Introduction

Geotechnical and geophysical assessment of dam sites is very important for dam safety (Blyth and De Freitas 1978). Most of the geotechnical problems in dam sites are related to geological problems (Tabwassah and Obiefuna 2012). Most of these problems are associated with groundwater seepage, fracturing, and karstification (Biswas and Charttergee 1971; Olorunfemi et al. 2000). Such fractures and karst features change the physical properties of the rock materials and hence make geophysical methods useful for detecting these features (e.g., Butler et al. 1990; Titov et al. 2000; Panthulu et al. 2001) and consequently investigating dam safety. Accordingly, geophysical tools play an important role in acquiring subsurface geological and geotechnical information at dam sites (Karastathis et al. 2002; Al-Omosh et al. 2008; Minsley et al. 2011; Moustafa Sayed et al. 2012; Tabwassah and Obiefuna 2012; Youssef et al. 2012; Akinrinmade et al. 2013).

To the northwest of Riyadh city, the capital of Saudi Arabia, the El-Elb Dam was built across Wadi Hanifa. The dam rests on the limestone bedrock of the Jubaila Formation that represent a critical hazard to the dam safety due to the fracturing and karstification features that affect this rock (Touma and Bellerjeau 1981; Vaslet et al. 1988). In the present study, the possible existence of the near-surface fractures and karst features, such as cavities and voids, at the dam site were investigated and verified using electrical resistivity tomography (ERT) and ground-penetrating radar (GPR) techniques along with the field structural and geological verification. The expected success of these tools is based on the assumption that there is a contrast in electrical resistivity (in the case of electrical resistivity imaging) and in electromagnetic energy amplitude reflections (in the case of GPR) between air-filled cavities or sediment-filled collapse structures (or solution cavities) and the surrounding bedrock (McCann and Forde 2001; Polder 2001).

Hydrogeological and geological overview

Figure 1 illustrates the digital elevation model and the drainage network of the El-Elb dam watershed after locating the dam site as the outlet. The maximum elevation in the upstream is about 950 m above sea level and it decreases with slight slopes to 660 m at dam site, mean basin elevation is about 823 m. The general flow direction of the streams goes from west to east (Fig. 1). The catchment area of the El-Elb dam is about 1291 km2 with an average slope of 0.076 and the total length of the streams within the catchment is about 2515 km. The region receives an average annual rainfall of 94.22 mm/year and has a mean annual pan-evaporation rate of about 2893.5 mm/year (PME 2011).

Fig. 1
figure 1

Digital elevation model (DEM) and drainage network of the El-Elb Basin

Full size image

Daily records of rainfall and evaporation data were obtained from the Ministry of Water and Electricity (MWE) of Saudi Arabia for the duration of 15 and 45 years, respectively (MWE 2008). The data has been obtained from records of the meteorological station of Jubaila, which falls within the catchment area of the El-Elb dam. Figures 2 and 3 show the average monthly rainfall and evaporation for the El-Elb dam catchment.

Fig. 2
figure 2

Average monthly rainfall in the El-Elb dam catchment (PME 2011)

Full size image
Fig. 3
figure 3

Average monthly evaporation in the El-Elb dam catchment (PME 2011)

Full size image

Figure 4 shows the water level variation in the El-Elb dam reservoir from the 2nd to the 20th of May 2010. The average infiltration rate for the reservoir has been estimated from the records of the water level data during the period from the 14th to the 19th of May and was found to be approximately 8 mm/h (200 mm/day). The rate of decline in the water level was taken as a measure of the average infiltration rate in the reservoir only when there was a continuous water level decline trend, free from quick decline due to the opening of the gates for controlling the drainage outlet works (PME 2011 and MWE 2008).

Fig. 4
figure 4

Water level data for El-Elb Dam reservoir (PME 2011 and MWE 2008)

Full size image

Geologically, the study area is located in the stable shelf of the Phanerozoic Arabian Platform. The Arabian Platform, which dips toward the Arabian Gulf, is made up mainly of a sedimentary sequence resting non-conformably on Precambrian igneous and metamorphic rocks of the Arabian Shield. The El-Elb Dam was built across the Wadi Hanifa on carbonate rocks of the Upper Jurassic Jubaila Formation (Fig. 5). The Jubaila Formation rests dis-conformably on the Hanifa Formation and is overlain by the Arab Formation. The Jubaila and Hanifa Formations are Kimmeridgian in age, whereas the Arab Formation is assigned to the Tithonian (Powers et al. 1966). From bottom to top, the Jubaila Formation is composed of hard ledges of limestone alternating with soft calcareous clayey beds (Fig. 5). The succession of the Jubaila Formation has been ascribed to a carbonate ramp model in the Central Arabian Arch (El-Asmar et al. 2015). The formation dips toward the east. The development of drainage channels in outcrops of the Jurassic limestone has been controlled by fracture zones and joints. This led to the formation of the steep-sided valleys of the Wadi Hanifa and its tributaries. The bottoms of these valleys, which are partially covered by coarse alluvial sediments, may reach a depth of 100 m.

