Introduction

Groundwater is considered one of the most critical water resources in Tunisia. Due to rain shortage and high population in the last years, well excavation and groundwater use for different purposes have increased without planning. This is a great challenge for this region.

Moreover, this scarcity of water resources in this region is unfortunately combined with severe deterioration of groundwater quality due to salinity. The main objective of this study was to identify the main sources of groundwater salinity and the major hydro-geochemical processes controlling groundwater chemistry in the El Berka region. For example, the presence of evaporates in some cases causes salinization of the water table (Mascle et al. 1986). In coastal aquifers, marine intrusion may occur due to over-exploitation of the aquifers (Mamou 1990; Shtivelman et Goldman 2000; Fedrigoni et al. 2001; Panteleit et al. 2001; Pulido et al. 2003; Trabelsi et al. 2005). Other aquifers could be contaminated by salinization from the bedrock (Zouari et al. 2003; Krimissa et al. 2004). The encroachment of brines from sabkhas (Brayant et al. 1994) and shallow water could result in the mineralization of groundwater (Zammouri et al. 2007; Tarki et al. 2011; 2012). Ascending salt water from a deep aquifer across faults is also possible (Zouari et al. 2003; Kraiem et al. 2015). Under some conditions, evaporation may increase groundwater salinity (Mokadem et al. 2015; Tizro and Voudouris 2008; Moussaoui et al. 2018).

This paper integrates geochemical behavior and geophysical data (electrical tomography, used for the first time in the area) to determine the source and the spatial extent of salinization in the Pliocene–Quaternary aquifer in the El Berka region. This paper also develops a conceptual model of the salinization phenomenon.

Geology and hydrology setting

The investigated area is located in the Southern Atlas Mountains in Southwest Tunisia, approximately 40 km northwest the town of Gafsa (Fig. 1). An arid climate characterizes this region, with evaporation that exceeds precipitation on average (CRDA Gafsa, 2018). It is the site of increasing agricultural activities, which results in an increased demand for water resources and a gradual deterioration in the chemical quality of these waters (Karaouli et al. 2008; Gouasmia 2008; Hamed et al. 2014). Pliocene–Quaternary sands from the primary surficial aquifer have become progressively more saline during recent decades. The study area is located in Moulares-Redayef mining basin (Fig. 2), where phosphate is mined from the Paleocene-Eocene strata (Sassi 1974). It covers about 300 km2 (Omrane et al. 2016) and is bounded to the north by the city of Moulares, to the east by Douara region, to the south by mountain Alima, and to the west by the region of Tabeddit (Fig. 2).

Fig. 1
figure 1

Geological map and location of the study area (extract from the geological map of Tunisia to 1/500.000 publication of the national geological service)

Full size image
Fig. 2
figure 2

A shaded-relief topographic map with the morphology of the study area

Full size image

It is part of the Atlas fold and thrust belt, characterized by E-W trending anticlines and synclines and NW-trending strike-slip faults (Soumaya et al. 2020). The strike-slip faults of Gafsa, Negrine, and Metlaoui are active and prone to strong earthquakes (Soumaya et al. 2015). From the distribution of the epicenters and focal mechanism solutions, the dominant activity is centered on the NW-trending Gafsa and Metlaoui faults (Soumaya et al. 2018).

The structure of the entire Southern Atlas is linked to the interplay of two major deformation corridors (Zargouni 1984):

  • The corridor of the recess of Gafsa’s fault is in the north, which extends from the chain of Tebessa in Algeria to the coastal plain of Gabes (Fig. 1).

  • The corridor in the south of Chebika, whose existence is linked to the Tozeur-Negrine accident (Fig. 1).

During the Cenozoic era, this area was affected by strike-slip tectonic reactivation, resulting in pull-apart basins bordered by E-W and NW–SE trans-tensional faults (Fig. 3) (Soumaya et al. 2020).

