openENTRANCE/openentrance

Additional regional aggregations (REMIND)

Renato-Rodrigues opened this issue · 7 comments

Request to add support for REMIND regions aggregation for use in the ECEMF project.

Additional regional aggregations:

World:
   definition: 'all World countries'

EU27+UK:
   definition: 'EU27 plus United Kingdom' 

UKI:
   definition: 'United Kingdom and Ireland'

NEU:
   definition: 'European countries outside EU (minus UK)'
   
NEN:
   definition: 'Non European Union North countries'
   
NES:
   definition: 'Non European Union South countries'

ECE:	
   definition: 'European Union Center East countries'

ENC:
   definition: 'European Union North Center countries'

ECS:	
   definition: 'European Union Center South'

ESC:	
   definition: 'European Union South Center countries'

ESW:	
   definition: 'European Union South West countries'

EWN:	
   definition: 'European Union West North countries'

You can find the detailed mapping of the above mentioned regions in the below file:
https://github.com/remindmodel/remind/blob/develop/config/21_regions_EU11/regionmapping_21_EU11.csv
region names (column 2), iso-3 codes (column 2):

I agree some region definitions woudl be great to be added esp. also for ECEMF.

Not sure if all models' native modelling regions are required to be added? In case this could be a large super set for all models.

I would also add
-R52 macro regions (region definition see here: https://tntcat.iiasa.ac.at/SspDb/dsd?Action=htmlpage&page=10#regiondefs)

EU27+UK is already implemented as EU28 I think? (and the main analysis dimension to be used in WP1 as I understand?)

Already existing aggregated regions are listed in aggregate-regions.yaml
Additionnal can of course be added, but please be carefull so that the definition is not ambiguous. We have defined new aggregated regions by listing the already existing defined regions, eg
EU27:
definition: 'AT + BE + BG + HR + CY + CZ + DK + EE + FI + FR + DE + GR +
HU + IE + IT + LV + LT + LU + MT + PL + PT + RO + SK + SI + ES + SE + NL'
notes: 'membership as of February 1, 2020'
Here the list of countries (AT,....) are defined in countries.yaml
Moreover the definition tag is being used in automated python scripts (which with the above example is very easy to do)
As an example from the list above, UKI should be
UKI:
definition: 'UK+IE'
EU28 could be defined as
EU28:
definition:'EU27+UK'
Best
Sandrine

EU27+UK is already implemented as EU28 I think? (and the main analysis dimension to be used in WP1 as I understand?)

I don't see the EU28 region in the aggregate-regions.yaml file. Could you direct me to which region should I use for that?

Not sure if all models' native modelling regions are required to be added? In case this could be a large super set for all models.

It would be great if someone with experience with the interaction between the IIASA explorer and the nomenclature could comment on this - whether it is necessary for all relevant model-internal regions to be added?
So if we would want the IIASA database to automatically calculate e.g. EU27 results from the native regions in REMIND, which partially are on national level, but mostly consists of several nations grouped together (eg ESW, as in Europe South West, = Portugal + Spain), would we first need to add ESW to the openEntrance nomenclature, or does the explorer handle this anyway?

Best,
Robert

Re the comment by @JohannesEmm: The R5-macro regions were added, see #122.

Re the question by @Renato-Rodrigues about "EU28" versus "EU27+UK" - this is not (yet) defined in the nomenclature.

Re the question by @robertpietzcker "we would want the IIASA database to automatically calculate ...", we are currently implementing the workflows to do this as part of the (new) ism-nomenclature package (see https://github.com/iamconsortium/nomenclature).

The "complete workflow" will be as follows:

  1. A modelling team uploads the scenario data at model-native-region resolution
  2. The scenario data is renamed from e.g., "UKI" to "REMIND v2.2|UKI" (to safeguard against collisions between different model-native-region names and make regions more easily findable in the Scenario Explorer - this will be then be shown as an expandable list, see the example below)

image

3. The scenario data are aggregate to compare-regions commonly used by all/many modelling teams (e.g., EU27+UK).

@Renato-Rodrigues, if I understand you correctly, you request these additional regions as "native-model regions", plus "EU27 & UK" added as a "compare region"?

"EU27 & UK" (or EU28) is not model specific, so you are perfectly right and my request was to add it directly to the "aggregate-regions.yaml" nomenclature definition.

The other regions I mentioned above are REMIND specific region aggregations. Although, they can be added only as remind reporting regions, I would also see as an important asset to be able to compare them directly against model results that provide country level data to the database.

By the way, thank you for all the support provided!

Closing as implemented