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
Current daily data processing (as introduced in #9) has to make some pretty rough assumptions about sensible time chunks.
The current chunking uses a mixture of 360-day and Julien years and can only assume 30 day months. It would be more user friendly to chunk into calendar months and years as per the underling calendar.
As far as I can tell the ESGF metadata doesn't include calendar information and this would have to be discerned from the time coordinate metadata in each file. I think it's safe to assume the calendar is the same for multiple files with the same ID, but will differ by model and may well differ between experiments and activities for each model.
The text was updated successfully, but these errors were encountered:
Current daily data processing (as introduced in #9) has to make some pretty rough assumptions about sensible time chunks.
The current chunking uses a mixture of 360-day and Julien years and can only assume 30 day months. It would be more user friendly to chunk into calendar months and years as per the underling calendar.
As far as I can tell the ESGF metadata doesn't include calendar information and this would have to be discerned from the
time
coordinate metadata in each file. I think it's safe to assume the calendar is the same for multiple files with the same ID, but will differ by model and may well differ between experiments and activities for each model.The text was updated successfully, but these errors were encountered: