.png){kind=icon}
.png){kind=icon}
Return
Investigation into the activation features of representative faults and their triggering mechanisms for geohazards in the Barkam area
.png){kind=icon}
PRE
AI
E
W
Y
W
Y
C
H
J
DOI：10.1080/01431161.2026.2693201.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
The 2022 Ms 6.0 Barkam earthquake occurred on the Songgang Fault, located inside the Bayan Har block along the eastern margin of the Tibetan Plateau, and triggered numerous geological hazards. Based on GF-2 and Jilin-1 high-resolution remote sensing imagery, DEM, InSAR data, deep learning recognition results, and field investigation data, we interpreted the activity characteristics of the Songgang Fault and analysed the post-earthquake geohazard evolution. The results indicate that the Songgang Fault generally trends NW – SE, dips towards the NE with a dip angle of approximately 50°–70°, and is characterized by tectonic landforms such as fault scarps, linear troughs, drainage offsets, and ridge grooves along its length. It is therefore interpreted as a Late Quaternary active fault with primarily reverse motion and a subordinate left-lateral strike-slip component. Automatic recognition of post-earthquake landslides was performed based on the U‑Net and ResNet50 model, which yielded an overall accuracy of 90.37% and an F1‑score of 75.63%. In total, 293 landslides were detected. Through the integration of field investigations, radiocarbon dating, and multivariate statistical analysis, the geohazard evolution in the Barkam region exhibits a staged characteristic: ‘tectonic pre‑conditioning – coseismic triggering – post‑seismic adjustment’. Although the activity of the Songgang Fault affects the spatial distribution of landslides via rock mass fracturing, differential topographic formation and seismic triggering effects, hazard development is not governed solely by the fault; instead, it is the joint outcome of faulting, slope angle, lithology, river erosion, and human‑induced disturbance. This research can serve as a reference for recognizing intra‑block active faults on the eastern margin of the Tibetan Plateau, for rapid post‑earthquake geohazard extraction, and for regional risk mitigation.
Keywords:
Remote sensing
Songgang Fault
deep learning
landslide recognition
geohazard evolution
Journal
IF:
2.6
Papers:
1.2W
Citations:
2.7W
Organization
Cited Papers