Groundwater for Sustainable Development 10 (2020) 100296 Available online 4 November 2019 2352-801X/© 2019 Elsevier B.V. All rights reserved. Research paper Assessment of groundwater quality and the main controls on its hydrochemistry in some Voltaian and basement aquifers, northern Ghana Yvonne Sena Akosua Loh a, Bismark Awinbire Akurugu a,b,*, Evans Manu b, Abdul-Samed Aliou a a Department of Earth Science, University of Ghana, Legon, Accra, Ghana b Council for Scientific and Industrial Research -Water Research Institute - (CSIR-WRI), Accra, Ghana A R T I C L E I N F O Keywords: Sawla-Tuna-Kalba Hydrochemical analysis Groundwater quality Geostatistics A B S T R A C T Groundwater resources play the single most important role in the delivery of potable water to rural communities in northern Ghana, especially during the long dry season and where surface water sources are polluted or non- existent. This study sought to assess the quality and main controls on groundwater chemistry in parts of Sawla- Tuna-Kalba District in the Savannah Region of Ghana. Multivariate statistical analysis and conventional hydrochemical plots were employed in the analysis of 112 groundwater samples from the study area. Conven- tional graphical methods, R-mode Hierarchical Cluster Analysis (HCA) and Principal Component Analysis (PCA) identified dissolution of silicates and the influence of agrochemicals and domestic wastewaters as the main sources of variations in the hydrochemistry in the study area. Q-mode HCA coupled with Stiff diagrams identified Ca–HCO3 water type in recharge areas, and Mg–Ca–HCO3 water type, which evolves into a Ca–Na–K–HCO3 water type in discharge areas in the groundwater flow regime. Mineral stability diagrams indicate the groundwater is stable in kaolinite, which suggests little or no restricted groundwater flow conditions. Groundwater quality for domestic purposes was assessed using the weighted arithmetic index approach. The computed water quality indices (WQIs) from the data suggest that 94% of the sampled boreholes provide groundwater of “excellent” quality for drinking purposes, whereas 5% and 1% present water of “good” and “poor” quality respectively. Spatial interpolation of the estimated WQIs suggests the quality of the groundwater in the study area is suitable for domestic purposes. The assessment of the groundwater quality for irrigation purposes suggests the water is of “excellent” to “permissible” quality and may be used for irrigation without prior treatment. 1. Introduction Groundwater is being adopted as a preferable source of potable water supply in arid and semi-arid climates. This is due to its minimal treatment requirements, spatial availability and capacity to balance large rainfall variability as well as associated water demands during droughts and when surface water resources exceed their sustainable levels (Treidel et al., 2012; Bloomfield et al., 2019). A comprehensive groundwater assessment must reflect the several factors and processes influencing groundwater chemistry in space and time, and is prerequi- site to fully understanding the groundwater system for a proper man- agement of the resource. In this regard, researchers have adopted several strategies to assess the several sources of variation in groundwater chemistry. Most of these studies are well documented in literature (e.g. Saana et al., 2016; Hwang et al., 2017; Telahigue et al., 2018; Boateng et al., 2019; Sunkari et al., 2019). Advanced geostatistical techniques for instance have been at the forefront in this regard in recent times, especially in hydrochemical studies where they have been used to characterise groundwater flow regimes, discriminate hydrochemical facies, make predictions at unsampled locations, generate statistical models, identify sources of variations and characterise groundwater evolution (Adomako, 2011; Jing and Yufei, 2011; Kumar et al., 2011; Yidana et al., 2011, 2012a; Nur et al., 2012; Sarukkalige, 2012; Hassan, 2014; Sharma et al., 2015; Abanyie et al., 2018; Sunkari et al., 2019). In an attempt to assess the main sources of hydrochemical variation and suitability of groundwater for drinking purposes, Kumar et al. (2011) employed factor analysis, hierarchical cluster analysis (HCA) and geo- spatial techniques on major physicochemical parameters in the Palar river basin, India. Their study concluded that effluent discharge in the Palar river basin degraded groundwater quality in the northeast and southeast parts of the river basin and groundwater along such areas were unsuitable for drinking. HCA when combined with conventional * Corresponding author. Department of Earth Science, University of Ghana, Legon, Accra, Ghana. Tel: þ233202900233. E-mail address: bismarkakurugu@yahoo.com (B.A. Akurugu). Contents lists available at ScienceDirect Groundwater for Sustainable Development journal homepage: http://www.elsevier.com/locate/gsd https://doi.org/10.1016/j.gsd.2019.100296 Received 25 May 2019; Received in revised form 14 September 2019; Accepted 30 October 2019 mailto:bismarkakurugu@yahoo.com www.sciencedirect.com/science/journal/2352801X https://http://www.elsevier.com/locate/gsd https://doi.org/10.1016/j.gsd.2019.100296 https://doi.org/10.1016/j.gsd.2019.100296 https://doi.org/10.1016/j.gsd.2019.100296 http://crossmark.crossref.org/dialog/?doi=10.1016/j.gsd.2019.100296&domain=pdf Groundwater for Sustainable Development 10 (2020) 100296 2 graphical techniques can help explain the evolution of groundwater and the minerals dissolved in it as it moves through the rock matrix. Belkhiri et al. (2011) demonstrated this when they employed Q-mode HCA to distinguish recharge zones from discharge areas in the groundwater flow regime in the Ain Azel plain, Algeria. Geostatistical techniques offer a wide array of powerful statistical models and tools to effectively explore and analyse hydrochemical data. However, employing these techniques requires an in-depth understanding of the prevailing environmental conditions in the study area such as geology, hydrogeology, topography, weather conditions and dominant anthropogenic activities. Fig. 1. Location map of the study area. Fig. 2. Geological map of the study area showing sampled points. Y.S.A. Loh et al. Groundwater for Sustainable Development 10 (2020) 100296 3 Geostatistical techniques have been employed extensively in groundwater studies in Ghana. For instance, with the aim of unveiling the key factors influencing fluoride concentrations in some parts of Northern Region Ghana, Yidana et al. (2012a) utilised principal component analysis (PCA) coupled with cluster analysis to explain the hydrochemistry and factors influencing fluoride enrichment and other ions of the groundwater in the middle Voltaian aquifers. They also adequately distinguished the various facies in the groundwater flow regime using Q-mode HCA. Furthermore, Sunkari et al. (2019) com- bined multivariate statistics and mass balance techniques to assess the drivers of groundwater chemistry in Ga West, Ghana. Their study revealed silicate and carbonate weathering, seawater intrusion and anthropogenic activities as the main controls on groundwater chemistry in the district. Other similar studies in Ghana are well documented in Helstrup et al. (2007); Banoeng-Yakubo et al. (2009); Yidana et al. (2010); Loh et al. (2016); Abanyie et al. (2018). It follows therefore, that when applied properly, with adequate knowledge of the terrain, geostatistical techniques can be used to satisfactorily characterise groundwater systems. This study adopts similar methodology in addition to conventional and mass balance hydrochemical models to unveil the relationships among water param- eters and the main influence on groundwater chemistry from the base- ment aquifers in the Sawla-Tuna-Kalba District in northern Ghana as well as assesses its suitability for domestic and irrigation purposes. 