In the days leading up to Hurricane Matthew, researchers used satellite maps of soil moisture to help forecast, with 91% accuracy, where the power would go out along the U.S. East Coast.
The model used was developed by scientists from the Ohio State University (OSU), the University of Michigan and Texas A&M University with the aim of curtailing outages by helping power companies allocate equipment and crews in advance of storms. It is premised on the notion that healthy trees that receive just the right amount of moisture are less prone to storm damage, so soil moisture levels are a good indicator of where outage crews will be needed.
“We see increased numbers of outages at both ends of the spectrum—wherever soils are too wet or too dry,” says Steven Quiring, OSU professor of atmospheric sciences. “Drought makes tree branches more likely to snap off, and over-saturation makes trees more likely to be uprooted.”
To make their forecast model, the researchers relied on data from NASA’s Soil Moisture Active Passive satellite mission, which they cross-referenced against data on population density, land use, average wind speed and the duration and intensity of storms.
For Hurricane Matthew, which hit the U.S. coast in October 2016, the team was able to forecast five days ahead of time that 4.5 million people would be without power in Georgia, North Carolina, South Carolina and Virginia. The actual number turned out to be around 4.1 million, so the researchers overestimated the extent of outages by around 9%.
Quiring says the researchers are expanding the project to include outages caused by thunderstorms, winter storms and wind storms, which impact a much larger portion of the United States than do hurricanes. They are already working with several power companies along the East Coast and hope to form partnerships with companies in the Midwest and South.