FEASTS Compared with Simulations: Abnormally Irregular and Extended HI Morphologies at a Column Density of $10^{18}\,\text{cm}^{-2}$ in TNG50 and Auriga

TL;DR

This paper investigates how HI morphology at low column density ($N_{ ext{H}} \sim 10^{18}$ cm$^{-2}$) around Milky Way–like galaxies compares between FEASTS observations and cosmological simulations (TNG50 and Auriga). The authors generate fair mock HI cubes, matching FEASTS galaxies in stellar and HI mass and applying a consistent analysis to both observed and simulated data, including a dedicated postprocessing pipeline and morphology metrics $I_{18}$, $\log(S_{18}/S_{19})$, and $\log(S_{19}/S_{20})$ based on contours at $N_{ ext{H}}$ = $10^{18}$, $10^{19}$, and $10^{20}$ cm$^{-2}$. They find that more than a third of TNG50 galaxies exhibit abnormally irregular and extended low-$N_{ ext{H}}$ HI compared with FEASTS, with stronger deviations at higher stellar mass and lower HI fraction, and that stellar feedback is the dominant driver of these discrepancies, whereas AGN feedback acts more indirectly. The Auriga results suggest magnetic fields may help regularize HI morphology, while simply increasing mass resolution does not fully resolve the discrepancies, highlighting the importance of CGM physics and magnetic effects in shaping low-density HI. Overall, the study demonstrates the potential of using deep, low-$N_{ ext{H}}$ HI observations to constrain and refine future galaxy-formation simulations.

Abstract

With new atomic-hydrogen (HI) observations of FAST Extended Atlas of Selected Targets Survey (FEASTS), we present the first statistical comparison of HI morphology between observations and cosmological simulations, focusing on low-column density ($10^{18}\,\text{cm}^{-2}$) regions of Milky Way-like central galaxies. We select a 330-galaxy sample from IllustrisTNG50 (TNG50) matched to 33 FEASTS galaxies by stellar and HI masses, and mock observe them to the FAST resolution and depth at corresponding inclinations and distances for a fair comparison. In contrast to FEASTS, abnormally irregular and extended morphology is found in more than one-third of TNG50 galaxies, especially those massive and HI poor. Stellar feedback is the property that most significantly correlates with the HI morphological deviation from observations, although these deviations mostly occur at a high stellar or black-hole mass. These results indicate that in TNG50, stellar feedback significantly influences the HI morphology at $10^{18}\,\text{cm}^{-2}$, while active galactic nucleus (AGN) feedback has not so direct a role as expected. With an additional sample from Auriga, we find that the magnetic field may help HI to be more regular in its morphology, while improving the mass resolution does not alleviate the discrepancy from observation. This study reveals the potential of constraining future simulations of galaxies by observing low-column density HI.

Figures (11)

• Figure 1: The H mass ($M_{\text{H \xspace,}}\xspace$) and stellar mass ($M_\ast$) of our observational sample (filled green triangles), TNG50 matched sample (empty pink squares), and the Auriga sample (empty blue triangles). The $M_{\text{H \xspace,}}\xspace$--$M_\ast$ relation of star-forming galaxies in xGASS is plotted as the gray line, with the [-NoValue-]∼0.3 scatter indicated by dotted lines.
• Figure 2: The H moment-0 atlas of FEASTS galaxies (the single panel above) and their distance-matched TNG50 cubes sorted by the values of morphological parameters in descending order (ten panels below). In the full Figure Set, FEASTS galaxies are sorted by $M_\ast$ in descending order. All panels have a sidelength of [-NoValue-]250, and the beam size of ;3.24; is indicated by the circle at the lower left corner. On the left of FEASTS moment-0 map, we list the NGC name and morphological parameters; on the right, we list the stellar and H masses and the inclination. In the TNG panels, we list their subhalo ID, AGN-feedback mode ("K" for kinetic mode and "T" for thermal mode), and morphological parameters. Morphological parameters listed in the atlas are not corrected for the distance. Contours are plotted at $N_{\text{H \xspace,}}\xspace\xspace=\text{\SIlist[-NoValue-]{1e18;1e19;1e20}{\per\cm\squared}}$ (solid lines) and [-NoValue-]1e21 (dotted lines).
• Figure 3: The cumulative distribution of three morphological parameters ($I_{18}$, $\log(S_{18}/S_{19})$, and $\log(S_{19}/S_{20})$ from left to right) and their dependence on distances $D$. The FEASTS observational sample and TNG50 distance-matched sample are represented by green dashed and pink solid lines, respectively. The $p$-value of a K--S test between the two samples is given in the legend.
• Figure 4: The absolute radial and vertical extensions of TNG50 H disks compared with observational relations. (a, b) The H size--mass relation with $R_{\text{H \xspace,}}\xspace$ and $R_{\text{H \xspace,001}}\xspace$, the radii at H surface densities of [-NoValue-]1;0.01 after correcting for projection effects. Inclinations are required to be less than 90. Kinetic- and thermal-mode TNG50 galaxies are labeled as purple crosses and orange pluses, respectively, and fiducial Auriga galaxies are plotted as blue triangles. The observed relations from 2016MNRAS.460.2143W2025ApJ...980...25W are plotted as solid lines for reference, and a scatter of [-NoValue-]∼0.06 is indicated by dotted lines. (c) The extension $z_{18}$ perpendicular to the H disk at a column density of [-NoValue-]1e18, measured using edge-on mock observation. The observed relation from 2025ApJ...984...15Y is plotted as the solid line for reference with a scatter of [-NoValue-]∼0.1. The binned median values and the 16th and 84th percentiles are plotted as purple dotted line and shading for kinetic mode and as orange dashed--dotted line and shading for thermal mode in panels (b) and (c).
• Figure 5: The relations between two morphological parameters at FEASTS distance ($I_{18}$ and $\log(S_{18}/S_{19})$; two rows) and three physical properties ($M_\ast$, $f_{\text{H \xspace,}}\xspace$, and $\text{sSFR}$; three columns). FEASTS observational sample and TNG50 matched sample are represented by green triangles and pink squares, respectively. The Spearman rank correlation coefficients ($r_{\text{s}}$) are listed above every panel for each sample, along with the $p$-values; their uncertainties are obtained with bootstrapping. The running median values of the morphological parameters are plotted as pink solid (TNG50) or green dashed (FEASTS) profiles. The widths of median filters are chosen as [-NoValue-]20 of dynamic ranges in physical properties, i.e., [-NoValue-]0.32;0.28;0.41 for $M_\ast$, $f_{\text{H \xspace,}}\xspace$, and $\text{sSFR}$, respectively, and the profile steps are [-NoValue-]5 of the ranges. Only bins with at least four data points are plotted. We remove the incomplete bins for TNG50 profiles, and indicate them as thin lines for FEASTS. The uncertainty of median values is indicated by translucent shadings.