Closed cecilehannay closed 1 year ago
The MT/L93 gives an error:
trcdata_init: data type: INTERP_MISSING_MONTHS file: tracer_cnst_halons_3D_L70_
1849-2015_CMIP6ensAvg_c180927.nc
(GETFIL): attempting to find local file
tracer_cnst_halons_3D_L70_1849-2015_CMIP6ensAvg_c180927.nc
(GETFIL): using
/glade/p/cesmdata/cseg/inputdata/atm/cam/tracer_cnst/tracer_cnst_halons_3D_L70_
1849-2015_CMIP6ensAvg_c180927.nc
open_trc_datafile:
/glade/p/cesmdata/cseg/inputdata/atm/cam/tracer_cnst/tracer_cnst_halons_3D_L70_
1849-2015_CMIP6ensAvg_c180927.nc
trcdata_init: file%has_ps = T
HALONS is not an invariant
ERROR: tracer_cnst_init
@fvitt or @tilmes: do you have guidance about that error?
@cecilehannay The case is using trop_mam4
chemistry rather than ghg_mam4
.
Should we change the FMTHIST
to use ghg_mam4
?
@fvitt I think the question would be if we should use trop_mam4 (as it is now) for both FMTHIST and FLHIST. Is there a reason why the ghg_mam4 does not use the same setting as trop_mam4?
Differences are: trop_mam4 includes HALONS as GHGs, and loss of methane and loss of N2O, etc. The specific additional reactions maybe not be relevant for the low-top model, but they may also not hurt to have, in order to have consistency between the two cases.
trop_mam4 has 4 extra tracers.
I could be mistaken, but I thought the plan was to use ghg_mam4
. FLHIST
uses ghg_mam4
.
@cecilehannay can you add 'OCN_FLUX_DMS' for FMTHIST?
@fvitt not sure who decided that, but sure. I am not sure what the differences are, maybe we should ask Dan Marsh? Not sure what his Github name is.
@dan800
I could be mistaken, but I thought the plan was to use ghg_mam4. FLHIST uses ghg_mam4.
I also thought that. Since FLTHIST has CAM%DEV%LT%GHGMAM4
I would've expected FMTHIST to be identical except with an "MT", but instead it's this CAM%DEV%MT
. Maybe we decided to punt due to uncertainties in UBC and gravity wave settings at MT? But those are just namelists; I think we should've set the -chem flag to that in FLTHIST, as that's been settled, right?
@adamrher I am not sure if I was part of the discussion what chemistry to use. Can someone point me to that discussion? Sorry, maybe there was a long github issue and I should refresh my mind. @cecilehannay this chemistry does not include _a5 aerosol, since it is not interactive.
%GHG_MAM4 turns on -chem ghg_mam4. I think most of the discussions is in here https://github.com/ESCOMP/CAM/issues/762 and here https://github.com/ESCOMP/CAM/issues/765.
@dan800 as stated before: Differences are: trop_mam4 includes HALONS as GHGs, and loss of methane and loss of N2O, etc. The specific additional reactions maybe not be relevant for the low-top model, but they may also not hurt to have, in order to have consistency between the two cases.
The advantage of using at least the production of H2O would be that we don't have to produce the external forcing files for water vapor from the full chemistry version. I don't see a discussion on the other tracers like N2O and CFCs. Would you want to comment on that?
@fvitt @cecilehannay @adamrher I am fine to use the -chem ghg_mam4 version for now (but would be great to get Dan's answer).
All physics configurations not controlled by a namelist variable (inc. chemistry) are identical to LT for MT. To begin with just use the LT namelist with a different IC (ncdata) file with MT levels. The default namelist should be pretty much identical to LW except:
1) Non-orogragraphic GWs: Specific tuning of the GW settings for MT vs LT should be discussed here: https://github.com/NCAR/amwg_dev/discussions/248#discussioncomment-5722514
2) The flux boundary condition on H2O should be should be changed to zero flux.
3) Other stuff I may have forgotten about that will show up when we try to run :-)
I could be mistaken, but I thought the plan was to use ghg_mam4. FLHIST uses ghg_mam4.
I also thought that. Since FLTHIST has
CAM%DEV%LT%GHGMAM4
I would've expected FMTHIST to be identical except with an "MT", but instead it's thisCAM%DEV%MT
. Maybe we decided to punt due to uncertainties in UBC and gravity wave settings at MT? But those are just namelists; I think we should've set the -chem flag to that in FLTHIST, as that's been settled, right?
No punting! - we start from LT configuration for MT for all physics modules. Not changes for surface boundary conditions, emissions, anything except changing UBC conditions for H2O (zero flux) and tau_0_ubc = True
No punting! - we start from LT configuration for MT for all physics modules. Not changes for surface boundary conditions, emissions, anything except changing UBC conditions for H2O (zero flux) and tau_0_ubc = True
OK! Then let's plan to discuss how to get this on the trunk quickly at the SE priorities mtg tmrw a.m. @cacraigucar.
Ill note separately that the BMT/BLT compsets we are currently making, are setting GHG_MAM4 for both https://github.com/ESCOMP/CESM/pull/223#pullrequestreview-1417118514. Looking like this B case PR will probably go into a cesm tag today.
As was just pointed out in https://github.com/ESCOMP/CESM/pull/223#pullrequestreview-1417118514, @JulioTBacmeister, @cecilehannay and I have put a hold on making a CESM tag until there are use_cases for BLT1850_v0b and BMT1850_v0b. The discussion for the use_case settings is here: https://github.com/NCAR/amwg_dev/discussions/261
Issue with the FMTHIST compset in cam6_3_109. This run was replaced by #263.
Description: FMTHIST with f.cam6_3_109 + tuning "F"
compset: FMTHIST_v0b add cosp output use clm5.2 dataset
Tuning:
user_nl_cam
user_nl_clm
Case directory: Locally (if still available): /glade/p/cesmdata/cseg/runs/cesm2_0/f.cam6_3_109.FMTHIST_v0b.ne30.tuningF.001
On github: https://github.com/NCAR/amwg_dev/tree/f.cam6_3_109.FMTHIST_v0b.ne30.tuningF.001
Sandbox: Locally (if still available): /glade/work/hannay/cesm_tags/cam6_3_109
On github: hash: 962921a (https://github.com/ESCOMP/CAM/tree/cam6_3_109)
Diagnostics: ADF diags (if available) https://webext.cgd.ucar.edu/FMTHIST/f.cam6_3_109.FMTHIST_v0b.ne30.tuningF.001/atm/
Contacts: