C. R.

Geoscience 341 (2009) 21–35

http://france.elsevier.com/direct/CRAS2A/

Surface geosciences (hydrology-hydrogeology)

Seawater intrusion and associated processes:
Case of the Korba aquifer (Cap-Bon, Tunisia)
Lamia Kouzana a,*, Abdallah Ben Mammou a, Mennoubi Sfar Felfoul b
a
Laboratoire des ressources minérales et environnement, département de géologie, faculté des sciences de Tunis,
université de Tunis El Manar, 2092 Tunis, Tunisia
b
Institut national agronomique de Tunis, 43, avenue Charles-Nicolle, 1082 Tunis, Tunisia
Received 27 September 2007; accepted after revision 26 September 2008
Available online 18 November 2008

Presented by Ghislain de Marsily

Abstract
It has been established that intensive agricultural activities, usually, increase the risk of groundwater quality degradation through
high groundwater pumping rates. In fact, the uncontrolled groundwater extraction causes a modification of natural flow systems and
induces seawater intrusion from the coast and causes the groundwater quality deterioration. The Korba aquifer is located in the
North-East of Tunisia, where a semi-arid Mediterranean climate prevails. The dry season is pronounced and this aggravates the
situation, given that the highest water demand usually coincides with the drought period (dry weather conditions). The principal aim
of this study is to characterize the hydrochemistry of this coastal aquifer, identifying the main processes that occur in the system, and
to determine the extent of marine intrusion in the aquifer. In order to achieve this aim, geophysical and chemical parameters were
measured, such as vertical electrical soundings (VES), electrical conductivity, pH, temperature, anions and cations concentrations.
The analytical results obtained in the hydrochemistry study were interpreted using ion correlations with chloride and SO42-/Cl and
Mg2+/Ca2+ ratios, in conjunction with calculations of the ionic deviations and the saturation indexes. Saturation indexes are
calculated with the PHREEQC 2.8 software used for mineral saturation modelling of the aquifer seawater-freshwater mixture. The
high groundwater salinity anomaly observed in Diar El Hajjej, Garaet Sassi and Takelsa-Korba zones was explained by the presence
of seawater intrusion in these areas. This hypothesis is based on high chloride concentrations, the inverse cation exchange reactions,
and the lower piezometric level compared to sea level. To cite this article: L. Kouzana et al., C. R. Geoscience 341 (2009).
# 2008 Académie des sciences. Published by Elsevier Masson SAS. All rights reserved.

Résumé
Intrusion marine et les processus associés : le cas de l’aquifère de Korba (Cap Bon, Tunisie). La nappe côtière du Plio-
Quaternaire de la région de Korba a été très sollicitée par les agriculteurs, ce qui a engendré, probablement, une inversion du
gradient hydraulique et, par conséquent, l’avancée du biseau salé. Cela se traduit par une évolution spatio-temporelle de la
piézométrie et de la qualité chimique des eaux. L’intrusion marine est mise en évidence par des méthodes géo-électriques (sondages
électriques verticaux de type Wenner) et par l’étude de la conductivité électrique et des éléments chimiques majeurs. Pour identifier
les processus et les réactions chimiques qui gouvernent sa salinisation et déterminent la limite de l’intrusion marine, nous avons eu
recours à l’interprétation des résultats en utilisant la corrélation des éléments majeurs avec les chlorures, la variation des rapports
SO42/Cl et Mg2+/Ca2+, le calcul des écarts ioniques et des indices de saturation. Les indices de saturation sont calculés par le

* Corresponding author.
E-mail address: [email protected] (L. Kouzana).

1631-0713/$ – see front matter # 2008 Académie des sciences. Published by Elsevier Masson SAS. All rights reserved.
doi:10.1016/j.crte.2008.09.008
22 L. Kouzana et al. / C. R. Geoscience 341 (2009) 21–35

logiciel PHREEQC 2.8, utilisé pour la modélisation de l’état de saturation des minéraux à l’état de mélange eau douce–eau de
mer dans un système aquifère. Les résultats de ces analyses montrent que les zones à fortes salinités de Diar el Hajej, Garaet
Sassi et Tazerka-Korba sont contaminées par les eaux marines. L’intrusion d’eau marine est mise en évidence par les fortes
teneurs en chlorures, par la présence des réactions d’échange cationiques inverse et par une piézométrie négative. Pour citer cet
article : L. Kouzana et al., C. R. Geoscience 341 (2009).
# 2008 Académie des sciences. Published by Elsevier Masson SAS. All rights reserved.

Keywords: Korba aquifer; Seawater intrusion; Hydrochemistry; PHREEQC 2.8 software; Ionic deviation; Saturation indexes; Model

Mots clés : Nappe côtière de Korba ; Intrusion marine ; Hydrochimie ; Logiciel PHREEQC ; Écarts ioniques ; Indices de saturation ; Modélisation

