Fichier PDF

Partage, hébergement, conversion et archivage facile de documents au format PDF

Partager un fichier Mes fichiers Convertir un fichier Boite à outils PDF Recherche PDF Aide Contact



Siegfried2007Salinization .pdf



Nom original: Siegfried2007Salinization.pdf
Titre: Author template for journal articles
Auteur: Tobias Siegfried

Ce document au format PDF 1.6 a été généré par Acrobat PDFMaker 7.0.7 for Word / Acrobat Distiller 7.0.5 (Windows), et a été envoyé sur fichier-pdf.fr le 12/11/2012 à 12:42, depuis l'adresse IP 41.228.x.x. La présente page de téléchargement du fichier a été vue 977 fois.
Taille du document: 1.5 Mo (37 pages).
Confidentialité: fichier public




Télécharger le fichier (PDF)









Aperçu du document


1

Salinization of Groundwater in the Nefzawa Oases

2

(Tunisia): Results of a Regional-Scale Hydrogeologic

3

Approach

4
5

Mounira Zammouria, Tobias Siegfriedb, Tobias Fahemc, Samiha Kriaad, Wolfgang Kinzelbache

6
7

a

8
9

b

10

c

11

d

12
13

e

Ecole Superieure des Ingénieurs de l’Equipement Rural de Medjez el Bab, Tunisia

Institute of Hydromechanics and Water Resources Management, Swiss Federal Institute of
Technology, Zürich, Switzerland
Institute of Geology, University of Bonn, Bonn, Germany
Ecole Nationale d’Ingénieurs de Tunis, Tunis, Tunisia

Institute of Environmental Engineering, Swiss Federal Institute of Technology, Zürich,
Switzerland

1

1

Abstract

2

A rise of the water salinity of the Complexe Terminal aquifer has been observed as a

3

consequence of the increasing agricultural abstraction during the last decades. The nature of the

4

salinization of the groundwater in the oases is far from understood. Brine from the Chott El

5

Djerid, saline underlying aquifers, as well as agricultural drainage water may contribute to

6

varying degree. This paper looks into all sources of contamination using hydrochemical data

7

available from the beginning of the 1980's. Complementary water samples were taken from

8

different drains and observation wells tapping the CT and the phreatic aquifers. The samples were

9

analyzed with regard to chemistry, temperature, isotopes and other environmental tracers. Based

10

on this, detailed conclusions with regard to the local salinization mechanisms are drawn.

11

Furthermore, a finite difference model was developed to simulate the groundwater flow and

12

contaminant transport in the oases. Calibration for the 1950 - 2000 period was carried out in order

13

to adjust geological and chemical system parameters. The simulation of the planned extraction

14

projects predicts a worsening of the present situation. Maintenance of present abstraction is not

15

sufficient to reduce or stop the salinity increase.

16
17

Keywords: Arid regions, groundwater management, hydrochemistry, numerical modeling,

18

salinization

2

1

1. Introduction

2

Negative effects from groundwater mining are observed globally. They threaten future supply

3

locally. Especially in semiarid to arid regions, where aquifers are the sole freshwater resource,

4

this is problematic and can lead to an excessive rise of provision costs and hamper economic

5

productivity of irrigated agriculture. Both, soil salinization and groundwater quality degradation

6

are the main threats to sustainable utilization of natural resources. For example, on a global

7

average, it is estimated that approximately 77 M ha or about one third of irrigated land is affected

8

by salinization to varying degrees (Gardner 1996). The Sahara oases of Northern Africa are no

9

exception (Ghassemi, Jakeman, and Nix 1995).

10
11

The process of soil salinization consists of an accumulation of salt in the productive root zone of

12

the soil. It is caused by increased salinity of phreatic aquifers and rising groundwater levels that

13

result due to the application of excess irrigation water as well as insufficient drainage of soils

14

(WHO-UNEP 1989). If no adequate measures are taken, complete soil sterilization and, with that,

15

a loss of yield, will result. Mitigation measures include application of excess water so as to flush

16

the soil as well as the extension and / or improvement of drainage systems. Yet, the former

17

technique is only applicable if the latter is ensured. Otherwise, shallow groundwater tables will

18

rise at an even faster pace and further aggravate the situation (Hillel 2000). Economic resilience

19

to such adverse development is weak where natural resources including productive soil and

20

freshwater as well as capital are scarce. Therefore, proper resource management in these areas is

21

crucial.

22
23

One such location is the Nefzawa oases region in Southern Tunisia situated at the northern fringe

24

of the Sahara (see Figure 1). There, climate is arid to hyper-arid with less than 100 mm/a of

25

rainfall with high variability. Daily mean temperatures vary between 10°C in the winter to 32°C

26

in the summer with August being the hottest month. Yearly open water evaporation is in the

27

range of 2’500 mm (Fahem 2003; Mamou 1990). The Nefzawa oases region is famous for the

28

production of the high quality Deglet Nour date. At the turn of the century, Tunisia was selling

29

more than 20'000 metric tons on the world market which accounted for more than half of the total

30

dates export of Africa or 10% of the total Tunisian agricultural export market value (Food and

31

Agricultural Organization 2005).
3

1

Apart from the export of cash crops and fuelled by rising population numbers ( 4.3 ⋅ 106 in 1960,

2

9.8 ⋅ 106 in 2003), agricultural policy in Tunisia in the last fifty years was determined by

3

considerations of food security, self-sufficiency and import substitution practices (Perennes

4

1993). As a result, the irrigated area has increased by a factor of four over the same period of

5

time. In the year 2000, it extended over an area of more than 381’000 ha (Food and Agricultural

6

Organization 2005). This extension was mainly accomplished by the large scale allocation of

7

fossil groundwater resources from the Complexe Terminal (CT) and the Continental Intercalaire

8

(CI). It involved a gradual transition from traditional groundwater abstraction by hand dug wells

9

or springs to high volume abstraction by motor pumps and deep artesian wells. The technological

10

transition was governmentally supported.

11
12

Until 1907, exploitation of the CT in the area of the Nefzawa was accomplished by utilizing

13

discharge from natural springs (see map in Figure 3). These springs were fed by artesian water

14

from the CT. The first deep well with artesian outflow was drilled in 1907. From the 1950’s

15

onward, the flow from the natural springs diminished because of the continuous lowering of the

16

piezometric level by the installed artesian wells. Deep wells were gradually equipped with pumps

17

after 1970. Figure 2 shows the water quantities extracted from the CT between 1950 and 2002.

18

The total abstraction increased more than sevenfold from 1950 with 1.4 m3/s to 10.5 m3/s in

19

2002. The natural outflow of the springs reached 0.47 m3/s in 1950 and practically vanished in

20

1995. Exploitation through pumping was increased from 0.45 m3/s in 1976 to 2.7 m3/s in 2002.

21

The installation of the so-called illicit deep wells brought another notable, yet governmentally not

22

approved increase of exploitation. These wells belong to private farmers settling in the vicinity of

23

existing oases. There, they reclaim land and extract water from privately funded wells. At the

24

present time, exploitation by illicit wells exceeds the officially approved one. It rose from only

25

0.025 m3/s in 1981 to about 6 m3/s in 2002. Piezometric levels in the CT aquifer have fallen at an

26

average rate of 1 m/a over the last 30 years in the region (Mamou and Hlaimi 1999).

27
28

With increasing abstraction activity, a gradual salinization of soil and groundwater resources was

29

observed that threatens irrigated agriculture in the Tunisian South. In some places such as the El

30

Hsay oasis in the southern Nefzawa oases, yields and quality of the dates are diminishing at an

31

alarming rate due to the combined effects from the application of enriched irrigation water and

32

soil salinization.
4

1
2
3

Figure 1: Location of project area in Southern Tunisia is shown by the tilted box. Its outline corresponds to the flow
and transport model domain (see Chapter 4). Figure 3 shows a detailed map of the Nefzawa oases region.

4
5

The situation will likely worsen in the future. In the whole CT basin, planned groundwater

6

withdrawals will reach 89.7 m3/s in 2050 compared to 42.7 m3/s in the year 2000 (Observatoire

7

du Sahara et du Sahel (OSS) 2003). Over the same time period, an increase of 1.1 m3/s of

8

abstraction is projected in the Nefzawa oases region. Although this is a marginal increase locally,

9

the regional CT piezometric levels are influenced by the basin-wide, transboundary groundwater

10

abstraction activity in the aquifer leading in the combination to a considerable decline of

11

piezometric heads.

