MIT team finds that the ratio of component atoms is vital to performance.
Jennifer Mills pulled an all-nighter after the first fragment of comet Shoemaker-Levy 9 hit Jupiter last month. Why? The physics undergraduate was busy processing Hubble Space Telescope images of the collision in time for a NASA press conference the next morning.
Ms. Mills' involvement in the once-in-a-lifetime event began last spring when she started a UROP project on Neptune working with Dr. Heidi Hammel, a principal research scientist in the Department of Earth, Atmospheric and Planetary Sciences.
Neptune, however, was soon forgotten. NASA asked Dr. Hammel to be a team leader for Hubble observations of the impending crash, and "Jupiter started taking over Heidi's life," Ms. Mills said. As a result, Ms. Mills' work followed suit and the junior joined the 12-member Hammel team responsible for pictures of how Jupiter's atmosphere responded to the collision.
Preparations included creating a computer program to convert the raw data sent down from Hubble into clear images for analysis (among other things the program removed small scan lines). "NASA had such a program, but it took two days to get images out of it and that was unacceptable for this event," Dr. Hammel said. "Normally the time scale doesn't matter-if you're looking at a galaxy, you could take two days or two years to process the image and it wouldn't matter. But we knew that the comet fragments would be hitting on a scale of hours."
The Hammel team and colleagues eventually created a program, dubbed the calibration pipeline, that allowed them to process images in about seven minutes. Ms. Mills was more than instrumental to the effort: she wrote half the software. And once the comet fragments hit, the program proved invaluable. The team received and processed some 40 to 50 images a day.
The week of the collision was "pretty hectic," said Ms. Mills, who monitored the event from the Space Telescope Science Institute (STScI) in Baltimore with Dr. Hammel and three other team members. There the five scientists spent 15-hour days sending images through the calibration pipeline, analyzing them, selecting the best for distribution to the press, and more.
One of the most exciting moments for Ms. Mills was after the first fragment hit and Hubble sent back images of a plume of material rising above the edge of Jupiter. She recalled what happened when the first images of the plume started coming in: "I was standing outside the [control] room because they couldn't let everyone in, and the people inside started jumping up and down," she said. "It was kind of maddening-those of us outside all said `WHAT. what's happening?"
And then Ms. Mills' all-nighter began. "We had these beautiful images [of the plume] and the press and public wanted to see them," she explained. But although the team had developed a few computer programs to turn out enlarged images of expected phenomena like the impact sites immediately, the scientists hadn't expected the plume. So "we hadn't made any sort of [computer] template that could involve them," Ms. Mills said. As a result she spent the night preparing a photographic press release of the plume that showed, via a sequence of images, how it "rose, expanded, then fell into a flat pancake."
Ms. Mills' contributions to the Hammel team were critical, said Dr. Hammel. "We could not have accomplished what we did without Jennifer there" (at STScI). For her part, Ms. Mills said: "It's not something I'd want to do every day for the rest of my life-it was really stressful-but it was lots of fun.
"For a once-in-a-lifetime experience you couldn't beat it."
A version of this article appeared in the August 31, 1994 issue of MIT Tech Talk (Volume 39, Number 3).