Monitoring the Subsidence in Wan’an Town of Deyang Based on PS-InSAR Technology (Sichuan, China)

Abstract

In recent years, land subsidence has become a crucial factor affecting urban safety and sustainable development, especially in Wan’an Town. To accurately monitor and analyze the land subsidence in Wan’an Town, this study uses the PS-InSAR technique combined with an improved DEM for detailed research on land subsidence in Wan’an Town. PS-InSAR, or Permanent Scatterer Interferometric SAR, is suitable for high-precision monitoring of surface deformation. The natural neighbor interpolation method optimizes DEM data, improving its spatial resolution and accuracy. In this study, multiple periods of SAR imagery data of Wan’an Town were collected and preprocessed through radiometric calibration, phase unwrapping, and other steps. Using the PS-InSAR technique, the phase information of permanent scatterers (PS points) on the surface was extracted to establish a deformation model and preliminarily analyze the land subsidence in Wan’an Town. Concurrently, the DEM data were optimized using the natural neighbor interpolation method to enhance its accuracy. Finally, the optimized DEM data were combined with the surface deformation information extracted through the PS-InSAR technique for a detailed analysis of the land subsidence in Wan’an Town. The research results indicate that the DEM data optimized by the natural neighbor interpolation method have higher accuracy and spatial resolution, providing a more accurate reflection of the topographical features of Wan’an Town. The research found that the optimized DEM provided a more accurate reflection of Wan’an Town’s topographical features. By combining PS-InSAR data, subsidence information from 2016 to 2024 was calculated. The study area showed varying degrees of subsidence, with rates ranging from 6 mm/year to 10 mm/year. Four characteristic deformation areas were analyzed for causes and influencing factors. The findings contribute to understanding urban land subsidence, guiding urban planning, and providing data support for geological disaster warning and prevention.

1. Introduction

In recent years, geological disasters induced by surface deformation have garnered widespread attention in many places. These geological disasters are often caused by surface deformations resulting from earthquakes, volcanic eruptions, tectonic movements, and other factors. Land subsidence has become a significant environmental issue faced by multiple regions worldwide. Due to its multifaceted causes and influencing factors, numerous experts and scholars have conducted detailed analyses of land subsidence in various areas using different technical methods. According to previous research findings, the main causes of land subsidence include excessive groundwater extraction, large-scale urban construction, extensive mineral resource exploitation, and natural factors [1,2,3,4,5,6,7].

The Sichuan region, situated at the convergence of the Qinghai-Tibet Plateau, the Qinling Mountains, and the Sichuan Basin, is one of the most geologically active areas in China. Due to experiencing multiple and complex tectonic movements over time, surface deformation in Sichuan exhibits distinctive characteristics, including seismic activity, changes in topography, alterations in the lithosphere, fluctuations in groundwater levels, and risks of geological hazards [8,9,10,11]. Due to the presence of several significant fault zones such as the Yangtze River Fault and the Longmen Mountain Fault, which are the primary sources of frequent seismic activity, seismic-induced surface deformation is particularly pronounced in the Sichuan region. The terrain and landforms in Sichuan are diverse, including high mountains, canyons, and basins. Long-term tectonic movements have led to surface deformation such as uplift, subsidence, and uplift of the crust, thereby influencing the evolution of terrain and landforms. The frequent tectonic activity in the Sichuan region results in changes in the lithospheric mantle, including uplift, subsidence, and folding deformation of underground rock layers. Additionally, fluctuations in groundwater levels contribute to surface deformation. Sichuan boasts abundant groundwater resources, and groundwater extraction and recharge are influenced by crustal structures and surface topography, resulting in phenomena such as land subsidence and ground collapse.

Wan’an Town is located in Luojiang District, Deyang City, Sichuan Province. Situated on the northeastern edge of the Chengdu Plain, it borders Mianyang City to the east, the central urban area of Deyang City to the south, Guanghan City to the west, and Jianyang City to the north. With its advantageous geographical location and convenient transportation, Wan’an Town serves not only as an ecological barrier in the northeastern part of the Chengdu Plain but also as a crucial node in the Chengdu-Deyang-Mianyang economic belt. The total area of this region is 33.29 square kilometers, with a population of 52,771 people. Due to its prominent tectonic activity and abundant mineral resources, it possesses certain geological characteristics and risks of geological disasters, thus requiring a balance between resource development and ecological environment protection. Currently, the surface deformation monitoring work in this area primarily relies on traditional techniques, including leveling measurements, cadastral surveys, inclinometer and strain gauge measurements, groundwater monitoring, and depth measurements. Although GPS monitoring technology is gradually being widely used in surface deformation monitoring, there are still some problems. For example, the accuracy of leveling measurements is limited by human factors. Cadastral surveys are not only cumbersome but also inefficient. Inclinometer and strain gauge measurements usually only provide local relative deformation information. GPS monitoring may suffer from positioning errors, thus affecting measurement accuracy. Based on the above analysis, this paper proposes the use of PS-InSAR (Permanent Scatterer Interferometric Synthetic Aperture Radar) technology as a technical means to improve monitoring accuracy and make accurate predictions based on the results.

