AI reveals first direct observation of rupture propagation during slow quakes

Deep-learning approach extracts data from ‘noisy’ images to reveal how ground deforms during seismic event

December 9, 2020

Researchers from Los Alamos and the École Normale Supérieure in Paris are using artificial intelligence and satellite radar images to understand the physics driving earthquakes, from data gathered on the North Anatolian Fault in Turkey. Illustration credit: Dreamstime

LOS ALAMOS, N.M., Dec. 9, 2020—Using a trained neural network and data from the North Anatolian Fault in Turkey, a research team led by Los Alamos National Laboratory revealed the first direct observation of rupture propagation during a slow earthquake. The research will be presented Dec. 15, 2020 at the AGU Fall Meeting.

Applying AI to interferometric synthetic aperture radar (InSAR) images opens up a new way to unravel the physics behind tectonic faults and earthquakes, which is key to understanding the full spectrum of earthquake behavior.

“The deep-learning approach we developed makes it possible to automatically detect the small and transient deformation that occurs on faults with unprecedented resolution, paving the way for a systematic study of the interplay between slow and regular earthquakes, at a global scale,” said Bertrand Rouet-Leduc, a geophysicist at Los Alamos National Laboratory and leader of the research team.

The geophysicists from Los Alamos and the École Normale Supérieure in Paris trained the neural network to remove atmospheric “noise” in the data and extract ground deformation from InSAR data. The AI approach works without expert interpretation or previous knowledge of the fault system being studied.

InSAR is a satellite-based mapping technique that uses radar to create images of ground deformation. Unlike optical imaging systems, InSAR can “see” through clouds and works at night, but atmospheric turbulence and other distortions can skew the measurements of details on the earth’s surface. The new deep-learning tool corrects for those distortions, and learns to distinguish signal from noise. 

Bertrand Rouet-Leduc is available for an interview.

Presentation: “S056-07 - A deep learning approach for detecting transient deformation in InSAR (Invited),” AGU Fall Meeting,

Funding: Los Alamos National Laboratory, Laboratory Directed Research and Development (LDRD) program.

About Los Alamos National Laboratory

Los Alamos National Laboratory, a multidisciplinary research institution engaged in strategic science on behalf of national security, is managed by Triad, a public service oriented, national security science organization equally owned by its three founding members: Battelle Memorial Institute (Battelle), the Texas A&M University System (TAMUS), and the Regents of the University of California (UC) for the Department of Energy’s National Nuclear Security Administration.

Los Alamos enhances national security by ensuring the safety and reliability of the U.S. nuclear stockpile, developing technologies to reduce threats from weapons of mass destruction, and solving problems related to energy, environment, infrastructure, health, and global security concerns.