Fig. 3
figure 3

Structural framework of North Africa Alpine belt and its interpolate Atlas foreland (Soumaya et al. 2020)

Full size image

The stratigraphy of the study area is relatively well-known (Swezey 1996, 2003; Riley et al. 2011; Ahmadi et al. 2013; Ben Hassen et al. 2014; Gharbi et al. 2015; Masrouhi et al. 2019; Haji et al. 2020). The stratigraphic units exposed in the study area extend from the Upper Cretaceous to the Quaternary (Fig. 5). The oldest ones belong to the Cenomanian–lower Turonian Zebbag formation. They are composed of clay units at the base, interviewed with numerous carbonate and anhydrite horizons ranging from 580 to 650 m thick. Layers of gypsum marl and clay-limestone characterize the Lower Campanian Upper Turonian Aleg formation (450–500 m thick). The Campanian–Maastrichtian Berda formation (250–300 m) consists of three parts: the Lower Berda limestone and marl-limestone, the Middle Berda marls and limestone, and the Upper Berda white limestone and blackish green clays. The Paleocene El Haria formation (40–60 m thick) consists of limestone and clay stone. The Metlaoui group, which consists of three formations, represents the Upper Paleocene to Lower Eocene: the Upper Paleocene Selja formation (60–80 m thick) with dolomite and gypsum fossiliferous clay, the Lower Ypresian Chouabine formation (10–30 m thick) with phosphate and phosphate limestone, and the Upper Ypresian Kef Eddour formation (35–40 m thick) composed of fossiliferous limestone (Ahmadi et al. 2013).

Lutetian age series is composed of white dolomite at the base and massive gypsum of the Jebs formation (gypsum 20–60 m thick). The Oligocene Sehib formation (10–60 m thick) is composed of silt and clay sandstone with locally thick conglomerates at the base. The Miocene is represented by the Beglia formation (10–190 m), which consists of sands of various particle sizes with clay beds at the top. The following Segui formation is represented by a basal conglomerate surmounted by thick series of red clay, gypsum, and intercalation of sandy clay levels (10 to 20 m thick) (Ahmadi et al. 2013). The Quaternary forms thick accumulations of gravel and coarse sands in front of the mountain ranges and gradually thins towards the center of the basin, evolving towards finer sands and clays.

The aquifer of the Moulares-Redayef basin is composed of two main water tables whose piezometric levels have been determined (Farhat 1984; Gouasmia 2008; Karaouli et al. 2008; Hamed et al. 2014):

  • The Miocene sand aquifer (Figs. 4 and 5) (Beglia formation) in the western basin is with an average thickness of about 200 m and reaches 400 m at the center of the basin (Tabeddit region).

Fig. 4
figure 4

Cross section El Berka Tabeditt Basin

Full size image
Fig. 5
figure 5

Hydro-stratigraphic column of the Garâat Douza-Tebeditt basin (Mhamdi et al. 2022) with some field photographs

Full size image

  • The clay-sandy aquifer of Pliocene–Quaternary (Segui formation) (Figs. 4 and 5) in the eastern basin, with an average thickness of about 170 m and reaching 250 m thick at the level of Garâat Douza.

There is also a phreatic and a deep aquifer, but the primary aquifer is a shallow aquifer of Pliocene–Quaternary sands (Gouasmia 2008). Sandy-clay alluvial formations limit the groundwater. These levels captured by the surface wells form the same geological entity as the deep aquifer’s reservoir level, which is separated by impermeable intrusions. However, there is an alluvial layer in the coarse alluvial deposits of the wadi (valley), fed by surface water infiltration (Gouasmia 2008). The natural recharge of the Pliocene–Quaternary groundwater table in the study area is achieved by direct infiltration of rainwater and wadi (valley) flooding. We can conclude from our study that it allows us to identify and estimate the lithology and thickness of surface aquifer formations, delineating the saturated zone and define zones according to their estimated salinity from the measured resistivity (Fig. 6).

Fig. 6
figure 6

Flowchart of the methodological approach adopted in this work

Full size image

Materials and methods

Sampling and hydrochemical analysis

A total of 26 water samples were collected from the central hydrogeological units of the survey area (water wells in Pliocene–Quaternary sediments (Fig. 7)). All samples were collected from currently exploited wells in order to ensure that water is representative of actual aquifer conditions.

