|Introduction to the
The interpretation of observations of distant galaxy samples is limited
by our partial knowledge of the spectrophotometric properties of the
local field galaxy population. In particular, the range in properties
as a function of galaxy luminosity has not been studied systematically.
To remedy this situation, optical spectrophotometric and photometric
observations were obtained for 196 nearby galaxies that form a
representative sample of the local field population. These galaxies
span the entire Hubble sequence in morphological type, as well as a wide
range of luminosities and colors.
In the past decade deeper and higher resolution observations of the intermediate and high redshift universe have produced many surprising discoveries (see Ellis 1997 for a comprehensive review). These include (1) the large excess of faint blue galaxies compared to expectations (Broadhurst et al. 1988; Colless et al.1990), (2) moderate to strong luminosity evolution (Odewahn et al. 1996; Ellis et al. 1996; Lilly et al. 1995) which is morphologically segregated (Hammer et al. 1997; Driver et al. 1998), and (3) evolution of galaxy morphologies as evident from, e.g., the decline of barred spiral galaxies with redshift (Abraham et al. 1996), the increased rate of galaxy mergers (Glazebrook et al. 1995; Abraham et al. 1996; van Dokkum et al. 1999), and the appearance of ``chain galaxies'', very compact star forming galaxies and other apparently proto-galactic systems (Cowie, Hu & Songaila 1995; Steidel et al. 1996). At redshifts z 3 recent observations indicate a prevalence of small, proto-galactic, actively star forming systems (e.g., Stern & Spinrad 1999; Franx et al. 1997; Dey et al. 1998; Weymann et al. 1998; Spinrad et al. 1998; Chen, Lanzetta & Pascarelle 1999; van Breugel et al. 1999; Hu, McMahon, & Cowie 1999). The combined star formation rate in the universe appears to peak near a redshift of one, although the data do not exclude a constant high rate of star formation until z~1 and a steep decline ever since (Madau et al. 1996; Madau, Pozzetti & Dickinson 1998).
1. Comparison of Distant and Nearby Galaxies
The interpretation of these observations is seldom straightforward, due to the selection effects and biases involved in the comparison of distant and nearby galaxies (Koo & Kron, 1992). A major difficulty with the interpretation of distant spectroscopic data, for instance, is the lack of good nearby comparison samples (e.g., Koo & Kron 1992; Ellis 1997). Distant galaxies subtend small angles on the sky, comparable to typical spectrograph slit widths, and their spectra are -unavoidably- integrated spectra. The same size slits tend to sample only the nuclear regions of nearby galaxies. Galaxy morphology and surface photometry are sensitive to luminosity evolution, bandshifting, cosmological surface brightness dimming, and resolution effects. When discussing measurements at different redshifts, one needs to take the biases in the selection of nearby and distant galaxy samples into account, and ensure that a similar part of the galaxy population is sampled (e.g., Pritchet & Kline 1981; Bothun 1990; Bohlin et al. 1991; Koo & Kron 1992; Giavalisco et al. 1996; Frei, Guhathakurta & Gunn 1996; Brinchmann et al. 1998). A direct comparison of distant and nearby galaxies, therefore, is difficult.
2. Limitations of Current Nearby Samples
Despite the relative brightness of nearby galaxies, the effort involved in constructing a spectroscopic and photometric nearby reference sample is large. Whereas wide field CCD cameras are capable of imaging thousands of distant galaxies simultaneously and obtain optical spectra for up to several hundred such galaxies at a time (using aperture plates or fibers), observations of nearby galaxies are very inefficient because of the sheer area of sky that needs to be surveyed (Koo & Kron 1992).
Two pioneering efforts provided valuable local references. Gallagher, Bushouse & Hunter (1989) obtained large aperture spectra for 50 nearby galaxies with blue UBV colors. Although the range in morphological type and luminosity in this study are limited, a sizable portion of the luminosity function of blue galaxies is sampled. Kennicutt (1992) obtained integrated spectra for 90 nearby galaxies that spanned the entire Hubble sequence. These samples, particularly the latter because of the general availability of the spectra, continue to find wide use as a local benchmark and a zero-evolution reference. Uses for these data include:
Similar caveats, apply to the more recent galaxy spectral energy distributions obtained by Kinney et al. (1996), McQuade, Calzetti & Anderson (1995), and Storchi-Bergmann, Calzetti & Kinney (1994) for samples of star-forming, quiescent, and active galaxies. Although the apertures these authors used are smaller than Kennicutt's, their SEDs also include the ultraviolet regime.
