Pre- and post-failure spatial-temporal deformation pattern of the Baige landslide retrieved from multiple radar and optical satellite images

The Baige landslide in southwest China has been active for at least 50 years, and experienced two failure events on Oct. 10 and Nov. 3, 2018. The rock mass and debris of the slope failure formed a landslide dam. The discharge of the dammed lake plus the failure events caused significant socio-economic losses. Many slow-moving areas appeared after the two failure events and potential landslide risk still exists. Retrieving the long-term spatial-temporal pattern of landslide activity can help us to understand and mitigate landslide disasters. However, it is difficult to monitor all movement episodes of some landslides using single satellite dataset and method, due to the different deformation characteristics. Employing multiple radar and optical images may solve this problem. In this study, the Landsat 8, the ALOS PALSAR and ALOS-2 PALSAR-2 images are used to obtain the surface displacements before the first failure event by using the cross-correlation technique (for Landsat 8 images) and the pixel offset tracking technique. The displacement velocity map after the second failure is derived by the Multi-temporal Synthetic Aperture Radar Interferometry (MT-InSAR) method from the ALOS-2 PALSAR-2 and the Sentinel-1 images. The maximum radar line of sight (LOS) average velocity is 3 m/yr between Jan. 2007 and Feb. 2011, and 12 m/yr between Jul. 2015 and Jul. 2018. From Oct. 2013 to Feb. 2018, the maximum horizontal average velocity is 5.8 m/yr. The movement accelerated from Jul. 2017. A large deformation area appeared on the south of the failed slope after the second failure. The unstable area observed by the MT-InSAR is much larger than that obtained through the field measurements. In the first month after the second failure, the maximum LOS displacements are more than 20 cm. The depletion area and the neighboring zones of the landslide are unstable even one year after the second failure. This study shows how to use multiple satellite images and hybrid remote sensing techniques to investigate the landslide deformation history, and illustrates the benefits of landslide monitoring based on images from multiple satellite sensors.