Assembly History Of Galaxies Research Articles

Unveiling the (in)consistencies among the galaxy stellar mass function, star formation histories, satellite abundances, and intracluster light from a semi-empirical perspective

ABSTRACT In a hierarchical, dark matter-dominated Universe, stellar mass functions (SMFs), galaxy merger rates, star formation histories (SFHs), satellite abundances, and intracluster light (ICL), should all be intimately connected observables. However, the systematics affecting observations still prevent universal and uniform measurements of, for example, the SMF and the SFHs, inevitably preventing theoretical models to compare with multiple data sets robustly and simultaneously. We here present our holistic semi-empirical model decode (Discrete statistical sEmi-empiriCal mODEl) that converts via abundance matching dark matter merger trees into galaxy assembly histories, using different SMFs in input and predicting all other observables in output in a fully data-driven and self-consistent fashion with minimal assumptions. We find that: (1) weakly evolving or nearly constant SMFs below the knee ($M_\star \lesssim 10^{11} \, \mathrm{M}_\odot$) are the best suited to generate SFHs aligned with those inferred from MaNGA, SDSS, GAMA, and, more recently, JWST; (2) the evolution of satellites after infall only affects the satellite abundances and SFHs of massive central galaxies but not their merger histories; (3) the resulting SFR–$M_\star$ relation is lower in normalization by a factor of $\sim 2$ with respect to observations, with a flattening at high masses more pronounced in the presence of mergers; (4) the latest data on ICL can be reproduced if mass-loss from mergers is included in the models. Our findings are pivotal in acting as pathfinder to test the self-consistency of the high-quality data from, e.g. JWST and Euclid.

Characterizing tidal features around galaxies in cosmological simulations

ABSTRACT Tidal features provide signatures of recent mergers and offer a unique insight into the assembly history of galaxies. The Vera C. Rubin Observatory’s Legacy Survey of Space and Time (LSST) will enable an unprecedentedly large survey of tidal features around millions of galaxies. To decipher the contributions of mergers to galaxy evolution it will be necessary to compare the observed tidal features with theoretical predictions. Therefore, we use cosmological hydrodynamical simulations NewHorizon, eagle, IllustrisTNG, and Magneticum to produce LSST-like mock images of z ∼ 0 galaxies (z ∼ 0.2 for NewHorizon) with $M_{\scriptstyle \star ,\text{ 30 pkpc}}\ge 10^{9.5}$ M$_{\scriptstyle \odot }$. We perform a visual classification to identify tidal features and classify their morphology. We find broadly good agreement between the simulations regarding their overall tidal feature fractions: $f_{{\small NewHorizon}}=0.40\pm 0.06$, $f_{{\small EAGLE}}=0.37\pm 0.01$, $f_{{\small TNG}}=0.32\pm 0.01$, and $f_{{\small Magneticum}}=0.32\pm 0.01$, and their specific tidal feature fractions. Furthermore, we find excellent agreement regarding the trends of tidal feature fraction with stellar and halo mass. All simulations agree in predicting that the majority of central galaxies of groups and clusters exhibit at least one tidal feature, while the satellite members rarely show such features. This agreement suggests that gravity is the primary driver of the occurrence of visually identifiable tidal features in cosmological simulations, rather than subgrid physics or hydrodynamics. All predictions can be verified directly with LSST observations.

On the origin of globular clusters in a hierarchical universe

ABSTRACT We present an end-to-end description of the formation of globular clusters (GCs) combining a treatment for their formation and dynamical evolution within galaxy haloes with a state-of-the-art semi-analytic simulation of galaxy formation. Our approach allows us to obtain exquisite statistics to study the effect of the environment and assembly history of galaxies, while still allowing a very efficient exploration of the parameter space. Our reference model, including both efficient cluster disruption during galaxy mergers and dynamical friction of GCs within the galactic potential, accurately reproduces the observed correlation between the total mass in GCs and the parent halo mass. A deviation from linearity is predicted at low-halo masses, which is driven by a strong dependence on morphological type: bulge-dominated galaxies tend to host larger masses of GCs than their later-type counterparts. While the significance of the difference might be affected by resolution at the lowest halo masses considered, this is a robust prediction of our model and a natural consequence of the assumption that cluster migration into the halo is triggered by galaxy mergers. Our model requires an environmental dependence of GC radii to reproduce the observed low-mass mass distribution of GCs in our Galaxy. At GC masses $\gt 10^6\, {\rm M}_\odot$, our model predicts fewer GCs than observed, due to an overly aggressive treatment of dynamical friction. Our model reproduces well the metallicity distribution measured for Galactic GCs, even though we predict systematically younger GCs than observed. We argue that this adds further evidence for an anomalously early formation of the stars in our Galaxy.

