Deformation monitoring

1、 Overview of InSAR Technology

Interferometric Synthetic Aperture Radar (InSAR) technology is a spatial observation technique based on multiple interferometric images of the same area using Synthetic Aperture Radar (SAR), combined with signal processing and geodetic principles, to achieve high-precision inversion of surface three-dimensional deformation fields and terrain information. Its core advantages lie in large-scale coverage, high spatial resolution, and all-weather observation, breaking through the limitations of traditional ground monitoring methods (such as GPS and leveling) in terms of coverage and efficiency. The deformation monitoring accuracy can reach millimeter to centimeter level.

InSAR technology has gone through iterations from differential interferometry (D-InSAR) to multi temporal interferometry (MT InSAR, including PS InSAR, SBAS InSAR, etc.), gradually solving problems such as decoherence, atmospheric errors, and terrain residuals. It has been widely used in urban, mining, earthquake, volcano, glacier, permafrost, landslide, and infrastructure fields, providing key data support for geological disaster prevention, ecological protection, and engineering safety assessment.

2、 Core application areas and monitoring points

（1） Urban subsidence monitoring

Urban subsidence is the vertical downward displacement of the earth's surface caused by natural or human factors during the process of urbanization, which can damage underground pipelines, buildings, and roads in the long run and threaten urban safety.

1. Main causes of settlement

The core is dominated by human factors, among which the extraction of groundwater is the primary reason: large cities such as Shanghai and Beijing have overexploited groundwater due to population and industrial development, resulting in the compression of underground aquifers and causing surface subsidence - in some areas of Shanghai at the end of the 20th century, the annual subsidence exceeded 100mm, and in the plain areas of Beijing, multiple subsidence funnels were formed, accumulating subsidence

2. Monitoring core challenges

•Difficulty in phase unwrapping:In high-resolution SAR images, dense urban high-rise buildings cause interference fringes to overlap and break, directly affecting the accuracy of phase unwrapping and causing distortion of deformation results;

•DEM residual interference:The resolution of external digital elevation models (DEM) is low (such as SRTM 30m DEM), which cannot characterize urban micro topography. Residual terrain phase interference interferes with the extraction of settlement signals, and the impact is more significant in the suburban junction;

•The linear assumption is not applicable:MT InSAR technology defaults to linear (constant rate) deformation of PS/DS points, but groundwater recharge and engineering construction in cities can cause non-linear settlement (such as short-term acceleration and rebound), which can easily lead to misjudgment of deformation information.

（2） Mine deformation monitoring

Mining changes the underground rock structure, causing surface subsidence, cracking and other "goaf subsidence", threatening production safety and damaging ecology. InSAR technology focuses on the dual core of "deformation monitoring trend prediction".

1. Core monitoring content

•Three dimensional deformation monitoring of mining area surface:By fusing multi orbit SAR data (ascending/descending orbit, different incidence angles) and combining GPS calibration, the eastward, northward, and vertical three-dimensional deformation fields are inverted to locate the subsidence center, boundary, and displacement gradient of the goaf, supporting the stability assessment of the goaf;

•Deformation trend prediction:Based on 1-3 years of MT InSAR monitoring data, combined with coal seam burial depth, lithology, and mining progress, grey prediction (GM (1,1)), BP neural network and other models are used to predict deformation trends in the coming months to years and identify high-risk areas in advance.

2. Main technical challenges

The issue of coherence is prominent:The deformation in the mining area has a small range (ranging from tens of square kilometers) and a large gradient (in centimeters per day), far exceeding the InSAR phase entanglement threshold, resulting in a sharp drop or even close to zero coherence in the boundaries of goaf and fracture areas, making it impossible to obtain effective phases;

Application scenario limitations:The current technology focuses on "monitoring and prediction", and there is insufficient support for analyzing the settlement mechanism of mining areas (such as rock stress changes, filling effect correlation) and ecological restoration (land reclamation, vegetation restoration). The data has not been deeply integrated with ecological information.

（3） Seismic deformation monitoring

Earthquake deformation can be classified into three types: coseismic (when an earthquake occurs), post earthquake (stress redistribution), and inter earthquake (stress accumulation). InSAR technology provides key data for earthquake mechanism research and risk assessment.

1. Core monitoring content

•Seismic deformation monitoring:The magnitude of coseismic deformation is large (up to meters, such as the Wenchuan earthquake and the Tohoku earthquake in Japan), and InSAR can quickly obtain the line of sight (LOS) deformation field to identify fault direction and rupture length. However, due to the limitations of side view imaging, it cannot directly invert three-dimensional deformation and requires multi track or GPS fusion; Moreover, near-field deformation can easily lead to decoherence, affecting the accuracy of fault parameters;

•Post earthquake/inter earthquake deformation monitoring:The magnitude of the two is small (mm/year), and MT InSAR technology is required to suppress temporal incoherence (seasonal variation of vegetation) and atmospheric error. At present, it has been applied to the San Andreas fault zone in the United States, the North Anatolian fault zone in Türkiye, the Haiyuan fault zone in China and the Xianshuihe fault zone to support the analysis of seismic period.

