The geohazards associated with strongly deformed and reactivated large-scale ancient landslide are analyzed through a study of the Xiongba ancient landslide. The SBAS-InSAR method, combined with remote sensing interpretation, was used to obtain the Xiongba ancient landslide surface deformation characteristics, on the western bank of the Jinsha River, during the period from October 2017 to June 2020. Two large strong deformation zones were discovered in this study, H1 and H2, which were located at the front edge of the Xiongba landslide. The maximum cumulative deformation in the H1 deformation zone was approximately 204 mm, and the deformation in the H2 deformation zone was approximately 302 mm. Influenced by the Jinsha River erosion, the Baige landslide-dammed lake-dam breakage-debris (LDLDB) flow/flood hazard chains, which occurred 75 km upstream reaches in October and November 2018, and the erosion of the foot of the Xiongba ancient landslide foot resulted in notably enhanced deformation. The creep rate in the H1 deformation zone was 14~16 times that before the Baige landslide hazard chains occurred, and the hazard chains caused sliding in the H2 zone. The Xiongba ancient landslide is undergoing retrogressive reactivation. The Xiongba ancient landslide is currently experiencing continuously creep-sliding, and the deformation rate in some areas is accelerating, which may induce a large-scale reactivation of the Xiongba ancient landslide and an LDLDB hazard chain.
Field surveys and radiocarbon dating of detrital materials provide evidence that repeated landslides dammed the Yigong Tsangpo River ca. 3500 bc, 1300 bc, 1000 bc, 600 bc, and twice more recently. Together with historical slides in 1900 and 2000, these six older slides make for a total of eight known channel-damming landslide events at the same location over the past six millennia, indicating sub-millennia recurrence intervals over this time period. Together with the likely incomplete nature of the sedimentary record of past channel-damming episodes uncovered to date, our findings indicate late Holocene multi-century-scale recurrence intervals for large landslides at this location. Hence, the riverbed at and immediately upstream of this location may have been inundated by sediment, and therefore not incising, for much of the post-glacial period. Together with the location of this landslide complex at the head of the major knickzone defining the fluvial edge of the Tibetan Plateau, our findings support the hypothesis that repeated glacial and landslide damming in this region inhibited headward propagation of river incision into the Tibetan Plateau.