0.8 1.9 590 36.2 8.3 1.2 
w059 4.4 607.7 2.4 478.4 5.2 3.0 12.1 3,210 31.8 7.0 6.4 
Pumped Boreholes 
w126 9.8 118.9 236.2 520.5 7.4 11.4 7.1 1,155 26.7 7.4 2.3 
w017 8.3 22.2 73.2 299.3 3.5 2.5 5.6 802 26.7 8.2 1.6 
w006 8.1 29.3 102.5 311.2 3.1 2.1 4.0 688 29.1 8.3 1.5 
w004 7.6 102.2 298.4 39.8 0.3 0.4 0.1 748 26.8 8.2 1.5 
w145 7.6 39.6 244.2 168.4 4.9 1.1 2.5 884 29.7 8.4 1.8 
w151 6.7 96.4 1,199.6 40.6 2.2 29.2 38.2 447 27.3 6.6 0.9 
w096 6.0 111.2 0.2 240.2 4.5 3.0 4.4 1,297 21.6 8.6 2.6 
Wells 
w117 5.9 167.1 147.6 276.1 2.7 27.1 3.6 896 26.4 6.9 1.8 
w149 1.4 12.1 1,001.0 319.8 1.2 122.1 71.4 1,319 20.6 6.9 2.6 
w116 1.1 13.5 15.6 19.6 1.6 20.9 13.5 260 25.8 6.8 0.5 
w056 0.8 315.8 8.8 794.1 10.8 11.6 27.7 3,480 23.1 7.2 6.9 
w035 0.7 4.0 632.5 61.3 19.7 32.3 121.5 770 19.1 7.5 1.5 
w221 0.7 9.2 2.2 12.0 5.3 1.9 12.1 82 21.4 7.4 0.2 
w095 0.6 3.8 294.9 69.0 12.3 29.5 20.4 464 25.2 7.0 0.9 
Streams 
w255 2.9 2.8 69.7 11.8 2.8 24.1 48.4 284 21.5 7.3 0.6 
w100 0.5 8.2 504.8 288.2 17.3 41.3 8.4 720 23.8 7.5 1.4 
w092 0.4 6.4 118.0 89.1 5.8 33.4 12.3 546 19.5 8.3 1.1 
w088 0.4 11.4 272.2 99.2 20.2 62.9 71.5 816 22.9 7.9 1.6 
w053 0.2 3.9 16.2 69.6 3.8 16.8 23.1 310 25.8 7.2 0.6 
w089 0.2 0.1 14.6 29.9 4.2 20.1 10.5 264 23.3 8.0 0.5 
w064 0.1 59.9 0.5 75.2 15.4 4.0 6.1 229 21.6 7.5 0.4 

1TDS stands for total dissolved solids.

2EC stands for electrical conductivity.

More detailed statistical analyses of the water chemistry results are presented below.

Statistical peculiarity of water chemistry among water source types

A key assumption of the conventional one-way ANOVA is that variances among groups are homogenous (i.e. groups have approximately equal variances). However, Levene's F statistic showed that variances with respect to F, K+, Mg2+, and temperature among the water sources were significantly different, thus precluding the use of conventional ANOVA. Of the variables that passed Levene's test (Cl, SO42−, Na+, Ca2+, conductivity, TDS and pH), only Na+ and pH exhibited statistically significant differences among the different water sources at the 5% level.

F, K+, Mg2+, and temperature were subjected to the Welch ANOVA which is applied where the assumption of equal variances is not met. Only F and temperature values had statistically significant differences (P < 0.01) across the water source types. The pair-wise Games-Howell test shows that the only pairs of water source types that do not exhibit statistically significant differences in fluoride content are streams and wells, and streams and boreholes. In particular, the differences in fluoride content between artesian boreholes and any of the other three water source types are statistically significant at the 5% level.

Correlations of water chemistry variables within water source types

Correlations among water quality variables may suggest useful geochemical associations which may help identify geochemical environments associated with species of interest, in this case fluoride. For example, it is known that a negative correlation exists between fluoride and calcium in groundwaters (Handa 1975; Hem 1985; Edmunds & Smedley 1996). More broadly, groundwater geochemical correlations have aided geochemical source-rock deductions. As the four water source types in this study exhibit statistically significant chemical differences, correlations among the water chemistry variables were considered separately for each water source type. Statistically significant correlations found across water types in this study are between conductivity and TDS, Ca2+ and Mg2+, conductivity (and TDS) and Na+ or Cl, which are generally expected.

