The award for supercomputer work to advance tsunami warning times could mean better predictions in San Diego and worldwide.
A computational seismologist at UC San Diego’s Scripps Institution of Oceanography is part of a team that has won a prestigious award for their work in tsunami prediction, paving the way for faster warnings locally and globally.
Alice Gabriel/Scripps Institute of Oceanography
Called the Association for Computing Machinery Gordon Bell Prize and considered the “Nobel Prize of supercomputing,” the Gordon Bell Prize is an annual recognition of “the best science that uses the biggest computers,” said Alice Gabriel, who has been at SIO since 2022.
A physicist by training, Gabriel now specializes in simulating earthquakes and tsunamis using the largest supercomputers available.
Gabriel and her team members, who hail from the University of Texas at Austin and Lawrence Livermore National Laboratory, knew they’d been one of six finalist teams for the award and learned they’d won after presenting their work Nov 20 at the annual Supercomputing Conference.
“It’s pretty exciting,” said Gabriel, who had been a finalist for the same award 11 years ago as a new PhD graduate.
A supercomputer is a particularly powerful computer with a large number of cores, like chips, that work together to analyze mass amounts of data.
For Gabriel’s work, the supercomputers have Graphics Processing Units, or GPUs, similar to chips used by artificial intelligence companies.
In the prize-winning study, “we’ve been developing something that we call a ‘digital twin,’ … a computer model of something that exists in nature,” she said.
The “twin” in this case is a large, disastrous tsunami that could happen in the Cascadia subduction zone, a fault line in the Pacific Northwest.
“It’s one of the most hazardous fault systems that could affect the United States and Canada,” Gabriel said, adding researchers expect a large earthquake in that area that could potentially cause a tsunami.
A tsunami alert currently gives about 10 to 15 minutes’ advance notice, she said; the team’s research aims to develop a “real-time tsunami forecasting system that uses these large computers to give us, within seconds, a warning about that tsunami that’s imminent.”
The findings indicate such a system is now possible, Gabriel said.
Though this research focused on the Pacific Northwest, Gabriel said this kind of system would be applicable in San Diego and La Jolla, where the Kamchatka earthquake and tsunami triggered a tsunami advisory in July.
The local effects were about half an inch of elevated wave heights and dangerous rip currents, she said.
Tsunamis are a topic of interest here, Gabriel emphasized, as while the curvature of San Diego’s shoreline often shields the region from tsunamis on documented fault lines, there are other faults whose tsunami risks are unknown.
“We don’t have to know all the details” of the faults, she said. “We would just have to measure what happens on the sea floor, and we could issue a warning.”
What scientists need to put these findings into practice, Gabriel said, are more seafloor bottom sensors, noting her study assumes the distribution of about 600 sensors on the ocean floor, constantly measuring pressure.
When pressure changes due to an earthquake, “it changes in such an interesting way that we can calculate out of these measurements that there’s a tsunami about to hit the coastline,” Gabriel said.
The sensors exist in Japan, but not yet in the Pacific Northwest, she said. “A big goal would be to get these sensors in place to be able to implement that.”
The team’s Gordon Bell Prize opens many doors to further their research, as just being a finalist means the scientists have access to some of the largest supercomputers worldwide, an advantage that offers many resources to prove their approach is effective.
From the researchers’ simulations, “we now have a lot of follow up work,” she said, including exploring whether a minimum of 100 sensors would be sufficient for useful data.
Simplifying the approach to be more practical will lead to a better operational method of tsunami detection, Gabriel said. “This is a huge push for us.”
