It was the strongest storm to hit Guam in over two decades. Typhoon Mawar hammered the Pacific island with sustained winds over 140 miles per hour and torrential rain, knocking out power to most of the 150,000 residents and causing widespread flooding.
Fortunately, there were no reported fatalities or major injuries on Guam, which is also a critical hub for U.S. forces in the Pacific.
In the days since, the powerful storm moved towards the Philippines, Taiwan and Japan, without making landfall, while the Guam Power Authority worked to restore electricity and cleanup efforts continued.
The typhoon highlighted the perils of living in the Pacific islands, dangers that are expected to become increasingly more frequent due to climate change.
There is a critical need for emergency responders to have detailed and timely information on flooding and wind damage in the immediate aftermath of the storm.
One solution is based on satellite imagery, specifically synthetic aperture radar (SAR), a sensor that works in any weather conditions, day or night.
To demonstrate how SAR data can be utilized to detect flooding, we used the example of Typhoon Mawar.
Our friends at Umbra, a leading SAR provider, collected imagery over Guam and neighboring Rota Island.
We applied a proprietary algorithm, designed to detect the presence of water, to the SAR imagery.
The results were ingested into the analytics platform created by New Light Technologies Inc. (NLT), which assessed the images to detect potential flooding and impacts to vulnerable populations and assets in the islands.
NLT is a leading provider of integrated information technology, consulting, and research services, providing a broad range of integrated cloud, agile software development, cybersecurity, data science, geospatial, and workforce services and ready-to-use solutions for customers, including in the field of disaster management.
Ursa has partnered with NLT to leverage its Incident Management, Prioritization, Alerting and Coordination Tool (IMPACT). IMPACT provides a robust and scalable framework for disaster management alerting, prioritization and resource allocation, by considering the impacts on and needs of vulnerable communities to ensure timely delivery, deployment and allocation of resources to the populations most in need.
IMPACT also allows to identify and prioritize areas that require collection of satellite imagery, such as in the case of Typhoon Mawar. Once imagery is collected, IMPACT allows visualization of the derived insights and exposure analytics to further prioritize allocation of resources to the populations most in need.
The image below shows the flooding results over parts of Guam and Rota Island. We focused on 2 villages – Yigo, Guam and Songsong, Rota – though the same process could easily be scaled to include much larger areas.
Water pixels were extracted and classified separately from existing water bodies to obtain a flood extent map that was used to pinpoint the exact locations where flooding occurred. The results were overlaid on an optical base map.
As of May 28, a few days after the storm battered the islands, flooding was still apparent, though more so on the narrow peninsula of Songsong (Rota) than in Yigo, which is home to Andersen Air Force Base.
According to Dr. Ran Goldblatt, Chief Scientist of New Light Technologies, the collaboration between Ursa and NLT will allow Ursa to leverage the IMPACT platform to rapidly determine the areas most at risk and prioritize collection of satellite imagery in those locations where vulnerable communities may be impacted.
This will allow the rapid collection and delivery of high-resolution imagery – including SAR –to provide a rapid glimpse into the impacts of disastrous events such as demonstrated in the case of Typhoon Mawar. Brooke Hatcher, Senior Geospatial Data Scientist at NLT, adds that this example demonstrates the added value of high-resolution SAR imagery during disasters and the potential use this data to map floods in areas that have traditionally been inaccessible, hence provide essential intelligence to provide relief to the populations most in need.
“We talk a lot about solving real world problems, and this is exactly that,” said Chad Baker, Director of Data Partnerships, Ursa Space. “With NLT, we think of this scenario as a multiplier effect. Their intimate knowledge of vulnerable areas informs where Ursa can strategically task imagery for comprehensive situational awareness.”