identification of important vo spectral services benefiting from deployment on the grid
DESCRIPTION
Petr Škoda Astronomical Institute Academy of Sciences Ondřejov Czech Republic. Identification of Important VO Spectral Services benefiting from deployment on the GRID. EURO-VO DCA Workshop „GRID and the Virtual Observatory“ Garching 9.4.2008. Outline of the Talk. - PowerPoint PPT PresentationTRANSCRIPT
![Page 1: Identification of Important VO Spectral Services benefiting from deployment on the GRID](https://reader034.vdocument.in/reader034/viewer/2022050820/56814ffc550346895dbdc438/html5/thumbnails/1.jpg)
Identification of Important VO Spectral Services benefiting from deployment
on the GRID
Petr ŠkodaAstronomical Institute Academy of Sciences
Ondřejov Czech Republic
EURO-VO DCA Workshop „GRID and the Virtual Observatory“Garching 9.4.2008
![Page 2: Identification of Important VO Spectral Services benefiting from deployment on the GRID](https://reader034.vdocument.in/reader034/viewer/2022050820/56814ffc550346895dbdc438/html5/thumbnails/2.jpg)
Outline of the Talk
Current support for spectra in VOCurrent support for spectra in VO Clients (basic analysis)
The proper VO way The proper VO way Services Postprocessing of spectra – on server From individual programs to workflows Why GRID (not VOrg but workflows) Massive spectra reduction
![Page 3: Identification of Important VO Spectral Services benefiting from deployment on the GRID](https://reader034.vdocument.in/reader034/viewer/2022050820/56814ffc550346895dbdc438/html5/thumbnails/3.jpg)
Outline of the Talk
Advanced methods of spectra analysisAdvanced methods of spectra analysis Detection of ES planets in spectra Automatic spectra classification Disentangling of spectra Variability – Period analysis Doppler imaging, tomography Magnetic fields (Zeeman-Doppler imaging) Week signatures of Mg field (Lorentz force)
Conclusions for future development of VOConclusions for future development of VO
![Page 4: Identification of Important VO Spectral Services benefiting from deployment on the GRID](https://reader034.vdocument.in/reader034/viewer/2022050820/56814ffc550346895dbdc438/html5/thumbnails/4.jpg)
SPLAT-VO VO Client for analysis (SSA and local files) 1D VO Client for analysis (SSA and local files) 1D
FITS (CRVAL1, CDELT1)FITS (CRVAL1, CDELT1)
![Page 5: Identification of Important VO Spectral Services benefiting from deployment on the GRID](https://reader034.vdocument.in/reader034/viewer/2022050820/56814ffc550346895dbdc438/html5/thumbnails/5.jpg)
SpecView
VO Client for analysis (SSA and local files) 1D VO Client for analysis (SSA and local files) 1D FITS (CRVAL1, CDELT1), modelsFITS (CRVAL1, CDELT1), models
![Page 6: Identification of Important VO Spectral Services benefiting from deployment on the GRID](https://reader034.vdocument.in/reader034/viewer/2022050820/56814ffc550346895dbdc438/html5/thumbnails/6.jpg)
VOspec (SLAP,TSAP)
![Page 7: Identification of Important VO Spectral Services benefiting from deployment on the GRID](https://reader034.vdocument.in/reader034/viewer/2022050820/56814ffc550346895dbdc438/html5/thumbnails/7.jpg)
IAU XXVII General Assembly, Prague, SpS3
7
SDSS Spectrum ServicesPCA similar spectra
![Page 8: Identification of Important VO Spectral Services benefiting from deployment on the GRID](https://reader034.vdocument.in/reader034/viewer/2022050820/56814ffc550346895dbdc438/html5/thumbnails/8.jpg)
VO key feature Interoperability Interoperability
All current work of an astronomer done in ONE GUIAll current work of an astronomer done in ONE GUI
Transparent search, download, conversionTransparent search, download, conversion
Unified presentation - different relationsUnified presentation - different relations
Background computing on GRIDSBackground computing on GRIDS
results in WS or DB Remote control - batch observation - robotic telescopeRemote control - batch observation - robotic telescope
results DB ADASS quotation : The telescope is a database with very long ADASS quotation : The telescope is a database with very long
