Introduction

The Nile Delta is highly populated, with a high population density compared to other areas on Earth. The delta region accommodates a population of 50 million individuals, with a population density of 2300 persons per square kilometer. It is projected that this density will increase twofold within the next 50 years [1]. The Nile Delta holds significant strategic and economic importance for Egypt, primarily because of its ample natural resources, including oil, gas, black sand, and groundwater. Additionally, the region boasts fertile land and hosts several prominent archaeological monuments. The delta is currently experiencing degradation as a result of diverse environmental mechanisms, including subsidence of the earth and the increase in sea level, and shoreline flooding [2]. The convergence of escalating population growth and the repercussions of climate change, such as rising sea levels, portends a grim outlook for the deltas. Ground subsidence has been recurring frequently lately, due to the poor quality of limestone in Egypt, which in turn leads to ground subsidence with large cracks, as well as high temperatures, which “crinkle” the asphalt, There are other reasons that cause ground subsidence, including increased loads that exceed the strength and bearing of the road surface, Human activities such as groundwater extraction, oil and gas extraction, and land use change can also contribute to subsidence and subsidence in coastal areas. Groundwater extraction can cause underlying aquifers to collapse, causing subsidence in surrounding areas. Likewise, oil and gas extraction can cause subsidence due to the removal of fluids from the ground. Land use changes such as urbanization and deforestation can also lead to subsidence, as the weight of buildings and infrastructure compresses the underlying soil [3].

Climate change is a significant factor in the decrease. The combination of rising sea levels and intensified storm activity can result in erosion, which in turn can cause subsidence, particularly in the vicinity of coastal regions. Moreover, alterations in precipitation patterns have the potential to modify the quantity of water present in the soil, resulting in variations in soil density and subsidence [4].

Ground equipment, including mobile laser scanners, automated total stations, and ground sensors with sub-centimeter accuracy that are helpful for maintenance procedures, has historically been the basis for the quantitative evaluation of ground displacement altering infrastructure [5]. An additional option that may be seamlessly integrated into ground monitoring is space-borne differential SAR interferometry (DInSAR). Utilizing DInSAR for civil infrastructure monitoring is still in its infancy and is not yet being fully utilized [6]. Sentinel-1 is an interesting source of information due to its enormous volume of accessible data, which is captured more regularly and with high precision at cheap prices [7]. This work’s technology relies on sophisticated DInSAR methods [8, 9] which leverage long-term SAR acquisition sequences obtained in the same geometry. This allows us to generate time series with centimeter-to-millimeter-level precision [10] and extract valuable data on the temporal and spatial trends of displacement detected. An improved version of the InSAR technology, the DInSAR method, uses the difference in phase between a sequence of SAR pictures taken at various periods on an identical scene to quantify sub-centimeter ground movement with millimeter accuracy [11]. In many different geophysical contexts, including seismic, volcanic, and mass movement scenarios, DInSAR displacement time series have been extensively used with the dual goal of mapping and tracking observed displacements across huge regions [12, 13]. The ability of satellite-based DInSAR to comprehensively and consistently survey extensive areas from a remote location is one of its main advantages. This feature renders it a superb instrument for monitoring and managing transportation infrastructure networks. The utilization of the DInSAR approach is increasingly recognized as a highly efficient and cost-effective means of monitoring the condition and integrity of structures and infrastructure [14]. Satellite-derived DInSAR data are becoming more and more trustworthy for long-term, broad-area displacement monitoring. This technique’s primary contribution is the use of numerous point displacements that have been detected and are dispersed across the whole structure [15]. In the area of infrastructure, the initial uses stretch back more than 20 years. DInSAR is now employed to research curettage processes and structural failures using historical SAR data.

