Keywords

1 Introduction

Understanding the factors that causing natural disasters and hazards whether natural or anthropogenic factors have become essential to safe people life and assessing the development in arid and semi-arid regions. Land subsidence can substantially jeopardize human lives and infrastructures . This phenomenon represents a response to the differential subsiding of the ground surface due to the extent reduction of belowground materials by compaction or collapse of subsurface cavities (Ren et al., 1989). Land subsidence a result of volume reduction, either by natural or anthropogenic activities , is typically a slow process happening over nearly large areas. The rates of subsidence have been traditionally measured using ground-based surveying methods (e.g., geodetic leveling and global positioning systems ) (Galloway & Burbey, 2011). Yet, the advancement of the Interferometric Synthetic Aperture Radar (InSAR) technologies enabled cost-effective and timely estimations of land subsidence over large-scales (Gabriel et al., 1989). Examples of these applications include; some areas in California and Las Vegas such as Antelope, Santa Clara, and Las Vegas valleys (Galloway et al., 1998; Galloway & Hoffmann, 2007), The Nile Delta in Egypt (Becker & Sultan, 2009; Rateb & Abotalib, 2020) and central Arabia (Othman & Abotalib, 2019). Moreover, (InSAR ) technologies have been recently used to detect and quantify sluggish, large property displacements all over the globe (Hooper, 2006; Kampes, 2006), such as technologies enabled cost-effective San Francisco Bay Area (Bürgmann et al., 2006) and central Mexico (Chaussard et al., 2014). The relationship between radar-detected subsidence and related spatial - temporal measurements sheds light on the factors that cause the reported subsidence . (Higgins et al., 2014). These factors usually include groundwater or oil/gas extraction rates, rapid urbanization (Al-Jammaz et al., 2021), decline in groundwater levels , and the distribution of different lithologies . For example, significant subsidence of over 40 mm over 35 days in the oil fields over Lost Hills and Belridge in San Joaquin Valley, California was attributed to immoderate oil production from depths of about 700 m under the ground (Fielding et al., 1998). Subsidence and related deformations such as fissures , sinkholes , and earthquakes , have been reported in many arid and semi-arid regions, where excessive groundwater extraction is a common practice. In Las Vegas Valley, between 1992 to 1997, and by using geodetic and satellites applications such as GPS and InSAR data, deformational features, and land subsidence of up to six centimeters by year were discovered due to intensive pumping of groundwater (93 km3/year) (Bell et al., 2002). Significant land deformation and subsidence were also reported from Mashhad city northern Iran , where InSAR measurements indicated subsidence of∼28–30 cm/y between 2003 and 2005 along the axis of the Mashhad valley . These rates were set on as main factors caused by extensive extraction of the aquifer system as evident by analysis of piezometric records (i.e. 65 m of water table decline since 1960s) (Motagh et al., 2007). Moreover, Earth fissures were reported from the Sarir South agricultural project over the Nubian Sandstone Aquifer System (NSAS) in Libya , where water levels dropped by more than 6 meters leading to compaction of fine-grained sediments within the aquifer materials and development of earth fissures (El Baruni, 1994). In the same context, using GPS data from two stations (period: 2006 to 2009), subsidence rates of up to 10 cm per a year were reported from the Quetta Valley in Pakistan, where subsidence was also attributed to excessive groundwater abstraction to support progressive population growth (1975: 260,000; 2010: 1.2 million) (Khan et al., 2013). Groundwater-extraction-induced deformation was reported for both confined and unconfined aquifers (Galloway et al., 1998; Othman & Abotalib, 2019). The reported deformation was either uniform and extend over large areas, or differential, causing earth fissures , damage in buildings and infrastructures (Burbey, 2002) and might involve horizontal displacement in addition to the vertical component (Burbey, 2001; Galloway & Burbey, 2011).

