How to limit output

[Jakub_Sypek]

What version of the code are you using?
CESM 2.1.5


Have you made any changes to files in the source tree?
N/A


Describe every step you took leading up to the problem:
I ran a CESM case and the output was correctly archived into the plasmalab directory as I wanted. However, my carma history outputs are not in any of the netcdf files and it's still outputting a bunch of variables and tapes I didn't want:
(jupyter) [jjs1u25@login6001 hist]$ ls
satellite_everywhere.cam.h0.0001-01-01-00000.nc satellite_everywhere.cam.h1.0001-01-06-00000.nc satellite_everywhere.cam.h6.0001-01-01-00000.nc
satellite_everywhere.cam.h1.0001-01-01-00000.nc satellite_everywhere.cam.h1.0001-01-11-00000.nc satellite_everywhere.cam.h7.0001-01-01-00000.nc


I only wanted one history tape...

Here is the user_nl_cam file:
! Users should add all user specific namelist changes below in the form of
! namelist_var = new_namelist_value
empty_htapes = .true.
nhtfrq(1) = -24
mfilt(1) = 30
history_carma = .true.


I was hoping to only output some key variables pertaining to carma (history_carma) but not clutter my archive directory with a bunch of variables I don't need.

Describe your problem or question:
Does anyone know how empty_htapes works and how to stop all the default outputs of a given compset from outputting for a particular case? I imagine this is really simple but it seems I've misunderstood something!

 

[peverley]

Hi Jakub,

A few things/questions/requests:

1. What variables are included in the h0 file? Looking at the code, most CARMA fields are being added to the h0 file.

2. Can you attach your atm_in file (from your run directory)? I'd like to take a look at any fincl lists that are being set as a result of your configuration (empty_htapes = .true. SHOULD remove any variables, but there are exceptions).

3. As a note, you can add additional CARMA variables to your output (like LCFM, etc) by specifying "carma_diags_file = 1" (assuming you want the fields to end up on the h0 file with the rest of the CARMA variables)

4. What compset / grid resolution are you using? If you post your full create_newcase command, that would be plenty for me to try to reproduce any issues you're having. I also want to confirm that CARMA is indeed turned on in your configuration.

Courtney

 

[Jakub_Sypek]

Hi Courtney,

Thank you for offering to help!

There are hundreds of variables in the files, including the carma files I wanted now (I'm not quite sure what happened but when I attempted to run a CARMA case the following day it just worked.)

My history output tapes h0, h1, h6, and h7, despite me only requesting h0.

Here is the pertinent section of the user_nl_cam for the run:
! Users should add all user specific namelist changes below in the form of
! namelist_var = new_namelist_value
nhtfrq = -24
mfilt = 30
history_carma = .true.
carma_do_coag = .false.
carma_do_drydep = .false.
carma_do_wetdep = .false.

I have attached the atm_in file.
Ncview of the h0 file gives this:


This is far more variables than I would need for my current analysis and I am trying to save disk space because I am running several multi-year simulations.

The compset is FW2000climo with resolution f19_f19_mg17.

Best wishes,
Jakub

 

[peverley]

Ah, OK. Since you're running a WACCM case, you'll get additional output. And there appear to be other defaults for that configuration that are cluttering your output. In your atm_in file, I see:

history_aero_optics = .false.
history_aerosol = .false.
history_amwg = .true.
history_budget = .false.
history_carma = .true.
history_cesm_forcing = .true.
history_chemistry = .false.
history_chemspecies_srf = .false.
history_clubb = .false.
history_dust = .false.
history_eddy = .false.
history_scwaccm_forcing = .true.
history_vdiag = .false.
history_waccm = .true.
history_waccmx = .false.

Can you try setting the bolded variables to .false. in your user_nl_cam?

history_amwg = .false.
history_cesm_forcing = .false.
history_scwaccm_forcing = .false.
history_waccm = .false.

Once you do that, run "./preview_namelists" in your case directory and look again at the atm_in file. You should see a dramatic decrease in the finclX lists (which are what describe the output for each history file - where X = file# + 1). Any additional variables that are hanging around in your finclX lists you could add to your fexclX namelist variable to get the exact output you want.

Hope that helps!
Courtney

 

[Jakub_Sypek]

Hi Courtney,

This is perfect! Thank you :)

Best wishes,
Jakub