FREE ELECTRONIC LIBRARY - Dissertations, online materials

Pages:   || 2 |

«Astronomy is the science devoted to the study of planets, stars, galaxies and the Universe as a whole. It is closely coupled to astrophysics where ...»

-- [ Page 1 ] --

Entry for the Encyclopedia for Statistical Science (on-line edition)

Submitted April 8 2008 (revised May 13 2008)


Author: Eric D. Feigelson

Keywords: astronomy, cosmology, multivariate analysis, time series, spatial

point processes, truncation, measurement error, Poisson processes

Astronomy is the science devoted to the study of planets, stars, galaxies and

the Universe as a whole. It is closely coupled to astrophysics where laws governing physical processes (e.g., gravity, electromagnetism, quantum mechanics) are used to model observed astronomical properties. Perhaps more than other physical sciences, astronomy is frequently statistical in nature. The objects under study are inaccessible to direct manipulation in the laboratory, so the astronomer is restricted to observing a few external characteristics and inferring underlying properties and physics.

As the oldest physical science, astronomy has served the development of statistical methodology for centuries. During the 17-18th centuries, in particular, important foundations of modern statistical theory were formulated to address astronomical problems; the astronomers were the statisticians. The fields diverged during much of the 20th century, as astronomy turned to physics for insights and statisticians turned to applications in human sciences and industries. In the past two decades, the cross-disciplinary field of astrostatistics has reemerged to deal both with important astrophysical issues and to treat mega-datasets produced by high-technology observatories.


Quantitative measurements of celestial phenomena were carried out by many ancient civilizations (1,2,3,4). The classical Greeks were not active observers but were unusually creative in the applications of mathematical principles to astronomy. Greek natural philosopher Hipparchus started a discussion which lasted for centuries on how inconsistent measurements should be combined to quantify a phenomenon. Finding scatter in Babylonian measurements of the length of a year, defined as the time between solstices, he took the middle of the range -- rather than the mean or median -- for the best value. Ptolemy similarly estimated parameters of a non-linear cosmological model using a minimax goodness-of-fit method. The Persian 11th century scientist Al-Biruni recommended a three-point procedure, and discussed the dangers of propagating errors from inaccurate instruments and inattentive observers.

While some medieval scholars advised against the acquisition of repeated measurements, fearing that errors would compound rather than compensate for each other, the utility of the mean to increase precision was demonstrated with great success in the 16th century by Tycho Brahe.

Celestial mechanics in the eighteenth century, in which Newton’s law of gravity was found to explain even the subtlest motions of heavenly bodies, required the derivation of a few interesting quantities from numerous inaccurate observations. This required advances in the understanding of statistical inference and error distributions. Mayer, in his 1750 study of lunar librations, suggested a procedure of reconciling a system of 27 inconsistent linear equations in three unknowns by solving the equations in groups. Laplace, in a 1787 analysis of the influence of Jupiter’s gravity on Saturn’s motion, suggested a more unified approach that led to Legendre’s invention of the least-squares method in an 1805 study of cometary orbits. Shortly thereafter, in an 1809 monograph on the mathematics of planetary orbits, Gauss first presented the normal (or Gaussian) distribution of errors in overdetermined systems of equations using a form of Bayes’ theorem, though the actual derivation was flawed. Legendre developed L2 least squares parameter estimation to model cometary orbits. The least-squares method rapidly became widely used in European astronomy and geodesy.

Many other individuals also contributed substantially to both astronomy and statistics Galileo proposed an L1 procedure for estimation and discussed observational errors concerning the distance to the supernova of 1572. Huygens wrote books on optics and Saturn and on probability in games of chance. Halley, famous for his early contributions to celestial mechanics, laid important foundations to mathematical demography and actuarial science. Bessel, codiscoverer of stellar parallaxes and the binary companion of Sirius, introduce the notion of probable error. Quetelet, founder of the Belgian Royal Observatory, led early applications of probability theory to the social sciences. Airy, a late-19th century British Royal astronomer, is known both for his text on statistical errors and his study of telescope optics.

Relatively few connections between astronomy and statistics are notable in the first decades of the 20th century. A subfield called ``statistical astronomy'' concentrated on issues involving positional astronomy, star counts and Galactic structure (5). Astronomers were mainly familiar with normal distributions and least-squares estimation methods as presented in volumes written for physical scientists, particularly (6). Fourier analysis commonly used for time series analysis. The 1970s witnesses some broadening of methodology as maximum likelihood estimation and the nonparametric Kolmogorov-Smirnov goodness-offit test gained popularity. Bayesian methods have become increasingly important since the 1990s.

