You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Is your feature request related to a problem? Please describe.
ECMWF have updated the grib-to-netCDF converter used in downloading the netCDF version of ERA5 (and presumably also for ERA-Interim): https://confluence.ecmwf.int/display/CKB/GRIB+to+netCDF+conversion+on+new+CDS+and+ADS+systems.
This has changed the format of the ERA5 files, and the ESMValTool ERA5 on-the-fly cmoriser might need to be updated to work with the new format. I say might because there are a couple of problems with the format that I've requested be fixed at source:
Non CF-compliant variable expvar: https://forum.ecmwf.int/t/netcdf-variable-expver-not-cf-compliant-in-era5-new-converter/7148
Reading the new format data into ESMValTool produces a "coordinate not found" error, which could potentially be caused by the first of these. However, I propose that we wait for responses to these issues before taking action on the cmoriser.
In the meantime, here is some guidance to ESMValTool users downloading ERA5 data:
Avoid the last few months of the timeseries. My understanding (though I've not fully investigated) is that that if the data overlaps with the near-real-time "initial release" period (see https://climate.copernicus.eu/climate-reanalysis) then it adds an additional coordinate which breaks the ESMValTool cmorisor, and also will double the size of your netCDF file.
@ESMValGroup/obs-maintainers, where do you think would be the appropriate places to document the above interim advice? Is an issue the best place for advice that might evolve quite quickly?
Would you be able to help out?
Yes
The text was updated successfully, but these errors were encountered:
I would personally wait a little before adapting to the new CDS since there are still a few things that might change. We could potentially look into pinning one of these issues so it doesn't get drowned in between other things. That might also remind us to keep it more up-to-date.
Is your feature request related to a problem? Please describe.
ECMWF have updated the grib-to-netCDF converter used in downloading the netCDF version of ERA5 (and presumably also for ERA-Interim): https://confluence.ecmwf.int/display/CKB/GRIB+to+netCDF+conversion+on+new+CDS+and+ADS+systems.
This has changed the format of the ERA5 files, and the ESMValTool ERA5 on-the-fly cmoriser might need to be updated to work with the new format. I say might because there are a couple of problems with the format that I've requested be fixed at source:
expvar
: https://forum.ecmwf.int/t/netcdf-variable-expver-not-cf-compliant-in-era5-new-converter/7148Reading the new format data into ESMValTool produces a "coordinate not found" error, which could potentially be caused by the first of these. However, I propose that we wait for responses to these issues before taking action on the cmoriser.
In the meantime, here is some guidance to ESMValTool users downloading ERA5 data:
@ESMValGroup/obs-maintainers, where do you think would be the appropriate places to document the above interim advice? Is an issue the best place for advice that might evolve quite quickly?
Would you be able to help out?
Yes
The text was updated successfully, but these errors were encountered: