output date error - monthly history files shifted 1 month

[micah_wilhelm@env_ethz_ch]

Hi there,
According to my office mates, I'm not the first one to encounter this problem. I have found this issure at the CESM known bugs webpage, however they do not offer a solution.
The monthly history files (h0) that cesm1.04 outputs seem to have their dates shifted by one month into the future (ie filename: experimentname.cam2.h0.2000-10.nc  however 'cdo showdate' gives: 2000-11-01
This seems very strange to me because usually monthly model output would have a day of the month at 15 or 16. Maybe this bug has already been fixed, but are there any recommendations for me to fix this problem a) within the namelist settings/model configuration to avoid this problem in the future and b) in post-processing to correct the data that I already have outputted ?? Many thanks,
micah

 

[eaton]

This is not a bug.  The time coordinate corresponds to the end of the averaging period.  The bounds attribute of the time coordinate variable points you to the variable time_bnds which contains the exact bounds of the averaging period.  Each history file has a fixed output frequency, for example, once per day.  The time operator applied to the output data does not have to be the same for all the fields.  For example some fields could contain daily averages, some the maximum daily value, and others could be instantaneous.  The time coordinate for an instantaneous field will be exact, but for all the others which depend on the domain of the time operator, the time coordinate is located at the end of the time interval and the time_bnds variable tells you exactly what the interval is. 

 

[micah_wilhelm@env_ethz_ch]

Hi Eton,
Thanks for the clarification. Do you by chance have a workaround for cdo or some other post processing language that does not screw up the time bound interpretation? Cheers

 

[eaton]

I'm not familiar with the current status of which post-processing tools recognize and do something reasonable with the bounds attribute which has been part of the CF conventions for about 10 years.  Perhaps others have suggestions. 

 

[l_vankampenhout@uu_nl]

Hi Micah,see this new topic: https://bb.cgd.ucar.edu/external-tools-dont-cesm-time-coordinate

 

[mmills]

updated link for previous reply: https://bb.cgd.ucar.edu/cesm/threads/external-tools-dont-like-cesm-time-coordinate.4604/