S. V. Nghiem
NASA Jet Propulsion Laboratory
California Institute of Technology
Pasadena, California, U.S.A.

Super Typhoon Yagi directly hit Northern Vietnam as one of the most intense ever.  Yagi made landfall over two major coastal cities of Hải Phòng and Quảng Ninh, Vietnam, on 7 September 2024.  Torrential rains carried by Yagi confounded by water release from upstream dams in multiple rivers led to extensive flooding wreaking havoc and inflicting deaths and destructions in many areas. 

As an emergency response, the NASA Land Cover and Land Use Change Program has an urgent initial effort for flood mapping with Sentinel-1 (SN1) satellite synthetic aperture radar (SAR) data using the Depolarization Reduction Algorithm for Global Observations of inundatioN (DRAGON) developed and implemented at the NASA Jet Propulsion Laboratory (Nghiem et al., 2017).  SN1 SAR operates at C-band microwave  frequency, which allows observations of land surface regardless of cloud cover that was pervasive and extensive during Yagi.  The DRAGON inundation digital data products have a resolution of 10 meters across a swath width of 250 km, and results were hydrologically consistent and verified with known surface water from WorldView images and with other flood mapping products from the Global Flood Partnership (Nghiem et al., 2017).  Moreover, a comparison of flood mapping between DRAGON and a different and independent SN1 method at the Dartmouth Flood Observatory (DFO) at the University of Colorado shows consistent results; thereby, increasing the confidence level of the SN1 flood observations.

Figure 1 presents inundated areas on both sides of the Red River detected by the DRAGON method on 12 September 2024 in a southern Hà Nội region together with pre-existing surface water on 31 August 2014.  As of 12 September, there were 330 people dead or missing due to flooding from Typhoon Yagi across many districts and cities in northern Vietnam.  For examples, flooded areas were observed with SN1 SAR data over regions around Chương Mỹ lowland (Figure 2), Đình Chu in the east of the Lô River and along the Phó Đáy River (Figure 3), and Nam Định - Thái Bình about 25 km to the coast of the Gulf of Tonkin (Figure 4).  In addition, flood mapping products were obtained for other regions around Đan Thượng, Hải Dương, Hải Phòng, Tân Quang, Trung Giã, and Việt Long.  The flood maps were produced as soon as possible to provide urgent flood information for immediate use by colleagues in Vietnam. Users also include the U.S. Embassy in Hà Nội, sending a note of appreciation for the flood maps that the Embassy forwarded widely to the United States Agency for International Development (USAID) / Bureau for Humanitarian Assistance (BHA) team who are leading immediate relief response. It should be noted that information about inundated areas, where flooding had occurred on or before 12 September 2024 and might have waned recently, can be useful for damage assessment, for coordination and distribution of emergency relief resources, and for planning to cope with potential infectious diseases and aftermaths in flood affected areas due to Typhoon Yagi.

References

Nghiem, S. V., G. R. Brakenridge, and D. T. Nguyen, Hurricanes Harvey and Irma - High-Resolution Flood Mapping and Monitoring from Sentinel SAR with the Depolarization Reduction Algorithm for Global Observations of InundatioN (DRAGON), Abst. NH23E-2873, American Geophysical Union (AGU) Fall Meeting, San Francisco, Calif., 11-15 Dec. 2017.

Acknowledgments

The research the Jet Propulsion Laboratory (JPL), California Institute of Technology, is supported by the NASA Land Cover and Land Use Change (LCLUC) Program under a contract with NASA (80NM0018D0004).  Thanks to D. Nguyen at JPL for SN1 SAR data processing.  We acknowledge the Copernicus Programme of the European Space Agency for the use of Sentinel-1 SAR data.

Figures

In all figures below, yellow is for flooded areas on 12 Sep 2014 and blue is for pre-existing surface water on 31 Aug 2024 derived from Sentinel-1 SAR data using the JPL Depolarization Reduction Algorithm for Global Observations of inundatioN (DRAGON).  The inundated areas are overlaid with road network on the Google Earth base map.  SRTM Water Bodies data were used to make the known water bodies (circa 2000) transparent to reveal the underlaid features. due to the need for rapid data processing, results in the maps below have not yet fully accounted for topography effects, noise reduction, and isolated pixel misclassification.  The maps are in a 20-m (rather than 10-m) grid due to the large file size.

Figure 1. Inundation map around southern Hà Nội.

Figure 2. Inundation map around Chương Mỹ lowland.

Figure 3. Inundation map around Đình Chu.

Figure 4. Inundation map around Nam Định and Thái Bình.