Overview
Southeast Asia experiences severe monsoon flooding each year, affecting millions of people across countries like Vietnam, Thailand, Cambodia, and Myanmar. These floods cause significant damage to infrastructure, agriculture, and communities. Rapid and accurate flood mapping is essential for effective emergency response, humanitarian aid distribution, and damage assessment.
Satellite Used
Sentinel-1
Time Period
2021-2022 Monsoon Season
Challenge
A coalition of international humanitarian organizations and local disaster management agencies faced several challenges in responding to monsoon flooding in the Mekong Delta region:
- Rapidly mapping flood extent across vast and often inaccessible areas
- Persistent cloud cover during monsoon season limiting the use of optical satellite imagery
- Need for near real-time information to guide emergency response efforts
- Identifying affected communities and critical infrastructure
- Assessing agricultural damage to anticipate food security impacts
- Monitoring flood progression and recession to plan recovery operations
Traditional flood mapping approaches using optical satellite imagery were severely hampered by cloud cover during the monsoon season. Ground-based assessments were limited by accessibility issues and the vast areas affected. A more effective approach was urgently needed to support humanitarian response efforts.
Solution
We implemented a rapid flood mapping system using Sentinel-1 synthetic aperture radar (SAR) imagery, which offers:
- All-weather imaging capability, unaffected by cloud cover
- Day and night observation capability
- Regular 6-12 day revisit frequency
- 20m spatial resolution suitable for regional flood mapping
- Strong contrast between water and land in radar imagery
Key Components of the Solution:
1. Automated Flood Detection
We developed a specialized flood mapping algorithm that:
- Applied change detection between pre-flood and flood-time Sentinel-1 imagery
- Utilized the VV polarization for optimal water detection
- Incorporated terrain information to account for radar shadows
- Applied machine learning classification to distinguish between permanent water bodies and flood waters
- Generated flood extent maps within hours of Sentinel-1 data acquisition
2. Impact Assessment
We integrated flood maps with other geospatial datasets to assess impacts:
- Population data to estimate the number of people affected
- Land use/land cover maps to identify affected agricultural areas
- Infrastructure data to assess impacts on roads, bridges, and critical facilities
- Building footprints to identify affected structures
- Historical flood data to compare with previous events
3. Rapid Dissemination System
We developed a system to quickly deliver actionable information to responders:
- Web-based dashboard showing near real-time flood extent
- Mobile-friendly maps for field teams
- Automated reports for emergency operations centers
- Data exports compatible with common GIS platforms
- Regular updates as new satellite data became available
Sentinel-1 based flood extent map of the Mekong Delta region
Results
People identified in affected areas
Average time from acquisition to map
Mapping accuracy
The implementation of the Sentinel-1 based flood mapping system delivered significant benefits to the humanitarian response:
- Rapid Assessment: Flood maps were typically available within 12 hours of satellite data acquisition, compared to days or weeks with traditional methods.
- Comprehensive Coverage: The system successfully mapped flooding across the entire Mekong Delta region, including areas inaccessible to ground teams.
- High Accuracy: Validation against ground observations and high-resolution imagery showed an overall accuracy of 94% for flood extent mapping.
- Improved Response: Humanitarian organizations used the maps to prioritize aid distribution, reaching an estimated 1.2 million affected people more quickly than would have been possible otherwise.
- Agricultural Impact Assessment: The system identified approximately 820,000 hectares of flooded agricultural land, helping authorities anticipate food security impacts and plan recovery support.
"The Sentinel-1 flood mapping system transformed our response capabilities during the 2021-2022 monsoon floods. For the first time, we had a comprehensive, near real-time view of the situation across the entire affected region, regardless of cloud cover or accessibility issues. This allowed us to direct resources more effectively and reach communities that might otherwise have been overlooked."
Conclusion
This case study demonstrates the critical role of Sentinel-1 SAR imagery in disaster response, particularly for flood events in regions prone to persistent cloud cover. By leveraging the all-weather imaging capability and regular revisit frequency of Sentinel-1, we were able to develop a flood mapping system that provided timely, accurate, and comprehensive information to support humanitarian response efforts.
The success of this project highlights the value of radar remote sensing for disaster management, offering capabilities that complement and extend beyond what is possible with optical satellite imagery or ground-based assessments. The approach developed here has since been adapted for use in other flood-prone regions around the world, helping to enhance disaster preparedness and response capabilities globally.