Lei Xie

On 4 December 2025, nearly two years after the 2024 Mw 7.0 Wushi earthquake, an Mw 5.8 event struck the nearby county of Aheqi, southwestern Tianshan. Owing to the subparallel strikes of both nodal planes and the interspersed hypocenter locations among regional structures in the reported focal mechanisms, the exact fault geometry of this event remains unresolved, impeding a better understanding of regional tectonic activity and the associated seismic hazards. To resolve this, we applied Interferometric Synthetic Aperture Radar (InSAR) technique to map the coseismic deformation and invert for the fault geometry and slip pattern. Significant tropospheric delays are mitigated using a moving-window linear model and a multi-interferogram weighted averaging strategy. The result shows significant uplift (~5.0 cm for ascending track and ~6.0 cm for descending track), indicating thrust-dominated mechanism. Bayesian inversion reveals two possible fault models: a 31.6° north-dipping blind thrust or a 54.4° south-dipping back-thrust. While both fault planes fit the InSAR observations, integrated evidence from the absence of back-thrust development conditions, the surface deformation pattern, and regional topography indicates that the north-dipping Aheqi fault is the causative structure. Together with the steeper Maidan fault to the north, it forms the Orogen Basin boundary along the southern Tianshan piedmont. Our findings highlight that resolving moderate blind-thrust seismogenic structures using InSAR requires integration with pre-existing structural and geomorphic evidence. Furthermore, Coulomb stress calculations indicate a rupture-promoting effect from the Wushi earthquake, which occurred on a reactivated fault, onto the Aheqi event, with stress loading exceeding 2 bar at the hypocenter. Thus, the potential for stress-driven sequential rupture between reactivated and present-day active structures necessitates an updated seismic hazard assessment in the southern Tianshan.

Unwrapping synthetic aperture radar interferograms with extensive low-coherence regions remains challenging, even when the deformation signal has a low gradient, such as that due to interseismic displacement. Here, we present a novel algorithm to unwrap low-gradient interferograms more efficiently and reliably using a semantic segmentation neural network. We first partition large-scale interferograms into overlapping patches and employ a trained Segformer network, making full use of spatial features, to identify decorrelated pixels while predicting the wrap count for coherent pixels. In a further step, we correct wrap counts or re-unwrap certain patches, based on a reliability metric. Finally, the patches are mosaicked to reconstruct the fully unwrapped interferogram by leveraging overlapping areas. The Segformer model is trained on over 20,000 simulated samples with varying decorrelation noise and more than 10,000 real-world samples from the COMET-LiCSAR portal. Synthetic experiments show that our approach significantly reduces the mean absolute error compared to the classical minimum cost flow method. Further validation on thousands of interferograms from the western Altyn Tagh Fault and the western Haiyuan Fault confirms its superior performance in isolated regions separated by decorrelated noise and its ability to suppress the wide-ranging propagation of unwrapping errors, with the proportion of unwrapping errors reduced by 28-83% for real interferograms. These results highlight the potential of our method to be used in large-scale automated processing of straining regions, such as the Alpine-Himalayan Belt.

Landslides are recurrent and far-reaching natural disasters that exert significant impacts on infrastructure stability and human livelihood. Interferometric synthetic aperture radar (InSAR) has been widely adopted for landslide deformation monitoring and early hazard detection with its features such as broad spatial coverage and high accuracy. However, in mountainous canyon regions, which are more likely to the unstable rock and soil movements, InSAR is particularly susceptible due to elevation-correlated tropospheric delays. Existing correction methods, predominantly based on mathematical modeling of atmosphere-elevation relationships, exhibit limitations to estimate topography-correlated tropospheric delays with heterogeneous terrain in the spatial domain. This study proposes a new wavelet multiresolution analysis approach, namely quad-multiresolution analysis (Quad-MRA), which accounts for spatial heterogeneity of tropospheric delays in the frequency domain with a quad-tree segmentation strategy. The validation in simulated experiments indicates Quad-MRA method not only mitigates atmospheric delays but also enhances phase stability, reducing the standard deviation (STD) of interferograms from 3.8 rad to 0.7 rad. We further found that the applicability of this method is influenced by the proportion of landslide-affected pixels in the interferogram, with optimal performance observed when the pixel proportion is 2% or lower. This method also has been employed for the Zhongpai Landslide, Gansu, China, using the Sentinel-1 data, in which various lines of highways passby. The results demonstrate a 50% reduction in STD of phase fluctuation, and the root mean square error for InSAR and GNSS improves from 32 mm to 7 mm. The results also identified four distinct deformation transitions between 2021 and 2024, and each associated with accelerated deformation triggered by rainfall. These findings demonstrate the effectiveness of the proposed method in mitigating topographic related tropospheric delays in the mountainous terrain, and particularly benefit for the landslide monitoring, detection, and mechanism interpretation.

