3D Spectroscopy in the Virtual Observatory: Current Status
Igor Chilingarian, Francois Bonnarel, Mireille Louys, Ivan Zolotukhin, Frederic Royer, Isabelle Jegouzo, Pierre Le Sidaner, Pierre Fernique, Thomas Boch
aa r X i v : . [ a s t r o - ph ] N ov
3D SPECTROSCOPY IN THE VIRTUAL OBSERVATORY: CURRENT STATUSIgor Chilingarian , Franc¸ois Bonnarel , Mireille Louys , Ivan Zolotukhin , Fr´ed´eric Royer , Isabelle J´egouzo ,Pierre Le Sidaner , Pierre Fernique , and Thomas Boch LERMA Observatoire de Paris, France VO-Paris Data Centre, France Sternberg Astronomical Institute, Russia CDS Observatoire de Strasbourg, France LSIIT, ULP, Strasbourg, France GEPI Observatoire de Paris, France
ABSTRACT
Three cornerstones for the 3D data support in the Vir-tual Observatory are: (1) data model to describe them,(2) data access services providing access to fully-reduceddatasets, and (3) client applications which can deal with3D data. Presently all these components became avail-able in the VO. We demonstrate an application of theIVOA Characterisation data model to description of IFUand Fabry-Perot datasets. Two services providing SSA-like access to 3D-spectral data and Characterisation meta-data have been implemented by us: ASPID-SR at SAORAS for accessing IFU and Fabry-Perot data from theRussian 6-m telescope, and the Giraffe Archive at theVO Paris portal for the VLT FLAMES-Giraffe datasets.We have implemented VO Paris Euro3D Client, handlingEuro3D FITS format, that interacts with CDS Aladinand ESA VOSpec using PLASTIC to display spatial andspectral cutouts of 3D datasets. Though the prototype weare presenting is yet rather simple, it demonstrates how3D spectroscopic data can be fully integrated into the VOinfrastructure.Key words: Virtual Observatory; 3D Spectroscopy; DataArchives; Data Models.
1. INTRODUCTION
Integral field (or 3D) spectroscopy is a modern techniquein astrophysical observing that was proposed by GeorgesCourt´es in the late 60’s. The idea is to get a spectrum forevery point in the field of view of a spectrograph. Severalinstrumental approaches in the optical domain (as well asNIR and near-UV) exist: scanning Fabry-Perot interfer-ometry, image slicing and transforming two-dimensionalfield of view into a slit using Integral-Field Unit (see re-view in P´econtal-Rousset et al., 2004 for a description of different image slicing techniques).At present, nearly all large telescopes in the world areequipped with 3D spectroscopic devices, and rapidlygrowing volume of data produced by them pose a numberof questions regarding the data discovery and retrieval. Inthis paper we demonstrate how 3D data are handled in aframework of the International Virtual Observatory.All 3D spectroscopic observations result in datasets hav-ing both spatial and spectral information. They are usu-ally referred as “datacubes”, though sometimes (in caseof IFU) they are not regularly gridded in spatial dimen-sions.There are three cornerstones for the 3D data support inthe Virtual Observatory:1. data model – an abstract, self-sufficient and stan-dardised description of the data2. data access services – archives, providing access tofully reduced science-ready datasets3. client applications – data-model aware software thatis able to search, retrieve, and display 3D data, aswell as to give a possibility for sophisticated scien-tific data analysisAll these blocks became available, and we will reviewthem in the forthcoming sections.
2. DATA MODEL
An abstract, self-sufficient and standardised descriptionof the astronomical data is known as a data model. Sucha description is constructed in a way to become sufficientfor any sort of data processing and analysis. The DataModeling working group (DM WG) of the InternationalVirtual Observatory Alliance (IVOA) is responsible fordefinition of data models for different types of astronom-ical data sets, catalogues, and more general concepts e.g.”quantity”.To describe 3D spectroscopic data we use Characterisa-tion Data Model (Louys et al. 2007). One of the mostabstract data models developed by the DM WG, it gives aphysical insight to the dataset, i.e. describes where, howextended and in which way the Observational or Simu-lated dataset can be described in a multidimensional pa-rameter space, having the following axes: spatial , time , spectral , observed (e.g. flux, polarimetric), as well asother arbitrary axes. For every axis the three characteri-sation properties are defined: coverage , resolution , and sampling . There are four levels of details in the descrip-tion of the dataset: (1) location or reference value – av-erage position of the data on a given parameter axis; (2) bounds , providing a bounding box; (3) support , describ-ing more precisely regions on a parameter axis as a set ofsegments; map , providing a detailed sensitivity map.Details about applying Characterisation Data Model tothe 3D spectroscopic datasets are given in Chilingarian etal. (2006). The algorithm for the characterisation meta-data computation is described there as well.
3. 3D DATA ARCHIVES
We have developed two data archives providing accessto fully-reduced “science-ready” IFU and IFP datasets:ASPID-SR and GIRAFFE Archive. For both archives theIVOA Simple Spectral Access (SSA, Tody et al. 2007)interfaces are provided.
