• Python 3.7 or higher
• Python packages:
  • numpy
  • netCDF4
  • xarray
  • matplotlib
  • Optional: pylatex to prepare atlas pdf files.

To install SCM-atlas on a CNRM computer, with access to the CNRM Lustre system:

1. Get the installation script:

   wget https://raw.githubusercontent.com/romainroehrig/SCM-atlas/master/install.sh

2. Modify install.sh:

   • Set ATLAS_VERSION, e.g., ATLAS_VERSION=1.3
   • Set where you want to install SCM-atlas tools: default is DIR_ATLAS=$HOME/Tools/SCM-atlas/V${ATLAS_VERSION}
   • Set where you want to run SCM-atlas tools: default is DIR_RUN=$HOME/Atlas1D/V${ATLAS_VERSION}
   • Set where the reference datasets (e.g., LES) can be found: default is DIR_REF=/cnrm/amacs/USERS/roehrig/share/SCM-atlas/References/V1.0

3. Execute install.sh. A test is done at the end with provided SCM simulations with ARPEGE-Climat 6.3.1 for the ARMCU/REF, RICO/SHORT and SANDU/REF cases. Check the output atlas at the link provided at the end of the test.

1. Go in the DIR_RUN directory

2. Source setenv to have the right PATH, PYTHONPATH and SCM_REFERENCES environment variables

   source setenv

3. Prepare you own config file, following the example config/config_CMIP6.py

4. Run SCM-atlas:

   run_atlas1d.py -config config/YOUR_CONFIG_FILE

   You can add the option --pdf to export pdf files for each case.