As a service to the community, Frei et al. (1996) published a digital atlas of cosmetically uniform and calibrated g,r,i (Thuan & Gunn 1976) filter CCD images for 31 galaxies and of bJ and R (Gullixson et al. 1995) filter images for 82 galaxies. All galaxies in that atlas are nearby, bright, extend a large angle on the sky, show well resolved morphological features, and were selected to span the Hubble classification classes. The Frei et al. atlas has found wide application as local benchmark and as ``zero-evolution'' reference at higher redshifts (after artificial redshifting and simulated observation by instruments like the Hubble Space Telescope's WFPC2 camera [e.g., Abraham et al. 1996; Brinchmann et al. 1998; Bouwens et al. 1998; Abraham et al. 1999]). The Frei et al. sample, however, is known to be deficient in low luminosity and late type galaxies.
In summary, presently there exists in the literature no good representative nearby galaxy sample which is designed for comparison with observations at higher redshifts, and for which both large aperture integrated spectrophotometry, nuclear spectroscopy, as well as multi-filter surface photometry are available. The Nearby Field Galaxy Survey (NFGS), the work described in this thesis and its follow-up, is aimed to remedy this situation.
3. Nearby Field Galaxy Survey Project Goals
The purpose of the present study is to accurately describe the distributions in the local galaxy population of (1) the magnitudes and colors, (2) emission and absorption line strengths, (3) star formation rate and star formation history, (4) morphology, and (5) structural parameters.
To this end, integrated and nuclear spectrophotometry, as well as UBR surface photometry are obtained for a large sample of galaxies in the nearby field. The NFGS sample includes galaxies of all morphological types and spans as large a range in galaxy luminosity as was possible at the time this project was conceived. The present work thus aims to extend the work of Gallagher et al. (1989) and Kennicutt (1992) to lower luminosity galaxies across the Hubble sequence and to study the variation in star formation rates and history, and metallicity over a large range in absolute magnitude and type. The availability of uniform and accurate surface photometry is a valuable aid in the interpretation and analysis of the spectra.
The data may stimulate theoretical work and numerical modeling, once the relative volume densities per type and magnitude have been calibrated using larger complete samples (e.g., Carter et al. 1998, 1999).
4. Outline of this Thesis
The larger part of this thesis describes the acquisition, reduction, and calibration of an extensive spectrophotometric set of data on a representative local galaxy sample. The analysis presented here is a mere prelude to the many scientific applications the NFGS data will find. In the following a brief description is given of each of the chapters.
UBR Surface Photometry
In chapter 2 we describe the objective selection of the NFGS sample from the first CfA redshift survey (Huchra et al. 1983) and we discuss its merits and limitations. The selected 196 galaxies include all morphological types, span over 8 magnitudes in luminosity -reaching 5 magnitudes below the characteristic local galaxy luminosity-, and include several types of nuclear activity. With few caveats, this sample is shown to be representative of the local galaxy population.
We proceed to describe our observations at the F. L. Whipple Observatory's 1.2 m telescope, and the reduction and calibration of the U, B, and R filter data. Radial profiles of surface brightness and color are extracted from the images using an ellipse fitting technique (Franx, Illingworth, & Heckman 1989; Jørgensen, Franx, & Kjærgaard 1995). Isophotal and fractional (half-light and total) magnitudes, radii and colors are derived using these radial profiles. The quality of the data is evaluated and demonstrated to be high. The primary data products, the images and radial profiles of surface brightness and color, are presented as an atlas for future reference. The photometric measurements are presented in a table.
We conclude the chapter with a brief discussion of the correlations between morphological (Hubble) type, luminosity and color, and find that (1) Hubble type and galaxy color are strongly correlated: galaxies become progressively bluer towards later types, (2) lower luminosity systems tend to be bluer than luminous galaxies, (3) color-magnitude relations are useful for early-type systems and to a lesser degree for very late type systems, but not useful for intermediate-type spiral galaxies, (4) in a color-color diagram the galaxies lie on a well-defined relation with small scatter, despite the large range in galaxy properties, (5) the observed trend of redder colors with increasing galactocentric radius in half of the galaxies fainter than MB = -17 can be explained by star formation that dominates the local colors in the inner parts, and (6) both the average difference in color between inner and outer parts of a galaxy and the range in this difference from galaxy to galaxy are smaller in early-type than in spiral galaxies.