12
13

In contrast to the mechanisms of soil salinization, the sources of groundwater salinization in the

14

Nefzawa oases region have not yet been identified and remain ambiguous. The salinization

15

phenomena in the region are complex. Several possible causes for salinization exist: a) brine

16

intrusion from the Chott El Djerid, b) salt water upconing from saline underlying aquifers and c)

17

seepage of agricultural drainage water.

18
5

1
2
3
4
5

Figure 2: Development of CT water abstraction in the Nefzawa. Note that due to the pronounced growth of
unofficial (illicit) pumping, abstraction is effectively doubled. The leveling off in the case of the illicit wells is not
observed but rather an expression of lacking data after 1996 and certainly a conservative estimate of the actual
pumped quantity in illicit wells.

6
7

This paper attempts to find an explanation of the deterioration in CT water quality. The relevance

8

of processes will be discussed with regard to their present and future impact on CT water. For

9

this purpose, data available at the Tunisian agencies responsible for management of surface water

10

and groundwater resources, i.e. Commissariats Régionaux au Développement Agricole (CRDA),

11

Kebili and Direction Générale des Ressources en Eau (DGRE) in Tunis, have been collected and

12

analyzed. During numerous field trips in 2002, complementary water samples were taken and

13

hydrochemical as well as isotopic analyses carried out.

14
15

In the remainder of the paper, we first present the geology of the Nefzawa area. Second, different

16

potential causes of groundwater quality deterioration are highlighted and discussed. Third, results

17

from tracer analyses are presented and conclusions with regard to the origins of the salinization

18

drawn. Finally, a numerical groundwater flow and contaminant transport model of the area is

19

presented. The model helps to assess different future water allocation strategies with regard to
6

1

their potential impact on drawdown and water quality. Furthermore, it will help to sensitize local

2

authorities with regard to the formulation of future allocation strategies and their impact on the

3

salinization of the CT groundwater resource.

4

5
6
7
8
9

Figure 3: Overview of the Nefzawa oases region. Important oases are highlighted. The chemical cross sections are
shown in Figure 6. Numbered oases are located in the Presquîle de Kebili (PIK): 1) Tauoergha; 2) Oum Somma; 3)
Mennchia and Ziret Louhichi; 4) Bou Abdallah; 5) El Glea; 6) Oued Zira. CT piezometric levels in 1988 are shown
(after Mamou (1990)). Projection: UTM, Date: WGS 84, Zone: S 32.

10

7

1

2. Hydrogeology and Hydrochemistry of the Nefzawa Oases

2

Region

3

2.1 Hydrogeology

4

Among the different aquifer systems that exist in the study area, the CT and CI formations are the

5

two major ones with regard to extent and their relevance for yield. Phreatic aquifers of variable

6

quantity and quality are only used for backup supply in times of extreme water shortage. The

7

aquifer of the CT extends over an area of approximately 670'000 km2 in the northern Sahara. The

8

term Complexe Terminal describes a multi-layer aquifer which consists of the most recent

9

formations in the northern Saharan basin, i.e. the Senonian, the Eocene and the Mio-Pliocene (see

10

Figure 4). Evidence of hydraulic communication between these formations can be found in the

11

whole basin except in the Chott regions where the middle and upper Eocene are inserted.

12
13

In the Nefzawa oases region the thickness of the CT varies between 100 m in the north-east and

14

400 m towards the south-west. The depth to the top ranges from 50 m to 100 m. It is generally

15

increasing from east to west and towards the Chott. Over most of the Nefzawa oases region, the

16

CT is confined by shales and marls of the Paleocene, as well as evaporites of the middle Eocene.

17

The most recent cover consists of sands as well as shales of the Mio-Pliocene and the Quaternary

18

alluvium. In the Nefzawa, the Senonian limestones are primarily exploited due to their high yield

19

(Mamou and Hlaimi 1999; Observatoire du Sahara et du Sahel (OSS) 2003; UNESCO 1972).

20
21

Towards the east, these limestones are cropping out. The exposure to weathering has left them

22

highly karstified. There, in riverbeds, the CT gets recharged by direct infiltration through floods

23

descending from the Djebel Tebaga and Dahar mountain ranges located to the north and east of

24

Nefzawa. Whereas in the central basin, water ages dated by radiocarbon (C14) indicate

25

paleogroundwater of Pleistocene and early Holocene age (28’000 to 5’000 years), younger water

26

ages at the fringes of the basin are an indication of modern recharge (Edmunds et al. 1997).

27

However, present recharge is low due to the arid to hyper-arid conditions within the catchment. It

28

depends mostly on infrequent storm events. The groundwater flow in the CT aquifer is in general

29

directed from S and E towards the Chott region (see Figure 3). For the study area flow occurs

30

mainly from east and south-east towards west and north-west respectively. CT groundwater
8

1

shows a total dissolved solids (TDS) content ranging between 1 g/l to 2 g/l with locally increased

2

values of up to 6–7 g/l.

3

4
5
6

Figure 4: Generalized conceptual geological model of the Nefzawa oases region (not scaled). The homogeneous
sequence of rock layers is interrupted by the Kébili-Tozeur fault (based on (Mamou 1990; UNESCO 1972)).

7
8

As shown in Figure 4, the underlying Turonian dolomites constitute a separate aquifer over large

9

parts of the Nefzawa oases. The impermeable lagoonal Lower Senonian seals the bottom of the

10

CT. Towards the east in the Dahar mountain range, the Turonian dolomites are cropping out.

11

Along the Kebili-Tozeur fault the Turonian dolomites form part of the CT where they contain

12

freshwater. In contrast the Turonian aquifer is of little interest in the central and southern part of

13

the Nefzawa region due to high salinity with a TDS content of up to 7 g/l.

14
15

The CI is defined as the set of sedimentary layers that comprise mainly continental sandstone-

16

clay formations of the lower Cretaceous. Associated with it are post-Paleozoic and ante-

17

Cenomanian marine or lagoon sediments. It is one of the largest groundwater systems in the

18

world. In total, it covers an area of about 1'100'000 km2. It can be found at depths of 800 to 2'500

19

m with a thickness of around 300 to 1'200 m. Most of the wells tapping the CI show strong

20

artesian conditions. Similar to the CT, the groundwater of the CI contains mainly

21

paleogroundwater whichis dated back to the Pleistocene and early Holocene (UNESCO 1972).

22

Recent recharge is observed at the periphery of the Sahara basin (Edmunds et al. 1997). The
9

1

water is geothermal with temperatures between 60°C to 70°C. The CI, as the CT, is a multilayer

2

aquifer. The CI includes the different aquifers between the bottom of the Triassic and the top of

3

the Albian (UNESCO 1972). Its confining unit is clearly defined as the Cenomanian shales,

4

which separate the CI from the Turonian. In the Nefzawa, the main aquifers are situated in the

5

Lower Cretaceous from the Barremian to the Albian and the base is defined by the shales of the

6

Malm. The general flow of CI groundwater is directed from south-east towards north-west. CI

7

groundwater quality in the Nefzawa oases region is mediocre with TDS ranging from 2 g/l to 4

8

g/l.

9
10

The phreatic aquifers are found in the Mio-Plio-Quaternary sediments and the alluvial fillings of

11

the Wadis. With regard to exploitation, they are only of secondary importance because their

12

water is usually highly salinised with 2-10 g/l TDS (Fahem 2003; Mamou and Hlaimi 1999). In

13

the Presquîle de Kebili (PIK), which corresponds to the north-western part of Nefzawa, located

14

between the Djebel Tebaga and the Chott El, TDS observations in the 1960s showed salinity to

15

range from 5 to 12 g/l (Pouget 1966). This area is also called Presquîle de Kebili (PIK) as it is

16

reaching into the Chott el Djerid like a peninsula (see Figure 3). The phreatic aquifers are

17

recharged by vertical percolation from the confined CT and by infiltration of excess irrigation

18

water as well as water from flood events that descends from the surrounding mountains to the

19

Chott el Djerid. Generally, their piezometric level is elevated in the irrigated oases areas and

20

diminishing with increasing distance from the latter. Measurements of the water table depth show

21

an average depth below ground level from 1 m in the PIK to 5 m in Douz within the irrigated

22

perimeters. Outside the oases, lower water tables have been measured ranging from 9 m in PIK to

23

14 m in Douz (Direction Générale des Resources en Eau 2000). Flow direction of the phreatic

24

groundwater is generally directed towards the Chott El Djerid.

