Mapping groundwater recharge potential zones in arid region using GIS and Landsat approaches, southeast Tunisia

ABSTRACT

The rapid growth of population and agricultural and industrial activities has caused an increase in demand on the Jeffara aquifers of Gabes (southeast Tunisia). In fact, the over-pumping of this aquifer system has resulted in water-level declines ranging from 0.25 to 1 m/year during the past three decades. The aim of this study is to identify favourable artificial recharge sites of this aquifer system based on the combined use of remotely sensed data, a geographic information system (GIS), the Shuttle Radar Topography Mission (SRTM) product and a multiple criteria decision making (MCDM) technique. The delineation of artificial recharge zones shows high to moderate potential for groundwater recharge (40%) in the Gabes region, with high precision of good potential proposed sites. Recharge processes need to account for natural conditions and ecosystems.

KEYWORDS:

• GIS
• remote sensing
• SRTM
• MCDM
• artificial recharge
• arid zone

Editor R. Woods Associate editor A. Fiori

1 Introduction

Groundwater resources play a major role in our lives, as well as those of some terrestrial and aquatic ecosystems (Magesh et al. 2012, Shekhar and Pandey 2015), and thus in socio-economic development (Pinto et al. 2015). The estimation of the World Water Assessment Program in 2003 showed that groundwater provides approximately 50% of water used in global domestic water supplies, 40% of water used in industrial activity and 20% of water used in irrigated agriculture (Foster and Loucks 2006).

In recent decades, water demand has increased intensively, mainly in arid and semi-arid regions of the Middle East and North Africa. Furthermore, it is used for a variety of purposes such as domestic, industrial and agricultural uses (Schmidt et al. 2008, Chenini et al. 2010, Mohanty et al. 2013). Unfortunately, this natural resource has become both a blessing and a challenge in these areas due to excessive water exploitation (Murthy 2000, Ben Kabbour and Zouhri 2004, Ben Kabbour et al. 2005, Gaur et al. 2011). In these regions, many coastal water tables are over-exploited (Fedrigoni et al. 2001, Sophocleous 2002, Kouzana et al. 2009, Yeh et al. 2009, Wada et al. 2010, Zouhri 2010, Nag and Ghosh 2013, Zghibi et al. 2013). The over-use of groundwater is causing many problems, such as a decrease of piezometric levels, deterioration of water quality, water pollution and water exhaustion; hence a water stress condition has been created. To contend with these hydraulic crises, it is necessary to search urgently for an effective technique as well as time for planning, utilization, administration and management of groundwater resources (Singh et al. 2017).

In the past few decades, the water table of the Gabes region, which is the object of the present study, has been overexploited, and the natural recharge has decreased due to the apparent reduction of rainfall. Hence, the result is an imbalance between water resources and the recharge demand. This coastal region is characterized by drought conditions due to irregular variations of rainfall, high evaporation rates equal to 200 mm in July and, most importantly, water scarcity. In order to restore the water level to its normal state, the proposed methodology in this study integrates use of GIS, remote sensing, SRTM (Shuttle Radar Topography Mission) and MCDM (multiple criteria decision making) techniques, which are well known as powerful and useful methods where the main objectives are achieved by identification and delineation of favourable artificial recharge sites (Sirinivasa Rao and Jugran 2003). This methodology is also well known as a powerful tool for groundwater resources management, particularly in arid environments (Abdallah 2012). Various studies have been interested in the delineation of groundwater potential zones and groundwater evaluation using geoscience technologies. These processes form an important tool in water studies due to their ability in developing spatiotemporal information and their efficiency in spatial data analysis. Previous works include the implication of remote sensing in groundwater exploration (Raju et al. 1986) and especially in groundwater prospecting in hard rocks (Prasad et al. 2008) and soft rocks (Shahid et al. 2000), and geological and geomorphological features in groundwater exploration (Krishnamurthy and Srinivas 1995). Remote sensing data have been used with GIS techniques for land, management and water resources assessment (Chi and Lee 1994, Murthy 2000, Sener et al. 2005, Ganapuram et al. 2009, Nagarajan and Singh 2009, Ghorbani Nejad et al. 2017). The implication of remote sensing and GIS techniques has been identified in hydrological modelling processes (Saraf et al. 2004) and in the management and development of groundwater resources (Jha et al. 2007, Gaur et al. 2011). Remote sensing and GIS have been integrated in the assessment of the groundwater recharge potential zone (Ghayoumian et al. 2007, Oikonomidis et al. 2015, Yeh et al. 2016), in the identification of suitable areas for artificial recharge estimation (Subagunasekar and Sashikkumar 2012) and in disasters and risk reduction in groundwater (Ayazi et al. 2010, Senanayake et al. 2016).

