Moving generate_biomee/pmodel_driver_data.R from FluxDataKit to rsofun

[marcadella]

Related to issue #228 and the fact that we are homogenizing the interfaces: should we move rsofun/data-raw/generate_biomee_driver_data.R to DataFluxKit? It looks like it is related to FluxDataKit/analysis/05_generate_demo_rsofun_drivers.R

[fabern]

This would also affect the generation of pmodel drivers.

The README.md states that the use https://github.com/geco-bern/FluxDataKit/blob/main/analysis/05_generate_demo_rsofun_drivers.R. However, this script has disappeared in the meantime. For the phydro branch the forcing input is anyway extended with the columns: forcing, forcing_24h, forcing_daytime, and forcing_3hrmax

• use the phydro-updated data (generated with this script and the function fdk_format_drivers()) including the columns: forcing, forcing_24h, forcing_daytime, and forcing_3hrmax
• also include site properties AI and MAT (according to Add site meta info FluxDataKit#100)

[stineb]

The mentioning of rsofun/data-raw/generate_biomee_driver_data.R can be removed from FluxDataKit. The complete driver object for all sites and years is generated in FluxDataKit (analysis/02_batch_format_rsofun_drivers.R). In rsofun, let's keep the scripts that generate the demo data from that (subsetting to one site and a couple of years within that). This is done by data-raw/generate_biomee_driver_data.R and data-raw/generate_pmodel_driver_data.R. In that sense, I think there is no change needed for creating the P-model drivers. But then, the BiomeE drivers generated by rsofun/data-raw/generate_biomee_driver_data.R may be updated?

[fabern]

The way I understand it, there are the following tasks still to be done:

• remove mentions of rsofun data generation from the FluxDataKit repository [EDIT: following this PR: https://github.com/Minor cleanup FluxDataKit#101]
  • also consider treating/following up the issues: obs_eval_fluxnet2015.Rdata wrong data sofunCalVal#18 and fdk worksflow for preparing forcing data sofunCalVal#5
• regenerate drivers for p-model and BiomeE-model
  • generation scripts should be put: into the data-raw-folder of the repo rsofun
    • for p-model it should use analysis/b_02_batch_format_rsofun_drivers.R from the updated FDK PR
      • for p-model it currently uses the Zenodo files generated with analysis/b_02_batch_format_rsofun_drivers.R
    • for BiomeE there is already a script generate_biomee_driver_data.R
      • but is missing a script that generates CH-LAE_forcing.rda
  • the data-raw-folder of the repo rsofun also contains data needed for the generation
    • for BiomeE this is CH-LAE_forcing.rda
    • for p-model should we add a reduced FDK csv (containing only FR-Pue, Reichetal_Colorado, Reichetal_New_Mexico, Reichetal_Venezuela, and Reichetal_Wisconsin)?
  • updated drivers replace the *.rda-files in the data-folder of the repo rsofun