Fig. 5
figure 5

Location and geologic map of the investigated site

Full size image

Significant karstification develops most commonly in thick-bedded limestone (CaCO3). Fracture zones and joints in the limestone are considered to be effective elements controlling the circulation of fluids and the dissolution processes during karstification (Swinnerton 1932; Sweeting 1950; Ford and Ewers 1978; Palmer 1991). Structurally, the Jubaila Formation is dissected by nearly vertical joints trending in mainly E–W and N–S directions (Fig. 6). The nature and extension of fractures and karst features reveal that the dissolution of jointed Jubaila limestone beds may have occurred by downward chemical weathering caused by meteoric (H2O–CO2) solutions. The solutions infiltrating the Upper Jurassic limestone (Jubaila Formation) might have moved through joints and along bedding surfaces. This has led, in some places, to rock fall hazards (Fig. 7). The circulation of fluids through the limestone of the Jubaila Formation and the consequent dissolution process seem to be controlled by the areal distribution of joints and bedding planes (Fig. 7). Field investigations have revealed a relationship between joints and karstic cavities (Figs. 6 and 7). The karstic cavities are often highly developed along the intersections of bedding planes and joints, where infiltrating solutions and dissolution are focused.

Fig. 6
figure 6

Field photograph showing an E–W-striking nearly vertical joint in the limestone. Note that black arrows point to a crack along this joint in the cement wall of the draining sector of the dam

Full size image
Fig. 7
figure 7

Field photograph showing an E–W-striking joint (blue arrows), a karst feature (yellow arrow) in the limestone, and a collapsed limestone block (red arrow)

Full size image

Data acquisition and processing

In the present study, four ERT profiles were carried out on the downstream side of the El-Elb Dam (Fig. 8). The survey was on the downstream side of the dam, which is the side that is the main area of interest regarding seepage; in addition, the upstream area of the dam was occupied by water. Two profiles (A–A′ and B–B′) were acquired parallel to the dam structure in the east–west direction (Fig. 8) with a length of 160 m and electrode spacing of 2.5 m. Two other profiles (C–C′ and D–D′), each with a length of 360 m and electrode spacing of 5 m were acquired along the main course of Wadi Hanifa in a direction perpendicular to the dam structure (Fig. 8). The resistivity measurements were acquired using the Syscal Pro Switch-72 resistivity meter (IRIS Corp.).

Fig. 8
figure 8

Sketch map showing the locations of the surveyed profiles

Full size image

The field resistivity raw data have been subjected to different and successive processing and filtering operations. The raw data is uploaded into the Prosys II software for processing and filtering. Using this software, the raw resistivity data was filtered and converted to absolute values. The data was inspected for bad points (data points having too high/low apparent resistivity values) that have been removed when necessary. The filtered data is exported through Prosys II software into RES2DINV format for further editing and removing the bad data points which have wrong resistivity values using the RES2DINV software (Geotomo 2011). These bad data points appear to have too high or low resistivity values in comparison with the neighboring data points. The least-square method inversion with smoothing of the model resistivity was applied on the filtered 2D electrical resistivity data using RES2DINV software to obtain better and refined results with the minimum error (Loke and Barker 1996). The optimum solution of the resistivity model is calculated where the least difference between the measured and calculated apparent resistivity values with the minimum root mean squared (RMS) error was reached. The RES2DINV program uses an iterative method whereby starting from an initial model, the program tries to find an improved model whose calculated apparent resistivity values are closer to the measured values. The 2D inversion results were interpreted to delineate the subsurface structures within the investigation depth.

The radar data were collected using a GPR SIR 3000 system equipped with a 400-MHz antenna at a sampling rate of 50 scans per meter. The acquisition range for the 400-MHz antenna was set to about 30 ns, with the expectation of reaching depths of up to 3 m. The antenna transmission rate was set to 100 kHz (for the transmitter). The data were collected with a trace spacing of about 2 cm and with 512 samples per trace. These data were collected using a survey wheel, which enabled distance measuring mode for the GPR system. The wheel of the GPR survey wheel system was calibrated on site to reduce the distance measurement error due to surface irregularities.

