ORIGINAL PAPER

Comprehension and hydrogeological conceptualizationof aquifer in arid and semi-arid regions using integratedhydrogeological information system: case of the deepaquifer of Zéramdine-Béni Hassen (east-central Tunisia)

Fethi Lachaal & Ammar Mlayah & Makram Anane &

Mourad Bédir & Jamila Tarhouni & Christian Leduc

Received: 17 October 2011 /Accepted: 25 November 2011 /Published online: 5 January 2012# Saudi Society for Geosciences 2011

Abstract In arid and semi-arid regions, the groundwateroverexploitation caused drawdown in piezometric levelsand a degradation of chemical water quality. That is whythe groundwater monitoring needs a good comprehension ofthe hydrogeological aquifer properties. This is specially thecase of Zéramdine–Béni Hassen deep aquifer (east-centralTunisia). Seismic profiles interpretation highlights the exis-tence of the Zéramdine fault corridor, the Boumerdès anticline,the Moknine and Mahdia grabens that represent lateral bound-aries for the study aquifer. The outcrop of the aquifer is locatedin the Zéramdine, Béni Hassen and Ain Ben Jannet regions,where two lithostratigraphic sections were realized. The pie-zometric study shows that the principal groundwater flow isfrom west to east. A secondary flow is from NW to SE. Thehydrochemical study of 22 sample shows that the aquifer ischaracterized by freshwater, dominated by Na–Ca–Cl–SO4facies. The salinity increase is from the west to the east, whichcoincides with the principal water flow direction. The integra-tion of all results deduced from the hydrogeophysic, hydrody-namic and hydrochemical studies is developed to investigatehydrological processes of Zéramdine–Béni Hassen aquifer and

consequently to propose a conceptual model, which will helpto propose a rescue plan for the studied aquifer and to imple-ment a spatial hydrogeological database using the globalinformation system and then to characterize the complexaquifer system.

Keywords Groundwater . Conceptual model .

Geodatabase . Hydrogeological information systems .

Tunisia

Introduction

In the case of the complex aquifer and hydrologic data lack,the use of one investigation technique is not sufficient tounderstand and identify the aquifer system for its sustainablemanagement. This is the case of the Miocene aquifer systemin the Sahel of Tunisia (east-central Tunisia). In fact, previousworks were done to characterize the hydrogeological structureof this system (Lachaal et al. 2011a; Lachaal 2011). Theseworks show that all possible investigation methods and avail-able hydrodynamic and geochemical data must be integratedin order to clarify a schema of the aquifer’s distribution. Amethodology for characterizing a complex aquifer systemwasdeveloped and applied to the Miocene aquifer systems in theSahel of Tunisia region (Lachaal et al. 2011a; Lachaal 2011).The methodology consists in compiling and combining allavailable data, especially the geophysical (seismic reflectionand wireline logging of drilling wells), the hydrodynamic andthe hydrochemical ones. This studies show that Mioceneseries in the Sahel region is compartmentalized into Jemmel,Zéramdine and Mahdia–Jébéniana blocks under the influenceof the tectonic structures of the Zéramdine fault corridor, the

F. Lachaal :A. Mlayah :M. Anane :M. BédirWater Research and Technology Centre, Borj Cedria Ecopark,PO Box 273, Soliman 8020, Tunisia

F. Lachaal (*) : J. TarhouniWater Sciences and Technique Laboratory,National Agronomic Institute of Tunisia,43 Avenue Charles Nicolle, Mahrajène,1082, Tunis, Tunisiae-mail: lachaalfethi@yahoo.fr

C. LeducIRD, UMR G-EAU,BP 5095, 34196, Montpellier Cedex 5, France

Arab J Geosci (2013) 6:2655–2671DOI 10.1007/s12517-011-0498-x

Mahdia and Moknine grabens and the El Jem half-graben.These structures controlled the Miocene fluvio-deltaic sedimen-tation and led to the spatial distribution of Miocene reservoirsystems. In addition to the upper Miocene sandy level reservoir,other reservoir horizons were identified, and correlated, and anew distribution of the Miocene aquifers are presented (Lachaal2011). The Zéramdine–Béni Hassen Miocene aquifer (ZBH) isone of these aquifers, which needs a new hydrogeology charac-terisation and conceptualisation using the new structural results.

In addition, over the past 30 years, socioeconomic develop-ment within the region has been largely due to intensive use ofZBH Miocene aquifer resources. The groundwater extractionthat increased from 1.25 in 1973 (Beni-Akhy 1998) to3.12 Mm3 year−1 in 2005 (DGRE 2005) has provoked thedecrease in groundwater levels, reduction of water resourcesand degradation of its chemical quality. To overcome theseproblems, we propose a sustainable management program. It isimportant to understand its structure, geometry and operation,and to propose a hydrogeological conceptual model of thewater in question (Jagelke and Barthel 2005; Nastev et al.2008; Abdalla 2009; Gedeon et al. 2011; Osman and Abdalla2011).

On the other hand, understanding and managing ground-water resources require the integration of a large amount ofhigh-quality data from a variety of sources (Baalousha 2011;Chesnaux et al. 2011). Because of the difficulty to access tothe data related to groundwater and subsurface conditions,gathering available information is crucial when conducting asuccessful hydrogeological study. Global information system(GIS) is one of the most important tools for integrating andanalyzing spatial information from different sources or disci-plines (Ghayoumian et al. 2007; Tweed et al. 2006). It helps tointegrate, analyze and represent spatial information and data-base of any resource, which could be easily used for planningof resource development, environmental protection and scien-tific researches and investigations, although its importance iswidely known and implemented in many countries in theworld. Since the 1990s, several methodologies of groundwatermanagement and vulnerability assessment have been devel-oped using GIS and database environments (Mende andAstorga 2007; VanderPost and McFarlane 2007; Chesnauxet al. 2011).

