A Multi-source Fusion Landslide Inventory and Susceptibility Assessment Dataset for Gilgit-Baltistan, Pakistan

Read full text Summarise with VIRAT
Authors: Yaopeng, Xiong, Zelang Miao, Dai Huayan, Karamat Ali, Sayyid Amer Mahmoud, WU Lixin
Cited by: 0

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

As a pivotal section of the northern China-Pakistan Economic Corridor (CPEC), the Gilgit-Baltistan (GB) region exhibits complex conditions for landslide development due to frequent tectonic activities and significant topographic relief. However, restricted accessibility in high-altitude canyon areas has led to significant spatial sampling bias in existing landslide inventories, thereby compromising the reliability of regional disaster risk assessments. To address these issues, this study constructed a multi-source landslide inventory and a high-precision landslide susceptibility assessment dataset covering the entire GB region. The study employed a multi-scale segmentation method to delineate geomorphologically meaningful slope units. It systematically integrated historical public records, field survey data, and landslide points identified via Intermittent Small Baseline Subset (IStacking) time-series deformation analysis and verified by high-resolution optical imagery, constructing a spatiotemporally complementary multi-source landslide inventory system covering the period from 1970 to 2024. On this basis, the CatBoost algorithm, which possesses superior capabilities in handling categorical features, was selected to conduct landslide susceptibility modeling combined with a multi-source data fusion strategy, generating spatial distribution results of landslide susceptibility in the study area. This dataset contains 11 landslide conditioning factors, multi-source landslide inventory data, and corresponding landslide susceptibility assessment results, which to a certain extent mitigates the issues of missed detection and spatial sampling bias common in single-source data for high-altitude mountainous areas. The related results can provide a reliable data foundation and technical support for infrastructure route selection, disaster prevention and mitigation planning, and geological disaster risk management along the China-Pakistan Economic Corridor.

Fields of Study & Subjects

Landslides and related hazards Tree Root and Stability Studies Soil erosion and sediment transport

Publication details

Journal / source
ScienceDB
Publication date
1 March 2026
Language
ZH
Document type
Dataset