Feature – Grid helps to filter LIGO’s data

Workers at LIGO adjust the optic suspensions in situ. The suspension system is one of the physical measures LIGO takes to isolate the detector from false signals such as seismic activity. Courtesy of LIGO Lab.

Tiny ripples in the fabric of space-time may provide scientists with a way to study cosmic processes that are invisible to optical telescopes, such as the collision of two black holes.

In his theory of general relativity, Einstein predicted that such ripples, called gravitational waves, would be created when a mass accelerates. However, gravitational waves are so small – about one thousand times smaller than a proton – that even the relatively large ones generated by massive astrophysical events are very difficult to detect.

The Laser Interferometer Gravitational Wave Observatory (LIGO), which has sites in Washington and Louisiana, uses lasers to search for these minute cosmic ripples that carry information about the motion of objects in the universe.

Analysis of data from LIGO’s detectors is very computationally intensive, however, and researchers depend on the LIGO Data Grid and Open Science Grid to process large amounts of data to look for signals of gravitational waves.

When gravitational waves pass through an object, the object’s length fluctuates by an extremely small amount. It is these tiny fluctuations that scientists search for using LIGO detectors.

However, many things – a truck on the highway, an earthquake, or even someone dropping a hammer nearby – can create signatures in LIGO’s detectors that appear similar to those of gravitational waves. That’s why researchers need to weed out false signals.

A closeup view of a 35-watt laser at LIGO. Courtesy of D. Shoemaker LIGO Lab

LIGO instruments collect roughly one terabyte of raw data each day that researchers must sift through. In one type of search, the data is broken into smaller segments, and each segment is compared to tens of thousands of computer-generated signatures to identify candidate signals – a process that requires hundreds to thousands of CPUs, said LIGO researcher Kent Blackburn.

A second type of search looks at weeks of data at a time, but requires researchers to account for a slight glitch caused by the motion of the earth and the object producing the waves. Correcting for this glitch and searching for signals in the data is extremely computationally intensive.

LIGO researchers use the grid to filter the data quickly enough to keep up with the instruments’ high rate of data production.

They also run mock data challenges where they blindly introduce a fake signal into the LIGO instruments to enhance their analysis techniques and verify that their process for picking out real signals works.

Although the LIGO Scientific Collaboration researchers have not yet directly detected gravitational waves, they can estimate the rate at which gravitational waves are generated based on the fact that they did not detect any with instruments of a given sensitivity over a given period of time.

“Scientists have built telescopes that can observe the universe using infrared, x-rays, and gamma rays, but these are all types of light,” Blackburn said. “By using gravitational waves, it’s like creating a whole new set of eyes to look at the universe. We’ll be able to see processes that don’t give off light that we’ve never been able to see before.”

—Amelia Williamson, for iSGTW