25
26

In times of shortage during summer months, water from the phreatic aquifers is used for

27

irrigation to complement the official water allotment. Due to the mediocre water quality, their

28

waters have to be diluted with CT water to lower the TDS concentration. As of the year 2000,

29

exploitation was approximately 0.05 m3/s of which more than 70% is utilized in the southern

30

Nefzawa oases region (Douz and El Hsay oases).

31

10

1

The Chott El Djerid finally is covering an area of approximately 5'400 km2. It is a playa system

2

in a subsidence basin located between the southern flank of the Djebel Tebaga mountain range

3

and the northern part of the Sahara platform (see Figure 1). In the undisturbed state, the Chott

4

was the discharge area for water from the CT and the phreatic aquifers. Groundwater inside the

5

Chott is highly salinised often exceeding 100 g/l of dissolved salts (Gueddari 1980). The

6

hydrogeology of the Chott is not well known. No wells or piezometers exist inside the Chott.

7

Silty sediments and crusts formed by salts cover large parts of it. The sedimentation of the Chott

8

Djerid is evaporitic and lacustrine since the late Pleistocene. Geological log samples made in

9

different areas in the Chott show layers of clays and salty materials (Gypsum) to a depth of

10

around 60 m which is the maximum thickness of the Quaternary Chott sediments. Between 60 m

11

and 125 m, there are sandy sediments from the Mio-Pliocene with clayish and salty intercalated

12

layers (Meckelein 1977). Bedded marly sandstones or red shales are found at the bottom in the

13

center of the Chott up to a depth of about 900 m. The groundwater table is high, ranging from 0.2

14

to 0.5 m below the surface. Winter flood events can cause standing water in some areas of the

15

Chott.

16
17

An interesting but still not clearly understood natural feature in this context is the occurrence of

18

spring mounds on the fringes of the Chott El Djerid in the Nefzawa oases. In the past, CT water

19

evaporated in these spring mounds after having risen through preferential flow paths from the CT

20

towards the quaternary surface (Meckelein 1977). The lowering of the hydraulic heads of the CT

21

in recent years caused a complete drying up of the springs.

22
23

2.2 Complexe Terminal Hydrochemistry

24

Hydrochemical investigations in the Nefzawa started in the late 1950s. However, systematic and

25

continuous observations are only available from 1980 on. At this time, CT water of the PIK oases

26

were elevated in TDS (2 to 3.5 g/l) when compared to the rest of the Nefzawa oases (1 to 2 g/l in

27

average). During the last 20 years, an increase in TDS of 1 up to 2 g/l was observed in the PIK

28

(Ras el Ain, Bou Abdallah and Telmine). In the southerly Douz region, including the El Hsay

29

oasis, the TDS increased more than 3 to 4 g/l. As a consequence, some of the pumped water is no

30

longer suitable for irrigation. In contrast, oases situated in the central region as well as next to the

11

1

Chott El Djerid, e.g. Negga, Ghidma and Guettaya, do not show any deterioration of water

2

quality (see Figure 5).

3
4

The chemical N-S cross-sections through the study region depict the situation in 2002 (see Figure

5

6, the geographic location is shown in Figure 3). They show that the salinization chemistry is

6

changing towards the South. There, concentrations of sodium and chloride are relatively

7

enriched. This could be an indication of evaporation processes that influence the local

8

groundwater chemistry. However, it is noteworthy that the change in salinity is not

9

homogeneously distributed within the affected regions, neither temporally nor spatially. Even

10

inside the same oasis such as for example Ras el Ain, CT boreholes located in close proximity to

11

each other show different behavior.

12

13
14
15

Figure 5: Exemplary TDS development in the Nefzawa oases showing the different regional salinization trends of
CT water.

16

12

1
2
3
4

Figure 6: Chemical cross-sections through the study area. Cations are shown in the upper, anions in the lower cross
section. Quality deterioration is observed in the PIK and Kebili region as well in the southern Nefzawa (see also
Figure 3). Groundwater samples were taken in the field in 2002.

5
6

As mentioned before, three differing salinity sources are likely causes for the deteriorating

7

quality of CT water in the Nefzawa oases region. These sources and related salinization

8

mechanisms are shown in Figure 7 and discussed in the following.

9
10

The Chott El Djerid contains very saline groundwater within its tertiary and quaternary

11

sediments. There, TDS values range from 10 g/l up to 350 g/l (Fahem 2003; Gueddari 1980;

12

Kbir-Ariguib, Chehimi, and Zayani 2001). In the undisturbed state, the elevated CT head relative

13

to the shallow groundwater piezometry in the Chott prevented any significant downward

14

percolation of brine. Progressive groundwater mining lowered the CT head leading to an

15

inversion of the hydraulic gradient in the CT. Figure 7 shows how an infiltration of Chott water

16

into the CT may be caused by such a gradient inversion. Piezometric levels of the CT prior and

17

during exploitation are shown in a schematic way. It is conceivable that Chott water infiltrates
13

1

into the CT with the progression of the gradient inversion. In this case infiltrating water has to

2

pass the semipermeable aquitard of the Pleistocene clays. Although these clays cover most of the

3

Cretaceous in the Nefzawa oases region, their thickness varies. After sedimentation in the late

4

Paleocene, these clays were exposed to erosion on the surface and partially attenuated.

5

Percolation through these strata, further accentuated by density driven flow of the Chott's brine,

6

might be possible locally.

7
8

Turonian and CI water are a second potential source of salinity in the Nefzawa. Both show

9

increased salinity in this region. In the Turonian aquifer, measured TDS values range from 2.5 -

10

3.5 g/l in the PIK to 7 g/l in the south-eastern part of the Nefzawa. Similarly, TDS in the CI

11

increase from 2.5 - 3.5 g/l in the PIK to 4 g/l in Douz. It is likely that salinization of the CT

12

occurs where interactions between these aquifers exist. The higher pressure heads of the Turonian

13

and CI allow their water to percolate upward. With a lowered piezometric level of the CT due to

14

mining and a still quite unchanged level of the Turonian, this upwelling might be increased and

15

could lead to a density layering in the basal zone of the CT formation (see Figure 7). Under these

16

circumstances, pumping possibly triggers upconing which would explain a quality deterioration

17

of the pumped CT water.

18
19

Finally, agricultural drainage water with an increased TDS collects in terrain depressions, so

20

called Sebkhas, where it gets enriched due to strong evaporation. The TDS of drainage water

21

varies from 8 g/l in winter to 25 g/l in summer. In autumn 2001, the measured TDS in the Sebkha

22

of El Hsay reached 100 g/l (Fahem 2003). Especially in summer months, drainage waters do not

23

get flushed to the more distant Chott due to greatly reduced flow. These waters then collect in the

24

Sebkhas within close proximity of the oases. Consequently, they infiltrate to the underlying

25

shallow aquifer (Direction Générale des Resources en Eau 2000). As already stated, the confining

26

clays of the Paleocene are not a continuous layer of equal thickness (see Figure 7). They might be

27

thinning and forming lenses thus containing preferential infiltration pathways. Locally, in areas

28

of high exploitation where a lowering of the CT head below the head of the oases aquifer is

29

observed, either oases drainage water or highly salinised water from the Sebkhas might infiltrate

30

into the CT.

14

1
2
3

Figure 7: Conceptualization of the three different salinization processes. A: Infiltration of brine from Chott; B:
Upconing of water from the CI/Turonian aquifer; C: Salinization by agricultural drainage water (Fahem 2003).

4
5

3. Determination of Salinization Origins

6

To get an understanding of the complex salinization phenomena observed in the Nefzawa oases,

7

data on chemistry, temperature and isotopes available at the CRDA, Kebili were obtained for the

8

period of 1957 to 2000. This data set was complemented by sampling groundwater and

9

temperature at 53 sites from deep wells in the CT and CI, shallow wells, drains and Sebkhas in
15

1

2002. Subsequently, chemistry and radioisotopes (²H, 3H, 18O) were analyzed. All sampling was

2

coordinated with local representatives. Data are provided in Table 4 in Appendix 1.

3
4

First, correlations between the pointwise CT salinity increase and pumping, piezometry as well as

5

structural properties were investigated. No significant correlation was found between salinity

6

increases on the one hand and pumping amounts or decline of hydraulic heads in the CT on the

7

other. However, a strong relationship between the increase of the TDS and the depth of the CT

8

top exists. The latter is generally coincident with the screen top for most of the boreholes.

9

Boreholes tapping the CT near to the surface show a pronounced increase of the TDS whereas

10

deeper CT boreholes do not show any sign of a deterioration of the pumped water quality (see

11

Figure 8). The phenomenon is visible in the northern (PIK, Kebili) and southern Nefzawa region

12

(Douz, El Hsay). This result indicates possible salinization by the third salinization source, i.e.

