The study area is located in the middle-west of Senegal between longitudes 16°38′ to 17°00′ W and latitudes 14°09′ to 14°26′N with an estimated total area of about 880 km² (Figure 1). Annual average precipitation and evaporation in the study area are ~600 and 1,600 mm, respectively. The climate is of tropical type, characterized by two climatic seasons, namely, a rainy season from June to October and a dry season from November to May. The annual temperature ranges from 20°C in December to 40°C in May, and the relative humidity ranges from 60 to 96%. The topography inclines from the northwest to the southeast, and the surface elevation varies from 0 m in the estuary system to more than 20 m above mean sea level inland (Figure 1). The savannah covers almost the entire region with the exception of the littoral and estuarine zones where a particular halophyte vegetation develops. Several soils exist in the study area (Maignien, 1965; Sarr, 1982): the black to gray-black clays cover depressions; dune sands, locally called “Joor,” extend well to the east, south, and extremity northwestern of the region; the saline soils, locally named “tanne”; and the black or black-gray salted soils saturated around the tributaries of the Saloum estuary.

The lithological sections (Sarr, 1999) show, at very variable levels, detrital soils with a predominance of clay and sand that constitute the layers of the Continental Terminal (Mio-Pliocene) and the Quaternary (Figure 2). The Eocene is generally characterized by marly or clayey facies enriched in limestone in the upper part (Tessier, 1954; Flicoteaux and et Lappartient, 1972). The base of Eocene consists of clayey limestone, marls and clays with quartz, flint, and phosphate (Sarr, 1982; Saint-Marc and Sarr, 1984). The Paleocene that outcrops around the town of Mbour sinks east and south-east under the Eocene formations. Paleocene is characterized by a rather great homogeneity of the calcareous and marly limestone facies, often shells (Monciardini, 1966). These shell limestones, slightly marly, karstified, or fissured, often sandstone, occupy the middle and upper horizon of the Paleocene. In the whole region, the base of the Paleocene is constituted by hard limestones and gray marly limestones, which can be sandstone (Tessier, 1954; Sarr, 1999). Beneath this formation lies the clays and sands of the Maastrichtian. Various shallow and deep aquifer systems are exploited in the study area. The Mio-Pliocene and Quaternary aquifer constituted by the clayey sands represents the marly limestone aquifer of the Upper Ypresian, the superficial system. In some areas (Figure 1: Ngueniene and Thiadiaye logs), the Eocene and Miocene-Quaternary layers form a single aquifer. In the Mio-Plio Quaternary aquifer, hydraulic conductivity is variable with mean values of 1.5 × 10⁻⁴ m/s and effective porosity of 20% (Geohydraulique OMS, 1972; Madioune et al., 2014). The deep system, essentially tapped by boreholes, wells, and wells-boreholes, is made up of the deep aquifer of the Paleocene and the Maastrichtian. The Paleocene water table is currently the most exploited with depletion or salinization in some sectors of the surface water tables (Tine et al., 2011). The hydraulic conductivities in Paleocene aquifer range from 6.6 × 10⁻⁶ to 2.0 × 10⁻² m/s with a storage coefficient that varies from 1 × 10⁻⁴ to 7 × 10⁻². Currently, network functioning depends on the importance of the annual rainfall. This hydrographic network is divided into two groups: the small marigots flowing toward the ocean (MBaling, Warang, Nianing, and Mbodiene) and the tributaries of the Saloum (Figure 1). The latter (Foua, Balabougou, Joal, and Ngueniene) collect all of the water from the thalwegs. The Saloum River is the main collector of the region under marine water pressure.

A total of 42 samples from 22 wells and 20 boreholes (ranging in groundwater level depth from 2 to 27 m) were collected in polyethylene bottles in September 2019. Water bottles were rinsed three times prior to sampling, while electrical conductivity (EC), pH, and total dissolved solids (TDS) were measured in the field by a multi-parameter WTW (Model: Multi 350i/SET, Germany). Alkalinity measurements were also performed using a digital titrator (HACH). Samples for major ions analysis (Cl⁻, NO3-, SO42-, HCO3-, Na⁺, K⁺, Mg²⁺, and Ca²⁺) were collected in two 30 ml polyethylene bottles after filtration through 0.45 μm nitrocellulose membranes. Samples for cations analysis were acidified with ultrapure HNO3- at 70%. Bottles were filled airtight with no bubbles and kept at 4°C until analyses. Chemical analyses were conducted at the Laboratory of Chrono-Environment, University of Bourgogne Franche-Comte, France, using an ion chromatographic system (ICS) for anions and an inductively coupled plasma-mass spectrometry (ICP-MS) for cations. Ionic balance calculations in 95% of the samples were within ±10%.