2. Study area The study area (Fig. 1) is located in the western part of the Savannah Region of Ghana. It lies between latitudes 8�400 and 9�400 north and longitudes 1�500 and 2�450 west. The study area shares common boundaries with Wa West and Wa East Districts to the North, Bole Dis- trict to the South, West Gonja and North Gonja Districts to the East, and La Cote d’Ivoire and Burkina Faso to the West. It has a total land area of about 4226.9 km2 and a population of 99,863 (Ghana Statistical Service, 2010). Generally, the topography of the study area is undulating, with elevation ranging between 245 m and 350 m. The Black Volta River runs through the study area along the western border and drains the area through several tributaries and streams. The area lies within the Tropical continental climate, and generally is one of the driest parts of the country. It experiences a single rainfall season that occurs between early May and late October, with the highest rainfall experienced between August and September (Dickson and Benneh, 1995). Relative humidity is high during the rainy season (65–85%) but may fall to as low as 20% during the dry season. Monthly mean rainfall ranges between 200 mm and 300 mm, with an annual average of about 1100 mm (Sawla-Tuna-- Kalba District Assembly, 2014). 2.1. Geology and hydrogeology The study area is principally underlain by about 80% Birimian rocks with the other 20% being granitoids and Kwahu-Morago Group of the Voltaian Supergroup (Fig. 2). The Birimian consists of metamorphosed volcanic and sedimentary rocks, which form five subparallel belts of volcanic rock separated by broad “basins” of sedimentary rocks (Kesse, 1985). Rocks of the Birimian Supergroup have been divided into met- asedimentary and metavolcanic rocks (Junner, 1935, 1940; Bates, 1955). The Birimian metasedimentary are made up of greywackes with turbidite features, phyllites, slates, schists, weakly metamorphosed tuffs and sandstones whiles the Birimian metavolcanics comprise rock types such as lava flows and dyke rocks of basaltic and andesitic composition. Most of these rocks have now been metamorphosed to hornblende actinolite-schists, calcareous chlorite schists and amphibolites (Kesse, 1985). The Birimian Supergroup is intruded by granitoids of the Ebur- nean and ‘Tamnean’ Plutonic Suites. In Ghana, four main types of granitoids are recognized to be associated with the Birimian Supergroup. These include; the Winneba, Bongo, Dixcove and the Cape Coast type granites. The Cape Coast and Dixcove granites are seen to be present in the current study area. The Cape Coast type granitoids occur only within the Birimian metasediments. This group also includes gneisses, which are especially well developed in the metasediments (Banoeng-Yakubu et al., 2011). Dixcove-type granitoids are metal- uminous and typically dioritic to granodioritic in composition, and intrude the Birimian metavolcanic rocks. They are typically hornblende bearing and are commonly associated with gold mineralisation where they occur as small plutons within the volcanic belts (Kesse, 1985). The granitoids are massive in outcrop, do not have a compositional banding or foliation, and are thus generally considered post-deformation (Banoeng-Yakubu et al., 2011). The Kwahu-Morago Group of the Vol- taian Supergroup consist mainly of sandstone, which are white, medium-grained, cross-bedded, flaggy, and quartzose. Hydrogeologically, the study area falls within three main hydro- geological provinces, namely Birimian Province, Voltaian Province, and Crystalline Basement Granitoid Complex Province (Dapaah-Siakwan and Gyau Boakye, 2000). Groundwater in the Birimian Province occurs mainly in the saprolite, saprock and in the fractured bedrock. The most productive zones in terms of groundwater delivery in the Birimian Province comprise the lower part of the saprolite and the upper part of the saprock, which usually complement each other in terms of perme- ability and storage (Carrier et al., 2008). The upper, less permeable part of the saprolite can act as a semi-confining layer for this productive zone, while the lower, usually saturated part of the saprolite is characterised by lower secondary clay content, thus creating a zone of enhanced hy- draulic conductivity (Banoeng-Yakubu et al., 2011). Generally, areas underlain by the Birimian rocks display deeper weathering than areas underlain by the granitoids. Banoeng-Yakubo (1989) identified three types of basement aquifers. These are the weathered rock aquifers, which are fracture related, the fractured quartz-vein aquifers and the fractured unweathered aquifers. Banoeng-Yakubu et al. (2011) underscored that in the Birimian Province aquifers, the saprolitic zone is a combination of the topsoil, the under- lying lateritic soil, the highly weathered zone and the moderately weathered zone. Successful boreholes drilled through the rocks of the Birimian and the Tarkwaian range between 35 and 55 m with an average of 42 m (Agyekum, 2004). Carrier et al. (2008) also recorded borehole depth in a similar range of 35 and 55 m with an average of 50 m in the granitoids. Data collated from WRI-CSIR shows that the depth to water table falls within 4–37 m with an average of 21 m. There is no well-defined local groundwater flow regime, however, the groundwater flow within the study area follows the regional flow regime, which is from north to south and to a large extent toward the south-western parts of the study area. The aquifer transmissivity of the productive zones of the Birimian Province has been reported by Banoeng-Yakubu et al. (2011) to range between 0.2 m2/d and 119 m2/d, with an average of 7.4 m2/d. The Kwahu-morago Group, which is composed of ‘Yabraso’ sandstone formation, has been identified as good aquifer zone based on available data from drilling projects within the area. The basement rocks are characterised by little or no primary porosity, and therefore the hydrogeology of the area is controlled by secondary permeabilities resulting from weathering and fracturing of the rock which has enhanced the storage and transmissive properties of the rocks to form groundwater reservoirs (Darko et al., 2006; Yidana et al., 2012b). 3. Methodology Data of groundwater samples collected from boreholes in the study area were obtained from Hydronomics Ghana limited in Accra. Standard protocols for water sampling and storage as prescribed in APHA (2005); USGS (2006) and Canada CCME (2011) were adopted. One hundred and twelve (112) samples collected 500 ml in pre-cleaned sterilized poly propylene plastic bottles were stored in cool boxes (about 3 �C) and transported to the Water Research Institute of the Council for Scientific Y.S.A. Loh et al. Groundwater for Sustainable Development 10 (2020) 100296 4 and Industrial Research (WRI-CSIR) laboratory for major and trace cations and anions analyses whiles pH, electrical conductivity (EC) and Temperature were determined on the field with the aid of a multi-parameter portable meter (PELI 1521) from HANNA Instruments. In the laboratory, the samples were stored at 4 �C in a refrigerator for about a week before they were analysed. Different methods (i.e. gravi- metric (TDS), titrimetric (Ca, Total hardness, alkalinity), SPADNS (F� ), turbidimetric (SO4 2� ), argentometric (Cl� ), atomic absorption spec- trometry (As, Mn, Fe, HCO3 � ), flame photometry (K, Naþ, Mg2þ), hy- drazine reduction (NO3 � ), Stannous chloride (PO4 2� )) were used to analyse the major, minor and trace elements. The dataset was subjected to an internal consistency test using the charge balance