Too for parallel retrieve of ECMWF ERA5 from the Climate Data Store.
It is built on cdsapi and uses CDO and its python bindings for the postprocessing. Netcdf4 and xarray are also recommended altough not stricly required. Parallelization is done on yearly basis. Postprocessing provides daily and monthly files using CDO.
You can configure the ERA5_retrieve_postproc.py script (no command line interface but manual configuration required) and run it using python3.
It is recommended to run it under conda to do not mess with dependencies
conda env create -f environnment.yaml -n CDS
conda activate CDS
./ERA5_retrieve_postproc.py -c config.yaml -n 2
This will run with 2 processors the predefined configuration from config.yaml. Please modify the config.tmpl according to your needs
You can configure:
- Temporary and storage directories:
tmpdirandstoredir - Variable to be downloaded:
var - First and last year to be extracted:
year1andyear2 - Number of processor for parallel download:
nprocs(it uses themultiprocessingpython package) - Level you want to download:
levelout(it supports surface and a few predefined pressure levels) - Grid on which you want to download:
grid - Area on which download:
area(it could be global or sub-selected according to CDS vocabulary)