Piper (1944) and Chadha (1999) diagrams were used to identify the processes controlling the groundwater hydrochemistry and groundwater types. These diagrams were constructed using DIAGRAMME Software (Simler, 2009). Multivariate statistical methods were also used to classify groundwater and interpret hydrogeochemical processes (Brown, 1998; Wang et al., 2015). The software STATISTICA was used for statistical analysis of the hydrochemical data (StatSoft Inc, 2004). Several studies have used this technique for water types classification and determined the major factors controlling groundwater chemistry (Melloul and Collin, 1992; Farnham et al., 2000; Güler et al., 2002; Güler and Thyne, 2004; Hussein, 2004; Helstrup et al., 2007; Cloutier et al., 2008; Yidana et al., 2008; Montcoudiol et al., 2014). Among the multivariate statistical methods, the hierarchical cluster analysis (HCA) was used for this project. The HCA allows the identification of groups of groundwater samples on the basis of their similar hydrogeochemical components (Grande et al., 2003). Ward's linkage method was applied, and the Euclidean distance was chosen as the similarity measure to generate the main clusters (Güler et al., 2002).

The processes controlling groundwater hydrochemistry have also been determined by binary diagrams. The base-exchange reaction was quantified by the calculation of chloroalkaline indices (CAI) (Schoeller, 1934, 1967; Ndoye et al., 2018). Schoeller (1934) established two equations for the calculation of the chloroalkaline indices (CAI-I and CAI-II) that allow the quantification of ions' exchange between groundwater and the host environment. In the following equations, the unit for the concentration of all ions is mEq/L.

The analytical results of groundwater samples are presented in Table 1. The groundwater pH ranges from 7.1 to 8.2 with an average value of 7.6, indicating neutral to slightly alkaline water. The highest pH values are observed in the central-east zone in the localities of Guedj (43), Ngueniene (11), Ndiagamba (27), Pethemakh (9), and Ngohe (25). The electrical conductivity (EC) displays a large variability from 167 to 8,880 μS/cm, with an average value of 1,519 μS/cm (Table 2), showing that very heterogeneous water chemistry. High EC values (values >1,500 μS/cm) are observed in wells located in the coastal zone such as Ngazobil (15), Nianing (39), Sidibougou (40), Roff K. Seck (16), and Mbodiene (13). High EC values are also observed inland to Ngueniene (11), Balabougou (19), Ndiol Khokhane (31), Bagana Serere (35), and Ndiemane (18). These wells have shallow depths of 3 to 7.8 m and are located near the Balabougou and Ngueniene stream, suggesting an influence of saline surface water. Total dissolved solids (TDS) vary between 120 and 5,382 mg/L, with an average of 1,017 mg/L. Most of the samples (80%) have TDS values less than the allowable limit of 1,000 mg/L by WHO drinking water standards (WHO, 2011). The elevated EC and TDS are ascribed to the soluble electrolytes and high salinity (Brahman et al., 2013; Raza et al., 2016). The concentrations of Ca²⁺, Mg²⁺, Na+, and K⁺ ions vary from 17.8 to 562 mg/L (mean value = 103 mg/L), 2.7 to 168 mg/L (mean value = 46.2 mg/L), 8.1 to 1,106 mg/L (mean value = 142 mg/L), and 0.3 to 10.7 mg/L (mean value = 3.2 mg/L), respectively (Table 2). The source of Ca²⁺ and Mg²⁺ in this region is the dissolution of carbonate and silicate minerals as observed in other regions (Saha and Safiur Rahman, 2020; Zhou et al., 2020). In 2020, a study by Marghade (2020) provides the origin of Na⁺ from the dissolution of Na-containing lithogenic minerals, cation exchange, and/or mixing with seawater. Among the anions, Cl⁻ is the dominant ion followed by HCO3-, SO42-, and NO3-. The chloride concentration in the groundwater varies between 11.5 and 2,626 mg/L (mean: 280 mg/L). Dissolved chloride is common in shallow groundwater (Hem, 1992; Faye et al., 2003). This element comes from atmospheric precipitation, dissolution of salt deposits, weathering of halite and evaporite minerals (Hussien and Faiyad, 2016), and/or mixing with seawater. During dry periods, white efflorescence of salt deposits is observed. This is dissolved by rainwater and percolates toward shallow waters leading to a significant increase in the Na+ and Cl contents.

