Global Flood Detections from Satellite Radar 2014-2024

Global flood detections derived from Sentinel-1 Synthetic Aperture Radar imagery using a deep learning change detection model. The dataset spans October 2014 to September 2024, offering a longitudinal view of flood-prone areas worldwide and accompanies a 2025 Nature Communications paper.

Use Cases

• Train deep learning models for flood detection using cloud-penetrating Synthetic Aperture Radar data.
• Analyze longitudinal trends in global flood-prone areas over the 2014-2024 time range.
• Validate and benchmark flood mapping algorithms against a global-scale detection dataset.

Strengths

• Derived from a decade of Sentinel-1 satellite radar data, covering October 2014 to September 2024.
• Provides a global spatial coverage for analyzing flood events.
• Accompanies a peer-reviewed publication in Nature Communications (2025).

Limitations

• Specific data volume, row count, and column structure are unknown from the provided input.
• The methodology relies on a specific deep learning change detection model, which may have inherent biases or limitations.
• The dataset's update frequency and long-term maintenance plan are not specified.

Provenance

Source

ai-for-good-lab

Collection Method

Derived from Sentinel-1 Synthetic Aperture Radar imagery using a deep learning change detection model.

Time Range

October 2014 – September 2024

Freshness

Last updated on 2025-10-28.

Geography

Global