2. Technical optimization direction

It is necessary to improve the accuracy of 3D deformation through multi-source data fusion (InSAR+GPS+optical remote sensing), while using high-resolution SAR data (Sentinel-1A/B 3m, TerraSAR-X 1m) to reduce near-field incoherent effects.

（4） Volcanic activity monitoring

Volcanic surface deformation (uplift/subsidence) is a precursor to eruption, and InSAR technology focuses on "early warning mechanism analysis" to support volcanic risk prevention and control.

1. Core monitoring objectives

•Early Warning for Eruption:By capturing high-frequency SAR data (such as Sentinel-1 12 day revisit), surface uplift caused by underground magma chamber filling (centimeter to meter level) can be identified weeks to months in advance to identify eruption risks (such as the 2018 Hawaii Kilauea volcanic eruption warning);

•Parameter inversion of magma chamber:Combining InSAR deformation field with physical models such as Mogi model, invert the position, depth, and volume change rate of magma chambers to determine the intensity of volcanic activity;

•Terrain influence correction:Volcanic areas (cones, lava flow plateaus) have complex terrain and require high-resolution DEM (such as ALOS World 3D 12.5m DEM) to correct geometric distortions and reduce errors caused by the angle between the line of sight and the deformation direction.

2. Application limitations

Volcanic ash and water vapor can reduce the coherence of SAR images, making it difficult to obtain effective data during the active eruption period; Remote volcanoes (Antarctica, Pacific islands) have low SAR data coverage frequency (30 day repetition period), which cannot meet the high-frequency monitoring needs.

（5） Infrastructure deformation monitoring

Infrastructure (bridges, tunnels, highways, transmission towers, tailings dams) are affected by surface deformation in the surrounding area, and damage to foundation stability can lead to structural collapse. InSAR needs to consider the monitoring requirements of "large-scale high-resolution".

1. Monitoring objects and requirements

•Linear facilities (bridges, highways):Monitor settlement and lateral displacement along the line, with a focus on bridge piers and roadbeds;

•Point like facilities (transmission towers):Monitor the vertical settlement and tilt of the tower base;

•Surface facilities (tailings dam):Monitor the overall settlement, displacement, and cracks of the dam body to prevent collapse.

2. Main technical challenges

Traditional SAR sensors have a small antenna area and low resolution (such as the early ENVISAT 25m), which cannot meet the local fine monitoring of infrastructure; Moreover, the coverage of a single track is limited, and large-scale monitoring requires image stitching, which increases the risk of error accumulation; In addition, vegetation and water around the facility will reduce coherence, and metal structures are prone to false phases, which can interfere with deformation identification.

（6） Glacier Movement Monitoring

Glaciers are the "indicators" of climate change, and their movement and material balance reflect the regional climate state. InSAR, due to its advantages of "penetrating ice and snow, all-weather", compensates for the shortcomings of optical remote sensing (affected by clouds, ice and snow cover).

1. Core monitoring content

•Glacier boundary extraction:The low coherence of SAR is caused by the mobility of ice and snow on the surface of glaciers, which differs from the high coherence of surrounding bare rocks and vegetation. It accurately divides the range of ice tongues and cirques, and is suitable for high-altitude and cloudy areas (such as Himalayan glaciers);

•Glacier velocity monitoring:Using D-InSAR technology to calculate phase changes at intervals of several days to months, invert LOS flow velocity, and combine multi orbit fusion to obtain two-dimensional plane flow velocity (accuracy: polar meters/year, Alpine centimeters/day);

•Monitoring of glacier thickness changes:Based on 5-10 years of InSAR elevation data (combined with DEM differencing), the thickness change rate is calculated, and different wavelengths of SAR (L/C band) are used to compare and invert the surface ice and snow thickness, supporting material balance research.

2. Technical features

The advantage is that it is not limited by the optical contrast of ice and snow, and can penetrate dry snow (several meters deep); The limitation is that wet snow absorbs SAR signals, and the monitoring accuracy decreases during the summer ablation period; High flow rate glaciers (over 1km/year) are prone to phase entanglement and require high revisit rate data optimization and disentanglement.

