Scholarlink

Background <p>Landslides pose a persistent threat in Indonesia, causing significant damage and loss of life. The Pesanggaran District in Banyuwangi Regency East Java, Indonesia is particularly susceptible to this geological hazard due to its complex terrain and environmental conditions.</p> Purpose <p>This study aims to quantify landslide susceptibility in the district using an integrated bivariate modeling framework that combines the Weight of Evidence (WoE) and Frequency Ratio (FR) approaches with remote sensing-GIS techniques.</p> Method <p>Landslide inventory data were derived through NDVI anomaly detection using Google Earth Engine and validated with field surveys. Eight conditioning factors, including slope, elevation, aspect, lithology, land use, rainfall, distance to rivers, and distance to roads were analyzed to generate landslide susceptibility maps classified into low, moderate, and high zones. Model performance was evaluated using Receiver Operating Characteristic (ROC) curves produced Area Under the Curve (AUC).</p> Results <p>The models demonstrated good predictive performance, with AUC values of 0.700 for WoE and 0.732 for FR. Both models consistently identified slope, lithology, and rainfall as the most influential factors controlling landslide occurrence. WoE results indicate Pesanggaran is dominated by moderate susceptibility at 44% (194.13 km²), while FR indicates dominance of high susceptibility at 42% (185.6 km²). High-susceptibility zones are mainly concentrated in the northern mountainous sector, including areas influenced by steep relief and intensive land-use pressures such as mining-related slope modifications.</p> Conclusion <p>The WoE-FR-based susceptibility assessment provides robust spatial evidence for land-use planning, development control in hazard-prone zones, and strengthened disaster risk reduction strategies to support sustainable regional management.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Landslide susceptibility assessment through bivariate models (weight of evidence and frequency ratio) in Pesanggaran, East Java, Indonesia

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年12月24日

查看全文
  • Syamsul Bachri,
  • Rajendra P. Shrestha,
  • Sugeng Utaya,
  • Sumarmi Sumarmi,
  • Mellinia Prastiwi,
  • Nanda Putri,
  • A. Riyan Hakiki,
  • Tabita Hidiyah

摘要

Background

Landslides pose a persistent threat in Indonesia, causing significant damage and loss of life. The Pesanggaran District in Banyuwangi Regency East Java, Indonesia is particularly susceptible to this geological hazard due to its complex terrain and environmental conditions.

Purpose

This study aims to quantify landslide susceptibility in the district using an integrated bivariate modeling framework that combines the Weight of Evidence (WoE) and Frequency Ratio (FR) approaches with remote sensing-GIS techniques.

Method

Landslide inventory data were derived through NDVI anomaly detection using Google Earth Engine and validated with field surveys. Eight conditioning factors, including slope, elevation, aspect, lithology, land use, rainfall, distance to rivers, and distance to roads were analyzed to generate landslide susceptibility maps classified into low, moderate, and high zones. Model performance was evaluated using Receiver Operating Characteristic (ROC) curves produced Area Under the Curve (AUC).

Results

The models demonstrated good predictive performance, with AUC values of 0.700 for WoE and 0.732 for FR. Both models consistently identified slope, lithology, and rainfall as the most influential factors controlling landslide occurrence. WoE results indicate Pesanggaran is dominated by moderate susceptibility at 44% (194.13 km²), while FR indicates dominance of high susceptibility at 42% (185.6 km²). High-susceptibility zones are mainly concentrated in the northern mountainous sector, including areas influenced by steep relief and intensive land-use pressures such as mining-related slope modifications.

Conclusion

The WoE-FR-based susceptibility assessment provides robust spatial evidence for land-use planning, development control in hazard-prone zones, and strengthened disaster risk reduction strategies to support sustainable regional management.