Landslide Susceptibility Mapping Using Machine Learning Algorithm Validated by Persistent Scatterer In-SAR Technique

Area of Science:

• Geosciences and Remote Sensing
• Geological Hazard Assessment

Background:

• Landslides pose significant catastrophic risks in mountainous regions, particularly impacting critical infrastructure like the Karakorum Highway (KKH).
• Accurate landslide susceptibility mapping (LSM) is crucial for mitigating risks and ensuring the safety of transportation networks in vulnerable areas.

Purpose of the Study:

• To identify and map landslide susceptibility along the Karakorum Highway in Northern Pakistan.
• To compare the predictive performance of multiple machine learning models for landslide susceptibility.
• To integrate Persistent Scatterer Interferometry (PS-InSAR) data for enhanced landslide hazard assessment.

Main Methods:

• Development of a landslide inventory map from 332 identified landslide locations along the KKH.
• Application and comparison of four machine learning models: Random Forest (RF), Extreme Gradient Boosting (XGBoost), K-Nearest Neighbor (KNN), and Naive Bayes (NB).
• Utilized thirteen landslide conditioning factors for susceptibility mapping and Receiver Operating Characteristic (ROC) curves with Area Under Curve (AUC) for accuracy assessment.
• Employed Persistent Scatterer Interferometry (PS-InSAR) technology to analyze slope deformation velocity.

Main Results:

• The Random Forest (RF) model achieved the highest accuracy (83.08%) in landslide susceptibility mapping, followed by XGBoost (82.15%), KNN (80.31%), and NB (72.92%).
• PS-InSAR analysis revealed significant deformation velocities in areas identified as sensitive by the susceptibility models.
• Integration of RF model results with PS-InSAR data produced an improved landslide susceptibility map for the study region.

Conclusions:

• The Random Forest model, enhanced by PS-InSAR data, provides a robust tool for landslide susceptibility mapping along the Karakorum Highway.
• The developed susceptibility map and methodology can significantly aid in mitigating landslide hazards and ensuring the safe operation of the KKH.
• This integrated approach offers valuable insights for hazard management in similar high-risk mountainous transportation corridors.