WorksPhysics and Astronomy
The HI Mass Fractions of Milky Way Analog Samples
Abstract
We estimate average HI mass fractions, f_HI = M_HI/M_star, for nine Milky Way analog samples (MWAs): galaxies selected on similarity to the MW in terms of various properties. We find most (seven) MWAs have lower <f_HI> than a recent estimate for the MW, although only the two largest have significantly lower estimates. We suggest that either the MW has an unusual f_HI for its properties, or we are overestimating its HI content, one of many challenging global measurements to make from our perspective inside our own Galaxy.
Introduction
Milky Way analog galaxy samples (MWAs, e.g. T. C. Licquia et al. [13], L15; A. Fraser-McKelvie et al. [7]), containing galaxies which resemble the MW, can place the properties of the MW in the context of similar galaxies in the Universe and/or provide us with valuable information about our own Galaxy that are otherwise challenging to obtain. The total H I mass fraction, () is one such property. Measurements of both total stellar and H I mass suffer from selection effects from our internal perspective ([25, 11]). There are multiple ways to define MWAs. Matching with stellar mass () and star formation rate (SFR) is common (e.g. L15, A. Fraser-McKelvie et al. 2019). S. Tuntipong et al. [22] emphasize the importance of combining multiple criteria, selecting , SFR, bulge-to-total ratio () and effective radius, (). This research note considers the distributions of several MWAs and compares them with current best estimates of .
Sample and Data
Most of the MWAs we consider are from the Mapping Nearby Galaxies at Apache Point Observatory (MaNGA) DR17 (Abdurro’uf et al. [1]); we also include some nearby galaxies from L. S. Pilyugin et al. (2023, P23). For MaNGA galaxies, is obtained from the GALEX-SDSS-WISE Legacy Catalog (GSWLC-X2, S. Salim et al. [19, 18]), while H I total masses and upper limits are from H I-MaNGA DR4 ([14, 20]). P23 provide for their sample, while was sourced from various works ([2, 4, 5, 8, 10, 9, 15, 16, 21, 23, 24]).
In total we consider nine MWAs:
two from the MaNGA MWA ancillary program - an SFR matched sample, which was designed to closely replicate L15 () and a matched sample based on T. C. Licquia & J. A. Newman [12], ()1.
four MaNGA-based samples from N. Boardman et al. [3] matched in SFR (), (), disk scale length, (, ), and morphology ()
matched MaNGA MWAs from S. Zhou et al. [26]
local gas-phase metallicity matched MWAs (P23), which we further limit to six galaxies with consistent with .
our own HI-MaNGA-based -matched subset, used to provide a reference point for for galaxies with stellar masses in the range ().
We compute cumulative distribution functions (CDFs) of for each of these MWAs including HI upper limits via the Kaplan-Meier estimator2, to find the median and 95% confidence interval (CI). These results are shown in Figure 1. We adopt , which is shown as the dashed line (grey region) based on C. E. Fielder et al. [6] who infer and Y. Wang et al. [25] who measure –.

Figure 1. Comparison of between MWAs and the MW (black dashed line and shaded CI). For all samples, we show median (crosses) and 95% CI (horizontal lines). For most samples, individual MWA galaxy measurements are overlaid with detections (dots) and upper limits (left arrows); we show only the range (filled triangles) for the large -matched sample. The red dashed line show the weighted mean of all medians.
We find that the median values of most MWAs are smaller than, but agree within error of . An selected MWAs (B20) has lower , but this may be poorly constrained due to a large fraction of upper limits. The two largest MWAs - the B/T-matched MWAs of Z23 and our simple mass matched MWAs have significantly below the MW. The inverse-variance weighted mean of all medians is , is also just below the 95% CI.
Discussion and Conclusion
Most MWAs are found to have lower than recent estimates for (C. E. Fielder et al. [6]; Y. Wang et al. [25]), although agree within error. The two largest MWAs have significantly lower median than the MW, and a weighted mean of all medians is found to be just below the 95% CI. Based on this, we conclude that either the MW is unusual in its HI content, or current measurements of are overestimates. Adding to this puzzle, J. Lian et al. [11] suggest current estimates of the MW stellar mass are too high, while Y. Wang et al. [25] note that their measurement of is an underestimate of up to 20–40% due to a variety of factors. Both suggestions would increase relative to these MWAs. Even in this era of precision astrophysics, estimating global values for the MW remains challenging.