1. Introduction Cap-Bon (Fig. 1). This aquifer is bounded to the south

by Oued Boulidine, to the north by Oued Lebna, to the
The delicate exploitation of a coastal aquifer usually west by Jebel Abdeerahmen and to the east by the
induces problems of quantity and quality [13]. Indeed, a Mediterranean Sea.
prolonged drought can be the origin of the rupture of the The outcropping formations of the study area are
seawater–freshwater interface equilibrium, which in represented by Mio-Plio-Quaternary sediments [5]. The
turn may cause seawater intrusion [17,18]. This can be lower part of the Middle Miocene corresponds to
seen through the space-time evolution of piezometric detrital deposits known as the Beglia Formation. The
heads and the water chemical quality of these aquifers. upper part is composed of lenticular sandstones and
Tunisia, like all Mediterranean countries, has experi- marls with lignite levels called the Saouaf Formation. In
enced considerable periods of drought during the last the study area, the Upper Miocene is absent. This is
two decades. The resulting water deficit was compen- believed to have been caused by the widespread
sated through an increased withdrawal from aquifers. erosional activity due to the Miocene orogeny [14].
However, aquifers close to coastal areas are the most The trangressional marine Pliocene sediments were
threatened as they are marked by a very strong borehole unconformably deposited on these folded and eroded
density resulting from an intense overexploitation, and formations. These deposits are mainly composed of
shown by an increase in the mineralization of their sandstone-sand-marl alternations topped by sandstones
waters [10]. This constitutes a limiting factor for their and sand [11]. This facies changes laterally to
use. In this work, we will try to demonstrate seawater argillaceous sands or to consolidated sandstones, more
intrusion and its extension in the Plio-Quaternary or less argillaceous.
aquifer of Korba, a Tunisian aquifer which is very The Quaternary deposits are usually composed of
heavily exploited by farmers. two units: the lower unit of marine facies corresponds to
The main purpose of our hydrochemical study is to sandy limestone with molluscs indicating the maximum
determine the spatial extension of the salinization and flooding surface (MFS) of the Tyrrhenian transgression.
also to identify the origins and the mechanisms governing The upper unit is mainly composed of a continental
its contamination. This will be shown through the use of facies [28] with the occurrence of oolitic limestone and
the chemical contents in conjunction with the different coprolites or pelloïds. These deposits form, nowadays,
processes, which modify the water mixture character- coastal consolidated dunes built by wind following
istics [33]. This change is due to the absence of balance marine regression [7,28]. The old consolidated dunes
between the aquifer and the water mixture. Indeed, the cover the Tyrrhenian deposits [28]. The encrusted
carbonates and clays contribute in the dissolution and the limestone extends over significant distances. They are
precipitation of some minerals, with cation exchange, very rich in calcite, silica, sometimes in gypsum and
which acts in opposition to the changes caused by the alumina and frequently coloured by iron salts. Finally,
marine intrusion. These processes are the principal The Holocene deposits are formed by recent alluvia of
factors modifying the geochemistry of the water salinized Oued Chiba, by sebkhas deposits and current dunes and
by the seawater intrusion [33]. beaches.

2. Geographic and geologic situation 3. Hydrogeology

The study area, covering 430 km2 [30], is located in The hydrogeological study of the Korba area shows
northeastern Tunisia, within the eastern coastal Plain of that the Plio-Quaternary detrital deposits constitute a
 L. Kouzana et al. / C. R. Geoscience 341 (2009) 21–35 23

Fig. 1. Geology and samples and VES location in the Korba aquifer (June 2006).
Fig. 1. Localisation de la géologie de l’échantillonnage et des sondages électroniques verbiaux dans l’aquifère de Korba (juin 2006).

potential shallow aquifer. The marls of the Middle limestone before reaching the sea. Indeed, the Tyrrhenian
Miocene form the impermeable substratum of this deposits constitute a good reservoir and a key point of
aquifer [14] (Fig. 2). coastal aquifer recharge. Meteoric water recharge of the
The aquifer recharge is primarily made through the aquifer system is thought to occur through the coastal
carved glacis rivers having a very porous lithology. The consolidated barriers, which characterize the Tyrrhenian
intercommunication between the Pliocene formations beach. This contribution is shown through its great
and recent alluvia is absent. The Quaternary recharge is infiltration capacity and its role as topographic obstacle
increased partly by Pliocene sediments infiltration. Part against surface run off. This helps water infiltration
of the groundwater flow occurs within the Tyrrhenian downstream in case it escapes infiltration upstream [14].

Fig. 2. Geological cross section B–B’.

Fig. 2. Coupe géologique B–B’.
24 L. Kouzana et al. / C. R. Geoscience 341 (2009) 21–35

Fig. 3. Piezometric map of Korba aquifer (June 2006).

Fig. 3. Carte piézométrique de l’aquifère de Korba (juin 2006).

The recent 2006 piezometric map (Fig. 3) shows and the different factors monitoring its space-time
piezometric depressions related to overexploitation of evolution.
the aquifer with negative piezometric levels on the order This article presents the analytical results of 28
of 12 m. This map (Fig. 3) shows a multidirectional samples, which were collected in June 2006. Twenty-six
flow mainly oriented to the piezometric depressions samples were collected from shallow wells and two
located at Diar El Hajjej-Garaet Sassi and to the east of were taken from the deep aquifer.
Tafelloune. Electrical conductivity, temperature and pH were
measured in situ. The cations (Na+, K+, Ca2+ and Mg2+)
4. Methods were analysed at ‘‘Laboratoire des resources minérales
et environnement’’ of the Faculty of Sciences of Tunis
A multidisciplinary approach is needed in order to using Atomic Absorption Spectrometry. Bicarbonates
establish marine intrusion monitoring. Previous studies were determined by the volumetric method, Cl was
[3,4,9,12,19–21,23–25,29,32,33,35,36,38,41], through determined using the Mohr method and finally SO42
analytical, geophysical and modelling methods, have was, spectrophotometrically, analysed.
approached the problem and attempted to position the The interpretation of the analytical results is shown
freshwater–seawater interface. numerically and graphically through the saturation
Other studies [1,15,20,22,26,27,34,38,39] defined index (SI), ionic deviations, Piper Diagram and binary
the chemical processes and reactions, which character- diagrams.
ize mineralization and would thus be responsible for SI and ionic deviations were calculated to better
groundwater chemical elements enrichment or impov- understand the hydrogeochemical processes that take
erishment. place in the aquifer. Calculation of the ionic deviations
Geochemical properties were determined through (D) corresponds to a comparison of the measured
different analyses, in order to characterize salinisation concentration of each constituent to its theoretical
 L. Kouzana et al. / C. R. Geoscience 341 (2009) 21–35 25

