An Ohio State geography professor and team of researchers across the country developed a method of predicting power outages during severe weather that could help communities better plan for hurricanes.
Professor Steven Quiring used the Spatially Generalized Hurricane Outage Prediction Model that predicts outage-prone areas based on rainfall, wind power and soil saturation to help major agencies, such as the Department of Homeland Security, plan ahead for mass power outages.
This model, developed by Quiring, along with professors and researchers from Texas A&M and the University of Michigan, was used for areas in Texas impacted by Hurricane Harvey, as well as East Coast areas struck by Hurricane Irma.
The system uses information from utility companies and data taken from hurricanes like predicted wind speeds, as well as the moisture and soil type of a region, in order to pinpoint areas that might have a high risk of trees falling, which can cause power outages.
The SGHOP model uses a machine learning system that actively records information during the severe weather event to gather information about what is going on in the midst of the hurricane. It then relays the gathered information to researchers, like Stephen Shield, a graduate student in geography, to analyze the accuracy of their system.
The model has accurately predicted, with a 10 percent margin of error, power losses during previous hurricanes such as Sandy and Matthew.
After a storm, data from the SGHOP model provides information that can help first responders effectively address recovery efforts, Quiring said.
“They need to know how many people to bring in, where to situate them and how many poles and transformers may break during an event,” he said.
The practical use of this method has increased due to an especially tumultuous Atlantic hurricane season.
Quiring said the model can be used for long-term planning to improve pre-storm preparation, thus speed up recovery time for communities impacted by storms.
There soon could be a model for severe weather events in Ohio, Shield said.
“I’m working on thunderstorms, and Jordan Pino [a doctoral candidate in geography at Ohio State] is working on a model for snowstorms and ice storms.”