Kaspar J. Willam Professor Hyongki Lee’s satellite-driven inundation forecasting framework, Forecasting Inundation Extents using REOF (Rotated Empirical Orthogonal Function) (FIER) has received funding from both NASA (for applications in Southeast Asia's Mekong River basin) and NOAA (for applications in the US) as the project seeks to continue expanding into broader use cases around the world.
The framework has been developed with the support of many component projects and partnerships, and multiple funding sources, including the NASA SERVIR program, NOAA JPSS program, and South Korea’s K-Water institute, totaling just under $1.3 million thus far.
“Our framework has been made possible with supports from many colleagues, including Asian Disaster Preparedness Center (ADPC), University of Washington and Brigham Young University, Vietnam’s National Center for Water Resources Planning and Investigation (NAWAPI), SERVIR Coordination Office (SCO), NOAA and GEOGLOWS,” said Lee.
FIER was originally conceived due, in part, to Lee’s own experiences during Hurricane Harvey. It “began as a fluvial flood forecasting tool and has since evolved to forecast compound flooding,” including flood scenarios that combine fluvial, pluvial and storm surge components.
“Even in Houston, which is the fourth largest city in the US, we don’t have an inundation forecasting system. We only have real-time monitoring systems. That night, when the water level was rising in front of my home, there was no information available other than the water gauges and the real-time flood maps — but even those are limited to simulation,” he added.
“After that, I was interested in developing an inundation forecasting system. Traditionally these are developed using the traditional hydrodynamic model, but it requires a lot of input data, and most of the time those datasets are limited. Because my expertise is remote sensing for water applications, I had the idea to extract spatial-temporal flood information from historical satellite imagery data.”
This inclusion of historical data enables the model to simulate future inundation risk more effectively. The model was first implemented in the Mekong River basin in Southeast Asia, where forecasting the extent of inundation can make a big difference, for example, in agricultural activity — particularly in the success of rice crops and the estimation of expected economic impact due to damage.
“We wanted to implement this in the US as well,” Lee continued. “The hydrologic regions of the Red River and Mississippi River basins are a little bit different than the Mekong River basin, which means that without adjustments, we would miss many of the peak flood period signals. Now, we plan to expand to US coastal areas that are frequently hit by hurricanes — which bring not only fluvial but also pluvial flooding.”
Next, he hopes to take the tool truly worldwide.
“Our ultimate goal is for end users — agencies like NOAA — to use this tool for better decision-making. The strength of our FIER framework is that it’s highly scalable, because satellite images are available all over the globe. It also doesn’t require such expert skills as the hydrodynamic model, [which is good] because we want the FIER framework to be implemented in other flood-prone regions around the world. By combining this model with satellite observations, communities can improve decision-making to enhance both safety and economic resilience.”