Landslide susceptibility mapping based on data mining models in Lesser Caucasus and Kura foreland basin (Armenia and Azerbaijan)

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F61988987%3A17310%2F25%3AA2603D1S" target="_blank" >RIV/61988987:17310/25:A2603D1S - isvavai.cz</a>

Výsledek na webu

<a href="https://www.tandfonline.com/doi/full/10.1080/19475705.2025.2537221" target="_blank" >https://www.tandfonline.com/doi/full/10.1080/19475705.2025.2537221</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1080/19475705.2025.2537221" target="_blank" >10.1080/19475705.2025.2537221</a>

Jazyk výsledku

angličtina

Název v původním jazyce

Landslide susceptibility mapping based on data mining models in Lesser Caucasus and Kura foreland basin (Armenia and Azerbaijan)

Popis výsledku v původním jazyce

Landslides are a major geological hazard causing significant loss of life and infrastructure damage worldwide. Landslide susceptibility mapping is a crucial, though developing, tool for understanding the spatial distribution of landslide hazard. This study addresses the absence of a comprehensive landslide inventory, limited understandingof causative factors and the lack of regional-scale susceptibility maps for the LesserCaucasus and Kura Basin (LC-KB). A landslide inventory was created for the LesserCaucasus of Azerbaijan and compiled with other inventories, documenting 3,659 landslidepolygons. Sixteen causative factors were analysed, and multicollinearity tests confirmedno significant correlations. Three Machine Learning (ML) models—LogisticRegression (LGR), Support Vector Machine (SVM) and Extreme Gradient Boosting(XGBoost)—were fine-tuned to create landslide susceptibility maps. Slope is consistentlythe most influential factor across all models. Results suggest stronger influenceof seismic factors than climatic ones. XGBoost achieves the highest accuracy (0.81) onthe testing data set, followed by SVM (0.80) and LGR (0.73). The first two modelsshow strong validation performance, with AUC values of 0.89 and 0.87, respectively,while LGR shows a lower AUC of 0.78. The results are vital for planning and disastermanagement, highlighting areas needing urgent mitigation.

Název v anglickém jazyce

Landslide susceptibility mapping based on data mining models in Lesser Caucasus and Kura foreland basin (Armenia and Azerbaijan)

Popis výsledku anglicky

Landslides are a major geological hazard causing significant loss of life and infrastructure damage worldwide. Landslide susceptibility mapping is a crucial, though developing, tool for understanding the spatial distribution of landslide hazard. This study addresses the absence of a comprehensive landslide inventory, limited understandingof causative factors and the lack of regional-scale susceptibility maps for the LesserCaucasus and Kura Basin (LC-KB). A landslide inventory was created for the LesserCaucasus of Azerbaijan and compiled with other inventories, documenting 3,659 landslidepolygons. Sixteen causative factors were analysed, and multicollinearity tests confirmedno significant correlations. Three Machine Learning (ML) models—LogisticRegression (LGR), Support Vector Machine (SVM) and Extreme Gradient Boosting(XGBoost)—were fine-tuned to create landslide susceptibility maps. Slope is consistentlythe most influential factor across all models. Results suggest stronger influenceof seismic factors than climatic ones. XGBoost achieves the highest accuracy (0.81) onthe testing data set, followed by SVM (0.80) and LGR (0.73). The first two modelsshow strong validation performance, with AUC values of 0.89 and 0.87, respectively,while LGR shows a lower AUC of 0.78. The results are vital for planning and disastermanagement, highlighting areas needing urgent mitigation.

Druh

J_imp - Článek v periodiku v databázi Web of Science

CEP obor

—

OECD FORD obor

10508 - Physical geography

Projekt

—

Návaznosti

S - Specificky vyzkum na vysokych skolach

Rok uplatnění

2025

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název periodika

Geomatics, Natural Hazards and Risk

ISSN

1947-5705

e-ISSN

1947-5713

Svazek periodika

—

Číslo periodika v rámci svazku

1

Stát vydavatele periodika

GB - Spojené království Velké Británie a Severního Irska

Počet stran výsledku

27

Strana od-do

—

Kód UT WoS článku

001550860600001

EID výsledku v databázi Scopus

2-s2.0-105013463104