Arab J Geosci (2013) 6:3545–3561 DOI 10.1007/s12517-012-0617-3

ORIGINAL PAPER

Geochemistry and quality assessment of groundwater using graphical and multivariate statistical methods. A case study: Grombalia phreatic aquifer (Northeastern Tunisia) Besma Tlili-Zrelli & Fadoua hamzaoui-Azaza & Moncef Gueddari & Rachida Bouhlila

Received: 17 November 2011 / Accepted: 25 June 2012 / Published online: 27 July 2012 # Saudi Society for Geosciences 2012

Abstract The Grombalia coastal aquifer, situated in Northeastern Tunisia, is a water source for public, agricultural, and industrial supplies in the region. The overexploitation of this aquifer, since 1959, and the agriculture activities led to the degradation, by places, of the water quality. The present study implemented graphical, modeling, and multivariate statistical tools to investigate natural and anthropogenic processes controlling Grombalia groundwater mineralization and water quality for promoting sustainable development. To attempt this goal, groundwater was collected from 33 observation wells in January 2004, and samples were analyzed for 10 physicochemical parameters (temperature, pH, salinity, Na+, Ca2+, K+, Mg2+, Cl−, HCO3−, and SO42−). Hydrochemical facies using Piper diagram indicates a predominance of a mixed facies, of the Na-Cl-HCO3 type, or Na-Ca-Cl-SO4 type, and, with less expansion, Na Cl type. The main factors controlling Grombalia groundwater mineralization seem to be mineral dissolution of highly soluble salts especially, the halite dissolution existing in the surface salty deposits and, with less importance, the ion exchange and reverse ion exchange process with clay minerals existing in the aquifer. The comparison of the major ions of the Grombalia groundwater, with the World Health Organization norms of potability (WHO 2004), reveals that these waters cannot be used for human consumption without any treatment. Most waters of the Grombalia aquifer, with a B. Tlili-Zrelli (*) : F. hamzaoui-Azaza : M. Gueddari Laboratory of Geochemistry and Environmental Geology, Department of Geology, Faculty of Mathematical, Physical and Natural Sciences, University Campus, Tunis, Tunisia e-mail: [email protected] R. Bouhlila Modeling in Hydraulic and Environment Laboratory, National Engineers School of Tunis, Tunis, Tunisia

relatively high salinity, are not suitable for irrigation, in ordinary conditions. Nevertheless, they can be used for permeable soils, with an adequate drainage and applying an excess of leaching water. Keywords Grombalia groundwater . Tunisia . Hydrogeochemistry . Statistical analysis . Water quality

Introduction Groundwater resources in arid and semi-arid regions play a vital role in the socioeconomic development (Carreira et al. 2010). During the last few years, an understanding of the chemical composition of natural waters has become increasingly important due to the increased anthropogenic influences (Magnus Land 1997). In Tunisia, the groundwater resources, estimated at 1,840 million m3 year−1, depend on the geologic configuration of the country, which is widely dominated by sedimentary formations, dry climate, and endorheic stream system. In the region of the Cap Bon, situated in the Northeast of Tunisia, there are five main aquifers: the Tazoghrane aquifer, the Haouaria aquifer, the oriental coast aquifer of the Cap Bon, the coastal plain’s aquifer of Nabeul-Hammamet, and the aquifer of Grombalia which is the object of this work. The overexploitation of the Grombalia aquifer, since 1959, has led to the degradation, by places, of the water quality and in a decline more or less generalized on the piezometric level. The proximity of Grombalia aquifer to the sea (coastal aquifer), the presence of brines and saline soils in the Grombalia plain, the agriculture activities and the nature and the permeability of the aquifer, are also other factors which can be at the origin of salinity of this aquifer.

3546

The chemical composition of groundwater is controlled by many factors that include composition of precipitation, geological structure and mineralogy of the watersheds and aquifers, and geochemical processes within the aquifer. The interaction of all factors leads to various water facies (Murray 1996; Rosen and Jones 1998). In the current work, we used different maps and graphical representations to classify and interpret the geochemical data in terms of distribution map of chemical parameters, Piper (1944) diagram, plots of chemical parameters, and geochemical modeling. In addition, multivariate statistical analyses such as principal component analysis (PCA) and hierarchical cluster analysis (HCA) were used. Indeed, the combined use of PCA and HCA enabled the classification of water samples into distinct groups based on their hydrochemical characteristics (Hamzaoui-Azaza et al. 2011). PCA can help classify groundwater and identify major mechanisms influencing groundwater chemistry (Abdul Halim et al. 2010). HCA was used to determine if the samples can be grouped into statistically distinct hydrochemical groups that may be significant in the geologic context (Güler and Thyne 2004). Three indices (sodium adsorption ratio (SAR), permeability index (PI), and sodium percentage (% Na)) were then used to determine the suitability of groundwater from this area for irrigation activities. The main issues that will be addressed by this study include (1) the relative importance of natural and anthropogenic factors in controlling the water chemistry (2) to demonstrate the usefulness of the statistical analysis to improve the understanding of the groundwater composition and (3) the assessment of the suitability within Grombalia’s groundwater for the irrigation and drinking purposes.