InSAR (Interferometric Synthetic Aperture Radar) technology is an effective method for monitoring surface deformation by utilizing radar signals. It measures slight changes on the Earth’s surface by comparing radar images captured at different times. This technology has wide-ranging applications and can provide information on phenomena such as ground subsidence, uplift, and the formation of surface cracks. PS-InSAR is a more advanced technique developed based on InSAR. Since its introduction in the late 20th century, Permanent Scatterer Interferometric Synthetic Aperture Radar (PS-InSAR) technology has demonstrated tremendous application potential across multiple fields, particularly in surface deformation monitoring, geological disaster early warning, and urban infrastructure deformation detection. This technology has now been widely adopted globally, especially in geologically active areas such as the southwestern mountainous regions of China and volcanic regions in Europe. Over time, the scope of PS-InSAR’s applications has continually expanded, extending from traditional geological disaster monitoring to include infrastructure monitoring, land subsidence detection, mining activity surveillance, and oil field development. In the context of urban development, ground subsidence monitoring and transportation network stability assessments have also become key application areas for PS-InSAR technology. In recent years, PS-InSAR has increasingly integrated with big data and artificial intelligence. By leveraging the deep analysis of vast amounts of radar data, machine learning algorithms can accurately identify potential geological disaster risks and predict trends and anomalies. This integration not only enhances the intelligence of deformation monitoring but also provides more precise and timely information for the prevention and response to geological hazards. However, despite the significant progress made, PS-InSAR technology still faces limitations, such as environmental conditions, data processing capabilities, and atmospheric interference. To overcome these challenges, the future development of PS-InSAR will require further technological innovations and international collaboration. Breakthroughs in areas such as atmospheric correction, data standardization, and efficient data processing are especially needed to further improve the accuracy and efficiency of PS-InSAR monitoring. Its main advantage lies in the utilization of permanent scatterers on the Earth’s surface for monitoring. Permanent scatterers are stable reflectors on the surface, such as buildings, rocks, and bridges, whose phase remains stable over multiple observations. PS-InSAR technology achieves higher precision and stability in surface deformation monitoring by continuously performing interferometric measurements of these permanent scatterers’ phases [12,13,14,15,16,17]. Since the geological disaster work in this area is currently not comprehensive, this paper selects this area as the research area. PS-InSAR technology is utilized to monitor and predict the surface deformation in this area, study its development mechanism, and lay the technical foundation for subsequent work.

2. Materials and Methods

2.1. Study Area

Wan’an Town is located in the foreland basin sedimentary position of the Longquan Mountain structural zone, and thus has undergone multiple tectonic movements. Approximately 800 million years ago, a powerful regional mountain-building event occurred, having widespread effects on the North China Plate and its surrounding areas. The crustal movements during the Triassic and Early Jurassic periods were of significant importance to the tectonic development of China and neighboring regions, potentially leading to further landmass accretion in East Asia and subsequent regional geological structural changes. The Yanshan Movement, a tectonic event that was widespread throughout China during the Jurassic and Cretaceous periods, had a significant impact on shaping the geological structure of the study area, characterized by frequent folding, faulting, and volcanic activity. The mountain-building movements of the Cenozoic Era, primarily marked by the collision and continued compression between the Indian and Eurasian plates, though predominantly affecting the Qinghai-Tibet Plateau and its periphery, also had important implications for this region as a globally significant tectonic event. Thus, the geological structural characteristics of Wan’an Town indicate a complex geological evolution, including crustal uplift and subsidence, and mountain formation. These tectonic movements collectively shaped the present-day geological structure of the Luojiang area. The tectonic characteristics of the region primarily manifest as extensional tectonic activities since the Late Permian, resulting in the formation of specific stratigraphic structures. The lithological features of the strata are primarily characterized by high-energy facies sedimentation such as bioherms, bioclastic shoals, as well as hydrocarbon source rocks such as microcrystalline limestone and siliceous mudstone deposited in deepwater environments. These features hold significant importance for oil and gas exploration in the region.

The terrain and landform features of Wan’an Town are mainly characterized by undulating hills, river valleys, gullies, and basin-edge landscapes. Influenced by the surrounding mountains, the area exhibits diverse elevations and forms of hilly terrain (Figure 1); thus, the terrain of Wan’an Town can be divided into three main parts: mountainous terrain, hilly terrain, and basin-edge terrain.

