CESM1_0_6 user_nl_cam values usurped

[adrianne]

I'm running a job using the B_1850_CAM5_CN compset, and setting numerous values in user_nl_cam. For the most part that works, however, when I set:
prescribed_ozone_datapath = '/project/projectdirs/ccsm1/inputdata/atm/cam/ozone'
prescribed_ozone_file = 'ozone_1.9x2.5_L26_1850-2005_c090803.nc'
prescribed_ozone_name = 'O3'
prescribed_ozone_type = 'INTERP_MISSING_MONTHS'

However, when I build the job, the following line shows up in both Buildconf/cam.buildnml.csh and $RUNDIR/atm_in
prescribed_ozone_cycle_yr = 1850

This results in this error:
ozone is prescribed in :ozone_1.9x2.5_L26_1850-2005_c090803.nc
ENDRUN:trcdata_init: Cannot specify data_cycle_yr if data type is not CYCLICAL

When I set:
tracer_cnst_datapath = '/project/projectdirs/ccsm1/inputdata/atm/cam/chem/trop_mozart_aero/oxid'
tracer_cnst_file = 'oxid_1.9x2.5_L26_1850-2005_c091123.nc'
tracer_cnst_filelist = 'oxid_1.9x2.5_L26_clim_list.c090805.txt'
tracer_cnst_specifier = 'O3','OH','NO3','HO2'
tracer_cnst_type = 'INTERP_MISSING_MONTHS'

However, when I build the job, the following line shows up in both Buildconf/cam.buildnml.csh and $RUNDIR/atm_in
tracer_cnst_cycle_yr = 1850

ENDRUN:trcdata_init: Cannot specify data_cycle_yr if data type is not CYCLICAL

In cesm1_0_6, it's easy enough to remove it from the cam.buildnml.csh file, and run. I'm very worried that this may not be fixed in later releases.

 

[santos]

I'm not sure whether or not this is addressed in other releases. But I think that one alternative solution is to set the cycle_yr variable to zero.

 

[santos]

I'm not sure whether or not this is addressed in other releases. But I think that one alternative solution is to set the cycle_yr variable to zero.

 

[eaton]

I agree that setting prescribed_ozone_cycle_yr and tracer_cnst_cycle_yr  to 0 is the best solution (although clearly it's not obvious).This is a feature, not a bug:)  The B_1850_CAM5_CN compset is designed for an 1850 climatology run which is a cyclic mode with respect to the boundary datasets.  The values of the cycle year (1850) are coming from the CAM use case file atm/cam/bld/namelist_files/use_cases/1850_cam5.xml which is used to implement the compset.  In order to change certain boundary datasets from cyclic to transient *all* the relevent namelist variables must be changed consistently.  The code is flagging the inconsistent settings and asking the user to fix them rather than silently making a choice which may not be what the user intended.Yes, I agree that our error checking and message to the user about what needs to be done to fix the problem could be improved.  

 

[eaton]

I agree that setting prescribed_ozone_cycle_yr and tracer_cnst_cycle_yr  to 0 is the best solution (although clearly it's not obvious).This is a feature, not a bug:)  The B_1850_CAM5_CN compset is designed for an 1850 climatology run which is a cyclic mode with respect to the boundary datasets.  The values of the cycle year (1850) are coming from the CAM use case file atm/cam/bld/namelist_files/use_cases/1850_cam5.xml which is used to implement the compset.  In order to change certain boundary datasets from cyclic to transient *all* the relevent namelist variables must be changed consistently.  The code is flagging the inconsistent settings and asking the user to fix them rather than silently making a choice which may not be what the user intended.Yes, I agree that our error checking and message to the user about what needs to be done to fix the problem could be improved.  

 

[adrianne]

Thanks for the suggestion. I'll start setting the cycle_yr variables to 0. 

 

[adrianne]

Thanks for the suggestion. I'll start setting the cycle_yr variables to 0.