Environ Earth Sci DOI 10.1007/s12665-013-2241-2

ORIGINAL ARTICLE

Groundwater management based on GIS techniques, chemical indicators and vulnerability to seawater intrusion modelling: application to the Mahdia–Ksour Essaf aquifer, Tunisia Salwa Saidi • Salem Bouri • Hamed Ben Dhia

Received: 18 November 2011 / Accepted: 11 January 2013  Springer-Verlag Berlin Heidelberg 2013

Abstract In arid and semi-arid countries worldwide, conflicts between human development activities and conservation of groundwater resources are widespread and attract many public debates. This research aims to propose groundwater management alternatives for a coastal aquifer by studying its vulnerability and in particularly the risk of seawater intrusion. An additional objective is to propose some agricultural policies aimed to conserve groundwater resources in Mahdia and Ksour Essaf. Intensive groundwater mining, for irrigation and for water drinking, has caused an overexploitation of the water resources. In addition, the degradation of water quality, caused by septic tanks and intensive agricultural activities, has given rise to notable crucial state of the groundwater resources. With the aim of tackling the groundwater degradation problem, integration into a common platform of vulnerability assessment, seawater intrusion modelling and hydrochemical analysis is proposed. This platform can considerably reflect the water resources state in order to propose some solutions reducing the contamination of the Mahdia–Ksour Essaf aquifer. The groundwater management alternatives, proposed in this study, were prepared within a geographical information system. Keywords Groundwater management  Hydrochemical indicators  Vulnerability  Seawater intrusion modelling  Mahdia–Ksour Essaf aquifer S. Saidi (&)  S. Bouri  H. B. Dhia Water, Energy and Environment Laboratory (LR3E), ENIS.Bpw 1172, Sfax 3038, Tunisia e-mail: [email protected] S. Saidi Geological Department, University of Sciences El Manar, Tunis 2092, Tunisia

Introduction Actually, Fresh groundwater stored in coastal aquifers constitutes an important resource for humans and the natural environment. However, many coastal aquifers are vulnerable to seawater intrusion which can significantly degrade water quality and reduce freshwater availability. The increasing demands for fresh water in coastal areas and the anticipated impacts of climate change (such as sea-level rise and variations in rainfall recharge) may result in increases in the incidence and severity of seawater intrusion (Werner 2010). In the case of this study, groundwater is under other pressures which include intensive agriculture and change of lifestyle. These pressures can cause severe degradation of the groundwater quality. The most efficient measure to protect groundwater from overexploitation and pollution is to make a plan for integral prevention. Therefore, it is important to identify the most vulnerable areas which require a water management plan (Saidi 2011). In recognition of the need for efficient methods for protecting groundwater resources from future contamination, and especially from seawater intrusion, resource managers have sought to develop different models and methods for determining which areas are more contaminated like GALDIT method (Agarwadkar 2005). The specificity of this method is the integration of the groundwater analysis and the existing seawater events. So, vulnerability mapping can provide a representative understanding of the state of the monitored area based on hydrogeological characteristics. But, it cannot provide an idea about water quality contaminants’ loads and water pollution. A combination of both vulnerability and water quality can be considered as an efficient tool for groundwater management (Saidi et al. 2009). Furthermore, this study focuses on agricultural policies which can play an efficient role in controlling water

123

Environ Earth Sci

resources exploitation and contamination. In fact, some researchers believe that approximately 50 % of potential water saving derives from water management practices (Shangguan et al. 2002). To reach this aim, some measures were undertaken: (1) Vulnerability overview of the study area based on the transmit time of contaminants using AVI method and physical and chemical characteristics in order to analyse the seawater intrusion state of the aquifer using GALDIT method (2) criteria for the groundwater management assessment on the basis of vulnerability maps, hydrochemical data and socio-economical parameters. Study area General characteristics The Mahdia–Ksour Essaf aquifer is situated in north eastern part of the Mahdia region which is on the eastern coast

of Tunisia (Fig. 1). It covers about 600 km2. It is also characterized by a semi-arid climate with an annual precipitation of 308 mm, an annual temperature of 19 C and a potential evapotranspiration rate of 1118 mm/year (INM 2007). Geology and hydrogeology The study site is located on an alluvial plain whose geology is dominated by Quaternary deposits (Fig. 2). The Mahdia and Ksour Essaf areas have relatively stable tectonics apparent in the tabular sedimentary structure. The coastal part (both of the aquifer and the vadose zone) of the Mahdia–Ksour Essaf region is located in Plio-Quaternary layer system which is constituted mainly by alluvial fan, gravel and sand with high permeability (Saidi 2011) (Fig. 3). Therefore, it results in an easy infiltration of contaminants and salt water in the groundwater.