The objective of this study is to build such a hydrogeologicalconceptual model to the ZBH Miocene deep aquifer, based oncompilation of all available geological, geophysical, piezomet-ric and chemical data arranged in integrated hydrogeologicalinformation system using GIS tools.

Study area characterization

The study area covers the Jemmel, Zéramdine and BéniHassen plains with a total surface area of 2,610 km2. It is

located between 3,902,500 and 3,956,000 north parallelsand the 640,000 and 700,000 east meridians (Fig. 1). Thestudy area is bordered to the West by the Zéramdine faultcorridor, to the South by the Mahdia–Jébéniana plain, to theNorth by the Moknine fault and platform and to the East bythe Mediterranean Sea (Fig. 1).

The study catchment is located in the Sahel of Tunisia,which is positioned in the northern part of the MesozoicSahel bloc (east-central Tunisia). This region is character-ised by a strong subsidence compared with central Tunisiain the Oligocene period. The Sahel bloc has been affected bya variety of E–Wand N–S trending faults and grabens, suchas the Mahdia and Moknine grabens, and the Zéramdinefault corridor (Fig. 2). The extensional tectonic regime in theSerravallian and Tortonian controlled the deposit formationof a fluvio-deltaic series consisting of silicoclastic sediment(Bédir et al. 1992, 1996). In the extensional tectonic regime,the Zéramdine fault corridor was active during the Miocene.It forms the limit between two basins: In the north is theJemmel bloc, and in the south is the Zéramdine bloc.

Geological outcrops in the studied area are mainly Plioceneand Quaternary deposits. The oldest sediment outcropping inthe Zéramdine fold is the Upper Miocene Oum DouilFormation (Burollet 1956; Gaaloul 1995; Mannaï-Tayech2009). It was formed by alternation of clay and sandy clay,and it is characterized by abundant lignite. The thickness andthe lithology of these sediments are variable. According to thepetroleum wells, the Miocene is made of:

– The Lower Miocene (200 m) mainly composed of clays,sandy clays, sands and silts; characterized by pyrite andlignite traces;

– The Middle Miocene: composed of two Langhian forma-tions: (1) the transgressive Ain Grab Formation consistsof 60 m (in the Jem-1 well) of an interbedded limestone–claystone sequence; and (2) the Mahmoud Formationwith planktonic clays and an average thickness of 50 m.It represents the substratum of the Miocene reservoirsystems;

– The Upper Miocene: represents an important thickness(800 m) of alternating sandstone with pyrite and lignitetraces, bioclastic sands, limestone, and sandy clays withbioclastic sands. This unit outcrops in the Zéramdineanticline with an average thickness of 75 m.

The Zéramdine and Béni Hassen sector (east centralTunisia) has a semi-arid to arid climate. The mean annualprecipitation for the period of 1980–2007 is 360 mm year−1,with 90% of rainfall occurring from March to September.The monthly temperature averages are 12.7°C and 27.5°C inJuly and August (Lachaal 2011). Water resources are limited,andmost of the water demand is supplied by both deep (UpperMiocene) and shallow (Pliocene and Quaternary) aquifers(Bouri and Ben Dhia 2010).

2656 Arab J Geosci (2013) 6:2655–2671

Materials and methods

In this paragraph, we describe a conceptual model for acomprehensive hydrogeological spatial database able tocharacterize the aquifers systems, and we present the mainoutputs of a hydrogeological spatial database developed forthe ZBH Miocene deep aquifer.

The spatial database structure and implementation

The development of a spatial database, whatever the domainof application, permits to gather all required data in onestructure making access easier and offering the possibilityto update it when needed (Ghayoumian et al. 2007; Tweedet al. 2006). In the case of hydrogeological field, this

Fig. 2 A–A’ seismic profile showing the influence of the Moknine and Mahdia grabens in the compartementation of the Miocene reservoirs(Lachaal et al. 2011a)

Fig. 1 Location map of the study area (a) and structural map (b) of the Jemmel–Zéramdine sector

Arab J Geosci (2013) 6:2655–2671 2657

database renders easier the aquifer characterization and thenpermits the identification of groundwater potential zones(Mende and Astorga 2007; VanderPost and McFarlane2007). However, to develop such a database, it is mandatorybeforehand to conceptualize a model that explains the kindand the structure of the data that should be included in thebase. In this work, we propose the conceptual model ofaquifers systems that is sketched in Fig. 3. This sketchexplains the input data and the process applied to get fromit the final data needed for modelling.

To facilitate and simplify the interpretation and the com-pilation of hydrogeological results deduced from differentinvestigation techniques, we propose to build a conceptualdatabase based in the GIS tools. Different spatial and alpha-numeric data were integrated to determine the groundwaterresource condition of the Zéramdine and Béni Hassen catch-ment. The integrated spatial data helped to obtain informationregarding the geological environment, the groundwater poten-tial, the aquifer system characteristics and abstraction ratesand the water quality conditions of the study area.

Used data collection and processing

Different types of data from different sources and indifferent supports were collected and integrated in the

spatial database (Fig. 3). Description of these data is givenhereafter.