Concentrations of HCO3- vary between 9.4 and 541 mg/L with a mean of 297 mg/L. The bicarbonate could originate from the same sources as Ca and Mg, i.e., carbonate precipitation and dissolution of silicate minerals (Belkhiri et al., 2012; Asmael et al., 2015; Chebbah and Allia, 2015; Saha and Safiur Rahman, 2020). HCO3- also could originate from the biodegradation of organic soil compounds (Winter et al., 1998; Magha et al., 2021). Sulfate concentrations vary between 1.2 and 527 mg/L, with a mean of 71 mg/L. Sulfate (SO42-) source comes from the weathering of rocks forming minerals such as gypsum and anhydrite (Datta and Tyagi, 1996), or from the mixing of groundwater with seawater. Additionally, high sulfate concentrations may be due to the influence of leaching from sulfate-acid soils found on the “tannes” and anthropogenic activity.

The concentration of NO3- in the groundwater ranges from 0.0 (< LOD 0.05) to 432 mg/L with a mean of 37.2 mg/L. Nitrate is a frequent contaminant in groundwater in arid and semi-arid regions and emanates from anthropogenic activities (Adimalla et al., 2019, 2020; Perez Villarreal et al., 2019; Mufur et al., 2021). High concentrations of NO3- in the studied area are observed in Nianing (39), Ndianda (34), Ngueniene Serere (12), and Mbourock Cisse (13) and are attributed to anthropogenic local pollution from agricultural practices (fertilizers typically), improper disposal of domestic waste, animal and human waste. The concentrations of the other ions are low.

The Piper (1944) trilinear plot (Figure 3) reveals three main water types in boreholes and wells in the studied area: calcium magnesium bicarbonate (Ca-Mg-HCO₃), calcium magnesium chloride (Ca-Mg-Cl), and sodium chloride (Na-Cl). Ca-Mg-HCO₃ water type accounts for 50% of analyzed samples. They are found essentially in zone 3 (Figure 4) located in the north-east of the studied area. This water type is attributed to the dissolution of calcite and dolomite. The sodium chloride (Na-Cl) water type is located in the vicinity of the sea, zone 1. This water type may originate from brackish water of surface and/or seawater intrusion. The Ca-Mg-Cl-SO₄ water type is encountered in the wells located in the middle part of the area (zone 2). For most groundwater samples in zone 2, the dominant cation Ca²⁺ and the dominant anion is Cl⁻.

Chadha diagram (Chadha, 1999) is used to verify the classification of water type (Figure 5). This diagram defines four fields, representing four types of hydrogeochemical processes (Raza et al., 2016; Kaur et al., 2019). It allows explaining the different hydrogeochemical processes in the aquifers. The Chadha diagram shows that the data points are plotted in fields A, B, and C. The water type classification by the Piper diagram is affirmed by results obtained with the Chadha plot. In this plot, 54.8% of the samples are represented by Ca-Mg-HCO₃ water type (Field A), 35.7% of the samples are represented by Ca-Mg-Cl water type (Field B), and 9.5% of the samples are represented by Na-Cl water type (Field C). Ca-Mg-HCO₃ water type reflects the dissolution of carbonates and/or the alteration of silicates. The facies Ca-Mg-Cl type in our case reveals a mixing zone where the dissolution process was observed (A type) combined with the diffusion of brackish waters. The facies Na-Cl can find its origin in the evaporation process or mixing with seawater.

Using HCA, three major clusters C₁, C₂, and C₃ including four subclusters C_{1 − 1}, C_{1 − 2}, C_{3 − 1}, and C_{3 − 2} are identified (Figure 6A). The C₁ cluster consists of two subclusters C_{1 − 1} and C_{1 − 2} and represents 28.5% of the water samples. The C_{1 − 1} subcluster includes wells 15, 16, 19, 23, 31, 33, and 40, and the C_{1 − 2} subcluster includes wells 11, 12, 13, 18, and 35. The mean EC is 1,948 μS/cm for C_{1 − 1} (Ca-Cl) and 2,670 μS/cm for C_{1 − 2} subcluster (Na-Cl). This water type is characteristic of highly saline water.

