The Mw 7.0 Dingri earthquake, which occurred on 7 January 2025, occurred at the southern end of the Xainza-Dinggyê Fault Zone within the South Tibet Rift (STR) system, in the Dengmecuo graben. It is the largest normal-faulting event in the region recorded by modern instruments. Using Sentinel-1A and Lutan SAR data combined with strong-motion records, we derived the coseismic surface deformation and slip distribution. InSAR interferograms and displacement vectors confirm a typical normal-faulting pattern. The slip model, based on an elastic half-space assumption, identifies the Dengmecuo Fault as the source fault, with an average strike of ~187° and a dip of ~55°. The rupture was concentrated within the upper 10 km, with a maximum slip of 4–5 m at ~5 km depth, extending to the surface with ~3 m vertical displacement. Partial rupture (≤2 m) in the southern segment (5–10 km depth) did not reach the surface, likely due to lacustrine deposits or possible post-seismic stress release. The rupture bottom intersects the fault plane of the South Tibet Detachment System (STDS), suggesting a restraining effect on coseismic rupture propagation. Considering stress transfer along the Main Himalayan Thrust (MHT), we propose that the 2025 Dingri earthquake is closely associated with stress transfer following the 2015 Gorkha earthquake in the lower Himalayas.

Interferometric synthetic aperture radar (InSAR) technology plays a crucial role in monitoring surface deformation and has become widely used in volcanic and earthquake research. With the rapid advancement of satellite technology, InSAR now generates vast volumes of deformation data. Deep learning has revolutionized data analysis, offering exceptional capabilities for processing large datasets. Leveraging these advancements, automatic detection of volcanic and earthquake deformation from extensive InSAR datasets has emerged as a major research focus. In this paper, we first introduce several representative deep learning architectures commonly used in InSAR data analysis, including convolutional neural networks (CNNs), recurrent neural networks (RNNs), generative adversarial networks (GANs), and Transformer networks. Each architecture offers unique advantages for addressing the challenges of InSAR data. We then systematically review recent progress in the automatic detection and identification of volcanic and earthquake deformation signals from InSAR images using deep learning techniques. This review highlights two key aspects: the design of network architectures and the methodologies for constructing datasets. Finally, we discuss the challenges in automatic detection and propose potential solutions. This study aims to provide a comprehensive overview of the current applications of deep learning for extracting InSAR deformation features, with a particular focus on earthquake and volcanic monitoring.

Focusing on stratified tropospheric delay correction in the small-amplitude coseismic displacement field of small-to-moderate earthquakes (<Mw 6.5), we develop a Simple-Stratification-Correction (SSC) approach based on the empirical phase-elevation relationship and spatial properties of the troposphere, via an equal-size window segmentation. We validate our SSC method using 23 real earthquakes that occurred from January 2016 to May 2021 with a moment magnitude (Mw) ranging from 4.5 to 6.5. We conclude that SSC performs well according to the amount of reduction in semi-variance and the root-mean-square value. This method primarily focuses on stratification delay correction; thus, it is especially useful in regions with complex terrain, while it can mitigate partial large-scale turbulence signals. We investigate three parameters that are empirically setup in the correction working flow and inspect their optimal settings, when implementing SSC for quick response after earthquake. Our method is ready to be integrated into an operational InSAR processing chain to produce a reliable atmospheric phase screen map, which can also serve as an auxiliary product to quickly and timely quantify stratification delays in coseismic interferograms. Through improved accuracy of the coseismic displacement field, the focal mechanism could be better constrained to facilitate the building and expansion of the geodesy-based earthquake catalogue.

