EUVIMEDEUVIMED entdecken
Crossref

Assessing Landslide Susceptibility in the Muthirapuzha Watershed, Idukki: A Random Forest Approach

Vara Laxmi Chitikela N., Padma Kumari Kollipara

Disaster Advances · 30.04.2026 · 6 (19) · 93

DOI: 10.25303/196da0930100 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Disaster Advances, online erschienen am 30.04.2026. Verfasst von Vara Laxmi Chitikela N. und 1 weiteren Autor:innen.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Landslides are a recurring hazard in the highland areas of Kerala, particularly within the Muthirapuzha watershed in Idukki district. Understanding the spatial distribution of landslide-prone zones is essential for effective land management and disaster mitigation. This study applies the Random Forest algorithm, a robust machine learning technique, to model landslide susceptibility in the watershed. A comprehensive database was developed by integrating historical landslide records with a range of conditioning factors such as elevation, slope, aspect, distance to road (DTR), rainfall, lithology, land use/landcover (LU/LC), normalized difference vegetation index (NDVI), distance to lineament (DTL), distance to drainage (DTD), topographic wetness index (TWI) and geomorphology. To ensure model robustness and minimize redundancy among predictors, multicollinearity was evaluated using Variance Inflation Factor (VIF) and all variables recorded VIF values below 5, confirming the absence of significant multicollinearity. The VIF values ranged from 1.044 (TWI) to 2.135 (elevation), indicating that all factors contributed independently to the model. The landslide susceptibility map generated from the model was classified into five classes: Very Low, Low, Moderate, High and Very High. The spatial distribution of susceptibility revealed 20% under high and 8% under very high susceptibility zones. Model performance was evaluated using the receiver operating characteristic (ROC) curve. The area under the curve (AUC) values were found to be 0.925 for the success rate and 0.927 for the prediction rate, indicating excellent predictive capability of the model. The results highlight critical areas vulnerable to future landslides and offer valuable input for local authorities in planning and disaster risk reduction (DRR) initiatives.

Zitieren

APA

N., V. L. C., & Kollipara, P. K. (2026). Assessing Landslide Susceptibility in the Muthirapuzha Watershed, Idukki: A Random Forest Approach. Disaster Advances, 6(19), 93. https://doi.org/10.25303/196da0930100

Vancouver

N. VLC, Kollipara PK. Assessing Landslide Susceptibility in the Muthirapuzha Watershed, Idukki: A Random Forest Approach. Disaster Advances. 2026;6(19):93. doi:10.25303/196da0930100

BibTeX
@article{N2026,
  author = {N., Vara Laxmi Chitikela and Kollipara, Padma Kumari},
  title = {Assessing Landslide Susceptibility in the Muthirapuzha Watershed, Idukki: A Random Forest Approach},
  journal = {Disaster Advances},
  year = {2026},
  volume = {6},
  number = {19},
  pages = {93},
  doi = {10.25303/196da0930100}
}
Kostenlos starten →