Conference Agenda

Under the framework of the Dragon Programme (ID: 95348), jointly initiated by ESA and the Ministry of Science and Technology of China (ESA-MOST), the National Institute of Geophysics and Volcanology (INGV) of Italy, in collaboration with Northeastern University, China University of Mining and Technology, Earthquake Agency of Jilin Province, and Jilin University of China, has been conducting collaborative research on surface deformation monitoring using multi-source remote sensing data. Building upon the Dragon-4 and Dragon-5 programmes, the Sino-Italian joint team aims to identify, quantify, and interpret surface deformations caused by volcanic dynamics, fluid migration, slope mass movement, and anthropogenic engineering activities. In the current year, the research was carried out along five directions: soil moisture interference correction in slope deformation monitoring, robust phase unwrapping for landslides with large deformation gradients, quantitative risk assessment of debris flow hazards, deformation-stress interpretation of a large cable-stayed bridge, and remote sensing detection of precursory signals prior to a submarine volcanic eruption. These studies employed multi-temporal InSAR, numerical simulation, and finite element analysis, integrating SAR imagery from Sentinel-1A, TerraSAR-X, and COSMO-SkyMed, as well as UAV photogrammetric data.

In slope deformation monitoring, periodic oscillation signals consistent with seasonal soil moisture variations commonly contaminate InSAR time series. This study selected the Huangnibazi landslide as a test site and conducted deformation monitoring using 132 Sentinel-1A and 44 TerraSAR-X images. A coupled model linking soil moisture, dielectric properties, penetration depth, and phase delay was established through microwave radiometer measurements, elucidating the complete process by which moisture variations generate spurious deformation signals. The correlation coefficient between Sentinel-1A deformation and soil moisture reached 0.638, and the response lag of deformation to moisture (2.13 days) was significantly shorter than that to precipitation (4.78 days), confirming soil moisture as the direct factor generating spurious deformation. The results have been published in IEEE TGRS.

Insufficient phase unwrapping accuracy in areas with large deformation gradients represents another critical bottleneck. This study proposed the Phase Gradient Rate constrained Minimum Cost Flow (PGR-MCF) method, which estimates the phase gradient rate through temporal stacking and incorporates it as a priori constraint into the MCF unwrapping network. Using the Guobu landslide as a test site, 264 interferograms from 91 Sentinel-1 images were used for validation. Evaluation at three GNSS points demonstrated that the RMSE was reduced by more than 69% compared to all existing methods. The results have been published in ISPRS Journal of Photogrammetry and Remote Sensing.

For debris flow risk assessment, taking the Wudaogou gully in Fushun as a case, a building vulnerability scheme combining MassFlow numerical simulation with Abaqus finite element analysis was proposed. Debris flow dynamics under different rainfall return periods were simulated using a 0.5 m resolution UAV-derived DEM, and a physical vulnerability function was constructed by coupling flow velocity and deposit depth. Under the 1000-year extreme rainfall scenario, approximately 57% of buildings would suffer damage, with 14% facing complete destruction. The results have been published in Landslides.

For structural health monitoring, the Sanhao Bridge in Shenyang was selected, and a deformation-stress assessment scheme integrating ascending-descending InSAR fusion with finite element modeling was proposed. Using 30 ascending TerraSAR-X and 29 descending COSMO-SkyMed images, seasonal thermal and long-term structural deformation were separated, and SVD-based fusion was employed to derive the two-dimensional deformation field. The InSAR-derived vertical deformation was input into a finite element model, revealing critical stress concentrations at pier supports, tower-beam connections, and cable-beam intersections, with discrepancy less than 1 mm. The results have been published in IEEE JSTARS.

For volcanic deformation monitoring, this study focused on the 2021–2022 HTHH submarine volcano eruption. A comprehensive approach utilizing Sentinel-2A, Sentinel-1A, and Planet SkySat imagery revealed the emergence of a new primary vent in the northeastern part of the island. SBAS-InSAR processing of 44 Sentinel-1A images showed that significant displacement with a maximum cumulative LOS deformation of 6.4 cm had occurred in the new crater area prior to the December 2021 eruption, and magma intrusion may have commenced as early as May 2020, lasting up to 19 months. This demonstrates that satellite monitoring of small islands at summits of submarine volcanoes can provide valuable eruption precursor information. The results have been published in the Bulletin of Volcanology.

The analysis of deformation associated with fluid extraction and injection has broader significance for natural hazard assessment and subsurface resource sustainability. Using the SBAS technique with Sentinel-1 descending images (2021–2025), our study of the Daqing Oilfield in the Songliao Basin revealed widespread ground uplift in extraction areas with velocities up to 40 mm/yr, and a clear shift from uplift to subsidence (−20 mm/yr) along a fault system. GNSS stations confirmed low horizontal velocities (~2 mm/yr), enabling projection of the LOS velocity to vertical for source modeling considering reservoir depths.

Finally, we have initiated a study on Hainan Island (South China), characterized by well-developed fault systems, volcanic activity, and seismic hazards. Haikou is undergoing rapid urbanization with critical infrastructure frequently located near active fault zones. We propose to apply MT-InSAR using COSMO-SkyMed (CSK) and COSMO Second Generation (CSG) high-resolution data covering Haikou and Sanya along ascending and descending orbits. The MT-InSAR processing will be performed in the third and fourth years of the Dragon-6 project.

This work summarizes the mid-term advances of the ReSeLMAIN project (ID: 95355 – Remote Sensing for Landslide Monitoring and Impact Assessment on Infrastructure), developed within the framework of the DRAGON-6 cooperation program, a joint initiative between the European Space Agency (ESA) and the Chinese Ministry of Science and Technology (MOST). It integrates AI-driven automated workflows, multi-temporal InSAR interferometry, and infrastructure impact analysis for the detection, kinematic characterization, and monitoring of geohazards across Spain and China. Firstly, we present new automated workflows for the detection of active landslides using deep learning-based phase gradient stacking and spatial clustering. This methodology significantly reduces the time required to identify potential landslides that may threaten infrastructure and improves their mapping. Regarding infrastructure impact, we analyze specific case studies such as the Beijing-Tianjin high-speed railway, where high-resolution TerraSAR-X data are used for structural health monitoring of the railway corridor affected by land subsidence. Furthermore, we explore the kinematic behavior of landslides using 3D time series and optical pixel offsets, providing critical data on slip surfaces (i.e., depth and shape) and volume evolution in Spain and China. Additionally, we monitor a sinkhole and a landslide in two different urban areas of Alicante (Spain) using open-access datasets from the European Ground Motion Service (EGMS) and evaluate their impact on infrastructure. Finally, high-resolution PAZ imagery is used to achieve millimeter-level precision displacements in urban environments and to monitor a collapsed bridge. Overall, these case studies illustrate the current status and achievements of the DRAGON-6 project ReSeLMAIN.