However, demarcation of potential sites for artificial recharge is based on knowledge-driven factor analysis, integrating different themes such as lineament density, drainage density, slope, soil types, land use/land cover, lithology, rainfall and geomorphology. The thematic layers mentioned in this study were enhanced using GIS, remote sensing, SRTM, MCDM and conventional data. These themes were assigned with proper weights based on their major and minor effects on the groundwater recharge potential area. These layers were then integrated in the GIS platform to manage, utilize and classify the results of remote sensing, to explore the diversity of recharges sites, to couple the themes of recharge potential, and finally to provide weight relationships (Krishnamurthy et al. 1996, Saraf and Choudhury 1998, Sener et al. 2005).

Consequently, a groundwater potential zones map was obtained to improve water resources management in Gabes. The area was divided into four parts: Very good (40%), Good (13%), Moderate (18%) and Poor (29%).

2 General description of study area

The study area (Fig. 1) is located in the southeast of Tunisia. Geographically, Gabes is situated between X = 550 000, X = 620 000 and Y = 3 700 000, Y = 3 770 000 (UTM geographic projection system). It belongs to the Jeffara basin, which extends from the Skhira region in the north to Libyan territory in the south.

A 30-year analysis of rainfall data in the Gabes region has shown an average of 162 mm/year (General Department of Water Resources 2015), of which more than 75% occurs from September to March. Analysis of the results shown in Figure 2(a)–(c) shows that the monthly rainfall distribution during the period 1985–2015 was seasonal. The precipitation in the study area was recorded from October to March, and the dry period varied between April and September. During the autumn and winter seasons, the rainfall rate is very significant, but it is moderate in spring and low to very low in summer.

Gabes is characterized by an average annual temperature of 19–22°C but sometimes exceeds this. The estimation of water balance parameters is based on a monthly calculation for each station (Table 1). The potential evapotranspiration (PET) estimated by the Thornthwaite method is very high (1066 mm/year) and largely exceeds the rainfall. The actual evapotranspiration (AET) is about 178 mm/year. The infiltration is equal to 45.36 mm/year. To calculate runoff, we apply the Tixeront-Berkaloff formula:

(1)

Table 1. Water balance parameters. R: rainfall; PET: potential evapotranspiration; AET: actual evapotranspiration; RUPS: reserve useful for plant survival (max. = 32).

Meteorological station	R (mm)	PET (mm/year)	AET (mm/year)	RUPS, initial	RUPS, final	Excess	PET (mm/d)
Gabes SM	198	1066	206.22	99.85	71.45	20.17	34.88
Gabes DRES	169	1066	185.22	86.71	63.32	7.17	34.88
El Mdou	154	1066	172.22	91.85	70.45	3.17	34.88
El Mida	105	1066	134.83	83.71	51.71	2.17	34.88
Ghannouch	164	1066	178.22	89.84	64.45	11.17	34.88
Matouia	188	1066	201.22	88.85	59.45	16.17	34.88
El Hamma PV	168	1066	176.22	94.85	74.45	12.17	34.88
El Hamma Téléphérique	93	1066	125	81.05	19.05	0	34.88
Mareth	209	1066	212.51	112.42	92.03	16.88	34.88
Matmata Delegation	176	1066	187.22	101.85	71.45	19.17	34.88
Average	162	1066	178	93	74	11	35
Water balance for the rainy year 2003 (Alaya 2014)	232.1	1173.09	212.53	–	–	19.57	–

where R is runoff (m), P is rainfall (m) and PET is potential evapotranspiration (m). Thus, from the above data, R = (0.162)³/3 × 1.066 = 0.00151 m/year = 1.51 mm/year.