time accesstime access
![Page 9: Identification of Important VO Spectral Services benefiting from deployment on the GRID](https://reader034.vdocument.in/reader034/viewer/2022050820/56814ffc550346895dbdc438/html5/thumbnails/9.jpg)
Architecture of VO
![Page 10: Identification of Important VO Spectral Services benefiting from deployment on the GRID](https://reader034.vdocument.in/reader034/viewer/2022050820/56814ffc550346895dbdc438/html5/thumbnails/10.jpg)
Spectra postprocessing
cutout service (ranges in SSAP )cutout service (ranges in SSAP ) echelle spectra – orders, overlap lacking certain line (in SSAP – names of photometric
regions - how wide ? ) Rebinning (change od dlambda in echelle)Rebinning (change od dlambda in echelle) Instrument profile (de)convolutionInstrument profile (de)convolution Broadening functions (rotation, limb dark)Broadening functions (rotation, limb dark) RV shift RV shift on server side (Pleinpot WWW pipeline)on server side (Pleinpot WWW pipeline)
![Page 11: Identification of Important VO Spectral Services benefiting from deployment on the GRID](https://reader034.vdocument.in/reader034/viewer/2022050820/56814ffc550346895dbdc438/html5/thumbnails/11.jpg)
Workflows - Pipelines
From programs on local PCFrom programs on local PC to distributed network (batch queue)to distributed network (batch queue) Condor - GRID?Condor - GRID? legacy applicationslegacy applications
parameters, filenames need to run in proper order
(setairmass - rvcorr)
waiting for files (usually quick) Workflows natural but conservatism !Workflows natural but conservatism !
![Page 12: Identification of Important VO Spectral Services benefiting from deployment on the GRID](https://reader034.vdocument.in/reader034/viewer/2022050820/56814ffc550346895dbdc438/html5/thumbnails/12.jpg)
Echelle Spectra Problems
![Page 13: Identification of Important VO Spectral Services benefiting from deployment on the GRID](https://reader034.vdocument.in/reader034/viewer/2022050820/56814ffc550346895dbdc438/html5/thumbnails/13.jpg)
Massive parallel spectra reduction
Workflows – Gasgano (UVES)Workflows – Gasgano (UVES) Using theoretical spectra to reduce echelle Using theoretical spectra to reduce echelle
spectra spectra continuum points – real 1.0 shape of blaze function in Balmer lines (convolution
with instrument profile) Complex echelle spectra reduction Complex echelle spectra reduction
(Piskunov 2002 – clustering, COP solved – not supposed)
Background model - HAMSCAT
![Page 14: Identification of Important VO Spectral Services benefiting from deployment on the GRID](https://reader034.vdocument.in/reader034/viewer/2022050820/56814ffc550346895dbdc438/html5/thumbnails/14.jpg)
Advanced Tools for Spectral Analysis
Simple tasks – done by VO clientsvisualisaton, overplot, RV, cont. fitting?experienced user, spectrum by spectrum
Require several variables from FITS
JD, time, epoch and derived variables
RV, line position, period – phase
and POST-PROCESSED spectrum
![Page 15: Identification of Important VO Spectral Services benefiting from deployment on the GRID](https://reader034.vdocument.in/reader034/viewer/2022050820/56814ffc550346895dbdc438/html5/thumbnails/15.jpg)
Dynamic Spectra
Quotient, Difference Quotient, Difference template (average)template (average)
For study of LPV For study of LPV (asteroseismology, winds)(asteroseismology, winds)
Requires Requires
time (JD) - winds
period (see Period analysis) - phase (LPV)
change of template (average, median)
removing bad data (interactive overplotting)
![Page 16: Identification of Important VO Spectral Services benefiting from deployment on the GRID](https://reader034.vdocument.in/reader034/viewer/2022050820/56814ffc550346895dbdc438/html5/thumbnails/16.jpg)
Dynamic Spectra
Interactive features, color cuts, LUT Interactive features, color cuts, LUT GRID (many stars)GRID (many stars) legacy packages, custom tasks legacy packages, custom tasks
D. Massa – IDL MIDAS TSA - (Stahl, Rivinius)
Multiple lines at the same timeMultiple lines at the same time Not yet molecular lines !
