The physical processes responsible for shaping how galaxies form and quench over time leave imprints on both the spatial (galaxy morphology) and temporal (star formation history; SFH) tracers that we use to study galaxies. While the morphology–SFR connection is well studied, the correlation with past star formation activity is not as well understood. To quantify this, we present Katachi (形), an interpretable convolutional neural network framework that learns the connection between the factors regulating star formation in galaxies on different spatial and temporal scales. Katachi is trained on 9904 galaxies at 0.02 < z < 0.1 in the SDSS-IV MaNGA DR17 sample to predict stellar mass (M*; root mean square error (RSME) 0.22 dex), current star formation rate (SFR; RMSE 0.31 dex), and half-mass time (t50; RMSE 0.23 dex). This information allows us to reconstruct nonparametric SFHs for each galaxy from gri imaging alone. To quantify the morphological features informing the SFH predictions, we use SHAP (SHapley Additive exPlanations). We recover the expected trends of M* governed by the growth of galaxy bulges, as well as those of SFR correlating with spiral arms and other star-forming regions. We also find that the SHAP maps of D4000 are more complex than those of M* and SFR, and that morphology is correlated with t50 even at fixed mass and SFR. Katachi serves as a scalable public framework to predict galaxy properties from large imaging surveys including Rubin, Roman, and Euclid, with large data sets of high signal-to-noise ratio imaging across limited photometric bands.
Star formation in galaxies is regulated by the interplay of a range of processes that shape the multiphase gas in the interstellar and circumgalactic media. Using the Cosmology and Astrophysics with MachinE Learning Simulations (CAMELS) suite of cosmological simulations, we study the effects of varying feedback and cosmology on the average star formation histories (SFHs) of galaxies at z ∼ 0 across the IllustrisTNG, SIMBA, and ASTRID galaxy formation models. We find that galaxy SFHs in all three models are sensitive to changes in stellar feedback, which affect the efficiency of baryon cycling and the rates at which central black holes grow, whereas the effects of varying active galactic nucleus (AGN) feedback depend on model-specific implementations of black hole seeding, accretion, and feedback. We also find strong interaction terms that couple stellar and AGN feedback, usually by regulating the amount of gas available for the central black hole to accrete. Using a double power law to describe the average SFHs, we derive a general set of equations relating the shape of the SFHs to physical quantities like baryon fraction and black hole mass across all three models. We find that a single set of equations (albeit with different coefficients) can describe the SFHs across all three CAMELS models, with cosmology dominating the SFH at early times, followed by halo accretion, and feedback and baryon cycling at late times. Galaxy SFHs provide a novel, complementary probe to constrain cosmology and feedback, and can connect the observational constraints from current and upcoming galaxy surveys with the physical mechanisms responsible for regulating galaxy growth and quenching.
Massive quiescent galaxies at high redshift show significantly more compact morphology than their local counterparts. To examine their internal structure across a wide redshift range and investigate potential redshift dependence, we performed spatially resolved spectral energy distribution fitting using piXedfit software on massive quiescent galaxies at 0 < z < 4 with public James Webb Space Telescope and Hubble Space Telescope imaging data from the Public Release Imaging for Extragalactic Research and the Cosmic Evolution Early Release Science Survey. We find that at z ∼ 3.5 the half-mass radius is about 5.4 times smaller than at z ∼ 0.5. This growth is driven by stellar mass buildup in the outskirts ( r > 4 kpc), while the central regions ( r ∼ 1 kpc) remain largely unchanged, with stellar mass surface density similar to local quiescent galaxies. The estimated star formation rates are too low to explain the stellar mass growth, indicating that an additional stellar mass accumulation process, such as mergers, is necessary. We parameterize the size–mass relation of the most massive galaxies in our sample as and find for z ⪅ 2, consistent with growth dominated by minor mergers, and for z ⪆ 2, consistent with growth dominated by major mergers. These results indicate that massive quiescent galaxies originate from compact quenched systems and grow through combinations of minor and major mergers.