error (CBE) to test its accuracy (equation (1)). A charge balance error value within �5% is generally acceptable and shows that the laboratory analysis of the parameters are a good balance of the cations and anions (Appelo and Postma, 2005). In view of this, samples that recoded CBE values above �5% were dropped from the analysis. C:B:E¼ P mcjzcj � P majzaj P mcjzcj þ P majzaj � 100 (1) Where mc and ma, and zc and za are respectively molar concentrations of major cations and anions, and charges of cations and anions. Similarly, the resulting datasets were subjected to normality test, since optimal multivariate statistical analyses assume Gaussian distri- bution of datasets. Datasets that were not normally distributed were log- transformed and/or standardised to their z score (equation (2)) values. z¼ x � μ s (2) Where x, μ and s are respectively the measured value, mean and stan- dard deviation of the parameter. The transformed datasets were subjected to R-mode and Q-mode hierarchical cluster analysis (HCA). The Q-mode HCA is used to discriminate the spatial associations or evolution of the groundwater into various types in space and/or time, whereas R-mode is used to determine and rank the sources of variation in the hydrochemistry. Although several similarity/dissimilarity and agglomerative techniques are available in HCA, the squared Euclidean distance and Ward’s agglomeration method were employed in this study, since a combination of these two have been identified to yield the best outcomes in HCA (Davis, 1986; Yidana et al., 2010, 2012a). For significant and optimal stochastic analysis, parameters with large missing datasets such as F, Fe, Mn and As in most cases were excluded from the cluster analysis. Principal component analysis (PCA) was also applied to the dataset to identify the main controls on groundwater chemistry. PCA is a data dimension reduction technique that reveals the significant components/ factors to aid interpretation of a large set of data and to visualise the correlations between the variables and hopefully be able to limit the number of variables. PCA was performed using the correlation matrix, which brings the measurements onto a common scale and the principal components sorted in a diminishing order of variance, such that the most important principal components are listed first. To ensure that the extracted components did not correlate with each other, an orthogonal (varimax) rotation technique was used. Assertions made in relation to the hydrochemical associations and processes based on the multivariate statistical techniques were sup- ported by employing conventional hydrochemical plots and mineral stability diagrams to further establish the main controls on groundwater chemistry and to identify the most stable mineral phases in the groundwater flow system. The quality of the groundwater was also assessed for drinking and irrigation purposes. Quality assessment for domestic purposes was based on a modified form of the water quality index (WQI) approach (Sahu and Sikdar, 2008), which is a weighted arithmetic index method. The resultant water quality indices estimated from this method were spatially interpolated using inverse distance weighting (IDW) technique. In order to minimise the errors associated with the spatial prediction, the interpolation was limited to places in the study area with optimal sample-point distribution, such that places with little/no sample points were excluded from the interpolation. Irrigation quality of the ground- water on the other hand was assessed using the Wilcox (1955) and United States Salinity Laboratory (USSL, 1954) diagrams. 4. Results and discussions The statistical summaries of the analysed parameters are presented in Table 1. Most of the major chemical parameters display high ranges of variance, suggesting variable sources and factors influence and contribute to the concentration of these parameters of the groundwater in the study area. Generally, concentrations of the major chemical pa- rameters are within the acceptable WHO recommended ranges for potable water (WHO, 2017). However, there are a few cases where some hydrochemical param- eters, which mostly occur as outliers, fall outside the acceptable WHO (2017) ranges for potable water. One instance is the occurrence of high fluoride levels recorded in some parts of Gindabo and Tuna in the central sections of the study area, which may be associated with the leaching and weathering of fluorite and/or the influence of alkaline water types in the area. Alkaline waters have been reported to retard the adsorption of fluoride onto the surfaces of clay minerals (Viero et al., 2009). pH presents the least variance among the major chemical parameters and has ranges falling outside (5.87–9.99) the WHO (2017) recommended range of 6.5–8.5, signifying a slightly acidic to slightly alkaline groundwater system. 4.1. Assessment of the main sources of variation in groundwater chemistry Results of Pearson correlation (r) analysis are presented in Table 2. Correlation analysis provides a quick way to visualise the relations be- tween two parameters for purposes of drawing inferences. It is obvious from the results that the main contributors to EC in the study area are Ca2þ, Naþ, HCO3 � , SO4 2� , Cl� and F� since these parameters have sig- nificant correlations (�0.50) with EC. Similarly, bicarbonate shows significant relation with Naþ and Ca2þ, which is consistent with silicate weathering (Freeze and Cherry, 1979), and possibly influenced by precipitation. R-mode hierarchical cluster analysis (HCA) and principal component analysis (PCA) have been used to further unearth the main chemical processes controlling groundwater chemistry in the study area. The Table 1 Statistical summary of the concentrations of the physicochemical parameters. Parameter Mean (X) Std. Dev (s) Minimum Maximum pH 6.86 0.61 5.87 9.99 EC 429 196 101 1040 TH 158.0 47.7 50.0 270.0 Alkalinity 178.3 57.5 70.0 366.0 Ca2þ 33.6 14.4 9.6 77.8 Mg2þ 18.0 6.4 3.4 32.5 Naþ 31.0 17.7 7.5 97.8 Kþ 4.9 1.7 1.4 11.0 HCO3 � 217.5 70.2 85.4 447.0 SO4 2- 12.8 14.5 2.5 65.2 Cl� 18.8 13.1 3.0 70.5 NO3 � 0.40 0.75 <0.01 5.3 PO4 2- 0.08 0.09 <0.01 0.39 F� 0.8 0.4 0.2 2.4 Fe 0.16 0.47 <0.01 4.13 Mn 0.06 0.07 <0.01 0.29 SiO2 59.1 11.6 28.0 87.6 As <0.01 0.01 <0.01 0.08 Y.S.A. Loh et al. Groundwater for Sustainable Development 10 (2020) 100296 5 dendrogram (Fig. 3) for the R-mode cluster analysis presents the visual associations among the parameters with a phenon line drawn at a linkage distance of about 22. Although the definition of clusters based on the distance of the phenon line is subjective, it is informed by the researcher’s understanding of the combining environmental factors such as the geology, hydrogeology and human activities which prevail in the study area and are likely to affect the chemistry of groundwater (Yidana et al., 2012a). The phenon line is drawn such that too many or too few clusters are not generated since the interpretation might be difficult, defeating the purpose of cluster analysis, or some important hydro- chemical processes might be omitted. Notwithstanding the semi-objectivity, two clusters were generated which represent two main groundwater associations and/or processes. Cluster 1 consists of Naþ, Ca2þ, EC, HCO3 � , pH, K, Mg and SiO2 which represents the dominance of rock-water interaction; dominated by dissolution of silicate minerals in the rocks of the area. The close association of EC, Ca2þ, HCO3 � , Naþ and pH in cluster 1a suggests the dissolution of feldspars in the sandstones found in the study area. The reaction of carbon dioxide in the atmosphere and precipitation and possibly within the soil zone results in the formation of carbonic