Even as researchers use supercomputers to probe the mysteries of earthquakes here on Earth, others are setting their sights on quakes just a little farther away. Researchers at ETH Zürich in Switzerland have applied supercomputing power from the Swiss National Supercomputing Centre (CSCS) to analyze marsquakes, illuminating the structure of Mars itself.
In 2018, NASA successfully landed a robotic detective on the Martian surface. The lander, called InSight (for INterior exploration using Seismic Investigations, Geodesy and Heat Transport), was there to investigate Mars’ seismic activity. To do that, it brought along Mars’ first seismometer, the Seismic Experiment for Interior Structure (SEIS), which was provided by ETH Zürich.
“Marsquakes have characteristics already observed on the Moon during the Apollo era, with a long signal duration of 10 to 20 minutes due to the scattering properties of the Martian crust,” said Domenico Giardini, the ETH Zürich professor leading the research project, in an interview with CSCS. Still, despite their equipment, researchers struggled to identify the direction of the waves, mostly capturing only their distance. Mars’ unfavorable conditions – including destabilizing high winds that shook the lander – only added to the difficulties.
InSight has now been on Mars for nearly 500 sols (which are slightly longer than Earth days). During that time, it has recorded over 450 seismic events, roughly one per day. That seismic information is sent to the Marsquake Service back on Earth, which is led by ETH Zürich. The goal of the research is to deepen understanding of Mars’ internal structure by analyzing its seismic waves.
In 2018, when InSight was still in flight, ETH Zürich was running a series of simulations to prepare for interpreting the data it would return. In essence, the researchers at ETH Zürich operated backwards, starting with around 30 different models of Mars’ internal structure and then propagating waves through those simulated structures. They then provided the resulting data to other researchers, asking them to interpret the structure from the wave data.
To run those simulations, ETH Zürich turned to CSCS’ Piz Daint supercomputer. Piz Daint is a Cray system comprising 5,704 XC50 nodes (each with an Intel Xeon E5-2690 CPU and an Nvidia Tesla P100 GPU) and 1,813 XC40 nodes (each with two Intel Xeon E5-2695 CPUs). Its two sections rate at 21.2 Linpack petaflops and 1.9 Linpack petaflops respectively, placing 6th and 185th on the most recent Top500 list of the world’s most powerful supercomputers. “Without a supercomputer like Piz Daint, simulating a single model on a laptop would have taken more than two years – so about four times as long as the lander’s journey to Mars,” said Christian Böhm, a senior research scientist in the Seismology and Wave Physics group at ETH Zürich, in an interview with CSCS.
With actual data in hand, the ETH Zürich researchers are now free to compare the simulations with reality, finding similarities between the seismic waves of their imagined structures and the seismic waves of Mars’ actual, mysterious structure. Using those comparisons, the researchers will refine and repeat the models, attempting to match reality as closely as possible.
Already, the researchers have drawn some conclusions from studying 174 of the recorded marsquakes, including that the uppermost ~5-7 miles of Mars’ crust is highly altered or fractured and that it contains more volatile elements (like oxygen and carbon dioxide) than the Earth’s moon.
Header image: one of Piz Daint’s simulated marsquakes. Image courtesy ETH Zürich.
For more information on ETH Zürich’s analyses of the InSight data, visit Michèle Marti’s article here.