[Paper Review] Populating HI gas in dark matter halos: I. method

Yi Lu, Xiaohu Yang+5 more|ArXiv.org|Aug 22, 2020

Galaxies: Formation, Evolution, Phenomena4 citations

TL;DR

This paper develops a halo-based model to predict HI gas content in dark matter halos by combining galaxy properties (stellar mass, g−r color) with halo mass-dependent HI survival efficiency. The model successfully reproduces observed HI detection rates, total HI mass, and satellite fractions—outperforming galaxy-only models that overpredict detections by up to a factor of two in massive halos.

ABSTRACT

We combine data from the Sloan Digital Sky Survey (SDSS) and the Arecibo Legacy Fast ALFA Survey (ALFALFA) to establish an empirical model for the HI gas content within dark matter halos. A cross-match between our SDSS DR7 galaxy group sample and the ALFALFA HI sources provides a catalog of 16,520 HI-galaxy pairs within 14,270 galaxy groups (halos). Using these matched pairs, we model the HI gas mass distributions within halos using two components: 1) {\it in situ} galaxy relations that involve the HI masses, colors $({ m g-r})$ and stellar masses 2) an {\it ex situ} dependence of the HI mass on the halo mass/environment. We find that if we solely use galaxy associated scaling relations to predict the HI gas distribution (solely component 1), the number of HI detections is significantly over-predicted with respect the ALFALFA observations. We introduce a concept for the survival of the HI masses/members within halos of different masses labelled as the `efficiency' factor, in order to describe the probability that a halo has in retaining its HI detections. Taking the above consideration into account we construct a `halo based HI mass model' which does not only predict the HI masses of galaxies, but also yields similar number, stellar, halo mass and satellite fraction distributions to the HI detections retrieved from observational data.

Motivation & Objective

• To establish a statistical model linking HI gas content to both galaxy properties and host halo environments.
• To address the overprediction of HI detections when relying solely on galaxy scaling relations.
• To quantify the suppression of HI gas in halos of increasing mass, particularly above log Mh ≥ 12.0.
• To develop a predictive framework for future HI surveys like FAST and SKA using empirical constraints from SDSS and ALFALFA.

Proposed method

• Cross-matched 16,520 HI-galaxy pairs from ALFALFA with SDSS DR7 galaxy groups to create a halo-resolved dataset of 14,270 groups.
• Developed a galaxy-centric HI mass estimator (Model 1) using stellar mass and g−r color, achieving a scatter of σ ≈ 0.25 in HI mass prediction.
• Identified a significant overprediction of HI detections in Model 1 when applied to the full SDSS catalog, especially in massive halos.
• Introduced an 'HI survival efficiency' factor to model the probability of retaining HI gas in halos of different masses.
• Constructed a halo-based model (Model 2) that combines in situ galaxy relations with ex situ halo mass dependence to suppress overprediction.
• Validated the model using mock catalogs and compared to ALFALFA’s α.100 catalog in terms of number counts, HI mass functions, and satellite fractions.

Experimental results

Research questions

• RQ1 Why do galaxy-only HI mass estimators overpredict the number of HI detections in massive halos compared to ALFALFA observations?
• RQ2 How does the total HI mass in galaxy groups evolve with increasing halo mass, particularly for log Mh ≥ 11.0?
• RQ3 What is the role of halo mass and environment in suppressing or retaining HI gas within galaxy groups?
• RQ4 Can a hybrid model combining in situ galaxy relations and ex situ halo dependence accurately reproduce observed HI statistics?
• RQ5 How does the satellite fraction of HI-detected galaxies vary with halo mass and stellar mass in the observed data?

Key findings

• Model 1, based solely on galaxy properties (stellar mass and g−r color), overpredicts the number of HI detections by a significant margin, especially in halos with log Mh ≥ 12.0.
• The total HI mass in groups remains nearly constant for halo masses log Mh ≥ 11.0, indicating a saturation or suppression effect in massive halos.
• The introduction of an 'HI survival efficiency' factor—dependent on halo mass—successfully suppresses overprediction and brings model results into agreement with ALFALFA observations.
• Model 2 (the halo-based estimator) reproduces the observed number of HI detections, total HI mass distribution, and satellite fractions across different halo mass bins with high fidelity.
• The mock catalog generated by Model 2 shows excellent agreement with the α.100 catalog in terms of number counts, HI mass functions, and satellite fractions, while Model 1 shows large excesses, especially in massive groups.
• The model is validated across multiple statistical metrics and is deemed suitable for predicting HI content in future surveys such as FAST and SKA.