The area is crossed by nine main rivers (Fig. 1): El Akarit, Gabes, Sourag I and II, Jir, El Fared, Zigzaou, El Malah and El Hamma, which are seasonal rivers. These hydrographic networks flow into the sea or into the endoreic depressions (Sabkhas). The topography of the study area is a plain with irregular landforms, and the relief of the area varies between 55 and 650 m a.m.s.l. The geological series presented in this area are from Triassic to Quaternary.

Figure 1. Location map of Tunisia and the Gabes region (projection system: TUN UTM; datum: WGS_1984; zone: 32).

Many normal faults affect the Jurassic, Cretaceous and Cenozoic geological series (Bouaziz 1995). Lithological study shows that the Gabes basin is formed of several lithological layers: sands and conglomerates, sandy clays, sands, limestone, dolomite, gypsum, halite and carbonates (see Section 4.3.7). Therefore, it can be deduced that the aquifer system of the Gabes region is a multilayered aquifer system. In this paper we are interested in the study of the phreatic groundwater table of a Mio-Plio-Quaternary aquifer. This aquifer is characterized by permeability varying from 0.09 × 10^–6 to 88.1 × 10^–4 m²/s, storage coefficient ranges from 1.2 × 10^–4 to 21 × 10^–2, and the transmissivity varies between 0.015 × 10^–4 and 35.2 × 10^–2 m² s⁻¹ (Mekrazi 1975, Ben Baccar 1982, Alaya et al. 2014). Piezometric map analysis of the Mio-Plio-quaternary aquifer in 2015 showed that the hydraulic gradient varies between 5 × 10^–4 and 1.6 × 10^–3 and the general groundwater flow direction is from west to east and, more precisely, to the Mediterranean Sea (Fig. 2(e)).

Figure 2. (a) Correlation of rainfall with temperature and years. (b) Monthly rainfall variability at the climatological station Gabes SM in the period 1985–2015. (c) Correlation of rainfall–PET from Gabes SM station. (d) Abstraction of Mio-Plio-Quaternary aquifer in the Gabes region. (e) Piezometric map.

In an arid area such as the Gabes zone the amount of groundwater resources is limited due to the irregularity of rainfall and increasing water demand. It may be seen in Figure 2(d) that the exploitation rate increased progressively between 1985 and 2015. In the study area the main zone of groundwater extraction from the Mio-Plio-Quaternary aquifer is the Gabes southern area (1852 wells in 2015), with an abstraction rate equal to 142% in 2015 (General Department of Water Resources 2015). In addition to the exploitation by wells, the groundwater of the Gabes region is exploited by several springs (El Akarit, El Hamma, Gabes and El Aouinet rivers and Ras el Ayoun) to meet the growing demand for water. In the early 1990s, most of the springs were depleted due to overexploitation (Alaya et al. 2014).

3 Methodology

The methodology adopted for this study is shown in a flow chart (Fig. 3). The main focus was on the delineation and selection of suitable artificial recharge sites using integrated GIS, remote sensing, SRTM products and MCDM techniques. However, in order to demarcate artificial recharge zones, a multiparametric dataset including satellite data, geological and topographic maps and other conventional data was used. The main map of Gabes was prepared from the digitization of geological and topographic maps on a 1/100 000 scale.