![Page 17: Identification of Important VO Spectral Services benefiting from deployment on the GRID](https://reader034.vdocument.in/reader034/viewer/2022050820/56814ffc550346895dbdc438/html5/thumbnails/17.jpg)
Spectral Disentangling
For blended spectra of binary (multiple) starsFor blended spectra of binary (multiple) stars Very powerfulVery powerful Requires good orbital coverage, estimate of Requires good orbital coverage, estimate of
orbital parameters (SIMBAD) orbital parameters (SIMBAD) Wavelength space disentangling - computing Wavelength space disentangling - computing
power, space (Simon&Sturm)power, space (Simon&Sturm) Fourier disentangling - perfect continuum, cut Fourier disentangling - perfect continuum, cut
regions, log lambda (Hadrava)regions, log lambda (Hadrava)
![Page 18: Identification of Important VO Spectral Services benefiting from deployment on the GRID](https://reader034.vdocument.in/reader034/viewer/2022050820/56814ffc550346895dbdc438/html5/thumbnails/18.jpg)
Spectra Disentangling in Fourier Space - KOREL
HD208905: Koubsky et al. 2006
![Page 19: Identification of Important VO Spectral Services benefiting from deployment on the GRID](https://reader034.vdocument.in/reader034/viewer/2022050820/56814ffc550346895dbdc438/html5/thumbnails/19.jpg)
KOREL - Many Spectra Overploted
V436 Per Janík 2003
![Page 20: Identification of Important VO Spectral Services benefiting from deployment on the GRID](https://reader034.vdocument.in/reader034/viewer/2022050820/56814ffc550346895dbdc438/html5/thumbnails/20.jpg)
Seaching ES Planets
Planet search on spectra (x photometry)Planet search on spectra (x photometry) precise RV precise RV
Iodine cells Changes in Line ProfilesChanges in Line Profiles
Bisector (different turbulence, flows, granulation) spots Another light !
![Page 21: Identification of Important VO Spectral Services benefiting from deployment on the GRID](https://reader034.vdocument.in/reader034/viewer/2022050820/56814ffc550346895dbdc438/html5/thumbnails/21.jpg)
Deconvolution of Iodine Lines
Desidera 1996
Models
Computation
Observation (different instruments - FTS)
![Page 22: Identification of Important VO Spectral Services benefiting from deployment on the GRID](https://reader034.vdocument.in/reader034/viewer/2022050820/56814ffc550346895dbdc438/html5/thumbnails/22.jpg)
Line Profile- Bisectors
Gray 2005Fiorenzano PhD 2006
![Page 23: Identification of Important VO Spectral Services benefiting from deployment on the GRID](https://reader034.vdocument.in/reader034/viewer/2022050820/56814ffc550346895dbdc438/html5/thumbnails/23.jpg)
Classification of Stellar Spectra
automatic classification of stellar spectraautomatic classification of stellar spectra direct chi^2 minimization (SFIT code, line
broadening – Jefferey 2001) Simplex AMOEBA or Levenberg-Marquardt)
Genetic Algorithm) Artificial Neural networks PCA – template spectrum + differences
example – Hot subdwarf filterexample – Hot subdwarf filter Ch. Winter 2006 – PhD thesis
![Page 24: Identification of Important VO Spectral Services benefiting from deployment on the GRID](https://reader034.vdocument.in/reader034/viewer/2022050820/56814ffc550346895dbdc438/html5/thumbnails/24.jpg)
WD models by manual fitting interpolation by experience
Kawka, Vennes 2005
![Page 25: Identification of Important VO Spectral Services benefiting from deployment on the GRID](https://reader034.vdocument.in/reader034/viewer/2022050820/56814ffc550346895dbdc438/html5/thumbnails/25.jpg)
Automatic classification engine
Winter 2006
Workflow
Parallel
![Page 26: Identification of Important VO Spectral Services benefiting from deployment on the GRID](https://reader034.vdocument.in/reader034/viewer/2022050820/56814ffc550346895dbdc438/html5/thumbnails/26.jpg)
Classification od hot subdwarfs
Winter 2006
![Page 27: Identification of Important VO Spectral Services benefiting from deployment on the GRID](https://reader034.vdocument.in/reader034/viewer/2022050820/56814ffc550346895dbdc438/html5/thumbnails/27.jpg)