acid, which dissolves such minerals during infiltration, thereby releasing HCO3 � and the associated ions. Cluster 1b on the other hand consists of SiO2, Kþ and Mg2þ, at a much shorter linkage distance than cluster 1a, and therefore presents the most similar members of the same cluster given the shorter linkage distance (Yidana et al., 2011), relative to cluster 1a and cluster 2. Cluster 1b probably represents silicate mineral weathering, especially micas such as biotite which are also present in the study area (Sawla Tuna Kalba District Assembly, 2014). The second cluster links NO3 � , SO4 2� and Cl� . Cluster 2 represents the influence of agrochemicals and domestic wastewaters, which are com- mon pollutants of groundwater in agrarian settlements such as the study area (Han et al., 2016). Table 2 Pearson correlation analysis of hydrochemical parameters. pH Ca2þ Mg2þ Naþ Kþ HCO3 � SO4 2- Cl� NO3 � PO4 2- F� Fe Mn SiO2 As EC pH 1.00 Ca2þ 0.30 1.00 Mg2þ 0.11 0.14 1.00 Naþ 0.32 0.51 0.04 1.00 Kþ � 0.02 0.35 0.18 0.14 1.00 HCO3 � 0.42 0.67 0.47 0.69 0.29 1.00 SO4 2- 0.06 0.50 0.00 0.54 0.20 0.15 1.00 Cl� 0.07 0.47 0.13 0.39 0.14 0.09 0.61 1.00 NO3 � � 0.08 0.24 0.13 0.21 0.19 0.08 0.35 0.42 1.00 PO4 2- � 0.02 � 0.15 � 0.23 � 0.28 � 0.11 � 0.31 � 0.13 � 0.04 � 0.17 1.00 F� 0.42 0.38 0.01 0.69 0.24 0.53 0.36 0.23 0.02 � 0.25 1.00 Fe 0.13 0.24 0.00 0.31 0.19 0.17 0.40 0.27 0.15 0.07 0.47 1.00 Mn 0.17 0.28 0.36 0.22 0.30 0.28 0.24 0.31 0.36 � 0.24 0.25 0.32 1.00 SiO2 � 0.27 � 0.40 0.00 � 0.45 � 0.10 � 0.42 � 0.26 � 0.16 � 0.04 0.12 � 0.44 � 0.14 0.00 1.00 As 0.04 0.00 0.16 0.01 0.06 0.10 � 0.05 � 0.04 0.03 � 0.07 0.03 � 0.02 0.01 0.00 1.00 EC 0.34 0.78 0.37 0.76 0.34 0.76 0.58 0.55 0.36 � 0.35 0.60 0.39 0.39 � 0.43 0.06 1.00 Fig. 3. Dendrogram from R-mode cluster analysis. Y.S.A. Loh et al. Groundwater for Sustainable Development 10 (2020) 100296 6 Results of principal component analysis (PCA) revealed two main components just like the R-mode cluster analysis, and these accounted for about 70% of the total variance in the hydrochemistry (Table 3). The factors were extracted based on variables with communalities of 0.5 and above (Kaiser, 1960), and variables below the set lower limit such as Kþ, PO4 2� , As, Mn and Fe were excluded from the analysis. Communalities measure the proportion of each variable’s variance that can be explained by the factors. Therefore, variables which loaded highly with more than one component such as Mg2þ and pH were dropped, since these vari- ables had duplicate effects and could not be used to explain a particular unique process in the hydrochemistry. The first component which accounts for over 40% of the total vari- ance in groundwater chemistry in the study area has high positive loadings for HCO3 � , EC, Naþ, Ca2þ and high negative loading for SiO2. It can be deduced from the correlation analysis, HCA and PCA that Component 1 represents the dissolution of silicates, particularly, the feldspars in the sandstones underlying the area, thus confirming the findings from cluster 1a (Fig. 3). The negative loading of SiO2 with component 1 also suggests that the silica in the groundwater does not contribute significantly to the electrical conductivity, which is in line with the results of the correlation analyses (Table 2). A plot of Na versus Cl (Fig. 4) suggests that the dissolution of halite is not the main source of Naþ in the groundwater and thus the dissolution of silicate minerals could be the source of this ion in the groundwater. The strong correla- tion between F� and Naþ (r ¼ 0.69) as shown in Table 2, possibly sug- gests the dissolution of minerals such as villiaumite and fluorapatite, which are common constituents of the bedrock of the study area (Yidana et al., 2012a). Table 3a Final factor loadings for the water quality parameters. Component 1 2 HCO3 � .923 -.050 EC .805 .491 Naþ .782 .328 Ca2þ .738 .397 SiO2 -.663 -.011 Cl� .175 .843 SO4 2� .305 .770 NO3 � -.009 .728 Table 3b Communalities of parameters on factor model. Initial Extraction EC 1.000 .889 Ca2þ 1.000 .702 HCO3 � 1.000 .854 Naþ 1.000 .720 SO4 2� 1.000 .686 Cl� 1.000 .741 NO3 � 1.000 .531 SiO2 1.000 .520 Table 3c Total variance explained. Component Initial Eigenvalues Extraction Sums of Squared Loadings Rotation Sums of Squared Loadings Total % of Variance Cumulative % Total % of Variance Cumulative % Total % of Variance Cumulative % 1 4.136 51.697 51.697 4.136 51.697 51.697 3.218 40.231 40.231 2 1.426 17.829 69.526 1.426 17.829 69.526 2.344 29.294 69.526 3 .737 9.208 78.733 4 .642 8.028 86.761 5 .504 6.297 93.058 6 .363 4.533 97.592 7 .129 1.608 99.200 8 .064 .800 100.000 Fig. 4. Molar concentrations of Naþ versus Cl� suggesting silicate weathering/ion exchange in the hydrochemistry of the study area. Y.S.A. Loh et al. Groundwater for Sustainable Development 10 (2020) 100296 7 Component 2 on the other hand loads highly with Cl� , SO4 2� and NO3 � and represents the influence of agrochemicals and domestic wastewa- ters, as these ions are common constituents of such pollutants. Since the study area is composed mainly of agrarian communities, with farming practices, which involves the use of organic and inorganic fertilizers such as manure and NPK, groundwater is likely to be polluted by such Fig. 5. Spatial distribution of the cluster from Q-mode HCA. Fig. 6. Dendrogram from the Q-mode HCA showing the spatial groundwater associations. Y.S.A. Loh et al. Groundwater for Sustainable Development 10 (2020) 100296 8 chemicals in the study area (Cobbina et al., 2012). Q-mode HCA resulted in a dendrogram (Fig. 6) showing three main spatial associations in the hydrochemistry in the study area. Average concentrations from the three clusters are presented with Stiff diagrams (Fig. 7). Generally, aquifers of the Birimian deliver groundwater of low ionic concentrations since their hydrochemistry is mainly influenced by silicate mineral weathering (Yidana et al., 2012a). Notwithstanding, the three clusters exhibit varying degrees of ionic concentrations and sug- gest a flow pattern in the groundwater system. It is obvious from Fig. 7 that on the average cluster 1 presents the least mineralised groundwater system. Cluster 1 is mainly composed of samples from around Kong, Sooma, Jilinkon, Nahari and Grungu, which are generally high elevation locations in the study area (Fig. 5). Conversely, cluster 3 presents the highest mean concentration of the major ions and TDS, and is found at locations with low elevations such as Babalayurin, Degwiwu, Kalba, and generally the south western portions of the study area, whereas cluster 2 occurs between clusters 1 and 3 (Fig. 5). In recharge areas, the concentration of ions in groundwater is usually low and similar to the precipitation in the area, but as the water travels through the subsurface and dissolves minerals and other materials, its ionic content increases with time. This explains why clusters 1 and 3 have been designated as recharge and discharge zones respectively, consistent with topographically-driven groundwater flow (Fitts, 2002), whereas cluster 2 samples are considered to represent a transition be- tween clusters 1 and 3. This is clearly demonstrated by the Stiff diagrams in Fig. 7. It is apparent from the Stiff diagrams (Fig. 7) that cluster 1 is a Ca–Mg–HCO3 water type, whereas cluster 2 is Mg–Ca–HCO3 water type which evolves into a Ca–Na–K–HCO3 water type in cluster 3. The dissolution of villiaumite and fluorapatite is plausible when the water had enough time to interact with the rocks as it evolves towards discharge areas thus increasing the concentration of Na in those areas. Similar water types were identified by Cobbina et al. (2012) in the Voltaian and Basement aquifers in the Savannah Region. This is