Fig. 7
figure 7

Localization map of water wells and ERT profiles

Full size image

Temperature (T°), pH, and electrical conductivity (EC) were measured in situ using multi-parameter equipment (Table 1). Water analyses were carried out in the laboratory of the Company of Phosphate of Gafsa (CPG) in Tunisia for the following major elements: Ca2+, Mg2+, Na+, K+, Cl−, SO42−, HCO3−, and NO3− (Table 1). The volumetric technique is used for the measuring Ca2+, Mg2+, Cl−, and HCO3−, and the atomic absorption spectrophotometry was used for Na+, K+, SO42−, and No3−.

Table 1 In situ measurement geochemical data and saturation indexes of ground waters
Full size table

Golden Software Surfer 10 is used as a mapping tool to display the spatial distribution of chemical elements. The standard methods of Schoeller-Berkaloff (Schoeller 1965) and the Piper classification (Piper 1994) were used to determine the main chemical facies and relations of waters. The principal elements were compared using Excel 7 to delineate the main groundwater salinization processes. Thus, the multivariate statistical approach helps in the interpretation of hydrochemical XLSTAT variables using software for the analysis of the main components (APC) and the hierarchical group (HCA) (Davis1986). The measurements of statistical levels of water wells are used to draw a piezometric map of the study area (Fig. 8).

Fig. 8
figure 8

Piezometric map of the study area

Full size image

Electrical resistivity tomography (ERT)

Generally, geophysical methods are satisfactory for exploring groundwater resources (Oyedele 2009). In particular, geo-electric methods have been widely used to assess groundwater quality hosted in different lithologies and under different conditions. Compared to other geophysical methods, geo-electric methods are characterized by light equipment, easy implementation, and faster and more economical data analysis (Ekine and Osobonye 1996; Ebraheem et al. 1997; Sikandar et al. 2010). Electrical conductivity and resistivity are intrinsic properties of groundwater chemistry in that they can provide information on the degree of contamination (Delleur et al. 2006; Sherif et al. 2006). Therefore, these methods can be used to delineate the distribution of fresh and saltwater in aquifer systems while identifying the potential for contamination (e.g., Song et al. 2007; Sikandar et al. 2010 a; 2010b; Chongo et al. 2011; Nejad et al. 2011; Mhamdi et al. 2015).

The geo-electrical data, together with the hydrochemical analyses, play a crucially important role in (1) identifying the impact of over-exploitation on the PQ water table, (2) characterizing and interpreting the spatial and vertical variability of the PQ aquifer resistivity features, (3) recognizing the hydrochemical groundwater facies, and (4) identifying the main groundwater mineralization processes. In light of the work undertaken in the study area, the investigation guides how regional decision-makers can prioritize policies and interventions to overcome water constraints in agriculture while ensuring efficient and sustainable water resource management and equitable access to water (Mhamdi et al. 2022).

Electrical imaging is the multidimensional representation or distribution of the intrinsic electrical properties of the medium studied using quadrupoles (Marescot 2004; Griffiths et Barker 1993). An electric current is injected into two electrodes, A and B. This current generates an electric field that depends on the conductivity distribution of the underlying terrains. Two electrodes, M and N, record the value of the potential difference (ddp) created by the electric field. It is then possible to reduce the apparent resistivity value of the basement, knowing the geometry of the electrodes, the intensity of the injected current, and the ddp. The apparent resistivity, ρa, represents the ratio between the potential measured in the field, and that is theoretically calculated under the same conditions (identical electrode geometry and injected current intensity) for a homogeneous resistivity field 1 (Marescot et al. 2008).

As part of this study, five electrical resistivity tomography (ERT) profiles are obtained using a Winner array to search for an essentially horizontal structure, which is then used to determine the salinization extent. The Winner array used an SYSCAL Pro resistivimenter (IRIS-Instruments-France), which measures the apparent resistivity, and an acquisition unit with RES2DINV switching software. The coordinates (X and Y) of each profile’s the first and last electrode are taken using a hand GPS unit (Fig. 7).

Results

Hydrogeology

The piezometric map of the Pliocene–Quaternary groundwater table of 2019 shows the variation of the piezometric levels between 374 and 344 m deep (Fig. 8).

There are three main directions of water flow, two of which tend to supply the El Berka region, directed respectively west–east and northeast-southwest (from Tabeditt wadi (valley) and Moulares to the El Berka plain). A direction northwest-southeast tends to supply El Berka wadi (valley). These new data have shown that the water table in the Pliocene–Quaternary aquifer of El Berka has dropped by 5 m compared with the data from the piezometric map published by Gouasmia (2008).