Eight thematic layers – drainage lineament, drainage density, land use/land cover, slope, soil type, lithology, rainfall and geomorphology – were used for the demarcation of artificial recharge zones.

The lineament and drainage densities were prepared using the line density analysis tool in ArcGIS. The slope map was prepared from the DEM data extracted from SRTM global 1-arcsecond (30-m) in Arc-info 10.1 software; slopes were calculated from the elevation contours. The rainfall map was obtained after the calculation of average annual rainfall using monthly data for our study area over 30 years (1985–2015) published by the General Department of Water Resources. Then, an interpolation technique using kriging on a Surfer platform was used to obtain the rainfall spatiotemporal distribution map. The study area lithological map was prepared by assembling, georectifying and digitizing six geological maps (Gabes, El Hamma, Mareth, Oglet El Marteba, Matmata, Tamazret) on a 1/100 000 scale, produced by the National Office of Mines. The soil type map was produced by the General Department of Water Resources and the geomorphological map was produced by the National Office of Mines, and was subsequently geo-rectified and digitized. The land-use/land-cover data were obtained following the processing of satellite images from the Landsat 8 Sensor on a scale of 1/500 000 (geo-coded, with UTM projection, spheroid and datum WGS 1984, zone 32N) by ENVI and then converted into raster format (30-m resolution) before being brought into the GIS platform. Groundwater recharge potential zones in the Gabes region were demarcated by overlaying all the thematic maps in terms of weighted overlay methods using the spatial modelling tools in ArcGIS 10.1. This was influenced by eight factors: lineament density, drainage density, rainfall, slope, land use/land cover, lithology, soil type and geomorphology. Diverse features of each factor were also assigned weights on a scale of 1–8, corresponding to their relative importance for recharge potential (Saaty 1980, Shaban et al. 2006). Qualitative evaluation of the different features of a given theme was performed based on the assigned weights as follows: Poor (weight = 1–2), Moderate (2–4), Good (4–6) and Very Good (6–8) (Saaty 1980, Shaban et al. 2006).

The weighted linear combination method was adopted to develop the groundwater recharge index of each pixel (Senanayake et al. 2016) in the Gabes region by:

(2)

where Pr is the groundwater recharge potential index, RF is the rainfall index, LG is the lithology index, GG is the geomorphology index, SG is the slope index, LD is the lineament density index, DD is the drainage density index, LC is the land-use/land-cover index, and SC is the soil cover index. The subscripts w and r refer, respectively, to the weight of a theme and the rate of individual features of a theme (Yeh et al. 2014, 2016).

4 Results and discussion

4.1 Weighted calculation of factors influencing the potential zones

In this study eight significant interdependent factors contribute in delineating groundwater potential zones.

In the Gabes region, the diagram of groundwater recharge potential is used (Fig. 4). The effect of each minor and major factor is assigned a weight of 0.5 or 1.0, where 1.0 is assigned for major interrelationship existing between two factors, while 0.5 is assigned for a minor interrelationship between two factors. Finally, the cumulative weight for each factor is considered in order to calculate the recharge potential factor weight:

(3)

Table 2 shows the scores assigned to the parameters influencing each recharge potential factor.

Table 2. Scores assigned to the parameters influencing each recharge potential factor (modified after Krishnamurthy et al. 1996, Edet et al. 1998, Shahid et al. 2000, Jaiswal et al. 2003, Shaban et al. 2006, Magesh et al. 2011, Kaliraj et al. 2014, Oikonomidis et al. 2015, Senanayake et al. 2016, Yeh et al. 2016)..

