Rapid Emergency Response Assessment of Earthquake-Induced Landslides Driven by Fusion of InSAR Deformation Data and Newmark Physical Models

• Published in: Remote sensing (Basel, Switzerland) Vol. 15; no. 18; p. 4605
• Main Authors: Zeng, Ying, Zhang, Yingbin, Liu, Jing, Wang, Qingdong, Zhu, Hui
• Format: Journal Article
• Language: English
• Published: Basel MDPI AG 01.09.2023
• Subjects: Aerial photography; Artificial satellites in remote sensing; China; Deformation; Disaster relief; Disasters; Earthquake damage; Earthquake prediction; earthquake-induced landslides; Earthquakes; Emergency preparedness; Emergency response; Environmental aspects; Equilibrium; Identification; Information systems; Landslides; Landslides & mudslides; Methods; Model accuracy; Newmark; Performance prediction; prediction; Prediction models; Remote sensing; Rescue operations; Safety and security measures; SBAS-InSAR; Search and rescue; Search and rescue operations; Seismic activity; Seismic response; the rapid emergency assessment of landslides; Unmanned aerial vehicles; China
• DOI: 10.3390/rs15184605
• ISSN: 2072-4292

Strong earthquakes induce a large number of secondary disasters, such as landslides, which bring serious challenges to post-disaster emergency rescue, and the rapid and accurate assessment of earthquake-induced landslide disasters is crucial for post-earthquake emergency rescue. This research aims to propose an emergency assessment model that is suitable for post-earthquake landslides, specifically targeting the first 72 h after an earthquake for emergency rescue guidance. The model combines remote sensing technology and the Newmark physical mechanics assessment model to form the InSAR Data–Newmark Physical Fusion Driver Model (IDNPM), which comprehensively considers the dynamic deformation of the ground surface and geological features. To validate the predictive performance of the IDNPM, the model is applied to the 5 September 2022 Luding earthquake event and the 8 August 2017 Jiuzhaigou earthquake event. The landslide qualitative evaluation, confusion matrix and Receiver Operating Characteristic (ROC) curve are utilized for quantitative assessment. The results show that the IDNPM can effectively reduce the false negative and false positive errors in landslide prediction by utilizing the SAR deformation information, and to a certain extent, it accounts for the dependence of the Newmark model on the accuracy of empirical formulas and geotechnical parameters. For the Luding earthquake event, the IDNPM shows an accuracy improvement of 10.296% compared to the traditional Newmark model. For the Jiuzhaigou earthquake event, there is also an improvement of 3.152%, with a promising generalization performance. The simplicity and ease of operation in constructing the model are accompanied by high reliability and accuracy. The research findings provide essential references for the development of post-earthquake landslide emergency prediction models and offer robust data support for emergency rescue and recovery efforts in earthquake-stricken areas in the future.