Statistically significant correlations between Na+ and SO42− are exhibited only by samples from artesian boreholes and streams. Significant correlations involving F are exhibited only by pumped boreholes: conductivity-F, Mg2+-F, pH-F and SO42−_F. Ca2+ and Cl are significantly correlated in water collected from wells and pumped boreholes.

Fluoride concentrations by water source types

The differences in F concentrations among the four water sources in this study are statistically significant. As shown in Figure 4, 16 water sources (13%) contain low fluoride (0–0.5 mg/L F), 75 (60%) optimal fluoride (0.5–1.5 mg/L F) and 35 (28%) high fluoride (>1.5 mg/L F). The figure shows that of the 35 high fluoride water sources in this study, 18 are pumped boreholes, 15 are artesian boreholes, one is a well and another one is a stream. Another piece of information in Figure 5 is the proportion fluoride content categories among water sources in each water source type. For example, all the 15 artesian boreholes (100%) are high fluoride water sources, whereas of the 58 pumped boreholes there are 18 (31%) in the high fluoride category, 34 (59%) in the optimal category and 6 (10%) in the low category. The spatial distribution of water sources with low, optimal and high fluoride concentrations is shown in Figure 5.
Figure 4

Fluoride content categories of water sources within each water source type. The numbers of water sources in each fluoride category are indicated on the bars.

Figure 4

Fluoride content categories of water sources within each water source type. The numbers of water sources in each fluoride category are indicated on the bars.

Figure 5

Distribution of waters sources with low, optimal and high fluoride concentrations by World Health Organization (2011) guidelines. The background imagery is Shuttle Radar Topography Mission from ESRI (2009).

Figure 5

Distribution of waters sources with low, optimal and high fluoride concentrations by World Health Organization (2011) guidelines. The background imagery is Shuttle Radar Topography Mission from ESRI (2009).

Concentrations of NA+, CL, SO42− and TDS, and PH

The World Health Organization currently (World Health Organization 2011) provides no health-based guideline values for concentrations of Cl, SO42−, Na+ and TDS in drinking water, but has some taste-based aesthetic guidelines. In this study, water from several water sources had concentrations of Cl, SO42−, Na+ and TDS above the World Health Organization (2011) aesthetic guideline values (Table 2). Using the general classification of salinity based on TDS where freshwater contains 0–1,000 mg/L TDS, brackish water 1,000–10,000 mg/L, saline water 10,000–100,000 mg/L and brine (>100,000 mg/L), 106 of the 126 water sources sampled issued fresh water. Of the 20 non-fresh water sources, 19 are brackish water sources and one (w198; TDS 14,750 mg/L) is saline.

Table 2

Concentrations of selected water quality variables in the study in comparison with World Health Organization (2011) guideline values

Variable Range (mg/L) WHO guideline value (mg/L) Water sources exceeding WHO guideline Remarks 
Cl 0–8,748 250 12 Cl > 250 mg/L can give objectionable taste and may increase metal corrosion rates resulting in increased metal concentrations in the supply (World Health Organization 2011
SO42− 0–1,200 500 May cause detectable taste and at >500 mg/L, health officials must be notified (World Health Organization 2011) due to possible gastro-intestinal effects. 
Na+ 0–3,468 200 31 Na > 200 mg/L may cause unacceptable taste (World Health Organization 2011). DEPa (2001) notes that Na > 160 mg/L aggravates sodium sensitive hypertension and diseases that cause difficulty in regulating body fluid volumes. 
TDS 1–14,750 1,000 20 Health-TDS relationship not known, but TDS > 1,000 mg/L causes objectionable taste in drinking water (World Health Organization 2011). 
Variable Range (mg/L) WHO guideline value (mg/L) Water sources exceeding WHO guideline Remarks 
Cl 0–8,748 250 12 Cl > 250 mg/L can give objectionable taste and may increase metal corrosion rates resulting in increased metal concentrations in the supply (World Health Organization 2011
SO42− 0–1,200 500 May cause detectable taste and at >500 mg/L, health officials must be notified (World Health Organization 2011) due to possible gastro-intestinal effects. 
Na+ 0–3,468 200 31 Na > 200 mg/L may cause unacceptable taste (World Health Organization 2011). DEPa (2001) notes that Na > 160 mg/L aggravates sodium sensitive hypertension and diseases that cause difficulty in regulating body fluid volumes. 
TDS 1–14,750 1,000 20 Health-TDS relationship not known, but TDS > 1,000 mg/L causes objectionable taste in drinking water (World Health Organization 2011). 

aDEP – Department for Environmental Protection.