The modern field of astrostatistics grew in the 1990s, stimulated by monographs on statistical aspects of astronomical image processing (7,8), galaxy clustering (9), Bayesian analyses (10) and astrostatistics in general (11).

The continuing conference series “Statistical Challenges in Modern Astronomy” (12,13) brought together astronomers and statisticians. Collaborations between astronomers and statisticians emerged such as the California/Harvard AstroStatistical Collaboration (14). The education of astronomers in statistical methodology remains weak. A Center for Astrostatistics has operated recently to run a methodological summer school for astronomers, provide computer codes, and otherwise liaise the communities (15). A selection of modern astrostatistical issues is reviewed in the remainder of this article.


Resources for astronomical measurements are increasingly focused on welldesigned large-scale surveys. These entail imaging large areas of the sky at various wavebands and/or obtaining spectra of large number of objects.

Perhaps the most influential recent project has been the Sloan Digital Sky Survey (16) which is producing terabytes of images, catalogs of ~200 million celestial objects, and spectra of ~1 million galaxies, quasars and stars. Several hundred publications have emerged since it began in 2000 on a wide range of astrophysical issues. Data from numerous surveys are freely available to the worldwide research community through the Virtual Observatory project (17).

A typical astronomical catalog is a table where rows represent individual objects and columns denote measurements of various properties: location in the sky (astrometry), brightness at different wavebands of light (photometry), morphological and categorical information. The structure of such tables can be fruitfully studied with well-established techniques of multivariate analysis and classification. However, astronomical multivariate databases differ from

those obtain in other fields due to three interrelated characteristics:

heteroscedastic measurement errors, censoring and truncation survey (18).

Based on careful calibration of the instruments and observation of source-free regions of the sky, astronomers directly measure the uncertainty of each photometric measurement. Thus, each brightness value is accompanied by a heteroscedastic measurement error with known variance. Analysis of the database thus needs multivariate methods with heteroscedastic weightings, which are generally not available. It is also common that some properties of a source are too faint to be detected, leading to left-censoring of the brightness based on the known measurement error (19). Astronomy thus needs extensions to survival analysis that permit censoring in any variable. Finally, the list of objects is often truncated by the sensitivity limits of the underlying.

Steps have been made in addressing these statistical challenges. Astronomer Lynden-Bell (20) and statistician Woodroofe (21) derived the nonparametric maximum-likelihood estimator for a randomly truncated sample. Other results include: a 2-sample test for randomly truncate datasets (22), an extension to Kendall’s tau correlation with censoring for a 3-variable database (23), a test for completeness in a truncated sample (24), a semiparametric estimator for an irregularly truncated sample (25), and a thorough Bayesian treatment of linear regression including heteroscedastic measurement error, censoring and truncation (26).


Despite the apparent constancy of celestial objects to the human eye, they exhibit an incredible array of variable phenomena when studied with modern telescopes (27, 28). Stellar rotations, pulsations, and orbits of binary stars and planetary systems produce periodic variations typically studied with Fourier analysis. Distant blazers produce aperiodic correlated variations. Accretion onto compact objects, including black holes, produces combinations of periodic and stochastic variations. Solar flares, supernovae, and gamma-ray bursts produce short-lived explosive variations. Variability is seen at radio, infrared, visible, X-ray and gamma-ray wavebands. Analysis is often hindered by unevenly-spaced and heteroscedastically weighted data points. An important method is a generalization of Schuster’s periodogram for unevenly-space data (29).

The analysis of binary star and planetary orbits has its roots in the analyses of Laplace based on Newtonian gravity. A binary star system follows an elliptical orbit but the modeling of its physical properties can be complex due to orbital eccentricity, inclination to the line of sight, tidal distortions, reflection of the other star’s light, and so forth. Binary star models, sometimes with dozens of nonlinearly related parameters, are typically found by least-squares procedure derived by (30). A similar recent problem is the characterization of planets orbiting other stars from sparse time series data. Figure 1 shows a model of three planets orbiting a nearby star similar to the Sun. This model, however, may not be unique. Sophisticated Bayesian model selection techniques based on Markov chain Monte Carlo calculations are now being applied to this important problem (31).



Certain fields of astronomy are devoted to the detection of individual particles or photons of light: cosmic ray and neutrino astronomy; X-ray and gamma-ray astronomy (33). The data are received as a sequence of events each with a position in the sky, energy and arrival time. Ancillary information is available from knowledge of the telescope and detectors. The goals are to identify and locate individual astronomical sources, to characterize their flux (intensity), spectrum and variability in the bands accessible to the instruments. An uninteresting background flux is present, either from noise in the detector or from cosmic emission which cannot be resolved into individual points.

