UO is leading an effort to speed up tsunami warnings

Tsunami warning sign

A sea change may be coming in the way tsunami warnings are issued worldwide, and University of Oregon researchers are steering the effort — with a project using GPS in the Cascadia subduction zone — to deliver them within five minutes of an earthquake.

If all goes according to plan, the National Oceanic and Atmospheric Administration will be able to meet that timing with precise forecasts of incoming wave sizes at particular locations, said project leader Diego Melgar, a professor in the Department of Earth Sciences.

Work began Sept. 1 under a four-year, $800,000 grant from NASA that is split between the UO and the University of Washington’s Brendan Crowell, a research scientist and member of the Pacific Northwest Seismic Network. The National Oceanic and Atmospheric Administration also is a key player in the project, which also involves a collaboration with Central Washington University.

“Our goal is to empower NOAA to be able to say more quickly the size of an earthquake and the expected tsunami it may generate in selected locations,” Melgar said.

Newly developed algorithms are being incorporated into some 500 GPS monitoring sites in Cascadia’s 600-mile-long coastal swath from Northern California to British Columbia. On campus, Melgar’s postdoctoral scholar Amy Williamson is crunching the mathematical formulas into GPS-based software to run thousands of simulations of strong earthquakes, such as a possible magnitude 9-plus event in the Pacific Northwest.

The idea is to determine the possibility of accurately forecasting tsunamis at time intervals ranging from one to 20 minutes after an earthquake.

“In Cascadia, it depends on where an earthquake starts,” Melgar said. “A full rupture, like the last one in January 1700, would take three to five minutes from start to finish. Tsunami impacts would likely begin as soon as 15 minutes, depending on location. With a five-minute warning, people could begin evacuations sooner and there would be more time to initiate protections for coastal infrastructure, such as ports.”

NOAA’s current standard is to issue tsunami warnings 20 minutes after a large earthquake, but doing so accurately is challenging. Quickly assessing both final magnitude of a quake and potential tsunami risk is difficult, Melgar said.

Such was the case with Japan’s 2011 magnitude 9.0 earthquake, in which initial size assessments failed to anticipate the ensuing catastrophic 30-foot-high tsunami that hit Sukuiso. In 2004, a magnitude 9.1 earthquake generated a surprise tsunami across much of Indonesia that killed more people than any other such event in recorded history.

“The technological approach we are using represents a sea change for NOAA,” Melgar said. “Amy is doing a lot of testing to show how it would work in an operational sense. What she is doing, in close collaboration with NOAA, is helping us understand how this approach may fit into the agency’s standard operating procedures.”

An improved approach using GPS data worldwide could provide a more robust and accurate warning system, Melgar said. Currently, warnings are not always followed by tsunamis, resulting in a false sense of security. The new approach, he said, would help NOAA quickly identify the size and type of earthquakes, resulting in enhanced accuracy of tsunami warnings.

The project is separate from efforts to build out ShakeAlert, the early earthquake warning system for the West Coast under development by the U.S. Geological Survey in collaboration with seven universities, including the UO, and a university-governed nonprofit.

Recent GPS-based research by Melgar and the USGS’s Gavin P. Hayes found early indicators in real-time GPS data that estimate an earthquake’s final magnitude just 10 to 15 seconds into an event. As a result, GPS monitoring is being added at ShakeAlert seismic stations. However, more testing is needed before the findings can be applied on a broader scale, including for tsunami warnings, Melgar said.

—By Jim Barlow, University Communications