skip to main content

Title: Examining the Limits of an Artificial Neural Network in Predicting the HI Content of Galaxies
The neutral hydrogen (HI) in galaxies provides the gas reservoir out of which stars are formed. The ability to determine the HI masses for statistically significant samples of galaxies can provide information about the connection between this gas reservoir and the star formation that drives galaxy evolution. However, there are relatively few galaxies for which HI masses are known because these measurements are significantly more difficult to make than optical observations. Artificial neural networks are a type of nonlinear technique that have been used estimate the gas masses from their optical properties (Teimoorinia et al. 2017). We present HI observations of 51 galaxies with gas and stellar properties that are rare in the Arecibo Legacy Fast ALFA Survey (ALFALFA, Haynes et al. 2018) which was used to train the Artificial Neural Network developed by Teimoorinia et al. (ANN, 2017). These sources provide a test of the Artificial Neural Network predictions of HI mass and include some rare and interesting systems including galaxies that are extremely massive in both stellar mass (log M_∗> 11.0) and HI mass (log M_HI> 10.2) with large HI line widths (w_50> 500 km/s). We find that this Artificial Neural Network systematically overestimates the gas fraction of more » the galaxies in our selected sample, suggesting that care must be taken when using these techniques to predict gas masses for galaxies from a broad range of optical properties. « less
Authors:
; ; ;
Award ID(s):
1637339
Publication Date:
NSF-PAR ID:
10097647
Journal Name:
American Astronomical Society, AAS Meeting
Volume:
233
Page Range or eLocation-ID:
260.01
Sponsoring Org:
National Science Foundation
More Like this
  1. To create a Baryonic Tully-Fisher Relationship (BTFR) for the ALFALFA galaxies, we require a corresponding catalog of optical photometry to produce reliable stellar masses to complement their HI masses. Different methods of calculating stellar mass have been used in derivations of the BTFR depending on the sky distribution, prior redshift information or multiwavelength photometry of the sample galaxies. In order to obtain stellar masses for all of the ALFALFA galaxies in the SDSS photometric footprint, we use a crossmatch file between the α-100 catalog and the SDSS DR14 database to produce a reliable catalog of optical photometric properties for use in the stellar mass calculation and inclination-dependent corrections. For the stellar masses, we use the methodology set forth in Taylor et al. (2011), specifically, the modeled mass to light ratio estimated from the g-i color index. In order to compare stellar masses derived in this way with results obtained by other authors in the literature who used Petrosian and Sersic masses available in the NASA Sloan Atlas (NSA) to derive stellar masses, we investigate the differences in stellar mass and the corrections for inclination and extinction between values complied in the NSA and ones obtained using our α-100-SDSS cross-match. Wemore »investigate systematic differences in stellar mass estimates based on photometric properties such as color, surface brightness and concentration and on others such as distance, velocity width and HI richness. This research is supported by the Brinson Foundation for the Arecibo Pieces-Perseus Supercluster Survey (APPSS) and NSF grant NSF/AST-1714828 to M.P. Haynes.« less
  2. The ALFALFA blind extragalactic survey has populated the faint end of the neutral hydrogen (HI) mass function with statistical confidence for the first time. Of particular interest is a subset of the ALFALFA detections, termed "ultra-compact high-velocity clouds" (UCHVCs). These systems, if located within ~1 Mpc, would populate the lowest-mass end of the HI mass function. Subsequent optical imaging has revealed that some of these UCHVCs harbor associated (though sparse) stellar populations, revealing that they may be some of the most extreme galaxies known in the Local Volume, with optical properties akin to ultra-faint dwarf galaxies but with significant neutral gas reservoirs. In this campaign, we investigate the neutral hydrogen properties of six UCHVC candidate galaxies using deep VLA HI spectral line imaging. A companion poster (Paine et al.) presents details on the data reduction, imaging, and resulting products. Here, we examine the morphological and kinematic properties of selected sources. We apply the modeling software 3D-Barolo to our deep HI images in order to derive the rotation curve and constrain the inclination angle for each source. Successful modeling allows us to determine the dynamical masses of these objects and thus to consider them in the context of various fundamental scalingmore »relations defined by more massive galaxies.« less