Correction to: Globular clusters contribute to the nuclear star clusters and galaxy centre γ-ray excess, moderated by galaxy assembly history

Correction to: Globular clusters contribute to the nuclear star clusters and galaxy centre γ-ray excess, moderated by galaxy assembly history

To be, or not to be: Balmer breaks in high-z galaxies with JWST

ABSTRACT Standard models of structure formation allow us to predict the cosmic timescales relevant for the onset of star formation and the assembly history of galaxies at high redshifts (z > 10). The strength of the Balmer break represents a well-known diagnostic of the age and star formation history of galaxies, which enables us to compare observations with contemporary simulations – thus shedding light on the predictive power of our current models of star formation in the early Universe. Here, we measure the Balmer break strength for 23 spectroscopically confirmed galaxies at redshifts 6 ≲ z ≲ 12 using public JWST NIRSpec data from the cycle 1 GO 1433 and GO 2282 programmes (PI Coe), as well as public spectroscopic data from the JWST Deep Extragalactic Survey (JADES). We find that the range of observed Balmer break strengths agree well with that of current simulations given our measurement uncertainties. No cases of anomalously strong Balmer breaks are detected, and therefore no severe departures from the predictions of contemporary models of star formation. However, there are indications of a number of outliers in the observed distribution which have weaker Balmer breaks than predicted by simulations.

Using the Two-point Correlation Function to Understand the Assembly Histories of Milky Way–like Galaxies

The two-point correlation function (2PCF) is a powerful statistical tool to measure galaxy clustering. Although 2PCF has also been used to study the clustering of stars on subparsec to kiloparsec scales, its physical implication is not clear. In this study, we use the Illustris-TNG50 simulation to study the connection between the 2PCF of accreted halo stars and the assembly histories of Milky Way-mass galaxies. We find, in general, that the 2PCF signal increases with the increase in galactocentric radii, r, and with the decrease in the pair separations. Galaxies that assemble late on average have stronger 2PCF signals. With z 1/4, z 1/2, and z 3/4 defined as the redshifts when galaxies accreted one-fourth, half, and three-fourths of their ex situ stellar mass today, we find they all show the strongest correlations with the 2PCF signals at r < ∼ 0.2R 200. z 3/4 shows the strongest correlations than those of z 1/4 or z 1/2. However, the correlations have large scatters. The 2PCFs in the velocity space show weaker correlations with the galaxy formation times within ∼0.35R 200 than real-space 2PCFs, and the scatter is considerably large. Both the real- and velocity-space 2PCFs correlate with the assembly histories of the host dark matter halos as well. Within 0.3R 200, the real-space 2PCF shows stronger correlations with the galaxy formation histories than with the halo formation histories. We conclude that it is difficult to use 2PCF alone to precisely predict the formation times or assembly histories of galaxies.

Globular clusters contribute to the nuclear star clusters and galaxy centre γ-ray excess, moderated by galaxy assembly history

ABSTRACT Two unresolved questions at galaxy centres, namely the formation of the nuclear star cluster (NSC) and the origin of the γ-ray excess in the Milky Way (MW) and Andromeda (M31), are both related to the formation and evolution of globular clusters (GCs). They migrate towards the galaxy centre due to dynamical friction, and get tidally disrupted to release the stellar mass content including millisecond pulsars (MSPs), which contribute to the NSC and γ-ray excess. In this study, we propose a semi-analytical model of GC formation and evolution that utilizes the Illustris cosmological simulation to accurately capture the formation epochs of GCs and simulate their subsequent evolution. Our analysis confirms that our GC properties at z = 0 are consistent with observations, and our model naturally explains the formation of a massive NSC in a galaxy similar to the MW and M31. We also find a remarkable similarity in our model prediction with the γ-ray excess signal in the MW. However, our predictions fall short by approximately an order of magnitude in M31, indicating distinct origins for the two γ-ray excesses. Meanwhile, we utilize the catalogue of Illustris haloes to investigate the influence of galaxy assembly history. We find that the earlier a galaxy is assembled, the heavier and spatially more concentrated its GC system behaves at z = 0. This results in a larger NSC mass and brighter γ-ray emission from deposited MSPs.