The largest number of subsidence cases were documented in arid lands, where precipitation was minimal, limited or no surface water resources were available, intensive irrigation programs were implemented mainly on groundwater as a source for freshwater (Maliva & Missimer, 2012) such as the Saharan-Arabian desert belt. Aquifers in the Saharan-Arabian deserts are fossil aquifers that were recharged mainly during the Quaternary wet climatic periods (Abotalib et al., 2016, 2019a). Currently, they receive only limited local recharge , which is way lower than the annual abstraction (Sultan et al., 2014). Examples of these aquifers include the Mega Aquifer System (MAS) in Saudi Arabia (Hoetzl et al., 1978; Sultan et al., 2014; Abotalib et al., 2019b), the Nubian Aquifer System in northeastern Africa (Abdelmohsen et al., 2020; Mohamed et al., 2017), the North-Western Sahara Aquifer System in northwestern Africa, and the Great Artesian Basin in Australia (Taylor et al., 2013).

Overstimulated exploitation of non-renewable groundwater in such regions, and for certain fossil aquifers , will be expected to cause artesian head downturn, high extensional stresses, and other problems.

In this study, we are investigating on the LMAS of Saudi Arabia , one of the world’s largest fossil aquifer systems, examine comparable cases elsewhere, and note the similarity with our results from the LMAS to those published from similar areas around the world. The Mega Aquifer System (MAS) (surface area: ~2 million km2) spans the Arabian Peninsula ’s Arabian Shelf in Saudi Arabia , United Arab Emirates (UAE), Qatar , Yemen , Oman , Jordan and Kuwait . The aquifer currently is the major water resource for mega-agricultural projects and within these reclaimed areas, several ground deformations and subsidence features were reported. Examples of these areas are the Al-Yutamah valley, Tabah village, and Najran valley (Roobol et al., 1985; Vincent, 2008; Youssef et al., 2014).

The government of Saudi Arabia started in the mid-eighties an ambitious agricultural development program in central Arabia (e.g. Al-Qassim and Ha’il regions central Saudi Arabia ), which led to a raise in groundwater extraction (1.9 × 109 m3/year in 1984 to 4.4 × 109 m3/year in 2004) and cultivated lands (from 213 × 103 hectares in 1984 to 316 × 103 hectares in 2004). Likewise, widespread agricultural projects were employed in northern Saudi Arabia (e.g. As-Sirhan valley northern Saudi Arabia ) leading to a raise in groundwater extraction from 0.095 x 109 m3/year to 2 × 109 m3/year, the cultivated areas have grown from 14 × 103 hectare in 1984 to 164 × 103 hectare in 2004 (MOAW, 1984, 2004; Abunayyan and BRGM, 2008).

Our integrated study includes a regional investigation over LMAS in northern and central parts of the KinSaudi Arabia and a local investigation over the Wadi As-Sirhan Basin (WASB) and Al-Qassim Region (Fig. 20.1). The regional investigation implemented field observations to examine deformational land features and applying spatial correlation for features with other relevant datasets in term of causal effects.

Fig. 20.1
figure 1

Location map of the regional study area and the investigated InSAR areas over Wadi As-Sirhan Basin and Al-Qassim Region. The boundaries and outcrops of the Lower Miga Aquifer System

Full size image

2 Geology and Hydrogeology of the Study Area

The Arabian Shield ’s basement complex crops out along Red Sea margins, creating the Red Sea mountain range (Red Sea Hills) (Fig. 20.2). Successions of Phanerozoic sedimentary overlie the basement complex unconformably, with east-dipping and approaching thicknesses of ten kilometers near the Arabian Gulf (Powers et al., 1966; Lloyd & Pim, 1990; Konert et al., 2001; Margat, 2007). The Red Sea mountains outcrops are younger to the eastward of the Gulf (Quaternary age) and older to the westward (Cambrian age) (Fig. 20.2). The Red Sea rifting exposed the basement complex and the overlaying Phanerozoic formation, which facilitates aquifer recharge within the surface exposures of these formations. The Red Sea opening rejuvenated many of the Proterozoic NW-trending Najd fault systems that affect the crystalline rocks of the Arabian Shield (Fairer, 1983; Morsy & Othman, 2020). The fault rejuvenation was dip-slip (e.g. normal faults and grabens ) in response to the extensional events associated with Red Sea rifting during the Oligo-Miocene time (Kellogg & Reynolds, 1983; Giannerini et al., 1988). These faults are represented by the Kahf fault system in northern and central Saudi Arabia (Othman et al., 2018). Similar observations of dip-slip displacement on Najd faults were also reported in the Egyptian side of the shield (Sultan et al., 2011).

