Multi-Sensor Framework for Continuous Flood Nowcasting in Pakistan
A new remote sensing system is being developed to enhance flood area mapping in Pakistan, aiming for nearly real-time results. This innovative tool incorporates data from multiple satellites, such as Sentinel-1 and Harmonized Landsat-Sentinel, alongside daily updates from MODIS and VIIRS, all processed through Google Earth Engine. The initiative will be operational for the upcoming monsoon season in 2025, which runs from June to December. This system aims to overcome the shortcomings of existing technologies like UNOSAT by offering continuous flood mapping, ultimately improving the effectiveness of disaster response efforts in affected regions.
Key facts
- Framework integrates Sentinel-1 SAR, HLS L30, HLS S30, MODIS, and VIIRS
- Uses Google Earth Engine for harmonized grid processing
- Tiered ensemble prioritizes higher-resolution sensors
- Applied to Pakistan's 2025 monsoon season (June–December)
- Addresses gaps in existing operational flood products like UNOSAT
- Generates near-real-time, continuous flood extent maps
- Preserves daily continuity at each sensor's native resolution
- Focuses on Pakistan's severe flood season in 2025
Entities
Institutions
- UNOSAT
- Google Earth Engine
- arXiv
Locations
- Pakistan
- South Asia