Fig. 1 Location of the study area and the anthropogenic activities in the Mahdia and Ksour Essaf Cities

123

Environ Earth Sci

Fig. 2 Geological map of the Mahdia region (Saidi 2011)

The hydrostratigraphic cross section shows that the PlioQuaternary detrital deposits constitute a potential shallow aquifer. The Quaternary recharge is increased partly by Pliocene sediments infiltration. Part of the groundwater flow occurs within the Tyrrhenian gravel deposits and Quaternary recent dunes before reaching the sea. Indeed, these deposits constitute a good reservoir and a key point of coastal aquifer recharge (Fig. 3). The aquifer has known overexploitation, which exceeds the exploited water resources of the aquifer, since 1995 (CRDA 2005). It has an estimated safe yield of 2.9,106 m3/yr but annual abstraction by pumping from 1,784 wells stands at 4.28 106 m3/yr (Saidi 2011). The recent 2007 piezometric map (Fig. 4) shows piezometric depressions related to overexploitation of the aquifer with negative piezometric levels on the order of -4 m which corresponds to the concentration of wells (Fig. 1). These areas correspond to high salinity. So, there is also evidence to suggest that the quality of groundwater supplies is under threat. Thus, salinity levels generally range from 0.6 to 3 g/L in the coastal aquifer, and exceed 7 g/L in some areas in the coast of the study area (CRDA 2007a, b). Hydrochemistry Water samples were collected from 25 wells (Fig. 1) over the study area. An additional sampling campaign is done to capture the variations in the different water quality parameters and the piezometric level. The geographical

coordinates of the wells were captured using handheld Garmin e-Trex GPS receiver. Water level data were also collected at each well during the sampling campaigns using Automatic Water level Recorders (AWLR). The water quality parameters analysed include salinity, bromide (Br-), sodium (Na?), magnesium (Mg2?), sulphate (SO2-) 4 , chloride (Cl-) and nitrate (NO3) (Tables 1, 2 and Fig. 1). Geoelectrical study Electrical resistivity data have been used to identify the geographical extent of salinization in the aquifer, as resistivity decreases with increasing salinity. The cross sections were developed from the interpreted layer thickness and sensitivities (Figs. 5, 6). The southwest– northeast cross section indicates that there are low resistivity zones associated to the areas surrounding the sebkhas and near the coast. Water quality data indicate that the low resistivity zones are a natural upcoming of mineralized waters high in sulphate derived from intergranular anhydrite and sebkhas evaporates near the aquifer. But the low resistivity zones adjoining the coast, however, are not the result of sebkha contaminations only and have higher chloride ion concentrations, indicative of seawater intrusion. These areas have a low resistivity wedge (0.9 Xm) extending from the sebkha to the Coast of the Mediterranean Sea. Water quality data indicate that groundwater in this zone is elevated in chloride ion content suggesting seawater intrusion.

123

Environ Earth Sci Fig. 3 Lithostratigraphic and hydrostratigraphic charts of study area (Saidi 2011)

Given these geophysical findings, a geochemical study was undertaken and the following presents the main results and interpretations. In fact, salinity measures in some samples near the coast exceed 6 g/l and suggest a seawater intrusion in this area (Saidi 2011). Anthropogenic activities Also, the study region has recently known an expansion of anthropogenic activities especially agricultural ones. In fact, the district has experienced a high increase in the

123

number of people engaged in agricultural activities (25 % of the total population implicated in these activities; Amari and Maaloug 2004). Indeed, the study area is mostly used for intensive cultivation of crops such as cereals and olive. The use of inorganic fertilizers on these crops has a great polluting effect, which adds to the environmental pressure on the groundwater of the Mahdia–Ksour Essaf aquifer (Saidi 2011). Furthermore, the septic tank and waste water effluents constitute additional factors of contamination of the Mahdia–Ksour Essaf groundwater (Fig. 1).

Environ Earth Sci

Fig. 4 Isopiestic map of the Mahdia–Ksour Essaf aquifer (Saidi 2011)

Methodology Groundwater vulnerability assessment Intrinsic vulnerability The name stands for Aquifer Vulnerability Index; AVI method was developed to estimate the aquifer vulnerability to contamination from potential sources at or near the ground surface (Van Stempvoort et al. 1993). This aquifer vulnerability assessment only requires knowledge of the local stratigraphy and the hydraulic conductivity (k) for each layer of the vadose zone. The AVI index, a transit times, is calculated from the thickness (h) and the

conductivity values (k) of each layer, above the water table. The index was calculated for each of the wells. After that data were interpolated using the inverse distance moving average technique. Data are presented in log of years. Short AVI times, in days, represent high vulnerability and long AVI times, in years, correspond to low vulnerability. So, the hydraulic resistance C is referred to a combination of two parameters: the thickness h of the confining layer above an aquifer and the vertical hydraulic conductivity k. It corresponds to an estimation of the travel time of a contaminant through the unsaturated zone (Van Stempvoort et al. 1993). The hydraulic resistance, in years or days, is calculated by means of the following expression:

123

Environ Earth Sci

C ¼ Rinoflayers hi=ki:

Table 1 Chemical analysis of the Mahdia–Ksour Essaf aquifer -

No. of well

HCO3

Cl

1

189

2

250

3 4

-

2-

-

2-

-

-

Cl /Br

-

SO4

Br

SO4 / Cl-

Cl / HCO3-

1,125

1,088

0.8

0.967

5.952

1,406.25

962

978

1.8

1.017

3.848

534.44

173 208

943 1,921

1,580 1,097

2 –

1.676 0.571

5.451 9.236

471.50 –

5

271

1,308

1,160

3

0.887

4.827

436.00

6

250

1,099

1,560

–

1.419

4.396

–

7

159

3,407

2,033

–

0.597

21.428

–

8

284

1,292

1,791

1.2

1.386

4.549

1,076.67

9

175

1,336

2,400

–

1.796

7.634

–

10

326

566

1,186

–

2.095

1.736

–

11

336

684

1,461

1.5

2.136

2.036

456.00

12

348

1,266

1,505

1.5

1.189

3.638

844.00

13

242

1,215

2,264

–

1.863

5.021

–

14

470

1,134

3,033

–

2.675

2.413

–

15

232

1,787

5,862

–

3.280

7.703

–

16

128

4,018

6,592

–

1.641

31.391

–

17

242

2,355

3,129

–

1.329

9.731

–

18 19

172 161

261 2,834

4,255 5,307

1.2 –

16.303 1.873

1.517 17.602

217.50 –

20

300

1,188

1,187

–

0.999

3.960

–

21

153

967

3,076

–

3.181

6.320

–

22

202

200

5,142

–

25.710

0.990

–

23

235

200

3,373

–

16.865

0.851

24

221

234

3,943

7.2

16.850

1.059

32.50

25

193

530

3,670

1

6.925

2.746

530.00

–

where (C) is the hydraulic resistance, hi and ki are, respectively, the thickness of the unsaturated zone and the hydraulic conductivity. The hydraulic resistance characterizes the resistance of an aquifer to vertical flow, either upward or downward. While it has the dimension of time, it does not represent the travel time of water or contaminants. The time for water to flow through a confining layer further depends on the porosity and vertical hydraulic gradient. Additional factors such as diffusion, density, decay and sorption will have to be taken into account when considering migration of a contaminant. Representative hydraulic conductivity values are listed in Table 3. To facilitate plotting and contouring of the hydraulic resistance data, the AVI has been defined as: AVI ¼ 10 Log (CÞ

123

Num

Nitrate (mg/l)

1

67

2

126

3

37

4

150

5

0.6

6 7

52 0.1

8

ð2Þ

where C is the hydraulic resistance Vulnerability to seawater intrusion modelling: The GALDIT method is chosen to evaluate the aquifer vulnerability to seawater intrusion in Mahdia–Ksour Essaf aquifer because it takes into account the hydrogeological conditions which can mainly influence the seawater intrusion potential in this zone. The physical characteristics that affect the seawater intrusion potential are also inherent in each hydrogeologic setting. The most important factors that control seawater intrusion are found to be the following: •

Table 2 Nitrate concentrations (Saidi 2011)

ð1Þ

Groundwater occurrence (aquifer type: unconfined, confined or leaky confined): Naturally, groundwater aquifers can be categorized as confined, unconfined, or leaky confined. The nature of groundwater occurrence has a high influence on the extent of seawater intrusion. Thus, an unconfined aquifer under natural conditions will be more affected

Table 3 Hydraulic conductivity (Rodriguez et al. 2001) Aquifer lithology

Hydraulic conductivity (m/s)

0.2

Sand

1.4 9 10-9

9

1

Sandy clay

5.1 9 10-6

10

0.17

Sand

2.1 9 10-4

11

0.1

Clay with sand

5 9 10-5

12

56

Thin/moderate sand

1 9 10-3

13

48

Thin sand

6 9 10-4

14

2

Moderate sand

1 9 10-4

15

0

Gravel

1.4 9 10-3

Environ Earth Sci

•

•

•

•

•

by seawater intrusion than a confined aquifer which has a pressure higher than atmospheric pressure. Also the confined aquifer may be more prone to seawater intrusion than a leaky confined aquifer because confined one is more vulnerable due to larger cone of depression after pumping whereas leaky confined maintains minimum hydraulic pressure by way of leakage from adjoining aquifers. Hence, the latter has got least susceptibility to saltwater intrusion (Chachadi and Labo Ferreira 2005). Aquifer hydraulic conductivity: It is defined as the ability of the aquifer to transmit water. This parameter has a high influence on the magnitude of seawater front movement; the higher the conductivity, the higher is the inland movements of the seawater front. Depth of groundwater Level above the sea: it is a very important factor in evaluating seawater intrusion because it determines the hydraulic pressure availability to push the seawater front back. So according to Ghyben–Herzberg relation for every metre of freshwater stored above the mean sea level, 40 m of freshwater is stored below it, down to the interface (Fetter 1994). Distance from the shore (distance inland perpendicular from shoreline): the impact of seawater intrusion generally decreases as one move inland at right angles to the shore (Chachadi and Labo Ferreira 2005; Agarwadkar 2005). Impact of existing status of seawater intrusion in the area: It can be assessed using the ratio of SO42-/Cl(Saidi 2011). Thickness of the aquifer, which is being mapped: it plays an important role in determining the extent and the magnitude of seawater intrusion in the coast; the larger the aquifer thickness, the smaller is the extent of seawater intrusion and vice versa (Agarwadkar 2005).