Hydrochemical properties

The reported pH ranges from 7.61 to 8.39, and the conductivity ranges from 1.87 to 16.94 ms/cm. The salinity map (Fig. 9) shows the spatial variation of the dry water residue from 2 to 13 g/l.

Fig. 9
figure 9

Salinity map of the study area

Full size image

Salinity tends to increase in the direction of the flow of the water. In this direction, water is increasingly enriched by other elements that contribute to high salinity in the region.

The distribution maps of the major elements (Fig. 10) have shown that the chemical composition of the Pliocene–Quaternary groundwater is chemically dominated by Ca cations and Cl anions. The relative abundance is in the order of Na > Ca > Mg for the cations and SO4 > Cl > HCO3 for the anions. Sodium is always present in natural waters in varying proportions. It originates from the leaching of NaCl-rich formations (halite).

Fig. 10
figure 10

Spatial distribution maps of the main major elements [a Ca2+; b Mg2+; c Na+; d Cl−; e HCO3−; f SO42−; g NO3.−]

Full size image

Observation of the Na distribution map indicates sodium concentrations (250 to 1200 mg/l) (Fig. 10a), with high concentrations in the southeast, and calcium levels (120 to 580 mg/l) (Fig. 10b), with high concentrations in the entire region. It is mainly due to the dissolution of carbonate (CaCO3) and evaporate (gypsum or anhydrite). Magnesium concentrations vary between 55 and 180 mg/l (Fig. 10c). The higher levels are recorded in the southeastern part of our study area. Cations (Mg2+ and Ca2+) come from carbonate dissolution formations rich in magnesium (dolomites). The high magnesium concentrations potentially result from the leaching of carbonates or evaporate of the Lutetian age Seugdal formation. Sulfate contents vary between 950 and 2150 mg/l (Fig. 10d). Their concentration is very high in all sides of El Berka Wadi. The high sulfate concentration is due to the simultaneous dissolution of anhydrite and gypsum from evaporative formations. Chloride concentrations are between 200 and 2000 mg/l (Fig. 10e), with a remarkable increase in the southeastern part of the study area.

The concentration of bicarbonates varies from 0 to 500 mg/l (Fig. 10f). This seems to be due to the dissolution of carbonate minerals. Besides, nitrate levels range between 5 and 31 mg/l (Fig. 10g). In addition, nitrate reflects the application of fertilizers during agricultural activities.

The representation of water points on the triangular Piper diagram for the El Berka region shows a trend towards the sodium pole in the sub-triangle cations. In contrast, the anions have shown a dominance of sulfate. The diagram shows sulfate chloride sodium and calcium facies (Fig. 11a).

Fig. 11
figure 11

a Piper diagram of the groundwater samples, b Schoeller Berkaloff diagram of the groundwater samples

Full size image

In addition, the Schoeller-Berkaloff diagram (Fig. 11b) shows that most samples are characterized by high sulfate, chloride, sodium, and calcium contents. So, according to the Piper and Schoeller-Berkaloff diagrams, salinity is generated by the dissolution of the sulfate, chloride, sodium, and calcium minerals.

The anion/Σ anions and cation/Σcation correlations (Fig. 12) show good correlations for sodium (R2 = 0.954) and calcium (R2 = 0.817). Similarly, sulfate shows a good correlation as a function of the sum of anions (R2 = 0.916) and chloride (R2 = 0.887). Therefore, this indicates that Na, Ca, SO4, and Cl play an essential role in the acquisition of the salinity of the groundwater.

Fig. 12
figure 12

Correlations of major element contents with the sums of cations and anions

Full size image

The examination of the Na+ versus Cl− plot has shown a good correlation between these two elements (Fig. 13a). In addition, the proportional evolution of the saturation index (SI) for halite and the sum of (Na + Cl) (Fig. 14a) strongly suggests that dissolution of halite seems to be a source for the chloride-sodium composition in the aquifer.