concentration of a known theoretical freshwater–sea- This seawater contribution was then used to calculate
water mixture calculated from the Cl concentration of the theoretical concentration of each ion:
the sample (see formula below) [16]: C i mix ¼ f sea :Ci;sea þ ð1  f sea Þ:Ci; f
DC i ¼ Ci; sample  C i mix ; These calculations take into account that Cl is a
conservative tracer [37]. In fact, Cl is not usually
where DCi is the ionic deviation of the ion i, removed from the system due to its high solubility [2].
Ci, sample is the measured concentration of the ion i in The only inputs are either from the aquifer matrix salts
the sample, and Ci mix is the theoretical concentration of or from a salinization source like seawater intrusion
the ion i for the theoretical (conservative) mixture of etc.
freshwater–seawater. The theoretical mixture concen- The PHREEQC code was used to model calcite,
trations were calculated taking into account seawater dolomite and gypsum saturation states of the fresh-
contribution ( f sea) based on the chloride contents in the water–seawater mixing process in a closed system. This
sample (CCl, sample), the freshwater Cl concentration code calculated the samples ionic strength and those of
(CCl,f ) and the seawater Cl concentration (CCl,sea): the theoretical mixtures between freshwater and sea-
water. Moreover, PHREEQC provided the geochemical
ðCCl; sample  C Cl; f Þ composition of these theoretical mixtures.
f sea ¼ ;
ðC Cl; sea  CCl; f Þ This software is a specialized code that uses the Mass
Balance and Electroneutrality equations to simulate
This allows calculation of seawater contents (%) in geochemical reactions, such as mixing of waters,
each sample. addition of net irreversible reactions to solution,

Fig. 4. Iso-resistivity contour map.

Fig. 4. Carte des contours d’iso-résistivité.
26 L. Kouzana et al. / C. R. Geoscience 341 (2009) 21–35

dissolving and precipitation phases to achieve equili- out between Korba and Oued Lebna. The outcome of
brium with the aqueous phase [31]. The modelled this geophysical survey is given below.
saturations states of these minerals and the geochemical
compositions of mixing waters will be compared to the 5. Geo-electrical study
measured ones. This comparison would confirm the
cationic exchange processes, precipitation, and mineral To evaluate the extension of the marine intrusion in the
dissolutions. coastal aquifer of the eastern coast of Cap-Bon, 38
Prior to these geochemical analyses, a geophysical vertical electrical soundings (VES) using the Wenner
survey combined to borehole logging data was applied process were carried out in the area of Korba. Lines AB of
on the study area. In fact, there is an empirical relation these electric surveys are directed parallel to the coast.
between the resistivity of water and its mineralization This orientation facilitates the quantitative interpretation
[8,25]. This geophysical method was implemented for of these electric soundings, carried out by the Winsev
the aquifer geometry study and the marine intrusion software. This geophysical study aims to construct an iso-
evaluation. In addition, the electrical resistivity of rocks resistivity map of seawater intrusion. The calibration of
depends on many factors such as groundwater salinity, the VES results with borehole data enabled us to establish
saturation, aquifer lithology and porosity. geological cross sections (Fig. 2), which shed more light
Electrical resistivity data have been used to identify on the geometry of the aquifer and estimate the vertical
the geographical extent of salinization in the aquifer, as and horizontal configuration of the seawater intrusion.
resistivity decreases with increasing salinity. Indeed, 38 Thus, the main purpose of this study is to highlight the
electric surveys using the Wenner process were carried spatial distribution of the salinization and its extension.

Fig. 5. Iso-chloride (ppm) contour map of Korba aquifer (June 2006).

Fig. 5. Carte des contours d’iso-valeur des chlorures (ppm) de l’aquifère de Korba.
 L. Kouzana et al. / C. R. Geoscience 341 (2009) 21–35 27

The iso-resistivity contour map (Fig. 4) shows zones of zones. This is to know the marine intrusion space
contamination which resistivity lies between 1 to 17 Vm. extension and to do a comparison between the
These contamination zones characterize the coastal zone geochemical phenomena observed in these areas.
and the Diar El Hajjej and Garaet Sassi areas. Some zones Two samples were taken from the deep aquifer.
of relatively high resistivity located along the coast are Water samples collected from these areas show high
the consequence either of the direct infiltration of chloride contents (Fig. 5).The determination of the
rainwater through the recent littoral sandy dunes or to the groundwater salinity origin was based on a spatial
rise of the substratum. This map shows the extent of the evolution of the total dissolved salt (TDS) and the major
marine invasion of the area of Diar El Hajjej compared to elements contents (Table 1). The high values of the TDS
that of Tefelloune. Given these geophysical findings, a of water of the aquifer relate to the zones characterized
geochemical study was undertaken and the following by significant piezometric depressions such as the areas
presents the main results and interpretations. Former of Diar El Hajjej-Garaet Sassi and Korba Tazerka.
studies have showed the complementarity of these two The Piper Diagram illustrates three chemical facies
methods [6,42]. (Fig. 6): Na-Cl facies, Ca-Mg-Cl facies and SO4-mixed
facies. The last one characterizes one sample of the zone
6. Hydrochemical results and discussion of Beni Khiar. The Ca–HCO3 water type is absent. This
facies characterizes the shorter residence time waters of
Water sampling was carried out covering the Korba- the detritic aquifers [1].
Diar El Hajjej-Garaet Sassi zone and the coastal zone. The calculation of seawater contents (%) presents
The geophysical method enabled to detect low values lower than 14% of seawater.
resistivities in these areas. However, during this study, Calcite, dolomite and gypsum SI were also
we had recourse to a sampling, which exceeds these calculated using the PHREEQC 2.8 software to verify