Study area Site description The Grombalia plain is located in the northeast of Tunisia between latitudes 36° 75′–36° 46′ north and longitudes 10° 72′–10° 30′ east, with a total area of 700 km2. It is situated about 40 km south of Tunis city, and it is limited by the Gulf of Tunis in the north, the Jebel Abderrahmen anticline and Oriental coastal plain in the east, Jebel Reba El Ain in the south, and Jebel Halloufa in the west (Fig. 1). The plain’s climate is semi-arid with mean annual precipitation of 448.2 mm with almost no precipitation between June and September, the average annual temperature is 18 °C with mean of 26.5 in summer (August) and 10.6 in winters (January) and potential evapotranspiration is 1,300 mm/year.

Arab J Geosci (2013) 6:3545–3561

Geology and hydrogeology Grombalia plain covers different stratigraphic units viz Jurassic, Cretaceous, Tertiary to Quaternary (Castany 1948; Sebei 2001). The Jurassic is observed in Djebel Mecella where it is formulated by limestone and in Djebel Ressas and constituted by dolomite, marl, and limestone. Cretaceous is mainly developed in northwestern part of Jebel Mecella–Jebel Ressas anticline and with the only Jurassic stratigraphic continuity in Djebel Ressas. In the studied area, the Tertiary formation is subdivided into four: (1) Eocene is constituted by limestone; (2) Oligocene is about 150 m in the north, this formation is composed of Sandstone, but in the south, is composed of clay; (3) Miocene, is about 20 m, characterizes a southwestern part of the plain’s border, and it is composed of gritty marl and gritty limestone; (4) Pliocene is about 20 m, and it is composed of sandstone. The Quaternary formation shows a heterogeneous continental detrital sedimentation which consists of sands, sandy clays, lagunal sand, marls, and sandstone. On the border of this ditch, there is a rectilinear fault, oriented approximately from N120 to N140, allocating as well the Oligocene and the lower Miocene: Sandstone of the Fortuna formation than those of the middle Miocene of the Saouaf formation (Ben Salem 1992). According to Ennabli (1980), three major aquifers are present within the Grombalia plain (Fig. 2). The oldest of which is confined with an average thickness of 200 m and, the overlying aquifer is also confined with average thickness of 100 m and the upper unconfined phreatic aquifer with an average thickness of 40 m. These aquifers are built of Quaternary sand and are interconnected. The study has focused on the unconfined phreatic aquifer having horizontal transmissivity values within the range of 25×10−4–2×10−2 m2/S and a storage coefficient of 5.5×10−3. Groundwater tables are generally shallow, usually within 1–5 m below the surface, with seasonal changes of about ±2 m, correlated to infiltration processes due to precipitation and irrigation. This aquifer is directly fed by infiltration of surface waters through the permeable formations or indirectly by the underlying aquifers. The main groundwater flow paths of the Grombalia aquifer start generally from south to north (Fig. 3). Discharge occurs mainly via evapotranspiration and artificial abstraction. In addition, two major rivers, Soltane and El Bey, with high slopes, discharge into Sebkhet el Melah (Ennabli 1970). The total output was estimated at 93 mm3 in 2000, that registered in 1990 was 89.7 mm3, for the renewable resources considered at 51 mm3/year, that led to an overexploitation of this aquifer with an overexploitation rate of 176 % (DGRE (General Direction of Water Resources) 1980–2000).

Arab J Geosci (2013) 6:3545–3561

3547

Fig. 1 Location map and geology of the Grombalia plain (Sebei 2001, modified)

Tunisia

Grombalia plain

Fig. 2 Schematic representation of the aquifer system of the Grombalia plain (Ennabli 1980)

3548

Arab J Geosci (2013) 6:3545–3561

Fig. 3 Piezometric map of shallow Grombalia groundwater (January 2004)

Since 1975, two types of groundwater artificial recharge have been commonly used: surface spreading and direct aquifer injection (Ennabli 1975). Since 1993, the adopted process has been the injection of Medjerda canal waters, with salinity lower than 1.5 g/l. The quantitative impact of the artificial recharge is an increase of the piezometric level which varies from 0.5 to 3.5 m (DGRE 2002).

Land use and soil type Land use in the study area is primarily agricultural in Bou Argoub, Menzel Bou Zelfa, and Soliman, with main crops produced consisting of cereals, fruit arboriculture (citrus, vine). On the other hand, industrial activities are developed in Soliman and Grombalia. According to the legend of the Food and Agriculture Organization (1990), most of the soils of the Grombalia plain are fluvisoils and vertisoils; the others are halomorphic, hydromorphic, calcimorphic, and rendzina.

Materials and methods Sampling and water chemical analysis Thirty-three groundwater samples were collected in January 2004. Groundwater samples were shallow, with piezometric level from 1 to 13 m, and were collected with the pumps installed in majority of wells. Water levels in all wells were measured manually by the use of electric probe. Salinity, pH, and water temperature were measured in situ. Samples were kept at 4°C for their subsequent chemical analyses. Afterwards, a split of the sample was filtered through a 0.45 μm Millipore filter. Chloride was determined by standard AgNO3 (Rodier 1984). Sulfate (SO42−) content was measured by the gravimeter method using BaCl2. Sodium (Na+) and potassium (K+) were measured by flame photometry and calcium (Ca2+) and magnesium (Mg2+) with atomic absorption. Alkalinity was determined by titration with HCl (Rodier 1984). The quality of chemical analysis was checked by doing an ionic mass balance, accepting an error lower than 10 %.