DISCUSSION

A previous study by Tobayiwa et al. (1991) found a 62% prevalence in dental fluorosis among school children in a portion of the current study area. In the current study, 35% of the 126 water sources sampled contained excessive fluoride which, in places, was sufficiently high to cause skeletal fluorosis. On the other extreme, some water sources (13%) issued fluoride-deficient water (<0.5% F), depriving communities of the dental benefits of fluoride in preventing dental caries. The remaining 75 (59%) water sources issued optimal-fluoride (0.5–1.0 mg/L F) water, thus potentially protecting communities against dental caries.

Thirty-three (94%) of the high fluoride water sources are pumped boreholes (18 of 58) and artesian boreholes (15 of 15), indicating that high fluoride levels are associated with deeper groundwater from the main LKA. Mamuse (2003) suggested that the fluoride contamination of the LKA emanates from carbonaceous and coaly lithologic units of the Lower Madumabisa Member and the Middle Wankie Member (Figures 2 and 3). This coaly or carbonaceous material consists of carbonaceous mudstone, carbonaceous shale and sub-bituminous to high-volatile bituminous coal, with the coal occurring as seams or dispersed lenses within the mudstones and shales (Lepper 1992). As the LKA is the main aquifer in this area which heavily relies upon groundwater, the plausibility of fluoride removal (defluoridation) should be investigated. New innovative defluoridation methods, such as the use of nanotechnology, should be explored because traditional fluoride removal methods, dominantly precipitation and sorption methods, have proved to be unsustainable, ineffective, or both, especially in rural areas in developing countries (Edmunds & Smedley 2013). This drawback has led to the general recommendation that in fluoride contaminated areas, provision of alternative low-fluoride water must be investigated as the first option (World Health Organization 2011; Edmunds & Smedley 2013). Consistent with this, the use of shallower groundwater above fluoretic aquifer materials, where and when available, should be encouraged in the Gokwe area. Thus there is need for prescription of well-depths, casing off of contaminant units, and regular water quality testing. The plausibility of applying other measures recommended elsewhere, such as rainwater harvesting and water importation (Edmunds & Smedley 2013), can be explored in the Gokwe area.

The potential health implications of fluoride and other water contaminants in the study area should be further explored so that holistic interventions, taking into account all health considerations, can be formulated. For example, concentrations of Cl, SO42−, Na+ and TDS are higher than World Health Organization (2011) taste-based aesthetic guidelines in many water sources. Although currently no WHO health guideline values exist for these variables (World Health Organization 2011), some authorities (e.g. Department of Environmental Protection 2001; Environmental Health Directorate 2012) caution communities on the potential aggravation of some conditions, such as aggravation of sodium-sensitive hypertension by exposure to >20 mg/L Na+ in drinking water. Another area of concern is microbial contamination, which is generally more common in shallower water sources.

ACKNOWLEDGEMENTS

This paper emanates from a Curtin University, Australia MSc research project by AM supervised by RW and financed through the Australian Agency for International Development scholarship. The Ministry of Mines, Zimbabwe provided financial and fieldwork support. Special thanks to Ministry personnel Forbes Mugumbate for proposing the project, Ait-Kaci Ahmed for field geological orientation, and Warren Makamure for field assistance.