Numerous research works have utilized geodetic methods [e.g., Global Positioning System (GPS) and Interferometric Synthetic Aperture Radar (InSAR)] to investigate the pace and patterns of deformations in the Nile Delta. Based on 34 pictures analyzed using persistent scatterer interferometry (PSI) from the European Remote Sensing (ERS-1 and -2) satellites between 1993 and 2000, Aly et al. [16] determined that the rate of deformation in Greater Cairo is − 7 mm/year. The identified causes of the sinking included tectonic activity, groundwater depletion, and sediment compaction. By using fourteen PSI-derived images from ERS-1 and -2 [17] discovered that the rate in Damietta was − 8 mm per year, which they ascribed to recent (i.e., within the past 2500 years) sediment compaction. By employing a total of 14 synthetic aperture radar (SAR) images from the ERS-1 and -2 satellites [18], utilized 37 photographs of both ERS-1 and ERS-2 images based on PSI taken from 1993 to 2000. These images were used to measure the rate of land subsidence among the cities of Mansoura and Mahla, which were found to be − 9 and − 5 mm/year, respectively. This subsidence is attributed to the extraction of groundwater in the delta’s plain’s central region. The authors argued that the fluctuation in groundwater levels during different seasons was concealing the regional sinking caused by the process of sediment compaction. By utilizing 49 PSI-based ENVISAT images captured from 2003 to 2010 [19], calculated that the subsidence rate in Alexandria ranges from − 0.4 to − 2 mm per year. It was hypothesized that the sinking of the city was primarily due to the compression of sediment occurring within shorter time frames, such as decades to centuries. Conversely, the subsidence over longer time spans, ranging from centuries to millennia, was believed to be more likely caused by fault collapse, tectonic activity, and earthquakes. Gebremichael et al. [20] utilized PSI stacking to perform an analysis on a total of 84 ENVISAT synthetic aperture radar (SAR) images obtained during the years 2004 to 2010 and determined that there was a consistent and slight subsidence of 0.4 mm/year in the southern portion of the delta. In addition, they discovered a gradual increase in elevation of approximately 2.5 mm per year in the flat areas of the delta. The Menoufia governorate experiences a significant subsidence rate of 9.7 mm per year and the expansion of arid region as a result of the removal of subsurface water. An identical pattern was observed in Abu Madi, located in western Alexandria, and it was associated with the extraction of gas. Conversely, a total of 73 ENVISAT images obtained from 2010 to 2004, along with consistent GPS stations, were analyzed using PSI. Without any evidence of improvement in the region of the Nile Delta [21], discovered different valuations of sea level change in several important cities: approximately − 6.4 ± 0.4 mm/year in Cairo. An annual rainfall rate of − 4.0 ± 0.6 mm in Tanta, − 4.8 ± 0.6 mm in Mahla, − 10 ± 1.2 mm in Mansoura, − 10.3 ± 1.3 mm in Damietta, and − 4.9 ± 1.6 mm in Port Said is recorded. The subsidence was attributed to the excessive extraction of groundwater. However [22], described localized rising in the middle Nile Delta stations of Tanta, Damanhur, and Hamoul using GPS data gathered between 2013 and 2017.

SAR image processing in areas with vegetation and wetlands is highly intricate and varies both seasonally and annually. This is primarily because pixels have a low degree of coherence, particularly in agricultural and wetland areas, as well as the limited temporal coverage provided by ERS-1, 2, and ENVISAT satellite images. Providing precise estimations of the rate of decline is challenging [23]. The methodology utilized in this study is founded on advanced DInSAR methodologies. These methodologies involve the analysis of SAR acquisition sequences collected over extended time periods with consistent geometry. This enables us to extract significant information regarding spatiotemporal patterns. Displacement is measured by generating time series with a precision of centimeters.

This study investigates the subsidence of the Nile Delta by employing novel approaches that deviate significantly from previous research. These approaches involve utilizing more comprehensive data and reducing decorrelation errors to constrain the InSAR velocity. Surface vertical displacements were investigated in a section of the delta, particularly on damaged roadways with high pressure that exhibit regular visible sinking, cracks, potholes and rutting in the pavement surface. Based on 25 Sentinel-1 satellite images, we examine the most recent deformations between 2020 and 2022, notably the route between Mansoura Gamasa and Mansoura Damietta. These pictures have tiny spatial baselines and shorter temporal baselines.

The purpose of the study was to determine if vertical displacement could be reliably detected using C-band Sentinel-1 data, especially in delta environments where thick vegetation predominates and provide a summary of the observed subsidence along the road network, and to evaluate Sentinel-1 data performance in locations with notable subsidence. Additionally, utilizing methods, in situ data, and findings from earlier research, it was feasible to confirm the results by comparing them with the more complex PSI scenario by StaMPS integrated with SNAP utilized the Sentinel-1 data.

Study area

The Nile Delta is situated in the northern region of Egypt as illustrated in Fig. 1. The study area is located in part of the Nile Delta and includes the urban road network that is subject to frequent subsidence due to the high pressure on it. In order to evaluate the soundness of the road network, we first conducted a comprehensive analysis of the entire study area, before focusing on the roadway with the highest decrease rate. A detailed assessment of the deformation pattern in the research region, which includes the highway between Mansoura and Damietta and the highway between Mansoura and Gamasa, was presented using Sentinel-1 data. These roads are among the busiest and have several visible fractures, rutting, and flaws which was monitored with smartphone images.