For mountainous terrain, the diverse forms of peaks and ridges result from geological tectonic movements and long-term weathering erosion. In mountainous landscapes, river erosion creates deep gorges and ravines, particularly prominent in areas of significant topographic relief. Moreover, mountainous terrain often exhibits rich vegetation cover, including forests, shrubs, and grasslands. This vegetation forms complex ecosystems crucial for soil stability and water source protection. Some relatively flat mountain plains exist between the mountains, formed by river alluvium or landslide sedimentation, serving as important areas for agriculture and human activities.

Regarding hilly terrain, compared to mountains, the contours of hillsides are generally gentler. Hilly areas typically possess fertile soil suitable for agriculture, thus often utilized for cultivation and crop planting. Additionally, hilly regions play an important role in maintaining soil stability and ecological balance. In transition zones between mountainous and hilly terrain, river channels and valleys formed by river erosion are common, with valleys often being deep and narrow. Basin terrain is typically characterized by low-lying flat areas surrounded by mountains or hills, relatively enclosed and sheltered by surrounding terrain. Due to the low-lying terrain, basins often accumulate water bodies such as lakes, swamps, or seasonal rivers. Sediment accumulation in basins, facilitated by low terrain, can create fertile farmland. Compared to other landforms, basin terrain is relatively flat and fertile, making it an important area for agricultural production (Figure 2).

Wan’an Town, located in the Longquan Mountain structural zone, has river systems originating from surrounding mountains and hilly terrain. Due to its complex terrain and geological diversity, the river system of Wan’an Town mainly consists of the Kaijiang River and the Mianyuan River, along with their tributaries and networks (

Table 1. List of runoff characteristic values in the main rivers in the combined survey area.

River Name	Rainwater Collection Area (km2)	Channel Length (km)	Average Slope (‰)	Average Runoff Depth Over Many Years (mm)	Average Annual Runoff (×108 m3)
Kai River	1140	100.3	1.0	396	4.51
Mianyuan River	1212	117.5	6.50	682	8.26

). The river system undergoes seasonal fluctuations influenced by rainfall and snowmelt. During the rainy season or snowmelt period, river discharge may increase, leading to rising water levels and accelerated flow velocity in the river channels. Over time, prolonged river erosion has shaped significant river valley landscapes, providing vital support for local water resources and ecological environments. In areas with gentle terrain, small-scale river valleys formed by rainwater erosion, resulting in gullies, though relatively small in scale, play a certain role in surface water circulation and landform formation processes.

Based on preliminary geological exploration data, the stratigraphic ages in Wan’an Town mainly include the Paleozoic, Mesozoic, and Cenozoic eras. The Paleozoic strata consist of the Cambrian, Ordovician, and Silurian periods, characterized by shale, limestone, and sandstone. Mesozoic strata, commonly found in Wan’an Town, comprise the Jurassic and Cretaceous periods, with lithologies primarily consisting of sandstone, mudstone, and coal seams, occasionally interspersed with volcanic rocks or intrusive bodies. The Cenozoic strata, including the Tertiary and Quaternary periods, predominantly consist of conglomerates, peat, volcanic rocks, along with minor glacial deposits and river sediments (Figure 3).

Ground subsidence refers to the process in which the soil or foundation in a specific area on or below the Earth’s surface moves downward or settles, which is a geological phenomenon constrained by underlying rock layers [18,19,20,21,22]. This subsidence may be caused by natural factors such as geological tectonic activity, surface sedimentation processes, or changes in groundwater levels, as well as by anthropogenic factors such as excessive groundwater extraction, underground mining, or construction activities. Tectonic movements during the Late Jurassic to Early Cretaceous period profoundly shaped the geological structure of the study area. Subsequent Cenozoic tectonic movements also influenced the geological structure and geomorphological evolution of the region, particularly the 6.8-magnitude earthquake that occurred in Luding County, Ganzi Prefecture, Sichuan Province in 2022. As an abrupt event of Neotectonic movement, it can cause significant regional vertical deformation in the short term, thereby exacerbating ground subsidence. Due to the influence of multiple periods of tectonic movement, the strata have undergone multiple vertical uplifts and subsidence, leading to localized ground subsidence or uplift. In addition, local geological tectonic activities, such as fault activity and strata deformation, can also result in changes in the geological structural features and topography of the region.

Currently, the method for monitoring regional subsidence still relies on traditional single-point subsidence estimation. This approach typically provides data for a limited area and fails to offer a comprehensive understanding of subsidence across a larger region. Consequently, it results in an incomplete understanding of subsidence in extensive areas and can lead to inaccurate measurements in complex geological environments or engineering structures [23]. InSAR technology can acquire three-dimensional surface information and provide extensive, continuous deformation monitoring for regions [24,25]. With the continuous development of microwave remote sensing technology, PS-InSAR has been applied in many countries for monitoring surface subsidence [26,27,28,29] and has been widely used in the field of geological disaster monitoring. This study focuses on detailed subsidence detection using radar imagery data from Sentinel-1, which cover the area of Wan’an Town. The imagery was acquired through the Copernicus Open Access Hub, a component of the European Space Agency’s (ESA) Copernicus Open Data Program. Sentinel-1, an Earth observation satellite equipped with C-band Synthetic Aperture Radar (SAR) technology, can capture high-quality radar images of the Earth’s surface regardless of weather conditions or the time of day, making it an ideal source for this study.

