arXiv subjects

A. J. Connolly

Publications and source records attributed to A. J. Connolly.

At least 37 records · Page 2Linked to original sources

arXiv2002

A feature at z ~ 3.2 in the evolution of the Ly-alpha forest optical depth

The effective optical depth in the Ly-alpha forest region of 1061 low-resolution QSO spectra drawn from the SDSS database decreases with decreasing redshift over the range 2.5 < z < 4. Although the evolution is relatively smooth, tau_eff ~ (1+z)^{3.8 pm 0.2}, at z ~ 3.2 the effective optical depth decreases suddenly, by about ten percent with respect to this smoother evolution. It climbs back to the original smooth scaling again by z ~ 2.9. We describe two techniques, one of which is new, for quantifying this evolution which give consistent results. A variety of tests show that the feature is not likely to be a consequence of how the QSO sample was selected, nor the result of flux calibration or other systematic effects. Other authors have argued that, at this same epoch, the temperature of the IGM also shows a departure from an otherwise smooth decrease with time. These features in the evolution of the temperature and the optical depth are signatures of the reionization of He II.

astro-ph↗

arXiv2001

A new source detection algorithm using FDR

The False Discovery Rate (FDR) method has recently been described by Miller et al (2001), along with several examples of astrophysical applications. FDR is a new statistical procedure due to Benjamini and Hochberg (1995) for controlling the fraction of false positives when performing multiple hypothesis testing. The importance of this method to source detection algorithms is immediately clear. To explore the possibilities offered we have developed a new task for performing source detection in radio-telescope images, Sfind 2.0, which implements FDR. We compare Sfind 2.0 with two other source detection and measurement tasks, Imsad and SExtractor, and comment on several issues arising from the nature of the correlation between nearby pixels and the necessary assumption of the null hypothesis. The strong suggestion is made that implementing FDR as a threshold defining method in other existing source-detection tasks is easy and worthwhile. We show that the constraint on the fraction of false detections as specified by FDR holds true even for highly correlated and realistic images. For the detection of true sources, which are complex combinations of source-pixels, this constraint appears to be somewhat less strict. It is still reliable enough, however, for a priori estimates of the fraction of false source detections to be robust and realistic.

astro-ph↗

arXiv2001

Early-type galaxies in the SDSS

A sample of nearly 9000 early-type galaxies, in the redshift range 0.01<z<0.3, was selected from the Sloan Digital Sky Survey using morphological and spectral criteria. The sample was used to study how early-type galaxy observables, including luminosity L, effective radius R, surface brightness I, color, and velocity dispersion V, are correlated with one another. Measurement biases are understood with mock catalogs which reproduce all of the observed scaling relations. At any given redshift, the intrinsic distribution of luminosities, sizes and velocity dispersions in our sample are all approximately Gaussian. In the r* band L ~ V^3.91, L ~ R^1.58, R ~ I^(-0.75), and the Fundamental Plane relation is R ~ V^(1.49) I^(-0.75). These relations are approximately the same in the g*, i* and z* bands. At fixed luminosity, the mass-to-light ratio scales as M/L ~ L^0.14. The g*-r* color scales as V^0.25. Color also correlates with magnitude and size, but these correlations are entirely due to the L-V and R-V relations. Chemical evolution and star formation histories are investigated using co-added spectra of similar objects in our sample. Chemical abundances correlate primarily with velocity dispersion. At fixed V, the higher redshift population is bluer, is weaker in Mg2, and is stronger in Hbeta than the population nearby. In addition, the population at higher redshifts is slightly more luminous. These differences are consistent with that of a passively evolving population which formed the bulk of its stars about 9 Gyrs ago. The Fundamental Plane suggests that galaxies in dense regions are slightly different from those in less dense regions, but the co-added spectra and color--magnitude relations show no statistically significant dependence on environment.

astro-ph↗

arXiv2001

Computational AstroStatistics: Fast and Efficient Tools for Analysing Huge Astronomical Data Sources