Fig. 1
figure 1

Footprint and area of interest of the chosen Sentinel-1 master scenes for each track

Full size image

Egypt possesses an extensive road network spanning 65,050 km, with 48,000 km being paved and 17,050 km remaining unpaved. Egypt has constructed an extensive network of highways that spans approximately 18,000 km, encompassing the entire country. Egypt also possesses a network of bridges, which aids in facilitating the movement of vehicles. Currently, there are a total of 181 bridges in the country, with 36 of them spanning the Nile and 85 being suspended. Egypt has experienced a recent surge in efforts to improve and expand its highway and main road networks, with the goal of providing comprehensive coverage across the entire country.

Primary paved, secondary paved, dual carriage, unpaved, and track and path are the different categories for the main roads in the Nile Delta area. Their lengths are 1636.16 km, 10,715.26 km, 877.38 km, 11,277.23 km, 20,711.73 km, and 1.36 km, respectively. This suggests that the primary road network in the area consists mainly of unpaved and secondary paved roads and tracks, but there is also a significant presence of well-developed, main paved roads. Road network lengths in the Nile Delta region.

The Nile Delta is situated to the south of the Nile Cone, a depositional structure created through a process of discharge and sediment accumulation over a span of 5 million years by the Nile and Paleo-Nile drainage systems [24]. The Pleistocene phase contains of up to eight hundred meters of deltaic sands with minimal clay layers [25]. The surface of the Nile Delta area is adorned with a variety of rocks, ranging from small stones to fine sand, interspersed with clay deposits. The deposits in the Nile Valley and Delta were categorized into three types: protonile (Q1), prenile (Q2), and neonile (Q3) deposits. The sedimentary deposits in the northern delta region, where the Nile river meets the sea, have a total thickness of 10 km. The Nile Delta is divided into three distinct zones: the southern, middle, and northern zones. The southern zone primarily consists of coarse Nile sediments, predominantly sand deposits. The middle zone is composed of sediments of higher granularity in comparison with the southern zone, thus it is presumed to serve as a transitional zone connecting the southern and northern Delta zones. The highest quality neonile sediments in the three zones were found in the northern region. A bay was created in Northern Egypt, surrounded by the cliffs of Cretaceous and Eocene rocks to the south and southwest, which formed during the Middle Miocene period. The sediments in the Nile Delta comprise shale and clays with sandstone interbeds, suggesting a swift and uninterrupted deposition in a basin that is gradually sinking. The Modern Delta is classified as Late Pliocene and Quaternary in age, making it a relatively youthful feature that is approximately 3 million years old. The process of sediment accumulation in the modern Delta began with the deposition of the EL-Westani Formation. The EL-Westani Formation is of Late Pliocene age. It is overlain by the Pleistocene Mit Ghamr Formation, which is then followed by the Holocene Bilqas Formation [26]. In the Nile Delta, a variety of man-made and natural physical processes contribute to land lowering, including: (1) From the late Miocene epoch (8500 Ka) until the present, there has been Holocene fluvial sediment sinking as a result of several sedimentation phases [27]. The impact of this process differs across the east–west axis and is associated with both the fluctuation of global sea levels caused by climate change and local tectonic activities, as well as downslope mass movements in Mesozoic bedrocks. Additionally, the cycle of discharge in the Nile River is another physical process that affects the Nile Delta’s rates of deformation. The process of recharging and discharging facilitates the transportation of sediments from different climatic regions in the Nile, while also inducing variations in soil loading caused by the fluctuation of the water table. Additionally, human activities such as dam construction and urbanization reduce the sediment flow in the Nile River, thereby intensifying land subsidence. An example of this is the construction of the Aswan High Dam in southern Egypt in the 1970s, which has led to a substantial decrease in the rate of sediment accumulation. Currently, upstream riparian nations construct mega-dams, such the Garand Ethiopian Renaissance Dam, which can impede the laden sediments from completely moving into the delta [28]; (2) Over time, regional geological processes may potentially help to decrease the Nile Delta. Several tectonic regimes are present in the Nile Delta, including as ENE folding, NEW-SE contraction, and NWW normal faulting. (3) Excessive absorption of groundwater, oil, and gas, which permits sediments to reorganize and causes isotactic compaction [29].

