For University of Houston professor Hyongki Lee, living firsthand through the calamity and destruction caused by Hurricane Harvey in 2017 has shaped his research interests, and brought to the forefront how important the need for a forecast flood extent can be.
Lee, an associate professor in the Cullen College of Engineering's Civil and Environmental Engineering Department, is the principal investigator for a grant, “Forecasting Inundation Extents Using VIIRS and SAR Imagery with Streamflow Forecasts from NOAA’s River Forecasting Centers/National Water Model and GEOGloWS.” The three-year project, tentatively budgeted for $513,804, was selected by the National Oceanic and Atmospheric Administration in December 2020.
In the statement of need for the project, Lee described how the lack of information in 2017 had a personal effect on him and his family.
“During the night of August 30, 2017, [Lee] had to anxiously monitor throughout the night the rapidly rising water level over his front porch every hour to make an evacuation decision,” he wrote. “The local TV news had announced that the U.S. Army Corps of Engineers (USACE) was starting to release water from the Addicks and Barker Reservoirs located upstream of his home in order to reduce the risk of a dam breach. However, there was a sudden hydrologic information blackout which lasted five days as the in-situ gauge to monitor water levels had washed out. The next morning, PI Lee and his family evacuated by boat from the floodwaters, thanks to the voluntary rescuers from the neighborhood using their own boats. If there was a source of predicted hydrologic-hydraulic conditions, PI Lee could have avoided his high-risk decision for his family.”
Ideally, the project will allow for more accurate forecasting of flooded extents by applying the new so-called Forecasting Inundation Extents using REOF analysis (FIER) technique, developed by Lee’s group. The technique was first developed using synthetic aperture radar images that aren't effected by cloud cover during the rainy season.
The researchers will focus on three test areas in United States – the Mississippi River Basin around New Madrid, Missouri for riverine flooding; the Red River Basin for snowmelt-induced flooding; and southeastern Texas, including the Houston metropolitan area, for pluvial (rainfall-based) flooding.
“This NOAA project is to experiment with our idea of applying the FIER technique to a stack of historical VIIRS imagery that have cloud covers, and extract the inundation signal only over three flood-prone regions in the U.S. as test studies,” Lee said. “Eventually our goal is to provide cloud-free forecasted inundation extents, being coupled with forecasted streamflows from NOAA’s National Water Model and GEOGloWS streamflow forecasting system, or forecasted river levels from National Weather Service (NWS) River Forecast Centers. Once our test studies are successfully performed, our long-term goal is to have this eventually implemented for operational uptake by NOAA and provide forecasted inundation extents along with currently operational streamflow/river level forecasts.”
Co-investigators include Gustavious Williams and E. James Nelson at Brigham Young University, and William Straka III from the Cooperative Institute for Meteorological Satellite Studies at the University of Wisconsin at Madison.