ibm-esa-geospatial/ImpactMesh-Flood

  

ImpactMesh-Flood

ImpactMesh is a large-scale multimodal, multitemporal dataset for flood and wildfire mapping, released by IBM, DLR, and the ESA Φ-lab. It integrates Sentinel-1 SAR, Sentinel-2 optical, Copernicus DEM, and high-quality annotations from Copernicus EMS. The technical report is released soon. You find the wildfire subset here: https://huggingface.co/datasets/ibm-esa-geospatial/ImpactMesh-Fire.

Features

• —Multimodal: SAR, optical, DEM
• —Multitemporal: Four time steps (pre-month, pre-event, event, post-event)
• —Global coverage: 200+ flood events
• —Scale: 24K samples
• —License: CC-BY 4.0

Modalities

Splits

Covering 206 flood events.

test_holdout (562 samples) is a subset of test, listing the samples whose event does not appear in train or val. Use it to evaluate generalisation to entirely unseen events.

Versions

The current release is v1, which we generally recommend using.

Compared to the initial v0, we performed extensive quality cleaning to improve training and reduce label noise. This removed a substantial number of noisy samples, so v1 is smaller but of higher quality. The v0 version is still available on request, please open a discussion in the Community tab.

Quick Start

Download the dataset:

hf download ibm-esa-geospatial/ImpactMesh-Flood --repo-type dataset --local-dir data/ImpactMesh-Flood

# Only download a single modality (e.g., S2L2A)
hf download ibm-esa-geospatial/ImpactMesh-Flood --repo-type dataset --include "*/S2L2A.tar" --local-dir data/ImpactMesh-Flood

# Only download a single split (e.g., validation)
hf download ibm-esa-geospatial/ImpactMesh-Flood --repo-type dataset --include "val/*" --local-dir data/ImpactMesh-Flood

Untar the samples:

mkdir data/ImpactMesh-Flood/data
for f in data/ImpactMesh-Flood/*/*.tar; do
  echo "Extracting $f"
  tar -xf "$f" -C data/ImpactMesh-Flood/data
done

The samples from all splits are saved in shared folders data/ImpactMesh-Flood/data/{modality}. After extracting, you can delete the tars:

rm -r data/ImpactMesh-Flood/train
rm -r data/ImpactMesh-Flood/val
rm -r data/ImpactMesh-Flood/test

We use TerraTorch for the model fine-tuning and provide data modules for ImpactMesh. You can download the code and configs for the fine-tuning from https://github.com/IBM/ImpactMesh.

Alternatively, you can install the data loading code with:

pip install impactmesh

terratorch fit --config configs/terramind_v1_tiny_impactmesh_flood.yaml

Citation

Our technical report is released soon!

Acknowledgement

ImpactMesh was developed as part of the FAST‑EO project funded by the European Space Agency Φ‑Lab (contract #4000143501/23/I‑DT).

Sentinel-2 Level-2A data were downloaded from Microsoft Planetary Computer and are provided under Copernicus Sentinel license conditions (© European Union 2015–2025, ESA) (https://planetarycomputer.microsoft.com/dataset/sentinel-2-l2a).

Sentinel-1 Radiometrically Terrain Corrected (RTC) SAR data were retrieved from Microsoft Planetary Computer (calibrated to GRD and terrain-corrected using PlanetDEM) under Copernicus Sentinel license terms (© European Union 2014–2025) (https://planetarycomputer.microsoft.com/dataset/sentinel-1-rtc).

The DEM data is produced using Copernicus WorldDEM-30 © DLR e.V. 2010-2014 and © Airbus Defence and Space GmbH 2014-2018 provided under COPERNICUS by the European Union and ESA; all rights reserved.

Annotations were sourced from the Copernicus Emergency Management Service (© European Union, 2012–2025), available at https://emergency.copernicus.eu/.