-
Notifications
You must be signed in to change notification settings - Fork 22
Alignment of OSCAR Requirements (O R) and OSCAR Space (O Space) with WMD vocabulary for variables
WMDS defines a metadata model for the description of observations that is designed to serve the description of all WIGOS observations (=capabilities), both surface-based and space-based, in support of the WIGOS RRR process. The standard makes copious use of controlled vocabularies. O/R lists a set of requirements for information on biogeophysicochemical properties of the Earth system. A gap analysis between requirements and capabilities is only possible if a clear link can be made between the two. To that end, TT-WIGOSMD is charged with evolving the vocabularies necessary to describe observations comprehensively. OSCAR/Surface already uses the official vocabulary, while O/R and O/Space don't fully comply with existing code lists yet.
The objectives are to
- identify and eliminate mismatch and gaps in the names and definitions of observed variables between O/R (ref0) and the official and approved WMDR vocabularies (ref1-5),
- reference the WMO notation for variables that are defined in WMDR in O/R, OSCAR/Space to facilitate gap analysis and further evolution of specific requirements,
- engage the O/R ans OSCAR/Space stakeholders in an orderly process of evolving vocabulary in case there is a need.
We concede that there may be variables/quantities for which requirements have been defined that
-
cannot be observed or retrieved from observations and for which the WMDR vocabulary may not apply, or
-
represent broader concepts that are not covered with any specific observation alone but rather with a 'family' of observations.
Previous attempts dating back to ca. 2017 to consolidate the differences in vocabularies have led to amendments of the WMDR code lists, but the task was not fully concluded. As a result, these prior analyses are of limited value and the current vocabularies need to be re-assessed and compared again. Issue #190 provides information on more recent work.
The comparison of vocabularies requires similar concepts for the definition of both. The OSCAR/Requirements vocabulary consists of a table, whereby each row is assigned an id and a number of properties. The WMDR code lists, specifically the vocabulary for variables have fewer elements.
| WMDR 1.0 | OSCAR/Requirement | Comment |
|---|---|---|
| Notation | ID | |
| Name | Name | Name currently may contain a notion of geometry (point, line, area, total column, volume) of the observation. This will be strictly eliminated in WMDR2.0 and will be represented as a separate property. |
| (partly implicit in name/path; will be a separate property in WMDR2.0) | Domain | The notion of 'domain' exists in WMDR, but refers strictly to physical domains. A new code list will be submitted to FT2021-2. |
| (partly implicit in name/path) | Cross-cutting Theme | |
| Definition/Description | Definition | |
| n/a | ||
| Measuring Units | ||
| n/a | Uncertainty Units | |
| n/a | Stability Units | |
| n/a | Hor. Res. Units | |
| n/a | Ver. Res. Units | |
| n/a | Comment | |
| n/a | Used in Applications | |
| (partly implicit in name/path; will be strictly a separate property in WMDR2.0) | Layers | |
| Matrix | The concept of 'matrix' can be used to specify 'sub-domains' as well. A new code list will be submitted to FT2021-2. | |
| Tags | Tags are not part of the specification of a variable, but serves to group variables and to facilitate search. |
The proposed approach consists of mapping the existing vocabularies, of identifying exact and close matches, of identifying missing elements in the WMDR vocabulary and preparing amendment proposals for the WMDR. It is suggested that the additional properties found in the flat table of the O/R variables vocabulary are not considered as constituting defining properties of a variable. Specifically, units should not be considered to be part of the definitions, as the same quantity can be expressed in different units.
-
Verify if ref6 reflects the current status of O/R and WMDR vocabulary (after FT2021-1 amendments)
-
Create new separate issues to address subsets labeled
- 'Exact match': Agree if definitions in O/R can be changed by dropping reference to units to match definitions in WMDR (ca. 187 entries)
- 'Not found in WMDR': Discuss and implement what needs to be done (ca. 69 entries)
- 'Discussion needed': Discuss and implement what needs to be done (ca. 62 entries)
-
Agree if subset 'Not found in O/R' (ca. 402 entries) needs further action - A priori, it is okay if variables are defined for which no requirements have been specified (yet).
ref0: https://space.oscar.wmo.int/variables
ref1: https://github.com/wmo-im/wmds/blob/master/tables_en/1-01-01.csv
ref2: https://github.com/wmo-im/wmds/blob/master/tables_en/1-01-02.csv
ref3: https://github.com/wmo-im/wmds/blob/master/tables_en/1-01-03.csv
ref4: https://github.com/wmo-im/wmds/blob/master/tables_en/1-01-04.csv
ref5: https://github.com/wmo-im/wmds/blob/master/tables_en/1-01-05.csv
ref6: https://github.com/wmo-im/wmds/files/7059753/variable_mapping_WMDR_OR.xlsx
ref7: https://github.com/wmo-im/wmds/issues/190