Abstract
To mitigate the adverse effects of concealed landslide hazards in densely vegetated mountainous regions on the full life cycle of highways, this study leveraged the vegetation-penetrating capability of airborne LiDAR to obtain a high-precision DEM with 0.2m spatial resolution for the under-construction G0321 Qimen to Anhui-Jiangxi Border expressway section (K2+430~K14+750 segment). A remote sensing interpretation methodology applicable to densely vegetated mountainous areas was formulated, a landslide hazard sample database for the study area was constructed, and the accuracy of remote sensing interpretation was validated through manual field verification. By comprehensively considering topographic and geomorphic conditions, meteorological and hydrological factors, and field investigation findings, eight predisposing factors were selected, and the Information Value coupled MaxEnt model (IV-MaxEnt) was employed to perform landslide susceptibility assessment in densely vegetated mountainous regions. The assessment results demonstrate that the empirical AUC value of the Receiver Operating Characteristic (ROC) curve for the evaluation model reaches 0.901, indicating high predictive accuracy, which bears significant engineering implications for the route selection, construction, and operation phases of highway projects.
Full Text
Preamble
Title: Landslide Identification and Susceptibility Assessment Methods for Mountain Highways in Dense Vegetation Areas Using Airborne LiDAR
Authors: ZHU Xiaoqiang¹, CAO Hao², DONG Xiujun¹
Affiliations:
¹State Key Laboratory of Geohazard Prevention and Geoenvironment Protection, Chengdu University of Technology, Chengdu 610059, China
²Anhui Transportation Holding Group Co., Ltd., Hefei 230000, China
Abstract
To mitigate the adverse effects of concealed landslide hazards in densely vegetated mountainous regions on highway life-cycle performance, this study leverages the vegetation-penetrating capability of airborne LiDAR to acquire high-precision Digital Elevation Models (DEMs) with 0.2 m spatial resolution for the under-construction G0321 expressway section from Qimen to the Anhui-Jiangxi border (K2+430–K14+750). A remote sensing interpretation methodology tailored for densely vegetated mountainous terrain was developed to establish a landslide hazard sample database, with interpretation accuracy validated through comprehensive field verification. Integrating topographic, geomorphic, meteorological, and hydrological factors with field survey data, eight hazard evaluation factors were selected to implement an Information Value-Maximum Entropy (IV-MaxEnt) coupled model for landslide susceptibility assessment. The model achieved a Receiver Operating Characteristic (ROC) curve AUC value of 0.901, demonstrating high predictive accuracy and significant engineering implications for highway route selection, construction, and operation.
Keywords: Mountain highways; Airborne LiDAR; Landslide identification; IV-MaxEnt coupled model; Susceptibility assessment