NEWS RELEASE, 3/16/99
UC Berkeley team challenges cosmologists to analyze
simulation of cosmic microwave background - and win a case of champagne
By Robert Sanders, Public Affairs
BERKELEY-- Scientists love a challenge, and now a group at the University of California, Berkeley, has one for the cosmologists: try to find the primordial seeds of the universe amidst simulated noise from the cosmos.
The challenge, thrown down March 15, is designed to prepare scientists for the flood of data that will be sent back after the launch late next year of the Microwave Anisotropy Probe (MAP), a National Aeronautics and Space Administration satellite that will measure the microwave radiation left over from the big bang.
The satellite will produce a microwave map of the universe much more detailed than its predecessor, the Cosmic Background Explorer (COBE) satellite, which gave scientists in 1992 their first evidence of clumps and wrinkles in the very early universe. These clumps of matter and energy presumably evolved into the clusters of galaxies we see today.
The UC Berkeley group, which calls itself WOMBAT, for Wavelength-Oriented Microwave Background Analysis Team, has produced a detailed simulation of what MAP will see, complete with all the confounding contamination and noise of the universe.
The teams that can pull the picture out of the noise will win a case of champagne or a toy stuffed animal - a wombat.
"We want to see if the analysis methods people have produced will really work," said Eric Gawiser, a graduate student in physics at UC Berkeley. "The idea is to get physicists to think about how they will deal with the uncertainties in the data."
In fact, when the conclusion of the COBE team was first announced in 1992, some scientists and statisticians attacked the data analysis as too weak to support their claims. Subsequent and more thorough analysis of the data convinced most scientists that the minute fluctuations seen in the cosmic microwave background are real.
"We are trying to go through a process now that will create faith in the final results," Gawiser said.
Joining Gawiser on the WOMBAT team are Doug Finkbeiner, also a graduate student in physics, Andrew Jaffe, a research physicist at UC Berkeley's Center for Particle Astrophysics, and a number of other UC Berkeley researchers working under the supervision of professors Joseph Silk, Marc Davis and George Smoot of the departments of physics and astronomy.
Jaffe is deeply involved in analysis of data from a smaller balloon-borne experiment that is measuring the cosmic microwave background one small patch of sky at a time. Called the Millimeter Anisotropy eXperiment IMaging Array (MAXIMA), this NASA-funded UC Berkeley experiment first flew last summer out of Palestine, Texas, and will fly again this summer.
WOMBAT's challenge also includes simulations of what MAXIMA will see, which is 50 times more detailed than what COBE measured, though in a piece of sky only 10 degrees on a side.
The cosmic microwave background is the radiation that escaped from the fireball of the big bang about 300,000 years after the birth of the universe, when the growing cloud of matter and energy finally became transparent enough for light to wiggle out. In the 14 to 15 billion years since then, the radiation has cooled from several thousand degrees to a frigid 2.7 Kelvin, equivalent to 455 degrees below zero Fahrenheit.
COBE was set up to look for structure in the apparently even glow of the microwave radiation and discovered fluctuations of about one part in 100,000 - that is, some areas were about 0.00002 degrees warmer or colder than others. Smoot, at the time a researcher at Lawrence Berkeley National Laboratory and now also a professor of physics at UC Berkeley, was the principal investigator on that experiment.
The MAP satellite will scan the entire sky with 30 times greater resolution than COBE, yielding much finer detail about the early universe. Physicists and astronomers hope to glean from the structure of the microwave fluctuations some important details of the universe: its age and shape, how fast it is expanding, when and how the first galaxies formed, and whether it will expand forever.
Though the COBE data were tough to analyze, the MAP data will be even tougher, according to the WOMBAT team. Microwave emission by distant galaxies will confuse the picture much more, as will microwave emission from the dust of our own galaxy.
The WOMBAT simulation was constructed by Gawiser, Finkbeiner and Jaffe based on their estimates of foreground contamination by dust and galaxies and clusters of galaxies that emit microwave radiation in the same wavelength range (around 90 gigahertz) as the microwave background.
In fact, Finkbeiner, Gawiser and Jaffe have put together the most realistic full-sky maps so far of all the major microwave foreground components, with estimated uncertainties. This is essential to any analysis of the microwave background.
To create the simulation they also threw in instrument noise and a few other complications, including a recently proposed source - microwave emission by spinning grains of dust.
"The point of the challenge is to see if, given a signal screwed up with small sources and noise, physicists can reconstruct the signal that made it," Jaffe said.
The UC Berkeley team expects perhaps six to eight groups around the world to take up the challenge. Among them, they hope, will be the large NASA team preparing for the MAP satellite mission and the team working on an even more advanced satellite called the Planck Surveyor. Planck is being built by the European Space Agency and will be launched in 2007.
Distinguishing between the cosmic microwave background radiation and other foreground sources is possible because of their different microwave spectra, Gawiser said. The background radiation peaks at the center of the microwave band and is analogous to green in the visible light spectrum of the rainbow. Radio emission is low-frequency and analogous to red in the spectrum, while thermal dust emission is high-frequency, analogous to blue.
All of these components are very noisy, however, and trying to subtract three forms of noise to get to the underlying noise of the density variations in the early universe should pose quite a challenge, he said, even for some of the world's most resourceful astrophysicists.
While several methods for eliminating this foreground emission have been proposed by researchers around the world, the WOMBAT challenge provides the first chance to test those analysis techniques on realistic full-sky simulations.
The WOMBAT group is offering two prizes to those who respond by July 15: a case of champagne for the team that most successfully determines the density, age and rate of expansion of each simulated universe from its cosmic microwave background radiation; and a stuffed toy wombat for the team that succeeds in the opposite task - eliminating the microwave background to learn about the microwave emission of our own galaxy and other galaxies.
The team is offering five simulations of the MAP sky, as well as five simulations of the small patches of sky that MAXIMA would see.
Each MAP simulation covers the sky with 12 million pixels, each of which contains information on intensity and frequency. Though Gawiser and Finkbeiner simulated the sky maps on a desktop workstation, the best analysis will require a supercomputer, they said.
"In all its glory, the problem is so hard that our largest dataset to date - 400,000 pixels from the Boomerang balloon flight - is already too big for supercomputers to handle," Jaffe said. "In the future, we will just have to become more clever."
Send comments to: email@example.com