Hydrochemical characterization of shallow and deep groundwater in Basement Complex areas of southern Kebbi State, Sokoto Basin, Nigeria

Abstract

Groundwater under Basement Complex areas of southern Kebbi has been characterized in order to determine its suitability for drinking and irrigation use. Water samples were drawn from shallow groundwater (hand-dug shallow wells < 5 m) and deep groundwater (boreholes > 40 m). Physical parameters (i.e., temperature, TDS, pH, and EC), were determined in situ, using handheld meters. Discrete water samples were obtained for determination of chemical parameters. Results from several-sample ANOVA (Kruskal–Wallis test) suggested that heterogeneity in water table appeared to exert significant influence on groundwater chemistry which is characterized by a significant difference in pH, EC TH, Na⁺, Zn²⁺, Mg²⁺, PO₄³⁻, Cl⁻, HCO₃⁻, SO₄²⁻, and NO₃⁻ concentrations. Also, ions including Fe³⁺, Zn²⁺, Mg²⁺, Na⁺, PO₄³⁻, and SO₄²⁻ are above World Health Organization (2011) and National Standard for Drinking Water Quality (2007) reference guidelines. Most of the groundwater sources are moderately hard. Groundwater classification based on chloride, EC, and TDS revealed water of excellent quality for all types of uses. However, groundwater classification based on nitrate pollution revealed water of poor quality. Rock mineral is the major mechanism controlling water chemistry, as revealed by the Gibbs model. Most of the water sources have positive Scholler index, indicative of overall base exchange reactions in the underlying aquifers. Such condition was well explained by Piper trilinear diagram, which revealed two types of faces: Ca–Mg–HCO₃ and Ca–Mg–SO₄–Cl. The HCA categorized wells into three groups according to their hydrogeochemical physiognomies. Despite the significant difference in ions concentration and chemical indices, groundwater composition is more influenced by rock weathering than anthropogenic inputs. Groundwater evaluation for irrigation use indicates a significant difference in SAR level which is related to poor permeability index in shallow groundwater. Higher values of Kelly’s index and magnesium adsorption ratio threatened groundwater suitability for irrigation use in the study area.

Introduction

Groundwater is one of the most important environmental reserves exploited for industrial, agricultural, and domestic uses (Wagh et al. 2016). The composition of groundwater tends to be good in most natural aquifers, and as a result, groundwater is increasingly been exploited as it meets the basic requirements for most uses. Groundwater quality is primarily influenced by the aquifer rock mineral and recharge pathways (Kohlhepp et al. 2017; Liu et al. 2017; Gu et al. 2017; Kumar et al. 2019). As water passes through its recharge pathways from recharge to discharge points, several other types of hydrogeochemical processes alter its physical and chemical properties and in some aquifers, water may be unsuitable for drinking and agriculture (Abd El-Aziz 2017; Dehnavi et al. 2011; Farid et al. 2017). At least, one-third of the global population rely on groundwater for drinking (Panaskar et al. 2016). While groundwater is increasingly exploited in response to population pressure (Bertrand et al. 2016; Shang et al. 2016; Parisi et al. 2018; Cavalcante Júnior et al. 2019), increased urbanization (Minnig et al. 2018; Tam and Nga 2018; Hughes 2019), industrialization (Zheng et al. 2019) and irrigation farming (Panaskar et al. 2016; Fabbri et al. 2016; Pulido-Bosch et al. 2018), only about 22% of the Earth’s 37 km³ (freshwater) is found in aquifers as groundwater (Panaskar et al. 2016). Remarkably, about 97% of this quantity is accessible for human use (Panaskar et al. 2016). Overexploitation of groundwater in key sedimentary aquifers is a global phenomenon (Joshi et al. 2018). The rate of annual global groundwater withdrawal, even though may vary with climate and geography, is about 1500 km³, which is above the normal rates of annual global groundwater recharge or replenishment (Joshi et al. 2018).

Growing demand for water driven by economic growth (Hertel and Liu 2016; Kummu et al. 2016; Shahzad et al. 2017; Flörke et al. 2018), urbanization (Kulabako et al. 2007; Sperling and Sarni 2019), generation of electricity (Powell et al. 2019), irrigation (Erban and Gorelick 2016; Wu et al. 2016), and domestic uses has inflamed global groundwater withdrawal (Shahid et al. 2015; Wada et al. 2016; Veldkamp et al. 2017; Agarwal and Garg 2015; Thomas et al. 2019), causing water table declines in aquifers (Joshi et al. 2018). Shallow groundwater is often the most affected by small-scale irrigation farming (Li et al. 2018), particularly in developing countries. In alluvial basins (e.g., Sokoto Basin), groundwater pollution is primarily derived from reinfiltering groundwater from irrigation fields (Selck et al. 2018) and municipal and industrial sewage (Wali et al. 2018a, b). Geophysical investigations showed alluvial aquifers (e.g., Sokoto Basin) are often in hydraulic conductivity with surface water (Emujakporue et al. 2018; Kudamnya and Andongma 2017; Nwankwo 2015; Nwankwo and Shehu 2015; Onuigbo et al. 2017), with shallow aquifers serving as conduits through which pollutants from surface water are transported to deep groundwater reservoirs.

There are immeasurable number of studies comparing groundwater quality between shallow and deep aquifers around the world: Basharat (2012), assessed the groundwater quality by demarcating the zones where brackish water exists in the form of depths and zones; Brancelj et al. (2016), showed variations in physical and chemical composition between shallow boreholes in distinct aquifers situated a few kilometers away from each other; Das and Mukherjee’s (2019), depth-dependent groundwater response to coastal hydrodynamics, showed variability in salinity between 1 and 4 ppt at > 100 mbgl depths, indicating a mixing between chemically discrete groundwater aquifers; Hubalek et al. (2016) showed that deepest aquifers had the least taxon richness and unexpectedly held Cyanobacteria; Pandith et al. (2017) revealed differences in fluoride absorption in shallow (basalt − 10 to 167 m) and deeper (granite less than 167 m) aquifers; Rajmohan et al. (2017) also discovered variability of water quality of deep and shallow wells; Holbrook et al.’s (2019) analysis of the relationship between physical and chemical showed that weathering occurs from a 65-m-deep well over earth’s critical zone; and Long et al. (2019) outlined a cavernous rotation and regularity of groundwater movement between aquifers. Evaluations of this kind are imperative since they help to describe the connections between groundwater tables, and this affects both the quality of shallow and groundwater aquifers (Hubalek et al. 2016; Park et al. 2018; Holbrook et al. 2019).

However, groundwater has been studied using different techniques: geothermal techniques (Nwankwo 2015; Nwankwo and Shehu 2015; Olatunji and Musa 2013; Toyin et al. 2016), isotope techniques (Adelana et al. 2003; Geyh and Wirth 1980; Fillion et al. 2014, 2018; Selck et al. 2018; Kattan 2018; Alemayehu et al. 2019; Besser et al. 2019), multivariate statistics (Mondal et al. 2010; Azhar et al. 2015; Yidana et al. 2018), univariate statistics (Marghade et al. 2010; Selvakumar et al. 2017; Wali et al. 2018a, b), modeling (Ebrahim et al. 2019; Locatelli et al. 2019), and chemical indices (Marghade et al. 2010; Panaskar et al. 2016; Wagh et al. 2016). Results obtained from these studies showed groundwater quality is influenced by both the natural geogenic processes and anthropogenic activities. In some environments, the groundwater quality is largely dependent on the water table depths (Dhar et al. 2008; Han et al. 2013; Deng et al. 2014).

Nigeria is the fastest growing country in sub-Saharan Africa (SSA) in terms of the human population (Akombi et al. 2019; Young 2019). Improved water supply which is one of the essentials for a healthy living has been constrained by uncontrolled anthropogenic activities (Vijay et al. 2011; Qin et al. 2013; Chen et al. 2016; Chitsazan et al. 2019) and by lesser extent natural conditions (Izah et al. 2016; Makinde et al. 2017; Zadawa and Omran 2018). Characterization of groundwater in Nigeria, is further constrained by lack of data, especially from Basement Complex areas of northern Nigeria (Akinluyi et al. 2018; Olubusola et al. 2018; Oyedele and Olayinka 2019; Tajudeen et al. 2019), owing to difficulties associated with groundwater exploration and accessibility (Betzler et al. 2017; Muhammad and Saad 2018; Oyeyemi and Aizebeokhai 2018; Ahmed and Mansor 2018).