Table 1
Sampling Analytical results of Korba aquifer (June 2006) (ppm).
Tableau 1
Résultats analytiques des échantillons d’eau de l’aquifère de Korba (juin 2006) (ppm).
No. Na+ K+ Ca2+ Mg2+ Cl HCO3 SO42 Error % Mg/Ca SO4/Cl TDS Ionic strength
1 594 41 388 168 878 325 1557 2.03 0.71 1.31 3951 7.36E-02
2 302 60 198 87 457 343 568 2.26 0.73 0.92 2014 3.86E-02
3 440 18 296 133 1124 412 323 0.16 0.74 0.21 2746 5.62E-02
4 518 44 219 106 983 331 438 1.16 0.80 0.33 2639 5.15E-02
5 215 24 128 45 351 308 160 4.59 0.58 0.34 1231 2.40E-02
6 979 61 240 85 1510 436 431 3.57 0.58 0.21 3741 6.98E-02
7 204 9 90 56 386 198 165 1.58 1.02 0.32 1106 2.23E-02
8 743 24 314 110 1370 360 605 0.45 0.58 0.33 3526 6.81E-02
9 437 9 333 60 983 215 490 0.75 0.30 0.37 2528 5.07E-02
10 457 10 208 54 808 383 290 0.24 0.43 0.26 2209 4.19E-02
11 279 4 159 103 737 285 318 5.66 1.07 0.32 1885 3.77E-02
12 166 9 61 48 281 180 163 0.69 1.30 0.43 908 1.81E-02
13 1138 23 426 301 2458 360 872 1.43 1.16 0.26 5577 1.13E-01
14 279 212 179 90 457 459 502 4.73 0.83 0.81 2178 3.96E-02
15 391 20 216 109 843 325 316 2.23 0.83 0.28 2220 4.51E-02
16 259 16 266 127 597 354 478 4.07 0.79 0.59 2097 4.35E-02
17 849 0 346 128 1475 244 935 0.28 0.61 0.47 3977 7.67E-02
18 340 14 136 167 843 354 229 1.81 2.03 0.20 2081 4.39E-02
19 514 18 278 269 1440 372 292 5.41 1.59 0.15 3183 7.12E-02
20 1198 112 226 270 2845 459 57 0.25 1.97 0.01 5166 1.05E-01
21 402 15 149 143 913 343 219 1.57 1.58 0.18 2184 4.50E-02
22 337 10 173 187 878 354 203 5.62 1.78 0.17 2142 4.73E-02
23 333 12 271 201 913 267 466 5.91 1.22 0.38 2464 5.44E-02
24 560 27 470 193 1756 227 433 1.71 0.68 0.18 3666 8.01E-02
25 563 14 385 159 1545 238 356 2.00 0.68 0.17 3260 7.01E-02
26 536 12 492 199 1861 238 372 0.35 0.67 0.15 3711 8.21E-02
27 542 15 179 154 983 378 377 4.38 1.42 0.28 2628 5.30E-02
28 511 15 120 94 632 627 207 5.71 1.30 0.24 2207 3.59E-02
28 L. Kouzana et al. / C. R. Geoscience 341 (2009) 21–35

Fig. 6. Water sampling analytical results plotted in the Piper Diagram (June 2006 survey).
Fig. 6. Représentation des résultats analytiques des échantillons d’eau dans le diagramme de Piper (juin 2006).

the salinization phenomena, the exchange processes, activity coefficients up to ionic strengths of about 1 M,
the precipitation and dissolution of some minerals, and which would include most solutions of geological
the dolomitization processes. Gypsum SI were mod- interest, including seawater. The Davies equation is
elled for different mixing rates to assess the possibility slightly more accurate in the range of 0.1 to 1 M ionic
of SO42 reduction or an alternative SO42 source (such strength. Above these concentrations, both the Davies
as gypsum dissolution). Fig. 7 shows calcite, dolomite and Debye-Hückel equations are increasingly inaccurate
and gypsum saturation index versus seawater in the [40]. The ionic strength values calculated lie between
mixture. Also shown is the theoretical line of 1.81E-02 and 7.30E-01 (Table 1). The last value is that of
conservative mixing. seawater. This means that the use of the PHREEQC code
The Mg/Ca ratios of the aquifer water are modelled is validated in the case of the explored salinity ranges.
by the PHREEQC software and calculated from the
concentration of chlorides. According to the propor- 6.1. Diar El Hajjej and Garaet Sassi
tions of seawater, the Mg/Ca ratios modelled by the two
methods have the same shape, which enhances the The relationship between Na+ and Cl (Fig. 8a)
reliability of these methods for the calculation of demonstrates low Na+ contents in these samples. In fact,
theoretical concentrations. the Diar El Hajjej Garaet Sassi Korba samples are
The PHREEQC code was also used to calculate the positioned lower than the theoretical mixing fresh-
ionic strength. water–seawater line. These samples are especially
Debye-Hückel has proved to be remarkably success- controlled by an inverse cation exchange between Na
ful in predicting activity coefficients in dilute solution. and Ca-Mg. This behaviour is documented by a
The extended Debye-Hückel Equation is most useful at majority of samples enriched in Ca2+ and/or Mg2+
concentrations less than 0.1 M, which includes many with respect to the theoretical mixture (Fig. 8b and c).
natural waters and provides adequate approximation for Divergences from the lines could be attributed to an ion
 L. Kouzana et al. / C. R. Geoscience 341 (2009) 21–35 29

Fig. 7. SI of calcite (a), gypsum (b), dolomite (c) and halite (d) of the Korba aquifer samples. SI of the conservative mixture of freshwater and
seawater.
Fig. 7. Index de saturation de la calcite (a), du gypse (b), de la dolomite (c) et de l’halite (d) dans les échantillons de l’aquifère de Korba. Index de
saturation du mélange conservatif eau douce/eau de mer.