Seventy-three well sheets have been delivered by theGeneral Direction of Water Resources, Tunisian Ministryof Agriculture (DGRE). Geographical location of each wellis identified and introduced with the descriptive data in aspatial layer. Outcrop data are obtained from a scannedgeological map (1:50,000). This map has been georeferencedand digitalised, and a vector map was obtained. Twenty-fourprofiles have been provided by the Tunisian Company ofPetroleum Activities (ETAP) for petroleum exploration in1971, 1984 and 1993 (Fig. 1). In this study, only the Plioceneand Miocene units are studied. Several reflectors of base andtop of the sandy layers were identified in the seismic sections,and two-way travel time structural maps were drawn for thedifferent layers using GeoFrame Charisma software (v3.8,Schlumberger). The seismic horizons were calibrated usingthe time–depth conversion curve of the petroleum wells andthe outcrops in the region. The obtained spatial data wereintroduced in the spatial database. The piezometric data usedin this study were provided by the DGRE. It consists ofmonthly water level measurements in 17 wells between1995 and 2008. The topography was obtained from a topo-graphic map (1:50,000) from which the level curves weredigitalised and the vector layer obtained and introduced in

Dat

a C

olle

ctio

nD

atab

ase

Co

nst

ruct

ion

Seismic-reflexionprofiles

Topography Outcrop area (lithology)

Digitizingcontours

TINmap

DEM

Lithologymap

Data classification

Data modelling

Attribute data

editing

Aquifer structure map

Rainfall

Rainfallmap

Sp

atia

l Dat

aA

nal

ysis

Dat

aM

od

elin

g

Attribute data

editing

Reservoir depth

Boreholes

Attribute data

editing

Petroleum wells

Recharge map

Structure Depth reservoir characterization

Top and button aquifer reflectors

Seismic shot point

Two-way traveltimestructural maps (TWT) of the reservoir layers

Seismic speed map

Groundwater geometric map

(depth )

Attribute data

editing

Piezometer

Piezometrichead

Attribute data

editing

Goundwaterwells simples

Hydrochimicalmap

Attribute data

editing

Digitizingcontours

Attribute data

editing

Generated layers for groundwater system modelling and monitoring

Fig. 3 Logical model of the applied GIS database

2658 Arab J Geosci (2013) 6:2655–2671

the database. Chemical data were obtained from the analysis ofwater samples pumped from 22wells in July 2008; temperature(T), electrical conductivity (EC) and pH were measured in situ;SO4

2- concentration was determined using the gravimetricmethod; Cl- was analyzed by titration (Mohr method), andHCO3

- and CO32- by titration with sulphuric acid. Ca2+, Na+,

Mg2+ and K+ were analyzed by atomic absorption spectrome-ter. Data from previous sampling campaigns were also includedin the analysis.

Outputs from the hydrogeological database

Several kinds of information could be derived from thedatabase from which we can indicate two-way travel timestructural maps of the different reservoir layers of the ZBHaquifer (Figs. 4 and 5), structural and piezometric headmaps, hydrogeological 3D-model and chemical water qual-ity maps such as spatial distribution of the salinity.

The construction of 3Dmodels should help to better predictthe extension and thickness of aquifer reservoirs and welllocations. The geodatabase will also serve as an importantinformation source for groundwater conceptual model.

All required GIS data for conceptual modelling is con-verted to the appropriate formats for numerical modellingusing the ArcGIS 3.2 software package (ESRI Inc.). Thesubsurface catchment, the geometric boundary of the modeledaquifer, the aquifer base and top, the recharge zone (Mioceneoutcrop), the piezometer and the pumping wells positions aresaved as shape files.

Hydrogeologic setting and aquifer structure

The groundwater resources in the Zéramdine and Béni Hassenregions are formed by two hydrogeological systems: shallowand deep groundwater. The shallow system is composed bythe ZBH and Jemmel-Bembla (JB) aquifers, with depths lessthan 50 m. They occur within the sandy and sandy-claydeposits of the Plio-Quaternary (Bouri and Ben Dhia 2010).The deep aquifer system is composed of the ZBH and JBMiocene aquifers (Hubert 1968; Beni-Akhy 1998; Lachaal etal. 2011b), which are composed of clay and sandy clay seriesthat was deposited in the Serravallian and Tortonian period(Gaaloul 1995; Mannaï-Tayech 2009).

Both the water table and water flow in the Mioceneaquifers are influenced by the tectonic structures. Mahdiagraben acts as a hydraulic screen between two hydraulicsystems: the ZBH and Jébéniana-Ksour Essef deep aquifers,the latter being the continuity of the Sfax deep aquifersystem. Water flow in each block (north and south) is totallyindependent of the other (Lachaal et al. 2011a).

The ZBH aquifer is one of the most important aquiferwithin Sahel of Tunisia. A lot of research works has shown

the hydrogeology and hydrodynamics complexity of thisaquifer (Hubert 1968; Beni-Akhy 1998; Lachaal 2011) dueto the strong tectonic activity in the region (Bédir 1989; Bédiret al. 1992, 1996). The ZBH deep aquifer is unconfined in theZéramdine area and confined close to the Mediterranean Sea(Hubert 1968; Beni-Akhy 1998).

From the 1970s, the overexploitation caused drop ingroundwater levels and a deteriorating chemical water qualitymarked by saltwater intrusions of Mediterranean Sea andMoknine salt lake (Bouri and Ben Dhia 2010). That way thedeep aquifers resources are the most adopted solutions and themajor source of water supply for domestic, agricultural andindustrial purposes.

The deep reservoir in this sector is formed by the Zéramdine–Béni Hassen Miocene aquifer (ZBH), which yielded3.93 Mm3 in 2002 (DGRE 2002). It is located in thefluvio-deltaic deposits, composed mainly by a sandy-clayseries of the Serravalian and Tortonian (Beni-Akhy 1998;Lachaal 2011; Lachaal et al. 2011a).

Results and discussion

Aquifer lithology and boundaries

The aquifer lithology characterization is based on the use oftwo investigation techniques: The lithostratigraphic cross-sections that were established in the Miocene outcrop locat-ed in the ZBH aquifer upstream and the interpretation ofsubsurface data (seismic reflexion profiles and petroleumand hydraulic wireline logging) for the central and down-stream parties.

In the Miocene outcrop, we established two lithostrati-graphic cross sections: Gamgouma and Ain Ben Jannetsections (Figs. 1 and 6).