Catalogue of model star clusters in the Milky Way and M31 galaxies

ABSTRACT Detailed understanding of the formation and evolution of globular clusters (GCs) has been recently advanced through a combination of numerical simulations and analytical models. We employ a state-of-the-art model to create a comprehensive catalogue of simulated clusters in three Milky Way (MW) and three Andromeda (M31) analogue galaxies. Our catalogue aims to connect the chemical and kinematic properties of GCs to the assembly histories of their host galaxies. We apply the model to a selected sample of simulated galaxies that closely match the virial mass, circular velocity profile, and defining assembly events of the MW and M31. The resulting catalogue has been calibrated to successfully reproduce key characteristics of the observed GC systems, including total cluster mass, mass function, metallicity distribution, radial profile, and velocity dispersion. We find that clusters in M31 span a wider range of age and metallicity, relative to the MW, possibly due to M31’s recent major merger. Such a merger also heated up the in-situ GC population to higher orbital energy and introduced a large number of ex-situ clusters at large radii. Understanding the impacts of galaxy mergers and accretion on the GC populations is crucial for uncovering the galaxy assembly histories.

Siena Galaxy Atlas 2020

We present the 2020 version of the Siena Galaxy Atlas (SGA-2020), a multiwavelength optical and infrared imaging atlas of 383,620 nearby galaxies. The SGA-2020 uses optical grz imaging over ≈20,000 deg2 from the Dark Energy Spectroscopic Instrument (DESI) Legacy Imaging Surveys Data Release 9 and infrared imaging in four bands (spanning 3.4–22 μm) from the 6 year unWISE coadds; it is more than 95% complete for galaxies larger than R(26) ≈ 25″ and r < 18 measured at the 26 mag arcsec−2 isophote in the r band. The atlas delivers precise coordinates, multiwavelength mosaics, azimuthally averaged optical surface-brightness profiles, model images and photometry, and additional ancillary metadata for the full sample. Coupled with existing and forthcoming optical spectroscopy from the DESI, the SGA-2020 will facilitate new detailed studies of the star formation and mass assembly histories of nearby galaxies; enable precise measurements of the local velocity field via the Tully–Fisher and fundamental plane relations; serve as a reference sample of lasting legacy value for time-domain and multimessenger astronomical events; and more.

In situ or accreted? Using deep learning to infer the origin of extragalactic globular clusters from observables

ABSTRACT Globular clusters (GCs) are powerful tracers of the galaxy assembly process, and have already been used to obtain a detailed picture of the progenitors of the Milky Way (MW). Using the E-MOSAICS cosmological simulation of a (34.4 Mpc)3 volume that follows the formation and co-evolution of galaxies and their star cluster populations, we develop a method to link the origin of GCs to their observable properties. We capture this complex link using a supervised deep learning algorithm trained on the simulations, and predict the origin of individual GCs (whether they formed in the main progenitor or were accreted from satellites) based solely on extragalactic observables. An artificial neural network classifier trained on ∼50 000 GCs hosted by ∼700 simulated galaxies successfully predicts the origin of GCs in the test set with a mean accuracy of 89 per cent for the objects with $\rm [Fe/H]\lt -0.5$ that have unambiguous classifications. The network relies mostly on the alpha-element abundances, metallicities, projected positions, and projected angular momenta of the clusters to predict their origin. A real-world test using the known progenitor associations of the MW GCs achieves up to 90 per cent accuracy, and successfully identifies as accreted most of the GCs in the inner Galaxy associated to the Kraken progenitor, as well as all the Gaia-Enceladus GCs. We demonstrate that the model is robust to observational uncertainties, and develop a method to predict the classification accuracy across observed galaxies. The classifier can be optimized for available observables (e.g. to improve the accuracy by including GC ages), making it a valuable tool to reconstruct the assembly histories of galaxies in upcoming wide-field surveys.

