Open ehogan opened 5 days ago
could be cf_units
- latest release is from about a month ago https://anaconda.org/conda-forge/cf-units
@ehogan Would you be able to share a bit more of the stack trace?
@ehogan Would you be able to share a bit more of the stack trace?
Apologies for the delay; I have updated the first comment of this issue with the full traceback 👍
It looks like this happens for variable rsus
. On Levante, the rsus
data is e.g. /work/bd0854/DATA/ESMValTool2/RAWOBS/Tier3/ERA5/v1/mon/rsus/era5_mean_surface_net_short_wave_radiation_flux_1995_monthly.nc. In the ERA5 documentation, Mean surface net short-wave radiation flux
is listed in Table 4 and not in Table 3 which lists accumulated variables. Yet, in the fix it is assumed that this is an accumulation: https://github.com/ESMValGroup/ESMValCore/blob/60aec3e5abe9a58db2f7256e979e2f4a888a0296/esmvalcore/cmor/_fixes/native6/era5.py#L285.
Therefore I suspect that either this is the wrong input data (maybe rsus maps to another variable?) or the fix is wrong (the call to fix_accumulated_units
should be removed).
Describe the bug
recipe_climwip_brunner2019_med.yml
ran successfully when testing ESMValCore v2.11.0rc1, see https://github.com/ESMValGroup/ESMValCore/issues/2421.@ESMValGroup/technical-lead-development-team would any of you have any indication of whether one of the recently added PRs may have caused this error? 🤔