In this study, the GPR survey was applied to image the concrete structures, steel reinforcing bars of the stilling basins in-front of the dam structure, and the possible cavities, voids, and/or degree of fracturing of the rock mass beneath these basins (Fig. 8). The stilling basins are located on the downstream side of the spillway of the dam that is connected to a main wadi channel. The stilling basins are a conventional method of energy dissipation of water flow, because of the high velocities of water at the end of a spillway, through the use of energy dissipaters, such as baffle blocks (Fig. 9). In this respect, six radar profiles were measured along the stilling basins and two profiles extending from east to west were measured just outside of the stilling basins site. During the acquisition process, the data were filtered using frequency and spatial filters to attenuate noise and eliminate electromagnetic interference. In addition, the acquired radar data was subjected to different and successive signal processes in order to get the most appropriate and interpretable subsurface image (Daniels 2004). In this study, the software Reflex 7.5 (Sandimeier 2014) was used for the data processing. The applied processing steps include time cut, background removal, subtract-mean, band pass frequency, automatic gain control, and average x-y filter.

Fig. 9
figure 9

Image showing stilling basins and El-Elb dam structure

Full size image

Data interpretation and discussion

The present study was planned for two main purposes. The first purpose is to image and delineate the possible underground cavities, voids, and/or fractures of the rock masses beneath the dam site using electrical resistivity tomography (ERT) along with the field structural and geological verification. The second is to image the stilling basins and the rock beds underneath in front of the dam structure using ground-penetrating radar (GPR) technique. The expected success of these tools is based on the assumption that there is a contrast in electrical resistivity (in the case of electrical resistivity imaging) and in electromagnetic energy amplitude reflections (in the case of GPR) between air/clay-filled cavities, soil-filled collapse structures, and the fractured rock masses and the surrounding massive bedrock (McCann and Forde 2001; Polder 2001).

Electrical resistivity tomography

Basically, the electrical properties of rocks are affected by porosity, permeability, temperature, clay, and water contents. In case of the geoelectrical resistivity survey, the current flow through the rock material by electrolytic process, therefore, voids/cavities and fractured features in the limestone creates a quantifiable disturbance in the electrical conductivity with a magnitude directly proportional to the filling material, degree of fracturing and degree of weathering (Ward 1990; Andrews et al. 2013). Accordingly, the interpretation of the present ERT data is done based on the expected ranges of resistivity values of the common karstic void features and the expected filling materials, e.g., air-filled or clay-filled cavities or sinkholes; in addition to the hardness, clayey content and degree of fracturing of the limestone bedrock. Based on the abovementioned facts and the geologic implication of the range of resistivity values obtained from the resulted 2D resistivity sections, the 2D resistivity sections are interpreted and explained.

Figure 10 presents the two east–west 2D resistivity sections (A–A′ and B–B′) that show more details of the subsurface structures behind the dam body. Along the 2D resistivity section A–A′, a relatively disconnected low resistivity (about 20 Ωm) thin layer of alluvial deposits covers the underlying limestone bedrock. Along this section, an abrupt change in resistivity is noticed at the distance of about 100 and 127 m, where the resistivity drops suddenly to about 30 Ωm to a depth of about 20 m. This feature could be interpreted as a collapsed faulted zone that could be filled by wadi alluvial sediments. Relatively fractured limestone bedrock intercalated with clayey facies is dominate the ERT section with resistivity that reach 200 Ωm with the exception of a significant high resistivity anomaly that exceeds 800 Ωm at eastern side of the section, which is indicative of a hard massive limestone bedrock. These karstic features and clay intercalations are noticed in the bedrock of the Jubaila Formation that outcrops on the wadi banks (Fig. 11). Moreover, a 7.5-m-wide wedge-like anomaly is apparent near the right end of the section. This anomaly is interpreted as an abnormal wedge filled with clayey soil or possibly as result of vertical water infiltration through the karstic limestone.