13

the phreatic aquifer where the shielding of the CT by its confining unit is less pronounced.

14

15
16
17
18

Figure 8: CT salinity development between 1980 and 1997, i.e. ΔTDS = TDS(1997) - TDS(1980), at different
locations in the Nefzawa in relation to the depth to the CT top (Kriaa 2003). CT water temperatures are shown for
selected oases (see also map in Figure 3 for the location of the oases within the Nefzawa oases region). An average

16

1
2

geothermal gradient of 5.5°C / 29 m = 20°C/km can be calculated. Note that Negga 2 bis is located several
kilometers south of the Kebili-Tozeur fault in the PIK and not affected by CI water upwelling.

3
4

In order to further substantiate this finding, bromide / chloride and sulfate / chloride ratios were

5

plotted against each other. Chloride and bromide are the two most conservative constituents in

6

waters, i.e. they are little affected by chemical reactions such as mineral precipitation or

7

adsorption. Similarly, sulfate is generally conservative during the mixing of different waters, and

8

is usually much more conservative than dissolved cations (Whittemore 1984; Whittemore 1988).

9

Results are shown in Figure 9. CT samples from Douz and El Hsay Oases show similar ion ratios

10

as water from drains and surface wells located in close proximity, i.e. Douz 2 bis / Puit Surface,

11

Douz SE as well as El Hsay 5 bis and Drain El Hsay and finally Douz Ouest / Puit Surface Douz

12

NO (see Figure 9).

13
14

A downward percolation of TDS enriched shallow groundwater can only occur if the piezometric

15

head of the oases aquifer is locally higher than the CT head. By 1996, CT artesianism around

16

Douz and El Hsay retreated up to 10 km west of these oases in the direction of the Chott El

17

Djerid. CT piezometric levels are up to 20 m below ground. Hence, conditions conducive for

18

direct infiltration prevail in the southern Nefzawa region. In contrast, CT boreholes in the PIK

19

and northern Nefzawa region have no affinity to surface water samples, i.e. Drain Telmine. The

20

hypothesis of surface water contamination cannot be maintained there.

21
22

Rather, an influence of the geothermal CI water in the PIK seems evident from the observed

23

temperature anomaly of the CT groundwater. Average water temperature of the central and

24

southern CT in the Nefzawa oases region is around 24 o C . In the northern Nefzawa and the PIK,

25

average temperature increases to approximately 27 o C . Figure 8 shows CT water temperatures

26

for selected oases. Temperatures of boreholes not affected by contamination as well as El Hsay

27

and Douz 2 bis in the southern Nefzawa region follow nicely the regional geothermal gradient of

28

about 20°C/km (Ben Dhia and Bouri 1995). In contrast, the temperature anomalies of the

29

boreholes El Glea (27°C) and Oued Zira (26.7°C) located in the PIK are clearly visible. They

30

cannot be explained by borehole screen depth alone. There, mixing of CT groundwater with

31

water from the CI is a likely salinization source (on this topic see also (Mamou 1990)).

32
17

1
2
3
4
5

Figure 9: The plot shows Br-/Cl- vs. SO42-/Cl- ion rations. The two CT boreholes Douz 2 bis and El Hsay 5 bis show
similarities in their ion ratios and with the ion ratio of the corresponding surface water samples from nearby
locations. Although of different ionic characteristics compared to the other samples in the southern Nefzawa region,
Douz Ouest and Puit Surface Douz are chemically related.

6
7

Isotopic analysis further supports the above results. Stable isotope samples from the CI obtained

8

in 2002 are grouped together with the ERESS samples (UNESCO, 1972) along the local meteoric

9

water line LMWL (see Figure 10). In contrast, CT samples mostly cluster below the LMWL and

10

are shifted to the right. Their isotopic compositions follow a characteristic evaporation line which

11

was interpolated from the CT data set. The CT evaporation line crosses the LMWL at -7,5 ‰

12

δ18O and –55,1 ‰ δ2H. This intersection is in the vicinity of CI samples. Original CT and CI
18

1

waters were formed by precipitation during the pluvial ages and are of equal origin. The shifting

2

of CT samples along the evaporation line shows that the CT receives younger recharge waters

3

that were subject to evaporation.

4

5
6
7
8

Figure 10: Relationship of δ²H and δ18O for samples from CT and CI taken during the field campaign 2002 (see
Table 4 in Appendix 1) . CT samples are from different areas. The evaporation line was interpolated from all
available isotope data of the area.

9
10

Apart from direct infiltration of local drainage water from the phreatic aquifer, a more distant

11

source of saline phreatic water has to be considered as well. In their report, (Mamou and Hlaimi

12

1999) present stable isotope measurements of phreatic water samples taken to the South and the

13

East of the study area (see Figure 10). They show a strong evaporative nature, being aligned

14

along the same evaporation line as the CT samples in the south. 30 to 50 km to the south and east

15

of the Nefzawa oases region, the CT is partly outcropping and partly covered by dunes. In this

16

region, enriched phreatic water can easily infiltrate the CT. Since the general groundwater flow

17

direction is south-east to north-west towards the Chott, it is evident that the CT values of the

18

southern and eastern Nefzawa region are enriched in isotopes due to mixing with recent water

19

subject to strong evaporation. Recent water has not yet replaced the paleowater in the CT. In

20

other words, the momentary CT waters’ isotopic pattern in this area marks the transition in

19

1

climatic conditions during the Holocene. Given time, all paleowater will eventually be replaced

2

by modern water.

3
4

In contrast, the relative shift of stable isotope values for the northern Nefzawa towards the

5

LMWL is most likely caused by CI groundwater upwelling. Direct recharge alone through

6

preferential pathways in the fractured limestones and dolomites on the southern flank of the

7

Djebel Tebaga could not explain the observed temperature anomaly.

8
9

No correlation between distance of borehole to Chott and TDS development was found (Kriaa

10

2003). Oases adjacent to the Chott El Djerid such as Negga, Gueliada and Ghidma do not show

11

any signs of deterioration of CT water quality. A present day influence of the Chott's brine is not

12

visible and its relevance as a source of salinization is refuted. However, this may not hold true for

13

the future when a regional scale gradient inversion will occur. The pertinence of this salinization

14

source with regard to planned aquifer exploitation will be discussed in Chapters 4.2 and 5.

15
16

4. Groundwater Modeling Approach

17

4.1 Flow Model

18

In order to assess the future impact of pumping on the CT aquifer in the Nefzawa over the next

19

fifty years, a quasi 3D finite-difference flow model was developed using Modflow-2000, a

20

modular finite-difference groundwater model code (Harbaugh et al. 2000). For model pre- and

21

post-processing, Processing MODFLOW was utilized (Chiang and Kinzelbach 2001). The

22

development and calibration of the model are based on previous work by (ARMINES and ENIT

23

1984; Observatoire du Sahara et du Sahel (OSS) 2003; UNESCO 1972) and described in detail in

24

(Kriaa 2003).

25
26

The study area covers an area of about 5’500 km2 and was horizontally discretized into square

27

cells of 780 m side length (see Figures 11 and 12). The north-eastern boundary of the model

28

follows the CT basin boundary. The other limits are artificially chosen with general head

29

boundary conditions that represent the hydraulic connection to the greater CT basin. These time-

20

1

variable third-type or Cauchy boundary conditions define the flux and head over the domain

2

limit.

3

4
5
6
7
8

Figure 11: Nefzawa model extent and discretization of phreatic aquifer layers. Gray cells: Chott drain boundary
conditions; Black cells: inactive cells; Green cells: General head boundary cells indicating the agricultural drains.
Blue cells: recharge. Piezometric contour lines are shown for the steady state in 1950 (see also Figure 1 for the
location of the project area).

9
10

Vertically, the upper model layer corresponds to the phreatic aquifers which are laterally

11

connected to the Chott aquifer. Layer geometry is shown in Figure 11. Only very limited

12

information about those phreatic aquifers, i.e. their extent and formation, is available (Mamou

13

and Hlaimi 1999). Combining the oases aquifers, the Wadi aquifers as well as the Chott aquifer

14

within one model layer, we assume that this aquifer extends over the whole study region in the

15

Mio-Plio-Quaternary sediments. Evapotranspiration occurs from this layer only. Recharge from

16

precipitation at the southern flank of Djebel Tebaga is modeled by prescribed inflow. There, an

17

estimated 105 m3/a is assumed (see Figure 11). This corresponds to the value presented in

18

(Mamou and Hlaimi 1999).