Factor	Major effect	Minor effect	Calculation process	Rate	Score assigned
Lineaments	2	0	2 + (0 × 0.5) = 2	2.0	12
Land use/land cover	1	3	(1 × 1) + (3 × 0.5) = 2.5	2.5	15
Lithology	4	0	(4 × 1) + (0 × 0.5) = 4	4.0	24
Drainage	1	1	(1 × 1) + (1 × 0.5) = 1.5	1.5	9
Slope	2	1	(2 × 1) + (1 × 0.5) = 2.5	2.5	15
Soil	1	0	(1 × 1) + (0 × 0.5) = 1	1.0	6
Rainfall	1	1	(1 × 1) + (1 × 0.5) = 1.5	1.5	9
Geomorphology	2	0	(2 × 1) + (0 × 0.5) = 2	2.0	12

4.2 Establishing potentially related groundwater recharge factors

4.2.1 Lineament density factor

Lineaments are simple and complex linear properties of geologic structures such as faults, fractures and discontinuity surfaces detected by remote sensing (O’Leary et al. 1976).

Using lineaments in the present study is very important in terms of the hydrogeology because they provide the pathways for groundwater movement (Abdalla 2012). The lineament density factor is given by:

(4)

where is the total length of the lineaments [L] and A is the unit area [L²].

A high lineament-length density provides better targets for groundwater recharge than joints, with high levels of potential groundwater recharge (Edet et al. 1998, Haridas et al. 1998, Yeh et al. 2016). In the study area, lineaments were extracted from the SRTM data and the geological map using the Geomatica 2015 platform.

The lineament density map was prepared using the line density analysis tool in ArcGIS 10.1. For the Gabes region, the lineament density map (Fig. 5) shows that high lineament density is observed in the relief, with values ranging from 1 to 1.5 km/km².

The statistical processing of lineaments enumerates about 320 lineaments. To represent the results of their arrangement, we realized a rose diagram of directional families grouped into 10° classes. The rose diagram (Fig. 6) shows the clear dominance of a major NNW–SSE direction of lineaments. This corresponds to the general direction of deformation stress related to the compressive tectonic phase of Miocene age responsible for the structuring of this region (Zargouni 1985, Dlala 1995, Msaddek et al. 2016).

4.2.2 Drainage density factor

Drainage density indicates the measure of the total length of all orders per unit area (Shekhar and Pandey 2015). In this study, a drainage network of the Gabes zone was created manually from the topographic maps (1:100 000 scale). In order to calculate drainage density, the drainage line density analysis tool in ArcHydro 10.1 (in ArcGIS software) was used:

(5)

where is the total length of the drainage network [L] and A is the unit area [L²].

The drainage density is correlated with the groundwater recharge, because the groundwater potential recharge is related to permeability and surface runoff (Krishnamurthy et al. 1996, Magesh et al. 2012, Huang et al. 2013). The zone with high drainage density has a high level of groundwater recharge (Shaban et al. 2006). The drainage density map in the study can be divided into four classes: Very Good (2.5–3.5 km), Good (1.5–2.5 km), Moderate (1–1.5 km) and Poor (0–1 km). High drainage (2.5–3.5) was recorded in the relief of the study area. The drainage density potential evaluation is shown in Figure 7.

4.2.3 Slope factor

The slope gradient is a very important factor for the identification favourable sites for artificial recharge because it directly influences the infiltration of surface water (Selvam et al. 2015). On steep slopes, the level of recharge is low as water flows rapidly downwards, having insufficient time to infiltrate (Machiwal et al. 2011). However, slight slopes facilitate groundwater recharge, due to the extensive high retention of rainwater.

In the present study area, the slope map (Fig. 8) was prepared based on SRTM data using the spatial analysis tool in ArcGIS 10.1. Based on the slope map, the Gabes zone was subdivided into three categories. The first category (0–2) is assigned Very Good for artificial recharge because of the nearly flat terrain and high infiltration; in the second category (2–4), slopes are considered as Moderate for groundwater storage with some runoff. The third category, considered as Poor, could show a difference due to the runoff and infiltration variations; it is categorized with an intensity of 4–10 in the presence of high runoff and low infiltration and with an intensity exceeding 10 in the presence of relatively very high runoff.

4.2.4 Soil types factor

The study of soil cover is an essential factor in delineating groundwater recharge potential zones (Punmia and Jain 2005).

