BOOMERanG Analysis Finds Flat Universe
December 12, 1999
Newly released data from the 1997 North American test flight of BOOMERanG, which mapped anisotropies in the cosmic microwave background radiation (CMB) in a narrow strip of sky, show a pronounced peak in the CMB "power spectrum" at an angular scale of about one degree, strong evidence that the universe is flat.
Analyzed at the Department of Energy's National Energy Research Scientific Computing Center (NERSC) at Lawrence Berkeley National Laboratory, the new data also suggest the existence of a cosmological constant, a form of countergravitational "dark energy" thought to fill the universe.
BOOMERanG stands for "balloon observations of millimetric extragalactic radiation and geophysics." The collaboration includes over two dozen researchers from seven countries; its principal investigators are Andrew Lange of the California Institute of Technology and Paolo de Bernardis of the University of Rome, "La Sapienza."
Phillip Mauskopf of the University of Massachusetts is first author of a letter to Astrophysical Journal Letters announcing the measurement of the power-spectrum peak. In a second article, the BOOMERanG collaboration presents new, independent limits on the magnitude of any cosmological constant, a property of space that offsets gravitational attraction.
Albert Einstein proposed the cosmological constant in 1917 but later retracted the idea. It was dramatically resurrected in 1998 when the international Supernova Cosmology Project based at Berkeley Lab, along with the High-Z Supernova Search Team centered in Australia, discovered that the expansion of the universe is not slowing down as it would be if gravity were not offset, but instead is speeding up.
The BOOMERanG data were acquired in August, 1997, during a six-hour flight from National Aeronautics and Space Administration's National Scientific Balloon Facility in Palestine, Texas. The data were analyzed by NERSC's Julian Borrill, who worked closely with colleagues at the National Science Foundation's Center for Particle Astrophysics, located at the University of California at Berkeley, and at the Canadian Institute of Theoretical Astrophysics at the University of Toronto.
Borrill notes that the BOOMERanG North America data set was so large — a partial map of the sky covering more than 200 square degrees and containing some 26,000 pixels — that a one-gigahertz serial processor would have required three months of continuous operation to extract the power spectrum, which is a measure of the structure of CMB anisotropies.
By employing the parallel processing power of NERSC's "Mcurie" system (a Cray T3E supercomputer) and using the MADCAP software package he developed at NERSC, Borrill was able to shorten the running time to a matter of hours. MADCAP is short for the "microwave anisotropy dataset computational analysis package."
Big as it is, Borrill says, the BOOMERanG North America data is only the first in a parade of data sets "of unprecedented quality and ever-increasing size." The MAXIMA 1 and MAXIMA 2 balloon flights have already produced substantially more data; their analysis is not complete.
Late last year BOOMERanG LDB (for "long duration ballooning") completely circled the South Pole in ten and a half days, yielding a map of virtually the whole sky visible from the pole and a data set, whose analysis is still underway, that is more than 17 times as large as the test flight's. Satellites to be launched early next century will produce more data by orders of magnitude; the European Space Agency's PLANCK will map the sky in 10,000,000 pixels.
"The memory required to process these CMB experiments increases as the square of the number of pixels, and the time increases as the cube," Borrill says. "Without new computers and new computational strategies, we won't be able to derive the detailed measurements of basic cosmological parameters the CMB experiments are designed to reveal."
Borrill adds that "because of the power of our parallel machines and the depth of our experience with cosmic microwave background studies, NERSC is becoming the computing center of choice for analyzing CMB data from experiments all over the world. We want to maintain that status, but it will take hard work and fresh ideas."