Persistent Scatterer–Soil Moisture Analysis for Risk & Triggering
PS-SMaRT v1.0 is an open-source Python model for the automatic detection and clustering of unstable slopes by fusing Persistent Scatterer Interferometric SAR (PS-InSAR) time-series deformation with soil-moisture anomalies and terrain indices. It provides both a scripted backend and a graphical interface for landslide-hazard assessment, developed for publication in Geoscientific Model Development (GMD).
- LOS → downslope projection of PS-InSAR velocities using incidence, heading, slope, and aspect.
- Automated filtering by slope and velocity thresholds.
- DBSCAN clustering to delineate coherent unstable zones.
- Cluster statistics: mean, variance, extrema, and polygon geometry.
- Soil-moisture anomaly overlap with χ² and Matthews correlation (φ).
- Topographic Wetness Index (TWI) analysis via Welch’s t-test.
- Hazard index generation through robust percentile scaling (P5–P95) and blending into
Low / Moderate / High classes. - Streamlit interface for interactive runs, visualization, and output management.
# 1. Clone the repository
git clone https://github.com/<yourusername>/ps-smart-v1.0.git
cd ps-smart
# 2. Create environment and install dependencies
pip install -r requirements.txt
# or
conda env create -f environment.yml
conda activate ps-smart
## Usage
##Command-line
python unstable_slope_hazard_pipeline.py --help
##Streamlit Application
streamlit run app.py