However, some scholars have pointed out that in the process of using remote sensing technology, it is necessary to integrate external Digital Elevation Models (DEMs). DEMs contain only terrain information and have relatively low spatial resolution. Therefore, when using medium- to low-resolution SAR imagery data, accurate registration can be achieved in areas with significant terrain undulations and minimal or no coverage. Nevertheless, it becomes challenging to achieve precise registration in areas with less pronounced terrain undulations and higher surface cover or vegetation density. DEMs can be used to correct the terrain phase in interferograms, but systematic errors in DEMs may be misinterpreted as input signals, resulting in discrepancies between research results and actual conditions [30,31,32,33,34]. To address this issue, this study employs PS-InSAR technology to monitor and study ground subsidence in the study area, considering its unique terrain characteristics. PS-InSAR is a synthetic aperture radar interferometric measurement technique capable of monitoring small surface deformations. In this technology, stable scattering points (typically buildings, rocks, or other distinct surface features) are utilized to measure vertical surface deformations through multiple radar observations of the same area. Therefore, this study applies this technology to select effective control point data and preprocess SAR images using filtering and other techniques. This study employs three methods: high-precision deformation measurement, fusion of multi-source data, and optimization of models and algorithms, all aimed at enhancing the accuracy of multi-source DEMs. Additionally, it performs resampling and height correction to generate high-quality DEM models. Through these processes, not only the quality of interferograms generated by SAR technology is enhanced, but also effective monitoring of deformation areas is ensured while minimizing noise interference [35,36,37,38]. This study statistically analyzes the results of extracted undulations and investigates the relationship between undulations and surface deformations to understand the characteristics of surface deformations in different undulation areas.

2.2. Data and Preprocessing

PS-InSAR is an advancement derived from traditional InSAR. It is a technique that utilizes Synthetic Aperture Radar (SAR) data for monitoring surface deformations [39,40,41,42]. This technology combines SAR and InSAR techniques, utilizing the polarization information of radar waves, which refers to the wave’s polarization state, to enhance the accuracy of surface deformation monitoring. By analyzing radar images in different polarization directions, more information can be obtained, thereby improving the precision and reliability of deformation monitoring. Therefore, this technology has significant applications in areas such as geological hazard monitoring, urban subsidence monitoring, and glacier change detection. Sentinel-1 is equipped with a C-band SAR, featuring four dedicated imaging modes with different resolutions and coverage areas. Specific information can be found in

Table 2. Different working modes of Sentinel-1 data.

Imaging Mode	Width (km)	Incidence Range	Polarization
SM	80	18.3–46.8°	HH/VV, HH + HV/VV + VH
IW	250	29.1–46.0°	HH/VV, HH + HV/VV + VH
EW	410	18.9–47.0°	HH/VV, HH + HV/VV + VH
WV	250	21.6–25.1°	HH/VV

.

Due to the topographical features of Wan’an Town, which primarily consist of undulating hills, river valleys, gullies, and the edges of basins, along with high vegetation coverage, to reduce the errors of external digital models, this paper uses 1:100,000 DEM data obtained based on InSAR technology. Through iterative processing, the high-precision registration of DEM data with SAR image data is improved to extract the terrain undulation of the joint survey area, with a sampling grid of 5 m. Using an n × n pixel rectangle as the template operator, the maximum and minimum values within the grid are calculated for each pixel. Here, n represents the size of the square neighborhood (n × n pixels) surrounding a specific pixel. Through field sampling, 100 points were selected. After optimizing the DEM, the comparison between the difference before and after optimization can be visually observed (Figure 4). It can be observed that after interpolation, the accuracy has significantly improved. The difference between the maximum and minimum values is then used to determine the topographic undulation results and the area proportions for each n × n window (Equation (1)). Based on this structure, the elevation of the study area is refined into categories: <50 m, 50–100 m, 100–150 m, 150–200 m, 250–300 m, 300–350 m, 350–400 m, 400–450 m, 450–500 m, 500–550 m, 550–600 m, 600–650 m, 650–700 m, 700–750 m, 750–800 m, 800–850 m, 850–900 m, 900–950 m, 950–1000 m, and >1000 m.

$$\mathrm{R}\mathrm{F}={\mathrm{H}}_{\mathrm{m}\mathrm{a}\mathrm{x}}-{\mathrm{H}}_{\mathrm{m}\mathrm{i}\mathrm{n}}$$

(1)

where RF is the topographic undulation within the analysis window, H_max is the maximum elevation value within the analysis window, and H_min is the minimum elevation value within the analysis window.

