Teams from Fermilab and Berkeley Lab used galaxies from wide-ranging SDSS Stripe 82, a tiny detail of which is shown here, to plot new maps of dark matter based on the largest direct measurements of cosmic shear to date. Image credit: SDSS
Two teams of physicists at the U.S. Department of Energy’s Fermilab and Lawrence Berkeley National Laboratory have independently made the largest direct measurements of the invisible scaffolding of the universe, building maps of dark matter using new methods that, in turn, will remove key hurdles for understanding dark energy with ground-based telescopes.
The teams’ measurements look for tiny distortions in the images of distant galaxies, called "cosmic shear," caused by the gravitational influence of massive, invisible dark matter structures in the foreground. Accurately mapping out these dark-matter structures and their evolution over time is likely to be the most sensitive of the few tools available to physicists in their ongoing effort to understand the mysterious space-stretching effects of dark energy.
Both teams depended upon extensive databases of cosmic images collected by the Sloan Digital Sky Survey, which were compiled in large part with the help of Berkeley Lab and Fermilab.
“These results are very encouraging for future large sky surveys. The images produced lead to a picture of the galaxies in the universe that is about six times fainter, or further back in time, than is available from single images," says Huan Lin, a Fermilab physicist and member of the SDSS and the Dark Energy Survey.
Read the rest of the press release issued jointly by Fermi National Accelerator Laboratory and Lawrence Berkeley Lab.
Read more from Fermilab at Quantum Diaries.
Read more from Berkeley Lab below.
Seeing in the Dark
This week at the American Astronomical Association’s 219th meeting in Austin, Texas, teams from Fermilab and Berkeley Lab are both announcing maps of the distribution of dark matter in a big chunk of the universe, based on their independent analyses of an enormous collection of galaxies recorded by the Sloan Digital Sky Survey. Their methods demonstrate that it will be possible to achieve the required accuracy for a new understanding of the role and nature of dark energy in the evolution of the universe, using the next generation of large, ground-based surveys. What follows is an overview of the Berkeley Lab approach.
Three-quarters of the universe is invisible dark energy, and over 20 percent is invisible dark matter. The only hope of “seeing” this invisible stuff is to look at its effects on the matter and radiation we can see, namely galaxies.
In principle it’s possible to calculate the large-scale distribution of dark matter with precision, because it interacts with ordinary matter through gravity. The gravitational fields of clumps of dark matter act as lenses, bending the light from galaxies behind them. One effect is to flex the shapes of distant galaxies and distort how they appear from Earth – so-called cosmic shear.
Over large stretches of sky, gravitational lensing is weak and can only be calculated statistically. Nevertheless, cosmic shear makes if possible to build a map of the universe’s invisible dark matter. At Berkeley Lab and UC Berkeley, Eric Huff and his colleagues have compiled a “shape catalog” of millions of galaxies by measuring cosmic shear – in effect, mapping how dark matter is distributed across a wide equatorial stripe of the sky.
Huff’s team includes Christopher Hirata of Caltech, Rachel Mandelbaum of Princeton and Carnegie Mellon, Tim Eifler of Ohio State, Berkeley Lab’s David Schlegel, and Uroŝ Seljak of Berkeley Lab, UC Berkeley, Ehwa Womans University in Seoul, and the University of Zurich. To construct their map they used images of galaxies collected from 2000 to 2009 by Sloan Digital Sky Surveys I and II (SDSS I and II), using the 2.5-meter telescope at Apache Point Observatory in New Mexico. Updated calibrations come from SDSS III, which continues today.
The galaxies lie in SDSS Stripe 82, a continuous ribbon of sky stretching for over 100 degrees along the celestial equator, recorded in images captured in multiple passes over many years. Huff and his team processed the images and added them together to improve their signal-to-noise ratio. Stripe 82 covers more sky area than any other cosmic-shear survey yet completed, but more ambitious surveys are already in the works.
Weak lensing measures the presence of dark matter almost directly (or, as Huff and his colleagues put it, provides “the least indirect observation”), but indirectly it can do more. Clumps of dark matter not only distort the images of galaxies behind them, they determine how galaxies cluster around them. With redshift data from these galaxies it’s possible to trace how the distribution of matter in the universe has evolved over time.
Thus weak lensing gives insight not only into the nature of dark matter but dark energy as well, a method ultimately of great strength. Powerful ground-based sky surveys such as the Dark Energy Survey and the Large Synoptic Survey Telescope plan to exploit cosmic shear to build models of the cosmos whose fundamental goal is to track and explain the expansion history of the universe.
"The community has been building towards cosmic shear measurements for a number of years now," says Huff, "but there's also been some skepticism as to whether they can be done accurately enough to constrain dark energy. Showing that we can achieve the required accuracy with these pathfinding studies is important for the next generation of large surveys."
The technical challenges are formidable. Galaxies are randomly oriented to begin with, creating noise in any attempt to measure distortions of their shapes statistically. Much worse for ground-based telescopes are the shape distortions suffered just by peering through the changing atmosphere; in an image these can stretch over significant distances and mimic weak lensing.
There are many more sources of error. Huff and his colleagues sum them all up with the term “point-spread function,” at base a term describing how the imaging system itself distorts or blurs an image, but for Huff’s team also including optical effects, atmospheric effects, glitches in telescope motion, and other factors.
To attack point spread they used a rounding kernel, a mathematical technique custom-made to “homogenize” the point-spread function in every SDSS image. Not all fluctuations in image quality could be eliminated this way, so the team ruthlessly cut away potentially flawed data, rejecting both individual cases and entire image fields.
Layering photos of one area of sky taken at various time periods, a process called coaddition, can increase the sensitivity of the images six-fold, by removing errors and enhancing faint light signals. The image on the left shows a single picture of galaxies from SDSS Stripe 82. The image on the right shows the same area after layering, increasing the number of visible, distant galaxies. Image credit: SDSS
The processed images were then “coadded” to construct a rough-draft galaxy catalogue. Stars, galaxies whose shape could not be measured, and other spurious objects were thrown out. They were able to include 2 million galaxies along a ribbon of sky covering 150 square degrees.
Finally the team applied shape measurement techniques to determine the cosmic shear along SDSS Stripe 82, and rigorously tested their results. They showed that remaining errors in weak-lensing distortions of galaxy shapes, after their corrections to the point-spread function, were neglible compared to the actual cosmic shear signal itself. Further inaccuracies owed more to statistical than systematic (“real”) errors, and are likely to be corrected by the massive, dedicated cosmological surveys coming soon.
For these surveys, the point-spread function corrections developed by Huff and his team, and their colleagues at Fermilab, should prove a valuable tool for the next generation of weak-lensing surveys. In Huff’s estimation, however, perhaps the team’s major accomplishment was to push the SDSS data past its design limitations to achieve precise measures of the cosmos, a task it was never built to take on.
- Paul Preuss