Nuclear and Integrated Spectrophotometry
The spectrophotometric part of the survey is described in chapter 3. We obtained nuclear and integrated spectra of all but a few of the target galaxies with the FAST spectrograph (Fabricant et al. 1998) at the F. L. Whipple Observatory's 1.5 m Tillinghast telescope. A controlled drift-scanning technique was used to obtain integrated spectra that, on average, sample 82 6% of the light within the B26 isophote, or 76 8% of the total galaxian light. All spectra include the 3550-7250 Å range and have a resolution of ~6Å (FWHM).
We describe the reduction and calibration of the data and critically evaluate the internal and external accuracy of our spectrophotometry. A comparison of photometric colors with synthetic colors measured in the spectra, demonstrates that our relative spectrophotometry is accurate to ~ 6%. The primary data products, the nuclear and integrated spectra are presented as a spectrophotometric atlas. Measurements of the strongest emission lines and the synthetic colors are presented as tables.
A short discussion focussing on the correlations between H and [O II]3727Å, morphological type, luminosity and color, as well as between nuclear and integrated emission-line strengths concludes the chapter. The main conclusions are, that (1) the range in continuum shape and emission line characteristics within a given morphological type class is large when one includes galaxies over a wide range in luminosity, (2) emission-line strengths increase and optical continua become bluer as the luminosity decreases, (3) the H//[O II] ratio systematically varies with galaxy luminosity, and (4) nuclear and integrated H emission-line strengths do not correlate well with one another.
[O II]: How Well Traceth the Tracer?
The forbidden [O II]3727Å doublet is an important tracer of the rate at which massive stars are formed (SFR) in a galaxy, and is often used as an alternative to H in distant galaxies, where H has shifted out of the optical wavelength regime or into a part of the spectrum where the earth's atmosphere hinders observations. Its excitation mechanism, however, is not as directly coupled to the intensity of the UV radiation from hot, young stars in H II regions as the H recombination line.
In chapter 4 we investigate the range and systematic variation in the observed H//[O II] emission-line ratio as a function of galaxy luminosity. The observed H//[O II] ratio is inversely correlated with galaxy luminosity (see also chapter 3), and varies near MB*, the characteristic absolute magnitude of the local luminosity function, by a factor of seven. It is shown that reddening and excitation differences are the main contributors to both scatter and systematic trend. Both are strongly correlated with absolute magnitude and are likely caused by the systematic variation in metallicity with galaxy luminosity. Excitation models (McCall, Rybski & Shields 1985) match the dust corrected line ratios within the accuracy of our data. The total variation in the ratio of the observed [O II] flux to the reddening corrected H flux is a factor 25.
The observed difference between the [O II]-SFR calibrations of Gallagher et al. (1989) and Kennicutt (1992) can be explained qualitatively as due to the difference in galaxy luminosity between the two samples. Systematic variations in reddening by dust as a function of galaxy luminosity plays an important role when interpreting emission line strengths in terms of SFRs.
When corrections for metallicity and dust are not possible, the use of [O II] fluxes to measure star formation rates may result in an overestimate of a factor of 3 if local calibrations for luminous galaxies are used. H is demonstrated to be a significantly better tracer of star formation than [O II]. Some empirical trends presented in this chapter, however, may help to estimate the likely range in reddening and may be useful in the high redshift regime.
The morphological properties of large samples of high redshift galaxy images observed with the Hubble Space Telescope's WFPC2 camera have been analyzed quantitatively in terms of two parameters, A and C (Abraham et al. 1996; Brinchmann et al. 1998). The former measures the deviations from axisymmetry and the latter traces the bulge/disk ratio. As several galaxy properties correlate with morphological type, automated morphological classification offers a quick way to obtain basic information on large numbers of galaxies. The A, C classification method has been calibrated using the Frei et al. (1996) sample of local galaxies, which sample (as mentioned) is deficient in low luminosity and late type galaxies. It also was selected subjectively in terms of the appearance of a galaxy to the eye.
In chapter 5 we present an alternative calibration using the NFGS sample as local reference. We investigate the dependence of A and C on the bandpass and isophote of measurement, on the spatial resolution of the images, on internal reddening by dust, and on star formation activity. We apply the results to the A, C measurements of Abraham et al. (1996) in the HST Medium Deep Survey and Hubble Deep Field.