Each of the six parameters has a predetermined fixed weight that reflects its relative importance to seawater intrusion. The GALDIT index is calculated by the following equation: GALDIT ¼ 1  G þ 3  A þ 4  L þ 2  D þ 1  I þ 2  T According to this index, it is possible to identify areas that are more likely to be susceptible to seawater intrusion than other areas; the higher the index, the greater is the seawater intrusion potential (Chachadi et al. 2003; Agarwadkar 2005). Chemical indicators Water samples were collected directly from 40 domestic wells, 2–3 m below the water table, by using mobile pumping stations. The analytical methods used to evaluate the water chemical composition are reported in Table 1.

Chloride to bromide mass ratio as an indicator of septic tank effluent in groundwater The impact of wastewater from septic tanks on groundwater quality is assessed in this study by comparing a targeted group of wells that have a septic tank influence with a notargeted group of wells that do not. These two groups were defined using Cl/Br mass ratios and chloride concentrations (Brian et al. 2011). So, Cl-/Br- ratios are used to distinguish wastewater (septic tank) sources from other potential sources of Cl and Br in groundwater (Table 1). Potential sources of nitrate contamination Nitrate is often seen as an agricultural pollutant of groundwater and so is expected to be at higher concentrations in the groundwaters surrounding a city than in those beneath it (Wakida and Lerner 2005). However, in this study another nitrogen sources are illustrated from septic tank and sewage systems from urban areas (Table 2). Potential sources of NO3 in the Mahdia–Ksour Essaf aquifer include imported water from the septic tank effluents, the fertilizers used in agriculture, the leaching of natural soil NO3 by recharged water, and NO3-rich irrigation-return flow. In fact, to evaluate these possible sources, water samples were collected from targeted wells which are near septic tank and compared with groundwater quality samples collected from wells far from septic tank and urban zones. Potential markers of the impact of existing status of seawater intrusion The area under mapping is invariably under stress which has already modified the natural hydraulic balance between seawater and fresh groundwater. This fact should be considered in the characterization of the pollution degree of the study area. The ratio of Cl-/(HCO3- ? CO32-) was used to show seawater intrusion and nine rates were assigned in terms of vulnerability to seawater intrusion degree (Chachadi and Labo Ferreira 2005) Table 4. Another ratio (SO42-/Cl-) was used as a natural tracer of the residence time of seawater intrusion into the coastal aquifers (Pulido-Leboeuf et al. 2003) (Table 5). Table 4 Used analytical methods Chemical element

Analytical method

Cl-(mg/l)

Titration with AgNO3

SO4-, 2 Br (mg/l)

Chromatography liquid phase

NO3- (mg/l)

Atomic absorption spectrometer

HCO3- (mg/l)

Titration with HCl

123

Environ Earth Sci Table 5 Impact of seawater intrusion indicators according the Chachadi and Labo Ferreira (2005) and Saidi (2011) classifications Note

Impact of the seawater intrusion Cl-/HCO3-

SO42-/Cl-

1

\1

4–25.71

2

1–1.5

3–4

3 4

1.5–1.75 1.75–2

2.75–3 2.5–2.75

5

2–2.25

2.25–2.5

6

2.25–2.5

2–2.25

7

2.5–2.75

1.75–2

8

2.75–3

1.5–1.75

9

3–4

1–1.5

10

4–31.39

\1

Table 6 Vulnerabilty index according the AVI method (classification of Van Stempvoort et al. (1993)) Vulnerability class

AVI index (days)

AVI index (log c)

Very low vulnerability

[200

[2.3

Low vulnerability

100–200

2–2.3

Moderate vulnerability

50–100

1.69–2

High vulnerability

10–50

1–1.69

Very high vulnerability

\10

\1

Agricultural policies It consists to analyze the impacts of the price of irrigation water and the change of some agricultural practices on the water conservation and preservation. The assumption of

this study is that an increase in irrigation water price, and the implementation of water saving-oriented agricultural policies under different scenarios, will encourage farmers to shift their crop mix away from high water consuming crops, and will in turn indirectly reduce groundwater exploitation for agricultural purposes (Hadipuro and Indriyanti 2009). Based on all synthetic documents and the agricultural policies, different Scenarios of groundwater management and practical solutions were proposed. Techniques Geographical information system (GIS) is a powerful tool and has great promise for use in environmental problem solving. Most environmental problems have an obvious spatial dimension and spatially distributed models can interact with GIS (Goodchild 1993). Troge (1994) reported that this computer-based tool has allowed successful integration of water quality variables into a comprehensible format. Hence, it is evident that many factors must be incorporated into vulnerability and risk evaluations, and GIS is a perfect tool for this kind of preliminary studies due to its ability to manage large volume of spatial data from a variety of sources (Saidi et al. 2010). In this study, GIS has been used in the interpolation of the different databases used in all methods’ parameters, using the analysis extension of Arc Gis9.3. Furthermore, in the map classification according to the rates and weights of each parameter, GIS is used as a database system in order to prepare maps of water quality according to concentration values of different chemical constituents (Skubon 2005; Yammani 2007). GIS is also utilized to prepare layers of maps to locate