Materials and methods

Dataset description

Sentinel data

The data from the Copernicus Sentinel-1 Synthetic Aperture Radar (SAR) was utilized, with descending geometries, from January 7, 2020, to December 22, 2023. We restricted our study’s scope on the data from Sentinel-1A with a 12-day repetition cycle, as indicated in Table 1. We deemed this dataset sufficient considering the projected ground displacements and the presence of an extensive archive encompassing the interest region. It is important to note that due to the lack of spatial synchronization in Sentinel-1 products, as the start and finish timings may change within every orbital path, several scenes are frequently necessary to completely encompass our designated area. The benefits of Sentinel-1 derive from its wide-ranging coverage, which spans a distance of 250 km when operating in interferometric wide mode, and its sufficient spatial resolution, which measures 5 m by 20 m in range versus azimuth. On top of that, the TOPS (Terrain Observation with Progressive Scans) imaging technology ensures uniform image quality throughout the entire range [30].

Table 1 Sentinel-1 data was used in the processing
Full size table

For the PSI method to produce accurate findings, at least 20 SAR time-series photos are needed. Consequently, time series including 25 images for a descending pass were chosen when applying the PSI to Sentinel-1 data (Table 1, Fig. 2). This includes the time series for the whole study years’ worth of descending pass geometry. To maintain the time-series Bps within the specified limit of approximately 200 m for Sentinel-1, the temporal baselines and perpendicular (Bp) of the data stack were examined. In the event that a Bp is greater than the value of the critical point, geometrical decorrelation results, making the interferometric phase unusable. Figure 2 illustrates that the time-series’ Bps typically did not exceed 200 m.

Fig. 2
figure 2

Images of star graphs displaying the perpendicular and temporal baselines of time series

Full size image

UAV data

The research used data from images taken by an unmanned aerial vehicle (UAV) and took place in Sherbin City, Egypt (Table 2). A UAV mission was initiated to efficiently capture the intricate details of the study area. Afterward, the aerial photographs obtained during the flight were analyzed using specialized software and image processing techniques, leading to the creation of photogrammetric results. The first stage of image processing involves aligning the images, while the next step includes creating a dense point cloud that contains both the point cloud and mesh. The concluding phases of this procedure involve generating a digital surface model (DSM) and an orthomosaic, which encompasses both the DSM and the digital terrain model (DTM).

Table 2 Characteristics of the UAV image data
Full size table

Smartphone images

We affixed an iPhone 13 Pro Max to the vehicle’s dashboard and captured photographs at a resolution of 600 × 600 pixels every second of roads where we observed frequent subsidence from Sentinel images over the course of the study.

Methodology

As a generic methodology, DInSAR encompasses techniques ranging from single-pair interferometry to more sophisticated methods. All of them, nevertheless, have the ability to make use of the data found in the interferometric phase, which is defined as the difference in phase between two pictures taken from nearly identical satellite positions at two separate periods. The following contributions are included in the interferometric phase [7].

$$\Delta \varphi = \Delta \varphi_{{{\text{flat}} }} + \Delta \varphi_{{{\text{disp}}}} + \Delta \varphi_{{{\text{noise}}}} + \Delta \varphi_{{{\text{atm}}}} + \Delta \varphi_{{{\text{height}}}}$$
(1)

where \(\Delta \varphi_{{{\text{height}}}}\) denotes topography and \(\Delta \varphi_{{\text{flat }}}\) is the phase contribution of the flat earth. \(\Delta \varphi_{{{\text{noise}}}}\) is residual noise, \(\Delta \varphi_{{{\text{atm}}}}\) is the phase contribution from the delay of radar wave propagation in the atmosphere, and \(\Delta \varphi_{{{\text{disp}}}}\) is the portion of the phase that reflects the ground deformation measured along the line of sight (LOS). As previously stated, DInSAR aims to eliminate all factors contributing to the interferometric phase other than displacement. For the most part, conventional DInSAR can handle them. However, the atmospheric and noise factors in Eq. (1) can only be addressed by using more reliable MT DInSAR methods, such the PSI employed in study.

PSI processing using SNAP-StaMPS

To perform the analysis using C-band data, the PSI necessitates a minimum of 20 SAR images. The PSI (persistent scatterer interferometry) method tracks changes in the Earth’s surface over extended periods of time, taking into consideration atmospheric conditions, topographic, and impacts of signal noise. The PSI makes use of interferograms in conjunction with a single master scene (Fig. 2). Only coherent pixels with steady phase or amplitude are handled in this approach. In densely vegetated areas, the quantity of resulting points may be relatively small. The roadway surface in this investigation was sufficiently extensive to be captured by numerous S-1 pixels, thus ensuring an ample number of coherent PSI pixels.