Arab J Geosci (2013) 6:3545–3561 Table 1 Statistical summary of hydrochemical parameters of the study area

3549

HCO3− (mg/l) Cl− (mg/l) SO42− (mg/l) Na+ (mg/l) Ca++ (mg/l) Mg++ (mg/l) K+ (mg/l) pH cond (ms/cm) sal (mg/l) T (°C)

Min

Max

Mean

SD

WHO (2004)

488.16 393.97 431.32 261.64 162.63 48.39 0.16 7.12 2.24 1,000 13.7

1,586.52 2,519.86 1,386.65 1,848.21 643.31 430.69 48.48 7.80 10.36 5,700 21.4

1,073.77 1,155.75 812.84 657.85 340.11 209.19 15.81 7.39 4.80 2,480 17.7

235.52 492.35 228.20 304.87 99.39 93.63 12.47 0.16 1.70 1,030 1.97

380 250 400 200 200 150 30 6.5–8.5 1.5 1,500

Statistical study The multivariate statistical methods is a quantitative and independent approach for classification of groundwater samples according to their geochemical characteristics and may allow Fig. 4 Salinity distribution map of Grombalia groundwater (January, 2004)

to simplify and organize large datasets providing meaningful insight that may be helpful for water quality assessment (Cloutier et al. 2008; De Andrade and Stigter 2009). In this study, two multivariate statistical techniques were used: the HCA and the PCA. A statistical computer code

3550

Arab J Geosci (2013) 6:3545–3561

Arab J Geosci (2013) 6:3545–3561

ƒFig. 5

Spatial distribution maps of a sodium, b chloride, c calcium, and d sulfate (January 2004)

“Andad” (CVRM 2000) was used for PCA and HCA analysis.

Hierarchical cluster analysis The HCA is a data classification technique. There are different clustering techniques; however, the hierarchical clustering is the one most widely applied in earth sciences (Davis 1986). Both Q-mode and R-mode were performed on the hydrochemical parameters (Bruce Banoeng 2009). The Q-mode HCA was used to classify the samples into distinct hydrochemical groups while the R-mode HCA is linking variables (De Andrade and Stigter 2009; Belkhiri et al. 2011). To perform CA, an agglomerative hierarchical clustering was developed using a combination of the ward’s linkage method and squared Euclidean distances as a measure of similarity. All the variables were log transformed and standardized. The standardized data are obtained by application of the equation: zi 0xi − mean/s (Davis 1986) where zi indicates the standardized value, xi represents the data for variation, s the standard deviation of the same variable. Data standardization is required in HCA as the calculation of the Euclidean distances will be influenced most severely by the parameters with the largest variances in their distribution if the raw data are used. Furthermore, scale standardization eliminates the influence of different units of measurements and renders the data dimensionless (Dillon and Goldstein 1984). For the Q-mode HCA, the data for each parameter were standardized to a range of −1 to 1 (Yidana et al. 2010). The result of the HCA was given as a dendogram.

3551

Principal component analysis PCA is used for data reduction and for deciphering patterns within large sets of data (Davis 2002; Farnham et al. 2003). According to Davis (1986), this method is considered as the R-mode technique. PCs provide information on the most meaningful parameters, which describes a whole dataset affording data reduction with minimum loss of original information (Helena 2000). PCs were extracted on the symmetrical correlation matrix which consists on interrelations between variables (Cloutier et al. 2008); these PCs were subjected to varimax rotation (raw) generating. Geochemical modeling In order to investigate thermodynamic controls on mineral– water interactions, the geochemical computer code PHREEQC (Parkhurst and Appelo 1999) was used to calculate mineral saturation indices. Saturation indices provide some indication up on the equilibrium state between groundwater and the surrounding mineral rocks assemblages (Rouabhia et al. 2010). Changes in saturation state are useful to distinguish different stages of hydrochemical evolution and help identify which reactions are important in controlling water chemistry (Güler and Thyne 2004).

Results and discussion Groundwater chemistry A statistical summary of hydrochemical parameters with (mean, max, and mean values) is given in Table 1. In the study area, pH values ranged between 7.12 and 7.8 with mean of 7.38 indicating that groundwater was neutral to slightly

Fig. 6 Plots of majors ions; a Cl versus Na, b HCO3 versus Ca, c Ca versus SO4

3552

Fig. 7 Plots of major ions with TDS

Arab J Geosci (2013) 6:3545–3561

Arab J Geosci (2013) 6:3545–3561

3553

Fig. 8 Scatter diagram of (Ca+Mg) versus (SO4 + HCO3−)

alkaline in nature, well within the permissible limit of 6.5–8.5 (WHO 2004). The electrical conductivity (EC) varies from 2.24 to 10.36 ms/cm with mean of 4.8 ms/cm. Total dissolved solids (TDS) also showed a wide variation from 1,230 to 5,698 mg/l with mean of 3,464 mg/l. Based on this value, groundwater is classified into Brackish water “TDS ranged between 1,000 and 10,000 mg/l” (Rabinove et al. 1958). Chemical analysis of water samples indicated that the order of abundance of the major cations is Na>Ca>Mg>K. The concentration of major cations Na+, Ca2+, Mg2+, and K+ ranged from 261.64 to 1,828.41, 162.63 to 643.31, 48.39 to 430.69, and 0.16 to 48.48 mg/l with mean of 657.85, 340.11, 209.19, and 15.81 mg/l, respectively. All samples of the study area exceeded the permissible limit of Na for drinking water (200 mg/l; WHO 2004). In 70 % of the samples, the permissible limit for Ca and Mg (300 and 150 mg/l respectively; WHO 2004) has been exceeded. Fig. 9 Scatter diagram of (Ca+Mg)−(SO4+HCO3-) versus (Na−Cl)