19
20

Phreatic groundwater, generally following the terrain gradient, flows towards the Chott where it

21

gets removed by evaporation. On-farm drainage systems in the oases are represented by general
21

1

head boundary conditions which allow to model exchange with the phreatic aquifer. The location

2

and geometry of the drain network were obtained by the superposition of a Landsat image

3

(Landsat-7-ETM, 16.08.2000) with the model grid. For steady-state calculations, only the ancient

4

oases have been accounted for, whereas the many new extensions of the irrigated perimeters from

5

1950 up to today are implemented in the transient model. Due to lacking knowledge of hydraulic

6

properties, i.e. hydraulic conductivities and storage coefficients, the confining clays, marls and

7

evaporites between the two aquifers have not been modeled explicitly. Instead, the CT and Mio-

8

Plio-Quaternary sediments are connected by a leakage term.

9

10
11
12
13
14

Figure 12: Nefzawa model extent and discretization of CT. Black cells: inactive cells; Red cells: pumping boreholes;
olive green cells: General head boundary cells to account for the regional CT development; Grey cells: drains at the
southern flank of Djebel Tebaga (corresponds to the location of the ancient sources). Piezometric contour lines are
shown for the steady state in 1950.

15
16

Accordingly, the lower model layer then corresponds to the CT aquifer as depicted in Figure 12.

17

It incorporates the stratigraphic units of the upper and lower Senonian formations. Its geometry

18

was interpolated on the model grid from borelog data. Springs are implemented as drains whose

19

flow should be reproduced after model calibration (see (Harbaugh et al. 2000) for more

20

information on the Modflow-2000 Drain package). The connection to the Turonian is modeled by

21

a general head boundary condition. Its representation should help to determine inflow via the
22

1

Turonian to the CT in steady and transient state. The prescribed heads that connect this regional

2

model to the greater CT basin vary in time in relation with the historical regional decline of the

3

CT piezometric levels. The latter is observed in observation wells that are located both inside and

4

outside of the study area.

5
6

The period taken as reference for steady-state calibration of the model is 1950, in which the CT

7

was considered in a state of equilibrium on a basin wide level (Observatoire du Sahara et du

8

Sahel (OSS) 2003; UNESCO 1972). Steady-state calibration consisted of reproducing the general

9

piezometric maps for the phreatic and CT aquifers as well as the discharge rates at Nefzawa

10

springs as measured in 1950 (see Figure 13). The time period from 1951 to the year 2000 was

11

taken as reference period for transient calibration. The reproduction of the drawdown trend as

12

well as the temporal reproduction of the total measured spring discharge were taken as calibration

13

criteria. The calibration parameters are horizontal transmissivities, storage coefficients and

14

specific yield for the phreatic aquifer, leakage between the latter and the CT as well as between

15

the CT and the Turonian and finally, the hydraulic resistances between the phreatic aquifer and

16

the soil surface at the level of springs and agricultural drains. The calibration process was

17

manual. Note that the availability of discharge data makes the calibration exercise a well-posed

18

problem.

19

20
21
22

Figure 13: Left plate: Scatter plot of steady-state calibration. Right plate: Temporal reproduction of spring discharges
(1950 - 2000)

23
23

1

The comparison of the available observed and calculated piezometric values of the CT aquifer in

2

1950 indicates a satisfactory model calibration with a correlation coefficient of R 2 = 0.71 and a

3

root mean squared error σ = 3.77 m . As shown in Figure 14, the calculated drawdowns over the

4

period 1951 – 2000 compare well to the observed ones and do not show any systematic deviation

5

(Kriaa 2003). Finally, we consider the temporal reproduction of the total measured spring

6

discharge as acceptable (see right plate in Figure 13).

7

8
9
10
11

Figure 14: Comparison of observed vs. calculated drawdowns at observation boreholes from 1950 to 2000. Data
from 5 observation boreholes were available i.e. Douz, Douz Nord, Kebili Sud, El Faouar and Noueil (see Figure 12
for the location of the observation boreholes).

12
13

The water budget (see Table 1) confirms the ongoing aquifer mining. As of the year 2000, strong

14

changes in the hydrodynamic regime of the aquifer systems compared to the steady state situation

15

are calculated. The loss of water by evaporation decreases from 3.0 to 1.2 m3/s. This includes

16

water being removed by the drainage nodes inside the Chott El Djerid as well as by

17

evapotranspiration from the phreatic aquifer. The inflow from the saline Turonian in 2000 (4.1

18

m3/s) is 12 times the value calculated for 1950 (0.3 m3/s). In the steady state, exchange through

19

the semi-pervious clay layer intercalated between the CT and the phreatic aquifer was solely

20

upward, i.e. from the CT towards the phreatic aquifer. Contrary to that, a reversed leakage from
24

1

the TDS enriched phreatic aquifer to the CT amounting to 0.5 m3/s is found in the year 2000. The

2

comparison of the CT vs. phreatic aquifer piezometry in 2000 indicates that the hydraulic head of

3

the former is still above the critical level of the Djerid Chott. A gradient inversion in a regional

4

sense therefore has not yet taken place. However, the strong drawdown caused by both, the

5

artesian and pumped wells, slowly progressed westward towards the Chott over the last fifty

6

years. The future implications of that will be discussed in Chapter 5.

7
Phreatic Aquifer

1950

2000

Reservoir depletion
Exchange with TC aquifer
Recharge
Total

3.15
0.01
3.16

1.29
0.45
0.01
1.75

Outflow [m 3 /s]
Exchange with TC aquifer
Evaporation and Aioun springs
On-farm drains
Total
Terminal Complex (TC)

0
3.05
0.11
3.16

0.51
1.23
0.01
1.75

0
4.87
0.35
5.22

0.66
0.51
6.02
4.14
11.33

3.15
0.87
0.72
0.48
5.22

0.45
9.27
1.6
0.01
11.33

3

Inflow [m /s]

Inflow [m 3 /s]
Reservoir depletion
Exchange with phreatic aquifer
Exchange with artificial limits
Exchange with Turonian aquifer
Total
Outflow [m 3 /s]
Exchange with phreatic aquifer
Pumping
Exchange with artificial limits
Nefzawa Springs
Total

8
9

Table 1: The water balances calculated for 1950 and 2000 according to the regional flow model.

10
11

4.2 Transport Model

12

Based on the satisfactorily calibrated flow model, a transport model was built by using the

13

transport model code MT3DMS (Zheng and Wang 1999). The goal was twofold: to cross-

14

validate conceptualization by the comparison of the historical salinity development with the

15

calculated one and to get an understanding of the future relevance of the three salinization

16

mechanisms discussed above. For this purpose, the layers of the flow model were further

17

discretized vertically in order to avoid instantaneous vertical mixing. The phreatic aquifer was

18

divided into 3, the CT into 10 layers. Vertical and horizontal permeabilities, the geometry as well

19

as the boundary conditions of the 13 layer model were correspondingly adopted to conform to the

20

2 layer flow model.
25

1

To arrive at a more or less realistic initial salinity distribution, crude but nevertheless

2

representative values were assigned to each of the water bodies, fluxes as well as boundary

3

conditions. First, a uniform CT porosity value of 0.2 was chosen. After calibration, this parameter

4

was finally set to 0.1 in the whole domain. Calibration revealed a high model sensitivity of the

5

CT porosity (see below). The porosity of the phreatic aquifer was uniformly set to 0.2. The initial

6

salinity distribution of the CT was taken from a 1950 map with values ranging between 1.5 g/l

7

and 3 g/l (UNESCO 1972). In the phreatic aquifer, we distinguish between the Chott region and

8

the oases and Wadi aquifers. In the latter, we adopted a uniform salinity of 8 g/l corresponding to

9

averaged observation values. In the Chott, salinity decreases gradually from 175 g/l in the north-

10

west to 10 g/l on its fringes according to actually measured concentration values (Gueddari

11

1980).

12
13

Information on water quality of the Turonian is only available in the PIK and in a borehole drilled

14

to the south-east of the Nefzawa oases region. There, a TDS of 7 g/l was measured. Lacking more

15

detailed knowledge, the boundary condition representing the Turonian was assigned a uniform

16

TDS of 7 g/l as fixed concentration. The same holds true for the CT boundary conditions with

17

fixed concentrations ranging from 1.5 - 3 g/l. The imposed concentrations in the agricultural

18

drains correspond to averaged observed TDS values which are regularly measured by the CRDA-

19

Kebili. They range from 4 g/l in winter to 25 g/l in summer. On average, representative values are

20

16 g/l in Douz-El Hsay, 12 g/l in the PIK and Kebili and 8 g/l in the other drains. In the Sebkhas

21

present in the vicinity of the oases, a TDS of 100 g/l is imposed. Based on these assumptions, a

22

more consistent initial concentration was subsequently calculated by flushing the aquifer in the

23

undisturbed state over several thousands of years to arrive at an approximation of the state in

24

1950.