in tandem with the assertions made above from Table 3 and Fig. 4, that dissolution of silicates plays a significant role, explaining over 40% of the total variance in groundwater chemistry in the study area. A Gibbs (1970) diagram plotted for all three clusters clearly indicates rock dominance as the main source of the dissolved ions in groundwater from the study area (Fig. 8). However, the hydrochemistry of cluster 1 samples is largely influenced by the chemistry of precipitation in the area, which is characteristically low in dissolved ions and high in Fig. 7. Stiff diagrams made from average concentrations of major ions from the three clusters from Q-mode HCA. Fig. 8. Gibbs diagram showing the main sources of variation in groundwater chemistry for the study area. Y.S.A. Loh et al. Groundwater for Sustainable Development 10 (2020) 100296 9 bicarbonate as a result of the CO2-charged rainfall. Although samples from clusters 2 and 3 plot within the rock dominance section, it can be seen that samples from cluster 3 plot much closer to rock domi- nance–evaporation-crystallization line than the rock domi- nance–precipitation line. This supports the argument that waters from cluster 3 are more enriched in dissolved ions as a result of the longer interaction of the groundwater with the rock matrix. Generally, groundwater from the study area is clearly a mixed cation fresh water type dominated by bicarbonate (HCO3>SO4þCl), as demonstrated by the Stiff diagrams (Fig. 7). Durov (1948) diagram has been used to discriminate the hydrochemistry further. From Fig. 9, it is apparent that most of the groundwater samples (42%) are dominated by Ca–HCO3, implying the dominance of alkaline earths over alkali (thus Ca þ Mg > Na þ K), whereas 36% and 20% are respectively Mg–HCO3 and Na þ K–HCO3 fresh water types. Cluster 3 samples plot closer to the Na þ K field and further away from the Ca and Mg fields, with corre- sponding high TDS and pH values (Fig. 9), confirming that these waters have had longer time to interact with the rock matrix and the sur- rounding environment. The low pH values mainly contributed from precipitation, as seen in cluster 1 samples, are neutralized as the water interacts and dissolves the rock material as it transits the recharge zones and evolves into a more ionically enriched water in cluster 3. The mineral stability diagrams for NaO–Al2O3–SiO2–H2O and Al2O3–SiO2–H2O are presented in Fig. 10. Both plots indicate that the most stable silicate phase in the aquifers is kaolinite, indicating the incongruent weathering of silicate minerals such as biotite and musco- vite, and suggests that the groundwater in the aquifer is relatively young to intermediate in its flow regime and age. This implies that there is little or no restricted groundwater flow such that the residence time is not long enough to allow for significant silica leaching into the groundwater (Edet and Okereke, 2005; Yidana et al., 2008a,b; Loh et al., 2016). 4.2. Groundwater quality evaluation for domestic use The quality of the groundwater has further been examined for its potability, using the water quality index (WQI) approach, modified after Brown et al. (1972). This approach is a weighted arithmetic approach, which involves a series of steps to arrive at a single value (index), which describes the overall quality of water in space and time. The suitability of groundwater in this study was based on the standards set by the World Health Organisation (WHO, 2017) for domestic drinking water. In the first step, water quality parameters of concern are assigned weights (wi) based on their importance and health implications when present in drinking water. F� , NO3 � and As were assigned a maximum of 5 based on their relative importance in drinking water. The rest of the parameters (pH, TDS, TH, Ca2þ, Mg2þ, Naþ, Kþ, CI� , SO4 2� , PO4 3� , Fe and Mn) were assigned weights of 1–5 also based on their relative importance and health implications when present in drinking water. Secondly, the Fig. 9. Durov diagram showing the predominant groundwater types in the study area. Fig. 10. Mineral stability diagrams for the (a) Ca–Al–SiO2–H2O system, (b) Na–Al–SiO2–H2O system at 1 atm and 25 �C. Table 4 Standards, weights and relative weights used for WQI computation. Chemical Parameter Objective to be met (Si) (mg/l) Weight (wi) Relative weight (Wi) pH 7.5 4 0.0943 TDS 500 4 0.0755 TH 200 4 0.0755 Ca2þ 200 2 0.0377 Mg2þ 150 2 0.0377 Naþ 200 2 0.0377 Kþ 30 2 0.0377 Cl� 250 3 0.0566 SO4 2- 250 3 0.0566 NO3 � 50 5 0.0943 PO4 3- 0.7 4 0.0755 F� 1.5 5 0.0943 Fe 0.3 3 0.0566 Mn 0.4 4 0.0755 As 0.01 5 0.0943 Y.S.A. Loh et al. Groundwater for Sustainable Development 10 (2020) 100296 10 relative weight (Wi) of each parameter was computed as a fraction of the total weights computed from the fifteen parameters (Table 4) based on equation (3). Wi wi P wi (3) Lastly, a quality rating scale (qi) was calculated for each parameter by dividing its concentration by the set objective (WHO standard) and multiplying by 100 (equation (4)). qi¼ Ci Si � 100 (4) Where Ci and Si are respectively parameter concentration and set objective. The sub-index and WQI are determined respectively using equations (5) and (6). Si¼Wi� qi (5) Fig. 11. Spatial distribution of water quality indices reclassified after Sahu and Sikdar (2008). Fig. 12. Groundwater quality classification for irrigation in the study area (USSL, 1954). Y.S.A. Loh et al. Groundwater for Sustainable Development 10 (2020) 100296 11 WQI¼ Xn n¼1 Si (6) The calculated WQIs are categorised according to Sahu and Sikdar (2008) as “excellent water” (WQI <50); “good water” (WQI ¼ 50–100); “poor water” (WQI ¼ 100–200); “very poor water” (WQI ¼ 200–300); and “unsuitable for drinking” (WQI >300). The computed WQIs have been spatially interpolated based on an inverse distance weighting (IDW) technique, which proved to be the most appropriate interpolation technique as a result of the smallest root mean square error associated with it (RMSE ¼ 0.852). The output was reclassified according to Sahu and Sikdar (2008) (Fig. 11). The esti- mated WQIs ranged from 16.64 to 132.15, with about 94% of the water classified as being “excellent” for drinking purposes, whereas 5% and 1% are considered “good” and “poor” quality respectively (Fig. 11). The poor water arises from two samples, STKD_086 and STKD_93 located at Sawla and Soomia, which are characterised by elevated levels of fluo- ride, arsenic and Fe. These samples generally have high TDS values, and the contamination could be attributed to localised leaching of such ions into the groundwater system. 