1. The first section is located in the Zéramdine region. It iscomposed by three levels, from bottom to top:

– A sandy level (5 m) of very fine sand with clayintercalations.

– A clay level (7 m), formed by two bundles ofblackish clay, intercalated with sandy clay. Thesummit is a two centimetre sandstone layer. Thefirst is green, rich in jarosite while the second isgrayish. These two layers are sandwiched by alignite layer.

– A sandy level (21m), formed by two sandbanks inter-spersedwith a clay layer. These sands are yellow, veryfine, with oblique stratification.

2. The Ain Ben Jannet section is located in the Jemmelregion. It shows a sandy series, dated from Serravallianto Tortonian (Gaaloul 1995). It begins with a centimeter

Arab J Geosci (2013) 6:2655–2671 2659

bench of fine sand, whitish, topped with thin clay, andsand intercalations. The series continues with fine sand,white, well sorted, with oblique stratification. The thick-ness of this series is about 23 m.

The lithostratigraphic cross-sections show that the ZBHdeep aquifer is a multi-layer aquifer formed by sand bedswith thin clay intercalations. Laterally, thicknesses anddepths of these layers present strange variability that is

Fig. 4 Two-way travel time structural maps (TWT) of the different reservoir layers in the Zéramdine block (Lachaal et al. 2011a)

2660 Arab J Geosci (2013) 6:2655–2671

Fig. 5 a Two-way travel time structural top map (TWT) of the third reservoir layer and b Two-way traveltime structural base map (TWT) of theforth reservoir layer in the Zéramdine block (Lachaal et al. 2011a)

Arab J Geosci (2013) 6:2655–2671 2661

explained by the fluvio-deltaic nature of the Miocene sedi-ments and their rapid facies variations (Lachaal et al. 2011a).

The Zéramdine-1 petroleumwell presents the section of thefour aquifer layers. From bottom to top, the four sand reservoirlayers are:

– The first reservoir layer (730–820 m) is formed by sandwith clayey intercalations.

– The second (235–340 m) consists of four sandy levels,with clayey intercalations.

– The third (115–130 m) is a sandy layer.

– The fourth reservoir layer forms the ZBH upper aquiferwhich outcrops in the Zéramdine, Béni Hassen and AinBen Jennet regions, whereas, in the Sidi Bannour area,it is buried to a depth of 250 m.

The aquifer boundaries were determined using the tec-tonic map (Fig. 1) and the two-way travel time base map ofZBH and JB deep aquifer (Figs. 4 and 5), wireline loggingof petroleum and water wells, the geology cross-section, thehydrogeological correlations of water wells (Figs. 7 and 8),and the piezometric map.

Fig. 6 a Oued Gamgouma and b Oued Ain Ben Jannet lithostratigraphic cross-sections (Lachaal et al. 2011a)

2662 Arab J Geosci (2013) 6:2655–2671

Northern limit The Moknine graben is considered as thenorthern border of Zéramdine platform (Beni-Akhy 1998;Lachaal et al. 2011a) (Fig. 2). It was identified by theelectrical discontinuity deduced from the electrical prospec-ting conducted by SEREPT and CGG in 1950 (Hubert 1968)and seismic interpretation (Bédir et al. 1992, 1996). Thisgraben is composed of some NW–SE and WNW–ESE faults.The Moknine normal fault (Fig. 1) connects the ZBH sandlevels with the most recent series (Plio-Quaternary). The wellNo. 1, located in the north of the Moknine fault, crosses1,000 m of clay and marls. These faults constitute awaterproof screen delineating the northern boundary ofthe aquifer.

Western limit The Zéramdine fold axis represents the westernboundary of the aquifer (Beni-Akhy 1998; Lachaal et al.2011b). This fold axis coincides with the eastern boundaryof the Zéramdine fault corridor previously shown in thetectonic map (Fig. 1) and the B-B’ hydrogeological cross-section (Fig. 7). This corridor creates connections betweenMiocene and Plio-Quaternary sediments. Therefore, it separatestwo different hydraulic systems: the JB and ZBHMiocene deepaquifers.

Southern limit The Mahdia graben affects the Miocene se-ries in the region and subdivides the study area into threeblocks: the Zéramdine block to the North and the Mahdia–Jébéniana to the South and the centre graben of Mahdia

(Fig. 2). The Mahdia faults range along a few tens of kilo-metres from the Mediterranean Sea to Boumerdès region(Lachaal et al. 2011a).

Two wells were drilled in the central part of the graben(Fig. 1): the first, Amiret el Fhoul (no. 2), with a depth of460 m; and the second, Henchir Ben Zineb (No. 3) with adepth of 102 m. Both wells indicate a clay sedimentation,which forms the collapsed part of the graben. Only a superfi-cial reduced level is formed by sandy clay at a depth between33 and 58 m with a very high salinity of 12.4 and 13.4 g l−1,respectively, measured at well nos. 2 and 3 (Lachaal et al.2011a). This superficial reduced level forms the Ksour Essefshallow aquifer. That way, the Mahdia fault is consider assouthern limit of the ZBH deep aquifer.

Eastern limit Two-way travel time structural maps (TWT)of the base of ZBH and JB deep aquifers (the Upper Mio-cene layer) done by Lachaal (2011b) shows an intensefracturing separates these areas into high and collapsedsurfaces. There is a high surface found in the Zéramdine,Béni Hassen and Ain Ben Jennet areas, where the ZBHaquifer is shallow. To the east, it becomes deeper (Fig. 8).Strong collapses can be observed along the Mahdia fault. Infact, near the Mediterranean Sea, the series are very deep(more than 1,250 ms (TWT) to the upper reservoir level(Fig. 4 and 5)) indicating that the Mediterranean Sea is thenatural discharges of the ZBH aquifer. This is the exsur-gence aquifer and forms an imposed head boundary. The

Fig. 7 B–B’ Hydrogeological cross section showing the influence of Zéramdine fault corridor in the differentiation of the Zéramdine–Béni Hassenand the Jemmel–Bembla Miocene deep aquifer

Arab J Geosci (2013) 6:2655–2671 2663

lack of productive wells to the East of Sidi Bannour isexplained by the great depth of the aquifer (more than 1 sin time double) and not by its absence.

