Perpetual Run Issue

[jshaman]

Hi,

I'm running CAM 4.0 on a linux cluster with Intel compilers. I'm setting up some perpetual runs using the namelist declarations 'perpetual=.true.' and 'perpetual_ymd=20000101'. The run works fine if I set the perpetual date to 'perpetual_ymd=20000101'--the seasonal cycle is lost and a perpetual January ensues; however, if I change the date to August ('perpetual_ymd=20000801') the run collapses during the first integrations of the land model--a soil balance error develops that causes the run to abort.

Is there a specific initial file (I've tried several, as well as ' ', but these just lead to other errors) that I should use to get around this? Or a surface file (fsurdat)?

Thanks,
Jeff

 

[eaton]

I'm afraid that perpetual mode is not currently working out of the box. It was working in CAM3, but since then several aspects of the model have been updated in a way that doesn't automatically recognize the perpetual mode flags. In particular, the code that interpolates the CAM4 prescribed ozone and aerosol datasets doesn't recognize the perpetual flag, but instead requires namelist settings to fix the interpolation date. See the prescribed_aero_* and prescribed_ozone_* variables. The perpetual flag is fixing the solar zenith angle to the specified date (a diurnal cycle is repeated for the fixed date). Our current use of the perpetual flag is mainly for aqua-planet experiments, so I'm not sure how CLM4 is treating this flag. The bottom line is that you should carefully check that all the forcings are what you expect if you want to use this option.

 

[jshaman]

Thanks for the response.

Jeff

 

[pkelly]

Hi Jeff,
Did you ever discover a work-around for this issue?

Thanks.

 

[yuwei2341@berkeley_edu]

I'm having the same problem of balance error in CLM when conducting perpetual run.
Anyone has any idea?
Thanks

 

[auestad]

I'm afraid that perpetual mode is not currently working out of the box. It was working in CAM3, but since then several aspects of the model have been updated in a way that doesn't automatically recognize the perpetual mode flags. In particular, the code that interpolates the CAM4 prescribed ozone and aerosol datasets doesn't recognize the perpetual flag, but instead requires namelist settings to fix the interpolation date. See the prescribed_aero_* and prescribed_ozone_* variables. The perpetual flag is fixing the solar zenith angle to the specified date (a diurnal cycle is repeated for the fixed date). Our current use of the perpetual flag is mainly for aqua-planet experiments, so I'm not sure how CLM4 is treating this flag. The bottom line is that you should carefully check that all the forcings are what you expect if you want to use this option.

Is a diurnal cycle repeated also for the aerosols and ozone (and all the other options), when fixing the date using the prescribed_* options in the namelist?