4.3. Assessment of groundwater quality for irrigation purposes The presence of sodium ion in groundwater is of grave concern due to its ability to affect the permeability of soil as well as offset the osmotic pressure of plants. An offset in the osmotic pressure of the plants affects the rate of water intake by plant roots and its subsequent use in the process of photosynthesis. High concentrations of Naþ in soil therefore affect the metabolism of the plants and this can translate negatively to crop yield (Mohan et al., 2000; Zaidi et al., 2015). Saleh et al. (1999) concluded that excessively high salinity waters also have the tendency to affect the ability of the plants to absorb nutrients and water. Fig. 12 is based on the United States Salinity Laboratory (USSL, 1954), which represents a plot of sodium adsorption ratio (SAR) against EC (salinity). All the samples from the study area plotted within the “low sodicity” (S1) region and “low to high salinity” ranges (C1 – C3). About 18% of the samples fell within C1– S1 (low salinity—low sodicity) and are mainly composed of waters from cluster 1 members of the Q-mode HCA. Sam- ples from cluster 1 have been categorised as recharge zones, and have low ionic content. Most of the samples (72%) plotted within the C2–S1 region, “medium salinity—low sodicity” irrigation waters, whereas only 6% of the samples plot within C3–S1. All three regions present water of good quality and may be used for irrigation without prior treatment. This assertion is corroborated by the Wilcox (1955) plot, in which all the water samples plot in “excellent to permissible” category. It is obvious from Figs. 12 and 13 that the salinity of the groundwater in the study area increases as the water travels from recharge (cluster 1) to discharge (cluster 3) areas. Cluster 1 samples are typically low in both sodium and salinity as a result of the short residence time and interac- tion with the geology. Cluster 2 samples appear to present the best balance of sodium and salinity ranges (SAR ¼ 1–5 and EC ¼ 500–900 μ/cm) (Banoeng-Yakubo et al., 2009) and hence have the best quality for irrigation. But as the groundwater interacts and dissolves the rock ma- trix, alongside getting polluted by anthropogenic sources such as agro- chemicals and domestic wastewaters, its quality for irrigation deteriorates, as seen in cluster 3 samples (Fig. 12). 5. Conclusion The key factors controlling groundwater chemistry and quality in some Voltaian and Basement aquifers in portions of Sawla-Tuna-Kalba District, northern Ghana have been assessed. Statistical techniques such as HCA, PCA and correlation analysis coupled with conventional hydrochemical plots suggests the dissolution of silicate minerals, particularly feldspars and micas in the rocks, as well as the influence of agrochemicals and domestic waste as the main factors controlling groundwater chemistry in the study area. Even though silicate mineral weathering was identified as the dominant process controlling the groundwater chemical composition, the concentration of SiO2 is low and did not contribute significantly to the EC. Mineral speciation calcula- tions derived from the analysis also suggests that the most stable silicate mineral phase is kaolinite, which suggests that groundwater in the area is at intermediate stage, and flow is not restricted mainly due to the occurrence and pervasiveness of secondary permeability. Q-mode HCA identified three (3) groundwater flow regimes in the study area, recharge zones (cluster 1) located in high elevated areas, characterised by low TDS, and dominated by Ca–HCO3 water type; transition zones (cluster 2) dominated by Mg–Ca–HCO3 water type; and Ca–Na–K–HCO3 water type (cluster 3) in discharge areas. Also, groundwater quality for domestic purposes was assessed using a weighted arithmetic index approach. The computed water quality indices (WQIs) from the data suggest that 94% of the sampled wells provide groundwater of “excellent” quality for drinking purposes, whereas 5% and 1% present water of “good” and “poor” quality respectively. The poor quality presented by some of the wells is attrib- uted to elevated levels of fluoride, arsenic, Fe and TDS and may have Fig. 13. Groundwater quality assessment for the study area based on Wilcox (1955) diagram. Y.S.A. Loh et al. Groundwater for Sustainable Development 10 (2020) 100296 12 resulted from the leaching of such ions into the groundwater. The study also finds groundwater in the area to be permissible for irrigation based on the USSL and Wilcox plots. Funding This work was supported by the Catholic Missions/Government of Ghana project. Declaration of competing interest The authors confirm that there are no known conflicts of interest associated with this publication and there has been no significant financial support for this work that could have influenced its outcome. Acknowledgement The authors would like to acknowledge the support of Hydronomics Limited Ghana for the valuable data they provided for this research. References Abanyie, S.K., Ampofo, S., Biyogue Douti, N., Boateng, T., 2018. Geospatial assessment of groundwater quality in the Savelugu-Nanton municipality, northern Ghana. A.J.A.R. 4, 93–105. https://doi.org/10.26437/ajar.04.01.2018.09. Adomako, D., 2011. Hydrogeochemical evolution and groundwater flow in the Densu River Basin, Ghana. J.W.A.R.P. 3, 548–561. https://doi.org/10.4236/ jwarp.2011.37065. Agyekum, W.A., 2004. Groundwater resources of Ghana with the focus on international shared aquifer boundaries. Tripoli, Libya, June 2002. UNESCO-IHP VI Series on Groundwater. In: UNESCO-ISARM International Workshop-Managing Shared Aquifer Resources in Africa. B. Appelgreen, vol. 8, p. 600. APHA, 2005. Standard Methods for the Examination of Water and Wastewater, twenty- first ed. American Public Health Association/American Water Works Association/ Water Environment Federation, Washington DC. Appelo, C.A.J., Postma, D., 2005. Geochemistry, Groundwater and Pollution, second ed. A.A Balkema, Rotterdam. Banoeng-Yakubo, B., 1989. Occurrence of Groundwater in Basement Complex Rocks of the Upper Regions of Ghana. Unpublished M.Sc. dissertation. Obefemi-Awolowo University, Geology Department, lfe, Nigeria, 1989. Banoeng-Yakubo, B., Yidana, S.M., Nti, E., 2009. An evaluation of the genesis and suitability of groundwater for irrigation in the Volta Region. Environ. Geol. 57 (5), 1005–1010. https://doi.org/10.1007/s00254-008-1185-y. Banoeng-Yakubu, B., Yidana, S.M., Ajayi, J.O., Loh, Y., Aseidu, D., 2011. Hydrogeology and groundwater resources of Ghana: a review of the hydrogeology and hydrochemistry of Ghana. In: McMann, J.M. (Ed.), Potable Water and Sanitation, vol. 142. Nova Science, New York, NY, USA. Bates, D.A., 1955. Geological Map of Ghana: Ghana Geological Survey Department, Scale 1:1,000,000. Reprinted by the Geological Survey Department in 1966. Belkhiri, L., Boudoukha, A., Mouni, L., Baouz, T., 2011. Statistical categorization geochemical modeling of groundwater in Ain Azel plain (Algeria). J. Afr. Earth Sci. 59, 140–148. Bloomfield, J.P., Marchant, B.P., McKenzie, A., 2019. Changes in groundwater drought associated with anthropogenic warming. Hydrol. Earth Syst. Sci. 23, 1393–1408. https://doi.org/10.5194/hess-23-1393-2019. Boateng, T.K., Opoku, F., Akoto, O., 2019. Heavy metal contamination assessment of groundwater quality: a case study of Oti landfill site, Kumasi. Appl. Water Sci. 9, 1–15. https://doi.org/10.1007/s13201-019-0915-y. Brown, R.M., McCleiland, N.J., Deiniger, R.A., O’Connor, M.F.A., 1972. Water quality index – crossing the physical barrier. In: Jenkis, S.H. (Ed.), Proceedings in International Conference on Water Pollution Research Jerusalem, vol. 6, pp. 787–797. Canadian Council of Ministers of the Environment (CCME), 2011. Canadian Environmental Quality Guidelines. Winnipeg, MB. Carrier, M.A., Lefebvre, R., Racicot, J., Asare, E.B., 2008. Northern Ghana Hydrogeological Assessment Project. 