(ORDO NEWS) — Earthquakes are more than just tumbling streets and overturning buildings. Seismic waves generated by earthquakes travel through the Earth, acting like a giant MRI machine and giving clues about what is inside the planet.
Seismologists have developed methods for taking wave signals from networks of seismometers on the Earth’s surface and reverse engineering the features and characteristics of the environment they pass through, a process known as seismic tomography.
For decades, seismic tomography was based on the theory of rays, and seismic waves were considered as light rays. This was a pretty good approximation and led to major discoveries about the interior of the Earth.
But to improve the resolution of today’s seismic tomographic models, seismologists need to account for the complexity of wave propagation using a numerical simulation known as full waveform inversion, says Ebru Bozdag, assistant professor of geophysics at the Colorado School of Mines.
“We are at a stage where we need to avoid approximations and corrections in our imaging methods to build these earth models,” she said.
Bozdag was the lead author of the first full waveform inversion model, GLAD-M15 in 2016, based on full 3D waveform modeling and global scale sensitivity of 3D data.
The model used the open source 3D global wave propagation solver SPECFEM3D_GLOBE and was created in collaboration with researchers from Princeton University, University of Marseille, King Abdullah University of Science and Technology (KAUST), and Oak Ridge National Laboratory (ORNL).
This work has received high praise in the press. Its successor, GLAD-M25 (Lei et al. 2020), was released in 2020 and has provided important features such as subduction zones, mantle plumes and hotspots to be seen to further discuss mantle dynamics.
“We have shown the feasibility of using full 3D wave modeling and the sensitivity of the data to seismic parameters on a global scale in our 2016 and 2020 papers. Now it’s time to use the best parameterization to describe the physics of the Earth’s interior in the inverse problem,” she said.
At the American Geophysical Union Fall Meeting in December 2021, Bozdag, postdoctoral researcher Ridvan Orsvuran, Ph.D. Armando Espindola-Karm. Student Armando Espindola-Carmona and computational seismologist Daniel Peter of KAUST and their colleagues presented the results of their efforts to perform global full waveform inversion to model attenuation – a measure of energy loss as seismic waves propagate inside the Earth – and azimuthal anisotropy – including how the wave speed varies with the direction of propagation in the azimuth direction, in addition to the radial anisotropy taken into account in the first generation GLAD models.
They used data from 300 earthquakes to build new global models of complete wave inversion. “We update these Earth models in a way that iteratively minimizes the difference between observational data and simulated data,” she said. “And we’re looking to understand how our model’s parameters, elastic and inelastic, fit together, which is a challenging task.”
The study is supported by the National Science Foundation’s (NSF) CAREER award and is conducted using the Frontera supercomputer at the Texas Center for Advanced Computing – the fastest among all universities and the 13th fastest in the world – and the Marconi100 system at Cineca, Italy’s largest computing center.
“Thanks to access to Frontera, public data from around the world, and the power of our modeling tools, we have begun to move closer to continental-scale resolution in our global total wave inversion models,” she said.
Bozdag hopes to get more accurate data on the origin of mantle plumes and the water content in the upper mantle. In addition, “to accurately locate earthquakes and other seismic sources, determine the mechanisms of earthquakes and more accurately correlate them with plate tectonics, it is necessary to have high-resolution models of the crust and mantle,” she said.
From the deepest oceans to outer space
Marsquake – Cerberus Fossae event (Mw 3.1). Visualization shows the speed of seismic waves (vertical component). The researchers used Frontera to simulate this event in collaboration with NASA’s InSight mission. Credit: Daniel Peter, KAUST
Bozdag’s work is relevant not only on Earth. She also shares her numerical simulation experience with NASA’s InSight mission as part of the Martian Interior Simulation Science Team.
Preliminary data on the Martian crust, limited to seismic data for the first time, were published in the journal Science in September 2021. Bozdag, along with the InSight team, continues to analyze quake data and elucidate details of the planet’s interior from crust to core using Frontera’s 3D wave simulation.
The work on Mars showed in perspective the lack of data in some parts of the Earth, in particular, under the oceans. “Now we have data from other planets, but getting high-resolution images under the oceans is still difficult due to the lack of instruments,” says Bozdag.
To solve this problem, she is working to integrate data from new instruments as part of her NSF CAREER award into her models, such as data from floating acoustic robots known as MERMAIDs (Mobile Earthquake Recording in Marine Areas by Independent Divers). These autonomous submarines can detect seismic activity in the ocean and rise to the surface to relay this data to scientists.
Seismic community access
In September 2021, Bozdag was part of the team that received a $3.2 million NSF award to create a computing platform for the seismological community known as SCOPED (Seismic COmputational Platform for Empowering Discovery), in collaboration with Karl Teip (University of Alaska-Fairbanks) , Marin Denollet (University of Washington), Felix Waldhauser (Columbia University), and Jan Wang (TACC).
“The SCOPED project will create a Frontera-backed computing platform that will provide data, computing and services to the seismological community to advance education, innovation and discovery,” said Wang, TACC Research Fellow and co-principal investigator for the project. “TACC will focus on developing the core cyber infrastructure that will serve both computational and data-intensive studies, including seismic imaging, waveform modeling, ambient noise seismology, and precision seismic monitoring.”
Another community-focused Bozdag group project is the recently released SphGLLTools project by graduate student Caio Ciardelli: a visualization toolkit for large seismic model files. Based on this toolkit, the package facilitates the plotting and sharing of global anchor point tomographic models with the community. The team described the toolkit in Computers & Geosciences in February 2022.
“We provide a complete set of computational tools to visualize our global joint models,” Bozdag said. “Someone can take our high-performance simulation-based models and convert them into a format that allows them to be visualized on personal computers and use collaborative notebooks to understand every step.”
Robin Reichlin, director of the NSF geophysics program, says that “with new, improved full wave models; tools that lower the bar for community data access and analysis; and a supercomputer-based platform that allows seismologists to unlock the mysteries of the deep interior of the Earth and other planets, Bozdag is advancing this area into a more precise and open area.”
Contact us: [email protected]