Fig. 20.2
figure 2

Regional geological map of Saudi Arabia . (After Powers et al., 1966)

Full size image

Phanerozoic sections contain one of the largest fossil aquifers in the world which formed from sandstone and limestone sequences (Figs. 20.1 and 20.2). The LMAS was primarily recharged during previous wet climatic conditions , but it is still receiving a modern recharge at the moment, especially in southwest Saudi Arabia , where AAP rates of up to 800 mm/y have been recorded (Alharbi et al., 2014). Groundwater Cl-36 ages attained close to 1 million years from some aquifers in the southeastern part of Saudi Arabia (Sultan et al., 2019). Those previous wet climatic conditions are completely different from current arid conditions with an average annual precipitation of 100 mm/y and minimal surface-water resources over Saudi Arabia (Barthélemy et al., 2007).

The main aquifer units within the MAS include formations of Paleozoic sandstone , Mesozoic-Cenozoic carbonates, Paleocene, and Neogene sedimentary formations (Wagner, 2011). The MAS geological have been split into two major groups separated by the anhydrite layer (Hith formation) (Othman et al., 2018). The topmost is the Upper Mega Aquifer System (UMAS), including Dammam, Rus, Umm Er Radhuma, Aruma, Wasia, Biyadh formations. Then Lower Mega Aquifer System (LMAS), which contains Dhruma, Minjur, Tawil, Tabuk , Wajid, and Saq formations. The area of MAS is ~236.3 × 104 km2, respectively, respectively.

3 Methodology

We briefly describe the technical approaches applied in this study to monitor the land deformation over Saudi Arabia using (1) Field observations and land deformation database, (2) TWS and GWS extracted from GRACE temporal monthly solutions, (3) Radar Interferometric data and (4) precipitation extracted from Tropical Rainfall Measuring Mission (TRMM) data. The spatial and temporal integration of these data sets with other relevant data sets (geologic, geochemical , remote sensing ) provides more evidences of the factors controlling the reported deformational features (e.g. sinkholes , fissures , and subsidence ) announced in the arid environmental lands over the fossil aquifers .

3.1 Gravity Recovery and Climate Experiment (GRACE) Data

GRACE is a satellite mission launched in 2002 by the German Aerospace Center (DLR ) and National Aeronautics and Space Administration (NASA) for the temporal global variations measurements in the Earth’s gravitational field (Tapley et al., 2004) that represents the temporal variation in the TWS (Wouters et al., 2014; Ahmed et al., 2016; Mohamed et al., 2017; Ahmed & Abdelmohsen, 2018; Sultan et al., 2019; Abdelmohsen et al., 2019, 2020). TWS refers to the vertically integrated water storage measurement that includes groundwater, soil moisture storage, surface water storage , snow water storage , and canopy water storage (Strassberg et al., 2007; Abdelmalik & Abdelmohsen, 2019).

Release (05; RL05), with a spatial resolution of 0.5° × 0.5°, temporal monthly constrain solutions (mascon; Save et al., 2016) covering the investigated time period (4/2002 to 6/2016) were utilized to estimate the long-term components of TWS changes over the lower MAS and the WASB. The mascon solutions were provided by Center for Space Research (CSR) at the University of Texas (Save et al., 2016). The mascon solution doesn’t require any post processing, analysis, filtering, or scaling factors (Save et al., 2016).

The long term trend TWS component was extracted by continuously fitting a linear regression function (Fig. 20.3) and the trend uncertinity was calculated using an approach described in Scanlon et al. (2016). The change in groundwater storage was calculated using a common equation of water mass:

$$ \varDelta GWS=\varDelta TWS\hbox{--} \varDelta SMS $$
(20.1)

Where ΔGWS and ΔSMS represent the variation in groundwater storage and soil moisture storage, respectively. Global Land Data Assimilation System (GLDAS) model was utilized to derive SMS. land surface modeling system developed by NASA as a simulation model for hydrologic and climatic variables (Rodell et al., 2004). Four GLDAS models to estimate the soil moisture:, Community Land Model [CLM], Noah, and MOSAIC, Variable Infiltration Capacity [VIC] were used to represent the soil moisture over the study area (Rodell et al., 2004; Dai et al., 2003). The uncertainty (σ_SM) were calculated from the standard deviation (STD) of the trends from the GLDAS four simulations . The trend uncertainty in GWS (σGWS) was calculated using standard error propagation equations:

$$ {\upsigma}_{GWS}=\sqrt{{\left({\upsigma}_{TWS}\right)}^2+{\left({\upsigma}_{SM}\right)}^2} $$
(20.2)
Fig. 20.3
figure 3