Although selecting a small sampling grid usually increases the complexity of data processing and computational demands, in cases where there is a lack of abundant foundational data and the study area has complex topographical features, it is necessary to optimize processing strategies to indirectly improve processing efficiency while maintaining a certain level of accuracy. This study, under the premise of meeting high-precision deformation measurement and multi-source data fusion requirements, employs the natural neighbor interpolation method to interpolate denser surface deformation information from sparse PS point data (Equation (2)), achieving the goal of finely predicting subsidence.

The natural neighbor interpolation method is a spatial interpolation algorithm based on Voronoi diagrams. On one hand, this algorithm effectively preserves local characteristics of the data, which helps maintain the authenticity and accuracy of the terrain. On the other hand, by calculating the weighted average of neighboring points to determine the weight of the point to be interpolated, it more reasonably allocates weights, making the interpolation results smoother and more accurate. This method is not only suitable for interpolating two-dimensional spatial data but can also be extended to interpolate three-dimensional spatial data [43,44,45].

$$f\left(x\right)={\sum }_{i=1}^n{\displaystyle \frac{s_i}{s}}\left(x\right)f_i$$

(2)

where f(x) is the interpolation result at the point x. s_i is the area of the polygon of sample point i remaining after being cut by the Voronoi polygon of the interpolation point x, and s is the total area of all sample point polygons. f_i is the value at sample point i, which is the known data point value.

This study focuses on ground subsidence in Wan’an Town, Deyang, using Sentinel-1A data collected from June 2016 to March 2024. The data consist of 15 scenes with a C-band, a revisit period of 12 days, a resolution of 15 × 15 m, and an IW imaging mode. During processing, the image from 14 February 2022 was selected as the master image, and 45 differential interferograms were generated. These were combined with an improved DEM model to predict ground subsidence in the study area.

2.3. Research Method

SAR is an active remote sensing observation system. Different SAR imaging modes produce different polarization images (HH, HV, VV, and VH). Depending on the varying sensitivity of different terrains and objects (such as buildings, mountains, ravines, rivers, etc.) to each polarization, the appropriate polarization method must be selected. Among these, the echo intensity of HH is greater than HV, and VV polarization is superior to HH [46]. Therefore, this study primarily uses VV polarization, with other polarization methods used for supplementary validation. In this study, Sentinel-1 data were used, with descending track InSAR deformation fields serving as the simulation conditions for ground subsidence in the study area. PS-InSAR time-series analysis technology was employed to extract and process ground subsidence deformation data for Wan’an Town from 2016 to 2024. For N + 1 SAR images, after undergoing processes such as registration, radiometric calibration, PS detection, and interferometric processing, the amplitude deviation method was used to select and extract phase changes of target points with high coherence and stable reflection characteristics as PS points (as shown in the formula below). This resulted in N interferograms and differential interferograms, multiple high-coherence PS points, and the differential interferometric phase sets of each PS point in each differential interferogram. After correcting for DEM errors and removing atmospheric coherence, each determined PS point was obtained. Additionally, based on the actual situation of the work area, phase unwrapping of PS discrete points was performed to separate the topographic phase on these target points, thereby obtaining the time-series deformation rate [47,48]. By separating and unwrapping the topographic phase of the selected PS points, surface deformation information was extracted. This technology effectively avoids the problems of temporal decorrelation and spatial decorrelation, ensuring higher accuracy of ground deformation information.

For N + 1 SAR images, after processes such as registration, radiometric calibration, PS detection, and interferometric processing, the amplitude deviation method is used to select points with high coherence and stable reflection characteristics as PS points and to extract phase changes (Equation (3)).

This results in N interferograms and differential interferograms, along with multiple high-coherence PS points. For each PS point in each differential interferogram, a differential interferometric phase set is obtained. After correcting for DEM errors and removing atmospheric coherence, each PS point is determined. Additionally, based on the actual conditions of the work area, phase unwrapping is performed on the PS discrete points to separate the terrain phase from these target points, thereby obtaining the time-series deformation rate [49,50,51,52]. By separating and unwrapping the terrain phase of the selected PS points, surface deformation information is extracted. This technique effectively avoids temporal and spatial decorrelation issues, ensuring higher precision in land deformation information. The technical flowchart is shown in Figure 5.

$${\sigma }_{\phi }=D_A={\displaystyle \frac{{\sigma }_A}{{\mu }_A}}$$

(3)

where ${\sigma }_{\phi }$ and μ_A represent the standard deviation and mean of the amplitude, respectively, and σ_A and D_A represent the standard deviation of the phase and amplitude deviation, respectively.

