Our understanding of the physical properties of galaxies could be wrong

Up until recently, astronomy was reliant entirely on electromagnetic waves. While that changed with the confirmation of gravitational waves in 2016, astronomers had developed fundamental frameworks in the electromagnetic spectrum by that point.

One critical framework broke the spectrum into three categories based on their wavelength—infrared, optical, and ultraviolet. To astronomers, each of these categories was created by a different physical phenomenon, and monitoring each gave its insight into what that phenomenon was doing, no matter what the other spectra said.

This was especially prevalent when researching galaxies, as infrared and optical wavelengths were used to analyze different aspects of galaxy formation and behavior. However, Christian Kragh Jespersen of Princeton’s Department of Astrophysics and his colleagues think they have found a secret that breaks the entire electromagnetic framework—the optical and infrared are connected.

That is the simple title of a new paper the researchers released on the arXiv preprint server that describes their remarkable feat. They predicted the infrared values of galaxies surveyed by the Wide-Field Infrared Explorer (WISE) by using optical data collected by the Sloan Digital Sky Survey (SDSS).

There is an assumption underlying most galaxy astronomy—that different “components” of a galaxy (i.e., its super-massive black hole or the stars on the edge of its spiral arms) are “separable” from one another since they emit across different wavelengths. This separation is simplified into a value called a spectral energy distribution, or SED, fitting code. They are commonly used to characterize the physical properties of the galaxies they are used to describe.

Unfortunately, according to the paper’s analysis, the assumption underlying the SED codes of existing galaxies—that of the “separability” of galaxy components—is fundamentally wrong. As the paper says, “We conclude that [IR emitting and optical emitting] processes must be strongly coupled. This may or may not surprise the reader, but it violates the assumptions of current SED models.”

To prove their point, the researchers looked at data collected by the SDSS and WISE. They collected data on over 500,000 different galaxies and then, after some training and validation of an algorithm, used the optical data collected by SDSS to predict the infrared data collected from WISE for each individual galaxy. That process was made easier since there were already indexes tying the WISE and SDSS data together on a per-galaxy basis.

The results were striking—the algorithm could predict the infrared value based solely on the optical input with very little noise. To further prove their point, the authors utilized two SED coding tools, CIGALE and prospector. Both missed the mark widely when attempting to calculate the correct value, allowing the authors to label a series of graphs showing how bad their estimations were with “Overconfident and biased.”

On the other hand, their own data matched up precisely with the observed data from the WISE database. To the author’s credit, they point out some weak points in their argument, such as the fact that WISE and SDSS were taken with different aperture sizes, which could influence the fit of their algorithm. However, the overwhelming preponderance of evidence points to a simple conclusion—that the “separability” assumption that underpins much of our understanding of the physical properties of galaxies is wrong.

The paper is only in preprint on arXiv and hasn’t been accepted in a peer-reviewed journal yet. But if it is, it seems poised to impact our understanding of our universe fundamentally and, more importantly, what frameworks we use to understand it.