E+A galaxies represent a class of recently quenched objects, with spectra that show evidence of a previous substantial starburst and strong Balmer absorption lines indicative of A-type stellar populations. Using an SDSS-IV MaNGA–selected sample of E+A galaxies, we identify a matching sample in the TNG50 simulation to study their evolutionary histories. Additionally, we identify a sample of generic post-starburst (PSB) galaxies based on their star formation histories (SFHs) from the MaNGA Pipe3D value-added catalog. We find that PSB-like SFHs make up a similar fraction of galaxies in TNG50 (0.9% compared to MaNGA’s ∼1%). Matching galaxies based solely on their stellar masses and metallicities in TNG50 does not result in a sample with PSB-like histories. We analyzed the chemical enrichment histories of our selected simulated galaxy samples and found that both PSB and E+A galaxies exhibit a distinct episode of rapid enrichment when compared to galaxies in TNG50 with a similar stellar mass range. These galaxies are typically metal-poor before undergoing an extended ∼2 Gyr starburst, during which a phase of rapid chemical enrichment occurs over the first ∼300 Myr. The final systems are generally more metal-rich than the average galaxy in TNG50, while the MaNGA data do not show this trend. This suggests that PSB galaxies undergo unique evolutionary processes as they transition from starburst activity to quiescence. Further studies are needed to determine whether these galaxies originate from truly metal-poor progenitors and to better understand their subsequent evolution.
The buildup of dense stellar cores is expected to mark an important transition in the star formation and black hole growth of massive galaxies. Using spatially resolved spectral energy distribution (SED) fitting of James Webb Space Telescope near-infrared imaging, combined with a stacking analysis of Chandra X-ray data, we trace stellar mass buildup and average black hole accretion in the most massive galaxies at z < 5, selecting the 50 most massive galaxies per redshift bin at constant number density of ∼4.4 × 10 −5 cMpc −3 . To robustly constrain central stellar populations, we separate active galactic nucleus (AGN) components affecting the photometry using multiband morphological decomposition and SED analysis. We find that the sample selected with constant number density exhibits an evolutionary trend of rapid central compaction at z ∼ 4, during which the median central 1 kpc stellar mass increases by ∼0.60 dex over ∼400 Myr. The majority of X-ray-detected AGN (63% ± 12%) are hosted by galaxies undergoing the compaction, while we find neither individually detected X-ray sources nor a significant stacked X-ray signal at z > 4, indicating that substantial average black hole growth emerges primarily during, rather than before, the compaction. Following the compaction, central specific star formation rates (sSFRs) decline by ∼1.24 dex over ∼700 Myr at z ∼ 3 while remaining elevated galaxy-wide, signaling the onset of inside-out quenching. Despite this central suppression, the specific black hole accretion rate remains coupled to the total sSFR. Our results suggest that dense core formation in the most massive galaxies marks the onset of inside-out quenching and a transition toward an enhanced black hole to stellar growth ratio.
The physical processes responsible for shaping how galaxies form and quench over time leave imprints on both the spatial (galaxy morphology) and temporal (star formation history; SFH) tracers that we use to study galaxies. While the morphology-SFR connection is well studied, the correlation with past star formation activity is not as well understood. To quantify this we present Katachi, an interpretable convolutional neural network (CNN) framework that learns the connection between the factors regulating star formation in galaxies on different spatial and temporal scales. Katachi is trained on 9904 galaxies at 0.02$<$z$<$0.1 in the SDSS-IV MaNGA DR17 sample to predict stellar mass (M$_*$; RMSE 0.22 dex), current star formation rate (SFR; RMSE 0.31 dex) and half-mass time (t$_{50}$; RMSE 0.23 dex). This information allows us to reconstruct non-parametric SFHs for each galaxy from \textit{gri} imaging alone. To quantify the morphological features informing the SFH predictions we use SHAP (SHapley Additive exPlanations). We recover the expected trends of M$_*$ governed by the growth of galaxy bulges, and SFR correlating with spiral arms and other star-forming regions. We also find the SHAP maps of D4000 are more complex than those of M$_*$ and SFR, and that morphology is correlated with t$_{50}$ even at fixed mass and SFR. Katachi serves as a scalable public framework to predict galaxy properties from large imaging surveys including Rubin, Roman, and Euclid, with large datasets of high SNR imaging across limited photometric bands.