In the upcoming years, two U.S.-funded telescopes will make these observations possible. Ahead of their operational debut, a Duke researcher is spearheading an initiative under the OpenUniverse project to deliver the most accurate simulation of what these telescopes will reveal.

In his office on Duke's West Campus, Duke physics professor Michael Troxel showcases a preview of the Vera C. Rubin Observatory's capabilities, set to commence operations in 2025, and NASA's Nancy Grace Roman Space Telescope, launching by 2027.

The Rubin Observatory, funded jointly by the National Science Foundation and the Department of Energy, will employ a massive 8.4-meter telescope on a Chilean mountaintop to capture data across nearly half of the sky. Concurrently, NASA's Roman telescope will orbit a million miles from Earth, peering close to the edge of the observable universe to identify faint and distant objects with unprecedented clarity.

By integrating observations from these two telescopes, scientists aim to unravel mysteries like the universe's accelerating expansion. However, merging their data presents significant technical challenges.

"Once they're up and running, the two telescopes will produce unprecedented amounts of data," Troxel says. Astronomers expect a combined total of 80 petabytes over the instruments' lifetimes. "That's three times the digital footprint of the U.S. Library of Congress."

To prepare for this data influx, Troxel leads an effort to recreate what the telescopes will see, galaxy by galaxy, simulating the actual data as closely as possible. This allows scientists to explore and analyze it similarly to real data. With a few mouse clicks, Troxel displays a simulated image combining data from a small section of the sky.

The image represents a minimal portion of space-ten such images would barely cover the full moon. The actual surveys will be vastly larger. Nonetheless, the image contains about 80,000 galaxies and other objects.

Some of them are so faint and far away -- up to 20 billion light years -- that they're "hard to tell from a speck of dust on my screen," Troxel says.

Each dot or speck of light signifies a distant galaxy. These galaxies, instead of being made of gas and dust, exist as computer code within a virtual universe in the cloud.

Troxel notes that the images are so realistic, "that even experts can't always tell at first glance whether they contain simulated telescope data or the real thing."

Creating such authentic simulations is a monumental computational task. It required simulating the light from each star and galaxy and tracking its journey through space over billions of years to the telescopes.

The team used a supercomputer capable of performing thousands of trillions of calculations per second. Utilizing the now-retired Theta cluster at Argonne National Lab in Illinois, they generated around four million simulated images of the cosmos in nine days-a task that would have taken 300 years on a standard laptop.

The project consumed more than 55 million CPU hours and over half a year of work by dozens of experts, along with the coordination of approximately 1300 researchers across multiple cosmology teams, Troxel says.

Currently, they have released a 10-terabyte subset of the complete 400-terabyte dataset, with the remainder to be processed and released this fall.

Researchers will use these simulated images to conduct a dress rehearsal, testing new methods and algorithms they will apply once they have access to the actual data.

By resolving issues in simulation beforehand, they can efficiently handle the data influx when it arrives.

Research Report:OpenUniverse 2024 Simulated Roman and Rubin Images: Preview

Related Links
Vera C. Rubin Observatory
Understanding Time and Space