exchange phenomenon. In fact, many samples present Ca2+ and/or Mg2+ is released, thus water quality
an excess of calcium and/or Mg2+ and a deficit of changes from Na-Cl rich to CaCl2-rich water and/or
sodium compared to chlorides. When seawater intrudes Mg-Cl2[12].
a freshwater aquifer, an inverse cation exchange occurs This behaviour is confirmed by the calculation of
and sodium is captured by the exchanger (clay), while DNa+, DCa2+ and DMg2+.
30 L. Kouzana et al. / C. R. Geoscience 341 (2009) 21–35

Fig. 8. Na+, Mg2+, Ca2+/Cl relationships (meq/l) (a,b,c), DNa+, DCa2+ and DMg2+ relationships with % seawater (d), Mg2+/Ca2+ ratio relationship
with % seawater (e) and SO42/Cl ratio relationship with chlorides (f) in Diar El Hajjej Garaet Sassi area (June 2006).
Fig. 8. Relations Na+, Mg2+, Ca2+/Cl (meq/l) (a,b,c) de DNa+, DCa2+ et DMg2+ avec le pourcentage d’eau de mer (d), du rapport Mg2+/Ca2+ avec le
pourcentage d’eau de mer (e) et du rapport SO42-/Cl- avec les chlorures (f) dans la zone de Diar El Hajjej Garaet Sassi (juin 2006).

In fact, the DNa+ is usually negative and the DCa2+ an average Mg2+/Ca2+ ratio (between 0.668 and 0.681).
and DMg2+ are positive (Fig. 8d). The most logical This is caused by their enrichment in calcium, which
explanation for this excess of Ca2+ and Mg2+ is an originates by the inverse cation exchange between water
inverse cation exchange. This confirms the mixture of and clays of the substratum [38].
freshwaters with seawater. The higher the deviation (D), The samples of Diar El Hajjej-Garaet Sassi areas
the more important are the processes modifying the present SI higher than the modelled values of the
content of the mixing. conservative mixture. Differences between the theore-
The Mg2+/Ca2+ ratio increases proportionally to tical saturation states and the ones calculated for the
seawater in the mixture [34,38,39]. The SO4/Cl ratio samples arise due to the non-conservative nature of
decreases as the seawater proportion in the mixture dissolution and precipitation of the calcite and the CO2
increases [34,38,39]. flux (PCO2 ) [33]. The Diar El Hajjej-Garaet Sassi
This seawater contamination hypothesis is con- samples plot above the calcite SI line for the
formed by the fact that the highest Mg2+/Ca2+ ratio conservative mixing and all these samples are over-
values (Fig. 8e) and the weakest SO42/Cl ratios saturated in this mineral. These samples are over-
(Fig. 8f) characterize the Diar El Hajjej and Garaet Sassi saturated with respect to gypsum. Gypsum SI are
areas samples. generally higher than the theoretical conservative
Some Diar El Hajjej area samples are characterized mixing line. This requires a source in SO4 other than
by low SO42/Cl ratio (between 0.148 and 0.182) and seawater like the dissolution of gypsum present in the
 L. Kouzana et al. / C. R. Geoscience 341 (2009) 21–35 31

catchment area or the evaporation of the irrigation water 6.2. Coastal area
excess. The dolomite SI are higher than 0.
The basic requirement for the deposition of dolomite The Na+ and Cl relationship demonstrates that the
is that the [Mg]/[Ca] ratio has to exceed 1. This coastal area samples plot above the seawater-freshwater
condition is met in the majority of the samples. These mixture line (Fig. 9a). This is confirmed by the fact that
ratios are characterized by values higher than those DNa+ is positive in most samples of this area. The most
modelled suggesting the presence of another Mg2+ logical explanation for this excess of Na+ is that a direct
source different from seawater. This water reached the cation exchange is taking place with the clay substratum.
precipitation condition of dolomite. This exchange produces Na+ release to the solution and
The Mg/Ca ratios of three samples are lower than Ca2+ and/or Mg2+ capture from it. This process is caused
those modelled with proportion higher than 5% in by the water flushing. Fig. 9b and c show that most of the
seawater. These samples are characterized by dolomite samples plot above the mixing line. Two samples taken
SI > 0.This confirms the assumptions of the dolomite from the coastal zone have positive DNa+ and negative
precipitation causing thus the Mg2+/Ca2+ ratio reduction DCa2+ and/or DMg2+ (Fig. 9d). This shows the
or the inverse cation exchange Na+-Ca2+ and the contribution of the coastal dunes to the recharge of the
salinization. aquifer and the presence in this zone of clay layers
The halite SI follows the conservative mixing line playing the role of barrier against seawater intrusion.
direction. This confirms the marine origin of this In general, DNa+ DCa2+ and DMg2+ are positive
mineral. (Fig. 9d). This confirms the hypothesis of the aquifer

Fig. 9. Na+, Mg2+, Ca2+/Cl relationships (meq/l) (a,b,c), DNa+, DCa2+ and DMg2+ relationships with % seawater (d), Mg2+/Ca2+ ratio relationship
with % seawater (e) and SO42/Cl ratio relationship with chlorides (f) in the coastal area (June 2006).
Fig. 9. Relations Na+, Mg2+, Ca2+/Cl (meq/l) (a,b,c) de DNa+, DCa2+ et DMg2+ avec le pourcentage d’eau de mer (d), du rapport Mg2+/Ca2+ avec le
pourcentage d’eau de mer (e) et du rapport SO42/Cl avec les chlorures (f) dans la zone côtière (juin 2006).
32 L. Kouzana et al. / C. R. Geoscience 341 (2009) 21–35

