Geohazard monitoring with satellite radar observations : applications to earthquakes and landslides

Abstract

Over the past few decades, satellite radar observations have developed into a powerful means of monitoring geohazards. What makes it distinguishable is that both transient and long-term deformation involved in geohazards are measurable in detail using the well-established interferometric synthetic aperture radar (InSAR) technique. However, recent and advanced applications encounter limitations due to the decorrelation problem of InSAR when observing large-gradient transient deformation and the lack of subsurface information. In addition, considering the large amount of InSAR observations and the widespread error sources, how to automatically and adaptively identify geohazard-related risk areas on a large spatial scale is also worthy of attention. This thesis aims to address the above-mentioned challenges and thereby improve the application of satellite radar observations to two typical geohazards: earthquakes and landslides, which occur frequently around the world and are interrelated in the sense that landslides can be triggered by earthquakes. Firstly, this thesis overcomes the InSAR decorrelation problem by combining InSAR with image offset tracking techniques so that the complete coseismic deformation of the 2019 Mw 7.5 New Ireland earthquake can be recovered and modelled. Secondly, this thesis incorporates seismic noise measurements to invert subsurface information (e.g., landslide depth) that is hard to obtain with InSAR alone and such a solution was used to investigate a landslide in Bolivia. Finally, a novel InSAR-based automated landslide detection method was developed to detect earthquake accelerated landslides (EALs) following the 2016- 2017 Central Italy earthquake sequence. These EALs responded to coseismic or post-seismic stress disturbances differently from extensively studied coseismic landslides and were typically activated with considerably accelerated ground displacement velocities compared to their preearthquake levels, without acute failures/collapse. This is the first time that an inventory of EALs against catastrophic coseismic landslides was established, which enables a systematic analysis of the spatiotemporal characteristics of EALs and a comprehensive understanding of the prolonged legacy effects of earthquakes on landslides. These works in this thesis provide detailed solutions for monitoring geohazards of different spatial scales and magnitudes using satellite radar observations. Also, they open a new perspective for assessing long-term earthquake-induced landslide risks, which could have important implications for hazard management in seismically active areas.