The PS-InSAR technology process used in this study is mainly divided into two parts: data preprocessing and data processing. (1) By statistically analyzing the intensity values and coherence of the pixels, PS candidate points with low coherence coefficients are removed. A threshold is determined from the range of values to serve as a reference for selecting PS points. After obtaining PS candidate points, phase stability is used as the criterion to select the best PS points, ultimately determining the PS phase stable points. (2) While generating interferograms, to ensure the accuracy of deformation phases, the influence of topographic phases needs to be removed using both the processed external DEM and the generated DEM, thus avoiding topographic phase errors caused by DEM inaccuracies. (3) After eliminating the effects of atmospheric phase (APS), DEM errors, and other influences, all PS phase points need to be unwrapped and recalculated to obtain more precise deformation rates of PS points. This allows for the derivation of true deformation values after atmospheric correction. The results of the spatiotemporal baseline estimation are shown in Figure 6.

To ensure the accuracy of the PS-InSAR technique and effectively remove terrain residuals and atmospheric phase delays, a series of differential interferometric processing steps need to be performed. From N SAR images, generate N-1 interferometric phase maps through interferometric processing of adjacent images. Using corrected DEM (Digital Elevation Model) data, perform differential interferometric processing on the N-1 interferometric phase maps to make surface deformation more evident. Conduct radiometric calibration processing on the N SAR images and select candidate PS (permanent scatterer) points using the coherence coefficient threshold method and amplitude threshold method. Using the selected PS points and the N-1 differential interferometric phase maps, establish the differential interferometric phase time-series for the PS points. Based on the ground deformation situation, establish a model equation system for the relevant parameters and differential phases. On this basis, remove the linear deformation component and DEM error component of the PS points from the original differential interferometric phase, obtaining the residual phase. Extract more PS points based on the residual phase, and finally generate the differential interferogram (Figure 7).

To obtain settlement information for the study area from 2016 to 2024 through calculations, grid image processing was conducted based on PS point deformation rates to improve measurement accuracy. According to the spatial distribution of average surface deformation settlement rates in the study area (Figure 8), it is evident that the settlement trends may vary in different years within the same region. The yellow portions indicate areas with surface settlement trends, where colors closer to red signify higher annual average settlement rates, while shades closer to green indicate higher annual average uplift rates. Most parts of the study area exhibit varying degrees of settlement trends, but settlement rates generally range from 6 mm/year to 10 mm/year.

To ensure data accuracy and completeness, the GPS data obtained from field measurements were compared with the calculated subsidence results on an individual basis (Figure 9). The comparative analysis indicates that the elevation differences between the two are minimal, with error values evenly distributed within the statistical range. This not only demonstrates the reliability of the research methodology in terms of precision but also validates the scientific rigor of the data processing and analysis procedures. Furthermore, these comparison results reinforce the effectiveness of the research model, providing a reliable data foundation and technical support for subsequent subsidence monitoring and evaluation.

From 2016 to 2024, surface subsidence observations in the study area have revealed that with the acceleration of urbanization, extensive groundwater extraction and the construction of high-rise buildings significantly increase the risk of ground subsidence. Industrial and mineral resource extraction also lead to the formation of underground cavities, which in turn cause ground subsidence. Additionally, infrastructure development (such as subways, tunnels, bridges, and other large-scale infrastructure) impacts ground stability. According to the research results (Figure 10), subsidence areas are concentrated in regions with human activities.

3. Results

Due to the high vegetation coverage and complex terrain features in the study area, human activities critically impact land use [53,54]. Therefore, this study randomly selected several real PS points within different regions of the study area and conducted densification using natural neighbor interpolation (Figure 11) to analyze surface deformation based on the selected PS feature points.

To ensure data continuity and quality in processing the interferometric phase of PS points, low-pass and high-pass filtering techniques are employed to suppress atmospheric delay effects. Low-pass filtering is used to eliminate high-frequency noise while preserving the main signals of surface deformation. High-pass filtering, on the other hand, removes low-frequency signals such as atmospheric delays and slow-varying components. This filtering approach effectively mitigates the impact of atmospheric delay phases on the interferometric phase, enabling more accurate extraction of deformation information. Following filtering, this study analyzed the time-series cumulative deformation of PS points (Figure 12). The figure illustrates cumulative deformation from 2016 to 2024, highlighting significant long-term and large-scale subsidence in the central part of the study area, which is densely populated by humans from 2022 to 2024. Additionally, notable subsidence is observed in the northwest, southwest, and southeast parts of the study area.

Based on the actual survey data, Area A has experienced significant land subsidence due to prolonged geological changes and human activities.

This has been accompanied by large-scale deforestation, leading to a marked reduction in vegetation cover and further degradation of the ecological environment. Area B, as the primary mining zone, is subject to frequent mineral extraction activities, resulting in severe surface destruction and varying levels of impact on the surrounding ecosystem. Area C is undergoing extensive development and construction, involving multiple infrastructure projects, causing frequent surface disturbances that may lead to localized soil erosion and environmental stress. Area D primarily engages in landfilling operations, enhancing land utilization and reshaping the regional terrain, though it may pose potential impacts on soil structure and local water systems (Figure 13).