Aquifer geometry

The aquifer system geometry was determined using thecombining results obtained from various data specially,geophysical (seismic reflexion profiles, oil and water welllogging: gamma gamma, spontaneous potential and electri-cal resistivity), hydrogeological cross-sections and structuralmap of the region.

The depth of the top of the upper aquifer layer is variableand depends on the tectonic architecture of the region. Itvaries from 150 to −300 m to sea level (sl). Figure 9 showsthe presence of high zones in the regions of Béni Hassen andZéramdine and other collapsed along the fault Mahdia. Thedepth of the base of the upper aquifer layer varies between50 and −350 m (sl) (Fig. 10). The thickness of the firstaquifer layer varies between 50 and 100 m. In Zéramdine,Béni Hassen and Ain Ben Jannet, the first reservoir layersoutcrop in the surface and show a thickness of 50 m, is therecharge area. However, in the region of Sidi Bannour, theaquifer thickness increases and reaches 100 m.

The depth of the top and base of the other tree aquiferlayers are deduced from the two-way travel time (TWT)maps of the different reservoir layers published by Lachaal(2011a) by conversion of the TWT in milliseconds to depthin metres, using the interval seismic speed. The TWT mapschow an intense fracturing separates these areas into highand collapsed surfaces. There is a high surface found in theZéramdine, Béni Hassen and Ain Ben Jennet areas, wherethe ZBH aquifer is shallow.

Hydrodynamics of the Miocene aquifer systemof the Zéramdine Béni Hassen

The piezometric map and water flow of the ZBH aquifer

The piezometric head contour map in the ZBH deep aquiferis illustrated in Fig. 11. Groundwater head ranges between 0and 85 m. Water flow follows two directions. The major oneis from west to east, with a hydraulic gradient varyingbetween 15.5‰ and 1.5‰ in the upstream and downstreamparts of the aquifer, respectively. The second water flow direc-tion is from northwest to southeast, from the Ain Ben Jennetregion to the Boumerdès region. The hydraulic gradient isabout 0.8‰. Variations in the hydraulic gradient are relatedto the spatial variation of the reservoir geometry. The rechargearea of the ZBH deep aquifer is situated in the Zéramdine, BéniHassen and Ain Ben Jennet regions (Hubert 1968; Beni-Akhy1998). The discharge fields are situated in the seashore.

The piezometric evolution and depression of the ZBHMiocene aquifer

The study of piezometry over the 1995–2008 period revealedthree piezometric types (Fig. 12):

– The seasonal piezometric fluctuation is very significant inthe ZBH upstream that represents the recharge area. Itexceed 1 m (in the 19635/4 piezometer (P. 1)) (Fig. 12a).The higher piezometric head was observed during thehumid season (September to February), when 83% of theannual precipitation was registered. The lower piezometricheadwas registered during the dry season (April toAugust),when the rainfall did not exceed 10% of the annual

Fig. 8 C–C’ hydrogeological cross-section showing the W–E evolution of the upper layer of the Zéramdine–Béni Hassen deep aquifer

2664 Arab J Geosci (2013) 6:2655–2671

precipitation. The seasonal piezometric fluctuation isexplained by the direct effect of rainwater infiltration inthe recharge area.

– An occasional seasonal piezometric fluctuation: ob-served in the 19573/4 piezometer (P. 2) (Fig. 12b),which is near the recharge zone, an occasional seasonalfluctuation of 0.5 m was observed during 1997 and1998. This third piezometric category is located in therecharge area, where the infiltration was locally excep-tional. That reflects a strong inter-annual variability ofrecharge.

– A continuous piezometric depression was identified in allthe piezometres especially in 19832/4 piezometer (P. 3)(Fig. 12c). The piezometric decrease varied from0.37 m year−1 (19833 piezometer (P. 4)) to 1.97 m year−1

(19826/4 piezometer (P. 5)). The high piezometric depres-sion is observed in the middle and downstream of the ZBHaquifer. However, it is lower in the Miocene outcrops, dueto the recharge influence. In the Chiba region (10654bis/4

piezometer (P. 5)), the piezometric head has decreased by18.2 m during the last 43 years because of intensivepumping and a limited recharge.

Precipitation and groundwater extraction

The intensive groundwater extraction is localised mainly incentral and downstream for agriculture (70%) and drinkingwater (30%). It increased from 0.97 in 1985 to 4 Mm3 year−1

in 2002. The well number has increased from four in 1973 to26 in 2007 (Fig. 13). The groundwater extraction evolutioncan be divided into three periods (Fig. 14):

– First period: until 1989, when aquifer extraction has notexceeded 2 Mm3 year−1. Along the 1973–1989 period,the water extraction average was 1.48 Mm3 year−1. Thepumping wells were limited to ten with an average of0.175 Mm3 year−1 well−1.

