Scripps researcher wins Gordon Bell Prize for real-time tsunami forecasting

A UC San Diego Scripps Institution of Oceanography seismologist has won an award often called the “Nobel Prize of supercomputing” for developing a real-time tsunami forecasting system.

Alice Gabriel, originally from Dresden, Germany, is part of a team that won the 2025 Association for Computing Machinery Gordon Bell Prize last week at the Supercomputing Conference in Missouri.

Gabriel’s team developed a “digital twin” framework, a computational replica combining physics-based models with real-time observations, reducing what normally requires 50 years of supercomputing time to less than a second.

“Using the largest supercomputers in the world, we developed new algorithms using seafloor sensors to forecast tsunami wave heights and their uncertainties in fractions of a second, 10 billion times faster than conventional algorithms,” said Gabriel, who has been at Scripps since 2022.

Traditional tsunami warning methods first attempt to characterize the earthquake, then use simplified models to determine threats. This takes precious time and sometimes produces false alerts that erode public trust, particularly in developing countries, according to Gabriel.

Gabriel’s team, which includes researchers from the University of Texas at Austin and Lawrence Livermore National Laboratory, applied their system to the Cascadia subduction zone in the Pacific Northwest. This 700-mile region stretching from Northern California to British Columbia hasn’t experienced a major earthquake since 1700, but seismologists estimate a 37% probability of a magnitude 8.0 or higher quake in the next 50 years.

When a tsunami strikes, coastal communities, including San Diego County, may have only 10 to 15 minutes of warning. This new research could greatly improve the speed and accuracy of early warnings for coastal areas at risk of earthquakes and tsunamis.

“Saving minutes would make a huge difference for saving lives,” Gabriel said.

Beyond speed, the system provides something else of value: uncertainty estimates. Rather than just forecasting a wave height, it can indicate confidence levels. This helps critical infrastructure operators make informed decisions and reduces the likelihood of false alerts, improving cost efficiency.

However, implementation faces a significant obstacle, Gabriel said. The United States doesn’t currently have the necessary seafloor sensors deployed off the Pacific Northwest coast. The prototype assumes 600 sensors measuring pressure on the sea floor.

“We showed this is possible, but we would need the sensors on the sea floor to really implement it,” Gabriel said.

Scientists have advocated for sensors since 2017, but implementation depends on research funding priorities at agencies such as the National Science Foundation. The team’s success in winning the Gordon Bell Prize may help advance these efforts.

Japan does have sensors in place and Gabriel’s team is now adapting its framework for Japanese waters, exploring questions such as the minimum number of sensors needed for effective warnings.

For Gabriel, the recognition is particularly meaningful. In 2014, while still in Germany, she was a finalist for the Gordon Bell Prize for modeling an earthquake in California but didn’t win. Eleven years later, she is proud geohazard research received the award among fierce competition.

“Any kind of model could win. It’s really open,” she said. “I think it’s great to have a geohazard topic winning because tsunamis and earthquakes affect so many communities.”

Gabriel credits Scripps with supporting the interdisciplinary collaboration required for such work.

“The biggest challenge in blending seismologists and supercomputers is we have a different language,” she said. “Like in every good relationship, you have to communicate.”

“And, of course, from my office window, I see the ocean every day, so I can think a lot about tsunamis.”