Based on the processing results using PS-InSAR technology and detailed geological data combined with optical imagery, this study successfully identified and extracted four significant deformation features, known as deformation target areas. Deformation Target Area A (Figure 14A) is located in the southern part of the study area, encompassing multiple regions with pronounced surface deformation. The deformation curves of surrounding points exhibit high consistency, with a maximum subsidence of up to 167 mm. Continuous monitoring of cumulative deformation shows a persistent downward trend, indicating ongoing surface deformation without relief.

Deformation Target Area B (Figure 14B) is located in the southeastern part of the study area. Compared to Deformation Target Area A, the subsidence deformation here is relatively smaller, reaching up to 168 mm, and it shows a gradually stabilizing trend. The subsidence in this target area is primarily induced by geological tectonic movements. These movements may involve crustal compression, stretching, or shearing, resulting in surface deformation. Human activities have also played a role in promoting subsidence, particularly through activities such as water resource development and infrastructure construction.

Deformation Target Area C (Figure 14C) is located in the central part of the study area and exhibits significant surface uplift, with a maximum uplift height reaching 203 mm. This area is characterized by intensive human activities. Changes in groundwater levels can alter the mechanical equilibrium of the soil layers. The region experiences extensive groundwater extraction activities, such as groundwater pumping and well drilling. Prolonged groundwater extraction may lead to a decline in groundwater levels, resulting in subsidence or localized surface uplift. Furthermore, in areas of intense human activity, such as those with dense traffic and construction activities, soil compaction can occur due to increased soil density and reduced volume. Long-term soil compaction may lead to surface subsidence, but in some cases, compacted soil may experience elastic rebound or localized uplift. Additionally, infrastructure development activities (such as roads, bridges, and buildings) in human activity zones can alter the morphology and stability of the land surface, leading to surface uplift. Major engineering projects like reservoirs and tunnels can have significant impacts on the geological environment during construction. Moreover, environmental factors such as climate change and changes in rainfall patterns can also influence surface uplift.

Deformation Target Area D (Figure 14D) is located in the northern part of the study area and exhibits surface uplift, indicating a noticeable increase in ground elevation compared to surrounding areas. The deformation magnitude in this target area is 170 mm, and it shows a sustained uplift trend. The period of accelerated uplift is similar to Deformation Target Area C, suggesting that both areas may be influenced by similar geological tectonic activities or driven by common external factors such as climate change or seismic activity. This uplift phenomenon not only affects the surface morphology but may also impact local geological structures, ecological environments, and human societies.

According to the above research results and actual geological surveys, the surface deformation in the study area is primarily caused by crustal movement, which leads to subsidence of the strata, resulting in structural settlement. Additionally, activities such as groundwater extraction and large-scale mineral extraction can cause significant surface subsidence in poorly consolidated or semi-consolidated soil layers due to soil consolidation and compaction. Furthermore, excessive land development, large-scale land rehabilitation projects, and the construction of hydraulic engineering projects can alter the original state of the ground, leading to changes in surface morphology, including potential uplift.

4. Discussion

In the analysis of surface subsidence characteristics in Wan’an Town, this paper discovered significant subsidence phenomena in specific areas. These areas are often associated with factors such as excessive groundwater extraction, loose geological structures, or urban construction. Traditional subsidence monitoring methods are often limited by the number and distribution of monitoring points, long data acquisition cycles, and limited accuracy. PS-InSAR technology, as an active microwave remote sensing technology, has the advantages of high precision, wide coverage, and continuous monitoring, enabling the capture of minute surface deformation information. At the same time, the application of the natural neighbor interpolation method further improves the accuracy of DEM, making subsidence analysis more accurate and reliable. Given the lack of research on surface subsidence in Wan’an Town, Deyang, and the singularity of existing monitoring and prediction methods, this study adopts PS-InSAR technology combined with the natural neighbor interpolation method to enhance DEM accuracy, achieving the goal of identifying and predicting surface subsidence in the study area.

In the study of surface subsidence in Wan’an Town, Deyang, the current data and research are relatively limited. This study extracted high-precision surface displacement information using PS-InSAR technology and generated a high-precision DEM using the natural neighbor interpolation method. Combined with these data, a detailed analysis of surface subsidence in Wan’an Town was conducted. The research results indicate that this method can effectively capture the surface subsidence characteristics of Wan’an Town, especially in areas with intensive groundwater extraction and construction activities, providing a more accurate subsidence distribution map. Moreover, PS-InSAR technology offers high spatial resolution and high precision, enabling the monitoring of minute surface deformations over large areas. By combining it with the natural neighbor interpolation method, it is possible to extend the discrete PS point information to the entire study area, thereby effectively enhancing the overall accuracy of the DEM. Compared to traditional interpolation methods, the natural neighbor interpolation method can better preserve the spatial characteristics of PS points, improving the accuracy and reliability of the interpolation results. This is crucial for the accuracy and comprehensiveness of surface subsidence monitoring.