Fig. 13
figure 13

Bivariate diagrams between major elements: a Na+/Cl−; b Ca2+/SO42−

Full size image
Fig. 14
figure 14

Mineral saturation indexes versus representative species relationship: a SI halite/(Na+  + Cl−), b SI gypsum/(Ca2+  + SO42−), c SI anhydrite/(Ca2+  + SO4.2−)

Full size image

The excellent correlation between Ca2+ and SO42− (R2 = 0.829) (Fig. 13b) points to a gypsum and/or anhydrite source. This is further confirmed by the parabolic and proportional evolution between SI and the sum of the ions (Ca2+ and SO42−), resulting from the possible dissolution of these two minerals (Fig. 14b and c).

The saturation indices for all the samples (Table 1) have values close to zero and are slightly positive for aragonite, dolomite, and calcite. Indicating that the water is saturated concerning to these minerals and is negative for the evaporate minerals (anhydrite, gypsum, halite), which shows that the investigated waters are under-saturated relative to these minerals. This suggests that the dissolution of the latter is quite likely. Thus, we can conclude that the mineralization of the El Berka Pliocene–Quaternary groundwater can be due to the dissolution of evaporates.

Statistical analysis

The statistical analysis has enabled us to determine the elements responsible for groundwater mineralization. In this study, both the principal component analysis (ACP) and hierarchical cluster analysis (HCA) have been performed to constrain better the interpretation of chemical data (pH, salinity, Ca2+, Mg2+, Na+, K+, HCO3−, Cl−, SO42−, and NO3−). Analysis of the correlation matrix (Table 2) shows that sodium, sulfates, chloride, calcium, and magnesium are highly correlated with TDS. The correlation coefficients are respectively 0.965, 0.954, 0.949, 0.917, and 0.832, which indicate the contribution of these elements in the mineralization of the groundwater of El Berka region.

Table 2 Correlation matrix of the groundwater samples from El Berka region
Full size table

The main components analysis is conducted on ten physico-chemical parameters from the analysis of water sampled over the entire study area (Table 2, Fig. 15a,b,c). The eigenvalues and the percentage of variance explained by each vector are listed in Table 3. We highlight the associations between the different variables and the possible existence of subpopulations. A correlation matrix of 10 parameters shows fascinating solid and positive correlations of the waters of the Pliocene–Quaternary (Table 3) showing that the first axis (F1) is around 60.05% of the total variance and the second 13.78% (F2).

Fig. 15
figure 15

Statistical analysis [a plot of principal component analysis of variables, b projection of groundwater samples onto the factorial plane (F1/F2), c hierarchical cluster analysis (HCA): dendrogram of the groundwater samples from El Berka region]

Full size image
Table 3 The principal component analysis results
Full size table

The sum of these two principal Eigenvalues equals 73.83% (Table 2). It is correlated with high loads of TDS, Ca2+, Mg2+, Na+, K+, Cl−, and SO42− (Fig. 15a). Therefore, it shows the natural processes of mineralization of groundwater. The spatial graphical representation of the statistical units of factorial plane F1-F2 (Fig. 15b) shows a competition, according to the F1 axis, between chlorinated and sodium sulfate waters on the positive side of the axis and calcium bicarbonate waters on the negative side. Water from calcareous and marly carbonate aquifers is low salinity, while water from aquifers containing evaporative saliferous or gypsum rocks is more saline. This salinity is remarkable in the samples going from upstream to downstream of the basin, confirming the hypothesis of contamination of the waters by the phenomenon of dissolution.

The hierarchical upward classification shows groundwater samples are subdivided into three groups (Fig. 15c). Group (C1) consists of 7 wells. This water group includes, on the one hand, the least mineralized waters. It is characterized by low TDS and low concentrations of Ca2+, Mg2+, Na+, HCO3−, Cl−, and SO42− compared to groups 2 (contains 12 wells) and 3 (includes 7 wells) (Table 4) of medium to high concentration. They are mainly characterized by strongly mineralized waters or influenced by the water–rock interaction and the process of hydrolysis in acquiring the mineralization of the waters during their infiltration.

Table 4 Mean values of the chemical constituents of groundwater clusters
Full size table

Tomography results (ERT)

The choice of the geo-electrical method is motivated by the satisfactory results obtained for the characterization of surficial aquifers in Tunisia (Abdelkader et al. 2018; Mhamdi et al. 2006; 2015; 2013; Gouasmia et al. 2018, Moussaoui et al. 2018). The ERT profiles were realized in a parcel between Wadi El Berka and Wadi Tabeditt (Fig. 7) to determine the salinization extent. The obtained results allow classifying the ERT profiles in two groups (Fig. 16).

