Scientists at Lawrence Livermore National Laboratory (LLNL) have developed a real-time tsunami forecasting system that uses the computational power of El Capitan, currently recognized as the world’s fastest supercomputer. The new system is designed to enhance early warning capabilities for coastal areas near earthquake zones.
El Capitan was built with support from the Advanced Simulation and Computing program at the National Nuclear Security Administration. Before transitioning to classified work, researchers used El Capitan in an offline phase to generate a large library of physics-based simulations that link earthquake-driven seafloor movements to resulting tsunami waves. This precomputation process used more than 43,500 AMD Instinct MI300A Accelerated Processing Units.
The research team, which includes partners from the Oden Institute at the University of Texas at Austin and Scripps Institution of Oceanography at UC San Diego, produced what they call a tsunami “digital twin.” This model integrates real-time pressure sensor data with advanced simulations to forecast tsunami behavior and quantify uncertainty.
“This is the first digital twin with this level of complexity that runs in real time,” said LLNL computational mathematician Tzanio Kolev, co-author on the paper. “It combines extreme-scale forward simulation with advanced statistical methods to extract physics-based predictions from sensor data at unprecedented speed.”
The approach enables rapid predictions in seconds using smaller GPU clusters during an actual event. Researchers report that their method can solve a billion-parameter Bayesian inverse problem in less than 0.2 seconds—a significant improvement over previous models.
Researchers highlight that current warning systems often rely on simplified models based on seismic and geodetic data, which may not capture complex fault ruptures and can result in late or inaccurate alerts. The new method instead uses detailed seafloor pressure sensor data and full-physics modeling to improve accuracy and speed.
As more seafloor sensor networks are deployed along earthquake-prone coasts, researchers see potential for widespread use of this approach in future warning systems.
“This framework represents a paradigm shift in how we think about early warning systems,” said Omar Ghattas, senior author of the study and professor at UT-Austin’s Oden Institute. “For the first time, we can combine real-time sensor data with full-physics modeling and uncertainty quantification — fast enough to make decisions before a tsunami reaches the shore. It opens the door to truly predictive, physics-informed emergency response systems across a range of natural hazards.”
The system relies on MFEM, LLNL’s open-source finite element library, which allows scalable simulations for phenomena like acoustic-gravity wave propagation in oceans. Running these simulations on El Capitan required solving equations with 55.5 trillion degrees of freedom—a record for unstructured mesh finite element simulation.
“MFEM’s high-order methods and GPU readiness, developed under the ASC program at LLNL and the Department of Energy’s (DOE) Exascale Computing Project, made it possible to scale to the full machine,” Kolev said. “This was really a first-of-its-kind demonstration of how we can use that power not just for raw performance, but also for mission-relevant, time-critical decisions in many MFEM-based applications.”
Kolev noted that after initial precomputations are done on El Capitan, subsequent forecasting steps can be performed quickly on smaller GPU clusters because of efficient algorithm design.
“This work is important because it shows that we can solve an inverse problem of enormous size — not for 10 or 15 variables, but for millions, or even billions of variables, very quickly,” said Kolev. “In the past, you’d either have a fast model that’s not accurate, or a full-physics model that takes hours or days. Now we’re showing that we can do both — accurate and fast — using principled mathematics and modern computing.”
Kolev added that this Bayesian inversion framework could also be adapted beyond tsunamis to areas such as wildfire tracking or space weather forecasting where quick analysis is needed.
Collaborators included Veselin Dobrev and John Camier from LLNL; Omar Ghattas, Stefan Henneking, Milinda Fernando and Sreeram Venkat from UT-Austin; and Alice-Agnes Gabriel from UC San Diego.