Automatic EW measurement
Normalisation !(highest points)
tricky on hot stars
(on echelles) require synt. spectrum
Zhao et al. 2006
![Page 28: Identification of Important VO Spectral Services benefiting from deployment on the GRID](https://reader034.vdocument.in/reader034/viewer/2022050820/56814ffc550346895dbdc438/html5/thumbnails/28.jpg)
Changes of EW in Time
Omi Cas : Koubsky et al. 2004 HD6226 : Slechta and Skoda 2004
Batchmode processing, workflowsBatchmode processing, workflows Intrinsically parallel – simple algorithmIntrinsically parallel – simple algorithm Result – one number – plot for each line in timeResult – one number – plot for each line in time
![Page 29: Identification of Important VO Spectral Services benefiting from deployment on the GRID](https://reader034.vdocument.in/reader034/viewer/2022050820/56814ffc550346895dbdc438/html5/thumbnails/29.jpg)
Time variability - Current VO Support
Period analysis – NONE but Period analysis – NONE but SIGSPEC (P. Regen) – used to MOST Period04 (P. Lentz)
Power spectrum FT Theta statistics
![Page 30: Identification of Important VO Spectral Services benefiting from deployment on the GRID](https://reader034.vdocument.in/reader034/viewer/2022050820/56814ffc550346895dbdc438/html5/thumbnails/30.jpg)
C. SchrijversC. Schrijvers
Non Radial Pulsations Modes
![Page 31: Identification of Important VO Spectral Services benefiting from deployment on the GRID](https://reader034.vdocument.in/reader034/viewer/2022050820/56814ffc550346895dbdc438/html5/thumbnails/31.jpg)
= 8, m=3
Tim
Bedding
= 2, m=1
Non Radial Pulsation
![Page 32: Identification of Important VO Spectral Services benefiting from deployment on the GRID](https://reader034.vdocument.in/reader034/viewer/2022050820/56814ffc550346895dbdc438/html5/thumbnails/32.jpg)
Measured Pulsations
Rho Pup – del Sct type Eps Cep - del Sct type
![Page 33: Identification of Important VO Spectral Services benefiting from deployment on the GRID](https://reader034.vdocument.in/reader034/viewer/2022050820/56814ffc550346895dbdc438/html5/thumbnails/33.jpg)
Doppler ImagingFrom LPV due to rotation stellar Spots - darker, brighter – chemical patch
![Page 34: Identification of Important VO Spectral Services benefiting from deployment on the GRID](https://reader034.vdocument.in/reader034/viewer/2022050820/56814ffc550346895dbdc438/html5/thumbnails/34.jpg)
Doppler Imaging - NRP
Vogt & Penrod -80sZet Oph
Requires high SNR (>300-500) Perfect rotation coverage Artefacts otherwise NRP or solar spot
Doppler Tomography Accretion jets in Algols Orbits in RV phase space
![Page 35: Identification of Important VO Spectral Services benefiting from deployment on the GRID](https://reader034.vdocument.in/reader034/viewer/2022050820/56814ffc550346895dbdc438/html5/thumbnails/35.jpg)
Doppler Imaging
Different elements
temperature distribution
![Page 36: Identification of Important VO Spectral Services benefiting from deployment on the GRID](https://reader034.vdocument.in/reader034/viewer/2022050820/56814ffc550346895dbdc438/html5/thumbnails/36.jpg)
Periodic spatial filter (PSF) concept for NRP mode detection (2-D concept)
(Mkrtichian 1994, Solar Physics, 152, p.275)
“roAp star as a Sun” observations:
Si abundance spots
Si pseudo-mask
Cr pseudo-mask
![Page 37: Identification of Important VO Spectral Services benefiting from deployment on the GRID](https://reader034.vdocument.in/reader034/viewer/2022050820/56814ffc550346895dbdc438/html5/thumbnails/37.jpg)
Mg Field - Polarimetry
![Page 38: Identification of Important VO Spectral Services benefiting from deployment on the GRID](https://reader034.vdocument.in/reader034/viewer/2022050820/56814ffc550346895dbdc438/html5/thumbnails/38.jpg)
Simulation of Stokes
![Page 39: Identification of Important VO Spectral Services benefiting from deployment on the GRID](https://reader034.vdocument.in/reader034/viewer/2022050820/56814ffc550346895dbdc438/html5/thumbnails/39.jpg)
Simulated Magnetic stars - spectra
