Introduction

The Mamora coastal plain is located in the northwest of Morocco on the Atlantic littoral coastal region (Fig. 1). This region has an area of about 390 km2, 50% of which is covered by one of the most important cork oak forests in the world.

Fig. 1
figure 1

Situation of the studied area and its local geology

It has a Mediterranean climate with temperatures varying from 13 °C in winter to 27 °C in summer. The mean annual precipitation is about 500 mm.year−1. The Sebou River (3 billion m3.year−1) crosses the region in the north. Ocean tide effects on the river were felt up to tens of kilometers away from the shore (Combe 1975).

Escalating stresses due to anthropogenic activities are affecting the Plioquaternary aquifer, formed by calcarenitic and sandstone rocks. Indeed, the socio-economic development of the area is generally dependent on groundwater pumping. Parallel growth of population, agriculture and industry has occurred especially near the Atlantic Ocean shore and the Sebou River estuary. The Kenitra Province, involving the study area, has 1 million inhabitants; 56% of which live in rural areas. Agriculture is the principal activity in the region and 53% of industries are involved the agri-food sector (DRPE 1989). Pumping discharge is 38 million m3.year−1 used with a proportion of 75% for drinking and industrial needs and 25% for agriculture (DRPE 1989). Kenitra city is supplied by three municipal pumping stations (Fig. 2).

Fig. 2
figure 2

Coastal Mamora land use map as deduced from topographic maps, aerial photography and field missions

Though many pumping wells were abandoned because of high salinity during previous accentuated dry periods, presently the groundwater salinity remains moderate at a regional scale. However, an increase of groundwater salinity towards the ocean and over tens of kilometers along the Sebou River estuary is observed (Fig. 3, Table 1).

Fig. 3
figure 3

Areal distribution of electrical conductivity (July 2000)

Table 1 Observation wells and the relevant electrical conductivity (July 2000)

In this study, the Direct Current resistivity method is used to measure the electrical properties of the coastal aquifer system of Mamora. This method, which in recent years has undergone a resurgence in interest due to advances in data acquisition technology, better mathematical algorithms and more interactive modeling software (Merrick 1997), has given good results in exploring groundwater productivity (Bose and Ramakrishna 1978), salinity and saltwater intrusion (Fretwell and Stewart 1981), geothermal properties (Thanassoulas and others 1986) and anthropogenic pollution (Adepelumi and others 2001, Kundu and others 2002). The DC resistivity method can be used successfully in a variety of different sedimentary and non sedimentary rocks. However, some difficulties can arise in clay rich or faulted formations. Those difficulties may be obviated by combining this method with others such as the electromagnetic (Arora and Bose 1981) or induced polarization (Seara and Granda 1987) methods. In the present study, the DC resistivity method is applied to explore the water table aquifer in the Plioquaternary, faulted, consolidated sandstone with no clay content in the saturated zone. In areas with faults, a technique inspired from offset Wenner profiling (Hobbs and Reading 1987) is used. Hence, the main objective of the research is to determine the geological structure, the depth of the marly base of this aquifer and to assess groundwater salinity distribution in both lateral and vertical directions.

Correlation between geoelectrical, hydrogeological and land use data, allow us to determine factors involved in groundwater vulnerability to marine intrusion and to establish an optimal network of electrical soundings for monitoring this saline contamination.

Geohydrology of coastal Mamora

The coastal part of Mamora plain is limited to the south by the Meseta, which is a Paleozoic tabular and stable bedrock (Flinch 1993), and to the north by the Rharb plain, which is a recent, and subsiding basin belonging to the Rifan domain (Fig. 1). The Paleozoic bedrock is covered by Miocene blue marls, Pliocene calcarenites, then by Quaternary sandstones and sands.

The Mamora coastal plain has a phreatic, consolidated and faulted aquifer. Its substratum is constituted by Miocene blue marl. Its saturated and unsaturated zones are generally made of calcarenitic and sandstone rocks, with high transmissivities (up to 1.5 10−01 m2.s−1). The as yet unpublished modeling of Mamora groundwater flow shows that values for hydraulic conductivity and effective porosity range respectively from 1.2E-04 m/s up to 2.25E-03 m/s and between 3 and 15%. The aquifer may involve different types of flow circulation. Indeed, it has been found that groundwater velocities around the Sidi Taibi pumping station fluctuate as follows (ONEP 1995):

  • from 125 to 342 m/day−1, attributed to flow through karstic channels;

  • from 34 to 75 m/day−1, attributed to flow through fissures;

  • from 9 to 17.5 m/day−1, attributed to flow through interstitial porosity.

These measurements were performed by following the aquifer geoelectrical properties after tracing groundwater with salt injections inside a selected network of wells.

