Eye of the Storm
Len Pietrafesa didn’t know if he was going to make it. He and his three colleagues had thrown everything overboard that wasn’t bolted down—furniture, supplies, even the refrigerator—trying to keep the 48-foot launch afloat. Here he was, a 22-year-old physicist barely out of college, on his first job with a solid future ahead of him… preparing to go down with the ship in a savage typhoon off the coast of Australia.
The Boston engineering firm he worked for had sent him down under to Port Hedland on the west Australian coast to conduct a study for a dredging project. The report was nearly finished when Pietrafesa decided he needed more data sets, so he arranged a one-day trip offshore. He and his team set out at 4 a.m. and expected to be back by midnight.
“There were no satellites then,” recalls Pietrafesa. “When we got a message from the Air Force about a monster storm heading toward the northwest coast, we were too far out to get back to port in time.”
After tossing the cargo over the side as gale winds pitched the vessel into the crashing waves, they strapped themselves down in the boat. And prayed.
Fortunately, they survived the storm, but it blew them hundreds of miles out to sea. It took them four days to get back to port. All they had was drinking water and some sandwiches. When they finally limped back to land, Pietrafesa was sunburned, hungry and humble.
The incident might well have ended his life, but in fact it was the real beginning of his career. It was the catalyst for a newfound interest in the physics of weather that led him to pursue a Ph.D. in fluid physics and geophysical fluid dynamics.
Pietrafesa has spent most of his career as a professor and scholar at N.C. State University, where over the years he has built a worldwide reputation for his authoritative research on climate, weather and the ocean, and their impacts on society. His expertise has brought him honors and invitations to serve on every important committee and board in the world of weather, including the chairmanship of the NOAA Science Advisory Board. He has also appeared before committees of the U.S. Senate and the House of Representatives as an expert witness on climate and weather issues.
Today Pietrafesa is on the faculty of CCU’s School of Coastal and Marine Systems Science (CMSS). Now retired from N.C. State, he joined CMSS last year after a decade-long professional association with Paul Gayes, longtime director of CCU’s Center for Marine and Wetland Studies and now director of CMSS, on a study of offshore wind power potential along the Carolina coast.
The two scientists share a common approach to coastal research. They have devoted their respective careers to the intensive study of the complex physical interrelations of multiple systems—ocean, land and atmosphere. The hurricane model that Pietrafesa has spent a career perfecting is a natural fit for CCU,
where for the past 25 years Gayes has built an internationally respected program known for its research on coastal erosion, sediment patterns, seafloor mapping, wind power and other studies relating to geophysical processes.
The other members of the team are Tingzhuang Yan, a meteorological oceanographer who was a key player in developing the hurricane forecasting system for NOAA’s National Centers for Environmental Prediction (NCEP), and Shaowu Bao, a computational expert specializing in atmospheric and oceanic climate systems. Both men have worked closely with Pietrafesa through their careers, beginning as students and rising to prominence in the field of hurricane climatology.
Their new model for hurricane prediction was unveiled Aug. 1 at perhaps the most significant media conference in CCU history. The reason the announcement was so newsworthy—grabbing headlines from here to Zimbabwe (200 media hits in all, according to CCU’s media service)—is because the new model is the only one that promises to deliver accurate data on that most vital piece of the puzzle: landfall.
Some of the other hurricane predictive instruments do include landfall in their matrix, but by and large they base their landfall prognostics on a regression model that assumes a percentage of the total activity. The one that Pietrafesa and his team created is different.
They call it HUGO, an acronym for the Hurricane Genesis and Outlook Project and a nod to the memory of the last big storm to hit South Carolina. Pietrafesa describes it as an “end-to-end” model. The front end is the part that predicts the large seasonal picture: the number of tropical storms, the number and likely intensity of hurricanes, how many will affect the Eastern seaboard versus the Gulf coast. The back end is the micro-picture. The track of the approaching hurricane is defined with greater precision than other models, and with better lead time: five days from landfall.
How is the HUGO team able to make such bold claims for its model? One reason is the breadth of empirical data that is fed into its algorithms. HUGO’s methodology employs a uniquely robust array of statistical analysis involving 22 separate climate factors. The HUGO deterministic model interactively couples other existing models, including NOAA’s Hurricane Weather Research and Forecasting (HWRF) and Wave Watch III models, the Regional Ocean Model System (ROMS) and Office of Naval Research’s Shallow Water Ocean Wave Model (SWAN) via the National Center for Atmospheric Research’s Earth System Modeling Framework.
One of HUGO’s most interesting and conclusive features is its intensive use of historical data from recorded storm systems. HUGO scientists have coded the characteristics of Atlantic hurricane activity since 1950 as a key methodology for predicting the likely track of future developing storms.
“When we tested our method against history, hindcasting the data for each year since 1950, we found it to be remarkably accurate,” says Pietrafesa. “It was either spot on or deviated from the historical record only by about one hurricane per year.”
The HUGO team updates its predictions periodically during the season as new data becomes available (see chart below). When a storm develops, in addition to detailed outlooks of its track and its characteristics and its ability to reduce the cone-of-uncertainty, the HUGO model system will predict the depth, time and location of storm surge and water inundation, information that emergency management officials say could mean the difference between life and death in evacuation planning.
The model provides statistical representations (in quality state-of-the-science GIS visualization) of water depths indicating the likelihood of flooding in the range from 0 to 100 percent probability. This will provide specific information to emergency management personnel, hospitals, local municipalities, the highway patrol, evacuation planners and other decision-makers on whether or not to evacuate, when to evacuate, where to evacuate to, which routes to take, etc.
“All the existing models offer linear data on storm surge,” says Randy Webster, director of emergency management for Horry County. “But CCU’s model system is vertical, with more information about depth and locations of surge. This is a tool we can use. For every mile of coastline you evacuate, it’s a million dollar impact.”