MMFlood: Multimodal Dataset for Flood Delineation from Satellite Imagery

Provider
LINKS Foundation (Montello, Arnaudo, Rossi)
License
CC-BY-4.0 (source)
Size
15 files
Format
JSON (in sampled files)
Stored at
Hugging Face (https://huggingface.co/datasets/links-ads/mmflood)
Access
Free and public. No account with the data provider is needed.
DOI
10.5281/zenodo.6534637
Category
Earth Observation & Geospatial
Source
https://zenodo.org/records/6534637

About MMFlood: Multimodal Dataset for Flood Delineation from Satellite Imagery

Multimodal dataset for flood delineation from Sentinel-1 synthetic aperture radar, built from Copernicus Emergency Management Service (EMS) flood activations worldwide. Each activation area (folders named by EMS code, for example EMSR548-0) is tiled into GeoTIFFs with four aligned modalities: Sentinel-1 VV-VH backscatter (s1_raw), a digital elevation model from MapZen/TileZen, an OpenStreetMap hydrography map where available, and the ground-truth flood mask rasterised from EMS polygons. An activations.json file holds per-activation metadata and the train, validation and test split used in the paper. The listing (measured 2026-11-14) shows the dataset packed as eleven activations.tar.*.gz.part pieces (about 11.2 GB) that are concatenated and extracted before use. Open the tiles with rasterio or xarray. Used for SAR flood segmentation, water mapping and multimodal fusion with elevation and hydrography. The Zenodo record from the authors releases MMFlood under CC-BY-4.0; cite Montello, Arnaudo and Rossi.

How to load this dataset in Python

from huggingface_hub import hf_hub_download
path = hf_hub_download(repo_id="links-ads/mmflood", filename="activations.tar.000.gz.part", repo_type="dataset")
print(path)

Data structure

File types in sampled files: JSON

Files

Example file: activations.tar.000.gz.part

Analyze this dataset without downloading it

On Scigantic, MMFlood: Multimodal Dataset for Flood Delineation from Satellite Imagery opens in a hosted JupyterLab notebook with the data already mounted, so you can read files where they are stored instead of copying them first. Each notebook includes an AI assistant that can see the mounted files and write and run analysis code. Open the dataset page to launch a notebook, or browse the full catalog.

This page is also available as Markdown and as JSON.

Related Earth Observation & Geospatial datasets