Two hydraulic units may be distinguished (Fig. 4). The northern unit of Sebou where the flow seems to be controlled by the river level and where hydraulic gradients are very low. This unit is separated from the southern one by the hydraulic divide of Sidi Taibi. Within the latter, groundwater flows generally northwestwards to the Atlantic ocean following higher hydraulic gradients. Salinization processes are occurring nearby the coastline and the Sebou River estuary (Fig. 3). Within the affected areas, the groundwater chemical facies changes from a calcium bicarbonate type to a sodium chloride type (Ben Kabbour 2002) and therefore the hydraulic potential is gradually decreasing. Because of the pumping activities and the ocean boundary, the salinization seems to be mainly related to the effects of a seawater intrusion.

Fig. 4
figure 4

Piezometric map of the coastal aquifer of Mamora (July 2000)

Materials and methods

Most geological materials are composed of electrically insulating minerals. Rocks can then conduct electricity through water filled pores and fractures. It has been found that electrical resistivity of a formation is proportional to the electrical resistivity of the saturating pore fluid and inversely proportional to the porosity raised to a power (Archie 1942). More precisely the following relationship was developed:

$$ F = \rho _{f} /\rho _{w} = a \cdot \phi ^{{ - m}} $$

where F is the formation factor, ρ f is the formation resistivity, ρ w is the pore fluid resistivity, φis the porosity, a is the tortuousity and m the cementation factor; the two latter factors are specific to a formation. There are several factors which can influence the resistivity given by Archie’s law. First Archie’s law does not include the internal surface conductivity (clay content) and the presence of clay minerals in the pore space reduces the formation resistivity by their cation exchange capacity. This phenomenon is well pronounced for geological formations with a resistivity of less than 10 ohm m (Keller and Frischknecht 1966). Consequently, the clay layers make it difficult to discriminate between regions of the aquifer with marine intrusion and regions of clays, which had equally low resistivities; a combination of the DC resistivity method with the induced polarization method can give good results in areas with intercalations of clay-rich formations (Seara and Granda 1987). Second, the presence of faults and fractures in an aquifer can have two different impacts; if the fractures and faults are opened and filled with water they lower the predicted formation resistivity but if those fractures and faults are filled with insulator materials they can obstruct the passage of electrical current and then increase the formation electrical resistivity.

In the DC resistivity method a Direct Current is driven through the ground using four-terminal electrodes. The Schlumberger array was used in our example to study variations in resistivity with depth through vertical soundings. Current is driven through one pair of electrodes; the potential established in the earth by this current is measured with the second pair of electrodes. Using the current intensity and the generated potential, the apparent resistivity of the earth can be measured by using Ohm’s Law and the geometric characteristics of the spacing between the various electrodes (Griffiths and King 1965; Keller and Frischknecht 1966; Seguin 1985; Stadelhofen and Meyer 1991). The resistivimeter GEOTRADE (Geotrade 2000) and the computer program GRIVEL (BRGM 1988) are then used respectively for the acquisition and the compilation of the geoelectrical data. The survey, which covers about 390 km2 at a line and current electrodes spacing varying from 300 to 1000 m, counts more then 60 Vertical Electric Soundings (V.E.S.) located in a 5 km wide zone parallel to the ocean. We began by exploring the electrical properties of the geological formations by sounding directly on the outcrops and near wells with geological and geophysical logs.

Results and discussion

Vertical variation of the aquifer formation resistivity

The amount of data recorded provides an insight into the three-dimensional distribution of geoelectric parameters in the area. The resistivity is deduced from laboratory measurements of both the aquifer formation porosity and water resistivity.

The interpretation of the apparent resistivity curves was performed in two stages. During the first stage we used two and three layer model curves (Astier 1971) and with the auxiliary point method we determined an approximate geoelectric model. Next, we used the Grivel computer program to improve the above model until the RMS difference between observed and calculated data was less than 5%. The geoelectric model thus reflects a possible geoelectric section. The interpretation of the resistivity is shown for the traverses P7, P18 and P23 located respectively near the sites of Sidi Bouknadel, Sidi Taibi and Sebou River. As seen from those profiles, the following formations can be identified (Fig. 5):

Fig. 5
figure 5

NW–SE geological sections as deduced from electrical resistivity profiles and well logs (see location in Fig. 9)

  1. 1.

    Surface formations covering the investigated area present large lateral variations in their electrical resistivity and thickness. For surface sandy loam at the east of the profile P7, the mean resistivity varies from 20 to 80 Ohm m.

  2. 2.

    The next geological formation is almost uniform on all traverses, it was identified as sand and calcarenitic formations, with electrical resistivity values lying between 200 and 2000 Ohm m depending upon the relative amount of concentrated sand, gravels and conglomerates and depending upon the existence of faults and karsts.

  3. 3.

    The aquifer formation is identified with a mean resistivity of 250 Ohm m. This resistivity is lowest in the west of all the profiles and decreases too less than 25% of its initial values for the soundings near the sea. This decrease reflects the seawater intrusion impact on the aquifer formation resistivity.

  4. 4.

    The last and deepest formation has a very low electrical resistivity generally between 5 and 70 Ohm m. This formation is constituted of the marls and sandy marls of the aquifer bottom.