The concentration of major anion Cl−, SO4−, and HCO3− varies from 393.97 to 2,519.86 with mean of 1,155.75 mg/l, 431.32 to 1,386.65 with mean of 812.84 mg/l and 488.16 to 1,586.52 with mean of 1,073.76 mg/l, respectively. The order of abundance of these anion is Cl>HCO3−>SO4. All samples exceeded the permissible limit of Cl and SO4 for drinking water (250 and 400 mg/l, respectively; WHO 2004). Factors controlling salinity distribution Like electrical conductivity, salinity showed wide variation from 1,000 to 5,700 mg/l m with mean of 2,480 mg/l. All samples have salinity beyond the maximum permissible limit for drinking water as per the World Health Organization standard, i.e., 1,500 mg/l. The spatial distribution of salinity (Fig. 4) showed that a higher salinity characterizes

3554 Table 2 Saturated indices of Grombalia groundwater

Arab J Geosci (2013) 6:3545–3561

Well

SI (aragonite)

SI (calcite)

SI (dolomite)

SI (gypsum)

SI (halite)

1 2 3 4 5 6 7 8 9 10 11 12 13 14

0.6791 0.7424 0.6789 0.9133 0.7579 0.9978 0.9421 0.7402 0.8204 0.6719 0.4662 0.8266 0.6519 0.8676

0.8266 0.8899 0.8263 1.0607 0.9053 1.1452 1.0895 0.8876 0.9678 0.8193 0.6136 0.9741 0.7993 1.015

1.6619 2.0182 1.846 1.9249 1.9044 2.3489 2.5218 1.532 1.8078 1.6386 1.2909 2.1841 1.9449 2.0194

−0.5747 −0.6638 −0.501 −0.6467 −0.5203 −1.0874 −0.8587 −0.3641 −0.5362 −0.4968 −0.7712 −0.4951 −0.9885 −0.6965

−5.0151 −4.4678 −5.1678 −4.9701 −4.4432 −5.2724 −4.9284 −5.0495 −4.9699 −4.8964 −5.1803 −4.751 −4.8753 −4.9347

15 16 17 18 19 20

0.6746 1.1937 0.9809 1.1875 0.6582 0.9146

0.822 1.3412 1.1283 1.3349 0.8057 1.062

1.5802 2.6305 2.1973 2.9606 1.7312 1.8528

−0.5988 −0.7271 −0.5902 −0.7627 −0.7858 −0.7619

−5.1609 −5.2159 −5.0948 −4.5749 −5.0754

21 22 23 24 25 26 27 28 29 30 31 32

1.0021 1.1495 0.8039 0.844 0.7085 0.6613 1.0525 0.9083 0.8409 0.9743 0.8438 0.8571

1.1495 1.2969 0.9514 0.9915 0.8559 0.8088 1.1999 1.0557 0.9883 1.1218 0.9912 1.0046

2.6717 2.0575 2.0998 2.3233 1.699 1.8778 2.2176 1.866 1.7925 2.5461 1.8013 2.1193

−0.8094 −0.6794 −0.6504 −0.6134 −0.7758 −0.6134 −0.578 −0.5403 −0.4107 −0.607 −0.5986 −0.6792

−5.0125 −4.5975 −5.6412 −4.8672 −4.6572 −5.421 −4.4308 −4.8987 −4.9402 −4.5959 −4.0679 −4.4684 −5.0352

33

0.7191

0.8665

1.9061

−0.6747

−4.5368

two distinct zones: one situated in the north of the study area, in the east of Sebkhet El Melah in Soliman and another in the south of the aquifer, between Grombalia city and Nianou village. The elevated salinity values in the first zone would be attributed of leaching of salty waters from brine and Sebkhet El Melah, whereas in the second zone that would be related with the leaching of salts from saline soils. Return flow from irrigation water and use of fertilizers may contribute in higher salinity especially in Bou Argoub Menzel Bou Zelfa, where there are developed agriculture activities. The salinity of the groundwater is low near the recharge zone in Menzel Bou Zelfa–Beni Khalled (dilution effect). Leaching of salts from saline soils and brine, land use activities, proximity of recharge zone, and return flow from irrigation water do not completely explain distribution of

salinity; this spatial distribution would be controlled in addition by nature and by permeability of the aquifer. Indeed, the Grombalia aquifer is sandy and permeable in the east, clayed and less permeable towards west, which affects a replenishing water rate and, consequently, a water salt load. Process controlling major solute distribution in groundwater The results from the chemical analysis were used to identify factors controlling geochemical processes and spatial distribution of major ions in Grombalia aquifer. The spatial distribution of Na, Cl, and salinity show a wide similarity (Fig. 5 a, b), this explains that Na which is the dominant cation and Cl which is the dominant anion control a water salt load. Na was related to Cl with a high correlation

