Scholarlink

Abstract <p>Landslides inflict serious damage to humans, the environment, and socio-economics in the Indian Himalayas. Infrastructure development and public welfare projects such as dams, reservoirs, tunnels, bridges, housing, water storage facilities, etc., along with the adverse terrain conditions have alarmingly increased landslide incidences in the Himalayas. Landslide susceptibility zonation (LSZ) assessment is an important preliminary step towards alertness for landslide hazards. This study was undertaken with the objective of assessing landslide susceptibility conditions using thirteen causal factors for the slopes bounding the Tehri reservoir. Geospatial tools, field observations, and ancillary data were used to extract thematic layers used to assess susceptibility. A novel semi-quantitative method based on landslide frequency ratio (FR) was evolved to weight and rate the factors as well as their classes. A scaling method has been developed to assign weights to the factors and ratings for attributes to remove inherent biases induced in susceptibility assessment by other methods. Factors and associated attributes were arithmetically overlaid in a GIS environment to produce the landslide susceptibility index (LSI) map. In order to establish the robustness of the proposed method,&#xa0;a fuzzy logic approach was used to generate an LSI map and reclassified into five relative susceptibility zones. Comparison between the proposed and fuzzy logic methods using the area under the curve (AUC) showed 77.3% and 80.7% accuracy, respectively. The present study provided insights into the capability of a semi-quantitative model in predicting landslide-susceptible areas and successfully assigned weights/ratings to factors as well as their classes. It also eliminates the inherent bias in landslide susceptibility models that are based on training and testing data.</p> Research Highlights <p>Introduces a hybrid, semi-quantitative framework that combines the objectivity of data-driven frequency ratio (FR) analysis with a simplified, interpretable ranking and weighting scheme.</p> <p>Bridges gap between data-driven and expert-based LSZ methods with a practical, transparent approach.</p> <p>Identifies reservoir induced landslide prone areas with reasonable accuracy in highly fragile terrain of Lesser Himalaya.</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

A novel approach for identification of landslide susceptibility zones – a case study from Uttarakhand region

  • 期刊论文

  • 发表日期：

    2026年3月2日

查看全文
  • Rohan Kumar,
  • K Khusulio,
  • Ankita Agarwal

摘要

Abstract

Landslides inflict serious damage to humans, the environment, and socio-economics in the Indian Himalayas. Infrastructure development and public welfare projects such as dams, reservoirs, tunnels, bridges, housing, water storage facilities, etc., along with the adverse terrain conditions have alarmingly increased landslide incidences in the Himalayas. Landslide susceptibility zonation (LSZ) assessment is an important preliminary step towards alertness for landslide hazards. This study was undertaken with the objective of assessing landslide susceptibility conditions using thirteen causal factors for the slopes bounding the Tehri reservoir. Geospatial tools, field observations, and ancillary data were used to extract thematic layers used to assess susceptibility. A novel semi-quantitative method based on landslide frequency ratio (FR) was evolved to weight and rate the factors as well as their classes. A scaling method has been developed to assign weights to the factors and ratings for attributes to remove inherent biases induced in susceptibility assessment by other methods. Factors and associated attributes were arithmetically overlaid in a GIS environment to produce the landslide susceptibility index (LSI) map. In order to establish the robustness of the proposed method, a fuzzy logic approach was used to generate an LSI map and reclassified into five relative susceptibility zones. Comparison between the proposed and fuzzy logic methods using the area under the curve (AUC) showed 77.3% and 80.7% accuracy, respectively. The present study provided insights into the capability of a semi-quantitative model in predicting landslide-susceptible areas and successfully assigned weights/ratings to factors as well as their classes. It also eliminates the inherent bias in landslide susceptibility models that are based on training and testing data.

Research Highlights

Introduces a hybrid, semi-quantitative framework that combines the objectivity of data-driven frequency ratio (FR) analysis with a simplified, interpretable ranking and weighting scheme.

Bridges gap between data-driven and expert-based LSZ methods with a practical, transparent approach.

Identifies reservoir induced landslide prone areas with reasonable accuracy in highly fragile terrain of Lesser Himalaya.