The Basement Complex areas of southeastern Sokoto Basin are in southern Kebbi State (Fig. 1) and are underlain by Pre-Cambrian Basement Formation (Akujieze et al. 2002; Joseph and Bamidele 2018; Oseji and Egbai 2019). Groundwater in the study area is generally available in small quantity derived from fractures and tabular partings and from the regolith, just below the earth surface (Anderson and Ogilbee 1973; Edet 1990; Paul and Bayode 2012; Odukoya 2015). The fissures are usually most open above a depth of 91 m but even so, yields to boreholes are relatively low and cause high drawdown (Anderson and Ogilbee 1973; Offodile 2002). While boreholes are widely used in the study area as sources for improved water supply, shallow groundwater remained the most reliable source of drinking water especially in rural areas (Mohammed et al. 2007; Adelana et al. 2003; Aleke and Nwachukwu 2018; Olorunfemi and Oni 2019). However, the geological settings in addition to changes in land use combined with rock mineral influence groundwater composition (Dehnavi et al. 2011; Khatri and Tyagi 2014; Redwan and Abdel Moneim 2016; Venkatramanan et al. 2017; Jebreen et al. 2018; Mukate et al. 2019). Effluents from municipal, industrial, and crop fields are transported to groundwater by surface flows via infiltration (recharge) zones. These impurities are first collected into shallow aquifers (Kong et al. 2018; Narr et al. 2019), before their eventual downward movement to deep groundwater reservoirs, where they may stay for decades or even centuries.

Fig. 1

Map of the study area a Sokoto Basin; b Kebbi State; and c Study area

To highlight this problem, we look at the Basement Complex areas of southeastern Sokoto Basin. The study area is underlain by intrusive granite of igneous origin and deformed metamorphic rocks, chiefly gneiss, schist, hyalite, and quartzite (Toyin et al. 2016; Toyin and Adekeye 2019; Omolabi and Fagbohun 2019). Groundwater in this type of aquifer (often one-aquifer system), tend to be highly mineralized. Further, the topography of the study area presents another obstacle to groundwater development (Kogbe 1986). Therefore, groundwater is cheaply found along with low-lying areas and is largely hauled out using handlines from shallow wells particularly in rural areas (Anderson and Ogilbee 1973; Offodile 2002). Groundwater quality studies in Sokoto Basin (du Preez and Berber 1965; Anderson and Ogilbee 1973; Adelana et al. 2003; Amadi et al. 2015; Ekpoh and Ekpenyong 2011; Ette et al. 2017; Toyin et al. 2016; Wali et al. 2016, Wali et al. 2018a, b) revealed water of excellent quality and of Holocene age (100–10,000 years BP). These studies were carried out in Cretaceous and Cenozoic Sediment sections of north-western and central Sokoto Basin. Groundwater quality in the Basement Complex section of southeastern Sokoto Basin remained poorly reported in the literature (Anderson and Ogilbee 1973; Offodile 2002; Nwankwoala 2015), even though hydrochemical evaluation of groundwater over space and time proved to be an important technique for solving different hydrogeochemical problems (Edmunds 2009; Li et al. 2018; Yin et al. 2019). Understanding the aquifer hydrochemistry is important for effective utilization of, and development of this finite resource (Kashiwagi et al. 2006). Thus, the objective of this study is to assess the variability of groundwater quality between shallow and deep aquifers and evaluate its suitability for drinking and irrigation uses.

The study area

Location and climate

The Basement Complex areas of southeastern Sokoto Basin are in southern Kebbi State. The study area is situated between Latitudes 11°20″ and 11°40″N and Longitudes 4°30″E and 5°30″E (Fig. 1) and covers about 2411.69 km². This area overlays Fakai and Zuru local government areas (LGAs). From the conglomeration point along Koko-Mahuta road, just about 40 km before Mahuta, surface elevation increases steadily passing through Fakai and reaching over 400 m above sea level in Dabai. The rock outcrops formed a triangle of basement rock outcrop, which extends from Fakai to Zuru and Yauri.

The climate of the study area is hot, semiarid tropical (AW) in Koppen’s classification. It is dominated by two opposing wind systems: Tropical Maritime and Tropical Continental air masses (Gada 2014). These give the study area two contrasting seasons—wet and dry. The dry season results from continental air mass blowing from the Sahara Desert. The dry season lasts from October to April, whereas wet season lasts from May to October. From March onward, the temperature rises to over 35 °C (Fig. 2a). Temperature is generally high and showed marked seasonal variation. Mean maximum temperature is highest in April. Mean minimum temperature is the lowest in December < 20 °C. Monthly rainfall ranges from < 50 mm to over 250 mm. Most of the precipitation falls in July, August, and September (Fig. 2b).

Fig. 2

Source of Data: Climate-Data Org. Retrieved from https://en.climate-data.org/africa/nigeria/kebbi/zuru-380580/on 10/06/2019

A chart showing the variability of a temperature; and b rainfall in the study area.

Geological setting

Figure 3 shows the geological cross section of Sokoto Basin. In terms of geology, the Sokoto Basin is extensively studied (Emujakporue et al. 2018; Ette et al. 2017; Kogbe 1981; Moody and Sutcliffe 1991; Nwankwo 2015; Nwankwo and Shehu 2015; Onuigbo et al. 2017). The sediments of Sokoto Basin were accrued during three main stages of deposition, i.e., the first and last phases were the continental Mesozoic and Tertiary, with a second phase which is the marine Maastrichtian to Paleocene. Superimposing the Pre-Cambrian Basement unconformably are the grits and clays of the Illo and Gundumi Formations, forming part of the “Continental Intercalaire” of West Africa (Emujakporue et al. 2018; Kamba et al. 2018; Kogbe 1981; Moody and Sutcliffe 1991). These formations are superimposed unconformably by the Rima Group (Maastrichtian), consisting of mudstones and friable sandstones (Taloka and Wurno Formations), separated by the fossiliferous, shelly Dukamaje Formation (Emujakporue et al. 2018; Kamba et al. 2018; Kogbe 1981; Moody and Sutcliffe 1991). The Dange and Gamba Formations (mainly shales) of Paleocene are separated by the Kalambaina Formation (Calcareous). The Gwandu Formation (Continental Terminal) which covers these formations is of Tertiary age. These sediments dip gently and thicken gradually toward the northwest, with a maximum thickness of over 1200 m near the border with the Niger Republic (Emujakporue et al. 2018; Kamba et al. 2018; Kogbe 1981; Moody and Sutcliffe 1991).

Fig. 3

The geological cross section of Sokoto Basin

Hydrogeological setting

Lying below the sedimentary rocks of the Sokoto Basin and rising to the land surface in the highlands to the south and east of the basin are crystalline rocks of pre-Cretaceous age. These comprise of plutonic granite of igneous derivation and warped metamorphic rocks, mainly gneiss, schist, phyllite, and quartzite (Kogbe 1986). Groundwater in the high ground zones of crystalline rocks is usually obtainable in small amounts from fissures or other flat separations and from the regolith (worn rock) just underneath the ground surface (Anderson and Ogilbee 1973). The fissures are typically found above a depth of 91.44 m but, even so, yields to boreholes are rather low and cause high drawdowns (Anderson and Ogilbee 1973). The formation is Pre-Cambrian in age and is made up of schists, granite gneisses, phyllites, and quartzite. Of hydrogeological importance is the slight, north–south elongate belt of metasediments intersecting the basement rocks. The formation is splintered in places and occasionally deeply battered below ground surface. Likewise, the western peripheries of the basement rocks are marked by penetrating fracture systems (Anderson and Ogilbee 1973).

Hydrogeological studies of the pre-Cretaceous Basement Complex Formations of Nigeria (Akinluyi et al. 2018; Anderson and Ogilbee 1973; Olorunfemi and Oni 2019), revealed an average borehole yield of 14.67 gallons per sec (gps) from a mean depths of 37.49 m. Though, few boreholes in the unweathered rock, typically granite or gneiss produced more than a meager amount of water; in many wells, no water was found (Anderson and Ogilbee 1973). In contrast, wells tapping weathered granite and gneiss, where fully saturated, are found to yield the highest quantity of water (51.67 gps); the average yield for boreholes tapping battered rock was estimated to be around 23.33 gps per borehole (Anderson and Ogilbee 1973). Also, drawdowns during pumping were very high (62.48 m) from some wells. The normal depth of wells tapping the weathered materials is estimated to be 38 m and the normal depth to water stand at about 6 m (Anderson and Ogilbee 1973). Generally, groundwater condition in the basement section of Sokoto Basin is typical of the Nigerian Basement Complex described by (Akinluyi et al. 2018; Olorunfemi and Oni 2019).

