Coupler history file output

[mark@atmos_colostate_edu]

Recently, I discovered this answer from Jim Edwards in a thread from 2022 regarding a question about the coupler history file designations in env_run.xml:

"Coupler output frequency can be adjusted with the XML variables
HIST_OPTION and HIST_N for instantaneous output or
AVGHIST_N and AVGHIST_OPTION for averaged values."

I had previously been using only HIST_OPTION and HIST_N, but I really wanted averaged variables, so I activated AVGHIST_N and AVGHIST_OPTION. I wanted daily means, so I set:

<entry id="AVGHIST_OPTION" value="ndays">
<entry id="AVGHIST_N" value="1">

But the values in the resulting cpl.ha files look suspiciously like what I had been getting for the cpl.hi files using:

<entry id="HIST_OPTION" value="ndays">
<entry id="HIST_N" value="1">

With all four of these options set, I made a short run and compared the output. I've done this with both a B compset and C compset and they behave consistently. For the B compset (BHIST_ctrl using cesm2.2.2), if you compare xaoo_Faox_swdn from BHIST_ctrl.cpl.hi.1850-01-02-00000.nc, it is IDENTICAL to xaooavg_Faox_swdn from BHIST_ctrl.cpl.ha.1850-01-01.nc. And if you plot them, they both look like instantaneous daily fields rather than daily means.

Is there a bug in the code that handles this stuff, or am I doing something wrong?

Thanks,
Mark Branson

 

[dbailey]

Did you try looking at other variables? This version of the model may only be exchanging fluxes with the ocean once a day. Note that the AVGHIST options were removed in subsequent versions of the code. Pinging @jedwards here.

 

[jedwards]

I think that you may need to update to a newer version of the cmeps code.

 

[mark@atmos_colostate_edu]

The coupling frequency in both of my B and C compset tests is hourly (atm_cpl_dt = 3600). The other fields I looked at had the same problem.

Could I update to a newer version of cmeps within my cesm2.2.2 sandbox?