泸定地震典型滑坡震后快速形变监测研究
Post-Earthquake Rapid Deformation Monitoring of Typical Landslides Triggered by the Luding Ms6.8 Earthquake

地震滑坡是地震诱发的主要次生灾害之一，具有突发性强、持续性长等特点，对震后区域安全构成严重威胁。为实现震后滑坡的高效监测，本文以2022年四川泸定Ms6.8级地震为背景，选取典型滑坡区域，利用HyP3在线SAR数据，结合MintPy时序InSAR技术，构建了震后滑坡形变快速监测流程。针对传统InSAR方法中存在的解缠误差积累、参考点选取不合理和噪声干扰等问题，本文引入闭合相位法评估解缠误差，结合水体掩膜剔除潜在误差区域，并通过残差相位均方根(RMS)与中值绝对偏差(MAD)优化参考日期选取，从而提高了形变监测的精度与稳定性。结果显示，H1与H2两个典型滑坡区在地震后均表现出持续沉降特征，最大沉降速率分别达？13.2 cm/年和？16.7 cm/年，累积沉降量分别为？11.87 cm和？12.29 cm，滑坡体中部及坡脚变形显著，反映出滑坡体存在长期蠕动或进一步失稳的风险。研究表明，相比传统处理方式，HyP3-MintPy流程显著提高了数据处理效率和自动化程度，适用于震后应急滑坡监测的快速响应需求。
Earthquake-induced landslides are among the most significant secondary hazards caused by seismic events, characterized by sudden onset and long-term persistence, posing serious threats to post-earthquake regional safety. To achieve efficient monitoring of post-seismic landslides, this study takes the 2022 Luding Ms6.8 earthquake in Sichuan, China as the research background. Typical landslide areas were selected, and a rapid deformation monitoring workflow was constructed by integrating HyP3 online SAR data with MintPy time-series InSAR techniques. To address challenges in traditional InSAR processing—such as error accumulation during phase unwrapping, suboptimal reference point selection, and noise interference—this study employed the closure phase method to assess unwrapping errors, applied water body masking to eliminate potential error-prone areas, and used root mean square (RMS) of residual phases along with median absolute deviation (MAD) to optimize the selection of reference dates. These measures improved the accuracy and stability of deformation monitoring. Results show that both H1 and H2 typical landslide zones exhibited sustained post-seismic subsidence, with maximum deformation rates of ？13.2 cm/year and ？16.7 cm/year, and cumulative displacements of ？11.87 cm and ？12.29 cm, respectively. Significant deformation was concentrated in the central and lower parts of the landslide bodies, indicating prolonged creeping or potential instability. The study demonstrates that, compared with traditional processing methods, the HyP3-MintPy workflow significantly enhances data processing efficiency and automation, making it suitable for rapid-response monitoring of landslides in post-earthquake scenarios.