Arab J Geosci (2013) 6:3545–3561

3555

Fig. 12 Spatial distribution of the variables in the axes system F1–F2

Fig. 10 Geochemical facies of the groundwater of Grombalia

coefficient of 0.71 (Fig. 6a) indicating that these ions have a common origin: dissolution of halite present in saline soils or in saline surface deposits. Sodium and chloride were also strongly correlated with TDS with r00.82 and 0.9 (Fig. 7a, b) respectively, which confirms that groundwater salinities were mainly controlled by these ions. Spatial distribution of sulfate (Fig. 5d) Fig. 11 Dendogram for the groundwater samples

demonstrates that higher sulfate values characterize a limited zone between Soliman and Beni Khalled, whereas lower sulfate values are recorded in wells situated in proximity of recharge zone. A plot of Ca versus SO4 (Fig. 6c) indicates that sulfate would have a common origin with Ca via sulfato-calcic minerals, but it has different geochemical behavior with calcium especially in salt water. Besides, cross-plot of SO4 with TDS (Fig. 7d) showed a good correlation, indicating that sulfate participate, with Na and Cl, in water mineralization.

3556

Arab J Geosci (2013) 6:3545–3561

Table 3 Summarized the PCA results including the loadings and the eigenvalues of each PC

HCO3− Cl− SO4− Na+ Ca++ Mg++ KpH Cond Sal Eigenvalues % Variance explained % Cumulative variance

F1

F2

F3

0.48 0.94 0.56 0.91 0.41 0.86 0.16 −0.33 0.98 0.98 5.24 52.37 52.37

−0.66 0.036 0.18 −0.17 0.71 −0.15 −0.34 −0.6 0.01 0.02 1.51 15.14 67.51

0.01 0.053 0.016 0.023 0.01 0.16 −0.88 0.47 0.04 0.02 1.03 10.34 77.86

The spatial distribution of calcium (Fig. 5c) shows that higher calcium values characterize a southeastern part of the aquifer and a zone near of Sebkhet El Melah, and are respectively ranged between 330.6 and 414.8, and 451.5 and 643.3 mg/l. Higher calcium values would be related in the first zone to the alteration with leakage water of Belli-Nianou limestone crust; whereas in the second zone, they would be the result of alteration of both limestone crust and clayey gypseous formations. A cross-plot of HCO3− with TDS (Fig. 7e) demonstrated a good correlation in less loaded groundwater; but there were no linear relationship between them in more loaded groundwater samples indicating that bicarbonate partakes with calcium in transfer reactions with dissolved and solid phases. The plot of HCO3− with Ca++ (Fig. 6b) confirms the

ionic association between these ions, particularly in mineralized groundwater. The (Ca++ +Mg++) vs (HCO3− +SO4−) scatter diagram (Fig. 8) will be close to the 1:1 equiline if the dissolutions of calcite, dolomite, anhydrite, and gypsum are the dominating reactions in the system (Kumar2004; Vungopal 2009). Ion exchange tends to shift the points to right due to an excess happens to be of HCO3− and SO4− (Cerling et al. 1989; Fisher and Mulican 1997…). If reverse ion exchange is the process, it will shift the points to the left due to a large excess of (Ca++ +Mg++) over (HCO3− +SO4−; Belkhiri et al. 2011). Further, according to Fisher and Mulican (1997), a cross plot of Na+ −Cl versus (Ca+++Mg++)− (HCO3− +SO4−) should be linear with the slope of −1. In the study area, the plot of (Ca++ +Mg++) vs (HCO3− +SO4−; Fig. 8) shows that the groundwater samples are distributed on both sides indicating that Grombalia groundwater mineralization is controlled, in addition to mineral dissolution, by ion exchange and reverse ion exchange process with clay minerals present in aquifer. Besides, a plot of Na+ − Cl versus (Ca++ +Mg++)−(HCO3− +SO4−; Fig. 9) shows a slope of −1.1 reflecting that reverse ion exchange has taken place in few groundwater samples. Saturation index In order to deepen the study related to characteristics of Grombalia’s groundwater, the saturation indices (SI) of minerals calcite, dolomite, halite, gypsum, and aragonite were calculated using the PHREEQC code. The results are presented in Table 2. An SI less than zero indicate that the groundwater is undersaturated with respect to that particular mineral. An SI greater than zero specifies the groundwater being supersaturated with respect to that particular mineral phase and incapable of dissolving more of the mineral (Benony 2006). The precipitation of CaCO3 and CaMg (CO 3) can be assessed in terms of the SI (Al Agha 2005). The SI of the carbonate minerals is positive indicating that groundwater are supersaturated with carbonate minerals (dolomite, calcite and aragonite), this may justify the presence of calcareous crust. In the cases of gypsum and halite, SI values indicate undersaturation of these minerals suggesting that their soluble component Na, Cl, Ca, and SO4 concentrations are not limited by mineral equilibrium (Güler et al. 2002) Hydrochemical facies

Fig. 13 Spatial distribution of the variables and individuals in the axes system F1–F2

Piper diagram is the most widely used graphical form and it provides a convenient method to classify and compare water types based on the ionic composition of different water samples, where ion concentrations are expressed in

Arab J Geosci (2013) 6:3545–3561

3557

Fig. 14 Wilcox diagram for the study area

milliequivalent per liter (Kumar and Ramanathan 2008). In the current study, two hydrochemical facies were identified using a trilinear Piper diagram (Fig. 10): – –

a sodium–chloride facies, which characterizes the wells situated in the East of Sebkhet El Melah and in the zone situated in the North and the East of Grombalia a mixed facies, of Na-Cl-HCO3 type or Na-Ca-ClSO4, that characterizes the wells situated in the artificial recharge zone, as well as, in the south part of the aquifer

The first facies characterizes greatly mineralized waters, while the second characterizes less loaded waters.

