Robust Estimation of the Physical Properties of z~2 Emission-line Galaxies
Open Access
- Author:
- Bowman, Will
- Graduate Program:
- Astronomy and Astrophysics
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- May 13, 2022
- Committee Members:
- Sarah Shandera, Outside Unit & Field Member
Michael Eracleous, Major Field Member
Randy McEntaffer, Major Field Member
Rebekah Dawson, Program Head/Chair
Robin Ciardullo, Chair & Co-Dissertation Advisr
Joel Leja, Major Field Member
Donghui Jeong, Major Field Member
Caryl Gronwall, Dissertation Co-Advisor - Keywords:
- galaxy formation
galaxy evolution
spectral energy distribution fitting - Abstract:
- Our knowledge of galaxy formation and evolution has exploded over the past few decades and we are now in an era of large galaxy surveys consisting of millions of galaxies. These samples enable statistical characterization of galaxy populations, helping to empirically explore the manner in which galaxies evolve through time. Current and future missions will push the envelope further by identifying millions of galaxies, many of which will be selected via their strong emission lines. We are motivated to develop ever-more sophisticated statistical methods to extract the maximal amount of information from these surveys. In Chapter 2, I describe the details of identifying a large sample of z~2 galaxies selected from Hubble Space Telescope (HST) grism frames on the basis of their strong rest-frame optical emission lines, with [O III] being the strongest in the vast majority of these systems. I also present the basis physical properties of the sample, including their rest-frame UV and optical size, stellar mass, UV-based star formation rate (SFR), and dust content. In order to provide context for how this emission-line galaxy (ELG) sample relates to the broader galaxy population at this epoch, I identify a comparison sample selected on the basis of their photometric redshifts and compare the two samples' physical properties. The ELG sample has systematically lower stellar masses, SFRs, and dust contents compared to the photometric redshift sample. This comparison indicates that identifying galaxies on the basis of their strong emission lines is an efficient way to find low-mass systems but is biased against objects with large amounts of dust. In Chapter 3, I measure the luminosity function and star-formation rate density of the sample introduced in Chapter 2. This measurement directly informs the observing strategy that is required for upcoming missions, which will use rest-frame optical ELGs to measure the large scale structure of the Universe. Since the precision in this measurement is directly related to the number of sources that are identified, accurate estimates of the line luminosity function are essential for designing the observing strategies of these flagship missions. I introduce a new, flexible spectral energy distribution (SED) code, MCSED, in Chapter 4. This galaxy fitting tool is specifically optimized to allow for varying assumptions about the physics and interplay of stars, dust, and gas in galaxies. Since the properties of these various components change with redshift and host galaxy type, and are constantly being improved from both an empirical and theoretical standpoint, flexible SED fitting tools are essential for extracting the maximal amount of science from new surveys. I apply this code to the z~2 galaxy sample introduced in Chapter 2 using a flexible model with multiple parameters to describe the dust attenuation and star formation history. The sample exhibits clear evolution in the star formation histories, dust contents, and average spectra across the three orders of magnitude in stellar mass. As the stellar mass increases, the objects become redder due to both a higher dust content and a larger population of old, red stars. Finally, Chapter 5 presents an expanded framework for fitting the SEDs of galaxies that accounts for correlated and non-Gaussian uncertainties in the photometric flux measurements. Such covariances are nearly universal in modern techniques for measuring photometry, yet these correlations have not been taken into account in SED fitting until now. This method for propagating variances and covariances throughout the analysis pipeline will be particularly important for upcoming high-precision cosmology experiments that rely upon photometric redshifts.