33rd WEDC International Conference, Accra, Ghana. Cobbina, S.J., Nyame, F.K., Obiri, S., 2012. Groundwater quality in the Sahelian region of northern Ghana, West Africa. R.J.E.E.S. 4 (4), 482–491. Dapaah-Siakwan, S., Gyau Boakye, P., 2000. Hydrogeologic framework and borehole yields in Ghana. Hydrogeol. J. 8, 405–416. Darko, P.K., Mainoo, A.P., Dapaah-Siakwan, S., 2006. Borehole Inventory, Numbering and Functionality Survey in Sawla-Tuna-Kalba District. Final district specific preliminary hydrogeological report. CWSA-NR/AFD. Davis, J.C., 1986. Statistics and Data Analysis in Geology. John Wiley & Sons Inc., New York. Dickson, B., Benneh, G., 1995. A New Geography of Ghana, revised edn. Longman Group UK Ltd, UK. Durov, S.A., 1948. Classification of natural waters and graphical representation of their composition. Dokl. Akad. Nauk. USSR. 59 (1), 87–90. Edet, A., Okereke, C., 2005. Hydrogeological and hydrochemical character of the regolith aquifer, northern Obudu Plateau, southern Nigeria. Hydrogeol. J. 13 (2), 391–415. Fitts, C.R., 2002. Groundwater Science. Academic Press, New York, NY, USA. Freeze, R.A., Cherry, J.A., 1979. Groundwater. Prentice-Hall, Inc., Englewood Cliffs, New Jersey 07632, ISBN 0-13-365312-9. Ghana Statistical Service, 2010. Population and Housing Census Summary Report of Final Results. Gibbs, R.J., 1970. Mechanisms controlling world water chemistry. Science 170, 1088–1090. Han, D., Song, X., Currell, M., 2016. Identification of anthropogenic and natural inputs of sulfate into a karstic coastal groundwater system in northeast China: evidence from major ions, δ13CDIC and δ34 SSO4. Hydrol. Earth Syst. Sci. 20, 1983–1999. https:// doi.org/10.5194/hess-20-1983-2016. Hassan, J., 2014. A Geostatistical approach for mapping groundwater quality (Case study: Tehsil Sheikhupura). Indian J. Sci. Res. 3, 239–245. Helstrup, T., Jørgensen, N.O., Banoeng-Yakubo, B., 2007. Investigation of hydrochemical characteristics of groundwater from the Cretaceous-Eocene limestone aquifer in southern Ghana and southern Togo using hierarchical cluster analysis. Hydrogeol. J. 15, 977–989. Hwang, J.Y., Park, S., Kim, H., Kim, M., Jo, H., Kim, J., Lee, G., Shin, I., Kim, T., 2017. Hydrochemistry for the assessment of groundwater quality in Korea. J.A.C.E.N. 6, 1–29. https://doi.org/10.4236/jacen.2017.61001. Jing, J.I.N., Yufei, C., 2011. Assessment of groundwater quality based on principal component analysis method. Int. J. Civ. Struct. Eng. Res. 2 (2), 424–434. Junner, N.R., 1935. Gold in the Gold Coast: Gold Coast Geological Survey Memoir No. 4, Accra, p. 67. Junner, N.R., 1940. Geology of the Gold Coast and Western Togoland: Gold Coast Geological Survey Bulletin No., vol. 11. Accra, p. 40. Kaiser, H.F., 1960. The application of electronic computers to factor analysis. Educ. Psychol. Meas. 20, 141–151. Kesse, G.O., 1985. The Mineral and Rock Resources of Ghana. A.A. Balkema, Rotterdam. Kumar, P.J.S., Jegathambal, P., James, E.J., 2011. Multivariate and geostatistical analysis of groundwater quality in Palar River Basin. Int. J. Geol. 5 (4), 108–119. Loh, Y.S.A., Yidana, S.M., Banoeng-Yakubo, B., Sakyi, P.A., Addai, M.O., Asiedu, D.K., 2016. Determination of the mineral stability field of evolving groundwater in the Lake Bosumtwi impact crater and surrounding areas. J. Afr. Earth Sci. 121, 286–300. Mohan, R., Singh, A.K., Tripathi, J.K., Chowdhary, G.C., 2000. Hydrochemistry and quality assessment of groundwater in Naini industrial area, Allahabad District, Uttar Pradesh. J. Geol. Soc. India 55 (1), 77–90. Nur, A., Ishaku, J.M., Yusuf, S.N., 2012. Groundwater flow patterns and hydrochemical facies distribution using geographical information system (GIS) in Damaturu, Northeast Nigeria. Int. J. Geosci. 3, 1096–1106. https://doi.org/10.4236/ ijg.2012.35111. Saana, S.B.B.M., Fosu, S.A., Sebiawu, G.E., Jackson, N., Karikari, T., 2016. Assessment of the quality of groundwater for drinking purposes in the Upper West and Northern regions of Ghana. SpringerPlus 5, 1–15. https://doi.org/10.1186/s40064-016-3676- 1. Sahu, P., Sikdar, P.K., 2008. Hydrochemical framework of the aquifer in and around East Kolkata wetlands, West Bengal, India. Environ. Geol. 55, 823–835. Saleh, A., Al-Ruwaih, F., Shehata, M., 1999. Hydrogeochemical processes operating within the main aquifers of Kuwait. J. Arid Environ. 42, 195–209. Sarukkalige, R., 2012. Geostatistical analysis of groundwater quality in Western Australia. E.S.T.I.J. 2, 790–794. ISSN: 2250-3498. Sawla Tuna Kalba District Assembly, 2014. District Medium Term Development Plan under the Ghana Shared Growth and Development (Agenda). Draft report 2014- 2017. Sharma, P.K., Vijay, R., Punia, M.P., 2015. Geostatistical evaluation of groundwater quality distribution of Tonk district, Rajasthan. Int. J. Geomatics Geosci. 6 (2), 1474–1485. Sunkari, E.D., Abu, M., Bayowobie, P.S., 2019. Hydrogeochemical appraisal of groundwater quality in the Ga west municipality, Ghana: implication for domestic and irrigation purposes. Groundw. Sustain. Dev. 8, 501–511. https://doi.org/ 10.1016/j.gsd.2019.02.002. Telahigue, F., Belgacem, A., Souid, F., Kharroubi, A., 2018. Assessment of seawater intrusion in an arid coastal aquifer, south-eastern Tunisia, using multivariate statistical analysis and chloride mass balance. Phys. Chem. Earth 106, 37–46. https://doi.org/10.1016/j.pce.2018.05.001. Treidel, H., Martin-Bordes, Jose, L., Gurdak, J.J., 2012. Climate Change Effects on Groundwater Resources. A Global Synthesis of Findings and Recommendations. htt p://userwww.sfsu.edu/jgurdak/Publications/Treidel_etal_2011_ClimateChange-Gr oundwater_tableofcontents.pdf. (Accessed 25 August 2019). USGS (United States Geological Survey), 2006. National Field Manual for the Collection of Water-Quality Data: U.S. Geological Survey Techniques of Water-Resources Investigations book 9, chaps. A1-A10. http://pubs.water.usgs.gov/twri9A. USSL (United States Salinity Laboratory), 1954. Diagnosis and Improvement of Saline and Alkaline Soils. USDA Agric handbook, Washington DC. No. 60. Viero, A.P., Roisenberg, C., Roisenberg, A., Vigo, A., 2009. The origin of fluoride in the granitic aquifer of Porto Alegre, Southern Brazil. Environ. Geol. 56, 1707–1719. WHO (World Health Organisation), 2017. Guidelines for Drinking-Water Quality: Fourth Edition Incorporating the First Addendum, ISBN 978-92-4-154995-0. https://apps. who.int/iris/bitstream/handle/10665/254637/9789241549950-eng.pdf?sequence ¼1. (Accessed 13 May 2019). Wilcox, L.V., 1955. Classification and Use of Irrigation Waters. USDA, p. 19. Circular No. 969. Y.S.A. Loh et al. https://doi.org/10.26437/ajar.04.01.2018.09 https://doi.org/10.4236/jwarp.2011.37065 https://doi.org/10.4236/jwarp.2011.37065 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref3 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref3 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref3 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref3 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref4 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref4 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref4 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref5 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref5 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref6 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref6 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref6 https://doi.org/10.1007/s00254-008-1185-y http://refhub.elsevier.com/S2352-801X(19)30153-5/sref8 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref8 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref8 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref8 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref9 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref9 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref10 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref10 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref10 https://doi.org/10.5194/hess-23-1393-2019 https://doi.org/10.1007/s13201-019-0915-y http://refhub.elsevier.com/S2352-801X(19)30153-5/sref13 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref13 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref13 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref13 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref14 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref14 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref15 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref15 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref15 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref16 