Fig. 10 Isodepth base map of the upper layer of Zéramdine–Béni Hassen deep aquifer

Fig. 9 Isodepth top map of the upper layer of Zéramdine–Béni Hassen Miocene deep aquifer

Arab J Geosci (2013) 6:2655–2671 2665

19573/4 piezometer (P.2)

38

39

40

41

42

janv-

97

janv-

99

janv-

01

janv-

03

janv-

05

janv-

07

Time year

Pie

zom

etri

c h

aed

(m

) Piezometric depression of 0.30 m.y-1

Occasional seasonal piezometric fluctuation

19635/4 piezometer (P.1)

65

66

67

68

69

janv-

98

janv-

00

janv-

02

janv-

04

Time (year)

Pie

zom

etri

c h

aed

(m

)

0

50

100

150

200

250

300

Rai

nfa

ll (m

m)

Rainfall (mm)

piezometric head (m)

19833/4 piezometer (P.4)

67

68

69

70

71

janv-

98

janv-

00

janv-

02

janv-

04

janv-

06

janv-

08

Time year

Pie

zom

eter

hea

d (

m)

Piezometric depression of 0.37 m.y-1

19832/4 piezometer (P.3)

27

29

31

33

35

37

janv-

99

janv-

01

janv-

03

janv-

05

janv-

07

Time year

Pie

zom

etri

c h

ead

(m

)

Piezometric depression of 0.91 m.y-1

(a) (b)

(c) (d)

Fig. 12 Piezometric fluctuations of ZBH Miocene deep aquifer: a Typical seasonal fluctuation and rainfall at Monastir. b Typical occasionalseasonal fluctuation. c, d Typical piezometric depression

Fig. 11 Piezometric head (September 2007) maps of the Zéramdine–Béni Hassen deep aquifer

2666 Arab J Geosci (2013) 6:2655–2671

– Second period (1990–1998): In 1990, the groundwaterextraction has exceeded, for the first time, 2 Mm3 year−1.During this period, the extraction average is about1.84 Mm3 year−1. The pumping wells increased fromten to 17 with an average of 0.14 Mm3 year−1 well−1.

– Third period: dates from 1999 to the present. Thisperiod is characterized by annual average harvest to3.13 Mm3 year−1. Since 2002, aquifer extractionincreased to over 3.5 Mm3 year−1, and pumping wellshave increased from 17 in 1999 to 26 in 2007.

Hydrochemistry

Hydrochemistry can also be an effective tool to validate andconfirm the aquifer hydrodynamic hypothesis, previouslyproved by geophysical and hydrodynamic methods (Lachaalet al. 2010; Mlayah et al. 2011; Ayman and Mohamed2011). The aims of this section are to chemically characterizethe ZBH aquifer and to study the spatial evolution of ground-water quality in the study area.

Groundwater mineralization

Total dissolved salt of the ZBH deep aquifer waters variesbetween 0.92 g l−1 at the Béni Othman well (no. 4) and4.45 g l−1 at Chiba-3bis (no. 5) (Fig. 15). In the unconfinedpart, water salinity is low and does not exceed 1.5 g l−1. Thiswater is mainly used for drinking and agricultural irrigation.Salinity values increase from west to east, corresponding to themain groundwater flow direction, and exceed 4 g l−1 in somelocations. This salinity increase could be explained by a longercontact-time with rocks during water circulation (Bouhlassaand Aicahi 2004; Lachaal et al. 2010; Mlayah et al. 2011).

Hydrochemical facies

Piper and Stiff diagrams were generated using Diagramsoftware (Simler 2000). Two different geochemical faciesappear in the Piper diagram (Fig. 16). The ZBH deep aquifersamples are fresh water with a great variability but mainly ofthe Na–Ca–Cl–SO4 type.

Fig. 14 Extraction evolution ofthe Zéramdine–Béni Hassendeep aquifer (1973–2008)

0

50

100

150

200

250

300

350

0

5

10

15

20

25

30

35

1970 1975 1980 1985 1990 1995 2000 2005 2010

Ext

ract

ion

ove

rag

e p

er w

ells

(1

03 m3 y

ear-1

wel

l-1 )

Ext

ract

ion

wel

ls n

um

ber

Year

Extraction wells number

Extraction overage per well

Fig. 13 Evolution of the pump-ing well and extraction averagein the Zéramdine–Béni HassenMiocene deep aquifer

Arab J Geosci (2013) 6:2655–2671 2667

Origin of groundwater mineralization

Binary diagrams are effective tools for studying the origin ofdissolved salts in water. Dominant cations are sodium, chlo-ride and potassium. The first ratio diagram is Na/Cl whichseems to be the most relevant. Figure 17a indicates the ratiodiagram of Na/Cl for the deep groundwater in the region, onwhich was superposed the halite dissolution line (NaCl). Thegood correlation between the plotted ratios with the halitedissolution line confirms that halite dissolution is the originof water salinity. Another ratio diagram used was Ca/SO4

(Fig. 17b). It yielded a good correlation between the pointsand the line of gypsum dissolution (CaSO4, 2H2O) indicatingthat gypsum dissolution is a second source of minerals inwater. Ca/HCO3 ratio points are not correlated with the linesof calcite (CaCO3) and dolomite (CaMgCO3) dissolution(Fig. 17c). The dolomitic limestone is not correlated with thecalcium and magnesium (ratio diagram of Ca+Mg/HCO3;Fig. 17d). This indicates that the dissolution of gypsum andhalite are the dominant processes controlling water salinity.These minerals are abundant in the fluvio-deltaic Mioceneseries (Gaaloul 1995; Mannaï-Tayech 2009).