I present here a review of past and present multi-disciplinary research of the Pittsburgh Computational AstroStatistics (PiCA) group. This group is dedicated to developing fast and efficient statistical algorithms for analysing huge astronomical data sources. I begin with a short review of multi-resolutional kd-trees which are the building blocks for many of our algorithms. For example, quick range queries and fast n-point correlation functions. I will present new results from the use of Mixture Models (Connolly et al. 2000) in density estimation of multi-color data from the Sloan Digital Sky Survey (SDSS). Specifically, the selection of quasars and the automated identification of X-ray sources. I will also present a brief overview of the False Discovery Rate (FDR) procedure (Miller et al. 2001a) and show how it has been used in the detection of ``Baryon Wiggles'' in the local galaxy power spectrum and source identification in radio data. Finally, I will look forward to new research on an automated Bayes Network anomaly detector and the possible use of the Locally Linear Embedding algorithm (LLE; Roweis & Saul 2000) for spectral classification of SDSS spectra.

astro-ph↗

arXiv2001

Evidence for Reionization at z ~ 6: Detection of a Gunn-Peterson Trough in a z=6.28 Quasar

We present moderate resolution Keck spectroscopy of quasars at z=5.82, 5.99 and 6.28, discovered by the Sloan Digital Sky Survey (SDSS). We find that the Ly Alpha absorption in the spectra of these quasars evolves strongly with redshift. To z~5.7, the Ly Alpha absorption evolves as expected from an extrapolation from lower redshifts. However, in the highest redshift object, SDSSp J103027.10+052455.0 (z=6.28), the average transmitted flux is 0.0038+-0.0026 times that of the continuum level over 8450 A < lambda < 8710 A (5.95 150, and is consistent with zero flux in the Ly Alpha forest region immediately blueward of the Ly Alpha emission line, compared with a drop by a factor of ~10 at z(abs)~5.3. A similar break is seen at Ly Beta; because of the decreased oscillator strength of this transition, this allows us to put a considerably stronger limit, tau(eff) > 20, on the optical depth to Ly Alpha absorption at z=6. This is a clear detection of a complete Gunn-Peterson trough, caused by neutral hydrogen in the intergalactic medium. Even a small neutral hydrogen fraction in the intergalactic medium would result in an undetectable flux in the Ly Alpha forest region. Therefore, the existence of the Gunn-Peterson trough by itself does not indicate that the quasar is observed prior to the reionization epoch. However, the fast evolution of the mean absorption in these high-redshift quasars suggests that the mean ionizing background along the line of sight to this quasar has declined significantly from z~5 to 6, and the universe is approaching the reionization epoch at z~6.

astro-ph↗

arXiv2001

Spectroscopic Target Selection for the Sloan Digital Sky Survey: The Luminous Red Galaxy Sample

We describe the target selection and resulting properties of a spectroscopic sample of luminous, red galaxies (LRG) from the imaging data of the Sloan Digital Sky Survey (SDSS). These galaxies are selected on the basis of color and magnitude to yield a sample of luminous, intrinsically red galaxies that extends fainter and further than the main flux-limited portion of the SDSS galaxy spectroscopic sample. The sample is designed to impose a passively-evolving luminosity and rest-frame color cut to a redshift of 0.38. Additional, yet more luminous, red galaxies are included to a redshift of 0.5. Approximately 12 of these galaxies per square degree are targeted for spectroscopy, so the sample will number over 100,000 with the full survey. SDSS commissioning data indicate that the algorithm efficiently selects luminous (M_g=-21.4), red galaxies, that the spectroscopic success rate is very high, and that the resulting set of galaxies is approximately volume-limited out to z=0.38. When the SDSS is complete, the LRG spectroscopic sample will fill over 1h^-3 Gpc^3 with an approximately homogeneous population of galaxies and will therefore be well suited to studies of large-scale structure and clusters out to z=0.5.

astro-ph↗

arXiv2001

Global star formation history: A Local Group perspective

The global star formation rate (SFR) density is estimated from the star formation histories (SFHs) of Local Group galaxies. This is found to be broadly consistent with estimates of the global SFH from existing redshift surveys for two favoured cosmologies. It also provides additional evidence for a relatively constant global SFR density at high redshift (z>1).

astro-ph↗

arXiv2001

Towards a resolution of the discrepancy between different estimators of star formation rate