The stellar halo in Local Group Hestia simulations

Stellar chemical abundances and kinematics provide key information for recovering the assembly history of galaxies. In this work we explore the chemo-chrono-kinematics of accreted and in situ stellar populations, by analyzing six M31/Milky Way (MW) analogues from the HESTIA suite of cosmological hydrodynamics zoom-in simulations of the Local Group. We show that elemental abundances ([Fe/H], [Mg/Fe]) of merger debris in the stellar haloes are chemically distinct from the survived dwarf galaxies, in that they are [α/Fe]-enhanced and have lower metallicity in the same stellar mass range. Therefore, mergers debris have abundances expected for stars originating from dwarfs that had their star formation activity quenched at early times. Accreted stellar haloes, including individual debris, reveal [Fe/H] and [Mg/Fe] gradients in the E − Lz plane, with the most metal-rich, [α/Fe]-poor stars, which have formed in the inner parts of the disrupted systems before the merger, contributing mainly to the central regions of the host galaxies. This results in negative metallicity gradients in the accreted components of stellar haloes at z = 0, seen also for the individual merger debris. We suggest, therefore, that abundance measurements of halo stars in the inner MW will allow constraining better the parameters, such as the mass and merger time, of MW’s most massive merger Gaia-Sausage-Enceladus. The metallicity distribution functions (MDFs) of the individual debris show several peaks and the majority of debris have lower metallicity than the in situ stars in the prograde part of the E − Lz space. At the same time, non-rotating and retrograde accreted populations are very similar to the in situ stars in terms of [Fe/H] abundance. Prograde accreted stars show a prominent knee in the [Fe/H]–[Mg/Fe] plane, reaching up to solar [Mg/Fe], while retrograde stars typically contribute to the high-[Mg/Fe] sequence only. We find that the most metal-poor stars ([Fe/H] ≲ −1) of the HESTIA galaxies exhibit net rotation up to 80 km s−1, which is consistent with the Aurora population recently identified in the MW. At higher metallicities ([Fe/H] ≈ −0.5 ± 0.1) we detect a sharp transition (spin-up) from the turbulent phase to a regular disk-like rotation. Different merger debris appear similar in the [Fe/H]–[Mg/Fe] plane, thus making it difficult to identify individual events. However, combining a set of abundances, and especially stellar age, makes it possible to distinguish between different debris.

Different higher order kinematics between star-forming and quiescent galaxies based on the SAMI, MAGPI, and LEGA-C surveys

ABSTRACT We present the first statistical study of spatially integrated non-Gaussian stellar kinematics spanning 7 Gyr in cosmic time. We use deep, rest-frame optical spectroscopy of massive galaxies (stellar mass $M_\star \gt 10^{10.5} \, \mathrm{M_\odot }$) at redshifts z = 0.05, 0.3, and 0.8 from the SAMI, MAGPI, and LEGA-C surveys, to measure the excess kurtosis h4 of the stellar velocity distribution, the latter parametrized as a Gauss–Hermite series. We find that at all redshifts where we have large enough samples, h4 anticorrelates with the ratio between rotation and dispersion, highlighting the physical connection between these two kinematic observables. In addition, and independently from the anticorrelation with rotation-to-dispersion ratio, we also find a correlation between h4 and M⋆, potentially connected to the assembly history of galaxies. In contrast, after controlling for mass, we find no evidence of independent correlation between h4 and aperture velocity dispersion or galaxy size. These results hold for both star-forming and quiescent galaxies. For quiescent galaxies, h4 also correlates with projected shape, even after controlling for the rotation-to-dispersion ratio. At any given redshift, star-forming galaxies have lower h4 compared to quiescent galaxies, highlighting the link between kinematic structure and star-forming activity.

Initial sizes of star clusters: implications for cluster dissolution during galaxy evolution

ABSTRACT Massive star clusters are often used as tracers of galaxy formation and assembly. In order to do so, we must understand their properties at formation, and how those properties change with time, galactic environment, and galaxy assembly history. The two most important intrinsic properties that govern star cluster evolution are mass and radius. In this paper, we investigate 10 theoretically and observationally motivated initial size–mass relations for star clusters, and evolve populations of clusters through galaxy formation models. We compare our results to each other and to observations of cluster populations in M83, M31, and the Milky Way. We find that none of our size–mass relations agree with the observations after 6–10 Gyr of evolution. We can successfully reproduce the cluster mass functions with models that have a small range of initial radii, and which do not allow cluster radii to change with time. However, these models do not agree with our understanding of cluster evolution, which does involve radius evolution, and do not match the observed distributions of radii. We note that there is a region of parameter space where clusters are optimally protected from both tidal shocks and evaporation due to two-body relaxation. Clusters that are allowed to evolve into this parameter space will likely survive. An improved understanding of both mass and radius evolution of star clusters in realistic, time-varying galactic potentials is necessary to appropriately make the connection between present-day cluster properties and their use as tracers of galaxy formation and assembly.

