Update GRIB2ParameterLocalTable.csv

[feggleton]

Have added the 2 divisions for global model and 6 divisions for defence models. Division detail now in name and descriptions given. No mention of bins as these relate to the model. Please check carefully as to the sizes matching in name and description. Please let me know if description reads correctly and if any other information is required. I would like to link these to the stash codes, however, there is only 6 stash codes, so not clear if these only link to the 6 defence divisions??? https://reference.metoffice.gov.uk/um/_stash see Dust division 1 mass mixing ratio (1-6). In addition, need to check any current usage of 0-13-193 and 0-13-194 will be effected by the change in name to include the sizes

[feggleton]

m01s00i431 m01s00i432 m01s00i433 m01s00i434 m01s00i435 m01s00i436 - 6 defence stash codes

[feggleton]

m01s00i431 global division 1
m01s00i432 global division 2

[mgange659]

The latest WMO documentation has a Note above the category 13 aerosols Code Table: << This category is no longer populated. Please use "Product Discipline 0 - Meteorological products, parameter category 20: atmospheric chemical constituents". >> So perhaps we should be using (0,20,192), (0,20,193),(0,20,194), etc. for the new dust parameter local codes.

[bjwheltor]

@feggleton - The values and descriptions look correct to me (reviewed against https://www-nwp/~ppdev/stage/stage_project/parameters/parm_desc/dust.html), so I'm happy to approve on this basis.

However, I note that @mgange659 has raised the point that:

The latest WMO documentation has a Note above the category 13 aerosols Code Table: << This category is no longer populated. Please use "Product Discipline 0 - Meteorological products, parameter category 20: atmospheric chemical constituents". >> So perhaps we should be using (0,20,192), (0,20,193),(0,20,194), etc. for the new dust parameter local codes.

[laragunn]

@mgange659 and @bjwheltor although this may be something that is right, we need to have some way to tell WDH and make customers aware. we already output dust as a local code in one of the deterministic models. If we change codes around they have to tell customers and I get it in the neck because a change has happened and they don't understand. We need to have a way to communicate this correctly rather than just make changes and hope for the best. This is a breaking change in the existing model output so we might have to go with 13 for now then do some kind of release later than makes the change.

[feggleton]

@mgange659 I can't see this note on https://github.com/wmo-im/GRIB2/blob/master/GRIB2_CodeFlag_4_2_0_13_CodeTable_en.csv - could you link to where you saw this note?

[laragunn]

We could update the codes for the 5 dust sizes that are in the CAMs/DRMs and then in slower time update the deterministic model but that seems like a silly approach

[mgange659]

WMO Website link is https://library.wmo.int/records/item/35625-manual-on-codes-volume-i-2-international-codes

[bjwheltor]

@mgange659 and @bjwheltor although this may be something that is right, we need to have some way to tell WDH and make customers aware. we already output dust as a local code in one of the deterministic models. If we change codes around they have to tell customers and I get it in the neck because a change has happened and they don't understand. We need to have a way to communicate this correctly rather than just make changes and hope for the best. This is a breaking change in the existing model output so we might have to go with 13 for now then do some kind of release later than makes the change.

@laragunn - Is this actually a breaking change? Presumably, the updates to the local codes for the Global models can be applied to the existing local codes (this is not clear in this PR as the codes are not explicitly stated), so this would not be a breaking change, and I was assuming the the DRM local codes are new, so there is nothigng 'to break' yet. Have i misunderstood the situation?

[feggleton]

@mgange659 and @bjwheltor although this may be something that is right, we need to have some way to tell WDH and make customers aware. we already output dust as a local code in one of the deterministic models. If we change codes around they have to tell customers and I get it in the neck because a change has happened and they don't understand. We need to have a way to communicate this correctly rather than just make changes and hope for the best. This is a breaking change in the existing model output so we might have to go with 13 for now then do some kind of release later than makes the change.

@laragunn - Is this actually a breaking change? Presumably, the updates to the local codes for the Global models can be applied to the existing local codes (this is not clear in this PR as the codes are not explicitly stated), so this would not be a breaking change, and I was assuming the the DRM local codes are new, so there is nothigng 'to break' yet. Have i misunderstood the situation?

Yes the two existing local codes (global - 193, 194) have just been renamed and given descriptions and the other 6 are new for the DRM

[feggleton]

Currently EM-DRM and SEA-DRM actually uses 20 but global uses 13 because the existing local codes are 13. Let's discuss tomorrow.

[laragunn]

@bjwheltor it's a breaking change for the two existing local codes that are in the Global. We would have to change it in MDDA - not a massive thing but that doesn't break the service at WDH it just means that the glossary doesn't match. So WDH will claim that's a breaking change so I'll have to do a release i.e. put something into a CI branch for a few weeks then move it over.

[mgange659]

More Notes from the WMO Binary Codes Manual:
PDS Octet 10: GRIB master table version number (see Common Code table C–0 and Note 1)
PDS Octet 11: Version number of GRIB Local tables used to augment Master tables (see Code table 1.1
and Note 2)

(1) Local tables shall define those parts of the Master table which are reserved for local use except for the case described
below. In any case, the use of Local tables in messages intended for non-local or international exchange is strongly
discouraged.
(2) If octet 10 contains 255 then only Local tables are in use, the Local table version number (octet 11) must not be zero nor missing, and Local tables may include entries from the entire range of the tables.
(3) If octet 11 is zero, octet 10 must contain a valid Master table version number and only those parts of the tables not reserved for local use may be used.

[feggleton]

I have made the changes as discussed yesterday so please re-review @laragunn @bjwheltor @mgange659

[laragunn]

Looks good to me thanks