Different wavelength regimes and methods for estimating the space density of the star-formation rate (SFR) result in discrepant values. While it is recognised that ultra-violet (UV) and H-alpha emission line data must be corrected for the effects of extinction, the magnitude of the required correction is uncertain. Even when these corrections are made there remains a significant discrepancy between SFRs derived from UV and H-alpha measurements compared with those derived from far-infrared (FIR) and radio luminosities. Since the FIR/radio derived SFRs are not affected by extinction, and simple corrections to reconcile the UV and H-alpha measurement with these do not fully account for the discrepancies, a more sophisticated correction may be required. Recent results suggest that at least part of the solution may be a form of extinction which increases with increasing SFR (or luminosity, given the common assumption that SFR is proportional to luminosity). We present an analysis of the effects of a dust reddening dependent on star formation rate applied to estimators of SFR. We show (1) that the discrepancies between H-alpha and FIR/radio SFR estimates may be explained by such an effect, and we present an iterative method for applying the correction; and (2) UV-based estimates of SFR are harder to reconcile with FIR/radio estimates using this method, although the extent of the remaining discrepancy is less than for a non-SFR-dependent correction. Particularly at high redshift, our understanding of extinction at UV wavelengths may require a still more complex explanation.

astro-ph↗

arXiv2001

A New Very Cool White Dwarf Discovered by the Sloan Digital Sky Survey

Early data taken during commissioning of the SDSS have resulted in the discovery of a very cool white dwarf. It appears to have stronger collision induced absorption from molecular hydrogen than any other known white dwarf, suggesting it has a cooler temperature than any other. While its distance is presently unknown, it has a surprisingly small proper motion, making it unlikely to be a halo star. An analysis of white dwarf cooling times suggests that this object may be a low-mass star with a helium core. The SDSS imaging and spectroscopy also recovered LHS 3250, the coolest previously known white dwarf, indicating that the SDSS will be an effective tool for identifying these extreme objects.

astro-ph↗

arXiv2000

Star Formation in Galaxies between 0.7 < z < 1.8

We present an analysis of the star formation rate in galaxies between 0.7<z<1.8 using Near Infrared Camera and Multi-Object Spectrograph (NICMOS) grism spectral observations. We detect 163 galaxies in an area of ~4.4 square arcminutes, 37 of which show possible H-alpha emission. We extend the observed H-alpha luminosity function (LF) in this redshift range to luminosities a factor of two fainter than earlier work, and are consistent in the region of overlap. Using the H-alpha LF, we estimate a star formation rate (SFR) density in this redshift range of 0.166 M_sun/yr/Mpc^3 (H0=75 km/s/Mpc), consistent with other estimates based on emission lines, and supporting the order of magnitude increase in SFR density between z=0 and z=1. Our measurement of SFR density is a factor of ~2-3 greater than that estimated from UV data, comparable to the factor observed locally, implying little evolution in the relative extinctions between UV and H-alpha out to z~1.3.

astro-ph↗

arXiv2000

Fast Algorithms and Efficient Statistics: Density Estimation in Large Astronomical Datasets

In this paper, we outline the use of Mixture Models in density estimation of large astronomical databases. This method of density estimation has been known in Statistics for some time but has not been implemented because of the large computational cost. Herein, we detail an implementation of the Mixture Model density estimation based on multi-resolutional KD-trees which makes this statistical technique into a computationally tractable problem. We provide the theoretical and experimental background for using a mixture model of Gaussians based on the Expectation Maximization (EM) Algorithm. Applying these analyses to simulated data sets we show that the EM algorithm - using the AIC penalized likelihood to score the fit - out-performs the best kernel density estimate of the distribution while requiring no ``fine--tuning'' of the input algorithm parameters. We find that EM can accurately recover the underlying density distribution from point processes thus providing an efficient adaptive smoothing method for astronomical source catalogs. To demonstrate the general application of this statistic to astrophysical problems we consider two cases of density estimation: the clustering of galaxies in redshift space and the clustering of stars in color space. From these data we show that EM provides an adaptive smoothing of the distribution of galaxies in redshift space (describing accurately both the small and large-scale features within the data) and a means of identifying outliers in multi-dimensional color-color space (e.g. for the identification of high redshift QSOs). Automated tools such as those based on the EM algorithm will be needed in the analysis of the next generation of astronomical catalogs (2MASS, FIRST, PLANCK, SDSS) and ultimately in in the development of the National Virtual Observatory.