Secular trend (mm/y) for GRACE-derived TWS for the Arabian Peninsula extracted from temporal (04/2002–03/2016) CSR 1° × 1° Mascon solutions showing high TWS depletion rates over the study areas

Full size image

3.2 Radar Interferometric Data (Persistent Scatterer Interferometry )

Twenty-nine descending scenes from the European Remote Sensing ENVISAT satellite were implemented. Twenty-two of those scenes covered the Wadi As-Sirhan Basin (Tabarjal and Busayta area, tracks: 221, 351 and 493) and the remaining eight scenes were acquired over Al-Qassim area (Track: 49 and 278). The investigated period for the first set was from 2003 to 2012, while the second set covered the period between 2003 and 2005. The spatial baseline ranged from 15 to 855 m concerning the master scenes. The maximum temporal baseline was 1351 days for both investigated areas, and the ranges of permanent scatterers’ density from 10 to 251,351 days for both investigated areas, and the ranges of permanent scatterers’ density from 10 to 25 persistent scatterers/km2 with a coherence threshold value of 0.5 and an amplitude dispersion value of 0.4.

Persistent Scatterer Interferometry (PSI) techniques has been applied (Hooper et al., 2004, 2007) to examine the spatial variations in subsidence rates across the two sites on the Northern and southern parts of the study area (LMAS) (Fig. 20.1). The PSI technique restrict the unwrapping phase and analysis to coherent pixels , containing individual scatterers and remaining stable over the investigated period. Buildings, rock formations, rocks, houses, utility poles , and well platforms were classified as persistent scatterers in the study region using the Stanford Approach for Permanent Scatterers (StaMPS) algorithm (Hooper et al., 2007, 2012). The interferometric was processed using the StaMPS process (Hooper et al., 2004). The accurate orbit information was provided from the Delft Institute for Earth-Oriented Space Research (Scharroo & Visser, 1998).

3.3 Landsat Images

Multi-temporal Landsat (5, 7, and 8) images (path 172 and row 39) were been acquired for the following dates (February 1987, 1991, and 2000, March 2003, January 2014, and February 2017) then were analyzed to monitor the development of cultivated lands throughout the past three decades (1987 to 2017). Temporal changes in the agricultural developments areas were estimated using the NDVI (Normalized Difference Vegetation Index ). Using the reflectance values from the near-infrared and red bands, an NDVI value was estimated using standard model and process (Rouse et al., 1974):

$$ \mathrm{NDVI}=\left(\mathrm{near}\ \mathrm{infrared}\hbox{--} \mathrm{red}\right)/\left(\mathrm{near}\ \mathrm{infrared}+\mathrm{red}\right). $$
(20.3)

It is a positive relationship between the values of NDVI and density of vegetation. The cultivated land areas were extracted from the Landsat imagery using a threshold NDVI value of 0.3; visual comparisons with Google Earth images validated the threshold value.

4 Discussion and Findings

Our analysis of the land satellite images and GRACE solutions over the largest aquifer systems in Saudi Arabia revealed that land subsidence features can be targeted excessive extraction of groundwater. In the next parts, we going to describe and explain our findings over the selected areas in relation and relevant deformation factors.