REFERENCES

REFERENCES
Ait-Kaci Ahmed
A.
(in preparation)
The ecology of the area north of Gokwe, Gokwe and Kadoma districts, Zimbabwe Geological Survey Bulletin 105, Geological Survey Department, Harare
, p.
85
.
Bouska
V.
1981
Geochemistry of Coal.
Elsevier
,
New York
,
284
pp.
Day
F. H.
1963
The Chemical Elements in Nature
.
George Harrap & Co
,
London
, p.
372
.
Department of Environmental Protection
2001
Drinking water inorganic contaminants, DEP, Florida, USA. Web link: http://www.dep.state.fl.us/water/drinkingwater/inorg_con.htm.
Edmunds
W. M.
1995
Characterisation of groundwaters in semi-arid and arid zones using minor and major elements
. In:
Groundwater Quality
.
Nash
H.
McCall
G. J. H.
(eds),
Chapman & Hall
,
London
, pp.
19
30
.
Edmunds
W. M.
Smedley
P. L.
1996
Groundwater geochemistry and health: an overview
. In:
Environmental Geochemistry and Health
.
Appleton
J. D.
Fuge
R.
McCall
G. J. H.
(eds),
Geological Society Special Publication
,
13
, pp.
91
105
.
Edmunds
W. M.
Smedley
P. L.
2013
Fluoride in natural waters
. In:
Essentials of Medical Geology
.
Selinus
O.
(ed.),
Springer
,
Amsterdam
, pp.
311
336
.
Environmental Health Directorate
2012
Sodium in drinking water, Department of Health, Western Australia. http://www.nhmrc.gov.au/guidelines/publications/eh34.
ESRI
2009
ESRI® Data & Maps 9.3.1 Global imagery and shaded relief. Europe and Africa (DVD)
.
Fawell
J.
Bailey
K.
Chilton
J.
Dahi
E.
Fewtrell
L.
Magara
Y.
2006
Fluoride in Drinking Water
.
World Health Organization
,
Geneva
, p.
134
.
Fewtrell
L.
Smith
S.
Kay
D.
Bartram
J.
2006
An attempt to estimate the global burden of disease due to fluoride in drinking water
.
Journal of Water and Health
4
,
533
542
.
Goldschmidt
V. M.
1962
Geochemistry
.
Clarendon Press
,
Oxford, UK
,
730
pp.
Hem
J. D.
1985
Study and Interpretation of the Chemical Characteristics of Natural Water, US Geological Survey, Water-Supply Paper 2254
.
Hounslow
A. W.
1995
Water Quality Data: Analysis and Interpretation
.
CRC Press Inc.
,
New York
, p.
397
.
Koritnig
S.
1978
Fluorine
. In:
Handbook of Geochemistry
.
Wedepohl
K. H.
Correns
C. W.
Shaw
D. M.
Turekian
K. K.
Zemann
J.
(eds),
vol. II/1
,
Springer Verlag
,
Berlin
, pp.
9-C-1
9-O-3
.
Lepper
J.
1992
The Lower Karoo in the Mid-Zambezi Basin (Zimbabwe)
.
Geologishes Jabhrbuch
,
Hannover
, p.
38
.
Leyshon
P. R.
1969
The geology of the country around the Copper Queen. Rhodesia Geological Survey Bulletin 66. Geological Survey Department, Salisbury
, p.
147
.
Lister
L. A.
1987
Erosion surfaces of Zimbabwe. Zimbabwe Geological Survey Bulletin 95. Zimbabwe Geological Survey, Harare.
Mamuse
A.
2003
Fluoride Contaminated Drinking Water in Gokwe District (NW Zimbabwe): Spatial Distribution, Lithostratigraphic Controls and Implications for Human Health. MSc. Thesis, Curtin University of Technology, Western Australia. http://adt.curtin.edu.au/theses/available/adt-WCU20041004.123924/.
Ming-Ho
Y.
2000
Environmental Toxicology: Impacts of Environmental Toxicants on Living Systems
.
CRC
,
Florida
,
USA
, p.
255
.
Owen
P.
Jones
F. B.
1995
Main factors affecting the composition of natural waters
. In:
Trace Elements in Natural Waters
.
Salbu
B.
Steinnes
E.
(eds),
CRC Press
,
Tokyo
, pp.
1
20
.
Schultz
H.
Hattman
E. A.
Booher
W. B.
1973
The fate of some trace elements during coal pre-treatment and combustion
.
P. S. Babu (ed.). Trace elements in fuel. Advances in Chemistry Series 141, 139–153, American Chemical Society, USA
.
Tobayiwa
C.
Musiyambiri
M.
Chironga
I.
Mazorodze
O.
Sapahla
S.
1991
Fluoride levels in two districts in Zimbabwe
.
Central African Journal of Medicine
37
,
353
361
.
US National Research Council
1993
Health Effects of Ingested Fluoride
.
National Academy Press
,
Washington
, p.
181
.
World Health Organization
2011
Guidelines for Drinking-water Quality
.
4th edn.
,
WHO
,
Geneva
, p.
541
.