Fig. 16 Piper diagram for watersamples of Zéramdine–BéniHassen deep aquifer (June 2008)

Fig. 15 Spatial distribution of the salinity in the Zéramdine–Béni Hassen deep aquifer (June 2008)

2668 Arab J Geosci (2013) 6:2655–2671

Conceptual hydrogeological model

The term conceptual hydrogeological model is usually takento be a subjective understanding of the processes controllinggroundwater flow at the site (IAEA 2001). Lateral andvertical aquifer extension and its conceptual structure,which contribute to determine boundary conditions, arederived from integrating oil and water well data and seismicprofile interpretation. Indeed, the aquifer is a multilayer

system, formed by sand beds with local clay intercalations.The Zéramdine faults corridor, the Moknine and Mahdiagrabens, form the aquifer boundaries from the west, northand south, respectively. To the east, water flows to theMediterranean Sea, on which is imposed a constant head(Fig. 18). The aquifer geometry is complex and is influencedby the tectonic structures. The aquifer is shallow in the up-stream. The recharge area is situated in the Miocene outcrops(the Zéramdine, Béni Hassen and Ain Ben Jennet regions)

Fig. 18 Conceptual 3D-model of the Zéramdine–Béni Hassen deep aquifer

0

10

20

30

40

50

0 10 20 30 40 50

Cl (Meq/l)

Na

(Meq

/l)

0

2

4

6

8

10

12

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9

HCO3 (meq/l)

Ca

(meq

/l)

0

5

10

15

20

25

0.0 0.2 0.4 0.6 0.8 1.0

HCO3 (meq/l)

Ca

+ M

g (m

eq/l)

0

5

10

15

0 5 10 15 20

SO4 (meq/l)

Ca

(meq

/l)

Zéramdien-Béni Hassen aquifer

Halite dissolution lineZéramdien-Béni Hassen aquifer

Gypsum dissolution line

(a) (b)

(c) (d)

Fig. 17 Ionic ratios of the Zéramdine–Béni Hassen deep aquifer (June 2008)

Arab J Geosci (2013) 6:2655–2671 2669

with an area of 50 km2 (Lachaal 2011). In the southern part,the aquifer communicates with the Mahdia-Ksour Essef aqui-fer. For the central and downstream regions, the aquifer iscaptive. It deepens to the east, exceeding 300 m in the Chibaarea. The discharge fields are situated in the seashore.

The hydrodynamics of the water is very complex, influ-enced by the aquifer geometry and the tectonic structures.The groundwater flow converging from Miocene outcropsin two directions. The main direction is from west to east(from Zéramdine and Béni-Hassen to Chiba and Sidi Bannourareas). The secondary flow is from NW to SE. The ground-water hydrodynamics is characterized by three piezometricbehaviors: a seasonal fluctuation, an occasional seasonal fluc-tuations and a major depression. The groundwater extractionis intense in the central and downstream parts, and it increasesduring the last years.

The ZBH deep aquifer consists of fresh water with asalinity ranging from 0.92 to 4.45 g l−1. In the unconfinedpart, water salinity is low and does not exceed 1.5 g l−1.Salinity values increase from west to east, corresponding tothe main groundwater flow direction. The water facies is Na–Ca–Cl–SO4 type. The gypsum and halite dissolutions are thedominant processes controlling the water mineralization(Lachaal et al. 2010, 2011a).

Conclusions and perspectives

In this paper, we developed amethodology for hydrogeologicalcomprehension and conceptualisation of aquifer in arid andsemi-arid regions using integrated hydrogeological informationsystem. This method is applied to the ZBH deep aquifer. Thegeophysics techniques, especially seismic reflection and wire-line logging methods are used for better defining aquifergeometries (clarify the geological structure and their lateraland vertical extension). Hydrogeological characterization iscompleted with piezometric and geochemical studies. Hydro-dynamic characterization is used to follow the spatial andtemporal piezometric behaviour and also to determine ground-water flow. Groundwater chemistry may trace the water originand water–rock interactions. Therefore, the GIS-assisted data-base system helped in groundwater characterization accessingeasily to different types of hydrogeological data. It integratesand analyzes spatial information and manages groundwaterresources layers.

The produced groundwater related database and concep-tual hydrogeological model is on demand of all interestedinstitutions, researchers, groundwater practitioners, drillingcompanies and decision makers, etc. It offers an accessibleand easy-to-get data on the aquifer gathered in the samesupport. The construction of 3D conceptual models shouldhelp to better predict the extension and thickness of ground-water reservoirs, i.e., groundwater resources availability. The

geodatabase will also provide an orientation for future man-agement plans of groundwater resources. Indeed, it wouldserve as an important information source for future updatingof the Sahel groundwater simulation numerical model.

References

Abdalla O (2009) Groundwater modeling in semi-arid Central Sudan:adequacy and long term abstraction. Arab J Geosci 2(4):321–335.doi:10.1007/s12517-009-0042-4

Ayman AA, Mohamed HA (2011) Hydrochemical evolution and varia-tion of groundwater and its environmental impact at Sohag, Egypt.Arab J Geosci 4:339–352. doi:10.1007/s12517-009-0055-z

BaaloushaHM (2011)Mapping groundwater contamination risk usingGISand groundwater modelling. A case study from the Gaza Strip,Palestine. Arab J Geosci 4:483–494. doi:10.1007/s12517-010-0135-0

Bédir M (1989) Contrôle tectonique et images sismiques de méga-structures synsédimentaires miocènes dans le Sahel de Mahdia–Tunisie orientale. J Afr Earth Sci 9(3/4):657–663

Bédir M, Zargouni F, Tlig S, Bobier C (1992) Subsurface geodynamicsand petroleum geology of transform margin basins in the Sahel ofMahdia and El Jem (eastern Tunisia). Am Assoc Pet Geol Bull 76(9):1417–1442

Bédir M, Tlig S, Bobier C, Zargouni F, Aïssaoui N (1996) Sequencestratigraphy, basin dynamics and petroleum geology of Miocenefrom the eastern Tunisia. Am Assoc Pet Geol Bull 80(1):63–81

Beni-Akhy R (1998) Etude des impacts anthropiques sur les eauxsouterraines du Sahel oriental tunisien: caractérisation hydrogéolo-gique, modélisationmathématique et cartographie de la vulnérabilitéenvironnementale. PhD thesis, University of Tunis II, p. 377

Bouhlassa S, Aicahi A (2004) Origin of salinity of the Shallow Aquiferof Guelmime, South Morocco: hydrogeochemical study. WaterQual Res J Can 39(2):119–128

