Installation

1. Install julia (version 1.6.2 or above was used for this workflow), see instructions and download here.

Important:
Make sure julia can be executed from the shell, e.g. add julia to your PATH environment as described in the documentation.

2. From within the SynDe project directory open a shell and startup julia with the project environment enabled:

   SynDe$ julia --project=workflow/envs/
   

   Open the package manager to check if the environment is activated and build the packages (may take some time)

   julia> ]
   
   (envs) pkg> status
       Status `~/share/GitHub/SynDe/workflow/envs/Project.toml`
   [31bfc850] GlobalEnergyGIS v0.1.0 `https://github.com/niclasmattsson/GlobalEnergyGIS#master`
   
   (envs) pkg> instantiate
   

3. Configure GlobalEnergyGIS (GEGIS) with user account data and work folder location

   julia> using GlobalEnergyGIS
   
   julia> saveconfig(string(pwd(),"/","resources/gegis"), <your CopernicusID>, "<your Copernicus API key>", agree_terms=true)
   

   For instruction what to enter for <your ...> see the the first steps of the GEGIS documentation on how to setup GEGIS.

Region selection and definitions

The snakemake workflow describe below creates regions for GlobalEnergyGIS to work on. The regions are defined in config/regions.yaml. The definitions contain:

• The region name (e.g. Africa)
• The subregions associated with the region, with an entry for each region specifying
  • An uniquely identifying code (e.g. ISO 3166-2 two letter country code for countries)
  • The subregion name as recognised by GlobalEnergyGIS, i.e. as recognised by the GADM database (version 2.8 / 2018)
• example entry for a region with one subregion:

  my_region:
      CH: Switzerland
  

Note: Some subregions may not work as they break the GlobalEnergyGIS algorithm, which fails to determine population centers for very small or sparsely populated subregions. The following subregions are therefore missing from the definition and no synthetic demand is created for them:

• "Antarctica"
• "Bouvet Island"
• "Heard Island and McDonald Islands"
• "French Southern Territories"
• "Saint Helena, Ascension and Tristan da Cunha""
• "Cocos (Keeling) Islands"
• "Christmas Island"
• "British Indian Ocean Territory"
• "Korea, Republic of"
• "Lao People's Democratic Republic"
• "Macao"
• "Guernsey"
• "Jersey"
• "Monaco"
• "Svalbard and Jan Mayen"
• "Aruba"
• "Anguilla"
• "Saint-Barthélemy"
• "Bermuda"
• "Cayman Islands"
• "Saint Kitts and Nevis"
• "Saint Martin (French part)"
• "Montserrat"
• "Sint Maarten (Dutch part)"
• "United States Minor Outlying Islands"
• "Virgin Islands, British"
• "Virgin Islands, U.S."
• "American Samoa"
• "Cook Islands"
• "Micronesia, Federated States of"
• "Marshall Islands"
• "Norfolk Island"
• "Niue"
• "Nauru"
• "Pitcairn"
• "Tokelau"
• "Tuvalu"
• "Wallis and Futuna"
• "South Georgia and the South Sandwich Islands"

Caveats, Limitations, Manual interventions

• In the used version (2021-08-09) during the download_datasets() step, the protected land database (WDPA) did not download correctly, possibly due to a wrong URL. Changing the URL inside the GlobalEnergyGIS code to

  "WDPA"         ("WDPA (protected areas):", "WDPA.zip",
  "https://d1gam3xoknrgr2.cloudfront.net/current/WDOECM_Aug2021_Public_shp.zip")
  

  solved the issue.
• In the used version (2021-08-09) during the download_datasets() step, the extraction step of the WRI database did not work properly. The files were extracted but failed to be moved to the correct location, this was then done manually and the faulty step removed from the GlobalEnergyGIS, then the step was repeated.
• A fix to the timezone handling in GEGIS from this GitHub PR was applied.
• The demand projection does not work for some countries, e.g. Mauritania. Affected countries are shown in grey in the world map example below.
• SSP4 and SSP5 are not available in GEGIS and thus this workflow.

Example data

This workflow produces synthetic hourly resolved electricity demand under various scenarios.

Excerpt Germany hourly resolved electricity demand for different target years and ERA5 basis years for the SSP2-RCP2.6 scenario.

Statistical representation of German electricity demand for different target years and ERA5 basis years for the SSP2-RCP2.6 scenario.

Annual electricity demand by country for 2030 (ERA5 data for 2013) in the SSP2-RCP2.6 scenario:

License

Copyright 2021 by Johannes Hampp .

This code is licensed under MIT. The Readme.md file is licensed as CC-BY-4.0. The configuration files are licensed CC0-1.0.