The infiltration of surface water into an aquifer system is dependent on soil type. Based on the infiltration rate, texture, depth, drainage condition and water transmission capacity (Viessman et al. 1989), soils are classified into different hydrological soil classes: A, B, C and D (Table 3).

Table 3. Criteria for classification of hydrological soil group (Chow et al. 1988, Viessman et al. 1989).

Character	Hydrological soil group
Infiltration rate	A: High	B: Moderate	C: Slow	D: Very slow
Texture	Sand/gravel	Moderately coarse to moderately fine	Moderately fine to fine	Clay
Depth	Deep	Moderately deep to deep	Moderately deep	Shallow over an impervious layer of clay or clay pan of high water table
Drainage	Well drained to excess	Moderately well to well drained	Moderately drained to slow	Very slow
Water transmission	High	Moderate	Slow	Very slow
Remarks	Low runoff potential	Moderate runoff potential	Moderate runoff potential	High runoff potential

Soil study shows that the Gabes region belongs to hydrological soil groups A and D. Group A covers about three-quarters of the area while the rest of the surface is covered by D group soil (Fig. 9).

4.2.5 Rainfall factor

We collected rainfall data for 30 years (1985–2015) from the General Department of Water Resources for 10 raingauge stations. Annual average rainfall of the study area is around 162 mm/year.

According to the data from the meteorological station of Gabes SM, located in Gabes city (X = 601855.68; Y = 3750275.386), we can conclude that the rainfall increases near the coast, and decreases gradually as we move away from the coast (Fig. 10).

The correlation between precipitation and elevation shows that precipitation decreases progressively with increasing altitude (Fig. 11). Based on the rainfall map and the precipitation–elevation correlation, it can be seen that the higher the rainfall is the higher the groundwater potential becomes. The resulting map was classified into five major classes (Fig. 10): 10–50 mm/year (Poor), 50–100 mm/year (Moderate), 100–150 mm/year (Good), 150–200 mm/year (Good), and <200 mm/year (Very Good). In the study area, the rainfall factor was assigned a weight of 10% in the total groundwater potential value.

Figure 3. Schematic representation of the methodology adopted to delineate the groundwater recharge potential zone of the Gabes region.

Figure 4. Interrelationship between the multiple influencing factors concerning the recharge property.

Figure 5. Lineament density map of the Gabes region.

Figure 6. Rose diagram of the lineament orientation in the Gabes region.

Figure 7. Drainage density map of the Gabes region.

Figure 8. Slope map of the Gabes region.

Figure 9. Soil types map of the Gabes region.

Figure 10. Rainfall map of the Gabes region.

Figure 11. Correlation between precipitation (P) and elevation.

Figure 12. Land-use/land-cover map of the Gabes region.

Figure 13. Lithology map of the Gabes region.

Figure 14. Geomorphology map of the Gabes region.

Figure 15. Distribution of groundwater potential classes in the Gabes region.

Figure 16. Groundwater recharge potential zone map of the Gabes study area.

4.2.6 Land-use/land-cover factor

Land use/land cover is an important factor in the groundwater potential zone identification (Shaban et al. 2006). The use of this factor makes it possible to define the distribution of buildings and vegetation cover areas. These land-use classes were interpreted from Landsat 8 OLI/TIRS satellite imagery. In the satellite imagery, the soil deposit cover is identified with beige colour, the vegetation areas are identified with contrasting green and the built-up land is white.

Analysis of the land-use/land-cover map (Fig. 12) shows that around 90% of the total area is represented by bare land and built-up land, and the other 10% is represented by forests and agricultural lands.

4.2.7 Lithology factor

The type of lithology affects groundwater recharge by supervising water flow percolation (El-Baz et al. 1995), and afterwards it is associated with water permeability (Oikonomidis et al. 2015).