Statistical study Q-mode HCA The classification of samples in the Q-mode HCA into six dimensional spaces (salinity, Na+, Ca2+, Mg2+, Cl−, and SO42−) is based on the visual examination of the resulting dendogram which is using the ward’s linkage method and squared Euclidean distances. However, in the judgment of the results, researchers are guided by their best understanding of the geology, hydrogeology, and other criteria in the studied area (Cloutier et al. 2008; Yidana et al. 2008). In our study, two clusters

3558

Arab J Geosci (2013) 6:3545–3561

Fig. 15 USSL diagram for the study area

were identified (Fig. 11) than were classified based on location within the study area and corresponding characteristics: cluster 1 is represented by the wells 2, 5, 18, 21, 24, 26, 29, 30, 31, and 33 situated in the east of Sebkhet El Melah and in the zone situated between the city of Grombalia and the village of Nianou, and it contains 30.3 % of the water samples. These are greatly mineralized characterized by mean EC value of 6.95 ms/cm and present a Na-Cl geochemical facies. Cluster 2 less loaded waters with mean EC value of 3.87 ms/cm and it contains 69.7 % of the water samples presenting a mixed facies, of the Na-Cl-HCO3 type, or NaCa-Cl-SO4 type, that characterizes the wells situated in the artificial recharge zone and in the southern part of the aquifer.

PCA In this study, the PCA of chemical data (Fig. 12) was used to distinguish the contributions of natural and anthropogenic processes to the chemical composition of groundwater in the Grombalia plain. The variables for PCA were temperature, pH, salinity, Na+, Ca2+, K+, Mg2+, Cl−, HCO3−, and SO42−. Following this rule, three independent factors were extracted, which explained 77.86 % of the total variance. The first one was responsible of 52 % of the total variance and was best represented by salinity, Na+, Cl−, SO42−, and Mg2+. These elements contribute to the mineralization of waters in Grombalia aquifer. F2 explained 15 % of the total variance and was mainly represented by HCO3−, Ca2+, and

Arab J Geosci (2013) 6:3545–3561

3559

Fig. 16 A Doneen’s chart for the study area

pH. Additional 10 % of the total variance was explained in F3 and was represented by K+ (Table 3). Besides, spatial distribution of the variables and individuals in the axe systems, F1–F2 (Fig. 13) show the presence of two groups of waters: group I characterizes waters of wells strongly correlated with major ions participating in water mineralization, and it is situated in east of Sebkhet El Melah and in the zone situated between the city of Grombalia and the village of Nianou; group II characterizes waters of wells situated in the artificial recharge zone and in the south part of the aquifer.

Suitability of groundwater for irrigation The suitability of water for irrigation is assumed by many indices (SAR, % Na, potential salinity PS, ES effective salinity, magnesium ratios, Kelley’s ratio, residual sodium carbonate, total hardness, TDS, and EC (Delgado 2010; Kumar and Ramanathan 2008, 2009; Yidana et al. 2010). Among these indices, each one is plotted with the other in different diagrams to ameliorate water classification for irrigation suitability (Wilcox 1955 diagram; United States Salinity Laboratory (USSL) 1954 diagram); Doneen’s chart (Domenico and Schwartz 1990)…).

3560

Arab J Geosci (2013) 6:3545–3561

In the current study, USSL diagram, Wilcox diagram and Doneen’s chart are taken in consideration. Wilcox diagram Wilcox (1955) used % Na and specific conductance in evaluating the suitability of groundwater for irrigation. % Na is given in (Eq. 1) % Na ¼ ðNaþ =Naþ þ Caþþ þ Mgþþ þ Kþ Þ  100

ð1Þ

Sodium concentration is important in classifying irrigation water because sodium reacts with soil to reduce its permeability (Srinivasamoorthy et al. 2011). In the study area, a Wilcox diagram (Fig. 14) shows that all samples are unsuitable for irrigation. Indeed, when the concentration of sodium is high in irrigation water, sodium ions tends to be absorbed by clay particles, displacing Mg and Ca ions. This sodium process of Na in water for Ca and Mg in soils reduces the permeability and eventually results in soil with poor internal drainage (Tijani 1994; Yidana et al. 2010). USSL diagram The USSL diagram relates the SAR with the EC of groundwater and places the samples into irrigation water categories based on the combination of the two parameters. EC is the measured salinity of water and SAR is calculated from Eq. (2) where concentrations are reported in milliequivalent per liter.

