CHRS Data Portal

Key Foundational Data Precipitation

Global satellite precipitation data archive from UC Irvine's CHRS. Access PERSIANN system data (1983-present) with visualization and download tools for spatiotemporal precipitation statistics.

Visit Data Portal Visit CHRS

PyLandslide

Code Python Machine Learning

A Python tool for landslide susceptibility mapping and uncertainty analysis.

Visit Repository View Paper

LHASA Model

Code Python Machine Learning

Landslide Hazard Assessment for Situational Awareness (LHASA) Model - A NASA-developed system for near real-time landslide hazard assessment.

Visit Repository View Webpage

QGIS-FORM

Tool Susceptibility Assessment GIS

A QGIS framework for physically-based probabilistic modelling of landslide susceptibility.

Visit Repository View Webpage

Lisem (Lisem Integrated Spatial Earth Modeller)

Tool Physical Model Hydrological Modeling

Free open-source geospatial modeling tool for hydro-meteorological surface hazards. Features scripting environment, data viewer, and Python bindings for automation and integration.

Visit Official Website

LiCSBAS2

Tool Remote Sensing

LiCSBAS is an open-source package in Python and bash to carry out InSAR time series analysis using LiCSAR products (i.e., unwrapped interferograms and coherence) which are freely available on the COMET-LiCS web portal. LiCSBAS2 is the successor of LiCSBAS. Users can easily derive the time series and velocity of the displacement if sufficient LiCSAR products are available in the area of interest. LiCSBAS also contains visualization tools to interactively display the time series of displacement to help investigation and interpretation of the results.

Visit Repository View Paper

SALaD (Semi-Automatic Landslide Detection)

Tool Landslide Detection Machine Learning

Landslide mapping system from NASA. Uses object-based image analysis (OBIA) and a Random Forest classifier on optical imagery and a digital elevation model (DEM). Requires training polygons and runs in a Singularity container on Linux.

Visit Repository

OBIA + Random Forest Landslide Detector

Code Python Machine Learning

Scripts performing k-means segmentation and merging to extract candidate objects from optical imagery, followed by Random Forest classification. Built on Google Earth Engine and Python libraries; replicates a 2019 MSc thesis. Associated publication: Thesis: Landslide Detection using Random Forest Classifier (TU Delft, 2019)

Visit Repository

Sentinel-2 Rapid Mapping Script

Code Python

Python/GEE script for rapid landslide mapping using Sentinel-2. Calculates the Barren Soil Index (BSI), NDWI, and NDVI to highlight barren soil, water, and vegetation, then classifies landslides based on thresholds.

Visit Repository

Attention U-Net on SAR Data

Code Python Machine Learning

Code and sample data enabling training of an Attention U-Net segmentation model on Sentinel-1 SAR amplitude data. Based on data from the 2018 Hokkaido event. Associated publication: Rapid Mapping of Landslides on SAR Data by Attention U-Net (Remote Sensing 2022)

Visit Repository

HR-GLDD Deep Learning Dataset

Code Python Machine Learning

Repository with code and sample data for the HR-GLDD dataset. Enables training segmentation models using generalized DL for rapid landslide mapping on high-resolution satellite imagery. Associated publication: HR-GLDD: A globally distributed dataset using generalized DL for rapid landslide mapping on high-resolution satellite imagery (Remote Sensing 2022)

Visit Repository

SAR-LRA (Sentinel-1 SAR-based Landslide Rapid Assessment)

Code Python Machine Learning

Tool for rapid post-earthquake landslide detection using Sentinel-1 SAR. Implements deep neural networks on Google Earth Engine and local computing; aims for day-night, all-weather detection. Associated publication: Preprint: "Sentinel-1 SAR-based Globally Distributed Landslide Detection by Deep Neural Networks"

Visit Repository

DRIP/SLIP (NASA)

Code Python

Deprecated demonstration package from NASA for automated Landsat-based landslide detection in Nepal. Combines Landsat imagery with rainfall monitoring; depends on GDAL, SciPy etc. Associated publication: Automated Satellite-based Landslide Identification Product for Nepal (Journal of Applied Remote Sensing 2018)

Visit Repository

BRIGHT: a globally distributed multimodal building damage assessment dataset with very-high-resolution for all-weather disaster response

Key Foundational Data Loss and Damage

BRIGHT is the first open-access, globally distributed, event-diverse multimodal dataset that integrates very-high-resolution optical and SAR imagery, enabling truly all-weather, day-and-night building damage assessment. It fills a major gap in high-quality benchmarks by providing 0.3–1 m building-level imagery across 14 regions and seven disaster types, supporting the development of robust AI models for real-world disaster response. The dataset also includes standardized baselines from advanced AI models, offering a consistent foundation for research on transferability, domain adaptation, semi-supervised learning, and multimodal change detection.

Dataset and Code View Paper

USGS National Landslide Damages and Losses

Key Foundational Data Loss and Damage

A comprehensive database of landslide damages and losses across the United States, providing critical information for risk assessment, mitigation strategies, and understanding the economic and social impacts of landslide hazards.

Data Release Visit Webpage