Project snapshot

1. Objective: Produce near-real-time flood-extent maps and 72-hour flood-forecast layers for the LMB to support disaster-risk reduction and emergency response.
2. Key deliverables:

• 10 m-resolution seasonal flood-frequency layers (2015-2023)
• Web-based interactive early-warning dashboard updated every 6 hours during monsoon season
• Open-data API for national hydro-meteorological agencies

1. Methodology highlights:

• Sentinel-1 SAR backscatter thresholding for cloud-penetrating flood detection.
• Machine-learning (Random-Forest) classification of Sentinel-2 imagery to refine urban vs. agricultural inundation zones.
• Coupling of HEC-RAS hydraulic model outputs with GIS layers to forecast extent and depth.