Bouri S, Ben Dhia H (2010) A thirty-year artificial recharge experi-ment in a coastal aquifer in an arid zone: the Teboulba aquifersystem (Tunisian Sahel). Comptes Rendus Geosci 342(1):60–74

Burollet PF (1956) Contribution à l’étude stratigraphie de la Tunisiecentrale. Ann Mines Géol Tunis 18:350

Chesnaux R, Lambert M, Walter J, Fillastre U, Hay M, Rouleau A,Daigneault R, Moisan A, Germaneau D (2011) Building a geo-database for mapping hydrogeological features and 3D modelingof groundwater systems: application to the Saguenay-Lac-St-Jeanregion, Canada. Comput Geosci. doi:10.1016/j.cageo.2011.04.013

DGRE (2002) Annuaire de l’exploitation des nappes profondes. Techni-cal report, Direction Générale des Ressources en Eau. Tunis 2002.

DGRE (2005) Annuaire de l’exploitation des nappes profondes del’année 2005. Technical report, DGRE, Tunis.

ESRI, Environmental Systems Research Institute, Inc (1992–1999).Gaaloul N (1995) Les environnements siliciclastiques du Néogène du

Sahel de la Tunisie: palynologie et biosédimentologie. PhD thesis,University of Tunis II, Tunisia, p. 256

Gedeon M, Wemaere I, Labat S (2011) Characterization of groundwaterflow in the environment of the Boom Clay formation. J Phys ChemEarth. doi:10.1016/j.pce.2011.07.089

Ghayoumian J, Mohseni Saravi M, Feiznia S, Nouri B, Malekian A(2007) Application of GIS techniques to determine areas mostsuitable for artificial groundwater recharge in a coastal aquifer insouthern Iran. J Asian Earth Sci 30:364–374. doi:10.1016/j.jseaes.2006.11.002

Hubert C (1968) Etude hydrogéologique du Vindobonien dans larégion de Zéramdine–Béni Hassen. Technical report, DGRE,Tunis, p. 53

2670 Arab J Geosci (2013) 6:2655–2671

IAEA (2001) Characterization of groundwater flow for near surfacedisposal facilities, IAEA, VIENNA, 2001 IAEA-TECDOC-1199,ISSN 1011–4289. p. 61

Jagelke J, Barthel R (2005) Conceptualization and implementation of aregional groundwater model for the Neckar catchment in theframework of an integrated regional model. Adv Geosci 5:105–111

Lachaal, F (2011) Caractérisation hydro-géophysique, hydrodynamique,géochimique et modélisation du système aquifère miocène dans lesblocs de Zéramdine et Mahdia-Jébéniana (Sahel de la Tunisie). PhDthesis, INAT. Univ. Carthage. Tunisia, p 214.

Lachaal F, Bédir M, Tarhouni J, Leduc C (2010) Hydrodynamic andhydrochemical changes affecting groundwater in a semi-arid region:the deepMiocene aquifers of the Tunisian Sahel (central east Tunisia).IAHS Publ 340:374–381

Lachaal F, Bédir M, Tarouni J, Ben Gacha A, Leduc C (2011a) Charac-terizing a complex aquifer system using geophysics, hydrodynamicsand geochemistry: a new distribution of Miocene aquifers in theZéramdine and Mahdia–Jébéniana blocks (east-central Tunisia). JAfr Earth Sci 60:222–236. doi:10.1016/j.jafrearsci.2011.03.003

Lachaal F, Gabtni H, Bédir M, Tarhouni J (2011b) Seismic,gravity, and wireline logging characterization of the Zéramdinefault corridor and its influence in the deep Miocene aquifers distri-bution (east-central Tunisia). Arab J Geosci. doi:10.1007/s12517-011-0298-3O

Mannaï-Tayech B (2009) The lithostratigraphy of Miocene series fromTunisia, revisited. J Afr Earth Sci 54:53–61

Mende A, Astorga (2007) Incorporating geology and geomorphologyin land management decisions within developing countries: a casestudy in Southern Costa Rica. J Hydrol 334:109–124

Mlayah A, Ferreira da Silva E, Hatira N, Jellali S, Lachaal F, Charef A,Noronha F, Ben Hamza Ch (2011) Bassin d’oued Serrat : Terrilsminiers et rejets domestiques, reconnaissance des métaux lourdset polluants, impact sur les eaux souterraines (NW TUNISIE).Revue des Sciences de l’Eau 24(2):159– 175. http://id.erudit.org/iderudit/1006109ar

Nastev M, Morin R, Godin R, Rouleau A (2008) Developing concep-tual hydrogeological model for Potsdam sandstones in southwest-ern Quebec, Canada. Hydrogeol J 16:373–388

Osman A, Abdalla E (2011) Reply on the comment “Review ofGroundwater modeling in semiarid Central Sudan: adequacy andlong term abstraction (Abdalla 2009)” By R. Salama & A. Elamin(2009). Arab J Geosci 4:681–686. doi:10.1007/s12517-010-0126-1

Simler R (2000) Logiciel d’hydrochimie. Diagrammes (v5.0). Laboratoired’Hydrogéologie d’Avignon.

Tweed SO, Leblanc M, Webb JA, Lubczynski MW (2006) Remotesensing and GIS for mapping groundwater recharge and dischargeareas in salinity prone catchments, southeastern Australia. Hydro-geol J. doi:10.1007/s10040-006-0129-x

VanderPost C, McFarlane M (2007) Groundwater investigation insemi-arid developing countries, using simple GIS tools to facilitateinterdisciplinary decision making under poorly mapped conditions:the Boteti area of the Kalahari region in Botswana. Int J Appl EarthObserv Geoinform 9:343–359

Arab J Geosci (2013) 6:2655–2671 2671