In this study, interferometry was employed using open Synthetic Aperture Radar (SAR) data obtained from the Copernicus Open Access Hub, which is managed by the European Space Agency (ESA), utilizing an entirely open source SNAP-StaMPS workflow method. To handle complete stacks of images taken from the SLC Sentinel-1 in batch mode, we used the snap2stamps package, which offers a set of scripts. A MATLAB-based program called Stanford method for PS (StaMPS) was created at Stanford University and then upgraded at the Delft University of Technology and the Universities of Iceland, and the University of Leeds. Additionally, persistent scatterer interferometry is included in the package. It is designed to improve PSI processing performance in dry, non-urban environments with nonlinear deformation and no man-made buildings. It calculates the likelihood that a pixel will be a persistent scatterer (PS) using the amplitude and phase data. Calculations for the interferograms are done automatically. This application are wrapper for Python that utilize SNAP as an InSAR technology and produce PSI analysis results that are compliant with StaMPS. On GitHub, such script sets are accessible as open source [30]. In this section, a brief description of the main method (Fig. 3) and the details of the preparation process for the individual datasets utilized in this study are provided.

Fig. 3
figure 3

Simplified workflow of PSI processing between SNAP and STAMPS

Full size image

There are two different workflows for the PSI processing: (i) The processing of PSI using StaMPS, and (ii) the single master DInSAR processing employing ESA SNAP. Initially, due to the presence of orbit control in Sentinel-1, which guarantees the possibility of combining the data using interferometry, the selection of the master image is created using the initial data series. In addition, we expect that the amount of PS data collected will not be significantly affected by the temporal baseline, as our objective is to acquire PS points related to urban infrastructure. In order to guarantee precise identification of bursts that encompass our designated area of interest (AOI), we carry out two consecutive processes: master image splitting and orbit state vector updates. These processes are executed using the SNAP graphical user interface (GUI). These procedures are crucial because they reduce the amount of time and resources required for the remaining processing. The parameters for dividing the master image to select bursts within the area of interest (AOI) are provided in Table 3.

Table 3 Main attributes of the selected Sentinel-1A master scene
Full size table

The next step is to use the snap2stamps scripts to generate all single master interferograms, using an automatic processing scheme that is divided into four steps: Slave preparation, Slave splitting, Coregistration, interferogram calculation and the export of stamps. Then, use the script mt_prep_snap, which is included in the distribution, to import SNAP exports into StaMPS. Then, as outlined in the StaMPS User Manual [31], the StaMPS persistent scatterer interferometry (PSI) processing chain is executed sequentially, starting from step 1 and progressing through step 7. We utilized the integrated TRAIN method in this instance, incorporating a linear tropospheric correction technique to alleviate the atmospheric phase linked to topography. The assessment of the accuracy of the atmospheric phase screen (APS) estimation was conducted by analyzing the point sources’ (PSs) temporal coherence following graph inversion and APS removal. The analysis produced satisfactory results, as the majority of the chosen PSs exhibited a coherence value exceeding 0.9 (see Fig. 4).

Fig. 4
figure 4

Temporal coherence of selected PSs of time series

Full size image

Post-processing

We use Eqs. (2) and (3) for the purpose of determining the vertical component for each and every individual persistent scatter (PS) point by reference to PSI measurements, as explained in [32].

$$\left( {d_{{{\text{LOS}}}}^{{{\text{desc}}}} } \right) = A\left( {\begin{array}{*{20}c} {d_{{{\text{up}}}} } \\ {d_{{{\text{hold}}}} } \\ \end{array} } \right)$$
(2)
$$A = \left( {\cos \theta^{{{\text{desc}}}} \quad \sin \theta^{{{\text{desc}}}} } \right)$$
(3)

\(d_{{{\text{LOS}}}}\) is the displacement along the line-of-sight (LOS). The vertical displacement is denoted by dup, \(d_{{{\text{hold}}}}\) is the horizontal displacement projection in the descending azimuth look direction (ALD) and θ is the incident angle.

The application of the TRAIN linear phase-based correction occurs at the conclusion of the StaMPS processing. For further details regarding these procedures, please refer to the StaMPS/MTI Manual [33]. At this point, the user has the option to select a specific reference area or site and input an average velocity, typically obtained from CGNSS data. This reference velocity will be used to refer to the entire persistent scatterer dataset. StaMPS, conversely, defaults to utilizing the entire extent as a reference area. The average of all the persistent scatterer interferometric unwrapped phases is calculated and adjusted to a value of zero. The average stability behavior throughout the entire scene is presumed in this manner.