We find that (1) lower luminosity galaxies tend to be more asymmetric and their light less concentrated than high luminosity galaxies, (2) the galaxian light distributions are less concentrated and more asymmetric in the rest-frame UV than in redder bands, (3) the degree to which a galaxy image appears asymmetric depends on the number and spatial distribution of star forming regions in a galaxy, but to some extent lower surface brightness features (as can be found in galaxy mergers or due to intrinsic asymmetry of the stellar mass distribution) contributes as well, and (4) even when band-shifting, spatial resolution effects, and biases in the local calibration sample are taken into account, evolution of the galaxy population is needed to explain the high fraction of asymmetric galaxies in the MDS and HDF.
5. The Nearby Field Galaxy Survey and Beyond
A summary of the work to date is presented above. Here, I conclude with a brief description of the follow up observations and analysis currently under way, and an outline of my ideas regarding future paths of research and future observations to complement the available data.
Data in Hand
Star Formation History
The colors of galaxies relate to the star formation history. A more precise method to derive the recent star formation history (up to 3 Gyr ago) based on the slope of the combined continuum and absorption lines of H(H8)+Ca II and H10 has been succesfully applied to early type galaxies in the Coma cluster (Caldwell et al. 1993, 1995). An extension to later type galaxies may be possible if other spectral information is included, as well. This information can be used to study the relation between the current high mass star formation rate (as measured from H) and the star formation history. In particular in low mass galaxies evidence exists that the star formation history may be complex with multiple active and quiescent periods.
Radial Dependence of Emission-Line Strength
In chapter 3 it was shown that emission line strengths measured in the nuclear and in the integrated spectra can differ significantly. Other authors (e.g., Zaritsky, Kennicutt & Huchra 1994) have measured emission-line strengths in individual H II regions at different radii in nearby galaxies to study the radial dependence of metallicity. From our reduced frames (both nuclear and integrated) we can extract spectra as a function of radius. The radial emission line strength gradients combined with surface brightness and color information holds information on the formation of galaxies.
More luminous galaxies tend to have higher metallicities. A well known linear relation exist between metallicity and absolute magnitude over more than 10 magnitudes. The slope, however, is not as well established. The Nearby Field Galaxy Survey sample can be used to study this relation over almost 8 magnitudes in luminosity using a uniform data set.
Search for Dwarf AGN
It has been shown by Ho et al. (1995,1996,1998) that the majority of galaxies exhibits a form of active galactic nucleus with an energy output that is many orders of magnitude lower than that of classical AGN. Its signature is akin to that of LINERS but can be detected only in the nuclear spectra after careful modeling of the continuum and photo-ionized emission.
Observations in Progress
Using the FAST spectrograph at the F.L. Whipple Observatory's 1.5 m telescope, Kannappan et al. (1999) have obtained optical rotation curves and central velocity dispersions for galaxies with and without emission, respectively, for the entire NFGS sample. Combination of the surface photometry and high resolution spectra allow construction of radial mass profiles and the measurement of the radial dependence of the M/L ratio as a function of morphological and spectrophotometric properties, to compare gas and stellar kinematics, and investigate kinematic evidence of galaxy interactions, mergers and mass infall.
Measurements of the mass enclosed within the half-light radius in B would provide valuable contraints on galaxy formation models and the dark matter hypothesis. The kinematic portion of this survey, however, is beyond the scope of this thesis. Preliminary results have been presented in Kannappan et al. (1999) and Kannappan & Fabricant (2000).
Near-Infrared Surface Photometry
The relative fraction of light in bulges and in disks is not accurately known. Present estimates date from the late 50s and are accurate to 10% at best. This fraction is one of the key tests of galaxy formation scenarios. For most of the NFGS sample Pahre et al. (in prep.) have obtained near-infrared H and K surface photometry, using the near-infrared array camera at the F. L. Whipple Observatory's 1.2 m telescope.
Improved Abundances in Low Metallicity Systems
Empirical scaling relations (Pagel et al. 1984) adopted in this thesis start to break down at metallicities lower than 0.6 solar. Accurate measurements of the weak [O III]4363Å emission line are needed for a direct measurement of the oxygen abundance in the interstellar gas. The 6Å resolution in our spectra is insufficient to detect and measure [O III]4363. Additional high resolution spectroscopy would be needed for ~ 75 galaxies in the present sample.
The near infrared regime counts several very strong emission lines that would allow a determination of the abundance ratios of the different elements. As the optical and near infrared lines sample different physical regions in a galaxy, constraints can be placed on the abundance variations in the interstellar gas.