Kochukhov & Piskunov 2002
Abundances + Mg field Stokes parameters spectra
![Page 40: Identification of Important VO Spectral Services benefiting from deployment on the GRID](https://reader034.vdocument.in/reader034/viewer/2022050820/56814ffc550346895dbdc438/html5/thumbnails/40.jpg)
Zeeman Doppler Imaging
II Peg, Strassmeier 2007
![Page 41: Identification of Important VO Spectral Services benefiting from deployment on the GRID](https://reader034.vdocument.in/reader034/viewer/2022050820/56814ffc550346895dbdc438/html5/thumbnails/41.jpg)
Outline of the Talk
![Page 42: Identification of Important VO Spectral Services benefiting from deployment on the GRID](https://reader034.vdocument.in/reader034/viewer/2022050820/56814ffc550346895dbdc438/html5/thumbnails/42.jpg)
Outline of the Talk
![Page 43: Identification of Important VO Spectral Services benefiting from deployment on the GRID](https://reader034.vdocument.in/reader034/viewer/2022050820/56814ffc550346895dbdc438/html5/thumbnails/43.jpg)
Lorentz Force Signatures
Balmer lines variability
Shulyak 2007
![Page 44: Identification of Important VO Spectral Services benefiting from deployment on the GRID](https://reader034.vdocument.in/reader034/viewer/2022050820/56814ffc550346895dbdc438/html5/thumbnails/44.jpg)
Benefits of implementing as VO services
Unified data format (VO-Table, semantics of Unified data format (VO-Table, semantics of variables)variables)
Transparent data conversion, homogenization, Transparent data conversion, homogenization, rescalingrescaling
Powerful presentation with remote data (URI) + Powerful presentation with remote data (URI) + TVO resultsTVO results
Large spectral survey feasibleLarge spectral survey feasible Serendipitious research - click on star in the Serendipitious research - click on star in the
image of cluster to see its dynamic spectra image of cluster to see its dynamic spectra (many observation) (many observation)
![Page 45: Identification of Important VO Spectral Services benefiting from deployment on the GRID](https://reader034.vdocument.in/reader034/viewer/2022050820/56814ffc550346895dbdc438/html5/thumbnails/45.jpg)
Killer spectral applications
Use VO to find all stars with emission in given line (EW<0) – Use VO to find all stars with emission in given line (EW<0) – find the time when it was in em.find the time when it was in em.
Use VO to get 1000 spectra of the given object cut out regions Use VO to get 1000 spectra of the given object cut out regions arround given lines, plot the lines, make a gray dynamic arround given lines, plot the lines, make a gray dynamic spectrum folded in timespectrum folded in time
The same – search period, fold by periodThe same – search period, fold by period
Get the unknown line ID of piece of spectra from SLAP Get the unknown line ID of piece of spectra from SLAP overploted over SSA dataoverploted over SSA data
Create Light and RV curve for given period Create Light and RV curve for given period
Fit the grid of models (Teff, log g) to the observed spectrum – Fit the grid of models (Teff, log g) to the observed spectrum – for many starsfor many stars
![Page 46: Identification of Important VO Spectral Services benefiting from deployment on the GRID](https://reader034.vdocument.in/reader034/viewer/2022050820/56814ffc550346895dbdc438/html5/thumbnails/46.jpg)
Conclusions
VO clients – basic functionsVO clients – basic functions
Advanced work not Advanced work not supported in VO at all! supported in VO at all!
Server side servicesServer side services
switch to Workflowsswitch to Workflows
GRID GRID
Need of modelsNeed of models
TVO TVO
What to do ? - VO literacyWhat to do ? - VO literacy
More stellar spectra to VO More stellar spectra to VO archivesarchives
VO portal using local private VO portal using local private spectral serverspectral server
Write analysis tools with VO Write analysis tools with VO interfaceinterface
VO services for common VO services for common taskstasks
Ask astronomers !Ask astronomers !