flushing in the coastal area. The excess of Ca2+ and Mg2+ [Mg]/[Ca] ratio exceeds 1. This condition is met only in
is due to the dissolution of the carbonates and the gypsum two of these samples (No. 7 and 12). In spite of this, the
present in the catchment area and dispersed in the aquifer. dolomite and calcite SI are negative and dolomitization
This phenomenon is accompanied by a sulphate- is not reached. This situation is due to the low salinity
enrichment confirming the hypothesis of gypsum (TDS) of these water samples.
dissolution. The Ca2+ and Mg2+ excesses cause the The gypsum and halite SI are negative but they are
calcite, dolomite oversaturations. This is shown in the SI higher than the theoretical conservative mixing line.
of these minerals calculated by PHREEQC. The behaviour of the halite SI is explained by the excess
Two samples only (No. 3 and 4) have negative DNa+. of Na+. In fact, all the samples plot above the theoretical
These samples are located in the north of Beni Khiar. mixing line. The gypsum SI behaviour requires a source
They are characterized by an inverse cation exchange. in SO4 other than seawater like the dissolution of
The most coastal samples present average Mg2+/ gypsum present in the catchment area or the evaporation
Ca and SO42/Cl ratios values (Fig. 9e and f). These
2+
of the irrigation water excess.
samples are located in the coastal consolidated dunes of
the Tyrrhenian. In this area, the aquifer is strongly 6.3. Internal zone and deep aquifer
recharged by meteoric water [14]. This confirms the
water flushing theory of the aquifer in the coastal zone. The Na+ and Cl relationship indicates that the
The majority of samples present SI higher than the majority of the internal area samples approach the mixing
modelled values of the conservative mixture. The basic line (Fig. 10a) indicating a weak interaction between the
requirement for the deposition of dolomite is that the freshwater–seawater mixing and the substratum clay.

Fig. 10. Na+, Mg2+, Ca2+/Cl relationships (meq/l) (a,b,c), DNa+, DCa2+ and DMg2+ relationships with % seawater (d), Mg2+/Ca2+ ratio relationship
with % seawater (e) and SO42/Cl ratio relationship with chlorides (f) in the internal area and the deep aquifer (June 2006).
Fig. 10. Relations Na+, Mg2+, Ca2+/Cl (meq/l) (a,b,c) de DNa+, DCa2+ et DMg2+ avec le pourcentage d’eau de mer (d), du rapport Mg2+/Ca2+ avec le
pourcentage d’eau de mer (e) et du rapport SO42/Cl avec les chlorures (f) dans la zone interne et dans l’aquifère.
 L. Kouzana et al. / C. R. Geoscience 341 (2009) 21–35 33

Fig. 10b and c show the calcium and the magnesium 7. Conclusion
behaviour. These cations have the most distinctive
distribution pattern in the mixing water sample and The aquifer of Korba shows especially high values of
exhibit concentrations that exceed the theoretical TDS in the coastal zones, the Diar El Hajjej, Garaet
concentrations. This is produced by the inverse Sassi and Tazerka-Korba. These high TDS values are
exchange between Na and Ca-Mg caused by seawater due to contamination by sea water. This assumption is
intrusion and the dissolution of dolomite. justified by negative piezometric values, strong chloride
The DNa+ is usually negative but these values concentrations and weak electric resistivity. In the
approach 0 (Fig. 10d). This fact confirms the weak various studied diagrams, we observe a tendency
interaction between the mixture and the substratum. As between sample composition and marine water com-
an exception, samples No. 13 and 20 are characterized position, which gives an additional argument in favour
by negative weak DNa+ values and positive DCa2+ and of the presence of a marine intrusion in the coastal zone,
DMg2+ values. This is in favour with the inverse cation areas of Diar El Hajjej, Garaet Sassi and Tazerka-Korba.
exchange. This seawater intrusion is accompanied by other
DCa2+ and DMg2+ are positive in all the samples processes, which modify the hydrochemistry of the
except the samples No. 9 (Fig. 10d). The probable coastal aquifer. The most remarkable process is that of
origin of this excess of Mg2+ and Ca2+, in addition to the the inverse cation exchange, characteristic of the
inverse cation exchange, is the dissolution of the changes of the theoretical mixture of seawater-fresh-
gypsum and the dolomite. water, which is carried out between clays and the
Mg/Ca is higher than 1 or near to 1 (Fig. 10e). This aquifer water. This exchange consists in the release of
confirms the dolomitization phenomenon. Ca2+ and/or Mg2+ and the adsorption of Na+. This
Only samples No. 9 and 10 have low Mg/Ca values. process characterizes Diar El Hajjej-Garaet Sassi and
These values are lower than those modelled by the internal areas.
PHREEQC and by those calculated from the Cl The coastal zone is characterized by a low salinity
concentration. This resulted in weak dolomite satura- and weak concentrations chloride. These low (Cl)
tions indexes. Dolomite and calcite IS are usually values are due to the infiltration of meteoric water in the
positive in this area. Gypsum SI are higher than those Tyrrhenian consolidated dunes [14] and to the presence
modelled. Halite SI are on the mixture line. of clay layers playing as a barrier against seawater
In the deep aquifer, the Na+, Ca2+, Mg2+ and Cl intrusion (Fig. 2). Moreover, the direct cation exchange
(Fig. 10a–c) relationships demonstrates that these between Na+ and Ca2+ and/or Mg2+ (Na+ release and
samples plot above the seawater-freshwater mixture adsorption of Ca2+ and/or Mg2+) indicating a local
line. This is confirmed by positive DNa+, DCa2+ and flushing of water of the aquifer is detected in this zone.
DMg2+ (Fig. 10d). This Na+ excess is caused by a direct This phenomenon is accompanied by a sulphate
cation exchange with the clay present in the aquifer. enrichment. The most probable source of sulphates is
This process is caused by the water flushing. the dissolution of the gypsum dispersed in the aquifer.
Sample No. 14 presents close Mg2+/Ca2+ (Fig. 10e) Other processes such as dolomitization (dolomite
and SO42/Cl ratios values (Fig. 10f) (respectively, precipitation) exist in this aquifer, but they are not as
0.812 and 0.826). This is where the aquifer is recharged obvious as the cations exchange processes. Dolomitiza-
by meteoric water. This confirms the water flushing tion is indicated by an Mg2+/Ca2+ ratio higher than 1,
theory of the aquifer in this area. Calcite and dolomite negative values of DMg2+ and positive values of DCa2+
SI are positive in this sample. This explains the Mg/Ca and dolomite SI > 0.
ratio reduction. It is believed that due to the increased demand for
On the other hand, sample No. 28 is characterized by freshwater in agriculture, the area of saltwater intrusion
a high Mg2+/Ca2+ ratio and a low SO42/Cl ratio. In will eventually spread, particularly during dry periods.
spite of this, the dolomite and calcite SI are negative and Consequently, both groundwater monitoring and man-
dolomitization is not reached. This situation is due to agement and groundwater conservation are essential for
the low salinity (TDS) of this water sample. The efficient surveillance of the saltwater intrusion.
gypsum SI is negative but it is lower than the theoretical
conservative mixing line. References
The behaviour of the halite SI of the two samples is
explained by the excess of Na+. In fact, these SI are [1] D.M Allen, M. Suchy, Results of the Groundwater Geochemistry
slightly higher than those modelled. Study on Saturna Island, British Columbia. Rapport final préparé
34 L. Kouzana et al. / C. R. Geoscience 341 (2009) 21–35