Many wells were developed in the crystalline rocks particularly in the eastern parts of the basin (Gusau-Mafara axis). The borehole yield was found to be higher in the metasediments with splintered quartzites (Anderson and Ogilbee 1973). Yet borehole in most parts of the Basement Complex section of the basin is low. This kind of poor yields is also characteristic of Basement Complex Formation of Nigeria. Such low yield zone results from aquifers of weathered or fractured materials, with limited areal scope, such that during pumping or well development the spreading cone of depression hits the highly indurated frontier rocks in a short time (Akinluyi et al. 2018; Anderson and Ogilbee 1973; Olorunfemi and Oni 2019). The water level, or the drawdown, drops sharply and steadily decrease the yield per individual boreholes. Numerous boreholes and hand-dug wells have been drilled in the area and are still been drilled or dug in the Basement Complex zones of the basin for both the rural and urban water supply. High yields can be attained in some areas in the Basement Complex, contingent on the tectonics in those areas. Despite the poor yields, boreholes and hand-dug wells in Basement Complex zones are sufficient to sustain hand pumps, in rural water supply programs (Akinluyi et al. 2018; Anderson and Ogilbee 1973; Olorunfemi and Oni 2019). Figure 4 illustrates a typical lithologic section of the study area.

Fig. 4

Lithologic section of boreholes in the study area a Zuru and b Fakai

Groundwater quality

Geological work in Sokoto Basin dates to 1800 s. Reporting of fossil fuel localities was the main objective. A Comprehensive study of groundwater was carried out by du Preez and Berber (1965). Groundwater recharge (Adelana et al. 2002) is highly variable across the basin. Groundwater quality (Anderson and Ogilbee 1979; Uma 1993; Alagbe 2006; Graham et al. 2006; Wali et al. 2016; Wali and Bakari 2016; Wali et al. 2018a, b) is highly variable with TDS concentration ranging from 130 to 2340 mg/l. Sodium and nitrate concentrations exceed WHO reference guidelines in some locations. The hydrogeochemical faces (Alagbe 2006; Wali et al. 2018a, b) are predominantly of two types: calcium–magnesium–bicarbonate and calcium–magnesium–sulfate–chloride in nature. These faces perhaps are derived from dissolution of calcium and magnesium carbonates.

Materials and methods

Groundwater sampling and laboratory analysis

Forty groundwater samples were collected, 20 each from deep and shallow aquifers. Groundwater samples were collected mainly from shallow wells and boreholes which are currently in use. Samples were drawn from water sources constructed by Kebbi State Government. Because these sources are expected to meet all the necessary requirements for water supply, physical parameters—temperature, pH, EC, and TDS—were determined in situ using water quality probes (Table 1). Probes were first calibrated by deionized water and then by water from shallow wells and boreholes. Discrete water samples were collected in 1-l polyethylene bottles for determination of cations (K⁺, Na⁺, Ca²⁺, Cu²⁺, Fe³⁺, Zn²⁺ and Mg²⁺) and anions (Cl⁻, HCO₃⁻, CO₃²⁻, PO₄³⁻, NO₃⁻, and SO₄²⁻).

Table 1 Summary of field and laboratory methods

Samples were stored in insulated containers less than 5 °C. Prior to the collection of water samples, polyethylene bottles were washed twice; initially by using deionized water and then with the water from sampled boreholes and hand-dug shallow wells. Water samples were analyzed within 24 h, and as a result, no acid treatment for samples was made. All analyses were carried out in triplicates, and results were found reproducible within ± 5 error limit. Table 1 summarizes field and laboratory methodologies employed in this study.

Ions, including potassium, sodium, calcium, copper, iron, magnesium, and zinc, were analyzed using atomic absorption spectrometry. It is a method in which free gaseous atoms engross electromagnetic radiation at a specific wavelength to produce a calculable signal. The absorption of those free absorbing atoms in the optical path is proportional to the absorption signal (Fernández-Cirelli et al. 2009). Consequently, for AAS measurements, the analyte was first converted into gaseous atoms, typically by applying heat to a cell (i.e., atomizer). The nature of atomizer outlines the two main AAS-based analytical techniques: flame atomic absorption spectrometry (FAAS) that consistently offers analytical signals and electrothermal atomic absorption spectrometry (ETAAS) sending analytical signals in an intermittent mode (2–4/sample). In both approaches, liquid (or dissolved) samples are presented into the analyzer, as an aerosol as in the case of FAAS which is used in this study or as a fully marked low microliter volume in ETAAS (Fernández-Cirelli et al. 2009).

Monochromators, based on Czerny-Turner, Ebert, and Littrow designs are the most conventional wavelength selectors used in AAS. More recently, echelle optics is being fused to commercial AAS devices. The introduction of an echelle configuration in association with a powerful continuous lamp creates significant capabilities, for instance, analysis of the spectral background near the line and concurrent multi-elemental scrutiny (Fernández-Cirelli et al. 2009). The said ions were analyzed using Atomic Absorption Spectrophotometer (Model: Rs 12.5 Lakh/Piece). Nitrate and phosphate were analyzed using automated colorimetry (BluVisionTM discrete analyzer). In addition, sulfate concentration was analyzed using iron chromatography (Agilent 6890 Plus GC with FPD and 7683 Autosampler). Lastly, chloride, carbonate, and bicarbonate were determined by titration (Table 1). Table 2 presents the results of physicochemical composition of shallow and deep groundwater in the study area. Groundwater suitability for drinking was evaluated by comparing the results with World Health Organization (World Health Organization 2011) and Nigerian Standard for Drinking Water Quality (National Standard for Drinking Water Quality 2007) reference guidelines for drinking water quality (Table 2).

Table 2 Physicochemical parameters of groundwater samples from deep and shallow groundwater sources

Calculation of chemical indices

Chemical indices, including Ki, Mg, MR, PI, RSC, Si, SAR, Na% and Vi (Table 3), were also calculated to evaluate ion exchange, silicate weathering reactions. Also contained in the table is a formula for calculating total hardness.

Table 3 Summary of chemical indices and analytical methods

Statistical analysis

Groundwater data were standardized and summarized using basic descriptive statistics: mean, minimum, maximum, and standard error (Table 4). Pearson’s correlation (r) was used to test the relationship between physical and chemical elements of groundwater following Margahde et al. (2011). Prior to this, a nonparametric test (Kruskal–Wallis) was used to test whether there is a significant difference in groundwater composition between shallow and deep groundwater sources, using several samples ANOVA (Table 4). This method allows for comparing several independent random samples and can be used as a nonparametric substitute to the one-way ANOVA (Cruz et al. 2019; Montcoudiol et al. 2019; Morris et al. 2019). The Kruskal–Wallis test statistic for k samples, each of size n_i is defined viz:

$$T = \frac{1}{{s^{2} }}\left[ {\mathop \sum \limits_{i = 1}^{k} \frac{{R_{i} }}{{n_{i} }} - N\frac{{\left( {N + 1} \right)^{2} }}{4}} \right]$$

(1)

where N is the total number (all in) and R_i is the sum of the ranks (from all samples drawn) for the ith sample and:

$$S^{2} = \frac{1}{N - 1} = \left[ {\mathop \sum \limits_{\text{all}} R_{ij}^{2} - N \frac{{\left( {N + 1} \right)^{2} }}{4}} \right]$$

(2)

Table 4 Summary of physical and chemical properties of groundwater

The null hypothesis of the test is that all k distribution functions are equal. The alternative hypothesis is that at least one of the observations tends to yield larger values than at least one of the other observations (Cruz et al. 2019; Montcoudiol et al. 2019; Morris et al. 2019). Results obtained show that there was no significant difference in TDS, K⁺, Ca²⁺, and Cu²⁺ concentrations between shallow and deep groundwater.