25
26

Subsequently, the period 1951 - 2000 was simulated. Figure 15 shows the results obtained in

27

seven selected oases. These oases were chosen to be representative for the corresponding region,

28

i.e. the Chott-bordering area (Piezo Negga, El Fauoar 2), PIK (Fatnassa, Ziret Louhichi), the

29

Kebili region (Ras el Ain 4) as well the southern part of the Nefzawa (Douz 2bis, El Hsay 5b).

30

The calculated TDS values have been shifted vertically to make the first measured TDS value

31

coincide with the calculated value . This procedure allows to mask the calibration deviation in the

32

reconstitution of the initial concentrations.
26

1

Generally, the transport model confirms our understanding of the salinization processes in the

2

Nefzawa. Until now, the Chott is of no relevance with regard to an increase in TDS (e.g. Piezo

3

Negga). However, the fingerprints of upwelling from the confining deep layers in the PIK (Ras El

4

Ain 4) as well as backflow from agricultural drainage water Douz 2bis, El Hsay 5b) are visible.

5

In the latter cases, calculated trends of TDS development are too low. This may be an indication

6

of either underestimated fluxes in the model from the pollutant sources or an overestimation of

7

effective porosity ne or both. It may also be related to a conceptual issue of the model. For the

8

region of Douz and El Hsay, analysis of the available piezometric data indicates that the observed

9

piezometric levels in CT boreholes located in a sector of 3 km in diameter vary strongly. The

10

hydraulic head difference can reach 10 m corresponding to a hydraulic gradient of 5 ⋅ 10 −3 while

11

on average, the regional hydraulic gradient is 7.5 ⋅ 10 −4 . The drilling logs show a multilayer

12

structure of the CT aquifer that is composed of two chalk layers separated by a semi-pervious

13

layer of marls and clayey limestones. The deeper CT wells tap the second chalk layer whereas the

14

less deep CT wells tap the first one. This probably explains the existence of a vertical gradient of

15

hydraulic head and concentration. Unfortunately, there is not sufficient data available about top

16

and bottom of the screens in those boreholes to define a correlation between water salinity and

17

the lithostratigraphic nature of the screened formation. Structural issues of the CT and the

18

phreatic aquifer in the Douz region remain an area of future research.

19
20

Clearly, the assumption of regionally homogeneous porosity is not appropriate since the time-

21

scales of the observed and modeled phenomena do not match in the Douz area. A local zone of a

22

strongly decreased porosity could be introduced in the CT that would indicate a more clayey

23

nature of the aquifer. With this, the calculated TDS trend would increase naturally due to

24

increased transport velocity. Such modifications would also account for any unresolved

25

variations of the conductivities of the fractured limestones. Nevertheless, the lacking knowledge

26

of the spatial distribution of effective porosity does not justify a departure from our assumption of

27

a homogeneous regional parameter value. With regard to that, one has to keep in mind that our

28

modeling approach is regional covering an area of more than 5'000 km2. Despite localized

29

imperfections of the transport model, the threat to agricultural productivity due to increasing TDS

30

in the pumped water can be assessed for the whole of the Nefzawa.

27

1
2
3
4
5
6

Figure 15: Comparison of calculated with observed TDS values for selected oases from the year 1970 to 2000. The
filled dots are actual observations on TDS. Calculated TDS development is marked by the connected empty dots.
The calculated values have been shifted vertically to make them coincide with first observed TDS values. Screen
bottom depth. Fatnassa: 168 m, Ras El Ain 4: 115 m, Ziret Louhichi: 75 m, El Fauoar 2: 146 m, Piezo Negga: 200 m,
El Hsay 5b: 150 m, Douz 2bis: 67 m

7
8

5. Future Development

9

Here, we use the transport model to asses the impact of the long-term application of existing and

10

planned extraction projects on groundwater quality. For this purpose the groundwater simulation

11

model was extended to the year 2050 for various management alternatives: Scenario 1)

12

maintaining present withdrawals; Scenario 2) decreasing present groundwater extraction in

13

Tunisia and Scenario 3) increasing groundwater pumping over the whole basin according to the

14

planned extraction projects of Algeria, Libya and Tunisia (see Table 2). In the three scenarios and
28

1

over the period 2000-2050, the time-variation of the prescribed head along the artificial domain

2

limits is consistent with the results of the groundwater simulation model developed for the whole

3

SASS basin (Observatoire du Sahara et du Sahel (OSS) 2003).

4

5
6

Year

2000

Algeria
Tunisia
Libya
Total

20.9
14.4
7.4
42.7

2050
Scenario 1 Scenario 2 Scenario 3
20.9
20.9
53.6
14.4
12.2
17.7
7.4
7.4
18.4
42.7
40.5
89.7

Table 2: Country-wide pumping (in m3/s) in the CT according to withdrawal scenarios investigated.

7
8

Flow model results show that the maintenance of the present pumping rates as of 2000 over the

9

next 50 years (i.e. Scenario 1) provokes an average additional drawdown of 20 m in the Nefzawa

10

oases region. In the second scenario, it was assumed that irrigation water efficiency in the Djerid

11

and Nefzawa oases could be increased by 15% by means of capital investment to modernize

12

deficient irrigation networks and technology. This could then, of course, imply a corresponding

13

reduction of 15% in the total pumped quantity. In this case, the average additional drawdown in

14

the Nefzawa would amount to approximately 9 m. Scenario 3 finally, explores regional effects by

15

a cumulative increase of more than 100% in pumping over the whole CT basin. Model results in

16

this case yield an average additional drawdown of 34 m.

17

18
19

Fluxes
Reservoir depletion
Phreatic aquifer
Artifical limits
Turonian aquifer
On-farm drains
Pumping
Evaporation and Aioun springs

Scenario 1 Scenario 2 Scenario 3
1.71
1.18
6.09
1.09
0.36
1.98
1.86
2.45
-3.71
6.32
4.81
7.58
0.15
0.09
0.16
-9.27
-7.62
-9.37
-0.78
-0.92
-0.77

Table 3: Summary of the fluxes (in m3/s) in 2050 according to scenarios (>0: inflow into CT; <0: outflow from CT).

20
21

The results from the scenario analysis support the conjecture that all three pumping strategies

22

cause a major quality deterioration of the CT waters all over the Nefzawa (see Figure 16). The

23

already affected PIK (Ziret Louhichi, Ras el Ain 4) together with the Douz (Douz 2bis, El Hsay

24

5b) region show an even more pronounced TDS development than the Chott-bordering oases
29

1

(Piezo Negga, El Fauoar 2). According to Scenario 3, CT groundwater pumped in the southern

2

Nefzawa will become unusable for irrigation purposes with the TDS load close or even above 6

3

g/l towards 2050. In fact, it is in the southern part of the study region where, according to

4

governmental planning, abstractions should be increased most. Scenarios 1 and 2 do not show

5

any significant difference in their negative influence on water quality. The reduction in pumping

6

by 15% lowers the influx from the Turonian as well as the phreatic aquifer (see Table 3).

7

However, salinity still rises although at lower rate. Therefore, this measure is not sufficient to

8

stabilize salinity at present levels.

9

10
11
12
13
14
15

Figure 16: TDS development from 2000 - 2050 according to the scenarios. The filled dots are actual observations on
TDS. Calculated TDS development is marked by the connected empty dots. The calculated values have been shifted
vertically to make them coincide with first observed TDS values. The TDS rise in the Chott bordering area is not
induced by brine. Rather, it depicts the general TDS rise in the CT caused by upwelling of Turonian/CI water and
backflow of agricultural drainage water.

30

1

6. Conclusions

2

The simulation results show, that neither present nor planned pumping schemes for the Nefzawa

3

oases region are sustainable. All scenarios investigated show a strong decline in the general

4

piezometric levels. However, the main concern is due to quality issues. All over the study area, a

5

pronounced regional TDS increase will be observed. As in the case of the Douz region, the

6

locally pumped CT water would have to be diluted with fresher water in order to be still suitable

7

for irrigation purposes.