Table 7 Weights and rates proposed for GALDIT model of Mahdia-Ksour Essaf aquifer

Parameters

G

A

L

D

I

T

Groundwater occurence

Aquifer hydraulic conductivity

Depth of groundwater level

Distance from the shore

Impact of existing status of seawater intrusion

Thickness of the aquifer

Cl-/HCO3-

SO42-/Cl-

Aquifer type

m/day

m

m

1

3 0.08–10

4 15\

2 1,000\

1 \1

4–25.71

2

8–15

800–1,000

1–1.5

3–4

3–4

3

5–8

700–800

1.5–1.75

2.75–3

4–6

4–5

600–700

1.75–2

2.5–2.75

6–8

3–4

500–600

2–2.25

2.25–2.5

8–10

400–500

2.25–2.5

2–2.25

10–12

300–400

2.5–2.75

1.75–2

12–14

200–300

2.75–3

1.5–1.75

14–16

Weight rate 1

4

10–15

5 6 7

15–20

8

Semi confined

9

Unconfined

10

Leaky confined

123

20\

m 2

100–200

3–4

1–1.5

16–20

71–100

4–31. 39

\1

20\

Environ Earth Sci Table 8 Vulnerability classes according to GALDIT method using SO24 /Cl ratio (Saidi 2011) GALDIT index

Vulnerability degree

31–50

No vulnerable

50–70

Low vulnerability

70–90

Moderate vulnerability

90–97

High vulnerability

risked zones. GIS can also help to calculate vulnerability indexes by using raster calculator of Arc GIS 9.3 and gvSIG 1.11 as open-source software. This software is also used in the creation of the management scenarios by superposition of different maps. It facilitates the application and makes it possible to update the vulnerability, the hazard map on a regular basis (Nguyet and Goldscheider 2006).

Fig. 5 Electrical section ‘‘A’’ oriented west–east (Saidi 2011)

Fig. 6 Electrical section ‘‘B’’ oriented south west–north east (Saidi 2011)

123

Environ Earth Sci

Results and Discussions Vulnerability assessment AVI method The resultant vulnerability map was subdivided into five classes in relation to each degree of vulnerability according to the classification of Van Stempvoort et al. (1993) (Table 6 and Fig. 7). The highest classes of vulnerability (AVI \ 50 days) cover 50 % of the total surface in the East of the study area. This fact is due to the high hydraulic conductivity and low thickness of the vadose zone sediments which is constituted by

a Tyrrhenian gravel shallow groundwater. So, it results in a low capacity to attenuate the contaminants. The low and very low vulnerability (AVI [ 100 days), which is represented by 30 % of the total Mahdia–Ksour Essaf surface, is essentially due to the thick sediments of the vadose zone and especially the low hydraulic conductivity as well as the low recharge rate and deep water table (Saidi 2011). The moderate vulnerability (50 \ AVI \ 100) covers the 20 % of the study area. GALDIT method The GALDIT map shows four classes. In fact GALDIT Index ranges from 26 to 96. The highest class of vulnerability covers

Fig. 7 Groundwater vulnerability of the Mahdia–Ksour Essaf aquifer using AVI method (Saidi 2011)

123

Environ Earth Sci

particularly the north eastern part and the eastern part of the aquifer (the coast). This is due to the high permeability (unconfined aquifer), the high hydraulic conductivity ([20 m/day), the shallow groundwater table (\8 m), the high Cl- concentration (hence a low SO42 /Cl ratio), the thin aquifer (\10 m) and relatively the low distance separating wells from the sea (\1,000 m) (Fig. 8). As a consequence, in this area we should not allow either additional wells or high risk activities in order to preserve groundwater resource and reduce environmental pollution hazard. However, the north western part is highly vulnerable in spite of the great distance from the coast, the higher groundwater level and the thicker

aquifer. This can be explained by the high SO42 /Cl ratio. Consequently, the southern part of the aquifer will be more suitable for the implantation of potential anthropogenic activities and the northern part should be considered by the managers in order to minimise groundwater contamination by seawater intrusion (Tables 7, 8 and Fig. 9).

Groundwater contamination assessment by waterchemistry data analysis The examination of the Cl-/Br- ratio shows that Cl- and Br- concentrations in groundwater near septic systems

Fig. 8 Six GALDIT maps to compute the vulnerability to seawater intrusion index

123

Environ Earth Sci

Fig. 9 Superposition of vulnerability to seawater intrusion using GALDIT method and nitrate distribution

vary temporally; however, Cl-/Br- ratios remained within or near the targeted range (530–1,406.25) (Table 1). In no target wells far from septic tank, Cl-/Br- ratios varied from 32.5 to 534.4 because of the localisation outside the target range and there was no apparent seasonal effect (Fig. 10). To enhance the use of Cl-/Br- ratios as a screening tool, additional chemical would be desirable. In fact, Miller et al. (1999) found that the co-occurrence of multiple constituents (bacteria, wastewater compounds, and elevated concentrations of nitrate, chloride, or boron) provided strong evidence of contamination from septic tank effluents. Hence, an analysis of chloride and nitrate was done. The observation of these analyses

123

shows a high concentration in targeted wells which exceeds 50 mg/l of NO3- and 1,000 mg/L of Cl-. However, some wells (wells no. 17 and 6) which are far from target ones have high chloride concentrations. This suggests other kind of contamination. For this reason SO42-/Cl- and Cl-/HCO3- ratios were calculated to see the impact of seawater intrusion. So, the observation of both SO42-/Cl- and Cl-/HCO3- reveals, in the majority of the study area and especially in the coastal part, a high ratio of Cl-/HCO3- (Cl-/ HCO3- [ 30) and low ratio of SO42-/Cl- (SO42-/ Cl- \ 1). This corroborates the contamination of the Mahdia–Ksour Essaf aquifer by seawater intrusion (Figs. 11 and 12).