Principal component analysis

One of the major multivariate statistical method used in the interpretation of water chemistry is principal component analysis (Ayoko et al. 2007; Miguntanna et al. 2010). The PCA which is multivariate statistical method is applied to reduce the size of hydrochemical data (Alias et al. 2014; Hildebrandt et al. 2008; Machiwal and Jha 2015; Yidana 2010; Yidana et al. 2018). Reduction of the analytical data of individual sampling location is the primary purpose of applying PCA. The PCA tends to be intercorrelated with a less important set of ‘principal components’ (PCs) which can be interpreted (Machiwal and Jha 2015). The PC group inter-related absorptions together and can be related to some geogenic processes or pollution from the anthropogenic sources.

Generally, PCA comprises of two steps: standardization of data and extraction of PCs (Machiwal and Jha 2015). While some related data on the variability may be lost via transformation, the explanation of the system is significantly abridged, and it can be envisaged to derive suitable evidence on the relationship between parameters and observations (Hildebrandt et al. 2008). The PCA bilinear model can be rearranged following the matrix decomposition equation (Hildebrandt et al. 2008), and thus,

$$X = TP^{T} + E$$

(3)

where X represents a matrix of data which is compressed into T (scores of matrices) and PT (matrix of loadings), plus E (matrix residual) (Hildebrandt et al. 2008). The extracted PCs in this study showed that groundwater in both shallow and deep aquifers are more influenced by the natural geogenic process as compared to human activities.

Hierarchical cluster analysis

The vital role of HCA is to discrete the parameters in a raw data matrix or PC scores into separate classes designated by the user without any prior hypotheses. The HCA is an unverified outline identification technique that reveals inherent assembly or recognizing the pattern of a dataset without a prior hypothesis with regard to the data so that the objects of the system can be classified into clusters based on their resemblances (Lin et al. 2012; Machiwal and Jha 2015). The HCA is a widely used method which can form clusters consecutively, beginning with the most identical pair of parameters and forming complex clusters after each step which is repeated until a single cluster comprising all the observations is attained (Lin et al. 2012; Machiwal and Jha 2015). In this study, the Ward’s algorithmic gathering technique subsequent to the squared Euclidean distance was employed. This is measured as the most influential means of clustering (Lin et al. 2012; Machiwal and Jha 2015).

Before the clustering analysis, the detected hydrochemical data, x_ji, was standardized by z-scale transformation as given below (Machiwal and Jha 2015):

$$Z = \frac{{X_{{ji - \dot{x}_{j} }} }}{Sj}$$

where x_ji = value of the jth hydrochemical parameter measured at the ith location, ẋ_j = mean (spatial) value of the jth parameter and S_j = standard deviation (spatial) of the jth parameter (Machiwal and Jha 2015). The clustering achieved with standardized data is anticipated to be influenced less by the large and/or small variance of the hydrochemical data matrix. Also, the influence of diverse measurement units of the data can be removed by making the data dimensionless (Machiwal and Jha 2015). In this study, HCA was performed on a subset of 17 selected variables (pH, EC, Temp., TDS, TH, Ca²⁺, Mg²⁺, Na⁺, K⁺, Fe³⁺, Cu²⁺, Zn²⁺, CO₃²⁻, HCO₃²⁻, Cl⁻, NO₃⁻, PO₄³⁻, and SO₄²⁻), which represented the overall water chemistry outline. The HCA results are presented as a dendrogram (Fig. 10), which offers a graphic summary illustrating an image of the clusters and their closeness with an observed decrease in the dimensionality of original observations. This is measured as a good technique for presenting results of HCA (Lin et al. 2012; Machiwal and Jha 2015). This evidence allowed the construction of a dendrogram as a function of the water quality parameters and sampling locations. Therefore, applying raw data matrix into HCA is an excellent technique which helps classify hydrochemical data based on similarities of sampled parameters across the sampling sites.

Results and discussion

Suitability for drinking

Table 4 summarizes the physicochemical composition of groundwater obtained from boreholes and hand-dug shallow wells. Groundwater composition between the two sources showed a marked variability between the shallow and deep groundwater sources. Groundwater temperature varies significantly (H = 27.36, p = < 0.001) between shallow and deep aquifers. Temperature variability can be very critical especially where biochemical reactions are concerned. Because an increase in temperature by 10 °C in groundwater aquifer leads to doubling of chemical reactions (EPA 2001), the solubility of gasses, ion exchange capacity, redox reaction, sorption processes, complexation, speciation, EC and pH level are all affected by variations in temperature (Ngabirano et al. 2016).

Similarly, pH differs significantly (H = 9.86, p = 0.002) between the two sources of groundwater. Groundwater is slightly acidic to alkaline in the study area. While pH has less effect on consumers, it is fundamental to understanding the chemical composition of groundwater. Moderate pH level is required depending on the composition of groundwater and aquifer properties (EPA 2001). No significant difference in TDS concentration (H = 0.94, p = 0.33), but EC levels differ significantly (H = 21.15, p = < 0.001) between deep and shallow groundwater. Overall, the physical composition of groundwater in the study area indicates water of excellent quality for drinking. Mean pH, TDS, and EC are within World Health Organization (2011) and National Standard for Drinking Water Quality (2007) reference guidelines. Figure 5 illustrates the variability of physical parameters of groundwater in the study area.

Fig. 5

Variability in physical parameters between deep and shallow groundwater a temperature; b pH; c EC, d TDS; and e hardness. DGW deep groundwater, SGW shallow groundwater

Groundwater classification base on TDS, EC, and TH

Table 5 presents the groundwater classification based on TDS, TH, and EC. All the groundwater samples obtained from deep and shallow groundwater sources have TDS concentrations < 500 mg/l. This is especially required for drinking (David and DeWest 1966). Low TDS concentrations have been reported elsewhere in Sokoto Basin (along the Sokoto-Gusau road). Groundwater sources in this area have low TDS—28–79 mg/l (Anderson and Ogilbee 1973; Uma 1993; Alagbe 2006). Groundwater hardness also differs significantly between deep and shallow groundwater (H = 9.015, p = < 0.005). Further, classification base on hardness showed that most of the groundwater sources are moderately hard (Table 5). Current result concurs with previous reports on hardness in Sokoto Basin. Groundwater obtained from the Rima Group is moderately hard to hard in nature (Anderson and Ogilbee 1973; Uma 1993; Alagbe 2006). Groundwater classification based on EC showed that 95% of water sources fall in excellent class and 5% fall in a good class. The relatively low EC levels in the study area are consequent of low TDS, which is generally low in Sokoto Basin.

Table 5 Groundwater classification based on TDS, hardness, and conductivity

Groundwater classification base on chloride and nitrate

Table 6 presents the groundwater classification based on chloride and nitrate pollution in the study area. Based on chloride 75% of water samples fall in fresh class (i.e., 0.85–4.23 mg Cl), 22.5% fall in brackish salt class (i.e., 28.21–546.13 mg Cl) and 2.5% fall in fresh brackish salt (i.e., 4.23–8.46 mg Cl). However, based on nitrate pollution, 22.5% fall in moderate class (i.e., 5–30 mg NO₃) and 77.5% fall in severe class (i.e. > 30 mg NO₃). Since NO₃⁻ is mainly derived from the oxidation of ammonia and agricultural fertilizer, the observed high NO₃⁻ levels in both shallow and deep aquifer is perhaps consequent agriculture and/or other human activities (changes in land use, sewage), but new studies are required for further evaluation. Mean NO₃⁻ concentration is below 50 mg/l in both shallow and deep groundwater. This is an expected outcome since Basement Complex areas are often characterized by one-aquifer system. Therefore, effluents collected in shallow groundwater are transported to deeper groundwater reservoirs. Elevated NO₃⁻ levels in drinking water (> 50 mg/l) can be dangerous to infants—blue baby syndrome (EPA 2001; Wali et al. 2018a, b).

Table 6 Groundwater classification based on chloride and nitrate

The cation chemistry

Figure 6 shows the variability in K⁺, Na⁺, Ca²⁺, Cu²⁺, Fe³⁺, Zn²⁺, and Mg²⁺ concentrations between shallow and deep groundwater in the study area. Potassium concentration differs significantly (H = 4.41, p = < 0.005) between deep and shallow groundwater (Table 4). Elevated K⁺ level in groundwater is associated with toxicity. K⁺ in most aquifers is found in low concentrations and excessive intake is not associated with any health hazard (EPA 2001). Sodium concentration differs significantly (H = 25.41, p = < 0.001). Na⁺ is regulated in drinking water because of the joint effects it exercises with sulfate. High consumption is associated with hypertension (EPA 2001). Na⁺ absorption in an aquifer is dependent on the temperature of the solution and the associated anion. No significant difference in calcium concentration (H = 1.29, p = 0.25). Elevated Ca²⁺ level is often associated with hardness.

