A Novel Rapid Inundation Forecasting Technique: Forecasting Inundation Extents using Rotated empirical orthogonal function analysis (FIER) and Its Economic Benefits to Lower Mekong
| Meeting | VIWC2026 |
| Topic | OS2: Remote Sensing of water and inundation |
| Author | Chi-Hung Chang | Hyongki LEE | Son K Do | Tien Du | Faisal Hossain | Kel Markert | Thanapon Piman | Chinaporn Meechaiya |
| Organization | Department of Civil and Environmental Engineering, University of Houston, Houston, TX, USA | Department of Civil and Environmental Engineering, University of Houston, Houston, TX, USA | Department of Civil and Environmental Engineering, University of Houston, Houston, TX, USA | Department of Civil and Environmental Engineering, University of Houston, Houston, TX, USA | Department of Civil and Environmental Engineering, University of Washington, Seattle, WA, USA | Department of Civil and Construction, Brigham Young University, Provo, UT, USA / SERVIR Science Coordination Office, NASA Marshall Space Flight Center, Huntsville, AL, USA | Stockholm Environment Institute, Bangkok, Thailand/ SERVIR-Mekong, Bangkok, Thailand | Asian Disaster Preparedness Center, Bangkok, Thailand / SERVIR-Mekong, Bangkok, Thailand |
| DOI | 2022.1672421948 |
Abstract
We introduce a novel inundation forecasting system named \"Forecasting Inundation Extents using Rotational empirical orthogonal function analysis (FIER).\" The system is data-driven that forecasts the inundation extents based on the correlation between historical inundations observed from satellite images and river levels (or streamflow) and is computationally efficient, yet skillful, and highly scalable. Here, we present daily forecasted inundation extents over the Lower Mekong floodplains up to 18-day lead times, using forecasted river levels available from our in-house VIC-altimetry-based forecasting system. The results are validated using observed inundation extents from available satellite images. We also assess the economic benefits that our FIER-Mekong inundation forecasting system can contribute by delineating mature rice paddies in Lower Mekong which could have been saved from flood damages during the wet season harvest time by taking early actions, such as advising farmers when and where to reap the rice crops, using forecasted inundation depths derived from our forecasted inundation extents and a terrain model. Finally, we present a web application tool of FIER-Mekong (https://share.streamlit.io/skd862/fier_mekong/main/demo.py) which can facilitate its use in decision-making by end-users with hindcasted, nowcasted, and forecasted inundation extents, and inundation depths.
Keywords: Mekong River, Remote Sensing, Inundation Forecast, Flood Damage Prevention