Environ Earth Sci

High nitrate concentration was measured in no-targeted wells which are situated far from urban areas and from the influence of septic tank and industrial activities. But, these wells (wells no. 10, 13, 15 and 19) are situated in the vicinity and in the irrigated areas (Fig. 10). The high concentration is the consequence of the intensive application of fertilizers and especially the irrigation-return flow phenomena. Nitrate concentrations in drinking water are important because ingestion of nitrate by infants can cause low oxygen levels in the blood, a potentially fatal condition (U.S. Environmental Protection Agency 2006).

Water management solutions based on modelling scenarios The management scenarios proposed in this study should consider the interaction between the aquifer system and the different conditions such as external (water price, rate of exploitation, plant types, type of pollution…) and internal ones (permeability, water level, transit time of pollutants, aquifer conductivity…). Therefore, many alternatives should be considered. Each alternative, in turn, has various costs associated with it, thereby allowing decision-makers to determine which is the most effective in groundwater

Fig. 10 Bromides and Cl/Br ratio in the Mahdia–Ksour Essaf aquifer

123

Environ Earth Sci

Fig. 11 Cl-/HCO3- ratio in the Mahdia–Ksour Essaf aquifer (Saidi 2011)

management in the study area. This is done firstly by considering different factors affecting the aquifer status and secondly, by building different policy scenarios in order to propose practical solutions. The first step is to characterise the vulnerability degree based on the characteristics of the aquifer on one hand and the vulnerability to seawater intrusion and the effects of different prices on groundwater exploitation on the other hand. The second step is to consider the different policy scenarios in order to analyze the impacts of economical, agricultural, institutional, environmental instruments and aquifer characteristics on the agricultural sector and the water quality. This

123

can help to propose some conservation suggestions and efficient solutions in the study area. The baseline scenario indicates the existing condition of total groundwater demand by irrigated lands and drinking water consumption near to 6.5 Mm3. By continuing the existing condition of groundwater extraction through agriculture sector and drinking uses, there will be a negative volume fluctuation of groundwater which exceeds largely the safe yield of the aquifer. In this way the best solution is to increase irrigation water price and to implement water saving-oriented agricultural policies which will encourage farmers to shift their crop mix away from high

Environ Earth Sci

Fig. 12 SO42-/Cl- ratio in the Mahdia–Ksour Essaf aquifer (Saidi 2011)

water consumption crops. Therefore, it will in turn indirectly reduce groundwater exploitation for agricultural purposes. Note that in the total cost, the exploitation share is only 0.011–0.028 Tunisian Dinars per cubic metre (0.006–0.015 €). The price paid by the farmers for the dam irrigation water is 0.110–0.134 Tunisian Dinars per cubic metre, i.e. 0.06–0.075 €/m3 (Bouri and Ben Dhia 2010). So, this low water prices and the lack of payments by farmers for groundwater usage has encouraged farmers to inefficiently increase the number of wells and irrigated farms, with no consideration for groundwater resource conservation. Consequently, the groundwater level has decreased

continuously in recent decades, threatening the life of groundwater aquifers in this area. However, some zones require special suggestions and this is with consideration of the characteristics of the study area. For these reasons some alternative scenarios were proposed in order to meet the demands of sustainable groundwater management in the study area (Fig. 13): Scenario 1 In the east and especially the coastal part of the study area, which is characterized by high vulnerability, we should not allow additional high risk activities in order to obtain economic advantage and reduce environmental pollution hazard. We should not implant wells to avoid the seawater intrusion and the contamination of the aquifer. So,

123

Environ Earth Sci

in the areas affected by seawater intrusion (Fig. 9), a reverse-osmosis treatment facility and the artificial recharge can be used to brackish water treatment. We must ensure high distance between domestic wells and septic tank to avoid water contamination because of the high permeability of these areas. Previous studies show a positive correlation between septic tank density (in a 500 m radius buffer area around each well) and various chemical constituents. The use of a circular area (500 m radius circle) around each well is based on previous investigations that found significant correlations between groundwater quality and a variety of hydrogeologic and anthropogenic factors within circular buffer areas (Nolan and Hitt 2006; Squillace and Moran (2007)). So, in this scenario, we

advise to avoid in further development projects the implantation of new wells in the buffer zone for the security and safety of people. Scenario 2 The areas with moderate and low vulnerability are suitable for the implantation of feature wells to human consumption. But, we should not allow excessively exploitation to protect and conserve groundwater resources. In addition, we should substitute the most water intensive crops, which are also normally more labour intensive, by others with reduced water and labour requirements. In the irrigated areas we should also limit the usage of fertilizers in order to reduce the nitrate contamination of the groundwater. If they do, this would further support the idea that opportunities exist to improve environment and water

Fig. 13 Water management scenarios of the Mahdia–Ksour Essaf aquifer

123

Environ Earth Sci

quality without large sacrifices in technical efficiency (Andres and Gerald 2011). In the proposition of these scenarios different approaches and methods were used. The benefit of the application of several methods lies in the possible identification of processes that could erroneously affect the interpretation of data of a single method (Corcho Alvarado et al. 2005).