Fig. 10
figure 10

Interpreted 2D resistivity sections along profiles A–A′ and B–B′ (see Fig. 8 for profile locations)

Full size image
Fig. 11
figure 11

Field photograph showing an E–W-striking joint (blue arrows), a karst feature (yellow arrow) in the limestone, and thin layers of clay (black arrow) intercalated in the limestone

Full size image

The 2D resistivity section B–B′ shows to some extent the same resistivity features as the 2D resistivity section A–A′. The section starts with a disconnected thin layer of alluvial deposits that covers the limestone bedrock. This surficial topmost layer is characterized by its low resistivity that reach 20 Ωm. Along this section, the sharp and abrupt vertical change in resistivity that noticed along the A–A′ is also noticed along the section B–B′ between the distance of 92 and 125 m to a depth of about 20 m, where the resistivity drops suddenly to about 50 Ωm (Fig. 10). This feature could be interpreted as a collapsed fractured zone that is filled by wadi alluvial sediments and its continuation along the two sections may indicates its extension along the main course of the Wadi Hanifa course at this locality. Relatively fractured limestone bedrock is dominate the resistivity section (B–B′) with resistivity values that reach 400 Ωm with the exception of a significant local high resistivity anomalies that exceeds 700 Ωm at distances 45, 80, and 120 m along the section at a depth of about 8 m, which is indicative of a large air-filled cavities (Fig. 10). Furthermore, there is an abrupt change in the resistivity at the distance of about 60–65 m along the two E-W resistivity sections (A–A′ and B–B′); indicating a possible fault contact that might control the western side of the wadi course (Fig. 10).

Figure 12 shows the horizontal and vertical variations in resistivity along the two 2D resistivity sections (C–C′ and D–D′) that were measured along the main course of the wadi on the downstream side of the Dam. The two sections begin with a relatively thin low resistivity layer with variable resistivity values that range from 7 to 100 Ωm, corresponding to Quaternary alluvial sediments composed of clays, sands, and gravel that cover the wadi floor. This layer of alluvial deposits has irregular thickness (ranges from 0.5 to 2 m) that can be explained by the irregularity of the eroded surface of the underlying limestone bedrock.

Fig. 12
figure 12

Interpreted 2D resistivity cross sections along profiles C–C′ and DD′ (see Fig. 8 for profile locations)

Full size image

Along the resistivity section C–C′, the surficial thin layer overlies a moderate resistivity layer that attains a wide range of resistivity; range from 100 to 2000 Ωm (Fig. 12). This layer is interpreted as fractured limestone layer intercalated with clayey facies, where its wide range of resistivities could be referred to the degree of fracturing and the clayey content. This layer extends to an average depth of about 15 m, except at the distance between 110 to 285 m along the resistivity section, where the layer tends to deepen to about 60 m. The abrupt change in the layer thickness may indicate faulting affected this layer, especially the layer is surrounded by a relatively high resistive hard limestone layer (exceeds 2000 Ωm). To the western side of the section, a local and distinctive resistivity anomalies with relatively high resistivity reaching 5000 Ωm (Fig. 12), indicating a possible karstic void filled with air, are also noticeable at the distance of 275 m along the resistivity section C–C′. The resistivity section D–D′ shows local and distinctive low resistivity anomalies occur at about 5 to 10 m depths at distances of 125, 160, 225, 260, and 310 m along the resistivity section (Fig. 12). These anomalies are distinguished by their low electrical resistivity values (lower than 30 Ωm) that can be attributed to karstic cavities filled with wet alluvial clayey sediments developed within the limestone bedrock by the influence of the groundwater. Along this section, the limestone bedrock is characterized by its relative moderate resistivity that reach 500 Ωm that could be attributable to the degree of fracturing that affected the bedrock and its intercalation with clay as indicated by the field investigation of the limestone bed rock on the banks of the wadi (Fig. 13).

Fig. 13
figure 13

Field photograph showing an E–W-striking joint (yellow thick arrows) and thin layers of clay (blue arrows) intercalated in the limestone. Inset is a close view showing a swollen clay layer

Full size image

Ground-penetrating radar

After the radar data processing procedures applied as mentioned previously, the interpretation of the 400-MHz profiles was carried out on the basis of horizontal variations and appearance of reflections. These variations include strength, continuousness, and attenuation and lateral relations. The inspection of the processed radar sections that were measured on the stilling basins (Fig. 8) reveals the presence of a concrete zone that starts with a stratified appearance of reflections indicating the presence of a gravel-fill layer with a thickness of about 50 cm that includes coarse sand, gravel, crushed stone, and a concrete leveling layer (Figs. 14 and 15). This gravel layer overlies an anomalous high conductive zone with multiple zig-zag reflections, indicating the reinforcing steel mesh (Capineri et al. 1998) of the enforcement concrete zone (denoted C in Figs. 14 and 15).

