QC AT ESO

The main goals of ESO data quality control (QC) are:

• to know the instrument status (health checks)
• to monitor it over time (trending)
• to deliver data of known and defined quality (certification).

The QC process uses data processing pipelines. These support all VLT/VLTI instruments and all major modes. More ...

The ESO QC process is a distributed process. Currently data delivery to ESO headquarters in Garching takes 10 days on average. Therefore the most fundamental QC checks (so-called QC0) must be done on-site and in near-real time, by the daytime and night-time astronomers. They are supported by the Health Check monitor.

More complete and in-depth checks (QC1, trending) are done at ESO headquarters in Garching by the Quality Control Group. The complete trending history of each VLT instrument is also maintained here.

The QC0 process checks that

• raw data have a reasonable exposure level
• comply formally with the ESO standards
• calibration data are complete
• user-defined observational constraints have been respected during the observation

Typically the QC0 checks are done on raw data.

The Service Mode observers have defined constraints for their OBs. They receive a QC0 report with a comparison of their constraints and the actual conditions during observation. The data packages also include night log information, including an OB report containing the assessment by the on-site staff of the quality of the raw data.

The data are checked for the following constraints:

• airmass,
• moon distance,
• fractional lunar illumination (FLI),
• seeing.

Quality Control Level 1 consists of quality checks on pipeline-processed data. Quality is measured by numbers, the QC1 parameters. Main goals of the QC1 process are

• to monitor instrument stability and performance
• to monitor the atmosphere
• to monitor pipeline performance.

The checks are done both on calibration and on science data. In practice most quality information is extracted from the calibration data, while the science data can add QC information which can only be measured on sky (like image quality or zeropoints). Scientific usefulness cannot be assessed with this process.

The QC1 parameters are used

• to assess the quality of master calibration files (e.g. noise properties, gradients),
• to compare with expected values (e.g. gain value),
• to assess the performance of the instrument (e.g. throughput from photometric zeropoints),
• to monitor efficiency of master creation process (e.g. reduction of random noise).

The result of the QC1 process is used to

• decide whether master creation data are certified or rejected,
• assess the quality of the reduction process.

The QC1 parameters are archived. They are directly available for studies through the QC1 database interface (although in many cases the trending pages may be more interesting, see below).

Trending is monitoring QC1 parameters over time (e.g. effective resolution as function of time). A core task of QC Garching is to create trending pages which are used to monitor the instrument evolution and performance. Trending plots come in two flavours:

• as Health Check plot (HC), focusing on the current, most recent data set and comparing the latest Paranal-delivered QC1 parameters with the current set of QC1 parameters (typically from the last 60 or 90 days); the HC plots concentrate on the most crucial instrument properties and components;
• as classical trending plot, which is more complete both in the number and content of plots; trending plots also cover the full history of QC1 parameters.

The GIRAFFE HC plots are available on the HC home page.

The set of GIRAFFE trending plots is accessible here.

Both types of trending plots are static. A tool for creating dynamic trending plots and to perform research on instrument evolution is the QC1 plotter (GIRAFFE | All instruments).