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref16 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref60 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref60 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref17 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref17 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref17 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref18 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref18 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref19 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref19 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref20 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref20 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref21 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref21 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref22 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref23 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref23 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref24 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref24 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref25 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref25 https://doi.org/10.5194/hess-20-1983-2016 https://doi.org/10.5194/hess-20-1983-2016 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref27 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref27 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref28 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref28 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref28 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref28 https://doi.org/10.4236/jacen.2017.61001 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref30 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref30 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref31 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref31 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref32 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref32 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref33 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref33 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref34 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref35 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref35 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref36 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref36 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref36 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref37 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref37 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref37 https://doi.org/10.4236/ijg.2012.35111 https://doi.org/10.4236/ijg.2012.35111 https://doi.org/10.1186/s40064-016-3676-1 https://doi.org/10.1186/s40064-016-3676-1 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref40 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref40 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref41 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref41 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref42 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref42 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref43 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref43 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref43 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref44 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref44 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref44 https://doi.org/10.1016/j.gsd.2019.02.002 https://doi.org/10.1016/j.gsd.2019.02.002 https://doi.org/10.1016/j.pce.2018.05.001 http://userwww.sfsu.edu/jgurdak/Publications/Treidel_etal_2011_ClimateChange-Groundwater_tableofcontents.pdf http://userwww.sfsu.edu/jgurdak/Publications/Treidel_etal_2011_ClimateChange-Groundwater_tableofcontents.pdf http://userwww.sfsu.edu/jgurdak/Publications/Treidel_etal_2011_ClimateChange-Groundwater_tableofcontents.pdf http://pubs.water.usgs.gov/twri9A http://refhub.elsevier.com/S2352-801X(19)30153-5/sref49 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref49 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref50 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref50 https://apps.who.int/iris/bitstream/handle/10665/254637/9789241549950-eng.pdf?sequence=1 https://apps.who.int/iris/bitstream/handle/10665/254637/9789241549950-eng.pdf?sequence=1 https://apps.who.int/iris/bitstream/handle/10665/254637/9789241549950-eng.pdf?sequence=1 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref52 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref52 Groundwater for Sustainable Development 10 (2020) 100296 13 Yidana, S.M., Ophori, D., Banoeng-Yakubo, B., 2008. Hydrogeological and hydrochemical characterization of the Voltaian Basin: the Afram Plains area, Ghana. Environ. Geol. 53 (6), 1213–1223. Yidana, S.M., Ophori, D., Banoeng-Yakubo, B., Samed, A.A., 2012. A factor model to explain the hydrochemistry and causes of fluoride enrichment in groundwater from the middle voltaian sedimentary aquifers in the Northern Region, Ghana. ARPN- JEAS 7 (1), 50–68. Yidana, S.M., Alfa, B., Banoeng-Yakubu, B., Addai, M.O., 2012. Simulation of groundwater flow in a crystalline rock aquifer system in southern Ghana – an evaluation of the effects of increased groundwater abstraction on the aquifers using a transient flow model. Hydrol. Process. 28 (3), 1084–1094. https://doi.org/10.1002/ hyp.9644. Yidana, S.M., Banoeng-Yakubo, B., Akabzaa, T.M., 2010. Analysis of groundwater quality using multivariate and spatial analysis in the Keta basin, Ghana. J. Afri. Earth Sci. 58, 220–234. Yidana, S.M., Banoeng-Yakubu, B., Akabzaa, T., Asiedu, D., 2011. Characterisation of the groundwater flow regime and hydrochemistry of groundwater from the Buem formation, eastern Ghana. Hydrol. Process. 25, 2288–2301. https://doi.org/ 10.1002/hyp.7992. Yidana, S.M., Ophori, D., Banoeng-Yakubo, B., 2008. Hydrochemical evaluation of the Voltaian system-the Afram Plains area, Ghana. J. Environ. Manag. 88, 697–707. https://doi.org/10.1016/j.jenvman.2007.03.037. Zaidi, F.K., Nazzal, Y., Jafri, M.K., Naeem, M., Ahmed, I., 2015. Reverse ion exchange as a major process controlling the groundwater chemistry in an arid environment: a case study from northwestern Saudi Arabia. Environ. Monit. Assess. 187, 607. https://doi.org/10.1007/s10661-015-4828-4. Y.S.A. Loh et al. http://refhub.elsevier.com/S2352-801X(19)30153-5/sref53 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref53 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref53 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref54 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref54 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref54 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref54 https://doi.org/10.1002/hyp.9644 https://doi.org/10.1002/hyp.9644 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref56 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref56 http://refhub.elsevier.com/S2352-801X(19)30153-5/sref56 https://doi.org/10.1002/hyp.7992 https://doi.org/10.1002/hyp.7992 https://doi.org/10.1016/j.jenvman.2007.03.037 https://doi.org/10.1007/s10661-015-4828-4 Assessment of groundwater quality and the main controls on its hydrochemistry in some Voltaian and basement aquifers, north ... 1 Introduction 2 Study area 2.1 Geology and hydrogeology 3 Methodology 4 Results and discussions 4.1 Assessment of the main sources of variation in groundwater chemistry 4.2 Groundwater quality evaluation for domestic use 4.3 Assessment of groundwater quality for irrigation purposes 5 Conclusion Funding Declaration of competing interest Acknowledgement References