p SAR ¼ Naþ = ð2Þ Ca2þ þ Mg2þ =2 Figure 15 is the USSL (1954) diagram for samples from the study area. All the samples are within S1 low SAR category, without risk of destabilization of soil, and range in salinity between C4a (very high salinity) to C4b (extremely high salinity). This groundwater may not be used for irrigation without prior dilution with low salinity waters. Doneen’s chart The permeability index PI of a water sample is calculated from Eq. (3) where the concentration of all ions is in milliequivalent per liter. It measures the collective concentrations of Na+ and HCO3− relative to the total cations content of the water and gives an indication of the probable effects of the water on soil permeability when used for irrigation (Yidana et al. 2010; Kumar and Ramanathan 2008). p PI ¼ ððNaþ HCO3  Þ=ðNaþ þ Caþþ þ Mgþþ ÞÞ  100

ð3Þ

Permeability indices were plotted with the total ionic content of the groundwater samples on a Doneen’s chart

(Domenico and Schwartz 1990), which represent three different classes: CI with best water type for irrigation, CII water generally acceptable and CIII waters unacceptable. In the study area, the PI ranged from 25.48 to 58.85 with mean of 42. Figure 16 showed that all the samples fall within CI: best waters for irrigation purposes.

Conclusion Two major types of groundwater hydrogeochemical zonality were identified in the Grombalia aquifer: the first hydrogeochemical trend, greatly mineralized, characterizes the wells situated in the east of Sebkhet El Melah and in the zone situated in the north and the east of Grombalia, and presents a geochemical facies chlorinated sodique; the second trend, with low salinity, characterizes the wells situated in the artificial recharge zone and in the southern part of the aquifer, presenting a mixed facies of the Na-Cl-HCO3 type or Na-Ca-Cl-SO4 type. Thus, multivariate statistical analysis carried out by using CA and PCA of major hydrochemical ions provide confidence in identifying the extent of the two types of water and in process controlling their mineralization. Groundwater cannot be used for the human consumption without any treatment. For the irrigation purposes, most waters of the aquifer of Grombalia, with relatively elevated salinity, are not suitable, in ordinary conditions. However, they can be used for permeable soils, with an adequate drainage and while applying an excess of leaching water.

References Abdul Halim MD, Majumder RK, Nessa SA, Oda K, Hiroshiro Y, Jinno K (2010) Arsenic in shallow aquifer in the eastern region of Bangladesh: insights from principal component analysis of groundwater compositions. Environ Monit Assess 161:453–472 Al Agha MD (2005) Hydrogeochemistry and carbonate saturation model of groundwater, Khanyounis Governorate—Gaza Strip, Palestine. Environ Geol 47:898–906 Belkhiri L, Boudoukha A, Mouni L, Baouz T (2011) Statistical categorization geochemical modeling of groundwater in Ain Azel plain (Algeria). J Afr Earth Sci 59:140–148 Ben Salem H (1992) Contribution à la connaissance de la géologie du Cap Bon : Stratigraphie, Tectonique et Sédimentologie. Thèse.3éme cycle. Fac Sci Tunis, 203p Benony K (2006) Hydrochemical characterization of groundwater in the Accra plains of Ghana. Environ Geol 50:299–311 Bruce Banoeng Y (2009) Hydrochemical analysis of groundwater using multivariate statistical methods—the Volta region. Ghana KSCE J Civ Eng 13(1):55–63 Carreira PM (2010) Groundwater assessment at Santiago Island (Cabo erde): a multidisciplinary approach to a recurring source of water supply. Water Resour Manag 24:1139–1159

Arab J Geosci (2013) 6:3545–3561 Castany G (1948) Les fosses d’effondrement de la Tunisie Géologie et hydrologie, premier fascicule, plaine de Grombalia et cuvettes de la Tunisie orientale. 1er fascicule. Ann Mines Géologie No 3, 126p Cerling TE, Pederson BL, Damm KLV (1989) Sodium calcium ion exchange in the weathering of shales: implications for global weathering budgets. Geology 17:552–554 Cloutier V, Lefebvre R, Therrien R, Savard MM (2008) Multivariate statistical analysis of geochemical data as indicative of the hydrogeochemical evolution of groundwater in a sedimentary rock aquifer system. J Hydrol 353:294–313 CVRM (2000) Programa ANDAD. Manual do utilizador. CVRMCentro de Geosistemas. Instituto Superior Técnico, Lisboa Davis JC (1986) Statistics and data analysis in geology. Wiley, New York Davis J (2002) Statistics and data analysis in geology. Wiley, NewYork De Andrade A, Stigter TY (2009) Multi-method assessment of nitrate and pesticide contamination in shallow alluvial groundwater as a function of hydrogeological setting and land use. Agr Water Manag 96:1751–1765 Delgado C (2010) Quality of groundwater for irrigation in tropical karst environment: the case of Yucatan, Mexico. Agr Water Manag 97:1423–1433 DGRE (1972–2002) Annuaires de surveillance piézométrique DGRE(1980–2000) Annuaires pluviométriques DGRE (General Direction of Water Resources) (1980–2000) Annuaires d’exploitation des nappes phréatiques Dillon WR, Goldstein M (1984) Multivariate analysis methods and applications, 3rd edn. Wiley, New York, 587pp Domenico PA, Schwartz FW (1990) Physical and chemical hydrogeology. Wiley, New York, pp 410–420 Ennabli M (1970) Etat de l’évolution de la nappe de Grombalia. Rapp. Int. DGRE Tunis (18p) Ennabli M (1975) Recharge artificielle de la nappe de Grombalia Soliman. Note de présentation de projet 73/0.10.DGRE (3 p) Ennabli M (1980) Etude hydrogéologique des aquifères du Nord-Est de la Tunisie pour une gestion intégrée des ressources en eau .Thèse de Doctorat d’Etat. Nice, 570 p FAO (1990) Evaluation de l’emploi d’images SAR géocodées pour la cartographie de l’utilisation du sol en Tunisie : cas de Grombalia; Rapp.FAO/ESA. Direction des sols, 53p Farnham IM, Johannesson KH, Singh AK, Hodge VF, Stetzenbach KJ (2003) Factor analytical approaches for evaluating groundwater trace element chemistry data. Anal Chim Acta 490:123–138 Fisher RS, Mulican WF III (1997) Hydrochemical evolution of sodium-sulphate and sodium-chloride groundwater beneath the Northern Chihuahuan desert, Trans-Pecos, Texas, USA. Hydrogeol J 5:4–16 Güler C, Thyne GD (2004) Hydrologic and geologic factors controlling surface and groundwater chemistry in Indian Wells-Owens Valley area, southeastern California, USA. J Hydrol 285:177–198 Güler C, Thyne G, McCray JE, Turner AK (2002) Evaluation and graphical and multivariate statistical methods for classification of water chemistry data. Hydrogeol J 10:455–474 Hamzaoui-Azaza F, Ketata M, Bouhlila R, Gueddari M, Riberio L (2011) Hydrogeochemical characteristics and assessment of drinking water quality in Zeuss–Koutine aquifer, Southeastern Tunisia. Environ Monit Assess 174(1–4):283–298 Helena B (2000) Temporel evolution of groundwater composition in an alluvial (Pisuerga river, Spain) by principal component analysis. Water Res 34:807–816