Interferometric synthetic aperture radar (InSAR) technology has been widely employed in the rapid monitoring of earthquakes and associated geological hazards. With the continued advancement of InSAR technology, the growing volume of satellite-acquired data has opened new avenues for applying deep learning (DL) techniques to the analysis of earthquake-induced surface deformation. Although DL holds great promise for processing InSAR data, its development progress has been significantly constrained by the absence of large-scale, accurately annotated datasets related to earthquake-induced deformation. To address this limitation, we propose an automated method for constructing deep learning training datasets by integrating the Global Centroid Moment Tensor (GCMT) earthquake catalog with Sentinel-1 InSAR observations. This approach reduces the inefficiencies and manual labor typically involved in InSAR data preparation, thereby significantly enhancing the efficiency and automation of constructing deep learning datasets for coseismic deformation. Using this method, we developed and publicly released a large-scale training dataset consisting of coseismic InSAR samples. The dataset contained 353 Sentinel-1 interferograms corresponding to 62 global earthquakes that occurred between 2015 and 2024. Following standardized preprocessing and data augmentation (DA), a large number of image samples were generated for model training. Multidimensional analyses of the dataset confirmed its high quality and strong representativeness, making it a valuable asset for deep learning research on coseismic deformation. The dataset construction process followed a standardized and reproducible workflow, ensuring objectivity and consistency throughout data generation. As additional coseismic InSAR observations become available, the dataset can be continuously expanded, evolving into a comprehensive, high-quality, and diverse training resource. It serves as a solid foundation for advancing deep learning applications in the field of InSAR-based coseismic deformation analysis.

Three-dimensional coseismic surface deformation fields are important for quantifying the geometric and kinematic characteristics of earthquake rupture faults. However, traditional geodetic techniques are constrained by intrinsic limitations: Interferometric synthetic aperture radar (InSAR) can only extract far-field deformation fields owing to incoherence; global navigation satellite systems (GNSSs) can only acquire displacement at discrete points. The recently developed optical pixel correlation technique, which is based on high-resolution remote sensing images, can acquire near-field coseismic horizontal deformation. In this study, InSAR line-of-sight (LOS) and azimuth direction far-field deformation, horizontal near-field deformation determined using optical pixel correlation based on pre- and post-earthquake GaoFen (GF)-2/7 images, and vertical deformation determined by differencing pre- and post-earthquake GF-7 digital elevation models (DEMs) were combined to comprehensively provide the three-dimensional deformation field of the 2022 Mw 6.6 Menyuan earthquake. The results show that the near-field deformation field calculated by optical pixel correlation quantified displacements distributed over the rupture fault zone, which were not available from the InSAR deformation maps. We identified significant vertical displacements of ~1–1.5 m at a bend region, which were induced by local compressive stress. The maximum uplift (>2.0 m) occurred near the epicenter, on the southern sides of the main and secondary faults along the middle segment of the ruptured Lenglongling fault. In addition, surface two-dimensional strain derived from the displacement maps calculated by optical pixel correlation revealed high strain concentration on the rupture fault zone. The method described herein provides a new tool for a better understanding of the characteristics of coseismic surface deformation and rupture patterns of faults.

The coseismic surface three-dimensional deformation fields were significant for quantified the geometric and kinematic characteristics of earthquake rupture faults. Traditional geodetic techniques were constrained by their intrinsic limitations, such as InSAR, which could only extract far-field deformation fields due to its incoherence, and GNSS, which could only acquire deformations at discrete points. The recently developing optical pixel correlation technique based on high resolution remote sensing images could acquire the near-field coseismic horizontal deformation. In this study, the vertical displacement field of the 2022 Mw 6.6 Menyuan earthquake was obtained by differencing the pre- and post-earthquake DEMs from GaoFen-7 stereo satellite images. Further, the line-of-sight (LOS) and azimuthal (AZI) directions far-field deformations by InSAR, the horizontal near-field deformation by optical pixel correlation based on pre- and post-earthquake GF-2/7 images, and the vertical deformation by differencing the pre- and post-earthquake DEMs, were incorporated to comprehensively solve the complete three-dimensional deformation fields of the 2022 Mw 6.6 Menyuan earthquake. Firstly, the vertical displacement field by differencing DEMs indicated there were significant vertical displacements about 1 m at the bend region, which was induced by the local compressive stress. In addition, at the epicenter on the middle segment of ruptured Lenglongling fault (LLLF), the maximum lifting occurred on the southern sides of the main and secondary faults exceeded 2 m. Secondly, the three-dimensional deformation fields solved by multiple deformation data demonstrated that the near-field deformation field calculated by optical pixel correlation method could quantify the displacements distributed over the rupture fault zone, which were not available on the InSAR deformation maps. Finally, the surface 2D strain derived from the displacement maps calculated by optical pixel correlation revealed high strain concentration on the rupture fault zone. Our study focused on the complete surface three-dimensional deformation estimation from multiple far- and near-field deformation data, and provided a new perspective for a deeper understanding of the characteristics of coseismic surface deformation and the rupture pattern of the fault.