Debugging Galaxy Evolution with L-GALAXIES

February 01, 2025

The formation and evolution of galaxies are among the most complex challenges in astrophysics. Recent advancements with instruments like JWST and ALMA have shed light on high-redshift galaxies – those that existed billions of years ago. However, most theoretical models are tuned to match galaxies in the local universe. Researchers from the Max Planck Institute for Astrophysics and the University of Bonn now comprehensively evaluated the Munich semi-analytical model L-GALAXIES using the latest observations and found that while the model aligns well with the properties of local galaxies, it struggles with key aspects of high-redshift galaxies. Particularly, the study highlights critical issues with the model’s predictions of quenched galaxies, those that have ceased star formation. Their results suggest a need to revise the implementation of processes driving star formation quenching, including supermassive black hole feedback and galaxy mergers.
 

Observations from surveys such as SDSS, CANDELS, and COSMOS provide essential insights into galaxy properties and scaling relations. However, to uncover the underlying processes driving galaxy evolution, astronomers need to simulate the relevant astrophysical phenomena. The Munich semi-analytical model, L-GALAXIES, offers a self-consistent framework for tackling these challenges. Over the past three decades, L-GALAXIES has undergone continuous development, primarily at the Max Planck Institute for Astrophysics (MPA) in collaboration with international teams, establishing itself as a corner-stone tool for studying galaxy evolution. The model strikes a balance between computational efficiency and detailed physical modelling, making it a powerful complement to computationally demanding hydro-dynamical simulations.

The L-GALAXIES model builds upon its previous generation with a series of advancements that are motivated both by new observational data and a resulting deeper physical understanding of complex processes such as gas accretion and cooling, star formation, chemical enrichment, and stellar and black hole feedback. The most recent versions incorporate advanced environmental mechanisms like ram-pressure and tidal stripping. The models are calibrated using Monte Carlo Markov Chain (MCMC) techniques and constrained by low-redshift observational data. Together, these updates and calibrations represent the cutting edge of semi-analytical galaxy formation modelling.

Recent observational campaigns, particularly those utilizing advanced ground- and space-based instruments such as the Hubble Space Telescope (HST), the Atacama Large Millimeter/submillimeter Array (ALMA), and the James Webb Space Telescope (JWST), have provided unprecedented insights into the evolution of high-redshift galaxies. These observations reveal the size, compactness, and abundance of quenched galaxies at redshifts around z=2 (when the universe was just 3 billion years old) and beyond, offering a unique opportunity to rigorously test L-GALAXIES predictions well outside its original calibration regime. In particular, they are identifying areas where the model aligns with or deviates from observed trends, providing crucial guidance for improving its treatment of high-redshift galaxy populations.

The current study evaluates the latest version of L-GALAXIES alongside its two preceding iterations, focusing on their ability to reproduce the evolution of galaxy number density, size, and surface density across cosmic time. The analysis spans the history of the universe, from 500 million years after the Big Bang to the present day (~13.5 billion years later), with a specific focus on the first few billion years. It marks the first comprehensive comparison of L-GALAXIES predictions to high-redshift observations.

Galaxies were classified as star-forming or quenched based on their near-ultraviolet (NUV) to near-infrared (J-band) color. Sizes and surface densities were determined using methodologies consistent with observational studies. Additionally, X-ray data from instruments such as Chandra and XMM-Newton, along with microwave and longer wavelength data from Planck, were incorporated to examine baryon and gas distributions within host halos, shedding light on the interaction between baryonic matter and galaxy processes.

Although the model shows significant agreement with the properties of star-forming galaxies at both low and high redshifts, the study highlights significant discrepancies in the model’s predictions for quenched galaxies, particularly for Milky Way-mass and more massive systems at the times when the Universe was younger than 2 billion years old. The model underestimates the abundance of quenched galaxies by a factor of 60 and over-predicts the fraction of baryonic matter within galaxy clusters by around 15-20%. Moreover, the predicted sizes of galaxies are several times larger than observed, pointing to deficiencies in the modelling of star formation suppression mechanisms such as active galactic nucleus (AGN) feedback and galaxy mergers.

While L-GALAXIES successfully reproduces the stellar mass functions and cosmic star formation rates across a wide range of cosmic times, it fails to capture the size-mass relations and surface density trends of quenched galaxies at high redshifts. These findings suggest that the model’s simplified treatment of galaxy mergers and quenching mechanisms limits its ability to accurately predict the abundance and compactness of quenched galaxies.

Addressing these limitations will require recalibration efforts informed by data from JWST and other advanced observations. More critically, the model needs a thorough revision of its implementation of key physical processes relevant to galaxy quenching, such as AGN feedback and galaxy merger events. Additionally, leveraging state-of-the-art machine learning techniques to navigate the extensive parameter space offers a promising avenue for refining the model’s physical prescriptions and enhancing its predictive accuracy. These efforts will not only improve our understanding of galaxy evolution but also lay the groundwork for the development of next-generation L-GALAXIES models.

 

This interactive plot allows users to explore the formation of a massive galaxy cluster within a 20 Mpc cube. Users can zoom in, rotate the view, adjust the time slider to simulate the universe evolution, and toggle between star-forming and quenched galaxies at various cosmic epochs. Hovering over a data point reveals detailed information about each galaxy, including its mass and star formation rate, with larger markers representing more massive galaxies.

 

Other Interesting Articles

Go to Editor View