3561 Kumar M (2004) An integrated hydrogeochemical and isotopic study of NCR-Delhi, India [in English]. Mphil. Jawaharlal Nehru University, India Kumar S, Ramanathan AL (2008) Geochemical assessment of groundwater quality in vicinity of Bhalswa landfill, Delhi, India, using graphical and multivariate statistical methods. Environ Geol 53:1509–1528 Kumar S, Ramanathan AL (2009) Assessment of groundwater quality and hydrogeochemistry of Manimuktha River basin, Tamil Nadu, India. Environ Monit Assess 159:341–351 Magnus Land BJ (1997) Seasonal variations in the geochemistry of shallow groundwater hosted in granitic till. Chem Geol 143:205– 216 Murray KS (1996) Hydrology and geochemistry of thermal waters in the Upper Napa Valley, California. Ground Water 34: 1115–1124 Parkhurst DL, Appelo CAJ (1999) User’s guide to PHREEQC (ver. 2) A computer program for speciation, batch-reaction, onedimensional transport, and inverse geochemical calculations. USGeol Surv Water-Resources Invest Rept 99–4259 Piper AM (1944) A graphical procedure in the geochemical interpretation of water analysis. Trans Am Geophys Union 25:914– 928 Rabinove CL, Longford RH, Brookhart JW (1958) Saline water resources of North Dakota. U S Geol Surv Water Supply Pap 364:1418 Rodier J (1984) L’analyse de l’eau: eaux naturelles, eaux résiduaires, eaux de mer. Ed. Dunod, Paris, p 1135 Rosen M, Jones S (1998) Controls on the chemical composition of groundwater from alluvial aquifers in the Wanaka and Wakatipu basins, Central Otago, New Zealand. Hydrogeol J 6:264–281 Rouabhia A, Baali F, Fehdi C (2010) Impact of agricultural activity and lithology on groundwater quality in the Merdja area, Tebessa, Algeria Arab. J Geosci 3:307–318 Sebei A (2001) Impact des activités anthropiques sur la qualité des eaux de la plaine de Grombalia. DEA, Tunis, p 99, Fac Sci Srinivasamoorthy K, Vasanthavigar M, Vijayaraghavan SR, Gopinath S (2011) Hydrochemistry of groundwater in a coastal region of Cuddalore district, amilnadu, India: implication for quality assessment. Arab J Geosci. doi:10.1007/s12517-011-0351-2 Tijani MN (1994) Hydrochemical assessment of groundwater in Moro area, Kwara State, Nigeria. Environ Geol 24:194– 202 USSL (1954) Diagnosis and improvement of saline and alkali soils. USDA, Handbook 60:147 Vungopal T (2009) Environmental impact assessment and seasonal variation study of the groundwater in the vicinity of River of the Adyar Chennai, India. Environ Monit Assess Environ 149:81–97 WHO (2004) Guidelines for drinking-water quality, world health organization (p. 515, 3rd ed.) (vol. 1) Geneva, Recommendations Wilcox LV (1955) Classification and use of irrigation water. US Department of Agri., Circ, Washington, DC Yidana SM, Ophoria D, Banoeng-Yakubob B (2008) A multivariate statistical analysis of surface water chemistry data—the Ankobra Basin, Ghana. J Environ Manage 88:697–707 Yidana SM, Bruce Banoeng Y, Akabzaa TM (2010) Analysis of groundwater quality using multivariate and spatial analyses in the Keta basin, Ghana. J Afr Earth Sci 58:220–234