By Theo Schutt

Throughout the universe, two entities are playing an invisible cosmic game of tug-of-war: dark matter and dark energy. Their "rope" is the filamentary web of matter that weaves across the entire Universe (Figure 1). Dark matter makes up a majority of the web's structure and pulls all matter—both itself and the visible matter (galaxies and gas)—closer together through the force of gravity. On the other hand, dark energy, which drives the accelerated expansion of the Universe, stretches this web apart by expanding spacetime. Like game-day analysts, cosmologists have found myriad ways to dissect this match.

One promising tool is weak gravitational lensing. As light from a distant galaxy travels towards us, the gravity of the cosmic web bends its path, or "lenses," the image of the galaxy. In rare, dense regions, “strong” lensing distorts galaxies’ shapes into arcs. Most of the cosmic web is less dense and only causes "weak" (~1%) distortions. By correlating the shapes of neighboring galaxies, whose light passes through similar patches and is distorted in a similar way, we can map the cosmic web’s clumpiness, and thus, probe the nature of dark matter and dark energy.

Saved by the stars

Measuring these tiny distortions of galaxies on their own is already difficult. Now add the atmosphere, which causes the galaxies to shift and wobble as we’re taking measurements. It may seem impossible to figure out the true galaxy shape, but we have a trick up our sleeves to beat the tricks of the atmosphere: stars!

Stars are useful because they are very small; for almost all telescopes they are infinitesimally-small point sources of light. Thus, a star's image gives us a direct measurement of what astronomers call the "point-spread function" (PSF) — a description of how light is spread out in an image. Any spatial features, like the extended blurring or sharp "spikes" shown in Figure 2, are due to the atmosphere or optics, not from the star itself.

A star and a galaxy next to each other in an image are distorted similarly, so we can remove the distortion from the galaxy using information from the neighboring star! In this way, we can measure the galaxy's "real" shape, including the tiny distortion caused by lensing.

Modeling the PSF for the world's largest weak lensing survey

The Dark Energy Survey (DES) is a six-year survey conducted from 2013-2019 using the Dark Energy Camera (DECam) installed on the Victor M. Blanco Telescope at Cerro Tololo, Chile. For the past few years, I have worked on PSF modeling as part of weak lensing analysis using the full six years of DES data.

Of course, this isn't the first time we are looking at the data. Many KIPAC members have been heavily involved in DES, including in the most recent cosmology analysis using the first three years of data (see the KIPAC research highlights by Jamie McCullough, Jessie Muir and Justin Myles for more details on their work). While the Year 3 weak lensing analysis was very successful, some parts of the analysis could be improved for the full Year 6 analysis, including the PSF modeling. Because of inaccuracies in the PSF model, Year 3 was not able to use one of the color filters on the DECam, the g-band filter, equivalent to ignoring one-fifth of the images taken! I got the chance to tackle the challenge of recovering our g-band data for Year 6 as a big part of my PhD thesis.

We made two major changes to the PSF modeling with Year 6 data. One, we improved our algorithm for distinguishing between stars and galaxies. Year 3 PSF testing had shown that ~1-2% of the PSF "stars" were actually galaxies. Two, we introduced wavelength-dependence in our PSF model, to account for the stronger distortions at the bluest wavelengths (see Josh Meyers's KIPAC research highlight on this topic!), which are observed with the g-band filter. After verifying our new PSF model, we found that we could indeed use g-band data for weak lensing! The information from g-band helps us determine how distant the lensed galaxies are from Earth, a key component to the analysis, so this was a big win for the team.

Rubin on the horizon

Upcoming surveys like the Vera C. Rubin Observatory's Legacy Survey of Space and Time (LSST) will be leveraging the PSF software and analysis pipelines established for DES. While we are happy with our successful modeling for Year 6 analysis, the data volume from LSST—about 35 times more than DES!—will require even more precise PSF models to unlock the data's full potential to uncover information on dark matter and dark energy. Still, our work has placed us a major step forward, in time to tackle the exciting era ahead with Rubin/LSST!

Edited by Josephine Wong, Sanskriti Das, and Xinnan Du