astro-ph↗

arXiv2000

Computational AstroStatistics: Fast Algorithms and Efficient Statistics for Density Estimation in Large Astronomical Datasets

We present initial results on the use of Mixture Models for density estimation in large astronomical databases. We provide herein both the theoretical and experimental background for using a mixture model of Gaussians based on the Expectation Maximization (EM) Algorithm. Applying these analyses to simulated data sets we show that the EM algorithm - using the both the AIC & BIC penalized likelihood to score the fit - can out-perform the best kernel density estimate of the distribution while requiring no ``fine-tuning'' of the input algorithm parameters. We find that EM can accurately recover the underlying density distribution from point processes thus providing an efficient adaptive smoothing method for astronomical source catalogs. To demonstrate the general application of this statistic to astrophysical problems we consider two cases of density estimation; the clustering of galaxies in redshift space and the clustering of stars in color space. From these data we show that EM provides an adaptive smoothing of the distribution of galaxies in redshift space (describing accurately both the small and large-scale features within the data) and a means of identifying outliers in multi-dimensional color-color space (e.g. for the identification of high redshift QSOs). Automated tools such as those based on the EM algorithm will be needed in the analysis of the next generation of astronomical catalogs (2MASS, FIRST, PLANCK, SDSS) and ultimately in the development of the National Virtual Observatory.

astro-ph↗

arXiv2000

Five High-Redshift Quasars Discovered in Commissioning Imaging Data of the Sloan Digital Sky Survey

We report the discovery of five quasars with redshifts of 4.67 - 5.27 and z'-band magnitudes of 19.5-20.7 M_B ~ -27. All were originally selected as distant quasar candidates in optical/near-infrared photometry from the Sloan Digital Sky Survey (SDSS), and most were confirmed as probable high-redshift quasars by supplementing the SDSS data with J and K measurements. The quasars possess strong, broad Lyman-alpha emission lines, with the characteristic sharp cutoff on the blue side produced by Lyman-alpha forest absorption. Three quasars contain strong, broad absorption features, and one of them exhibits very strong N V emission. The amount of absorption produced by the Lyman-alpha forest increases toward higher redshift, and that in the z=5.27 object (D_A ~ 0.7) is consistent with a smooth extrapolation of the absorption seen in lower redshift quasars. The high luminosity of these objects relative to most other known objects at z >~ 5 makes them potentially valuable as probes of early quasar properties and of the intervening intergalactic medium.

astro-ph↗

arXiv1999

Reconstructing Galaxy Spectral Energy Distributions from Broadband Photometry

We present a novel approach to photometric redshifts, one that merges the advantages of both the template fitting and empirical fitting algorithms, without any of their disadvantages. This technique derives a set of templates, describing the spectral energy distributions of galaxies, from a catalog with both multicolor photometry and spectroscopic redshifts. The algorithm is essentially using the shapes of the templates as the fitting parameters. From simulated multicolor data we show that for a small training set of galaxies we can reconstruct robustly the underlying spectral energy distributions even in the presence of substantial errors in the photometric observations. We apply these techniques to the multicolor and spectroscopic observations of the Hubble Deep Field building a set of template spectra that reproduced the observed galaxy colors to better than 10%. Finally we demonstrate that these improved spectral energy distributions lead to a photometric-redshift relation for the Hubble Deep Field that is more accurate than standard template-based approaches.

astro-ph↗

arXiv1999

A Robust Classification of Galaxy Spectra: Dealing with Noisy and Incomplete Data