8
9

The northern part of the Nefzawa and the PIK region are mainly affected by upwelling of saline

10

waters from the CI. This process is promoted by the gradual lowering of the CT piezometric

11

heads. Since the overall pumping on the whole basin is responsible for this to occur, local

12

measures of reduced pumping are only of limited effect there. In contrast, in the region of Douz,

13

backflow of drainage water is the main cause for salinization. Here, a simple back of the

14

envelope calculation is instructive. The TDS of the pumped CT water derives from the mixture of

15

waters from the phreatic aquifer, the CT and the Turonian aquifer. By using calculated average

16

fluxes and concentrations of the individual contributions, it is easy to show that a 50% reduction

17

of drainage water backflow would cause a TDS reduction in the pumped water of 21%. A

18

backflow reduced by 75% would lead to a TDS reduction of 32% and result in an average

19

concentration of 3.5 g/l. Evidently, the replacement of the traditional furrow irrigation by PVC

20

tubes, a switch to precise irrigation practices as well as removal of drainage water from the area

21

are effective measures to stabilize TDS values.

22
23

Transport model results revealed problems in the south-eastern part of the model. Sensitivity

24

analysis showed the importance of the effective porosity parameter on results. Model results

25

could certainly be improved by better knowledge of the distribution of this parameter. All over

26

the Nefzawa oases region, salinity measurements at various depths would give important

27

information about the characteristics of the vertical salinity gradient.

28
29

Control measures to limit CT salinization would include the following. First, planned extraction

30

projects in the south of Nefzawa have to be abandoned. Second, pumping will have to be

31

significantly reduced within the oases and the extension of the irrigated perimeters halted so as to
31

1

minimize the contamination risk from the sources of salinity discussed above. Legal regulation

2

could help to avoid a further increase of groundwater utilization by private farmers which are not

3

subject to governmental planning. Third and finally, increasing irrigation efficiency as well as the

4

implementation of effective drainage measures should go hand in hand with a strategy of

5

moderation. From today's perspective, it seems inescapable that CT pumping will have to be

6

relocated to more distant places over the course of time and the pumped water subsequently

7

conveyed to the fertile oases soils.

8
9
10
11
12

Acknowledgements
The authors wish to acknowledge Dr. B. Ben Baccar and Prof. M. Besbes for their valuable help and comments in
the data analysis. The comments of three anonymous reviewers were greatly appreciated. Tobias Siegfried and
Tobias Fahem were partially supported by the Alliance for Global Sustainability (AGS).

13

32

1

Appendix 1: Hydrochemical and Isotope Data of Sampling Campaign in 2002

2

Table 4: Results of hydrochemical and isotope sampling campaign 2002 in the Nefzawa oases.
Sample name

Nefzawa
region

1

2

Coordinates
East
North
[m]
[m]

Na

+

[mg/l]

+

K

Ca

[mg/l]

2+

[mg/l]

2+

Mg

[mg/l]

Cl

-

2-

-

NO3

δH

2

δ O

[mg/l]

[mg/l]

[‰]

[‰]

SO4

F

Br

[mg/l]

[mg/l]

[mg/l]

-

-

18

CT boreholes
Bazma

N

501259

3724412

439,3

13,7

281,6

136,8

604,0

565,0

0,7

0,6

36,4

-47,3

-6,0

Bechelli 4

C

492487

3719580

301,4

10,0

217,5

95,8

429,3

401,3

0,7

0,5

32,7

-52,1

-6,2

Blidet 6

C

485701

3715046

290,3

11,7

228,2

101,3

520,4

521,5

0,8

0,5

25,0

-53,1

-6,1

Bou Abdallah 2

N

485655

3737590

666,4

20,1

657,9

240,0

854,9

1485,6

1,7

1,3

29,2

-

-

Chott Salhia 1

C

500366

3717523

280,9

9,2

190,1

83,6

399,4

371,8

0,7

0,5

36,2

-

-

Dar Kouskoussi 1 bis

N

497939

3729548

912,5

24,1

665,4

297,6

1229,5

1455,0

1,2

1,2

20,4

-45,2

-6,3

Douz 2 bis

S

494908

3703336

890,0

26,5

496,3

227,1

1600,5

1006,0

0,7

0,0

30,9

-

-

Douz Ouest

S

501320

3700841

1017,1

21,0

645,9

269,6

1165,2

1417,8

2,0

1,4

20,3

-51,3

-6,0
-6,5

El Gléa 3

N

486781

3736997

652,8

22,8

686,0

247,0

771,6

1400,8

1,6

1,2

23,5

-52,8

El Hsay 5 bis

S

494908

3703336

1714,5

31,1

793,7

357,1

2305,0

1257,3

1,0

2,0

31,2

-45,7

-5,5

Fatnassa 2

N

476116

3739473

546,9

22,5

463,3

157,7

878,9

1101,9

1,7

2,2

16,9

-50,4

-6,1

Ghidma 1

S

494908

3703336

266,3

11,3

180,8

80,5

436,0

416,4

0,7

0,5

36,8

-

-

Guettaya 4 bis

C

489630

3726877

309,7

10,6

261,8

119,8

507,5

560,4

0,9

0,5

19,8

-49,0

-6,0

Kelwamen

C

491628

3716641

389,2

12,2

298,1

134,2

497,1

472,9

0,8

0,5

35,7

-50,6

-6,2

Kelwamen (SONEDE )

C

490667

3716521

338,9

12,1

239,3

110,9

497,5

474,1

0,8

0,5

30,1

-

-

Klebia 2

C

488297

3707034

345,7

12,0

243,2

110,0

495,5

474,6

0,7

0,5

30,7

-50,4

-6,0

Ksar Tabeul 2

N

497815

3729008

672,3

18,6

481,6

214,4

956,9

1143,0

1,0

1,1

23,5

-

-

Mannsoura 2 bis

N

494120

3732695

389,2

14,5

327,3

151,0

573,4

740,7

0,8

0,6

20,9

-50,3

-6,0

Messaid 5

C

496731

3720075

330,7

11,3

232,2

105,2

449,6

426,7

0,7

0,5

33,8

-

-6,0

Negga 4 bis

N

484041

3733311

376,9

12,2

335,6

156,6

495,4

473,8

0,8

0,6

33,9

-49,6

Nouiel

C

488292

3707033

304,4

11,4

215,5

97,5

445,6

410,3

0,7

0,5

31,0

-

-

Oued Zira 3

N

490140

3735715

689,3

18,2

712,9

243,6

761,1

1447,8

1,9

1,2

23,5

-49,1

-6,7

3

S

473419

3705492

269,3

15,7

162,7

75,6

342,9

355,6

1,1

0,4

25,9

-49,7

-5,7

3

C

499701

3708107

233,3

14,4

49,6

5,1

257,3

43,7

0,4

0,6

38,8

-

-

PZ Dargine el Ameur
PZ Graad
Rahmat Foret

C

500191

3720715

342,6

14,0

239,1

109,2

440,1

406,2

0,7

0,5

38,5

-53,4

-5,9

Rahmat Sonede

C

500043

3721530

307,9

13,0

202,1

92,2

609,9

622,9

0,7

0,7

31,7

-48,9

-5,9

Scast 7

C

502018

3718463

609,3

17,3

408,4

184,6

799,9

903,7

0,6

0,9

33,1

-51,5

-6,7

Taouergha

N

494908

3703336

430,5

19,0

310,1

241,9

917,0

1757,7

1,4

1,3

nn

-55,4

-6,4

Telmine 3 bis

N

492180

3730799

474,5

15,8

360,2

161,3

621,8

669,8

0,9

0,7

27,9

-

-

Tifout

N

491576

3729535

302,3

10,7

241,4

112,9

497,1

517,2

0,8

0,5

23,2

-51,8

-6,1

Zaafrane 1

S

491358

3700400

264,9

9,2

177,8

78,6

346,1

338,1

0,6

0,4

38,5

-

-

33

1

Table 4 (continued): Results of hydrochemical and isotope sampling campaign 2002 in the Nefzawa oases.
Sample name

Nefzawa
region

1

2

Coordinates
East
North
[m]
[m]

Na

+

+

K

Ca

[mg/l]

[mg/l]

2+

[mg/l]

2+

Mg

[mg/l]

Cl

-

2-

-

-

SO4

F

Br

[mg/l]

[mg/l]

[mg/l]

[mg/l]

NO3

δH

2

δ O

[mg/l]

[‰]

[‰]

