Mapping between Data sets in archives and Deployment in OSCAR is undetermined

[tomkralidis]

(moved from wmo-cop/wmo-oscar#23)

Mapping between data sets in archives and Deployment in OSCAR is undetermined.

[joergklausen]

This is very true. The concept of deployment in WMDR is one of a homogeneous time series generated with one instrument, at least with only one method method. Expert judgement is required to decide if an observation should be documented as one or several deployments. The WMDS does not require that a homogeneous time series needs to be described with only one deployment, so annual datasets can in principle be described with multiple deployments.

I see various approaches depending on the situation:

1. OSCAR/Surface has nothing yet.
   I would suggest to create as many deployments as there are individual data sets /files /URLs in the archive.
2. OSCAR/Surface already has an observation with 1 deployment
   If the archive has only 1 data set, I suggest to update the existing one in OSCAR/Surface.
   If the archive has several data sets, I suggest to follow option 1 and (manually) migrate any existing additional information in OSCAR/Surface as far as possible, and then delete the existing deployment from OSCAR/Surface. This will require involvement of the station PI.
3. OSCAR/Surface already has an observation with several deployments
   This will require involvement of the station PI and a treatment on a case-by-case basis, I think.

@tomkralidis @NunesL How do you see this? What is the right forum to discuss and decide?

[tomkralidis]

Probably a good discussion for ET-ACDM to get an idea of how data centres model deployments and granularity.