Документ взят из кэша поисковой машины. Адрес оригинального документа : http://www.adass.org/adass/proceedings/adass02/reprints/P3-4.pdf Дата изменения: Wed Mar 12 01:45:55 2003 Дата индексирования: Tue Oct 2 09:48:50 2012 Кодировка: Поисковые слова: п п п п п п п п п п п п п п п

Astronomical Data Analysis Software and Systems XII ASP Conference Series, Vol. 295, 2003 H. E. Payne, R. I. Jedrzejewski, and R. N. Hook, eds.

XAssist: A System for the Automation of X-ray Astrophysics Analysis
A. Ptak Johns Hopkins University R. Griffiths Carnegie Mel lon University Abstract. XAssist is a NASA AISR-funded pro ject for the automation of X-ray astrophysics, with emphasis on galaxies. It is nearing completion of its initially funded effort, and is working well for Chandra and ROSAT data. Initial support for XMM-Newton data is present as well. It is capable of data reprocessing, source detection, and preliminary spatial, temporal and spectral analysis for each source with sufficient counts. The bulk of the system is written in Python, which in turn drives underlying software (CIAO for Chandra data, etc.). Future work will include a GUI (mainly for beginners and status monitoring) and the exposure of at least some functionality as web services. The latter will help XAssist to eventually become part of the VO, making advanced queries possible, such as determining the X-ray fluxes of counterparts to HST or SDSS sources (including the use of unpublished X-ray data), and add the ability of "on-the-fly" X-ray processing. Pipelines are running on ROSAT, Chandra and now XMM-Newton observations of galaxies to demonstrate XAssist's capabilities, and the results are available online (in real time) at http://www.xassist.org. XAssist itself as well as various associated pro jects are available for download.

1.

Introduction

We are currently in a renaissance for X-ray astronomy, with two ma jor missions, Chandra and XMM-Newton, currently operating. These missions are producing large amounts of archival data, which is supplementing existing databases from missions such as ROSAT and ASCA. Historically, only X-ray "experts" usually attempted the analysis of X-ray data. This is because there were fundamental differences in the analysis of X-ray data compared with other bandpasses, most notably the fact that individual photons are detected as opposed to the accumulation of (only) spectra or images. Most modern X-ray detectors are imaging spectrometers so each observation results in a photon list from which images, spectra and light curves can be extracted. In general the numbers of photons are small so Poissonian rather than Gaussian statistical methods must be used. The spectral and spatial resolution of most detectors is moderate at best and forward-fitting convolution methods are needed to properly fit the data. All of 465 c Copyright 2003 Astronomical Society of the Pacific. All rights reserved.


466

Ptak & Griffiths

these factors limit the accessibility of the X-ray data to non-experts. In addition each mission tends to have its own unique software package for the reduction and analysis of the data, and X-ray data often require reprocessing as the calibration improves. These latter two factors also limit the ability of experts to take advantage of all available data for a given pro ject, particularly large-scale surveys. We have developed a software package to address these concerns. XAssist is capable of performing data reduction and preliminary analysis for ROSAT, Chandra and XMM-Newton data. It is fully automatic making it well-suited for surveys, as well as for the reprocessing of existing data. Below we will discuss its capabilities and prospects for the future. 2. Capabilities

ist currently has the following capabilities: Downloads data Reprocesses data Creates exposure maps and detector masks (if possible) Detects sources (using built-in routine for ROSAT and ASCA, CIAO wavdetect for Chandra, etc.) · Fits each source with "simple" (i.e., not including point-spread function) model to establish source extent and (Poisson-correct) significance (using the stand-alone python program ximgfit1 ) · Flags extended, confused and problematic sources · Computes median (or mean) background level · Excludes times of high background · Extracts spectra, "postage stamp" images, and light curves of each source for more detailed analysis (a simple power-law model is fit to sources with more than 100 sources) · Optionally performs chip-by-chip analysis (relevant just for Chandra right now) · Analysis can be restricted to an energy band · Large emphasis on detailed reporting · Looks for correlations of X-ray sources with astronomy databases and provides links on the detailed source web reports to query Simbad (see below) Note that in the case of Chandra analysis, most of the data reduction steps are based on the "threads" reported in the Chandra web site.2 Most parameters controlling XAssist are read from IRAF/FTOOLS-style parameter file, and can be set on the command-line as well (allowingfor the automated setup of XAssist for surveys). XAssist can be run (and configured) interactively (with a textbased interface).
1 2

XAss · · · ·

http://www.xassist.org/xassist/Download.jsp http://cxc.harvard.edu


XAssist: Automation of X-ray Astrophysics Analysis 3. Sample Output

467

Figure 1 shows images created as part of the report for the processing of the Chandra observation NGC 1569. Figure 2 shows the detailed report generated for a source. While there are admittedly "warts" that occur (as in any automated system), this example demonstrates that even in moderately crowded fields the system performs well and continues to the point of fitting a power-law spectrum model to the source spectrum. Obviously, this opens a powerful possibility for virtual observatories, namely that searches could be performed on highlevel quantities such as spectral slope. Even though human inspection would of course still be necessary for science-grade results, an automated system such as this could cull samples and produce usable results for many of the sources, both of which would be particularly useful prior to observing proposal deadlines (especially if the data of interest had only recently become publically available and had not been published yet). Queries are also submitted to HEASARC to find correlations of X-ray sources with 2µass, USNO, FIRST, and other catalogs.

Figure 1. Chandra ACIS (left) and DSS (right) images of NGC 1569. Detected X-ray sources are marked in both images. The color code is: green = point source, red = problematic/questionable source, blue = extended source, magenta = asymmetric source (may be extended), cyan = estimated detector boundary.

4.

Pipelines

To demonstrate XAssist's capabilities, pipelines3 have been set up to run XAssist on ROSAT HRI and Chandra observations of RC3 galaxies. The ROSAT analysis is nearing completion, and crontab jobs are checking for the public release of Chandra observations that may containing RC3 galaxies in the FOV of
3

http://www.xassist.org/xassist/Pipelines.jsp


468

Ptak & Griffiths

Source 2

Source

Background

Results Source position = 4 30 48.1 64 50 50.58 Spatial Source region radii = 1.8", 1.5" Background annulus = 2.3"-4.5" Nh = 0.543 (0.46 0.637) Gamma = 4.05 (3.64 4.52) Final chisq(dof) = Spectral 69.171(50) F(0.5-2.0 keV) = 4.009e-14 F(2.0-10.0 keV) = 1.448e-14 K-S test against local background: d = 0.408762, prob = 0.000000 Temporal K-S test against global background: d = 0.468536, prob = 0.000000 Counterparts Return to source list Check SIMBAD

Plot

Detailed fitting results

Figure 2. Detailed web page report for a source in the Chandra observation of NGC 1569. view, and these data sets are downloaded (using the scr for download). XMM-Newton has recently been added and we are in the process of establishing a pipeline for pipeline products are searchable and work is in progress fields to be added to the processing lists. 5. Current and Future Work ipt cda.py also available as a supported mission, public XMM data. The to allow users to request

Tasks include: (i) finishing off details (especially for ASCA and ROSAT PSPC analysis), (ii) better installation and configuration support (including the creation of a GUI), (iii) web access to XAssist, and (iv ) web services: limited status reporting is already available using SOAP and it will be possible to request full or partial processing of data via SOAP requests.