Fig. 16
figure 16

Profiles of electrical resistivity tomography

Full size image

A high resistivity occurred in profiles 1 and 5 (more than 16Ω.m), which can be attributed to low groundwater salinity (less than 6 g/l). Conversely, low resistivity observed in profiles 2, 3, and 4 (less than 10Ω.m) seems to be related to high groundwater salinity (more than 7.5 g/l).

The ERT results confirm and complete the results obtained by geochemical interpretation and in particular the salinity map (Fig. 7). The tomographic profiles show the same trend as the salinity map, indicating that the Pliocene–Quaternary groundwater of the El Berka region contains a low electrical resistivity zone in the center with a relatively high resistivity approximately at the mountainous and wadi Tabeditt areas.

Discussions

The electrical resistivity tomography data allows the identification of the succession of the following water tables:

  • A highly resistant superficial layer can reach a thickness of 7 m, corresponding to conglomerates, pebbles, and alluvium.

  • An intermediate level of low resistivity with sands and clays that are saturated with salt water.

  • A lower relatively resistant level that may contain water of better quality than the intermediate level.

The characterization of the aquifer units was achieved by integrating geophysical data (ERT) and well data. The combination of the electrical resistivity tomography and geochemistry data shows that the salt concentration in the waters of the Pliocene–Quaternary aquifer of the El Berka region is the result of the following three successive processes:

  • Process A (ancient state of the El Berka basin): Significant amounts of rainfall on the Moulares and Metlaoui mountain ranges, specifically Jebel Billil in the north and Jebel Alima in the south (Fig. 2). These rainwaters are enriched in salts by leaching of evaporate rocks from the evaporate Seugdal formation. As this water then flowed towards El Berka, some of the water seeped into the groundwater, and the rest of the water evaporated and accumulated salts at the surface (Fig. 17a).

  • Process B (intermediate state of the El Berka basin): The arid climate of Southern Tunisia is the driving force behind evaporation. It is from these sources of water that the salts in solution get concentrated and then precipitate forming evaporated rocks such as gypsum, anhydrite, and halite from secondary sedimentary brines (Fig. 17b).

  • Process C (current state of the El Berka basin): During the diagenesis of the saline rocks, part of them will leach again into the groundwater, which will become increasingly concentrated in salts (Fig. 17c). The current state of the models shows an evolution in time and space of the aquifer levels. It is believed that all available results were combined to explain the salinization process in the study area.

Fig. 17
figure 17

Successive processes of the salt concentration in the waters of the Pliocene–Quaternary aquifer of the El Berka region [a (Processes A), the ancient state of the El Berka basin; b (Processes B), the intermediate state of the El Berka basin; c (Processes C), the current state of the El Berka basin]

Full size image

The existence of secondary evaporates in the superficial sediment is the result of dissolution rather than the evaporation of the primary evaporates (Bryant et al. 1994). So, the evaporative history is the result of a succession of very slow geological and geochemical events and the result of agricultural fertilizers (irrigation). Dissolution of evaporate rocks is known to be an agent of groundwater salinization in Southern Tunisia (Mascle et al. 1986; Bryant et al. 1994; Tarki et al. 2012; Kraiem et al. 2014; Mhamdi et al. 2015).

This happens especially when the ground has large volumes of evaporate rocks and/or when recharge water drains evaporated rocks and dissolves considerable quantities of salt following the chemistry of the water inflow in the Chott el Djerid (Bryant et al. 1994). Recharge also occurs via the leakage of upward flow from the deep aquifer (Houcine et al. 2014). The salts can be mobilized by lateral infiltration. Aquifer salinization may also occur due to saline water input from the underlying aquifer (Zouari et al. 2003; Krimissa et al. 2004), from contamination by hyper-saline aquifers and ascent of high salinity deep waters along faults (Zouari et al. 2006; Kraiem et al. 2015), and from Sabkha (chotts, salt lakes) at the surface (Mhamdi et al. 2006).

