The cause of earthquakes is known: tectonic plates scraping and clashing against each other. Scientists seeking a deeper understanding of the underlying causes of that plate movement turn to Texas A&M University professor Wolfgang Bangerth, a widely respected expert in computational mathematics and mathematical modeling.
Simulating the complex processes that unfold across millions of years hundreds of miles below the Earth’s surface requires not only an understanding of how rocks behave under immense pressure and high temperatures, but also software capable of describing the planet’s mantle to computers using mathematical models with billions of variables. Drawing on his proficiency in computational mathematics and broad understanding of the sciences, Bangerth has written ASPECT (Advanced Solver for Problems in Earth’s Convection), code that is used around the world and funded by a major facility in California at the epicenter of geodynamics research.
Bangerth’s ASPECT program is based off of deal.II, a modeling software library he developed starting as an undergraduate in Germany in the 1990s that is credited with helping researchers harness the power of supercomputers to provide a more accurate picture of complex problems and processes in many areas, from healthcare to environment to energy. In 2010, he was contacted by the Computational Infrastructure for Geodynamics, a National Science Foundation-funded facility at the University of California, Davis that supports the computing needs of the world’s geodynamics researchers. The facility awarded Bangerth a five-year, $800,000 subcontract to write software that allows researchers to simulate processes occurring within Earth’s interior.
Bangerth says his software — developed along with visiting assistant math professor Timo Heister — helps researchers who are studying the Earth’s mantle, the 1,800-mile-thick rocky shell between the planet’s crust and its core. Unlike the lava that pours out of volcanoes, the material in the mantle is largely solid due to the enormous pressure, but Bangerth notes it can move very slowly — at speeds up to 2 or 3 inches per year — like an extremely viscous liquid if observed across long enough time scales. It’s the processes within this region of the Earth that lead to tectonic plate movement, which during the course of hundreds of millions of years has led to the formation of the Rocky Mountains, Himalayas and much of the planet’s natural beauty. Researchers study the mantle to better understand these processes.
“Scientists want to use computers to simulate the Earth’s mantle to find out why continental plates or oceanic plates move,” Bangerth said. “If we understand how fast plates move, we can predict how often earthquakes happen and maybe even how strong they’ll be.”
Researchers have several theories about why the plates move. But in each case, they are studying events occurring deep below the Earth’s surface at a place they can’t see and that, over timescales of millions of years, can only be simulated on computers using specialized software which few scientists can write. Bangerth’s ASPECT program enables them to accurately describe how material flows in the Earth’s mantle, including how much heat is transported from the planet’s interior to the surface, the processes responsible for volcanoes in unlikely places such as Hawaii in the middle of an oceanic plate, and possible insights into parts even deeper in the Earth, such as its core.
While Bangerth says ASPECT is a powerful tool, he notes that earthquake prediction — exactly when an earthquake will happen — is still out of its reach. He likens it to grains of sand in an hourglass dropping onto the pile at the bottom. Researchers using software can figure out how many grains of sand fall per second, and even where, on average, they will land. But determining exactly which grain of sand triggers the mini-avalanche on the sand pile below isn’t possible yet.
Bangerth is using a recent $1.3 million NSF grant to expand on the possibilities for deal.II, improving its flexibility and ease of use in hopes of continuing to build on its growing global popularity among scientists studying a variety of topics, from plant root growth and glacier mechanics to heart muscle fiber simulation and even the impact of air pollution on Roman statues.
“It’s my desire to make the things that we develop on paper in mathematics available to applied scientists and engineers within academia, the national labs and industry,” Bangerth said. “Research at universities is taxpayer-funded, and I do sincerely believe that we have a responsibility to make what we develop broadly available to society.”
Source: Texas A&M University