When data needs to be visualized in a geographic information system (GIS), the coordinates of each persistent scatterer and an attribute table that contains the time series of slant-oriented displacements for relevant epochs can be exported in a vector shapefiles format. Additionally, the average velocity from a least squares regression analysis can also be exported in this format. For the purpose of determining the vertical and east–west displacement, we incorporated a vector decomposition algorithm that was developed in MATLAB. During the decomposition process, the variation in incidence angle across the entire study area was taken into consideration, and the appropriate value was assigned to each resolution cell.

Result

The final maps and histogram of displacement is presented in Figs. 5 and 6; the PSI Sentinel-1 data is used to depict the general deformation overview in LOS, vertical and horizontal direction.

Fig. 5
figure 5

a Average velocity map PSI Sentinel-1 for LOS, b Average vertical velocity, c histogram of displacement in time of series

Full size image
Fig. 6
figure 6

a Average velocity map PSI Sentinel-1 for horizontal displacement, b histogram of displacement in time of series

Full size image

In this study, LOS and vertical measurements of PSI deformation velocity values are made along satellites. From a descending track, this study estimated the true vertical motion components. According to InSAR data, the study area’s subsidence bowls are concentrated around the main thoroughfares and cities (Fig. 5). Between 2020 and 2022, the rate of land subsidence in Mansoura and Damietta ranges from − 19 to 20.7 mm/year.

For the datasets, histograms and fundamental statistics have been computed (Fig. 5c). As can be seen in Fig. 5c, the time series graphs show that the entire data set is split between negative values and positive values at 30 and − 30 mm/year, respectively.

In this study, horizontal measurements of PSI deformation velocity values are made along satellites. From a descending track, this study estimated the true horizontal motion components. According to InSAR data, the study area’s subsidence bowls are concentrated around the main thoroughfares and cities (Fig. 6a). Between 2020 and 2022, the rate of land subsidence ranges from − 26 to 28.4 mm/year.

For the datasets, histograms and fundamental statistics have been computed (Fig. 6b), As can be seen in Fig. 6b, the time series graphs show that the entire data set is split between negative values and positive values at 30 and − 30 mm/year, respectively.

It is evident that there are a variety of deformation patterns, which are primarily caused by the various subsidence mechanisms operating in the study area, especially roads that show many problems in nature such as cracks, potholes and rutting. There are several areas undergoing significant subsidence as shown in Fig. 7, such as (i) the road connecting Damietta and Ras Elbr; (ii) internal roads of cities along the Nile River; (iii) the road between Kafr Saad and Damietta and; (iv) coastal areas along the study area, including roads, while the rest of the region exhibits relatively low ground deformation rates. In the following sections, we provide a more detailed analysis of the PSI results.

Fig. 7
figure 7

Black insets highlight the four areas chosen for further investigation

Full size image

Area (a): the road between Kafr Saad and Damietta at km 9800–10,000

Samples were taken from PS points where high rates of subsidence were observed in Sentinel images, as well as at coordinates that were monitored in nature using a smartphone. A deformation pattern was identified from the data obtained from Sentinel-1, and the road between Kafr Saad and Damietta was analyzed. Subsidence is identified by downward geometry with respect to PSI (see Fig. 8). PSI (descending passage) measured a maximum vertical subsidence of roughly − 30 mm, whereas the average vertical subsidence of every PS was approximately − 15 mm during the observed time frame (2020–2023).

Fig. 8
figure 8

Combined results of the techniques applied in area (a) to identify land uplift and subsidence. The PSI results (time series), smartphone images taken on the ground and a graph showing the deformation of particular PS points from the time series—highlighted on maps with a white background—come from left to right

Full size image

Area (b): the road between Kafr Saad and Damietta, 10,400–10,600 km

The time series deformation for a PS point in the region is shown in Fig. 9. The temporal baseline in days (dated from January 2020 to December 2023) is represented by the X-axis in the figure, and the deformation value in millimeters is represented by the Y-axis. The deformation appears to have a linear motion trend at first glance, with a vertical displacement of roughly − 15 mm/year. The graph’s highest slope was recorded between October 21, 2020, and November 02, 2020. A high slope denotes a high rate of land movement acceleration. It also appears in the images taken from smartphone that there are cracks and rutting at the same coordinates.

