NAS Researchers Bring Asteroid Simulation Down to Earth
If you were standing on the southern coast of Spain just after midnight on February 22, 2020, you would have been dazzled by a brilliant fireball a hundred times brighter than Venus streaking across the night sky. Disappearing harmlessly 30 kilometers above the Mediterranean Sea, the meteor was one of millions that burn up in Earth’s protective atmosphere every day—most as small as dust particles.
Once in a while, a larger one gets far enough through the atmosphere to affect the planet’s surface, either explosively breaking up in the air, creating a blast wave that propagates to the ground, or impacting the ground itself. Predicting what happens next is where researchers working in the NASA Advanced Supercomputing (NAS) Division come in. As part of NASA’s Asteroid Threat Assessment Project (ATAP) team, these researchers run millions of hypothetical airburst and impact scenarios on supercomputers at the NAS facility, located at NASA’s Ames Research Center in Silicon Valley. Using their Probabilistic Asteroid Impact Risk (PAIR) model, they estimate the effects of blast waves, thermal radiation, tsunamis, and global climatic effects that could result from such impacts.
"The PAIR model integrates physics-based analysis of asteroid entry, breakup, and ground damage into a probabilistic risk simulation to estimate the likelihood of various damage levels for different sizes and classes of asteroids,” explained Lorien Wheeler, a researcher on the ATAP team. “We model hypothetical impact scenarios based on the close approach of real asteroids, with fictitious impact trajectories, as well as hypothetical asteroids of various sizes and properties."
From Hypothetical to Real World
Building on this work, the team recently turned their attention to supporting the development of the modeling capabilities that would be used if a real threat is identified. In support of the National Near-Earth Object (NEO) Preparedness Strategy and Action Plan, the ATAP team is leading a multi-agency asteroid impact modeling group (see sidebar) to “establish, exercise, and improve an integrated suite of computational tools for modeling NEO impact risks and mitigation techniques.”
The exercises are critical to helping ensure that we have the best possible modeling capabilities ready to go, in the event that an actual impact threat is ever discovered.
“Ultimately,” Wheeler said, “the results of these studies will help to assess what types and sizes of asteroids pose enough of a threat to warrant tracking, what level of mitigation is appropriate, and what strategies or technologies may be most effective at reducing the risks from impacts.”
A major part of this effort is conducting hypothetical asteroid impact exercises, in which a fictitious scenario is used as a test case to assess U.S. modeling capabilities. Among these are the probabilistic impact risk models and high-fidelity impact damage simulations that the ATAP team develops and runs on the Pleiades supercomputer. The group recently completed the first of these exercises this winter.
“The exercises are critical to helping ensure that we have the best possible modeling capabilities ready to go, in the event that an actual impact threat is ever discovered,” Wheeler explained. “They also help us identify ways to keep improving the models, and make sure that all the key agencies know how to work together to assess the threat and develop the most effective mitigation strategies.”
New Technologies to Burst Through Data Challenges
One of the biggest challenges in assessing asteroid impact risk and damage is insufficient information about the properties of meteors before they enter the atmosphere—such as size, speed, density, and strength—making it difficult to determine how those properties affect their entry and breakup, and the resulting damage.
To address this challenge, the ATAP team is working with NAS data analytics experts to develop machine learning methods and test them with their Fragment-Cloud Model, which estimates the energy deposited in the atmosphere during asteroid entry and breakup.
Machine learning will also help the researchers take advantage of a recently established data pipeline that sends information from satellite-borne instruments to the NAS facility’s Pleiades supercomputer for processing. Originally developed to detect lightning flashes, the Geostationary Lightning Mapper (GLM) sensors turn out to be very good at detecting bright meteors when they hit the atmosphere; machine learning methods can help distinguish the meteor data in the pipeline from the millions of other flashes caused by lightning or other sources. As this capability develops, it may provide the ATAP team with an increasing number of meteor observations to incorporate into their models.
Between the NASA supercomputers here on Earth, observatories in space, and the dedication of NAS researchers, data scientists, and their partners, we are adding more tools to improve our chances of lessening the damage of potential asteroid impacts, and hopefully even prevent them. In the next part of this series, we’ll look more closely at how new machine learning techniques and the GLM data pipeline are helping us get there.
—Michelle Moyer, NASA/Ames Research Center
A special issue of Icarus—The International Journal of Solar System Studies focuses on studies of the 1908 Tunguska meteor impact. Published in July, 2019, the issue was co-edited by ATAP scientist Darrel Robertson and Ames researcher David Morrison. Although Icarus requires subscription, several papers from the special issue authored by ATAP team members are available without subscription; see sidebar.