Fig. 6

Variability of cations a potassium; b sodium; c calcium; d copper; e iron; f zinc; and g magnesium

There is no significant difference in Cu²⁺ concentration (H = 1.39, p = 0.23). Unpleasant tastes can occur at levels above 1 mg/l (EPA 2001). High Cu²⁺ ingestion in drinking water is not harmful to humans, and therapeutic doses of ~ 20 mg/l are occasionally permitted (EPA 2001). There is a significant difference in Fe⁺ concentration (H = 8.7, p = 0.003). Elevated Fe³⁺ levels in water can be injurious to aquatic animals even though the degree of noxiousness can be reduced by the interactions between other elements. Zinc differs significantly (H = 29.37, p = < 0.001) between the two aquifers. At concentrations level of about 4 mg/l, the unfriendly taste can occur (EPA 2001). At levels between 3 and 5 mg/l, water might look opalescent and can form an oily film when boiled. There was a significant difference in Mg²⁺ concentration (H = 22.74, p = < 0.001). The significance of Mg⁺ in drinking water is that magnesium is the second major constituent of hardness (CaCO₃) (EPA 2001).

Anion chemistry

Figure 7 shows variability in PO₄³⁻, Cl⁻, HCO₃⁻, SO₄²⁻ and NO₃⁻ between deep and shallow groundwater in the study area. There is a significant difference in PO₄³⁻ concentration (H = 29.48, p = < 0.001). The relevance of PO₄³⁻ is mainly related to the rate of eutrophication in surface water bodies (EPA 2001). Chloride differs significantly (H = 29.39, p = < 0.001). Cl⁻ vary widely in natural waters, reaching a maximum level of ~ 35,000 mg/l. Excessive intake does not constitute a health hazard to humans, but at levels above 250 mg/l water will taste salty (EPA 2001). Chloride concentrations ~ 2000 mg/l in drinking water is consumed in arid and semiarid regions. However, elevated Cl⁻ levels in freshwater may render it unfit for irrigation use (EPA 2001). What is important is understanding in a sequence of outcomes from aquifers is that Cl⁻ values are not absolute, rather the relative levels from one sampling point to another. Elevated Cl⁻ levels of ~ 5 mg/l at one location in comparison with other sampling points might lead to the suspicion of groundwater contamination from sewage ejection, especially if ammonia levels are also elevated (EPA 2001). Bicarbonate differs significantly (H = 29.74, p = < 0.001). HCO₃⁻ in conjunction with Ca²⁺ and Mg²⁺ forms carbonate hardness. When groundwater designates high pH concentrations, it can be a sign of high content of carbonate and bicarbonate ions (EPA 2001). Sulfate concentrations differ significantly (H = 7.99, p = 0.004). High SO₄²⁻ in drinking water is associated with the emetic effect, particularly when joint together with Mg⁺ or Na⁺. Nitrate concentration differs significantly (H = 5.22, p = 0.02). High NO₃⁻ in groundwater lead to suspicion of past anthropogenic pollution or high application of composts slurries feast over the land and inorganic fertilizers (EPA 2001).

Fig. 7

Variability of anions a phosphate; b chloride; c bicarbonate; d sulfate; e iron; and f nitrate

Mechanisms controlling water chemistry

Geochemical facies

The origin of groundwater and the process which control groundwater chemistry is understood by the relationships between dissolved elements (Tables 7, 8). It is assumed that a sizable portion of HCO₃⁻ in aquifers originate from the dissolution of carbonate rocks by means of the action of infiltrating rainwaters enriched in CO₂. A Ca-HCO₃⁻ water type is produced when CO₂ is released into solution by the dissolution of carbonate (Marghade et al. 2010). HCO₃⁻ and Ca²⁺ were positively correlated in shallow groundwater (r = 0.32), suggesting that calcite rocks were a source of Ca²⁺. The negative correlation between these ions in the deep aquifer (r = − 0.27) suggests that Ca²⁺ was not exclusively derived from calcite mineral. Other sources of Ca²⁺ in groundwater aquifers are apatite, dolomite, fluorite, gypsum, limestone, and marble. A weak positive correlation between Ca²⁺ and SO₄²⁻ (r = 0.03) perhaps indicates that some parts of Ca²⁺ in shallow groundwater is perhaps derived from gypsum. The negative correlation between these ions in deep aquifer suggests that Ca²⁺ did not derive from gypsum. TDS correlates significantly with EC, Ca²⁺, and Cl⁻ in shallow groundwater and was significantly correlated with Zn²⁺ and Mg²⁺ in deep groundwater. Significant correlations between TDS and these ions suggest that large parts of the dissolved solids in the study area are derived from these ions.

Table 7 Correlation matrix of physicochemical parameters of deep groundwater

Table 8 Correlation matrix of physicochemical parameters of shallow groundwater

Weak correlation between HCO₃⁻ and Ca²⁺ (r = 0.32) in shallow groundwater and negative correlation (r = − 0.27), between these ions in deep groundwater, suggest that dissolution of gypsum may not be the source of Ca²⁺ and SO₄²⁻. Chloride correlates positively but weakly with Na⁺ in both shallow (r = 0.40) and deep groundwater (r = 0.20). Positive correlations between these two ions suggest that some parts of Na⁺ in the study area are derived from halite (Marghade et al. 2010).

Poor correlations (r = 0.11, r = 0.13), between Ca²⁺ and Mg²⁺ in both shallow and deep groundwater, suggest that the two ions might not have the same source. SO₄²⁻ and Mg²⁺ were positively correlated in shallow groundwater (r = 0.40), suggesting that the two ions might likely have the same origin. But deep groundwater the two ions were negatively correlated (r = − 0.36), indicative that SO₄²⁻ and Mg²⁺ are not derived from the same source. A charge equilibrium occurs between cations and anions when Ca²⁺, Mg²⁺, SO₄²⁻, and HCO₃⁻ + SO₄²⁻ originate from the simple dissolution of gypsum, dolomite, and calcite (Marghade et al. 2010).

Silicate weathering and ion exchange process

Silicate weathering reaction in the study area was evaluated using the Na⁺/Cl⁻ molar ratio (Table 9). About 55% of the analyzed water samples from deep groundwater sources have a molar ratio greater than 1, whereas 85% of water samples from shallow aquifer have a molar ratio greater than 1. This suggests that some parts of Na⁺ were derived from silicate weathering. However, the process of cation exchange Ca²⁺, Mg²⁺, and Na⁺ may produce higher levels of Na⁺. When groundwater samples have a molar ratio greater than 1, it indicates deficiency in Mg²⁺ + Ca²⁺ which is equivalent to Ca²⁺–Na⁺ cation exchange process, resulting in softening of water. More so, in aquifers having clay mineral, Na⁺ derived from the exchangeable sites can exchange with Ca²⁺ and Mg²⁺, causing elevated Na⁺ level (Marghade et al. 2010).

Table 9 Calculated chemical indices for evaluating rock weathering process

Conversely, Scholler index (Si) (Scholler 1965) can be used to evaluate the ion exchange process (Table 9). Water samples from both deep and shallow groundwater in the study area have positive Si, suggesting overall base exchange reactions in the underlying aquifer. In aquifers where alkaline rock minerals are exchanged with Na⁺ ions, (HCO₃ > Ca + Mg) indicates base exchange soft water. Hardened water is formed when Na⁺ ions are exchanged with alkaline rocks (Ca + Mg > HCO₃). Similarly, Versluy’s index (Versluys 1916), was positive in both shallow and deep groundwater, suggesting overall base exchange reaction in both shallow and deep groundwater (Table 9). Table 10 summarizes the geochemical aspects of groundwater samples (Panaskar et al. 2016). Two faces can be discerned. The alkaline earth (Ca + Mg) is greater than the alkalis in shallow groundwater, whereas the alkalis (Na + K) are greater than the alkaline earth in deep groundwater.

