search

NCAR / amwg_dev

Repo to store model sandboxes and cases used for CAM development
9 stars 2 forks source link

New baseline with cam6_3_041 #52

Closed cecilehannay closed 2 years ago

cecilehannay commented 2 years ago

Description of the run

New control with cam6_3_041.

All the development is now on the main trunk. We expect that the cam6_3_041 has the same climate as cam6_3_035.tphysac

Namelist modifications

se_rsplit = 3 se_nsplit = 2 se_hypervis_subcycle = 4

From Peter: The out-of-the-box is for WACCM with a ~140km top that requires more stringent time-stepping (for stability) than the ~80km top model. Peter recommends manually setting the split values so that the model runs faster

Source modifications

Will add source mods to output DQCORE.

Sandbox

cam6_3_041

Any other relevant information

Proposed output:

nhtfrq = 0, -24, -6, -3 interpolate_output = .true.,.true.,.true.,.true. interpolate_nlat = 192,192,192,192 interpolate_nlon = 288,288,288,288

fincl1 = 'U','V','Q','T','PSL','OMEGA','PS','PRECC','PRECL','UTGWORO','VTGWORO', 'TS', 'TAUX','TAUY','TAUBLJX','TAUBLJY','TAUGWX','TAUGWY','FLNT','FLNS','FSNS','FSNT', 'LHFLX','SHFLX','TMQ','FLDS','FSDS','FSDSC','SWCF','LWCF','PRECSC','PRECSL', 'DTCOND','Z3', 'PSL' 'PBLH','ZMDT', 'ZMDT', 'STEND_CLUBB','RVMTEND_CLUBB', 'ZMDQ','CMFMC_DP','DPDLFLIQ','DPDLFICE','PCLDTOP','CLDTOP', 'KVH_CLUBB', 'DTCORE','DQCORE','PTEQ','DCQ','QTGW','CT_H2O','DCQ','MPDQ'

fincl2 = 'TROP_T', 'TROP_P', 'Q', 'T','U850','V850','U200','V200','PRECT','PRECC','PRECL','PBLH', 'OMEGA500','FLUT','Z500','T500','PS','PMID:I','TREFHTMN:M', 'TREFHTMX:X','TSMN:M', 'TSMX:X', 'PBLH','ZMDT', 'ZMDT', 'STEND_CLUBB','RVMTEND_CLUBB', 'PRECT','PRECL','U200','V200','U850','V850','FLUT','Z500','TREFHT','TS','CLDTOT','SWCF', 'LWCF','PS','PSL','TMQ','LHFLX','SHFLX','T500','OMEGA500','CLDLOW'

fincl3 = 'U:I','V:I','T:I', 'OMEGA:I','PS:I','PMID:I', 'OMEGA500','PRECT','U200','U850','FLUT','Q:I', 'ZMDT:I','ZMDQ:I','CMFMC_DP:I','DPDLFLIQ:I','DPDLFICE:I','PCLDTOP:I','CLDTOP:I' , 'DTCORE:I','DQCORE:I','PTEQ:I','DCQ:I','QTGW:I','CT_H2O:I','DCQ:I','MPDQ:I'

fincl4 = 'PRECT','PRECC'

Contact info

@cecilehannay @JulioTBacmeister @adamrher

cecilehannay commented 2 years ago

@PeterHjortLauritzen and/or @jtruesdal

In the L58, we have been using the flags: -cppdefs -Dwaccm_debug in the CAM_CONFIG_OPTS

These don't come out of the box and we have been setting these for ever. I am not even sure what they do. Do you know?

We are now moving to cam6_3_041. I am not sure we still need the flags above. I thought I would ask you.

cecilehannay commented 2 years ago

@PeterHjortLauritzen

I have another question for you.

In our L58 runs, we have been using: se_rsplit = 3 se_nsplit = 2 se_hypervis_subcycle = 4

Now, we want to move to cam6_3_041. Out of the box, we get:
se_rsplit = 4 se_nsplit = 5 se_hypervis_subcycle = 2

Do you want the out of the box values or the values we used in our previous L58 runs?

cecilehannay commented 2 years ago

@adamrher

Please could you check you have all the output you want:

nhtfrq = 0, -24, -6, -3, -1, 1, -24,-24

fincl1 = 'U','V','Q','T','PSL','OMEGA','PS','PRECC','PRECL','UTGWORO','VTGWORO', 'TS', 'TAUX','TAUY','TAUBLJX','TAUBLJY','TAUGWX','TAUGWY','FLNT','FLNS','FSNS','FSNT', 'LHFLX','SHFLX','TMQ','FLDS','FSDS','FSDSC','SWCF','LWCF','PRECSC','PRECSL', 'DTCOND','Z3', 'PSL' 'PBLH','ZMDT', 'ZMDT', 'STEND_CLUBB','RVMTEND_CLUBB', 'ZMDQ','CMFMCDZM','DPDLFLIQ','DPDLFICE','PCLDTOP','CLDTOP', 'KVH_CLUBB', 'DTCORE','DQCORE','PTEQ','DCQ','QTGW','CT_H2O','DCQ','MPDQ' fincl2 = 'TROP_T', 'TROP_P', 'Q', 'T','U850','V850','U200','V200','PRECT','PRECC','PRECL','PBLH', 'OMEGA500','FLUT','Z500','T500','PS','PMID:I','TREFHTMN:M', 'TREFHTMX:X','TSMN:M', 'TSMX:X', 'PBLH','ZMDT', 'ZMDT', 'STEND_CLUBB','RVMTEND_CLUBB', 'PRECT','PRECL','U200','V200','U850','V850','FLUT','Z500','TREFHT','TS','CLDTOT','SWCF', 'LWCF','PS','PSL','TMQ','LHFLX','SHFLX','T500','OMEGA500','CLDLOW' fincl3 = 'U:I','V:I','T:I', 'OMEGA:I','PS:I','PMID:I', 'OMEGA500','PRECT','U200','U850','FLUT','Q:I', 'ZMDT:I','ZMDQ:I','CMFMCDZM:I','DPDLFLIQ:I','DPDLFICE:I','PCLDTOP:I','CLDTOP:I' , 'DTCORE:I','DQCORE:I','PTEQ:I','DCQ:I','QTGW:I','CT_H2O:I','DCQ:I','MPDQ:I' fincl4 = 'PRECT','PRECC' fincl8 = 'Z3:I','U850:I','V850:I','Z500:I','PS:I', 'PMID:I'

PeterHjortLauritzen commented 2 years ago

@PeterHjortLauritzen

I have another question for you.

In our L58 runs, we have been using: se_rsplit = 3 se_nsplit = 2 se_hypervis_subcycle = 4

Now, we want to move to cam6_3_041. Out of the box, we get: se_rsplit = 4 se_nsplit = 5 se_hypervis_subcycle = 2

Do you want the out of the box values or the values we used in our previous L58 runs?

The out-of-the-box is for WACCM with a ~140km top that requires more stringent time-stepping (for stability) than the ~80km top model. Hence I would recommend manually setting the split values so that the model runs faster

PeterHjortLauritzen commented 2 years ago

@PeterHjortLauritzen and/or @jtruesdal

In the L58, we have been using the flags: -cppdefs -Dwaccm_debug in the CAM_CONFIG_OPTS

These don't come out of the box and we have been setting these for ever. I am not even sure what they do. Do you know?

We are now moving to cam6_3_041. I am not sure we still need the flags above. I thought I would ask you.

We used this "#ifdef code" to get diagnostics for max. Courant number in each level. You don't need to set that (unless we are looking into an instability).

adamrher commented 2 years ago

fincl3 = 'U:I','V:I','T:I', 'OMEGA:I','PS:I','PMID:I', 'OMEGA500','PRECT','U200','U850','FLUT','Q:I', 'ZMDT:I','ZMDQ:I','CMFMCDZM:I','DPDLFLIQ:I','DPDLFICE:I','PCLDTOP:I','CLDTOP:I' , 'DTCORE:I','DQCORE:I','PTEQ:I','DCQ:I','QTGW:I','CT_H2O:I','DCQ:I','MPDQ:I'

You'll need to change CMFMCDZM -> CMFMC_DP. Otherwise it's fine.

It would be nice to clean up these fincl's a bit. These are mere suggestions; you can take it or leave it. Make all fincl3 vars instantaneous and all fincl2 averages. Then get rid of fincl8, so you can reduce the nhtfrq array to nhtfrq = 0, -24, -6, -3

cecilehannay commented 2 years ago

The fincl8 was on the native while the other history files were interpolated on the fly on a lat/lon grid.

If we get rid of fincl8, I want to make sure the person who uses this is ok with the interpolated history files. Is it you @isla?

islasimpson commented 2 years ago

Don't think it was me. I don't like things on the native grid...

adamrher commented 2 years ago

@cecilehannay FYI, I've confirmed with Mariana that cam6_3_041 already includes the shr_flux_mod changes, so you no longer have to SourceMod those in (apologies if you already knew this).

cecilehannay commented 2 years ago

Thanks. I had checked that too.

The only mods I need are the ones to output the water vapor tendency due to dycore (DQCORE)

cecilehannay commented 2 years ago

@PeterHjortLauritzen

I will use: se_rsplit = 3 se_nsplit = 2 se_hypervis_subcycle = 4 like in previous runs.

swrneale commented 2 years ago

I think ZMDT is on one fincl twice instead of ZMDQ

cecilehannay commented 2 years ago

This run is started as f.e21.FWscHIST.ne30_L48_BL10_cam6_3_041_control.hf.001.
See: https://github.com/NCAR/amwg_dev/issues/53