（7） Monitoring of frozen soil process

Permafrost (frozen rock and soil with a temperature below 0 ℃) is widely distributed in high latitude (Arctic, Siberia) and high altitude (Qinghai Tibet Plateau) areas. Climate change and human activities (oil and gas extraction, road construction) have led to its degradation, causing ecological and engineering problems.

1. Impact of permafrost degradation

•Ecological aspect:Soil melting leads to soil loosening, causing soil erosion (such as gully erosion on the Qinghai Tibet Plateau), damaging alpine meadows and coniferous forests, resulting in grassland degradation and desertification;

•Engineering level:The infrastructure in permafrost regions, such as the Qinghai Tibet Railway and the China Russia crude oil pipeline, relies on the stability of permafrost. Freezing and thawing can cause settlement and uneven deformation of the foundation, leading to road cracking and pipeline rupture.

2. InSAR monitoring characteristics

The deformation of frozen soil exhibits a "seasonal slow" pattern:Freezing uplift in winter and melting subsidence in summer, with an annual amplitude of centimeters. MT InSAR technology (such as PS InSAR) extracts seasonal deformation and long-term degradation trends from 1-3 years of data, identifies degradation hotspots (along engineering lines, significant climate warming areas), and has been applied in regions such as the Qinghai Tibet Plateau and the Arctic. However, summer vegetation (alpine meadows) can cause temporal inconsistency, and the PS point extraction algorithm needs to be optimized.

（8） Landslide monitoring

Landslides are sudden disasters caused by the sliding of rock and soil along weak surfaces, with great destructive power. InSAR technology is used for identifying hidden dangers and dynamic monitoring, but it is limited by technical adaptability and environmental interference.

1. Limitations of core technology

•Insufficient adaptability of MT InSAR:The existing technology is based on the development of ground subsidence and defaults to linear deformation, while landslides exhibit nonlinearity (millimeter level/month creep in the early stage, centimeter level/day acceleration in the later stage), and linear models cannot capture the acceleration stage, which can easily miss the warning window; Moreover, the landslide mass is loose rock and soil, making it difficult to extract sufficient PS/DS points and resulting in low data coverage;

•Terrain and vegetation interference:Landslides often occur in mountainous areas, and SAR image overlay and shadows result in some areas having no effective phase; The dense vegetation in mountainous areas (especially during the rainy season) can cause temporal inconsistency, and in severe cases, monitoring results cannot be obtained.

2. Optimization direction

We need to develop nonlinear deformation MT InSAR algorithms (such as Kalman filter dynamic models), integrate LiDAR (terrain correction), optical remote sensing (auxiliary boundary recognition), and use high-resolution, high revisit rate SAR data (Sentinel-1A/B, Gaofen-3) to improve the monitoring accuracy of small-scale high-risk landslides.

3、 Development Trends of InSAR Deformation Monitoring Technology

In response to the current technological bottleneck, we will break through towards the direction of "high precision, wide coverage, intelligence, and multi scenario" in the future:

（1） High resolution and high revisit rate satellite applications

Satellites such as the European Sentinel-1 constellation (6-day revisit), China Gaofen-3 (1m resolution), and the US NISAR (L/C dual band, 12 day revisit) will solve the problems of "balancing large-scale and fine scale" and "high gradient deformation decorrelation", supporting small-scale disaster monitoring such as small landslides and local mine subsidence.

（2） Deepening of multi-source data fusion

InSAR will be deeply integrated with GPS (calibrating 3D deformation), LiDAR (correcting terrain distortion), optical remote sensing (assisting coherence analysis), and geological/meteorological data (quantifying the impact of structure/climate on deformation) to enhance monitoring accuracy and mechanism interpretation capabilities.

（3） Intelligent data processing

Using AI algorithms (CNN, RNN) to achieve automatic screening and phase unwrapping of highly coherent scatterers, shortening processing cycles (from several days to several hours); Build an AI warning model for deformation anomalies, automatically identify risk areas and push warning information to improve decision-making efficiency.

（4） Expansion of Emerging Scenarios

Extending to areas such as monitoring sea level rise (distinguishing absolute sea level rise from surface subsidence), monitoring agricultural water resources (inverting soil moisture content), and monitoring carbon neutrality (assisting in ecological carbon sink assessment), expanding the value of technological applications.

4、 Summary

InSAR deformation monitoring technology has been widely applied in eight major fields, providing irreplaceable high-precision support for geological disaster prevention, ecological protection, and engineering safety. However, it still faces challenges such as decoherence, phase unwrapping, and model adaptability. With the development of high-resolution satellites, multi-source fusion, and intelligent technology, InSAR will further break through bottlenecks and play a greater role in intelligent disaster prevention, ecological civilization construction, major engineering safety, and other fields, providing key technical support for responding to global environmental change and disaster risks.