Post-observation processing
On this page:
- Pipeline processing of raw data from ESO
- Radial velocity
- Object classification
- Spectrum analysis
- Survey archive
- Quality control, Calibration, Verification
- Deliverable delivery
Target selection and OB development is described in Target identification, FPOSS and OB preparation. We consider here post-observation work.
Pipeline processing of raw data from ESO. Raw data reduction will be performed at CASU/Cambridge (GIRAFFE spectra) and Arcetri (UVES spectra), with all raw data and extracted spectra then being stored centrally at Cambridge. We deliberately process data by spectrograph type rather than by science target at this stage, to ensure field and clusters are treated similarly. The instrument-specific reduction pipelines are to be the current ESO systems. Private special-purpose pipelines capable of processing GIRAFFE data are also available. We will utilise these to investigate possible improvements to extant pipelines.
The Gaia-ESO project will utilise both an operational database and an internal archive, to hold all relevant information. Raw data are pipeline processed through the operational database, for delivery to the spectrum analysis teams. The operational database hosts the survey progress monitoring information, which is developed based on, and learning from, VISTA data reduction and monitoring experience. This keeps track of the status of all targets selected, through their several FPOSS allocations and GIRAFFE settings, the delivered signal to noise ratio (S/N) and quality flags in each, and the status of data processing and added value product determination. This system, building on the operational VISTA (UKIDSS, etc) systems is hosted at CASU, Cambridge.
Radial velocity. After pipeline processing to remove instrumental signature, we have extracted individual spectra. These are analysed (in one method, prior to extraction) to deliver radial velocities (& vsini where relevant), and associated error functions, using two pipelines for GIRAFFE, one for UVES. This process generates a quality control flag, and preliminary object classification parameters.
Object classification. Each spectrum, together with all its associated photometry from the target selection process, is then classified further, through dedicated systems (cross-correlation with templates, neural net, MPIA Gaia system) providing first-pass parameters for the spectrum analysis teams.
Spectrum analysis. All extracted spectra are processed through general purpose pipelines, to refine astrophysical parameters, and deliver elemental abundances to a level appropriate for the relevant stellar type and available S/N. These pipelines manage, respectively, hot, warm, cool stars, as well as pre-main sequence stars, and GIRAFFE and UVES spectra. It is a strength of this Gaia-ESO Survey team that it includes a majority of Europe's spectrum analysis groups, which between them have available expertise in many complementary and special-purpose methodologies. All have agreed to adopt a fixed set of atomic data and model atmospheres for the analysis of FGK stars, and to optimise their local expertise appropriately. Very considerable coordination between the teams has been underway for some months. They are already operating as a coherent community. This range of analysis excellence will be applied to the various stellar and data types as appropriate. Sanity checking will then deliver, for each star target, a “best” set of parameters and abundances, with corresponding random and systematic errors, and an explicit analysis of the effect of alternative analysis assumptions. All these results will be archived for later analysis, both in the operational database, and the Survey archive.
Survey archive. After abundance determination, the data are available for quality control, science verification, and preliminary analysis, by the survey team. This access and set of processes will be managed through an internal archive. This archive, building on expertise at AIP, Edinburgh and Madrid, will be Virtual Observatory (VO) compliant, and will serve the survey team with both survey and other available data.
Quality control, Calibration, Verification. The reduced data will be sanity checked, the calibration targets analysed, errors assessed and verified. The output first-pass deliverables will now be available for internal Gaia-ESO science verification, and quick-look analysis. Following this process, checked deliverable data products will be returned to the archive and the operational database, and prepared for delivery to ESO.
Deliverable delivery. The internal archive centre ensures all deliverables are appropriately formatted for ESO use, and documented. Following the agreed schedule, and Co-PI sign-off, value-added data are released to ESO, and to the Gaia-ESO Survey Co-Is for analysis.