Cluster C₂ includes wells 36, 34, 38, 26, 27, and 25 and accounts for 14.3% of the water samples. The mean EC for this cluster is 506 μS/cm, which is the characteristic freshwater. Groundwater in this cluster is of the Ca-HCO₃ type (Figure 6B). The dissolution of carbonates could be the process responsible for this type of water encountered in the central-eastern sector where the reservoir rock is essentially composed of argillaceous limestones flint, marl, and biogenic limestones. The third cluster (C₃) includes all the other samples divided into two subclusters C_{3 − 1} and C_{3 − 2} having a mean electrical conductivity of 1,000 μS/cm. The high chloride ion content in subcluster C_{3 − 2} (wells 1, 2, 3, 41, and 32) may be due to contamination by saline surface water in the vicinity of these wells (Figure 4). This type of water is interpreted by water–rock interactions as a result of dilution by salty surface water in contact with the carbonate formations of the aquifer. This water type is interpreted by water–rock interactions following dilution by salty surface water in contact with carbonate formations in the aquifer. The HCA gives a better precision of the distribution of facies and shows that the waters present a strong chemical heterogeneity unlike the Piper diagram, which illustrates three water types.

A plot of Ca²⁺ vs. HCO3- normalized to Na (Figure 7A) (Gaillardet et al., 1999; Belkhiri et al., 2010; Liang et al., 2018) shows that groundwater is influenced by silicate weathering and carbonate dissolution for most samples. This figure also shows the influence of evaporate dissolution for some samples [Ngazobil (15), Sidibougou (40), Keur Balla (1), and Bagana Serere (35)]. A Na-normalized plot vs. Mg (Figure 7B) shows that Mg comes from the silicates weathering and evaporates dissolution.

Figure 8 shows plots of (Ca²⁺ + Mg²⁺) vs. (HCO3- + SO42-) (Figure 8A), Mg²⁺ vs. Ca²⁺ (Figure 8B), HCO3- vs. (Ca²⁺ + Mg²⁺) (Figure 8C), and EC vs. Na+/Cl– (Figure 8D). The plot of (Ca²⁺ + Mg²⁺) vs. (HCO3- + SO42-) shows that some samples fall along the aquiline 1:1 (Figure 8A), which indicates that Ca²⁺, Mg²⁺, SO42+, and HCO3- were derived from the dissolution of calcite, dolomite, and gypsum. This corroborates the different facies obtained using the Piper and Chadha diagrams. Many other water samples are below this line (Figure 8A), suggesting a substantial excess of (Ca²⁺ + Mg²⁺) over (HCO3- + SO42-). In these samples, reverse ion exchange is the process that explains the presence of Ca-Cl facies. This figure reveals an excess of SO42- +HCO3- for some locations sampled, suggesting silicate weathering (Ndoye et al., 2018).

The plot of Mg²⁺ vs. Ca²⁺ (Figure 8B) indicates an excess of Ca²⁺ in some samples, an excess of Mg²⁺ in other samples, and a balance between Mg²⁺ and Ca²⁺ in the remaining samples. The plot of (Ca²⁺ + Mg²⁺) vs. HCO3- (Figure 8C) shows that most of the samples are above the equiline (1:1) and only three points are below this line. Using these two plots, we can say that excess of (Ca²⁺ + Mg²⁺⁾ over HCO3- is balanced by Cl- and SO42- (Su et al., 2009; Wang et al., 2013). This excess can be produced from silicate and carbonate weathering (Subramani et al., 2009). For HCO3-, apart from the carbonates zone, its enrichment (Figure 8C) can originate from carbonic acid-induced silicate weathering (Paul et al., 2019; Saha and Safiur Rahman, 2020) as a result of oxidation of organic matter or biogenic CO₂ dissolution (Liang et al., 2018). Figure 8D EC vs. Na⁺/Cl⁻ of groundwater samples shows that the trend line has a negative slope revealing a decrease in the Na⁺/Cl⁻ ratio with increasing salinity (EC). This plot indicates the removal of sodium by ion-exchange reaction (Subramani et al., 2009).

Na-Ca exchange is a significant process that can modify cation concentrations in groundwater and influence the evolution of hydrochemical compositions. When the chloroalkaline indices (CAIs) are positive, they indicate an exchange of Na⁺ and K⁺ from the water with Mg²⁺ and Ca²⁺ from the rocks (base-exchange reaction). It is reversed when CAIs are negative. The chloroalkaline index is observed for 45% of the samples, indicating the exchange of Na⁺ and K⁺ against the Ca²⁺ and Mg²⁺ ions of the aquifer rocks. The map of the distribution of chloroalkaline indices values (Figure 9) shows that negative values were observed in wells located in a central south-north band and in the north-eastern part of the study area. This observation indicates that the normal ion exchange and reverse ion exchange are geochemical processes controlling the composition of groundwater in a normal scenario. In this study, the pressure of marine waters close to the region modifies the chemical process of groundwaters.