The lithology map (Fig. 13) shows that the Gabes region is overlain by alluvium and Quaternary sediments, which are distributed over almost 70% of the total area. There is also sandy clay (sand and conglomerates) of lower Pleistocene to Mio-Pliocene age and coarse sand of Miocene age. The reliefs are covered by marly limestone and dolomites of Cretaceous age. In this study, the lithology factor was assigned a weight of 33% in the total groundwater potential value.

4.2.8 Geomorphology factor

The geomorphology of an area has a dominant role in evaluating the movement and storage of groundwater (Dinesh Kumar et al. 2007). The topography of the Gabes region is composed mostly of plains and low landforms, and the relief of the area varies between 165 and 605 m a.m.s.l. (Fig. 14). Mountains and high landforms are present in the southern part of the study area and in a limited region on the northern side. These characteristics favour the composition of eolian deposits, floodplain deposits, salt-evaporate deposits and pediment in a large part of the region. Rocky outcrops are present in a small portion of the study area. These geomorphological features have Good to Poor potential for groundwater recharge.

4.3 Delineation of the groundwater recharge zone

The groundwater recharge potential zone in the Gabes region was delineated by assigning weights to various classes in each factor based on their contributions in recharging. The resulting total weighting assessment is shown in Table 4. This approach has been used by several researchers, and it has been shown that this method gives precise results (Shahid et al. 2000).

Table 4. Classification of various factors for delineation of groundwater potential zones.

Factor	Domain of effect	Weight of effect	Rates	Weighted rating	Total	Descriptive recharge potential
Lineament density (km/km^2)^a	0–0.5	1	2.0	2	22	Poor
	0.5–1	4		8		Good
	1–1.5	6		12		Very good
Drainage density^b	0–1	2	1.5	3	16.5	Poor
	1–1.5	3		4.5		Moderate
	1.5–2.5	5		7.5		Good
	2.5–3.5	6		9		Very good
Land use/land cover	Buildings	1	2.5	2.5	38.75	Poor
	Bare land	3.5		8.75		Moderate
	Forest + agriculture	5		12.5		Good
	Streams	6		15		Very good
Lithology	Alluvium and Quaternary sediments	6	4.0	24	88	Very good
	Sand clay–sand	6		24		Very good
	Coarse sand	6		24		Very good
	Silty clay	2		8		Poor
	Marly limestone	2		8		Moderate
Slope	0–2°	6	2.5	15	27.5	Very good
	2–4°	3		7.5		Moderate
	4–10°	1		2.5		Poor
	>10°	1		2.5		Poor
Soil	Sandy and sandy silt	6	1	6	8	Very good
	Sandy + clay and silt	2		2		Poor
	Clay, clay sandy
Rainfall	10–50 mm	1	1.5	1	25	Poor
	50–100 mm	2		3		Moderate
	100–150 mm	4		6		Good
	150–200 mm	4		6		Good
	<200 mm	6		9		Very good
Geomorphology	0–110 (Plain)	6	2.0	12	28	Very good
	110–660 (Relief)	2		4		Moderate
	Rivers	6		12		Very good

^aLineament per km².

^bSegment per km².

The demarcation of groundwater potential zones of the Gabes region was based on the integration of thematic maps (lineament density, drainage density, slope, land use/land cover, lithology, soil types, rainfall and geomorphology) using GIS, SRTM and remote sensing data and the MCDM techniques. The different maps were considered as layers. ArcMap tools were used to overlay the different layers, after considering the weight/rate assessment. Finally, the resulting map reveals the groundwater potential zones in the Gabes region.