The integrated metallicity profile of the Milky Way

The heavy element abundance profiles of galaxies place stringent constraints on galaxy growth and assembly history. As the Milky Way is currently the only spiral galaxy in which we can measure temporally resolved chemical abundances, it enables insights into the origin of metallicity gradients and their correlation with the growth history of galaxies. However, until now, these abundance profiles have not been translated into the integrated-light measurements that are needed to compare the Milky Way with the general galaxy population. Here we report the measurement of the light-weighted, integrated stellar metallicity profile of our Galaxy. We find that the integrated stellar metallicity profile of the Milky Way has a ∧-like broken shape, with a mildly positive gradient inside a Galactocentric radius of 7 kpc and a steep negative gradient outside. This broken integrated metallicity profile of the Milky Way is not unique but is not common among Milky Way-mass star-forming galaxies observed in the MaNGA survey and simulated in the TNG50 cosmological simulation. Our results suggest that the Milky Way might not have a typical metallicity distribution for a galaxy of its mass, and thus offers valuable insight into the rich variety of galaxy enrichment processes.

Correction to: The fragility of thin discs in galaxies – II. Thin discs as tracers of the assembly history of galaxies

Correction to: The fragility of thin discs in galaxies – II. Thin discs as tracers of the assembly history of galaxies

The fragility of thin discs in galaxies – II. Thin discs as tracers of the assembly history of galaxies

ABSTRACT Thin galactic discs and nuclear stellar discs (NSDs) are fragile structures that can be easily disturbed by merger events. By studying the age of the stellar populations in present-day discs, we can learn about the assembly history of galaxies and place constraints on their past merger events. Following on the steps of our initial work, we explore the fragility of such disc structures in intermediate-mass-ratio dry encounters using the previously constructed N-body model of the Fornax galaxy NGC 1381 (FCC 170), which hosts both a thin galactic disc and an NSD. We dismiss major and minor encounters, as the former were previously shown to easily destroy thin-disc structures, whereas the latter take several Hubble times to complete in the specific case of FCC 170. The kinematics and structure of the thin galactic disc are dramatically altered by the mergers, whereas the NSD shows a remarkable resilience, exhibiting only a smooth increase of its size when compared to the model evolved in isolation. Our results suggest that thin galactic discs are better tracers for intermediate-mass-ratio mergers, while NSDs may be more useful for major encounters. Based on our simulations and previous analysis of the stellar populations, we concluded that FCC 170 has not experienced any intermediate-mass-ratio dry encounters for at least ∼10 Gyr, as indicated by the age of its thin-disc stellar populations.

Overmassive central black holes in the cosmological simulations astrid and Illustris TNG50

ABSTRACT Recent dynamical measurements indicate the presence of a central supermassive black hole (SMBH) with mass ${\sim} 3\times 10^6\, {\rm M_\odot }$ in the dwarf galaxy Leo I, placing the system ∼50 times above the standard, local MBH–M⋆ relation. While a few overmassive central SMBHs are reported in nearby isolated galaxies, this is the first one detected in a Milky Way satellite. We used the ASTRID and Illustris TNG50 lambda cold dark matter (LCDM) cosmological simulations to investigate the assembly history of galaxies hosting overmassive SMBHs. We estimate that, at the stellar mass of Leo I, ${\sim} 15~{{\ \rm per\ cent}}$ of galaxies above the MBH–M⋆ relation lie &amp;gt;10 times above it. Leo I-like systems are rare but exist in LCDM simulations: they occur in ${\sim} 0.005~{{\ \rm per\ cent}}$ of all overmassive systems. Examining the properties of simulated galaxies harbouring overmassive central SMBHs, we find that: (i) stars assemble more slowly in galaxies above the MBH–M⋆ relation; (ii) the gas fraction in these galaxies experiences a significantly steeper decline over time; and (iii) $\gt 95~{{\ \rm per\ cent}}$ of satellite host galaxies in overdense regions are located above the MBH–M⋆ relation. This suggests that massive satellite infall and consequent tidal stripping in a group/dense environment can drive systems away from the MBH–M⋆ relation, causing them to become overmassive. As the merging histories of overmassive and undermassive systems do not differ, we conclude that additional environmental effects, such as being in overdense regions must play a crucial role. In the high-z Universe, central overmassive SMBHs are a signature of heavy BH seeds; we demonstrate, in contrast, that low-z overmassive systems result from complex environmental interactions.

