appropriate control vocaburary for 'analysis' and 'reanalysis'

[yhonda21]

At level 10, 'analysis' and 'reanalysis' are proposed as control vocabularies. The definition of these terms are not clear enough and 'reanalysis' can be confused with 'climate reanalysis'.

In the case of global NWP, two type of analysis could be available according to the different cut-off time of observation ingest. One is the analysis with short cut-off time, which is processed to produce an initial field of a prognostic model. The other one is the analysis with long cut-off time, which is the best estimated current status of atmosphere. Normaly the latter analysis is considered as 'analysis' product. I wonder if the former analysis could be provided to users?

If the above explanation makes sense, I wonder if we can rename the current 'reanalysis' to 'analysis'? Then we need to find another appropriate name to replace the current 'analysis', e.g 'ealry-analysis'.

[yhonda21]

Please share how you call these two different analyses at your centres. That will be helpful to find the appropriate words.

[Yuanyuanliu2023]

For "early analysis", we refer to it as “short cut-off time analysis” .

[sebvi]

At ECMWF, an "analysis" is what we use in the context of operations or research regardless of the cut-off time usually to be used on its own as an single instance. The distinction between short and delayed cut off is made for the Data assimilation system producing the analysis, not the analysis itself. We use "reanalysis" in the context of creating a dataset from a long time in the past up to near real time. It is then use for verification, model version comparison, etc. In a sense, a "reanalysis" is a collection of "analyses".

[yhe-wmo]

TT-NWPMD Meeting 11.05.2023:

• @alexandreleroux will provide ECCC's perspective.
• more feedback needed from other TT-NWPMD members.

[alexandreleroux]

Here are the initial tentative definitions of analysis and reanalysis as used at the Meteorological Service of Canada :

• Analysis : An Analysis is the result of the use of data assimilation techniques that combine observational data with an NWP model to estimate current weather conditions.
• Reanalysis : A reanalysis is a retrospective reconstruction of past weather conditions using data assimilation techniques that combine historical observations with NWP models.

Not impossible these get adjusted, but might be a suitable starting point? Let's discuss at the TT-NWPMD meeting :-) Merci !

[yhe-wmo]

TT-NWPMD Meeting 2023.06.13 The team noted that "analysis"/"reanalysis" are used by different communities (e.g. weather, climate, ocean) with slightly different definitions and for various different systems. The meeting suggested keeping the four controlled vocabulary for this level (forecast, analysis, reanalysis, hindcast) as they are, and asking all members to try mapping their data to the topic hierarchy. Whenever concrete issues emerge during the pilot phase, the controlled vocabulary or their descriptions can be amended.