Telescopes in these fields are very expensive and frequently are launched into space to avoid absorption or emission from the Earth’s atmosphere. Examples include the space-bourne Chandra X-ray Observatory and INTEGRAL gamma-ray observatory, the Pierre Auger Observatory in Argentina for cosmic rays, and the IceCube experiment in Antarctica for neutrinos.

Statistically, the datasets emerging from these observatories can be considered tagged Poisson processes in 4-dimensions with heteroscedastic measurement errors of known variance. While strong sources are easily found, there is a challenge in detecting the faintest sources with only a few events or only barely above background levels. Methods include sliding windows, wavelet and adaptive filtering (7, 8, 34). Variable sources are identified by the Kolmogorov-Smirnov test and are can be modeled as a sequence of constant segments derived from a Bayesian likelihood analysis (35).

When a source is not evidently present above the background, astronomers seek an upper limit to the flux that can then be considered a left-censored data point in population studies. The Poisson upper limit problem has been widely discussed and no consensus has emerged on a best method (36). A common procedure used in the particle physics community where similar problems are encountered is the Feldman-Cousins maximum likelihood limit (37). A related matter involves the ratio of two Poisson-distributed variables;

this is often called the hardness ratio when the quantities are the counts in two spectral bands for a single source. Here the maximum likelihood solution is unstable (38). A Bayesian solution based on MCMC computations has been developed when background counts must be subtracted from both quantities (39).


Over the past century, several lines of astronomical inquiry powerfully indicate that the Universe began in a Big Bang explosion about 14 billion years ago (40).

The evidence includes: the redshifts of galaxies proportional to their distance from us (Hubble’s Law); the blackbody cosmic microwave background (CMB) radiation from the early radiation-dominated era; the cosmic abundances of helium and other light element isotopes; and the statistical distribution of galaxies in space.

The galaxy distribution can be viewed as a spatial point process in three dimensions, two from location in the sky and one from redshift as a surrogate for distance. Galaxy redshift surveys reveal remarkable nonlinear, anisotropic clustering. They are preferentially located along curved sheets surrounding voids in a pattern resembling a collection of soap bubbles (Figure 2 top, 41).

Pages:   || 2 |

Similar works:

«ASTRONOMY: C. G. ABBOT 82 PROC. N. A. S.THE LARGER OPPORTUNITIES FOR RESEARCH ON THE RELATIONS OF SOLAR AND TERRESTRIAL RADIATION' By C. G. ABBOT SMITHSONIAN ASTROPHYSICAL OBSERVATORY Communicated by the National Research Council, December 8, 1919 The earth is maintained at it§ present temperature by a balance between the solar radiation received and the terrestrial radiation emitted. The mean intensity of the solar radiation as it is at the earth's mean distance outside the atmosphere is...»

«Chapter 2 Filters and General Equipment for Astronomical Observing If you are an avid visual astronomer, it may come as little surprise to find that colored filters make a world of difference to your observing clarity, and their use brings out detail from the subtle shadings found on Solar System objects. On occasion, a telescope manufacturer will include a set of colored filters along with the telescope you have purchased, although many amateur astronomers tend to ignore their use, and the...»

«Jahresbericht 2009 Mitteilungen der Astronomischen Gesellschaft 93 (2010), 101–118 Bochum Ruhr-Universität Bochum Institut für Theoretische Physik IV Weltraumund Astrophysik Universitätsstraße 150, 44780 Bochum Telefon: +49 (234) 32-22032, Telefax: +49 (234) 32-14177 E-Mail: rsch@tp4.ruhr-uni-bochum.de WWW: http://www.tp4.ruhr-uni-bochum.de 1 Personal und Ausstattung 1.1 Personalstand Professoren und Privatdozenten Prof. Dr. Julia Becker [-23779] (seit 1.6.2009), PD. Dr. Horst Fichtner...»

«Astrological World Cycles By Tara Mata (Laurie Pratt) Published in Serial Format in 1932-33 IN a series of articles, of which this is the first, the writer proposes to demonstrate the profound connection of an astronomical phenomenon, known as the Precession of the Equinoxes, with the history of mankind and the great cycles of the world. The true Age, or Yuga, of the present world-period, in reference to the Grand Cycle of Time, symbolized by the stars in their courses, will be pointed out, and...»