Table 10 Geochemical characterization of groundwater samples

Hydrogeochemical faces

The Piper trilinear diagram (Fig. 8) (Piper 1944), shows that samples derived from deep groundwater fall in the class of Ca–Mg–SO₄–HCO₃, and Mixed Mg–Na–K–SO₄ water type, whereas in the shallow aquifer, groundwater falls in Ca–Na–K–Cl–HCO₃ and mixed HCO₃–Cl–SO₄ water type. The hydrogeochemical faces in Sokoto Basin are mainly of two types: Ca–Mg–HCO₃ and Ca–Mg–SO₄–Cl. These faces perhaps are derived from dissolution of Ca²⁺ and Mg²⁺ carbonates (Anderson and Ogilbee 1973; Uma 1993; Alagbe 2006). The observed slightly variability in the faces can be related to geology, as most previous reporting of groundwater, faces come from the Cretaceous and Tertiary section of the basin. But new studies comparing groundwater faces between the Basement Complex and Cretaceous and Tertiary sections of the Sokoto Basin are required for further evaluation.

Fig. 8

Piper diagram showing the chemical composition of shallow and deep groundwater

Gibbs model

The mechanism controlling water chemistry was further evaluated using a plot of weight ratio of TDS versus [Na + K]/Na + K + Ca] and [Cl]/[Cl + HCO₃] for cations and anions, respectively (Fig. 9). The model reveals that rock weathering is the dominant mechanism controlling groundwater chemistry in the study area (Gibbs 1970; Rakotondrabe et al. 2018). The observed mechanism is perhaps derived from the Geology and Geography of the study area, which is in the Sokoto Basin (semiarid). The lithology here is mainly comprised of sands and clays of different textural classes overlying a crystalline basement complex (CBC). The lithologic logs from CBC in southern parts of Kebbi State show the lithology is mainly comprised of gneisses, granites, quartzite, and schist (Anderson and Ogilbee 1973; Kogbe 1986; Offodile 2002).

Fig. 9

Gibbs plot showing major natural mechanisms controlling groundwater chemistry

Anthropogenic inputs

Variations of TDS in groundwater are due to contamination from anthropogenic sources (Marghade et al. 2010). Ions including Na⁺, Cl⁻, SO₄²⁻, and NO₃⁻ are mainly derived from anthropogenic sources—sewage ejections from municipal and industrial sources as well as the application of chemical fertilizer and manure. Correlations between Na⁺, Cl⁻, SO₄²⁻, and NO₃⁻ with TDS indicates how anthropogenic activities accelerate changes in groundwater composition (Marghade et al. 2010). TDS correlates positively with Na⁺, Cl⁻, Ca²⁺, SO4²⁻, and NO₃⁻. Negative correlations between TDS and Na⁺ in shallow and deep groundwater informs silicate weathering reaction was not the source of Na⁺ (Tables 7, 8). Municipal and industrial sewage and effluents from mining and engineering works can result in a positive correlation between Na⁺ and TDS (Dragon 2008). Significant correlations between TDS and Cl and positive correlation between TDS and SO₄ in shallow groundwater suggest input from anthropogenic origins. Negative correlations between TDS and NO₃⁻ in both shallow and deep groundwater (Tables 7, 8) suggest that groundwater in the study area is less influenced by anthropogenic activities.

However, the topographical setting of the study area may reduce the anthropogenic impact, since effluents may be washed off by a single storm. So, impurities may be transported downhill to the nearby river floodplains and transported downstream. Since Basement Complex aquifers are primarily recharged along cracks and other openings, the chances of getting contaminants may be low if the recharge zones are located away from the source of pollutants. This perhaps explains the low anthropogenic inputs in the study area. Yet new studies are required for further evaluation. Positive correlations between TDS and SO₄²⁻ also indicates contamination from the anthropogenic origin in the absence of geological inputs and this cannot be totally ruled out in the study area, owing to its geology. SO₄²⁻ and NO₃⁻ correlate negatively (r = − 0.29; r = − 0.06) in shallow and deep groundwater suggesting the different origin of the two ions, perhaps derived from both the rock mineral and anthropogenic activities (Dragon 2008). Weak correlations between Cl⁻ and NO₃⁻ (r = 0.07; r = 0.41) in both the shallow and deep groundwater suggest that the two ions might have originated from a different source. Negative correlations between Ca²⁺ and NO₃⁻ (r = − 0.23; r = 0.04), NO₃⁻ and SO₄²⁻ (r = − 029; r = − 0.06) and weak positive correlation between Cl⁻ and NO₃⁻ (r = 0.07; r = 0.41) in both shallow and deep aquifers made it very difficult to relate these ions to anthropogenic inputs.

Principal component analysis

In this paper, PCA was carried out using two sets of groundwater quality data, comprising of 20 groundwater samples each from shallow and deep groundwater sources, to identify and describe the factors that affect groundwater chemistry of the two groundwater environments. PCA was performed on a subset of 17 selected water quality variables, namely: Temperature, EC, pH, TDS, TH, K⁺, Na⁺, Ca²⁺, Cu²⁺, Fe³⁺, Zn²⁺, Mg²⁺, Cl⁻, HCO₃⁻, PO₄³⁻, NO₃⁻, and SO₄²⁻, which characterized the overall groundwater chemistry outline. The PCA results comprising the loadings, eigenvalues, and percentages of total variance are summarised in Table 11. The extraction of three factors each from shallow and deep groundwater was built on the proportion of variance accrued, which included a percentage greater than 80% (Lin et al. 2012; Selvakumar et al. 2017). The computed data indicate these three factors explained 58.63% and 54.98% of the total variance within the data matrix in shallow and deep groundwater, respectively. Based on ‘scree test,’ the three factors with typical factor loadings suggest three different noticeable contributions were involved in controlling the composition of shallow and deep groundwater in the study area. Factor 1 explained 27.195% and 27.045% of the variability of groundwater data in shallow and deep groundwater, respectively. In shallow groundwater, Factor 1 has corresponded to high positive loadings on pH, TDS, EC, K⁺, Ca²⁺, and Cl⁻ as well as TDS, EC, and Mg²⁺ in deep groundwater. Factor 1 can be related to rock weathering since all the parameters having high positive loading on these factors have a geologic origin.

Table 11 Varimax rotated R-mode factor loadings matrix

Potassium, for instance, is derived from many minerals through the weathering process. Such minerals include microcline and orthoclase feldspars, chlorine minerals, sylvite, carnalite, and clay minerals. Calcium in groundwater is primarily derived from rock minerals such as apatite, calcite, dolomite, fluorite, gypsum, limestone, and marble. In most natural aquifers, Cl⁻ occurs as a key ion, yet Cl⁻ is continuously added in freshwaters by anthropogenic activities in many receiving glasses of water (Gregory et al. 2015). Factor 2 explained 17.33% and 14.09% of the variability in shallow and deep groundwater, respectively. The factor corresponded to high positive loadings on Na⁺ and NO₃⁻ in shallow groundwater, pH, HCO₃⁻ and SO₄²⁻ in deep groundwater. A large part of Na⁺ comes from rocks and soil, whereas, NO₃⁻ is derived in aquifers from agriculture via runoff and oxidation of ammonia. Bicarbonate comes from carbonates dissolution and like nitrate, and SO₄²⁻ can be derived from both geogenic and anthropogenic sources. Therefore, Factor 2 can be related to both geogenic and anthropogenic influence. High positive loadings on HCO₃⁻ in deep groundwater inferred that it is expected to control the character of other parameters (Lin et al. 2012). Factor 3 corresponded to high positive loadings on K⁺ and Zn²⁺ in shallow groundwater, K⁺, Na⁺, and NO₃⁻ in deep groundwater water. Factor 3 can be related to rock weathering in shallow groundwater since both Na⁺ and Zn²⁺ are derive from rock weathering and soil. In deep groundwater, this factor can be related to both geologic and anthropogenic influence. However, negative loading on pH in PC1 in the deep aquifer was deemed reasonable since pH attained a converse relationship with rocks of carbonate origin (Lin et al. 2012).

