Project specific interfaces to REMIND / MAgPIE.

Reporting data to the AR6 DB is a multi-step process that can take a week if you do not have your model registered with the IIASA DB yet.

Information on the process can be found here:

https://data.ene.iiasa.ac.at/ar6-scenario-submission/#/about

Please check if the model version your runs are based upon is already registered with the AR6 DB. At the time of this writing I (Alois, 4.6.20) do not know of any automated way to do this. Maybe ask the DB admins.

For the REMIND model with global coverage, two registration steps are required: There is the rapid registration,

https://forms.office.com/Pages/ResponsePage.aspx?id=C7F5mwcAik6wcq0MjNwapZ2PcWg1iqFLp9EH2Rjz8hZUOTJTNzBaNjdWRUpPWERYWUJGNjEyODNIVi4u

and the more detailed model registration template which is part of the Download found under section 2 on the IIASA help page.

Usually you will be reminded after the first step to fill in the more detailed template and send it via mail.

This question is difficult to answer. The best way to judge is based upon the categories in the model registration template. If the setup of your runs differ substantially from the REMIND/MAgPIE defaults, e.g., you use different demand sector modules or there are major changes in the coupled models, it might be better to register a new model to avoid confusion.

Before submitting your data to the AR6 database please make sure that

• the policy costs in the reporting are calculated w.r.t. the appropriate reference run. For most runs, this will involve re-running the policy costs.

• idx: Index as given by the AR6 submission template, this is just for reference
• Variable_AR6: The AR6 reporting variable name as given by the submission template
• Unit: The AR6 reporting variable unit as given by the submission template
• r30m44: The REMIND/MagPIE variable name
• r30m44_unit: The REMIND/MagPIE
• r30m44_factor: Conversion factor multiplied with the REMIND/MagPIE value to obtain the AR6 compatible value
• r30m44_spatial: Restricts reporting to global level ("glo") or regional level ("reg"), empty values mean no restriction
• m44: MagPIE specific variable name
• m44_unit: MagPIE specific variable unit
• m44_factor: Conversion factor for MagPIE value
• internal_comment: Comment for internal reference
• Comment: Comment to be published with the AR6 submission template
• Definition: Variable definition, initial values copied from the AR6 submission template.

The corresponding columns for former REMIND v2.1, MAgPIE v4.2 are still available with the prefix r21m42.

The workflow to generate AR6-compatible output is as follows:

• the variable mappings are stored together with metadata in a template. You can, e.g., copy the ar6/mapping_template_AR6.csv and adjust it to your needs.
• from the template you generate a mappingfile.
• use the mapping and REMIND and/or MAgPIE output with iamc::write.reportProject to generate a MIF file with AR6 compatible variable names and factors.

A template is just a file which contains the mapping, optionally a factor and a weight column and some metadata. For example, for AR6 one should report model specifics regarding the accounting for variables in a tab called "comments". This column is included in the template for convenience.

When the mappingfile is produced from the template, the metadata is ignored.

To generate the mapping file for the iamc::write.reportProject function, open the ar6/generate_mappingfile.R script and adjust the column names to match the ones in your template and the path to the template. Then call (working directory is the ar6 folder)

Rscript generate_mappingfile.R

This will generate a mapping file, (e.g, mapping_r21m42_AR6DB.csv), to be used with iamc::write.reportProject on a combined MIF file. In the R shell:

iamc::write.reportProject("CombindedOutput.mif", "mapping_r21m42_AR6DB.csv", "AR6Output.mif", missing_log="missing.log")

If variables are not found because they are not present in the output, they will be ignored appended to missing.log. This happens, e.g., if you use output from standalone runs (MAgPIE variables missing) or runs which do not use the new industry or transport modules.

It is in any case recommended to check the missing variables to see which variables are ignored erringly.

If you have many MIFs to process, there is a little helper script, ar6/produce_output.R to assist you with the process.

You can specify the mappingfile and a directory with MIF files. It will produce AR6 compatible outputfiles with a given prefix and will also rename the model name (to be consistent with your model registration).

If required, the script will produce only one output file with all scenarios (the flag GENERATE_SINGLE_OUTPUT). This can easily be copied to the submission template.

The resulting AR6 compatible output files (MIFs) have to be either concatenated and imported as a single sheet (named data) into the submission template (this is the preferable way if you have a lot of scenarios to submit) or as multiple sheets with sheet names data1, data2 etc.

Regarding the meta-data in the submission template, please have a look at exiting (and recent) submissions in the ar6/submission_templates folder. Please add your submission to this folder as soon as you submitted, but please remove the data to not bloat this repository.

For SHAPE WP3, the harmonization of the input data across models is based on the AR6 reporting template which can be found here:

https://data.ene.iiasa.ac.at/ar6-scenario-submission/static/files/global_sectoral/IPCC_AR6_WG3_Global_sectoral_Pathways_scenario_template_v1.3.xlsx

The project partner requires us to provide additionally information on whether a variable is

• exogenous/endogenous (source column)
• of high/medium/low quality (reflecting the level of trust we have in the output)
• interesting for our own modelling effort (a useful input for our REMIND/MAgPIE?).

To facilitate working on the template and enabling version control, the variable list is put into a CSV file together with the additional columns (ar6/shape/shape-template.csv).

The github diff on large semicolon-separated csv files are often completely unreadable. To get a nice output, save the old version of the mapping and the use remind2::compareScenConf:

remind2::compareScenConf(fileList = c("oldfile.csv", "mappingfile.csv"), row.names = NULL)

and paste the text like that into the PR:

<details>
  <summary>Click to expand diff of `mappingfile.csv`</summary>
```
paste text here
```
</details>