This repository contains a workflow for the automatic ranking of landslide candidate areas at a regional scale using EGMS InSAR data for territorial planning and risk management.
A semi-automated methodology for identifying and ranking landslide candidate areas at regional scale using European Ground Motion Service (EGMS) InSAR data for territorial planning and risk management.
This repository contains Python scripts implementing the methodology described in the research paper for automatic ranking of landslide candidate areas using EGMS InSAR data. The approach transforms large-scale interferometric data into actionable insights for regional authorities responsible for landslide risk mitigation and territorial planning.
- Automated AOI Detection: Semi-automatic identification of Areas of Interest (AOIs) based on velocity anomalies from A-DInSAR analysis
- Multi-criteria Ranking System: Hazard and risk-oriented ranking combining intensity metrics and exposure factors
- Regional Scale Analysis: Capable of processing extensive areas with thousands of measurement points
- Integration Ready: Compatible with existing Italian landslide inventories (IFFI, PAI) and infrastructure databases
The implemented workflow consists of three main components:
- Selection of Landslide Candidates: Identification of AOIs based on Persistent Scatterer distribution from A-DInSAR analysis
- Intensity and Exposure Ranking: Hierarchical classification evaluating both hazard-oriented (velocity, area) and risk-oriented (buildings, infrastructure) criteria
- Validation Framework: Ground-truth validation through field surveys and geomorphological analysis
- EGMS Basic Products: Sentinel-1 time series and LOS velocity maps (ascending/descending)
- Digital Elevation Model: High-resolution DEM (10m resolution recommended)
- Landslide Inventories: IFFI Project maps or PAI Hazard Maps
- Infrastructure Data: Vector maps of roads, railways, and urbanized areas
- Anthropogenic Features: Quarries, mines, landfills, and industrial plants
The methodology was validated on the Lazio Region (Central Italy), identifying 4,811 unique Areas of Interest with 91% field validation success rate. The analysis revealed significant gaps in existing landslide inventories, with 68% of identified active areas not previously mapped.
- Velocity Threshold: ±2.5 mm/year for anomaly detection
- Interpolation Method: Inverse Distance Weighting (IDW) with 50m radius
- Minimum PS Requirements: 6 persistent scatterers within interpolation radius
- Clustering Distance: Maximum 200m between related deformation areas
- Slope Filtering: Areas with >5° slope inclination
If you use this methodology or code in your research, please cite:
@article{marmoni2024,
title={Automatic ranking of landslide candidate areas at a regional scale using EGMS InSAR data for territorial planning and risk management},
author={Marmoni, Gian Marco and Antonielli, Benedetta and Caprari, Patrizia and Di Renzo, Maria Elena and Marini, Roberta and Mastrantoni, Giandomenico and Mazzanti, Paolo and Patelli, Davide and Bozzano, Francesca},
journal={[Landslide]},
year={2025},
note={Corresponding author: gianmarco.marmoni@uniroma1.it}
}Corresponding Author:
- Gian Marco Marmoni (gianmarco.marmoni@uniroma1.it)
Contributors:
- Benedetta Antonielli¹
- Patrizia Caprari¹
- Maria Elena Di Renzo¹
- Roberta Marini²
- Giandomenico Mastrantoni¹
- Paolo Mazzanti¹,²
- Davide Patelli¹
- Francesca Bozzano¹
Affiliations:
- Earth Science Department of Sapienza University of Rome and CERI Research Centre for Geological Risks, P.le Aldo Moro 5, Rome, Italy
- NHAZCA S.r.l., Via Vittorio Bachelet, Rome, Italy
This study was conducted within the framework of:
- Institutional agreement between CERI Research Centre for Geological Risks and Lazio Region
- RETURN Extended Partnership (European Union Next-GenerationEU, NRRP Mission 4, Component 2, Investment 1.3)
- ReLUIS 2022-2024 & 2024-2026 Projects
EGMS, Automatic selection, landslide candidates, intensity ranking, interferometry, landslide risk, A-DInSAR, territorial planning, risk management
For questions or issues, please contact the corresponding author or open an issue in this repository.