The total weight in this study area was calculated using a weighted linear combination method (Equation (2)), as 22 + 38.75 + 80 + 16.5 + 27.5 + 8 + 28 + 25 = 245.75. Using a value of 246, this leads to the following percentages of factor effect on the recharge artificial capacity:

• Lineaments density = (22/246) × 100 = 9%

• Land use/land cover = (38.75/246) × 100 = 16%

• Lithology = (80/246) × 100 = 33%

• Drainage density = (16.5/246) × 100 = 7%

• Slope = (27.5/246) × 100 = 11%

• Rainfall = (25/246) × 100 = 10%

• Soil types = (8/246) × 100 = 3%

• Geomorphology = (28/246) × 100 = 11%

The distribution of groundwater potential classes in the Gabes region is presented in Figure 15 and Table 5.

Table 5. Groundwater recharge potential zones and categorization.

	Very Good	Good	Moderate	Poor
Area (%)	40	13	18	29
Area (km^2)	4607	1499	2061	3333
Average (%)^a	47.5	32.5	15	7.5

^aUN (1967).

The groundwater potential recharge zone map of the Gabes region (Fig. 16) shows that areas of Very Good groundwater potential occupy the largest part of the study area (4607 km²; Table 5).

However, to measure the quantity of water recharged to subsurface media in the Gabes region, a simplified calculation for the proposed recharge rates is used (UN 1967). The estimation of total recharged water volume (W) is calculated as:

(6)

where P is precipitated volume, R is the recharge ratio and A is % of area. Thus, W (m³/year) = 530 × 10⁶[(0.475 × 0.40) + (0.325 × 0.13) + (0.18 × 0.15) + (0.075 × 0.29)] = 530 × 10⁶ × 0.28 = 148.4 × 10⁶ m³/year.

The Mio-Plio-Quaternary aquifer is overexploited due to an increase in demand for water for irrigation, agriculture, domestic and industrial activities, and drought in the study region. This overpumping of groundwater resulted in a resources deficit of 7.60 hm³/year in the year 2015 (General Department of Water Resources 2015). According to our statistical study, 28% (148.4 × 10⁶ m³/year) of the precipitation received in the study area is infiltrating downward to recharge the groundwater reservoirs and water bodies, and the rest is lost in the form of either evapotranspiration or surface runoff.

5 Conclusion

The Gabes region is located in the southeastern part of Tunisia. In recent years, it has become known for both demographic and socio-economic growth. The excessive pumping of groundwater in the study area has resulted from a decrease in its piezometric level.

In this study, an artificial recharging technique was applied as an effective tool for water resources management. The present study was carried out to propose a methodology for delineating groundwater recharge potential zones in Gabes. An integrated approach for assessing the characteristics of groundwater recharge potential using remote sensing, GIS, SRTM and MCDM techniques was proposed, aiming to minimize time and money in order to be able to make quick decisions for sustainable water resources management by this powerful methodology. Landsat OLI/TIR data, geological and topographic maps were used to prepare thematic layers of land use/land cover, lineament density, drainage density, lithology, slope, rainfall, geomorphology and soil type maps. The various thematic layers were integrated on a GIS environment employing the weighted linear combination method to prepare a groundwater potential zone map. Finally, the results indicated that Gabes is characterized by different zones, namely, Very good, Good, Moderate, Poor. Around 40% of the total area corresponds to the Very good. The weighted linear combination method allows us to demarcate the favourable sites for artificial recharge in the study area using remote sensing, GIS and SRTM techniques as well as MCDM, despite the data scarcity in terms of quantity and quality. Furthermore, the result of the present study will serve as a guideline for planning future water resource management projects in the Gabes region in order ensure sustainable groundwater use. On the whole, it can be concluded that the method can be applied in other areas of the world experiencing water stress conditions for the exploration of suitable sites. The method applied to assess artificial recharge needs to be integrated with other processes that were neglected in this paper, when the research deals with the correlation between the excessive pumping of groundwater in the study area and the decrease in its piezometric level; this correlation needs to be quantified.

Acknowledgements

The research was supported by the Tunisian Ministry of Higher Education and Tunis El Manar University with the collaboration of the UniLaSalle-Beauvais Polytechnic Institute (France).

Disclosure statement

No potential conflict of interest was reported by the authors.