«US Particle Physics: Scientific Opportunities A Strategic Plan for the Next Ten Years Report of the Particle Physics Project Prioritization Panel 29 May 2008 Three frontiers of research in particle physics form an interlocking framework that addresses fundamental questions about the laws of nature and the cosmos.US Particle Physics: Scientific Opportunities A Strategic Plan for the Next Ten Years Report of the Particle Physics Project Prioritization Panel 29 May 2008 i In November 2007, at the...»

«Website; www.tauranga-astro.150m.com Correspondence to Secretary; 14 Hazel Terrace, Tauranga Issue No. 26 – March 2010 The Moon, taken through Andrew Walker’s 12-inch Telescope in his backyard observatory. 29 Brown Street, Tauranga, P.O. Box 1147 Phone 07 578 0059, Fax 07 578 2692 A visit to the CERN Large Hadron Collider and other interesting highlights of my July 2009 tour of Europe By Ursula Macfarlane To treat myself for my 50th birthday I impulsively decided to go on a trip to Europe....»

«University of Szeged (SZTE) Doctoral School of English and American Literature and Culture Nóra Borthaiser Performing New Cosmoses: Environmental Matters and Disability Issues in Disney’s Animations between 1990–2010 PhD Dissertation Supervisors: Réka M. Cristian, PhD Zoltán Vajda, PhD Szeged Spring, 2012 TABLE OF CONTENTS Table of Contents Acknowledgments Abstract / Összefoglaló 1. Introduction 2. From Mickey Mouse to the Desk Lamp Story-telling in the 20th and 21st Centuries:...»

«General Introduction to Ancient Astronomical Concepts and to Babylonian Methods -31Basic Concepts The four basic circles and how to locate things Terrestrial Equator: The great circle perpendicular to the pole of rotation of the earth. Terrestrial Longitude: The angular distance of a point on the earth projected onto the terrestrial equator. It is measured from a favored location (e.g. Alexandria). Since the end of the last century this is measured from Greenwich. In the nineteenth century,...»

«Sexagesimonoveno Período Ordinario de Sesiones de la Comisión 17 de enero de 1997 Lima – Perú DECISION 398 Transporte Internacional de Pasajeros por Carretera, sustitutoria de la Decisión 289 LA COMISION DEL ACUERDO DE CARTAGENA, VISTOS: El Capítulo XI del Acuerdo de Cartagena, la Decisión 289 de la Comisión, y la Propuesta 292/Mod. 1 de la Junta; CONSIDERANDO: Que el transporte internacional de pasajeros por carretera constituye uno de los instrumentos de ayuda eficaz para la...»

«Constellations, Fixed Stars and the Zodiac in Islamic Astronomy Author: Salim Ayduz IMPORTANT NOTICE: Chief Editor: Lamaan Ball All rights, including copyright, in the content of this document are owned or controlled for these purposes by FSTC Limited. In Production: Faaiza Bashir accessing these web pages, you agree that you may only download the content for your own personal non-commercial use. You are not permitted to copy, broadcast, download, store (in any medium), transmit, show or play...»

«Diamond Jubilee Symposium on Pulsars 14–17 March 1994 Raman Research Institute, Bangalore J. Astrophys. Astr. (1995) 16, 233–244 The Population of Binary and Millisecond Pulsars R. N. Manchester Australia Telescope National Facility, CSIRO, P. O. Box 76, Epping NSW 2121, Australia. Abstract. Recent searches for millisecond and binary pulsars are reviewed, with particular emphasis on the nearly complete Parkes southern survey. Correlations between several of the major parameters of these...»

«Space Radiation Risk L imits and E arth-Moon-M ars E nvironmental Models Francis A. Cucinotta1,*, Shaowen Hu2, Nathan A. Schwadron3, K. Kozarev3, Lawrence W. Townsend4 and Myung-Hee Y. Kim2 1,* F. Cucinotta (Corresponding Author) NASA Lyndon B. Johnson Space Center 2101 NASA Parkway, Houston, Texas 77058, francis.a.cucinotta@nasa.gov, M.Y. Kim, and S. Hu USRA Division of Space Life Sciences Houston TX 77058 myung-hee.y.kim@nasa.gov; shaowen.hu-1@nasa.gov N. A. Schwadron, Kamen Kozarev Boston...»

<<  HOME   |    CONTACTS
2016 www.dissertation.xlibx.info - Dissertations, online materials

Materials of this site are available for review, all rights belong to their respective owners.
If you do not agree with the fact that your material is placed on this site, please, email us, we will within 1-2 business days delete him.