The 2024 Mw 7.0 Wushi and 2020 Mw 6.0 Jiashi earthquakes, two recent large events in the southern Tian Shan, provide rare opportunities for understanding the seismotectonic deformation patterns in compressional orogenic regimes. Here, we utilize Sentinel-1 satellite radar images and Differential Interferometric Synthetic Aperture Radar techniques to map coseismic deformation fields for the 2024 Wushi and 2020 Jiashi earthquakes, occurring at the front and root of the Kepingtage fold-and-thrust belt, respectively. A modified Bayesian nonlinear method is utilized to simultaneously invert fault geometry parameters and slip distributions. Our findings show that both Wushi and Jiashi earthquakes are predominantly characterized by reverse/thrust motions with minor strike-slip components. The spatial continuity of the interferometric fringes and coseismic deformation fields suggests shallow (<5 km depth) folding deformation in response to deeper reverse/thrust faulting. Specifically, the Wushi earthquake produced a high-angle fault-propagation fold, with fault-slip penetrating into the crystalline basement, whereas the Jiashi earthquake formed a low-angle detachment fold developed within the cover layer. The Mw 5.7 aftershock following the Wushi mainshock compensated for the shallow coseismic slip deficit and generated positive flowerlike ruptures superimposed on the pre-existing mainshock fold deformation, explaining the observed opposite dip direction of the aftershock rupture compared to the mainshock mechanism. These two fold-type earthquakes demonstrate the dominance of folding deformation in accommodating crustal shortening and orogenic growth in the southern Tian Shan, elucidating compressional tectonic processes.

On 8 September 2023, an Mw 6.8 earthquake struck the High Atlas Mountains in western Morocco, where the tectonic regime has been poorly investigated due to its remoteness and weaker seismicity compared to the northern plate boundary. In this study, we combine the measurements from the Interferometric Synthetic Aperture Radar images and the seismic data to invert the coseismic slip model of the 2023 Morocco earthquake. The results show a predominantly reverse slip motion with a minor left-lateral strike slip. The rupture process lasts about 15 s and reaches the maximum of its seismic moment release rate at about 5 s. The coseismic slip is mainly distributed in a depth range of ~20–30 km, with the ~1.4 m maximum coseismic slip at a depth of ~25 km. The Coulomb stress change suggests a significant stress loading effect on surrounding faults. The high-angle transpressive rupture kinematics of the 2023 Morocco earthquake reveal steep oblique–reverse faulting of the Tizi n’Test fault within the western High Atlas Mountains. The slight left-lateral strike slip and focal depth anomaly of this event are largely attributed to differential crustal shortening and the rejuvenation of early rift structures inherited from the Mesozoic complex evolution.

Rupture characteristics and heterogeneity of large earthquakes are essential for seismic hazard assessments. We use relocated aftershocks, geodetic measurements, and seismic waveform data to distinguish contributions from closely separated fault structures of the 2024 M _w 7.5 Noto earthquake. We find that the initial rupture triggered slip on a complex fault of a preceding swarm and led to bilateral slow rupture there. The earthquake ruptured two fault segments with contrasting dip angles along the eastern and western segments. Aftershocks continued to rupture the preexisting swarm faults. A delayed rupture occurred southwest of the hypocenter, implying that substantial resistance caused by a barrier temporally hindered rupture propagation. Additional stress from surrounding slip eventually overcame the strength of the barrier fault section, leading to a compound rupture. The mainshock triggered a small earthquake swarm, in which the relatively larger events were not followed by abundant aftershocks. Our findings demonstrate the influence of strong asperities and complex geometry in the progression of cascading ruptures.