Fig. 9
figure 9

Combined results of the techniques applied to identify land uplift and subsidence. The PSI results (time series), smartphone images taken on the ground and a graph showing the deformation of PS point from the time series—highlighted on maps with a white background—come from left to right

Full size image

Previous study examined the assessment of land subsidence in the Nile Delta using GNSS measurements. Using differential GPS, eight GNSS stations from 2013 to 2015 were distributed throughout the Nile Delta and the Northern part of Egypt. Ten International GNSS Service IGS stations to monitor the spatiotemporal variations of land surface within the Nile Delta through time series analysis of the ellipsoidal height. Based on GNSS measurements carried out in 2015, the Nile delta’s land subsidence is approximately 10 mm/year in the north direction, containing the road between Kafr Saad and Damietta [34].

Area (c): the road between Damietta and Ras Elbr

The time series-deformation for a PS point in the region is shown in Fig. 10. The temporal baseline in days (dated from January 2020 to December 2023) is represented by the X-axis in the figure, and the deformation value in millimeters is represented by the Y-axis. The deformation appears to have a linear motion trend at first glance, with a vertical displacement of roughly − 5 to − 10 mm/year. The graph’s highest slope was recorded between December 12, 2021, and March 21, 2022. A high slope denotes a high rate of land movement acceleration. It also appears in the images taken from smartphone that there are cracks at the same coordinates.

Fig. 10
figure 10

Combined results of the techniques applied to identify land uplift and subsidence. The PSI results (time series), smartphone images taken on the ground and a graph showing the deformation of PS point from the time series—highlighted on maps with a white background—come from left to right

Full size image

During the years 2013 through 2017, three main geodetic techniques have been used to detect the rates of the horizontal and vertical movements: satellite altimetry time series, time-varying gravity models of the Gravity Recovery and Climate Experiment (GRACE) mission to monitor the temporal changes over the delta regions, and Global Positioning System (GPS) measurements of the Delta region. Their GPS-based results demonstrate that the northern portion of the Nile Delta region has obvious subsidence, particularly on the eastern and western sides, as indicated by the stations at Edfena, Port Said, Mansoura, Gamalyia, Alexandria, and Damietta [22].

Area (d): internal roads of cities along the Nile River (Sherbin city)

PS points cover a large area of the road. In this area, a drone flight was conducted due to frequent landings in this area, and this area is included among the villages most in need. In the figure, the ground motion of PS points is represented in RGB color scale, while red color indicates high ground motion. The result shows that PSs have a cumulative displacement of up to 40 mm in three years, with a movement speed of − 20 mm/year. The unmanned aerial vehicle (UAV) image shows cracks along most of the road. Figure 11 shows the PS distribution of the region.

Fig. 11
figure 11

Combined results of the techniques applied to identify land uplift and subsidence. The PSI results (time series), UAV images taken on the ground and a graph showing the deformation of PS point from the time series—highlighted on maps with a white background—come from left to right

Full size image

For the purposes of maintenance, the most important task is to continue to monitor the deformation of the road pavement. Potholes and cracks, which are the most common types of road deformities, are particularly problematic. In order to improve the efficiency of pavement inspections, new kinds of remote sensing data that do not cause damage to the pavement are beginning to be utilized more frequently in order to identify issues with the pavement. An approach that was proposed to extract road defects from images captured by unmanned aerial vehicles (UAVs) using machine learning, with a particular emphasis on data preprocessing, was presented. In the same region where we found evidence of subsidence using psi, we also found cracks, potholes, and other subsidence-related features [35].

During that period, a ground truth measurement campaign was organized in situ to verify the outcomes of PSI-SAR processing in area (d). Using a UAV DJI Phantom 4 RTK, ground truth data were gathered for a few chosen high coherence level PS points during the measurement campaign.

Table 4 presents a tabulation of the recorded results, comparing the elevation level, latitude (LAT), and longitude (LON) obtained from both UAV and PSI-SAR analysis. It demonstrates that the difference in elevation level between the techniques is less than 0.8% with average height error 0.41%.