4.1 The Lower Mega Aquifer System

The excessive groundwater abstraction of the LMAS and unsustainable utilization of the aquifers was reflected in TWS from GRACE solutions. Depletions in TWS and GWS derived from GRACE over LMAS (areas: A and B) were observed. The spatial and temporal changes in TWS anomaly over Saudi Arabia and entire the Arabian platform is shown in Fig. 20.3. Positive TWS trends indicate an increase in water mass with time and vise versa in the areas of the negative TWS trends. Figure 20.3 depicts that the study area over LMAS and WASB (Fig. 20.3 area A) are experiencing a significant decrease in TWS trend (red and orange), compared to the southern areas (Fig. 20.3 area B (Al-Qassim Region)) that extend from lower depletion (yellow and orange) to near-steady statuses (blue) over the Red Sea seaboard plain. Figure 20.3 shows the temporal differences in GRACE-derived TWS, and the secular trend over the WASB in LMAS (area A) (−8.7 ± 1.1 mm/year; −3.1 ± 0. 3 km3/year) and Al-Qassim Region (area B) (−9.5 ± 1.4 mm/year: −1.7 ± 0.2 km3/year). Using TWS derived from GRACE solutions combined with soil moisture derived from GLDAS applied to Eq. 20.1, GWS time series for both WASB and LMAS were derived (Fig. 20.4).

Fig. 20.4
figure 4

Time series and secular trends for TWS and GWS over the area A and area B in the regional study area (LMAS)

Full size image

The observed TWS and GWS seasonal periodical cycle were controlled mainly by groundwater abstraction. Temperatures in summers (June–August) are high (range: 28° to 37 °C), compared to winter season temperatures which range between (6° to 19 °C), and the rates of evaporation is between 25 to 28 mm/day in the summers season and as low as 0.4 mm/day in the winter seasons (Alsharhan et al., 2001). As a result, agricultural activities and groundwater abstraction got lower during the winter season (December–February) and decline in the summer seasons. Excessive abstraction in the winter season caused water-level declines, and surface land subsidence and the replenishment (gaining mass) during the summer season by infiltration and groundwater flow. Since there is a losing mass rate of −4.2 ± 0.7 km3/year in GWS over the study basin, the extracted groundwater does not fully compensated.

The fissures near the boundaries of the subsidence bowls surrounding Busayta area in WASB and in the southeastern part of Al-Qassim area (Fig. 20.5A and B) were most probably caused by curving beam transforming around the area of deformation , the regions have experienced high lateral extension (Bell et al., 2002; Othman et al., 2018). Fissures also can result from the lateral (horizontal) forces linked with sediment compaction or horizontal pressures caused by extraction; all of these mechanisms produce lateral (horizontal) strain extended along the weakness planes (Helm, 1994). This mechanism could have caused Earth fissures that are located proximal to faults. The fissures close to the Kahf fault system in Area A and B in Figs. 20.5 and 20.6 are good examples of this deformation mechanism.

Fig. 20.5
figure 5

“A” PSI results over Wadi As-Sirhan (Busayta and Tabarjal areas) “B” and the Al-Qassim region. The distribution of the major faults are also depicted with black lines. (Modified after Wallace et al., 2000)

Full size image
Fig. 20.6
figure 6

Field photos showing land fissures and sinkholes proximal to the land subsidence in the study areas

Full size image

Our research findings are consistent with land subsidence phonomena associated with aquifers depletion which reported from numerous places around the world (e.g., Poland, 1960; Galloway et al., 1998; Hoffmann et al., 2001; Burbey, 2002; Bell et al., 2002; Christiansen et al., 2007; Assumpção et al., 2010; Galloway & Burbey, 2011; González et al., 2012; Avouac, 2012; Martínez et al., 2013; Amos et al., 2014). The detected subsided lands in these areas were attributed to increased adequate pressure that causes layer compaction and land deformation and subsidence (Donaldson et al., 1995). The effective stress balanced the downward stress effect due to the overload the overlying rock, water effective stress and pore pressure balanced the downward stress from the overlying rock and water. The changes in groundwater extraction-induced stresses trigger earthquakes and land deformations that would have occurred later when natural stress is accumulated (Avouac, 2012).

The illustrated land subsidence in the northern central part of the Kingdom of Saudi Arabia has a similar patter to other global aquifer systems in terms of theresponse to overexploitation of groundwater and the significant associated drawdown of water table such as in the Mojave Desert, California (Galloway et al., 1998), central and northeast Iran (Motagh et al., 2008) and in North China Plain (Guo et al., 2015). Such aquifers are expecting to show near-steady GWS secular trends, and stable velocity (0 radar averaged velocity) (Chew & Small, 2014). The case from LMAS is different because the fossil aquifer was primarily recharged in previous wet climatic periods yet still receives a small amount of recharge (Sultan et al., 2008). LMAS is not in near-stable state condition according to two observations. Which groundwater abstraction from the LMAS was not compensated for by replenishment,: (1) the depletion in TWS (−4.8 ± 0.7 km3/year) and GWS (−4.2 ± 0.7 km3/year) over the investigated period (Figs. 20.3 and 20.4) the radar interferometric technique studies over WASB and Al-Qassim Regions produce deformation rates of −4 to −15 mm/year (Fig. 20.5A and B). The regional deformation in the study basin is attributed mainly to a rise in effective stress that accelerate compaction and land subsidence (Fig. 20.5). The balance between the downward stress from the overload of the upper layers and the upward buoyancy force by the groundwater is challenged by the overexploitation from the Saq aquifer, where the rate of annual groundwater withdrawal exceeds 100 times the rate of annual recharge (et Miniéres, B.D.R.G., 2006). This overexploitation reduce pore pressure in the aquifer layers causing expand effective stress, which causes compaction and land subsidence in this location (Donaldson et al., 1995).

The MAS in the upper part Saudi Arabia shows the vertical hydrological connection between the aquifer layers due to the structural control and preferred pathways for groundwater flows (e.g. Kahf fault system ) and the geological incompatibility associated with differential deposition of aquifer layers (Othman et al., 2018) (McGillivray & Husseini, 1992), (et Miniéres, B.D.R.G., 2006). The vertical connection between the deep gaseous-rich Saq aquifer (e.g. H2S and CO2; Kalin et al., 2001) and the upper Paleozoic and Triassic hyrologic layers leads to enhanced evaporation and leakaes from Saq aquifer and dissolution of limestones formations and forming the karst topography during the wet climatic periods and associated elevation of the groundwater level (Abotalib et al., 2016). These karst remain mechanically stable under different climatic conditions untilwater table drawdown, which causes instability over the pre-existing subsurface cavitis. This kind of mechanism was reported as the primary mechanism of formation of falling features in the Arabian Peninsula to the eastern region (Abotalib et al., 2019b). It can illustrate the the formation of local land fall features in the region of study. More evidence of support the latest mentioned mechanism is supported by the field observation/description of sinkholes northern of Buriydah City in Al-Qassim Region (Youssef et al., 2016). The abovementioned mechanisms, which induced by groundwater dynamics, however controlling the study basin doesn’t rule out the ifflluence of the geotechnical properties of the lithological units on the surface deformation in the study basin. The illustrated features of cracks and fissures in Fig. 20.6 might be affected by the existence of near surface approximately thick layer of up to 15 m with high plastic clay beds, causes upward rupture structure when sub-ground layers damping from close agricultural activities extends these beds (Al Fouzan & Dafalla, 2013). Consequently, the mixture of dissolution and subsequent sagging and landfall of the upper layers due to the drop down of groundwater level and the existence of high plastic clay layers with the hydrological and geological factors and the availability of suitable geological and hydrological conditions (e.g. loss of buoyancy) can cause the reported regional and local landfall features in the study basin.

5 Summary and Implications

The LMAS, one of the highly hydrologic conductive aquifers in the Middle East and North Africa , represents an significant source of groundwater abstraction for irrigation in the upper part of Saudi Arabia . Mainly, this sandstone fossil aquifer of the Paleozoic age has been inordinately mined throughout the past 50 years, causing regional groundwater table dropdown of up to 150 m. The subsidence locations inferred from the InSAR processing maps strongly correlate with GRACE-derived TWS results suggesting that the subsidence can result from excessive groundwater extraction resources. The rapid increase in the cultivated lands between 1987 and 2012 supports this phenomenon (Fig. 20.7) that continues to our present time. Findings from the present study support that excessive fossil groundwater abstraction is critical for forming land subsidence , collapse features, and earth fissures in the study areas. However, the formation of such features might be associated with other factors. The studied approach could provide a simplified and cost-effective method to assess the surface land deformations and landscape associated with anthropogenic and/or natural induced variations in the groundwater tables under arid /semi-arid to hyper-arid conditions.

Fig. 20.7
figure 7

Time series for Landsat 5, 7 and 8 scenes acquired (February 1987, February 1991, February 2000, January 2012) over Busayta area in Wadi As-Sirhan Basin (WASB). (Source: Earth Resources Observation and Science (EROS) Center)

Full size image