Despite the variability in water table depths, groundwater composition in the study area is controlled by rock weathering. Using scatter plot, water quality parameters cluster nicely along the spectrum of shallow and deep groundwater, respectively, in a biplot of PC1 and PC2 (Fig. 10a, b). The two components joint together explained 58.626% (shallow groundwater) and 54.975% (deep groundwater) of the variability within the data set. Interestingly the results are remarkable as the parameters form a comparatively tight cluster around the water sources. All the 17 originally analyzed variables (or parameters) are clearly visible on the biplot, even though these elements are selected on the basis of their absolute size (i.e., high elemental absorptions) which was established by reference to the raw data matrix. As a result, the significance or otherwise of these parameters is identified from this plot (Kokot et al. 1994, 1998; Kokot and Stewart 1995; Olsen et al. 2012; Zhang et al. 2015). Although a good clustering is observed in the biplot of PC1 and PC2, the biplot of PC2 and PC3 (Fig. 10c) indicates wide dispersity of groundwater quality parameters, making it difficult to identify any pattern. It is not clear which of these later components is important.

Fig. 10

a PCA biplots for examining relationships between groundwater quality parameters from 40 groundwater sources. b Biplot of PC1 versus PC2 of hydrochemical data matrix (y-mean centered) for 17 parameters analyzed from 40 groundwater samples. c Biplot of PC2 versus PC3 of hydrochemical data matrix for seventeen parameters analyzed from 40 groundwater samples

Hierarchical cluster analysis

The vital role of HCA is to discrete the parameters in a raw data matrix into separate classes designated by the user without any prior hypotheses. The HCA is an unverified classification technique that reveals inherent assembly or recognizing the pattern of a dataset without a prior hypothesis with regards to the data so that the objects of the system can be classified into clusters based on their resemblances (Lin et al. 2012; Machiwal and Jha 2015). In this study, Ward’s algorithmic clustering technique subsequent to the Euclidean distance was used. This is considered as the most powerful grouping tool (Dou et al. 2008; Shyu et al. 2011; Lin et al. 2012), because it is capable of minimizing the misrepresenting effect or sum of squared distances of centroids from two theoretical groups produced at each step (Lin et al. 2012). Using HCA, groundwater sources with comparable hydrochemical physiognomies can be grouped into the same cluster. The graphics collection of the clustering processes in this study is offered as a dendrogram. Base on the dendrogram (Fig. 11), three groups of groundwater sources can be distinguished.

Fig. 11

Dendrogram produced from cluster analysis based on sampling boreholes to understand the most important hydrochemical physiognomies in Basement Complex areas of southern Kebbi State. Three groups were distinguished viz. Group 1 (BH1, BH7, BH11, BH12, BH5, BH6, BH8, BH10, BH19, BH20, BH9 and HD15), Group 2 (BH2, BH14, BH3, BH18, BH4, BH16, BH15, BH17and HD1) and Group 3 (HD5, HD7, HD16, HD11, HD19, HD12, HD4, HD66, HD8, HD20, HD9, HD10, HD13, HD14, HD18, HD2, HD3 and HD17)

Group 1 (BH1, BH7, BH11, BH12, BH5, BH6, BH8, BH10, BH19, BH20, BH9 and HD15), Group 2 (BH2, BH14, BH3, BH18, BH4, BH16, BH15, BH17and HD1) and Group 3 (HD5, HD7, HD16, HD11, HD19, HD12, HD4, HD66, HD8, HD20, HD9, HD10, HD13, HD14, HD18, HD2, HD3 and HD17).Generally, the lithology of these boreholes is comprised of clay, ironstone, schists, and sands of different textural groups. Group 1 is characterized by a higher temperature, magnesium, pH, TDS, EC, and calcium. Group 2 corresponded to groundwater sources having higher potassium, chloride, sodium, sulfate, phosphate, zinc and bicarbonate. Group 3 is comprised of water sources having higher concentrations of copper, iron, and nitrate. Group 1 can be related to natural geogenic processes as it is comprised of ions derived primarily from rock minerals (notably Ca²⁺). Similarly, Group 2 can be related to natural geogenic processes. Lastly, Group 3 can be related to both anthropogenic (notably NO₃⁻ pollution) and rock weathering.

Suitability for irrigation use

The suitability of groundwater for irrigation can be evaluated using a total concentration of soluble salts (SAR) which is express in terms of specific conductance (Sadashivaiah et al. 2008; Rakotondrabe et al. 2018). Sodium percent was calculated following Kumar et al. (2014) and Rakotondrabe et al. (2018). Soil often reacts with sodium to decrease soil permeability, which results in high sodium concentration. This causes cation exchange between Mg²⁺ and Ca²⁺ in soil under wet conditions. Water and air circulations are reduced as a consequent to this process. In the same vein, Kelly’s index and magnesium adsorption ratio (MAR) were also calculated. Kelly’s index value greater than 1 designates water of excellent quality for irrigation use. However, index value less than 1 designates water which is unsuitable for irrigation, because of alkali hazard to crops (Kelly 1940). In calculating Kelley’s index, Ca²⁺ and Mg²⁺ are measured against Na⁺. Elevated levels of Mg²⁺ in groundwater disturb the soil quality by changing it to alkali, which subsequently decreases crop yield (Kumar et al. 2014). Also, values of MAR less than 50 in irrigation water are considered appropriate for irrigation use. Often in groundwater, Ca²⁺ and Mg²⁺ are found in a state of equilibrium. Elevated Mg²⁺ in groundwater at levels greater than Ca²⁺ fast-tracks the degree of Mg²⁺ saturation, which destroys soil structure, and consequently, reduces its productivity (Goswamee et al. 2015).

The chemical composition of groundwater for irrigation farming affects the soil quality and crop yield (Table 12). The suitability of groundwater for irrigation use hinges on the composition and concentration levels of ions in groundwater (Panaskar et al. 2016). To assess this, the properties of groundwater play an important role including (1) comparative amount of sodium to other cations, (2) total absorption of soluble salts, (3) the bicarbonate absorption in relation to the absorption of Ca + Mg (USSL 1954; Panaskar et al. 2016). Using this, an assessment of groundwater appropriateness for irrigation uses has been carried out in the study area using numerous salinity indices viz. sodium adsorption ratio (SAR), sodium percentage (Na%), Kelly’s ratio, magnesium ratio, residual sodium carbonate (RSC), and USSL diagram among others (Panaskar et al. 2016). Table 13 presents the calculated indices from the shallow and deep groundwater in the study area.

Table 12 Summary of chemical indices used in determining groundwater suitability for irrigation

Table 13 Summary of chemical indices of shallow and deep groundwater in the study area

Sodium adsorption ratio (SAR)

The foremost index applied for evaluation of groundwater suitability for agricultural purposes in relation to Na⁺ absorption which affect soil structure and permeability is SAR (Marghade et al. 2010; Panaskar et al. 2016). SAR is used to measure sodium and/or alkali hazards to crops (Table 13). It provides a clue relating to the magnitude of sodium adsorption by soil. Calcium and Mg²⁺ in the soil can be displaced as a result of the prolonged application of high Na-rich water. This causes damage to the soil structure and reduced crop yields (Marghade et al. 2010; Panaskar et al. 2016). SAR in deep groundwater ranged from < 0.001 to 1.9 with a mean value of 0.7, whereas, in shallow groundwater, it varies between 0.3 and 211.8 with an average value of 103.26. Figure 12 shows the variability in SAR levels between shallow and deep groundwater in the study area.

Fig. 12

USSL classification of groundwater samples from shallow and deep aquifers

Groundwater samples having SAR value within the range of 10 and 18 are classified as a good category; however, most of the groundwater samples derived from deep aquifer have SAR value > 18, with one groundwater sample (SAR < 18), classified as good.

Sodium percent (Na%)

Another index employed to identify the concentration of soluble sodium level in irrigation water is sodium percent (Na%). It is applied to reveal sodium hazard to crops (Panaskar et al. 2016). Calcium is replaced by sodium by process of base exchange which destroys soil structure and consequently reduces soil permeability (Table 12). Using Wilcox plot (Fig. 13), groundwater samples showed marked variability with shallow groundwater samples falling in good to excellent class. Whereas deep groundwater samples vary between good to excellent class and unsuitable to doubtful class, two samples from deep groundwater fall in unsuitable class. Generally, samples from shallow groundwater have very low to high sodium class. However, irrigation water of very low SAR and low salinity (< 200 µS/m) affects the rates of water infiltration into soils (Graham et al. 2006).

Fig. 13

Wilcox diagram showing groundwater classification for irrigation use

Permeability index (PI)

Prolonged application of irrigation water can interfere with soil permeability which is affected by bicarbonate, calcium, magnesium, and sodium (Li et al. 2016; Panaskar et al. 2016). Doneen (1962) developed a standard for evaluating irrigation water suitability using permeability index which is defined as contained (Fig. 14).