Fig. 14
figure 14

Example radar section recorded (GPR line-1) on the stilling basins using a 400-MHz antenna indicating zig-zag vertical reflections (C), the reinforcing steel of the reinforcement concrete layer and cavity and void features indicated by their multiple hyperbolae signatures, intra-bed attenuated reflections indicating collapsed fractured zone (FCZ), and discontinuities on the reflectors indicating that fractures affected the limestone bedrock

Full size image
Fig. 15
figure 15

Example radar section recorded (GPR line-4) on the stilling basins using a 400-MHz antenna indicating zig-zag vertical reflections (C), the reinforcing steel of the reinforcement concrete layer and cavity and void features indicated by their multiple hyperbolae signatures, intra-bed attenuated reflections indicating collapsed fractured zone (FCZ), and discontinuities on the reflectors indicating that fractures affected the limestone bedrock

Full size image

The radar sections were examined and interpreted for possible void and/or fracturing features underneath the concrete zone of the stilling basins as indicated by the pattern of fractures and associated karstic features of the rock outcrops (Figs. 12 and 13). The void and cavity features are detected in the radar sections by their characteristic diffraction patterns (Benson and La Fountain 1984), where the cavities often generate a reflection pattern consisting of multiple hyperbolae (Yelf 2007). Based on the signature of cavity and void features in the radar record, cavity features are delineated at depth of about 1.5 m beneath the enforcement concrete zone of the stilling basins (Figs. 14 and 15). In this respect, conductive fractured clay-filled zones are reflected as an attenuated and absorbing zone (Reynolds 2011) on the radar sections (denoted as FCZ in Figs. 14 and 15). Losing of coherency or continuity, bending and scattering of reflections are also considered indicators for existence of karstic features. The discontinuity and displacement of reflections on the radar section (Figs. 14, 15, and 16) indicates fractures affect the limestone bedrock (Grandjean and Gourry 1996; Grasmueck 1996).

Fig. 16
figure 16

Example radar record recorded (GPR line-6) in the front of the stilling basins area on the wadi floor using a 400-MHz antenna indicating cavity and void features shown by their multiple hyperbolae signatures and discontinuities on the reflectors indicating fractures affecting the limestone bedrock, in addition to intra-bed attenuated reflections indicating collapsed fractured zone (FCZ)

Full size image

Inspection of the other radar sections that were measured in the front of the stilling basins area on the wadi floor indicates that hyperbolic features are present at depth of about 1.5 m along the radar section, indicating the existence of karstic voids (Fig. 16). It also shows that around the sinkholes significant fractures extend through the limestone bedrock, appearing in the sections as discontinuities on the reflectors (Fig. 16).

Conclusions

In the present study, the subsurface existence of near-surface karst and fracture features and their lateral and vertical extent were investigated at the site of the El-Elb Dam using electrical resistivity tomography (ERT) and ground-penetrating radar (GPR) techniques. The dam generally rests on the Upper Jurassic Jubaila Formation, which consists mainly of limestone that is affected by fracturing and karstification, as indicated by the pattern of fractures and associated karstic features of the outcropped rock. These karst and fracture features represent a critical hazard to the safety of the dam as these features of potential weakness and could play a role as seepage paths. With this concern, four ERT profiles were measured on the downstream side of the El-Elb Dam; two profiles were acquired parallel to the dam structure, and others were collected at a perpendicular direction using a Syscal Pro Switch-72 resistivity meter. In addition, a GPR survey using a 400-MHz antenna and a SIR-3000 instrument was conducted along five profiles above the stilling basins on the downstream side of the spillway of the dam and one profile was measured outside of the stilling basins area on the course of the wadi. The measured geophysical data were interpreted carefully with consideration of the field-based structural and stratigraphic investigations of the surface outcrop rocks at the dam site on the banks of the wadi course.

The GPR imaging revealed the presence of karstic cavities and fractures affected the limestone bedrock at depth of about 1.5 m underlain the stilling basins. Furthermore, in some localities, the fractured zones are filled with clayey materials as indicated from the inter-bed attenuated reflection on the radar sections. The 2D resistivity sections revealed the presence of these local and small extended void features at shallow depths, as well as a possible collapsed fractured zone filled by wadi clayey sediment continued along the main course of the Wadi Hanifa course at this dam locality. The ERT sections indicate also a possible faulting affected the site and delimited the above interpreted collapsed fractured zone. The resistivity sections give indications on the presence of clayey intercalations within the limestone bedrock of the Jubaila Formation, especially in the western side of the wadi course. These karstic features near the surface and close to the dam body certainly represent suitable pathways for water leakage, especially when they have lateral and vertical extension as indicated from the ERT sections. In addition, the presence of these weakness features close to the dam body, especially under the stilling basins structure, might represent a threat to the safety and stability of the dam.