Country combinations: which countries are available, in which datasets, and where are combinations necessary?

[jkikstra]

For emissions harmonization, we mostly rely on CEDS data. They publish country-level timeseries.
GFED data will likely need to be aggregated from the grid ourselves.

For harmonization, a few more country-sets come in:

• SSP GDP data
• SSP Population data
• 7 IAM region-to-country mappings

The latest SSP data are currently version 3.1, with a few minor issues still to be resolved before these 'new SSP data' become final.

CEDS processing:
For previous SSP data, there were three sets of countries that were split out in CEDS, but only as aggregates in the SSPs:
country_combinations = {"isr_pse": ["isr", "pse"], "sdn_ssd": ["ssd", "sdn"], "srb_ksv": ["srb", "srb (kosovo)"]}

• With updated SSP data, we can most likely remove "isr_pse" and "sdn_ssd".

GFED processing:

• check the country-mask that is being used.

IAM region mapping

• check what countries are covered (start with the code available here: Create aggregate regions for each IAM #25 (comment) ), and which are not, and reach out to all IAM teams where necessary to fill the gaps.

SSP GDP and Population availability:

• add this to the notebook comparing missing data; should indicate that e.g. kosovo is currently missing

[jkikstra]

@eahoegner here's what we discussed yesterday, see also my comment here: #26 (comment)

[eahoegner]

great, i was already wondering about the aggregate country codes

[eahoegner]

here is the comparison of CEDS vs. iso_mask.nc (GFED4) based country mappings. the combined/not combined country codes aside, there is around 15 areas that CEDS apparently integrates into other countries. while i assume that f.ex. jersey and guernsey are probably counted as part of the UK, in cases such as andorra, monaco, or tuvalu, i am not sure where these have gone. the IAMC common definitions do include all of these separately when aggregating into regions, so this would provide an argument for applying the iso_mask also to gridded CEDS rather than using CEDS country definitions.

country source
0 srb_ksv ceds
1 isr_pse ceds
2 global ceds
3 sdn_ssd ceds
4 and gfed4
5 imn gfed4
6 mnp gfed4
7 sjm gfed4
8 aia gfed4
9 tuv gfed4
10 sdn gfed4
11 mco gfed4
12 ssd gfed4
13 nru gfed4
14 myt gfed4
15 srb gfed4
16 bes gfed4
17 pse gfed4
18 smr gfed4
19 jey gfed4
20 nfk gfed4
21 isr gfed4
22 ggy gfed4
23 shn gfed4
24 srb (kosovo) gfed4
25 pcn gfed4

[eahoegner]

using the IAMC common_definitions common.yaml file region definitions, this is missing even more countries.
compared to CEDS, it's missing:
{'asm',
'bmu',
'cuw',
'cym',
'dma',
'flk',
'fro',
'gib',
'grl',
'lie',
'msr',
'spm',
'sxm',
'tca',
'tkl',
'vgb',
'wlf'}

[eahoegner]

and compared to gfed additionally
'aia',
'and',
'bes',
'ggy',
'imn',
'jey',
'mco',
'mnp',
'nfk',
'pcn',
'shn',
'sjm',
'smr'