Fig. 14

Doneen’s chart showing groundwater classification of irrigation water based on PI

Values are express in meq/l. Shallow groundwater classification showed that PI values ranged from 2.6 to 86.2 with a mean value 29.8. In deep groundwater, PI values vary between 0 and 1.9 with an average value of 0.74. Based on permeability index, three classes can be discerned: Class I is classified as excellent for irrigation; Class II is classified as acceptable; Class III is categorized as unsuitable (Li et al. 2016). All the shallow groundwater samples fall in Class III. This is deemed reasonable since the prolonged application of irrigation water of very low salinity is related to poor permeability (Graham et al. 2006). In deep groundwater, 1 sample falls in Class III, 14 samples fall in Class II, and 5 samples fall in Class I. Water samples from deep aquifer can be used for irrigation with little permeability problem, but such water of high SAR level destroys soil structure (Li et al. 2016; Panaskar et al. 2016).

Kelly’s index (KI)

Kelly’s index is employed to measure the concentration of Na⁺ against Ca²⁺ and Mg²⁺ (Table 13). It is vital to evaluate groundwater suitability for agricultural use. Elevated Na⁺ level in aquifers produces an objectionable impact on changing characteristics of soil and its permeability; thus, extreme Na⁺ is used as an indicator of alkali hazard (Wagh et al. 2016). Kelly’s index is greater than 1 in deep groundwater, whereas, in shallow groundwater, indices are less than 1 (Table 13). Values less than 1 are considered suitable for irrigation, whereas KI values are greater than 1 in irrigation water can be classified as unsuitable for irrigation use (Wagh et al. 2016). KI values in shallow groundwater ranged from 0 to 0.4 with a mean value of 0.13. KI values are extremely high in shallow groundwater and vary between 0 and 92.9 with an average value of 28.06. The possible reasons for higher KI values in shallow groundwater need to be unveiled as it presents a serious environmental problem.

Magnesium adsorption ratio (MAR)

Magnesium hazard is also used to characterize groundwater and assess its suitability for irrigation use by revealing the amount of Mg²⁺ over Ca²⁺ (Ragunath 1987; Wagh et al. 2016). The MAR ratio as calculated in Table 13 shows that MAR values ranged from 13.0 to 90.8 with a mean value of 58.3 in shallow groundwater. In deep groundwater, MAR values vary between 4.8 and 76.9 with an average value of 30.17. Based on mean MAR, shallow groundwater in the study area is unfit for irrigation use. Usually, magnesium and calcium occur in a state of equilibrium in groundwater sources. The concentration of magnesium beyond calcium in irrigation water fast-tracks the magnitude of Mg²⁺ saturation and consequently damages soil structure and reduces crop yields. High magnesium in irrigation water impacts soil quality by converting it to alkali which eventually reduces its productivity. The variability of MAR in the study area is further illustrated using a radar chart (Fig. 15). Overall, groundwater evaluation shows deep aquifers in the study area hold water of good quality as compared to shallow groundwater sources.

Fig. 15

Radar chart showing the variability of magnesium adsorption ratio between shallow and deep groundwater

Implications for groundwater quality in Sub-Saharan Africa

Characterization of groundwater in the study area showed marked variability between aquifers. Groundwater parameters showed marked variability between the shallow and deep groundwater. To better understand the composition of groundwater in the rest of SSA, data from 50 locations from the literature was compiled (Table 14). Further Table 14 shows that evaluation of groundwater quality in SSA is constrained by lack of data. Results from the patchy data derived from the literature indicate that 16.7% (n = 24) of groundwater in SSA have temperatures above 30 °C, 90.0% (n = 40) have EC values above WHO reference guidelines (1000 µS/cm). At least, 38.6% (n = 44) of groundwater sources in SSA are acidic to alkaline in nature. Scholler and Versluy’s indices are positive, indicating overall base exchange reactions in aquifers underlying SSA. Kelly’s index is greater than 1 in about 10 locations, indicating water which is unsuitable for irrigation use.

Table 14 Compilation of the literature report of physical and chemical properties of groundwater in some parts of sub-Saharan Africa

Sodium percent is greater than 20 in most locations in SSA (Table 15). The high rates of sodium indicate the absence of ion exchange reaction between Ca²⁺ and Na²⁺. Wilcox plot indicates that most of the groundwater sources in SSA are suitable for irrigation use (Fig. 16). However, some parts of Na⁺ in groundwater aquifers across SSA were not derived from silicate weathering, because the molar ratio is greater than 1 in most locations. Therefore, groundwater in SSA is not deficient in Mg²⁺ + Ca²⁺. As a result, groundwater sources may be hard in most parts of the continent. Magnesium hazard is less than 50 in most locations.

Table 15 Summary of chemical indices of groundwater in some Sub-Saharan Africa

Fig. 16

Wilcox plot showing irrigation water classification in sub-Saharan Africa

Conclusion

The literature is unanimous on the importance of evaluation and understanding of the hydrochemical composition of groundwater under different environments. This study reveals that:

• 1. The concentrations of Fe³⁺, Zn²⁺, Mg²⁺, Na⁺, PO₄³⁻, and SO₄²⁻ are above World Health Organization (2011) and National Standard for Drinking Water Quality (2007) reference guidelines. Groundwater in the study area is acidic to alkaline in nature;

• 2. There is a significant difference in concentrations of temperature, pH, EC, hardness, Na⁺, Zn²⁺, Mg²⁺, PO₄³⁻, Cl⁻, HCO₃⁻, SO₄²⁻, and NO₃⁻ between shallow and deep groundwater sources;

• 3. Groundwater classification using TDS showed that water sources are suitable for drinking and irrigation use;

• 4. Groundwater classification base on hardness showed that most of the groundwater sources are moderately hard;

• 5. Groundwater classification based on EC showed that most of the water sources fall in excellent class. The relatively low EC levels in the study area are consequent of low TDS, which is generally low in Sokoto Basin;

• 6. Groundwater classification based on chloride revealed that most of the water samples fall fresh class, but most groundwater sources fall in severe class based on NO₃⁻ classification;

• 7. Evaluation of geochemical faces showed that groundwater composition is controlled by rock weathering, even though little inputs were derived from anthropogenic sources. This was further confirmed by the Gibbs model;

• 8. Result of Na +/Cl- molar ratio indicates that most of the sampling locations have molar ration greater than 1, which suggests that some parts of Na⁺ were derived from silicate weathering;

• 9. Most groundwater samples have positive Scholler index, suggesting overall base exchange reactions in the underlying aquifer;

• 10. Groundwater classification based on Piper diagram revealed two types of faces: Ca–Mg–HCO₃ and Ca–Mg–SO₄–Cl;

• 11. Significant correlations between TDS and Cl⁻, and positive correlation between TDS and SO₄ in shallow groundwater suggest input from anthropogenic sources;

• 12. Groundwater evaluation using PCA, showed water sources are more influenced by rock mineral as compared to anthropogenic inputs;

• 13. Groundwater analysis using HCA showed that Group 1 and 2 can be related to natural geogenic processes as it corresponds to ions of natural geogenic origin. Group 3 can be related to both anthropogenic (notably NO₃⁻ pollution) and rock weathering;

• 14. Most of the groundwater samples derived from deep aquifer have SAR value > 18, whereas, all the groundwater samples from shallow aquifer have SAR < 18;

• 15. Base on Na⁺% samples from shallow groundwater fall in very low to high sodium class, whereas, deep groundwater samples vary between good to excellent;

• 16. Based on Doneen’s chart, permeability indices are excellent in deep groundwater, whereas, in shallow groundwater, PI indices are affected by low salinity water type;

• 17. The possible reasons for higher KI values in shallow groundwater need to be revealed as it presents a serious environmental problem; based on mean MAR, shallow groundwater in the study area is unfit for irrigation use;

• 18. Many sources of groundwater in SSA are acidic in composition; Scholler and Versluy’s indices suggest an overall base exchange reaction in aquifers underlying SSA; and

• 19. Kelly’s index is less than 1 in most locations in SSA, indicating water which is suitable for irrigation use. MAR is less than 50 in most locations.

Despite the significant difference in water quality parameters and chemical indices, evaluation of shallow and deep groundwater in the study area indicates that groundwater composition is more a product of rock weathering than anthropogenic inputs.