pour : Islands Trust Victoria, B.C. Earths Sciences Simon Fraser [20] S. Grassi, G. Cortecci, Hydrogeology and geochemistry of the
University, 2001. multilayered confined aquifer of the Pisa plain (Tuscany –
[2] C.A.J. Appelo, D. Postma, Geochemistry, Groundwater and central Italy), Applied Geochemistry 20 (2004) 41–54.
Pollution. Balkema, Rotterdam, 1993, 536p. [21] C. Güler, G.D. Thyne, J.E. McCray, A.K. Turner, Evaluation of
[3] B. Arfib, J. Ganoulis, Modélisation physique de l’intrusion d’eau graphical and multivariate statistical methods for classification
de mer dans un aquifère karstique : cas de l’Almyros d’Héraklion of water chemistry data, J. Hydrogeol. 10 (2002) 455–474.
(Crète), C. R. Acad. Sci. Paris, Ser IIa 336 (2004) 999–1006. [22] P. Hudak, Sulfate and chloride concentrations in Texas aquifers,
[4] B. Arfib, T. Cavalera, E. Gilli, Influence de l’hydrodynamique J. Environ. Int. 26 (2000) 55–61.
sur l’intrusion saline en aquifère karstique côtier, C. R. Geos- [23] U. Kafri, A. Arad, Current Subsurface intrusion of Mediterra-
cience 338 (2006) 757–767. nean seawater. A possible source of groundwater salinity in the
[5] H. Ben Salem, Notice explicative de la carte géologique de la rift valley system, Israel, J. Hydrol. 44 (1979) 267–287.
Tunisie à 1/50.000 Menzel Bou Zelfa Feuille no 22, 1998. [24] S. Krimissa, J.L. Michelot, L. Bouchaou, J. Mudry, Y. Hsissou,
[6] M. Boughriba, A. Melloul, Y. Zarhloule, A. Ouardi, Extension Sur l’origine par altération du substratum schisteux de la miné-
spatiale de la salinisation des ressources en eau et modèle ralisation chlorurée des eaux d’une nappe côtière sous climat
conceptuel des sources salées dans la plaine des Triffa (Maroc semi-aride (Chtouka-Massa, Maroc), C. R. Acad. Sci. Paris, Ser.
nord-oriental), C. R. Geoscience 338 (2006) 768–774. IIa 336 (2004) 1363–1369.
[7] A. Chakroun, D. Zaghbib-Turki, A.M. Moigne, H. De Lumley, [25] L. Lebbe, K. Walraevens, P. Van Burm, W. De Breuck, L’évolu-
Découverte d’une faune de mammifère du Pléisticène supérieur tion de la distribution des eaux douces et salées dans la nappe
dans la grotte d’El Geffel (Cap-Bon, Tunisie), C. R. Palevol 4 libre de la plaine maritime aux environs de la frontière belgo-
(2005) 317–325. française, Ann. Soc. Géol. Nord, CIX (1989) 55–65.
[8] G. De Moor, W. De Breuck, De freatische waters in het Ooste- [26] F.S. Martos, P.A. Bosch, L.M. Sanchez, A. Angela Vallejos-
lijke Kustgebied en in de Vlaamse Vallei, Natuurwet. Tijdsch. 51 Izquierdo, Identification of the origin of salinization in ground-
(1969) 3–68. water using minor ions Lower Andarax, Southeast Spain, Sci.
[9] Z. Demirel, The history and evaluation of saltwater intrusion into Tot. Environ. 297 (2001) 43–58.
a coastal aquifer in Mersin, Turkey, J. Environ. Manag. 70 (2004) [27] S.B. Olobaniyi, F.B. Owoyemi, Characterization by factor ana-
275–282. lysis of the chemical facies of groundwater in the deltaic plain
[10] Direction Générale des Ressources en Eau, Annuaire d’exploita- sands aquifer of Warri, western Niger Delta, Nigeria, AJST Sci.
tion des nappes phréatiques, 1995. Eng. Ser. 7 (2006) 73–81.
[11] F. Damak Derbel, D. Zaghbib Turki, C. Yaich, Le Pliocène marin [28] A. Ozer, R. Pskoff, P. Sanlaville, A. Ulzega, Essai de corrélation
du Cap-Bon (Tunisie) : exemple de dépôts gravitaires, Géol. du Pléistocène supérieur de la Sardaigne et de la Tunisie, C. R.
Méditerranéenne tome XVIII (1991) 189–198. Acad. Sci. Paris, Ser. D 291 (1980) 801–804.
[12] A. El Achheb, J. Mania, J. Mudry, Mécanismes d’acquisition de [29] J.G. Paine, Determining salinization extent. Identifying salinity
la minéralisation des eaux souterraines dans le bassin Sahel- sources and estimating chloride mass using surface borehole and
Doukkala (Maroc occidental). Approche par des traceurs hydro- airborne electromagnetic induction methods, Water Resour. Res.
géochimiques, IGME, Madrid, ISBN, 2003, 84-7840-. 470-8. 3 (2003), p. 3-1-3-10.
[13] B. El Mansouri, Y. Loukili, D. Esselaoui, Mise en évidence et [30] C. Paniconi, I. Khlaifi, L. Giuditta, A. Giacomelli, J. Tarhouni,
étude du phénomène de l’upconing dans la nappe côtière du Modeling and analysis of Seawater Intrusion in the coastal
Rharb (NW du Maroc), IGME, Madrid, ISBN, 2003, 84-7840- aquifer of eastern Cap-Bon, Tunisia, Kluwer Academic Publis-
470-8. hers, Netherlands, Transport in Porous Media 43, 2001, pp. 3–28.
[14] M. Ennabli, Étude hydrogéologique des aquifères du Nord-Est [31] D.L. Parkhurst, C.A.J. Appelo, User’s guide to PHREEQC
de la Tunisie pour une gestion intégrée des ressources en eau, (version 2)-A computer program for speciation, batch-reaction,
Thèse Univ. Nice, 1980. one-dimensional transport, and inverse geochemical calcula-
[15] E. Farber, A. Vengosh, I. Gavrieli, A. Marie, T.D. Bullen, B. tions, U.S. Geological Survey Water-Resources Investigations
Mayer, R. Holtzman, M. Segal, U. Shavit, The origin and Report 99-4259, 1999, 312 p.
mechanisms of salinization of the Lower Jordan River, Geochim. [32] C.P. Petelas, I.B. Diamantis, Origin and distribution of saline
Cosmochim. Acta 68 (2004) 1989–2006. groundwater in the Upper Miocene aquifer system, coastal
[16] M.D. Fidelibus, Environmental tracing in coastal aquifers: old Rhodope area, northern Greece, J. Hydrogeol. 7 (1999) 305–316.
problems and new solutions, in: Coastal Aquifers Intrusion [33] P. Pulido-Lebœuf, Seawater intrusion and associated processes
Technology: Mediterranean Countries, vol. II, Publ. IGME, in a small coastal complex aquifer (Castell de Ferro, Spain),
Madrid, 2003, pp. 79–111. Applied Geochemistry 19 (2004) 1517–1527.
[17] R.A. Freeze, J.A. Cherry, Groundwater, Prentice-Hall, Engle- [34] P. Pulido-Leboeuf, A. Pulido-Bosch, M.L. Calvache, A. Vallejos,
wood Cliffs, NJ, 1979. J.M. Andreu, Strontium, SO42/Cl and Mg2+/Ca2+ ratios as
[18] N. Gaaloul, H. Alexander, D. Cheng, Hydrogeological and tracers for the evolution of seawater into coastal aquifers: the
hydrochemical investigation of Coastal Aquifers in Tunisia- example of Castell de Ferro aquifer (SE Spain), C. R. Geoscience
Crisis in overexploitation and salinisation, Second International 335 (2003) 1039–1048.
Conference on Saltwater Intrusion and Coastal Aquifers-Moni- [35] E. Sanz Escudé, Brackish springs in coastal aquifers and the role
toring, Modeling and Management, Merida, Mexico, March 30- of calcite dissolution by mixing waters: PhD Thesis, Department
April 2, 2003. of Geotechnical Engineering and Geo-Sciences (ETCG), Tech-
[19] K. Gemail, A. Samir, C. Oelsner, S.E. Mousa, S. Ibrahim, Study nical University of Catalonia (UPC), Institute of Earth Sciences
of saltwater intrusion using 1D, 2D and 3D resistivity surveys in ‘‘Jaume Almera’’ CSIC, 2007.
the coastal depressions at the eastern part of Matruh area, Egypt, [36] R.M. Spechler, Saltwater intrusion and quality of water in
EAGE 2 (2004) 103–109. Floridian aquifer system, Northeastern Florida: U.S. Geological
 L. Kouzana et al. / C. R. Geoscience 341 (2009) 21–35 35