In Southern Tunisia which is characterized by intense evaporation, rock-water interaction and local irrigation practices such as submersion irrigation and drainage basin are the main processes that affect groundwater composition (Moussaoui et al. 2018; Agoubi et al. 2013; Kraiem et al. 2012). Our research shows that it is also the case in our study area. It is worth mentioning that the same phenomenon has been observed in other regions of Southern Tunisia such as Gareet Douza (Moussaoui et al. 2018), the Jeffara aquifer in the southeast (Agoubi et al. 2013), as well as the in the oasis shallow aquifers of Nefzaoua region in Southwest Tunisia (Kraiem et al. 2012). The origin of water mineralization in central Tunisia is most probably related to the dissolution of evaporate minerals and the exchange of cations, as in the case of the Chougafiya basin (Farid et al. 2012). On the other hand, water–rock interaction processes and the backflow of irrigation water have led to high concentrations of nitrates in groundwater, especially in agricultural areas, due to the excessive use of fertilizers in the open aquifers of BouHafna-Haffouz (El Mejri et al. 2018). The groundwater of the Sidi El Hani basin (central-eastern Tunisia) is affected by the dissolution of halite, gypsum, and anhydrite with simultaneous precipitation of calcite and dolomite in association with ion exchange and inverse ion exchange with clay minerals (Lotfi et al. 2018).

Northern Tunisia is affected by seawater intrusion, just like in the region of Maâmoura (Chabaane et al. 2018) and the coastal aquifer of Nabeul-Hammamet (Trabelsi et al. 2016). The causes of salinization differ from one country to another. In the Wadi Nador coastal aquifer (Tipaza, Algeria), intensive good pumping, hydraulic connection to the sea, and high porosity of the Pliocene–Quaternary aquifer are the leading causes of groundwater salinization (Bouderbala and Remini 2014). The salinization in the Hassi Miloud area, Algeria, is related to the combined effects of irrigation and upwelling from the saturated zone (Semar et al. 2019). Dissolution of halite, carbonate weathering, and mixing of seawater are the main factors controlling groundwater in the Aseer region of Saudi Arabia (Mallick et al. 2018). Chemical alteration of rock-forming minerals controlled shallow groundwater in the Dongsheng Coal Basin in the Ordos Basin, China (Li et al. 2013). In coastal aquifers in Portugal, human-induced water quality degradation intensifies salinity directly through pollution or indirectly through agricultural activity over-pumping that leads then to the invasion of saline groundwater in freshwater aquifers (Carreira et al. 2014). In Egypt, for instance, the salinity of groundwater results from the precipitation of calcite, dissolution of dolomite and gypsum, as well as the exchange of sodium in groundwater with calcium and magnesium in the aquifer matrix along the groundwater flow path in the central part of the country (Mosaad and Kehew 2019), while northwest of the Nile Delta groundwater quality in shallow aquifers is controlled mainly by the dissolution of evaporate minerals and seawater intrusion, contamination of nitrogen compounds by anthropogenic activities, agricultural fertilizers, and domestic drainage water (Masoud 2014). In the coastal aquifers of Morocco, water salinity can be induced by marine intrusion due to the absence of red clays, favoring the penetration of seawater and the decrease in the recharge rate of aquifers caused by the decrease in rainfall under the effect of climate change. Moreover, groundwater chemistry is also influenced by water–rock interaction and anthropogenic activities (application of fertilizers, septic tank use) in the upstream area (Najib et al. 2017; Fadili et al. 2017; Ouhamdouch et al. 2019).

Conclusion

This study has aimed at integrating hydrochemical and geophysical (ERT) data for the study of the salinization of the Pliocene–Quaternary aquifer in the zone of El Berka (Southwest Tunisia). The hydro-geochemical analysis made it possible to determine that the high salinization of groundwater in the Pliocene–Quaternary sands resulted from the mineralization of waters from the mountain ranges and the dissolution of minerals (evaporates) existing on the plain. The ERT profiles allow to determine the extension of saltwater levels. Generally, evaporation enrichment is achieved by trapping groundwater in low-permeability materials. The high evaporation rate is combined with irrigation and induces salt precipitation and accumulation in the unsaturated zone. It should also be noted that water infiltration dissolves the rocks by rainfall.