-

-

18

Drains
Drain Douz

S

501534

3700669

276,9

87,4

850,8

439,1

4697,5

2309,2

2,8

5,1

0,0

-

Drain El Hsay

S

500146

3696139

1526,5

39,2

1042,3

338,1

3187,8

2225,6

2,9

2,5

15,7

-

-

Drain Kelwamen

C

490920

3716601

1155,4

53,4

916,2

325,6

1844,5

2090,1

2,2

9,5

nn

-52,2

-5,8
-

Drain Negga

N

484315

3733397

1559,9

90,5

1389,5

675,4

488,2

929,3

1,5

10,5

4,7

-

Drain Souk el Bayez

C

497950

3729063

6605,6

245,5

3301,7

0,0

268,5

2839,0

0,0

4,8

nn

-

-

Drain Telmine

N

491576

3729535

1305,6

54,0

1746,1

557,0

3078,7

2255,8

1,5

3,4

nn

-

-

Drain Zaafrane

S

491458

3701457

1127,5

51,5

859,5

398,2

1405,3

2120,0

3,3

1,4

3,2

-

-

Puit Surface Douz NO

S

501521

3702845

998,6

69,7

231,9

1201,4

1552,2

2,1

1,4

22,8

-48,6

-5,1

Wells
457,2

Puit Surface Douz SO

S

504054

3701307

937,4

30,7

400,7

158,6

1145,6

1067,1

1,5

1,8

25,8

-

-

Puit Surface Kebili Est

N

498102

3728438

1599,7

27,5

844,0

285,6

972,0

1868,4

2,2

0,0

3,5

-48,0

-5,5

Jemna CI-11

C

501224

3712763

632,0

41,4

323,2

76,8

899,2

752,8

0,4

1,8

nn

-81,1

-8,6

Kebili CI-10

N

498369

3728982

1763,2

122,6

1351,1

449,2

1641,2

2186,4

1,9

1,9

nn

-64,0

-8,3

Tenia CI 15

S

527565

3694688

1442,0

97,2

624,0

297,0

2059,0

2472,0

0,6

1,6

nn

-55,3

-7,7

CI boreholes

Chotts and Sebkhas

2
3
4
5
6

Chott el Djerid

-

446101

3756878

130256,0

6528,2

884,5

3370,0

187104,0

4725,0

0,0

82,1

61,5

-

-

Chott el Franigue

S

456306

3697012

273,6

9,2

199,8

86,2

401,1

371,7

0,6

0,6

44,4

-

-

Sebkha Douz-Laala

S

492274

3704959

15466,8

1206,0

3034,4

11909,0

20020,0

9251,4

0,8

7,1

56,3

-

-

Note:
1 - Each sample is assigned to a certain area within the Nefzawa oases region: N = North, C = Central, S = South. Exception: The Chott el Djerid is a feature of all
areas.
2 – Projection: UTM, Date: WGS 84, Zone: S 32
3 – PZ marks observation boreholes. These boreholes are not used for public supply

34

1

Bibliography

2
3
4
5
6

ARMINES, and ENIT. 1984. Modèle mathématique du Complexe Terminal Nefzawa-Djérid.
Tunis: Ministère de l'Agriculture.
Ben Dhia, H., and S. Bouri. 1995. Overview of Geothermal Activities in Tunisia. Paper read at
World Geothermal Conference 1995, at Florence, Italy.

7

Chiang, Wen-Hsing, and Wolfgang Kinzelbach. 2001. 3D-groundwater modeling with PMWIN a

8

simulation system for modeling groundwater flow and pollution. Berlin: Springer.

9
10
11

Direction Générale des Resources en Eau. 2000. Annuaire de la piézométrie des nappes de
Tunisie, edited by Ministère de l'Agriculture: Tunis.
Edmunds, W.M., P. Shand, A.H. Guendouz, A.S. Moulla, A. Mamou, and K. Zouari. 1997.

12

Recharge characteristics and groundwater quality of the grand Erg oriental basin, edited

13

by B. G. Survey. Keyworth, Nottingham.

14

Fahem, Tobias. 2003. Salinisation of Groundwater in the Nefzaoua Oases - South Tunisia,

15

Institute of Hydromechanics and Water Resources Management, Swiss Federal Institute

16

of Technology, Zürich.

17

Food and Agricultural Organization. 2005. FAOSTAT: Food and Agricultural Organization.

18

Gardner, G. 1996. Worldwatch Paper #131: Shrinking Fields: Cropland Loss in a World of Eight

19

Billion. In Worldwatch Papers, edited by T. W. Institute. Washington D.C.: The

20

Worldwatch Institute.

21

Ghassemi, F., A.J. Jakeman, and H.A. Nix. 1995. Salinisation of land and water resources:

22

human causes, extent, management and case studies. Edited by A. N. University.

23

Wallingford, Oxon UK: Wallingford, Oxon, UK : CAB International ; Canberra, ACT,

24

Australia : In association with Centre for Resource and Environmental Studies, Australian

25

National University, 1995.

26
27

Gueddari, M. 1980. Géochimie des els et des saumures du Chott el Djérid, Univ. de Paul
Sabatier, Toulouse.

28

Harbaugh, A. W., E. R. Banta, M. C. Hill, and M. G. McDonald. 2000. Modflow-2000, the U.S.

29

Geological Survey modular ground-water model - User guide to modularization concepts

30

and the Ground-Water Flow Process, edited by U. S. G. Survey. Reston, Virginia: U.S.

31

Geological Survey.
35

1

Hillel, D. 2000. Salinity Management for Sustainable Irrigation. Integrating Science,

2

Environment and Economics. In Environmentally and Socially Sustainable Development -

3

Rural Development, edited by T. W. Bank. Washington D.C.: The World Bank.

4

Kbir-Ariguib, N., D. B. Chehimi, and L. Zayani. 2001. Treatment of Tunisian salt lakes using

5
6
7
8
9
10
11
12
13
14
15
16
17

solubility phase diagrams. Pure and Applied Chemistry 73 (5):761-770.
Kriaa, S. 2003. Contribution à l'étude de la qualité des eaux de la nappe du Complexe Terminal
dans la Nefzaoua, Ecole Nationale d'Ingenieurs de Tunis, Tunis.
Mamou, A. 1990. Caractéristics et évaluation des ressources en eau du sud Tunisien, edited by
Dissertation. Paris: Université du paris-Sud Centre d'Orsay.
Mamou, A., and A. Hlaimi. 1999. Les nappes phréatiques de la Nefzaoua, edited by Direction
Générale des Ressources en Eau: Ministère de l'Agriculture, Tunis.
Meckelein, W. 1977. Geographische Untersuchungen am Nordrand der tunesischen Sahara.
Stuttgarter Geographische Schriften, 247-301.
Observatoire du Sahara et du Sahel (OSS). 2003. The North West Sahara Aquifer System: Joint
Management of a Transborder Basin. Tunis: OSS.
Perennes, J.-J. 1993. L'eau et les hommes au Maghreb: Contribution d'une politique de l'eau
Mediterranee. Editions Karthala ed. Paris.

18

Pouget, M. 1966. Etude pédologique des oasis de la PIK et du groupe Mansoura, edited by

19

Direction Generale des Resources en Eau: Ministère de l'Agriculture, Tunis.

20
21
22

UNESCO. 1972. Etude des Ressources en Eau du Sahara Septentrional, Final report and
appendixes. In ERESS, edited by UNESCO. Paris: UNESCO.
Whittemore, D. O. 1984. Geochemical identification of salinity sources. In Salinity in

23

Watercourses and Reservoirs( Proceedings of the International Conference on State-of-

24

the-Art Control of Salinity), edited by R. H. French. Stoneham, MA: Ann Arbor Science,

25

Butterworth Publishers.

26

Whittemore, D. O. 1988. Bromide as a tracer in ground-water studies: Geochemistry and

27

analytical determination. In Proceedings Ground Water Geochemistry Conference, edited

28

by National Water Well Association. Dublin.

29
30
31
32

WHO-UNEP. 1989. Chapter 10. In Global fresh-water quality A first assessment, edited by M.
Meybeck, D. Chapman and R. Helmer: Basil Blackwell.
Zheng, C., and P.P. Wang. 1999. A modular three-dimensional multi-species transport model for
simulation of advection, dispersion and chemical reactions of contaminants in
36

1

groundwater systems; documentation and user’s guide. In SERDP-99-1. Vicksburg, MS:

2

U.S. Army Engineer Research and Development Center Contract Report.

3
4

37


Documents similaires


siegfried2007salinization
index cfm
health hazard assessment of water hardness
transboundary massacre aquifer iah 2012
health risk assessment of fluoride
reconnaissance des ressources en eau du bassin d ouarzazate


Sur le même sujet..