Survey Water-Resources Investigations, Report 92-4174, 1994, [40] W. M. White Geochemistry, International Mine Water Associa-
76 p. tion, 2005, 700 p.
[37] J.H. Tellam, Hydrochemistry of the saline groundwaters of the [41] S.R. Wilson, M. Ingham, J.A. McConchie, The applicability of
lower Mersey Basin Permo-Triassic sandstone aquifer, UK, J. earth resistively methods for saline interface definition, J.
Hydrol. 165 (1995) 45–84. Hydrol. 316 (2006) 301–312.
[38] R. Trabelsi, M. Zaïri, H. Smida, H. Ben Dhia, Salinisation des [42] L. Zouhri, E. Carlier, B. Ben Kabbour, E.A. Toto, C. Gorini, B.
nappes côtières : cas de la nappe nord du Sahel de Sfax, Tunisie, Louche, Groundwater interaction in the coastal environment:
C. R. Geoscience 337 (2005) 515–524. hydrochemical, electrical and seismic approaches, Bull. Eng.
[39] A. Vengosh, E. Rosenthal, Saline groundwater in Israel: its Geol. Environ. 67 (2008) 123–128. , doi:10.1007/s10064-007-
bearing on the water crisis in the country, J. Hydrol. AMST 0101-6.
156 (1994) 389–430.