Although the method used in this study has demonstrated significant advantages in improving DEM accuracy and monitoring surface subsidence, there are still some limitations:

(1)

Impact of PS Point Distribution: The density and quality of PS point distribution directly affect the interpolation results. In areas where PS points are sparse or unevenly distributed, the interpolation results may contain errors, affecting the overall accuracy.

(2)

Lack of Deep Geological Information: PS-InSAR technology primarily monitors surface displacement and cannot directly obtain information about changes in deep geological structures. Therefore, its effectiveness in monitoring surface subsidence caused by deep geological structural changes may be limited.

To further enhance the accuracy of surface subsidence monitoring and prediction in Wan’an Town, Deyang, future research will focus on the following directions:

(1)

Advanced Interpolation Algorithms: Explore and apply more advanced interpolation algorithms, such as those based on machine learning and deep learning methods, to further improve interpolation accuracy.

(2)

Long-term Monitoring and Trend Prediction: Establish a long-term surface subsidence monitoring system to accumulate long-term time-series data. Utilize big data analysis techniques and predictive models to analyze subsidence trends and predict future changes, providing scientific basis for regional planning and disaster prevention and mitigation.

(3)

Multi-scale Analysis: Combine monitoring data from regional and local scales to conduct multi-scale analysis, revealing subsidence characteristics and mechanisms at different scales.

5. Conclusions

This paper conducts a study on ground subsidence in Wan’an Town, Deyang, using PS-InSAR time-series technology to obtain deformation information of the study area from June 2016 to March 2024. On this basis, the processing efficiency is indirectly improved through optimized processing strategies, and the natural neighbor interpolation method is used to interpolate denser surface deformation information from the sparse PS point data. This achieves effective monitoring and accurate prediction of surface subsidence in the study area.

With the acceleration of urbanization, extensive groundwater extraction and the construction of high-rise buildings increase the risk of ground subsidence. Industrial and mineral resource extraction also lead to the formation of underground cavities, triggering ground subsidence. Research results indicate that from 2016 to 2024, the average subsidence rate in the study area ranged from 6 mm/a to 10 mm/a, with certain areas showing a trend of continuous subsidence.

This paper analyzes four characteristic deformation target areas. Deformation Target Area A is a region with intense surface deformation activity, with a maximum subsidence of 167 mm and a continuing downward trend. To further understand the deformation mechanisms and risk situation in this area, more in-depth geological exploration and monitoring are required. The geological dynamics of Deformation Target Area B result from the combined effects of tectonic movements and human activities. In Deformation Target Area B and its surrounding areas, there may be extensive groundwater resource development activities, and the unreasonable exploitation of groundwater resources by human activities has accelerated the surface subsidence process. Deformation Target Areas C and D exhibit significant surface uplift phenomena, with a high degree of consistency in the uplift periods.

Innovations in this study using PS-InSAR technology to investigate ground subsidence in Wan’an Town, Deyang, mainly include the following aspects:

• Acquisition and Analysis of Long-Term Time-Series Data: Long-term surface deformation data from June 2016 to March 2024 were obtained using PS-InSAR time-series technology. This ample data support enhances the study’s temporal dimension, facilitating in-depth analysis of dynamic changes in ground subsidence.
• Application of Natural Neighbor Interpolation: The use of natural neighbor interpolation effectively interpolates denser surface deformation information from sparse PS point data. This method significantly enhances the spatial resolution of data, enabling finer and more comprehensive monitoring of ground subsidence. It provides a new technical means for high-precision monitoring of ground subsidence.
• Detailed Analysis of Characteristic Deformation Target Areas: Four characteristic deformation target areas were selected for detailed analysis, revealing the subsidence characteristics and potential influencing factors of these regions with high-precision data. This not only validates the effectiveness of the methods but also provides specific case studies and scientific basis for studying ground subsidence mechanisms in different regions.
• Integrated Analysis of Multiple Factors Contributing to Subsidence: When analyzing ground subsidence in characteristic target areas, a comprehensive analysis was conducted by integrating factors such as groundwater extraction, geological structures, and urban construction. This approach helps in comprehensively understanding the causes and complexities of ground subsidence, providing a more complete reference for scientific decision making.
• Integration and Innovation of Methods and Technologies: Integration of PS-InSAR technology, natural neighbor interpolation, and optimized processing strategies achieved efficient and high-precision monitoring and prediction of ground subsidence. This integrated approach and technological innovation provide valuable experiences and methods for similar studies on ground subsidence in other regions.