2. 3.

Limitations The combination of the results of the aquifer vulnerability method and the chemical data can be useful for proposing some management scenarios and projecting aquifer responses to various changes in aquifer stresses. However, a model is a highly idealized approximation of the actual system and it is based on average and estimated conditions. Perhaps the biggest limitations are the constraints and errors associated with the interpolation technique. The capability of the model to reliably project aquifer responses is also related to the accuracy of the input data used in the vulnerability mapping and the stresses related to socioeconomic development, which affect the study area.

Conclusions Using AVI method, 70 % of the total surface of the Mahdia–Ksour Essaf aquifer presents medium to very high vulnerability, which makes it susceptible to pollution and degradation by superficial sources of contamination. This fact was confirmed by chemical analysis and GALDIT method which demonstrate that the coastal part of the aquifer is highly susceptible to seawater intrusion. There is a high similarity between nitrate distribution map and vulnerability ones. In fact, areas with high vulnerability present high nitrate concentrations. Consequently GALDIT and AVI are validated. In addition, the aquifer suffers from other contamination deriving from septic tank effluents and seawater intrusion which are verified by Cl/Br, HCO3/Cl and SO42-/Cl-. This state is more drastic in the presence of some bad agricultural policies. So synthetic documents produced in this study (Figs. 7, 9, 13) recognize the current water quality of shallow groundwater wells in the Mahdia–Ksour Essaf and the impact of septic tank, seawater intrusion and agricultural practices on shallow groundwater quality. These data are useful to make future monitoring comparisons and to identify areas of potential concern for further investigation. However, to be more complete, this study should include other assessments. In particular it is recommended that: 1.

over-exploitation of the groundwater through installation of private boring wells needs to be strongly

4.

discouraged and this especially in the cosatal part of the study area; the vulnerability assessments should involve detailed socio-economic impacts; the interface freshwater saltwater can be pushed down by application of artificial recharge which implies that the reserve for temporal overdraft becomes greater. So, a detailed study to calculate the quantity of water, to be recharged artificially to push the interface towards the sea in high vulnerability zones, should be undertaken; the use of other kinds of water, such as recycled wastewater, as an alternative source of water supply for irrigation, tourism and industrial activities in order to minimise additional pumping of fresh groundwater in the coast.

References Agarwadkar YY (2005) Salinity mapping in coastal area using GIS and remote Sensing. Thesis report, Indian Institute Of Remote Sensing, National Remote Sensing Agency (NRSA), Department Of Space, Dehradun, India, p 65 Amari S, Maaloug I (2004) Portrait du gouvernorat de Mahdia, Institut Supe´rieur des Etudes technologiques de Mahdia. Rapport ine´dit, p 53 Andres GF, Gerald ES (2011) How might shadow price restrictions reduce technical efficiency? Evidence from a restricted DEA analysis of coffee farms in vietnam. J Agric Econ 62:47–58. doi: 10.1111/j.1477-9552.2010.00269.x Bouri S, Ben Dhia H (2010) A thirty-year artificial recharge experiment in a coastal aquifer in an arid zone: the Teboulba aquifer system (Tunisian Sahel). C R Geosci 342:60–74 Brian GK, Sandra ME, Leon JK (2011) Using Cl/Br ratios and other indicators to assess potential impacts on groundwater quality from septic systems: a review and examples from principal aquifers in the United States. J Hydrol 397:151–166 Chachadi AB, Labo Ferreira JP (2005) Assessing Aquifer vulnerability to sea-water intrusion using GALDIT method: Part 2— GALDIT Indicators Description. In: Proc of the Fourth Inter— Celtic Colloquium on Hydrology and Management of Water Resources. Guimaraes, Portugal, pp 12 Chachadi AG, Lobo Ferreira JP, Noronha L, Choudri BS (2003) Assessing the impact of sea-level rise on salt water intrusion in coastal Aquifers using GALDIT model. APRH/CEAS. In: ´ guas Subterraˆneas, Lisboa, Fev. Processing Semina´rio Sobre A 2003, pp 13 Corcho Alvarado JA, Purtschert R, Hinsby K, Troldborg L, Hofer M, Kipfer R, Aeschbach-Hertig W, Arno-Synal H (2005) 36Cl in modern groundwater dated by a multi-tracer approach (3H/3He, SF6, CFC-12 and 85Kr): a case study in quaternary sand aquifers in the Odense Pilot River Basin, Denmark. Appl Geochem 20:599–609 Fetter CW (1994) Geology of Ground Water occurrence, in chapter 9, Applied Hydrology, Third Edition, Prentice Hall, pp 370–371 Goodchild MF (1993) The state of GIS for environmental problemsolving. In: Goodchild MF, Parks BO, Steyaert T (eds) Environmental modeling with GIS. Oxford University Press, New York, pp 8–15