KLM and DVS acknowledge support from the US NSF under grant AST-2510740 ”Collaborative Research: Understanding the extreme diversity of atomic hydrogen depletion times in galaxies”.
References
- [1]Abdurro’uf, Accetta, K., Aerts, C., et al. 2022, ApJS, 259, 35, doi: 10.3847/1538-4357/ac4414
- [2]Ball, R. 1986, ApJ, 307, 453, doi: 10.1086/164435DOI
- [3]Boardman, N., Zasowski, G., Newman, J. A., et al. 2020, MNRAS, 498, 4943, doi: 10.1093/mnras/staa2731DOI
- [4]Braun, R., Thilker, D. A., Walterbos, R. A. M., & Corbelli, E. 2009, ApJ, 695, 937, doi: 10.1088/0004-637X/695/2/937DOI
- [5]Bush, S. J., & Wilcots, E. M. 2004, AJ, 128, 2789, doi: 10.1086/425876DOI
- [6]Fielder, C. E., Newman, J. A., Andrews, B. H., et al. 2021, MNRAS, 508, 4459, doi: 10.1093/mnras/stab2618DOI
- [7]Fraser-McKelvie, A., Merrifield, M., & Aragón-Salamanca, A. 2019, MNRAS, 489, 5030, doi: 10.1093/mnras/stz2493
- [8]Haynes, M. P., Giovannelli, R., Kent, B. R., et al. 2018, ApJ, 861, 49, doi: 10.3847/1538-4357/aac956DOI
- [9]Koribalski, B. S., & López-Sánchez, Á. R. 2009, MNRAS, 400, 1749, doi: 10.1111/j.1365-2966.2009.15610.xDOI
- [10]Koribalski, B. S., Staveley-Smith, L., Kilborn, V. A., et al. 2004, AJ, 128, 16, doi: 10.1086/421744DOI
- [11]Lian, J., Wang, T., Feng, Q., Huang, Y., & Guo, H. 2025, ApJL, 990, L37, doi: 10.3847/2041-8213/adfc73DOI
- [12]Licquia, T. C., & Newman, J. A. 2016, ApJ, 831, 71, doi: 10.3847/0004-637X/831/1/71DOI
- [13]Licquia, T. C., Newman, J. A., & Brinchmann, J. 2015, ApJ, 809, 96, doi: 10.1088/0004-637X/809/1/96DOI
- [14]Masters, K. L., Stark, D. V., Pace, Z. J., et al. 2019, MNRAS, 488, 3396, doi: 10.1093/mnras/stz1889DOI
- [15]Mishra, A., Kantharia, N. G., & Srivastava, D. C. 2012, BASI, 40, 515
- [16]Namumba, B., Koribalski, B. S., Józsa, G. I. G., et al. 2021, MNRAS, 505, 3795, doi: 10.1093/mnras/stab1524DOI
- [18]Salim, S., Boquien, M., & Lee, J. C. 2018, ApJ, 859, 11, doi: 10.3847/1538-4357/aabf3cDOI
- [19]Salim, S., Lee, J. C., Janowiecki, S., et al. 2016, ApJS, 227, 2, doi: 10.3847/0067-0049/227/1/2DOI
- [20]Stark, D. V., Masters, K. L., Avila-Reese, V., et al. 2021, MNRAS, 503, 1345, doi: 10.1093/mnras/stab566DOI
- [21]Theureau, G., Bottinelli, L., Coudreau-Durand, N., et al. 1998, A&AS, 130, 333, doi: 10.1051/aas:1998416DOI
- [22]Tuntipong, S., van de Sande, J., Croom, S. M., et al. 2024, MNRAS, 533, 4334, doi: 10.1093/mnras/stae2042DOI
- [23]Walter, F., Brinks, E., de Blok, W. J. G., et al. 2008, AJ, 136, 2563, doi: 10.1088/0004-6256/136/6/2563DOI
- [24]Wang, S., Wu, Z., Zhang, B., et al. 2024, A&A, 686, A172, doi: 10.1051/0004-6361/202348174
- [25]Wang, Y., Beuther, H., Rugel, M. R., et al. 2020, A&A, 634, A83, doi: 10.1051/0004-6361/201937095DOI
- [26]Zhou, S., Aragón-Salamanca, A., Merrifield, M., et al. 2023, MNRAS, 521, 5810, doi: 10.1093/mnras/stad853DOI