Observing descriptions & FITS header

There are many issues that concern FITS keywords in relation to observations and the resulting data. This is particularly relevant for things like automatic scripts as part of calibration plans, or data reduction and analysis through pipelines, but also for extended Virtual Observatory compatibility of our data and we had a specific working group meeting consisting of the astronomy staff and software group about this.

It was agreed that in addition to IMAGETYP and OBS_MODE to describe the type of observations a 3rd one will be created called IMAGECAT. To fill the 3 keywords with the correct settings we are looking at using the set defined by ESO, but some addition will be needed and need to be agreed. A next step will be to define the (combination of) values for the 3 `observing type' keywords for each and all of the different types of observations. The latter will also provide a basis to define how these keywords will be set. The general system will be such that at the end of an observation the 3 keywords will be set to default values. In all cases where the observing set-up fully defines any of the keywords, this will be set automatically during the observation, while for those that can not be defined the latest defined values will be used (which will be the default values if this is not specifically set). Beyond specifically setting the keywords, the idea would be to provide standard sequencer commands/scripts that as part of an observation set the keywords to the corresponding values (e.g., the command `expose' will always set the keyword IMAGETYP to `OBJECT', while the command `stdexp' will set it to `STD'; or more elaborate/sophisticated variations on such a scheme).

We also agreed to make a full review of the set of keywords that were are using for data from each instrument. This is basically a check of the relevant document that was made in 2004 by Saskia Prins, where the name and format should be defined of any keyword that is missing. Beyond the observing system related keywords it should be defined what `phase 1' and `phase 2' proposal information is required in the header, and if any additional keywords would need to be created for that purpose.

A point to consider for the observing block (OB) generator we are planning (see above) is related to the FITS keywords and their settings. As an OB describes a complete observation, it also should be able to define all the relevant keywords. I.e., the information to decide for each (set of) observation(s) in the OB to which values the keywords should be set (in particular the `observing type' keywords mentioned) should be included for each observation.

A specific change to the existing FITS keywords has been an update to keywords that contains information about the timing of exposures to an precision 0.1 second for ALFOSC, StanCam, FIES and NOTCam to reflect the accuracy of our timing.

Thomas Augusteijn 2010-05-27