InSAR-derived predisaster spatio-temporal evolution of a reactivated landslide

Slow-moving reactivated landslides can accelerate suddenly and fail catastrophically, posing a great threat to life and economy. Emerging synthetic aperture radar (SAR) technique for long-term monitoring of such landslides has been devoted to documenting precursory movements before the failure; however, understanding the complex spatiotemporal evolution of a slow-moving reactivated landslide in different parts remains a challenge. Here we present an exemplification of potential reactivation and spatiotemporal evolution based on a recent reactivated landslide in southwestern China. We conducted multi-temporal Interferometric SAR (MT-InSAR) using Sentinel-1 constellation data, spanning 4 years (2014~2018) period, and then retrieved the spatiotemporal deformation pattern for ascending orbit. Results indicated that prior to the large-scale failure on 19 July 2018, the landslide exhibited prolonged slow movement, with an annual line of sight (LOS) velocity reaching −67.2 mm/year. The intense and persistent precipitation in 2018 resulted in higher average LOS velocities compared to those observed during 2014-2017. We have also identified three potential deformation zones, and a detailed analysis of slope deformation in different sections unveiled that the leading edge exhibited the largest displacement, followed by the rear section. These findings strongly suggest that this landslide is most likely a compound event involving retrogressive and progressive failure modes. The 4-year vegetation indices unravel that the decline in values can be attributed to alterations in vegetation structure prior to the extensive failure, as no similar behaviors were observed in historical data. This study highlights the significance of SAR data for monitoring landslide reactivation and contributes to a more comprehensive understanding of the spatiotemporal evolution of slow-moving landslides.