Over the next few years new spectroscopic surveys (from the optical surveys of the Sloan Digital Sky Survey and the 2 degree Field survey through to space-based ultraviolet satellites such as GALEX) will provide the opportunity and challenge of understanding how galaxies of different spectral type evolve with redshift. Techniques have been developed to classify galaxies based on their continuum and line spectra. Some of the most promising of these have used the Karhunen and Loeve transform (or Principal Component Analysis) to separate galaxies into distinct classes. Their limitation has been that they assume that the spectral coverage and quality of the spectra are constant for all galaxies within a given sample. In this paper we develop a general formalism that accounts for the missing data within the observed spectra (such as the removal of sky lines or the effect of sampling different intrinsic rest wavelength ranges due to the redshift of a galaxy). We demonstrate that by correcting for these gaps we can recover an almost redshift independent classification scheme. From this classification we can derive an optimal interpolation that reconstructs the underlying galaxy spectral energy distributions in the regions of missing data. This provides a simple and effective mechanism for building galaxy spectral energy distributions directly from data that may be noisy, incomplete or drawn from a number of different sources.

astro-ph↗

arXiv1998

The Statistical Approach to Quantifying Galaxy Evolution

Studies of the distribution and evolution of galaxies are of fundamental importance to modern cosmology; these studies, however, are hampered by the complexity of the competing effects of spectral and density evolution. Constructing a spectroscopic sample that is able to unambiguously disentangle these processes is currently excessively prohibitive due to the observational requirements. This paper extends and applies an alternative approach that relies on statistical estimates for both distance (z) and spectral type to a deep multi-band dataset that was obtained for this exact purpose. These statistical estimates are extracted directly from the photometric data by capitalizing on the inherent relationships between flux, redshift, and spectral type. These relationships are encapsulated in the empirical photometric redshift relation which we extend to z ~ 1.2, with an intrinsic dispersion of dz = 0.06. We also develop realistic estimates for the photometric redshift error for individual objects, and introduce the utilization of the galaxy ensemble as a tool for quantifying both a cosmological parameter and its measured error. We present deep, multi-band, optical number counts as a demonstration of the integrity of our sample. Using the photometric redshift and the corresponding redshift error, we can divide our data into different redshift intervals and spectral types. As an example application, we present the number redshift distribution as a function of spectral type.

astro-ph↗

arXiv1998

Simultaneous Multicolor Detection of Faint Galaxies in the Hubble Deep Field

We present a novel way to detect objects when multiband images are available. Typically, object detection is performed in one of the available bands or on a somewhat arbitrarily co-added image. Our technique provides an almost optimal way to use all the color information available. We build up a composite image of the N passbands where each pixel value corresponds to the probability that the given pixel is just sky. By knowing the probability distribution of sky pixels (a chi-square distribution with N degrees of freedom), the data can be used to derive the distribution of pixels dominated by object flux. From the two distributions an optimal segmentation threshold can be determined. Clipping the probability image at this threshold yields a mask, where pixels unlikely to be sky are tagged. After using a standard connected-pixel criterion, the regions of this mask define the detected objects. Applying this technique to the Hubble Deep Field data, we find that we can extend the detection limit of the data below that possible using linearly co-added images. We also discuss possible ways of enhancing object detection probabilities for certain well defined classes of objects by using various optimized linear combinations of the pixel fluxes (optimal subspace filtering).

astro-ph↗

arXiv1998

Evolution of the Angular Correlation Function

For faint photometric surveys our ability to quantify the clustering of galaxies has depended on interpreting the angular correlation function as a function of the limiting magnitude of the data. Due to the broad redshift distribution of galaxies at faint magnitude limits the correlation signal has been extremely difficult to detect and interpret. We introduce a new technique for measuring the evolution of clustering. We utilize photometric redshifts, derived from multicolor surveys, to isolate redshift intervals and calculate the evolution of the amplitude of the angular 2-pt correlation function. Applying these techniques to the the Hubble Deep Field we find that the shape of the correlation function, at z=1, is consistent with a power law with a slope of -0.8. For z>0.4 the best fit to the data is given by a model of clustering evolution with a comoving r0 = 2.37 Mpc and eps = -0.4 +/- 0.5, consistent with published measures of the clustering evolution. To match the canonical value of r0 = 5.4 Mpc, found for the clustering of local galaxies, requires a value of eps = 2.10 +/- 0.5 (significantly more than linear evolution). The log likelihood of this latter fit is 4.15 less than that for the r0 = 2.37 Mpc model. We, therefore, conclude that the parameterization of the clustering evolution of (1+z)^-(3+eps) is not a particularly good fit to the data.

astro-ph↗