Breakwater dams are critical infrastructures that protect the safety of ports. However, these coastal structures are facing the compounding threats of sea level rise, storm surge, and dam subsidence. Heterogeneous deformations in these infrastructures arise from differential construction sequencing, sediment consolidation, and filling materials, yet traditional in situ monitoring remains spatially limited or even unavailable to trace back and continuously monitor deformation evolutions. In contrast, Interferometric Synthetic Aperture Radar (InSAR) offers valuable insights in providing the spatially and temporally covered dam deformation. In this study, we used two Sentinel-1 tracks from 2016 to 2025, and the persistent and distributed scatterers InSAR methods to map the long-term deformation of Xuwei Port, Lianyungang, China. We utilized six sites of leveling measurements to validate the InSAR-derived vertical deformation and indicate Root Mean Square Errors (RMSEs) ranging from −0.9–1.2 cm. We find, for the rock-sand filled section, the deformations show consolidating subsidence ranging from −63.8 cm to −40.6 cm. In contrast, the concrete tubular structure remains stable, with cumulative deformation ranging from −10.6 cm to −5.2 cm. The enclosing reclaimed land undergoes a period of accelerated settlement with subsidence rates of −64.9–−39.3 cm/yr, which are higher than original subsidence rates of −10.1–−9.7 cm/yr. Additionally, we integrated the consolidation model and tide gauge to quantify that the freeboard will decrease to 0.08–0.31 m in the following 100 years with the continuous sea level rise and dam subsidence. This study benefits our understandings of coastal dam and reclaimed land. It highlights InSAR as a valuable tool to evaluate the critical risk between sea level rise and coastal infrastructure subsidence.

Underground water leaks in urban networks result in significant resource loss, infrastructure degradation, and environmental damage—challenges that are particularly acute in high-density cities like Hong Kong, where aging and complex infrastructure complicates detection. Traditional methods such as acoustic sensing and manual inspections often fall short in efficiency and scalability. This study proposes the use of L-band SAR imagery from ALOS-2 combined with machine learning techniques to address these challenges. A robust leak detection framework was developed using six dual-polarized SAR images (HH and VV modes) alongside historical leak data from the Hong Kong Water Supplies Department (WSD). Features extracted via Gray-Level Co-occurrence Matrix (GLCM) metrics and backscattering coefficients were used to train various machine learning, deep learning, and ensemble learning models, with hyperparameter optimization performed using a grid search algorithm. Among these, the stacking algorithm delivered the best performance, achieving an accuracy of 80%. Despite these promising results, several critical issues remain unresolved—particularly data imbalance, the incorporation of physical leak characteristics, and the integration of additional environmental factors. Future research will focus on these challenges by exploring new data sources, such as four-polarization ALOS-2 images and Sentinel-1 C-band data, as well as advanced polarimetric and interferometric techniques, to further enhance the robustness and accuracy of leak detection models.

Understanding urban surface deformation from InSAR one dimensional line-of-sight (LOS) observation is challenging due to the interplay of various natural and anthropogenic driving factors. This study proposes a novel multi-scale InSAR deformation decomposition method that integrates two-dimensional fast Fourier transform (2D-FFT) with a multi-scale geographically weighted regression model (MGWR). We apply this method to analyze urban ground deformation in the Beijing Plain, China, using Sentinel-1 data from 2016 to 2023. The results identify five significant subsidence areas with the maximum subsidence reaching -360 mm. The decomposed large-scale component (4–8 km) correlates closely with the thickness of Quaternary sediments and variations of groundwater. The medium-scale component (1–3 km) is associated with the urban construction activities, while the small-scale deformation (<0.6 km) reveals localized subsidence along subway with tunnel construction. These findings demonstrate the effectiveness of the proposed method in capturing the multi-scale deformation patterns, especially the concealed small-scale deformation, through a single orbit of InSAR observation. Therefore, it shows high potentials to improve the urban infrastructure safety such as subway engineering.