TerraMosaic: A Community-Curated, Open-Access Global Landslide Platform

https://www.terramosaic.org/

Recent advances in machine learning (ML) and deep learning (DL) have significantly advanced landslide-related applications, including detection, early warning, and susceptibility mapping. Generative AI technologies further offer new opportunities to accelerate landslide research, particularly by supporting rapid prototyping and iteration of ML/DL workflows.

However, most existing models are trained on datasets specific to certain regions or landslide types, resulting in poor or untested generalization across different geographic and environmental settings. While there have been growing efforts to publish open-access landslide datasets, these resources remain fragmented across individual publications, institutional repositories, and project-specific websites. Researchers often spend substantial time locating, retrieving, and preparing data when building models for new regions. Progress remains constrained by the lack of high-quality, high-volume, standardized, and accessible datasets.

To address this gap, we present a community-curated, open-access, evolving platform that aggregates global landslide inventories and related geospatial data and provides detailed metadata for each dataset. Metadata fields include inventory type (e.g., point, polygon), record count, spatial resolution, geographic coverage, input features (e.g., optical imagery, elevation, land use), ML/DL models used, evaluation settings, and whether cross-regional generalization was tested. Users can search, filter, and download datasets through an interactive, map-based interface. The platform also encourages community contributions via an easy-to-use upload interface. It serves as a central hub for high-quality, globally sourced landslide and geospatial data, supporting the development and benchmarking of reliable, scalable, and generalizable AI models for both fundamental research and real-world applications.