SDSS-IV MaNGA: The Effect of Stellar Mass and Halo Mass on the Assembly Histories of Satellite Galaxies

We combine an unprecedented MaNGA sample of over 3000 passive galaxies in the stellar mass range 109–1012 M ⊙ with the Sloan Digital Sky Survey group catalog by Tinker to quantify how central and satellite formation, quantified by radial profiles in stellar age, [Fe/H], and [Mg/Fe], depends on the stellar mass of the galaxy (M *) and the mass of the host halo (M h ). After controlling for M * and M h , the stacked spectra of centrals and satellites beyond the effective radius (r e ) show small, yet significant differences in multiple spectral features at the 1% level. According to spectral fitting with the code alf, a primary driver of these differences appears to be [Mg/Fe] variations, suggesting that stellar populations in the outskirts of satellites formed more rapidly than the outer populations of centrals. To probe the physical mechanisms that may be responsible for this signal, we examined how satellite stellar populations depend on M h . We find that satellites in high-M h halos show older stellar ages, lower [Fe/H], and higher [Mg/Fe] compared to satellites in low-M h halos, especially for M * = 109.5–1010.5 M ⊙. These signals lend support to environmentally driven processes that quench satellite galaxies, although variations in the merger histories of central and satellite galaxies also emerge as a viable explanation.

The fragility of thin discs in galaxies – I. Building tailored N-body galaxy models

ABSTRACT Thin stellar discs on both galactic and nuclear, sub-kpc scales are believed to be fragile structures that would be easily destroyed in major mergers. In turn, this makes the age dating of their stellar populations a useful diagnostics for the assembly history of galaxies. We aim at carefully exploring the fragility of such stellar discs in intermediate- and low-mass encounters, using high-resolution N-body simulations of galaxy models with structural and kinematic properties tailored to actually observed galaxies. As a first but challenging step, we create a dynamical model of FCC 170, a nearly edge-on galaxy in the Fornax cluster with multiple galactic components and including both galactic-scale and nuclear stellar discs (NSDs), using detailed kinematic data from the Multi Unit Spectroscopic Explorer and a novel method for constructing distribution function-based self-consistent galaxy models. We then create N-body realizations of this model and demonstrate that it remains in equilibrium and preserves its properties over many Gyr, when evolved with a sufficiently high particle number. However, the NSD is more prone to numerical heating, which gradually increases its thickness by up to 22 per cent in 10 Gyr even in our highest resolution runs. Nevertheless, these N-body models can serve as realistic representations of actual galaxies in merger simulations.

ERGO-ML I: inferring the assembly histories of IllustrisTNG galaxies from integral observable properties via invertible neural networks

ABSTRACT A fundamental prediction of the ΛCDM cosmology is the hierarchical build-up of structure and therefore the successive merging of galaxies into more massive ones. As one can only observe galaxies at one specific time in the cosmic history, this merger history remains, in principle, unobservable. By using the TNG100 simulation of the IllustrisTNG project, we show that it is possible to infer the unobservable stellar assembly and merger history of central galaxies from their observable properties by using machine learning techniques. In particular, in this first paper of ERGO-ML (Extracting Reality from Galaxy Observables with Machine Learning), we choose a set of seven observable integral properties of galaxies to infer the stellar ex-situ fraction, the average merger lookback times and mass ratios, and the lookback time and stellar mass of the last major merger. To infer the posterior distribution for these parameters and hence estimate the uncertainties in the predictions, we use a conditional Invertible Neural Network (cINN). We find that the stellar ex-situ fraction and the time of the last major merger are well-determined by the selected set of observables, that the mass-weighted merger mass ratio is unconstrained, and that, beyond stellar mass, stellar morphology and stellar age are the most informative properties. Finally, we show that the cINN recovers the remaining unexplained scatter and secondary cross-correlations. Overall, this is a first step towards a tool that can be applied to large galaxy surveys in order to infer unobservable properties of the galaxies’ past, enabling empirical studies of galaxy evolution enriched by cosmological simulations.