  3. The “Survey of HI in Extremely Low-mass Dwarfs” (SHIELD) is a multiwavelength study of local volume low-mass galaxies drawn from the Arecibo Legacy Fast ALFA (ALFALFA) catalog. HST/Spitzer joint program GO-12658 revealed the stellar populations of the first 12 SHIELD galaxies (Cannon et al. 2011), allowing accurate distance measurements (McQuinn et al. 2014) and detailed studies of the patterns of recent star formation in each galaxy (McQuinn et al. 2015). These HST and Spitzer images are a critical interpretive benchmark for ground-based optical imaging and spectroscopy (Haurberg et al. 2015), as well as for sensitive VLA HI spectral line imaging of the SHIELD galaxies (McNichols et al. 2016; Teich et al. 2016). These results have furthered our understanding of the evolution of galaxies in a mass regime that was previously only sparsely populated. With the low-redshift ALFALFA catalog now complete, the scope of the SHIELD program has been expanded to include all 82 galaxies that meet distance, line width, and HI flux criteria for being gas-rich, low-mass galaxies. In HST program 13750, images of 18 more SHIELD galaxies have again set the physical scales for supporting HI spectral line imaging with both the VLA and the WSRT (Gordon et al.more »2016). Taken as a whole, the ongoing SHIELD program is one of the most comprehensive multiwavelength studies of the physical properties of low-mass galaxies outside of the Local Group.« less
  4. The ALFALFA blind extragalactic survey has populated the faint end of the neutral hydrogen (HI) mass function with statistical confidence for the first time. Of particular interest is a subset of the ALFALFA detections, termed "ultra-compact high-velocity clouds" (UCHVCs). These systems, if located within ~1 Mpc, would populate the lowest-mass end of the HI mass function. Subsequent optical imaging has revealed that some of these UCHVCs harbor associated (though sparse) stellar populations, revealing that they may be some of the most extreme galaxies known in the Local Volume, with optical properties akin to ultra-faint dwarf galaxies but with significant neutral gas reservoirs. In this campaign, we investigate the neutral hydrogen properties of six UCHVC candidate galaxies using deep VLA HI spectral line imaging. A companion poster (Bralts-Kelly et al.) presents 3D kinematic modeling of selected sources. Here, we show the imaging products and discuss the morphological and kinematic properties of the six chosen sources: AGC 198606, AGC 215417, AGC219656, AGC 249525, AGC 258237, and AGC 268069.
  5. The Arecibo Pisces-Perseus Supercluster Survey (APPSS) aims to observationally measure the dark matter mass density of Pisces-Perseus by detecting the peculiar velocities of galaxies falling onto the supercluster. To do this, APPSS will measure galaxies' distances using the Baryonic Tully Fisher Relation (BTFR), which relates a galaxy's baryonic mass and rotational velocity. Recovering the signature of infall as robustly as possible requires a careful choice of rotational velocity measurement, as the use of various velocity definitions changes the scatter and systematics of the relation. We introduce and compare multiple automated methods for measuring a galaxy's rotational velocity using its unresolved line profile. The velocities discussed include global HI profile width measures commonly reported in large surveys, velocity widths derived from best-fit parametrizations to profiles, and velocity widths derived using more novel methods including the spectral line's curve of growth and neural network-derived velocities which incorporate information about the profile's width and shape. We compare these velocity measures by finding best-fit BTFR relations for two samples of galaxies - the SPARC sample and a selected sample of gas-dominated ALFALFA galaxies (Papastergis et al. 2016). With these best-fit BTFRs, we compare intrinsic scatters and residual correlations with source properties to investigate howmore »velocity choice affects the absolute and systematic uncertainties of BTFR-derived galaxy distances. This research is supported by NSF/AST-1714828 and the Brinson Foundation.« less