GALSBI-SPS: A stellar population synthesis-based galaxy population model for cosmology and galaxy evolution applications

¹ Universitäts-Sternwarte, Fakultät für Physik, Ludwig-Maximilians-Universität München, Scheinerstr. 1, 81679 München, Germany
² Institute for Particle Physics and Astrophysics, ETH Zurich, Wolfgang-Pauli-Strasse 27, CH-8093 Zurich, Switzerland
³ Excellence Cluster ORIGINS, Boltzmannstr. 2, 85748 Garching, Germany
⁴ ICRAR, The University of Western Australia, 7 Fairway, Crawley, WA 6009, Australia
⁵ Swiss Data Science Center, Paul Scherrer Institute, Forschungsstrasse 111, 5232 Villigen, Switzerland

^⋆ Corresponding author: luca.tortorelli@physik.lmu.de

Received: 31 May 2025
Accepted: 21 September 2025

Abstract

Context. Next-generation photometric and spectroscopic galaxy surveys will enable unprecedented tests of the concordance cosmological model and of galaxy formation and evolution. Fully exploiting their potential requires a precise understanding of the selection effects on galaxies and biases on measurements of their properties, which are required, above all, for accurate estimates of redshift distributions. The forward-modelling of galaxy surveys offers a powerful framework to simultaneously recover galaxy redshift distributions and characterise the observed galaxy population.

Aims. We present GALSBI-SPS, a new stellar population synthesis (SPS)-based galaxy population model developed for cosmological and galaxy evolution studies. The model generates realistic galaxy catalogues, which we use to forward-model Hyper-Suprime Cam (HSC) observations in the COSMOS field.

Methods. GALSBI-SPS samples the physical properties of galaxies from analytical parametrisations informed by GAMA, DEVILS, and literature data, it computes galaxy magnitudes with the generative SED package PROSPECT, and it simulates HSC images in the COSMOS field with UFig. We measured photometric properties consistently in real data and simulations. We compared redshift distributions and photometric and physical properties to observations and to those from the phenomenological GALSBI model.

Results. GALSBI-SPS reproduces the observed g, r, i, z, y magnitude, colour, and size distributions down to i ≤ 23 with good accuracy. Median differences in magnitudes and colours remain below 0.14 mag, with the model covering the full colour space spanned by HSC data. Galaxy sizes are overestimated by ∼0.2″ on average and some tension exists in the g − r colour distribution, but the latter is comparable to that seen in the phenomenological GALSBI model. Redshift distributions show a mild positive offset (0.01 ≲ Δ¯z ≲ 0.08) in the mean. GALSBI-SPS qualitatively reproduces the stellar mass–star formation rate and size–stellar mass relations seen in COSMOS2020 data.

Conclusions. GALSBI-SPS provides a realistic, survey-independent description of the galaxy population at a Stage-III-like depth using only literature-based parameters. Its predictive power is expected to improve significantly when constrained against deep observed data using simulation-based inference, thereby providing accurate redshift distributions that satisfy the stringent requirements set by Stage IV surveys.