Marin grad uses AI to improve data about northern lights
For Jason Press, artificial intelligence is about more than chatbots. He thinks it could help scientists get better data about the aurora borealis.
Press, a Marin Catholic High School graduate, is using the technology to study northern lights as a data research project at Pepperdine University in Malibu. Clouds can compromise data collection on the glowing phenomenon.
“There are very few clear-sky nights with aurora data,” Press, 22, said in an email. “When clouds roll in, that data is usually lost. My job was to build a model that could see through the clouds.”
Press, a senior at Pepperdine, said he and the college research team are “literally recovering scientific data that was thrown out years ago.”
Press and his team were successful last summer in clearing up images from 90 minutes of satellite video of the aurora borealis. By December, they were invited to present their findings at the conference of the American Geophysical Union, a science association.
Fabien Scalzo, Press’ computer science professor, was pleased by the success of his student’s research.
“I’m thrilled by the work Jason has done,” Scalzo said in an email. “His curiosity and hard work at the intersection of computer vision and AI are opening vast areas of research in the analysis of Auroras.”
The aurora borealis, generated by the interaction of solar-charged particles and the Earth’s magnetic field, has been tracked by satellite for at least 10 years. However, at least half of the data from the satellite videos was useless because it was obscured by clouds, Press said.
Press’ chief mentor in the research, Pepperdine physics professor Gerard Fasel, said being able to read the aurora borealis more accurately could offer scientists the ability to predict catastrophic events caused by large solar flares or “solar wind” emissions.
In one such event, a solar wind blast in 1859 caused some of the world’s telegraph lines to burst into flames. In another, Quebec lost its power grid for a couple of days in 1989 because of a less severe solar storm, Fasel said.
“We’d like to be able to predict these storms so that we can power our spacecraft down so that we don’t lose the electrical circuits on our spacecraft,” said Fasel, who has been studying the aurora borealis since 1995.
“By being able to see more of the aurora, and to couple that with spacecraft data, it will give us a better understanding of the solar-terrestrial interaction,” Fasel said.
Press, who grew up in Mill Valley, is the middle child of five children of veterinarians Mary and Curtis Press of Belvedere. The family moved from Mill Valley to Belvedere when Press was 18, he said.
Chris Valdez, the principal at Marin Catholic, said the Press family was well known at the Kentfield high school.
“Jason Press is one of five Press siblings having attended Marin Catholic, each of them top academic students, but also balanced by activities in service, sports and music,” Valdez said.
“Creative endeavor is a family characteristic,” Valdez said. “It is no surprise that Jason is using his intellectual gifts and curiosity to not only explore this data recovery project, but that he is also using his creative ingenuity to develop tools to aid in the process.”