Spatial variation of the aquifer substratum geometry

The spatial variation of the aquifer substratum altitude for the whole of the investigated area (Fig. 6) shows that the zone with the deeper substratum is downstream of Sebou River where the substratum level is situated at more than 70 m below sea level. This results in a large thickness of aquifer saturated zone leading to a consequent hydraulic potential.

Fig. 6
figure 6

Structural map of Mamora coastal aquifer substratum

Many faults are detected upstream in the recharge area under the Mamora forest using a technique inspired from offset Wenner profiling (Hobbs and Reading 1987). Those faults have a direction parallel to the coast and can coincide with highly productive zones in the aquifer (Zouhri and others 2001).

Groundwater salinity distribution

The resistivity values are highly influenced by the proximity of the ocean. For example, from a distance of 200 m spacing between the current electrodes, the lowest resistivity is observed in the G3 soundings, which are 500 m from the sea, and the highest resistivity is observed in the G1 sounding, which is 2000 m from the sea (Fig. 7). The lower resistivity at the beginning of the eastern sounding is caused by a 5-m thick clay-rich layer on the surface separated from the aquifer by a 5-m thick high resistive unsaturated zone. Hence, the aquifer remains unconfined in this part of the studied area in spite of the existence of the surface clay-rich formation. The aquifer formation resistivity decreases too less than 50 Ohm m northeast and southwest of the region (Fig. 8), in areas where salinization processes are observed. If we relate this lowest resistivity to the sole impact of seawater intrusion, the coastal area downstream of Sidi Bouknadel and downstream of Sebou River are the most affected by seawater contamination (Fig. 8). The area of Sidi Taibi is also affected, but to a lesser extent. Vertically, the seawater wedge seems to penetrate 3 km inside the continent for the profile P23, 1 km for the profile P7 and is of a local extent for the profile P18 (Fig. 5).

Fig. 7
figure 7

Some vertical electrical soundings showing the impact of water salinity on the aquifer electrical resistivity

Fig. 8
figure 8

Interpreted resistivity of the Mamora coastal aquifer

Seawater intrusion factors and geophysical monitoring

In addition to pumping activities, several factors may explain the variable extent of seawater intrusion from sector to sector.

One of the several equations deduced from the Ghyben-Herzberg equation, describing the geometry of the fresh-saline interface in relation to the hydrogeologic characteristics of the aquifer and rates of fresh groundwater flow (Todd 1980; Domenico and Schwartz 1990), is as follows:

$$ L = \frac{{\rho _{f} }} {{\rho _{s} - \rho _{f} }} \times \frac{{K \cdot m^{2} }} {Q} $$

where ρ f is the density of freshwater and ρ s is the density of seawater.

It shows that the landward extent of the saline water wedge (L) is directly proportional to the hydraulic conductivity (K) and the square of the aquifer thickness (m), and inversely proportional to the flow of freshwater to the sea.

Consequently, in the area of Sebou River, the wide extent of seawater intrusion can be explained by the high depth of the aquifer substratum leading to the large thickness of the aquifer. High permeability could also be invoked because of the very low hydraulic gradients and the probable existence of fault directions perpendicular to the coast (Zouhri and others 2001).

On the other hand, ecological factors also make an important contribution. The Sebou River constitutes, indeed, a pathway for seawater into the mainland. Therefore, seawater contamination may occur directly from the ocean near the coastline as well as via the Sebou riverbed several tens of kilometers inland. Concerning the Sidi Taibi sector, the forest near the coastline (Fig. 2) is preserved by law, leading to a lower intensity of exploitation of groundwater for irrigation use and thus to less seawater intrusion.

Traditional shallow wells with small withdrawals as well as the probable karstic flow circulations would contribute to keeping groundwater salinity at generally acceptable levels.

An optimal network of electrical soundings (Table 2) can be chosen from all the soundings to monitor saltwater intrusion. We can use for the site of Sidi Bouknadel soundings G1, G2, G3; for the site of Sidi Taibi soundings Fr4, Sc1, Sc2; for the area closer to Kenitra city soundings Hd1, Brm1, Kh1 and finally soundings D1, NKh1 for the area of Ahmed Taleb (Fig. 8). These soundings are easy to access, located in areas where the aquifer basement is under the sea level and downstream of municipal pumping stations.

Table 2 Optimal network of electrical soundings for monitoring saltwater intrusion

Conclusions

The DC resistivity method works very well in the Mamora coastal aquifer even if the aquifer is consolidated and covered by very resistant surface soils.

Many geoelectric documents were produced. It has been shown that the zone with the deeper substratum is downstream of Sebou River, resulting in a large thickness of the saturated zone of the aquifer and then a consequent hydraulic potential. This thickness may, on the other hand, be an important factor leading to the wide extent of seawater wedge in this sector. This is, indeed, a significant example, which shows that in coastal areas, the productive zones are often the most vulnerable to seawater intrusions.

Ecological factors, such as the presence of the forest cover and the Sebou Estuary, are also controlling seawater intrusion. Therefore groundwater resources management should take into account all the Mamora ecosystem compounds for water development planning as well as for protection measures against groundwater salinization.