ASPID stands for the ”Archive of Spectral, Photometric,and Interferometric Data”. The world largest collectionof raw 3D spectroscopic observations of galactic and ex-tragalactic sources is provided. ASPID-SR (Chilingarianet al. 2007) is a prototype of an archive of heteroge-neous science ready data, fed by ASPID, where we tryto take full advantage of the IVOA Characterisation DataModel. Multi-level Characterisation metadata is providedfor every dataset. The archive provides powerful meta-data querying mechanism (Zolotukhin et al. 2007) withaccess to every data model element, vital for the effi-cient scientific usage of a complex informational system.ASPID-SR is one of the reference implementation of theIVOA Characterisation Data Model. The datasets areprovided in several formats: stacked spectra, regularly-gridded data cubes, and Euro3D FITS.A high level of integration between the archive WEB in-terface and existing VO tools is provided (see next sec-tion).
GIRAFFE Archive (Royer et al., this conference)contents fully reduced data obtained with theFLAMES/Giraffe spectrograph at ESO VLT. Dataobtained with all three observing modes of Giraffe:MEDUSA (multi-object spectroscopy), IFU (multi-IFUspectroscopy), and ARGUS (single IFU) are provided.Raw datasets are taken from the ESO archive afterthe end of their proprietary period and reduced inan automatic way using the Giraffe data processingpipeline. There is a possibility of accessing individualextracted 1D spectra from the multi-object spectroscopicobservations, as well as full datasets in the Euro3D FITSformat.
4. CLIENT SOFTWARE
Presently, there is a number of VO tools available thatdeal with images (such as CDS Aladin) and 1-D spec-tra (ESA VOSpec, SpecView, SPLAT). However, none ofthem is able to handle IFU datasets.In a framework of the VO Paris project (Simon et al.2006) we have developed VO Paris Euro3D Client specif-ically to deal with the datasets in the Euro3D FITS formatin a VO context. This tool interacts with CDS Aladin todisplay position of the fibers (or slit) on the sky and dis-play individual extracted spectra in ESA VOSpec. Cata-logue of positions of fibers (or slit pixels) can be exportedas VOTable.VO Paris Euro3D Client is an open-source Java package,including basic functions for the Euro3D FITS I/O and agraphical user interface.Individual or co-added spectra can be extracted from theEuro3D FITS file and exported as VOTable serializationof the IVOA Spectrum Data Model 1.0 (McDowell et al.2007). All the interaction between applications is doneusing PLASTIC (PLatform for AStronomical Tool Inter-Connection) – a prototype of the VO application messag-ing protocol.Presently VO-Paris Euro3D Client is used as an inte-grated data visualising software at ASPID-SR - Science-Ready Data Archive at the Special Astrophysical Obser-vatory of Russian Academy of Sciences.In Fig. 1 we demonstrate how the interaction between VOclient applications and the ASPID-SR archive interface isimplemented. There are several stages:1. Querying the characterisation metadata using WEB-interface (see Louys et al., this conference)2. Light-weight Java applet is integrated into theHTML pages, containing query response; it de-tects a PLASTIC hub, connects to it, and checkswhether other tools (Aladin, VOSpec, VO ParisEuro3D Client) are registered within it. If the appli-cations are not detected, they will be started usingJavaScript and Java WebStart.3. As soon as all the used applications have beenstarted and registered within the PLASTIC hub, asmall script is sent to CDS Aladin to display theDSS2 image of the area, corresponding to the po-sition of the IFU spectrograph. At the same time,the IFU dataset in the Euro3D FITS format is loadedinto VO Paris Euro3D Client.4. Positions of IFU fibers are sent from VO ParisEuro3D Client to CDS Aladin and overplotted onthe DSS2 image.5. User can interactively select either groups of fibersor individual ones using CDS Aladin. An extractedspectrum (or co-added spectra of several fibers) issent to ESA VOSpec using PLASTIC by clickingon the corresponding button in the user interface ofVO Paris Euro3D Client.
5. SUMMARY
In Chilingarian et al. (2006) we concluded that ”allthe necessary infrastructural components exist for build-ing VO-compliant archives of science-ready 3D data andtools for dealing with them”. Since that time there wasa substantial progress of VO standards and protocols.Now we are able to provide access to first two such VO-compliant archives. This not only a “proof-of-concept”,but the services that can be used for real scientific pur-poses. Another important conclusion that can be drawnis that the present state of VO standards (including PLAS-TIC – a prototype of the VO application messaging proto-col) is totally sufficient for dealing with complex datasetsin a VO framework without need to develop new clientapplications for every particular kind of data.
ACKNOWLEDGMENTS
IC is grateful to ESAC and VO-Spain for providing fi-nancial support to attend the workshop. Special thanksto John Taylor (Astrogrid) and Isa Barbarisi (ESAC) forhelp with many technical points related to PLASTIC im-plementation in VO Paris Euro3D Client.
REFERENCES