How to format and how to verify atmospheric forcing files for ORCHIDEE offline driver

Author: M. Guimberteau
Last revision: C. Ottle 2020/02/28

Several meteorological forcings are available in the land surface modelling community and are used by the ORCHIDEE team to run the model standalone (not coupled with LMDZ). To this end, the forcing datasets need to be adapted beforehand for their reading and the inter/extrapolation by the driver of ORCHIDEE. The purpose of this protocole is to build a systematic verification when a new forcing is going to be used.

1. NetCDF file construction rules

The ORCHIDEE driver requires specific informations and conventions in the structure of the NetCDF files of the atmospheric forcing to be able to read them. Therefore, it is required to format the raw files following all the informations described at this page: ​https://forge.ipsl.jussieu.fr/orchidee/wiki/Documentation/Forcings

• you can find here examples of scripts which have been used to format files of several forcing:
  • Jinfeng Chang's scripts used for GSWP3 v0 forcing formatting on obelix (LSCE): /home/surface7/jchang/GSWP3_3hr/.
    Warning: this script does not include the nco commands which are required to shift one step backward all the GSWP3 variables. Once you applied Jinfeng's script, you have to apply nco commands, for example (by Vladislav Bastrikov and Nicolas Vuichard):

    ncks -a -d tstep,1, GSWP3_XXX_2002.nc temporary_2002.nc
    ncks -a -d tstep,0 GSWP3_XXX_2003.nc temporary_2003.nc
    cdo -r cat temporary_2002 temporary_2003.nc modif_final_2002.nc
    

  • Chao Yue's scripts used for CRUNCEP forcing. It is originally developed by Bryce Nordgren (​https://github.com/bnordgren/pylsce). In case one day Bryce removed his repository, Chao downloaded them and put them here with addition of explanations in the header of the scripts: ​https://forge.ipsl.jussieu.fr/orchidee/attachment/wiki/Documentation/UserGuide/NewForcingVerif/chao_cruncep.zip

2. Verification of the good reading and inter/extrapolation by the driver

To know if the driver read and inter/extrapolate well the atmospheric forcing files, the best way is to perform one-year offline simulation with ORCHIDEE forced by this new forcing

• the meteorological variables (Tair, Qair, Psurf, Rainf,Snowf, SWdown and LWdown) have to be written in high-frequency (30min) outputs

The 30-min outputs are compared with the original forcing file

• for a given variable (SWdown is a special case, see below), the diurnal cycle should match between the ORCHIDEE 30-min output file and the forcing file
  • example of an IDL script (from Fabienne Maignan on obelix at LSCE) which plots an average seasonal cycle at 30min time step (the script has to be adapted before using it because it uses watchout files that have been removed a long time ago): home/users/maignan/IDL/ORCHIDEE/FORCAGES/CRUNCEP/TestForcagesCRUNCEP.pro

• for SWdown: as the driver also uses the zenith angle when it interpolates the time of SWdown, we need to check its diurnal cycle over a grid-cell at Greenwich, in a 24h-day of June for example, to ensure that SWdown peaks at noon exactly (as recommended at the end of this document: ​https://forge.ipsl.jussieu.fr/orchidee/attachment/wiki/Documentation/Forcings/Description_Forcing_Files.pdf)

3. Indexing of the forcing

All the informations of the new forcing have to be referenced at this page: ​https://forge.ipsl.jussieu.fr/orchidee/wiki/Documentation/Forcings.

4. Storage of the formatted forcing files

After the full check of the new forcing files, you can ask to store the formatted forcing files in the shared accounts in IGCM/SRF/METEO directory.