123

Environ Earth Sci Hadipuro W, Indriyanti NY (2009) Typical urban water supply provision in developing countries: a case study of Semarang City, Indonesia. Water Policy 11:55–66 INM (Institut National de Me´te´o) (2007) Tableaux climatologiques mensuels, stations de Mahdia (unpubl.). Annuaire de l’Institut de la Me´te´orologie Nationale, Tunis, Tunisia CRDA (Commissariat re´gional du de´veloppement agricole de Mahdia) (2005) Annuaires d’exploitation des nappes phre´atiques du gouvernorat de Mahdia, CRDA Mahdia, Tunisia CRDA (Commissariat re´gional du de´veloppement agricole de Mahdia) (2007a) Comptes rendus des forages et pie´zome`tres de surveillance (unpubl.). Arrondissement des Ressources en Eaux de Mahdia. CRDA, Mahdia, Tunisia CRDA (Commissariat re´gional du de´veloppement agricole de Mahdia) (2007b) Annuaires de surveillance de pie´zome´trie, de nitrate et de salinite´ (unpubl.). Arrondissement des Ressources en Eaux de Mahdia. CRDA, Mahdia, Tunisia Miller MW, Hay ME, Miller SL, Malone D, Sotka EE, Szmant AM (1999) Effects of nutrients versus herbivores on reef algae: a new method for manipulating nutrients on coral reefs. Limnol Oceanogr 44:1847–1861 Nguyet MTV, Goldscheider N (2006) A simplified methodology for mapping groundwater vulnerability and contamination risk, and its first application in a tropical karst area, Vietnam. Hydrogeol J 14:1666–1675 Nolan BT, Hitt KJ (2006) Vulnerability of shallow ground water and drinking water wells to nitrate in the United States. Environ Sci Technol 40:7834–7840 Pulido-Leboeuf P, Pulido-Bosch A, Calvache ML, Vallegos A, Andreu JM (2003) Strontium, SO42-/Cl - and Mg2?/Ca2? ratios as tracers for the evolution of seawater into coastal Aquifers: the example of Castell de Ferro Aquifer (SE Spain). C R Geosci 335:1039–1048 Rodriguez R, Reyes R, Rosales J, Berlin J, Mejia JA, Ramos A (2001) Estructuracion de mapas tematicos de indices de vulerabilidad acuifera de la mancha urbana de Salamanca Guanajuato. Technical report, Municipio de Salamanca, CEAG, IGF-UNAM, p 122

123

Saidi S (2011) Contribution des approches parame´triques, cartographiques et statistiques a` l’e´tude de la vulne´rabilite´ du syste`me aquife`re phre´atique de Mahdia (Tunisie Orientale), The`se de doctorat. Universite´ de Sfax, Tunisie, pp 198 Saidi S, Bouri S, Ben Dhia H, Anselme B (2009) A GIS-based susceptibility indexing method for irrigation and drinking water management planning: application to Chebba–Mellouleche Aquifer, Tunisia. Agric Water Manag 96:1683–1690 Saidi S, Bouri S, Ben Dhia H (2010) Groundwater vulnerability and risk mapping of the Hajeb-Jelma Aquifer (Central Tunisia) using a GIS-based DRASTIC model. Environ Earth Sci 59:1579–1588 Shangguan Z, Shao M, Horton R, Lei T, Qin L, Ma J (2002) A model for regional optimal allocation of irrigation water resources under deficit irrigation and its applications. Agric Water Manag 52:139–154 Skubon BA Jr (2005) Groundwater quality and GIS investigation of a shallow sand aquifer, Oak opening region, North West Ohio. Geol Soc Am Abstr Prog 37(5):94 Squillace PJ, Moran MJ (2007) Factors associated with sources, transport, and fate of volatile organic compounds and their mixtures in aquifers of the United States. Environ Sci Technol 41:2123–2130 Troge (1994) A GIS strategy for lake management issues, (National Conference on Environmental Problem-Solving with Geographic Information Systems, Cincinnati, Ohio, USA U.S. Environmental Protection Agency. (2006) Drinking water contaminants. http://www.epa.gov/safewater/contaminants/ Van Stempvoort D, Ewert LE, Wassenaar L (1993) Aquifer vulnerability index: a GIS-compatible method for groundwater vulnerability mapping. Can Water Resour J 18:25–37 Wakida FT, Lerner DN (2005) Non-agricultural sources of groundwater nitrate: a review and case study. Water Res 39(1):3–16 Werner AD (2010) A review of seawater intrusion and its management in Australia. Hydrogeol J 18:281–285 Yammani S (2007) Groundwater quality suitable zones identification: application of GIS, Chittoor area, Andhra Pradesh, India. Environ Geol 53(1):201–210