Daily and monthly-averaged values from CESM

[qpalccl]

I am having research with using daily and monthly-averaged precipitation values from CESM (CESM2 for CMIP6).

I noticed that I need to calculate monthly mean from daily values with one day lagged calendar days (i.e., 2nd JAN ~ 1st FEB 1990) for making identical provided monthly mean (i.e., JAN 1990) values from CESM.

However, at the first month after the model simulation start (initialize) (i.e., JAN 1850 for historical experiment), I could not obtain identical provided monthly-mean value (i.e., JAN 1850) by averaging the equal method (i.e., 2nd JAN ~ 1st FEB 1850 average) or other methods with a few modification (i.e., 1st JAN ~ 1st FEB or 1st JAN ~ 31th JAN 1850).

Please let me know how can I make identical monthly-averaged value from daily values, especially for first month after model simulation start.

Thank you !

 

[dbailey]

Keep in mind that for the model, the timestamp represents the end of the averaging period. So, the 2nd of January 1850 is actually the daily mean value for the 1st of January in the time variable. You can double check this by looking at the time_bounds variable. So, to get the actually monthly mean for January you would take the 2nd of January through the 1st of February. The file names should be correct. When we post-process for CMIP6, this time axis is corrected, but the raw history has a time axis that is the end of the averaging interval. @hannay @strandwg

 

[strandwg]

Dave is correct - the average for Jan 1850 from daily data are samples 2 (Jan 2) to 32 (Feb 1).