Table 4 Validation of measurement results using UAV 3D photogrammetry
Full size table

Discussion

This study assessed the effectiveness of using combined InSAR PSI analysis with Sentinel-1 (C band SAR) data to identify ongoing subsidence in the Nile Delta. The focus was on detecting subsidence, cracks, and rutting in the main roads, which have a significant impact on traffic flow. This approach yielded reliable outcomes when compared to on-site measurements conducted by unmanned aerial vehicles (UAV) and previous investigations utilizing global navigation satellite systems (GNSS) and global positioning systems (GPS). Moreover, it showcased its applicability in environments characterized by abundant vegetation. This approach successfully addressed the coherence limitations (citace) that were identified in previous research on the conventional DINSAR method. In previous studies, the combination of various synthetic aperture radar (SAR) analysis techniques, such as DSM change detection or displacement tracking, has been utilized to enhance the overall detection capabilities of subsidence. This study also employs this approach. We conducted an assessment of the current rates of displacement in specific areas of the northern Nile Delta, particularly in roads. This assessment was done using advanced persistent scatterer radar interferometry techniques applied to 25 sentinel1A scenes. The LOS displacement and the derived vertical displacement are utilized to ascertain road subsidence in the Nile Delta. While this study observed both subsidence and elevation features, subsidence features were more prevalent. Roads and buildings experience the most significant and relatively fast displacement. All four areas depicted in Figs. 7, 8, 9 and 10 exhibited low levels. The PSI showcased its usability in this particular case study.

From the analysis of the subsidence values mentioned in this study, it was proven that the Nile Delta suffers from a severe danger due to continuous subsidence. This subsidence affects buildings and roads greatly. It is possible that this subsidence is due to the poor quality of limestone in Egypt, which in turn leads to ground subsidence with large cracks as well as high temperatures, which “crunch” the asphalt. Other causes cause ground subsidence, including increased loads that exceed the strength and bearing of the road surface, and also the sedimentary soil of the delta consisting of a mixture of silt, fine sand, and silt. It is soil that contains voids filled with air or water and are therefore easily susceptible to compression under natural and human influences.

The degree of severity of vertical displacement greatly affects road safety, causing serious damage to infrastructure, buildings and natural resources. It can also lead to the formation of sinkholes, floods and seawater leakage, which can have disastrous effects. Subsidence poses a major threat to various types of infrastructure, including roads, bridges, pipelines and railways. As the ground subsides, these structures may deform, crack, or even fail completely.

Many effects of land subsidence were observed in the study area during the field visit. For example, cracks, grooves, and potholes, as shown in Figs. 7, 8 and 9, and also during the trip that took place in an area directly next to the Nile branch using drones, as this area is of interest to the government within the project to develop poor villages, as shown in Fig. 10.

In the case of the Kafr Saad-Damietta road, the average vertical displacement achieved by the PSI approach reached − 20 mm/year in vertical direction and − 26 mm/year in horizontal direction.

We also witnessed significant deterioration in the villages next to the Nile, where the average rate of displacement reached − 40 mm.

Conclusion

The study demonstrated that PSI technology can serve as a viable alternative to in situ techniques for analyzing subsidence along roads in the Nile Delta. The analysis of subsidence along roads exhibited a spatial distribution that was consistent with ground measurements. However, it also indicated a persistent escalation in subsidence rates, which is a cause for concern given the previous study. The analysis of a three-year continuous PSI time series using Sentinel-1 data allowed us to obtain information about temporal ground deformation, including velocity and displacement, through PSI-SAR processing. By conducting an analysis, various areas on the roads that experienced land subsidence were effectively identified. During the study period, the highway connecting Kafr Saad and Damietta experienced a maximum subsidence of 30 mm, while the internal road to the city of Sherbin next to the Nile had a maximum subsidence of 40 mm. To enhance map visualization accuracy, the geocoded land deformation results obtained from PSI-SAR analysis were verified using data gathered from in situ ground truth measurement using the 3D photogrammetry technique. Therefore, it can be inferred that the PSI-SAR technique is suitable for mapping and monitoring land deformation, while UAV-based 3D photogrammetry can serve as an alternative validation tool to assess the accuracy of ground survey.

The effective application of InSAR techniques in a highly vegetated region has demonstrated its capacity for further utilization in monitoring the development of linear structures. An analysis based on Sentinel-1 offers an exceptional and unmatched source of information due to its abundant data availability, reconsideration of time span, and optimal balance between spatial resolution and coherence features in vegetated areas, like the ones being studied. The presence of these attributes could potentially enhance the acceptance of DInSAR methods as a standard approach of validating in situ highway monitoring over extended time periods and larger geographical areas in the future.

Recommendation

We recommended monitoring the deformations of the Nile Delta using PSI and SBAS techniques, and making a comparison between the two techniques as the coverage of the SBAS technique is greater than the PSI technique to evaluate whether both are suitable for monitoring subsidence for a long period of time, and which of the two techniques is better.