Scholarlink

<p>The Western Ghats region of Idukki district in southern India is highly susceptible to rainfall-induced landslides due to steep topography, intense monsoons, and increasing land-use pressures. To address the need for localized landslide early warning system (LEWS), we developed the Slope Instability Predictor–Kerala (SLIP-K), a real-time system that integrates a physics-based landslide susceptibility model (Geographic Information System-Tool for Infinite Slope Stability Analysis (GIS-TISSA)) with empirical rainfall thresholds (RTs) quantified through data from eight automated weather stations (AWS). SLIP-K operates through an interactive web/mobile application, delivering 15-minute interval risk updates and user-friendly alerts to communities using Google Earth Engine-driven geospatial mapping. Beyond inventory-based or statistical models, SLIP-K offers physically interpretable outputs, community participatory reporting, and multilingual support. A unique aspect of this study is the multi-year, quantitative validation of SLIP-K using AWS data and fatal landslide inventories (2021–2024). Confusion matrix analysis across all AWS sites demonstrated high sensitivity (recall = 1.00), robust accuracy (0.91), and successful identification of all recorded fatal events. Additional assessments yielded a red alert success rate of 33.3% and an Area Under the Curve (AUC) of 0.82–0.88, comparable to national and international benchmarks. These findings establish SLIP-K as a transparent, statistically robust, and scalable landslide early warning framework, supporting risk reduction strategies across data-limited and topographically complex mountain environments.</p> Graphical Abstract <p>The main features and operations of the Slope Instability Predictor-Kerala (SLIP-K), a specific landslide early warning system (LEWS) created for the Idukki district, begins with the gathering of data from Automatic Weather Stations (AWS), which is essential for tracking rainfall in real-time. To determine the probability of occurrence of landslides, this rainfall data is sent to a web server and examined against a predetermined rainfall threshold (RT). A landslide susceptibility map in the Google Earth Engine platform gets meld with the rainfall data and changes dynamically, manifested as changes in a series of colour: from green to yellow when RT exceeds 25%, yellow to orange during 50%, and finally, to red when it crosses 75%. These changes are sent as push notifications to the user through SLIP-K mobile app.</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

Slope Instability Predictor-Kerala (SLIP-K): A mobile/web Application for Landslide Hazard Prediction in Idukki, India

  • 期刊论文

  • 发表日期：

    2025年10月9日

查看全文
  • Subhami Mohan,
  • A. Rajaneesh,
  • V. K. Krishnapriya,
  • K. S. Sajinkumar,
  • Cees van Westen,
  • Nikhil N. Vasu,
  • Yunus P. Ali,
  • Lina Hao,
  • C. L. Vishnu,
  • R. S. Ajin

摘要

The Western Ghats region of Idukki district in southern India is highly susceptible to rainfall-induced landslides due to steep topography, intense monsoons, and increasing land-use pressures. To address the need for localized landslide early warning system (LEWS), we developed the Slope Instability Predictor–Kerala (SLIP-K), a real-time system that integrates a physics-based landslide susceptibility model (Geographic Information System-Tool for Infinite Slope Stability Analysis (GIS-TISSA)) with empirical rainfall thresholds (RTs) quantified through data from eight automated weather stations (AWS). SLIP-K operates through an interactive web/mobile application, delivering 15-minute interval risk updates and user-friendly alerts to communities using Google Earth Engine-driven geospatial mapping. Beyond inventory-based or statistical models, SLIP-K offers physically interpretable outputs, community participatory reporting, and multilingual support. A unique aspect of this study is the multi-year, quantitative validation of SLIP-K using AWS data and fatal landslide inventories (2021–2024). Confusion matrix analysis across all AWS sites demonstrated high sensitivity (recall = 1.00), robust accuracy (0.91), and successful identification of all recorded fatal events. Additional assessments yielded a red alert success rate of 33.3% and an Area Under the Curve (AUC) of 0.82–0.88, comparable to national and international benchmarks. These findings establish SLIP-K as a transparent, statistically robust, and scalable landslide early warning framework, supporting risk reduction strategies across data-limited and topographically complex mountain environments.

Graphical Abstract

The main features and operations of the Slope Instability Predictor-Kerala (SLIP-K), a specific landslide early warning system (LEWS) created for the Idukki district, begins with the gathering of data from Automatic Weather Stations (AWS), which is essential for tracking rainfall in real-time. To determine the probability of occurrence of landslides, this rainfall data is sent to a web server and examined against a predetermined rainfall threshold (RT). A landslide susceptibility map in the Google Earth Engine platform gets meld with the rainfall data and changes dynamically, manifested as changes in a series of colour: from green to yellow when RT exceeds 25%, yellow to orange during 50%, and finally, to red when it crosses 75%. These changes are sent as push notifications to the user through SLIP-K mobile app.