Opened 2 years ago

Last modified 20 months ago

#330 assigned

PI control configuration

Reported by: Martin Dix Owned by: Martin Dix
Priority: major Component: ACCESS-CM2
Keywords: Cc:

Description

Setup proper PI control configuration

Change History (7)

comment:1 Changed 2 years ago by Martin Dix

Met Office PI control experiments are described in https://code.metoffice.gov.uk/trac/ukcmip6/ticket/8 with details in https://code.metoffice.gov.uk/trac/ukcmip6/wiki/ticket/8/SuiteSetup.

Standard suite is u-am927.

comment:2 Changed 23 months ago by Martin Dix

In this suite app/install_ancil/rose-app.conf has

[file:$ROSE_DATA/etc/um_ancils_gl]
source=fcm:ancil_data.xm_tr/ancil_versions/n96e_orca1/GA7.1/v1/ancils@4110

Other files with an explicit path in the suite include

SPECTRAL_FILE_DIR=$CMIP6_ANCILS/any/clim_1850-1873/Solar/v3.2_picontrol_cmip6_solar_244

and

ancilfilename='$CMIP6_ANCILS/n96e/timeslice_1850/OzoneConc/v1/mmro3_monthly_CMIP6_1850_N96_edited-ancil_2anc'

The CMIP6_ANCILS directory on JASMIN has these and also the other files that the suite depends on. This has now been copied to /g/data/access/TIDS/CMIP6_ANCILS.

The suite has

ukca_em_files='/data/d01/tkuhlbro/ancils/ukca_emiss/ukca_emiss_BC_biofuel_1850.nc',
             ='/data/d01/tkuhlbro/ancils/ukca_emiss/ukca_emiss_BC_fossil_1850.nc',
             ='/data/d00/hadtr/ancil/ukca_emiss/cmip5/1860/v2/ukca_emiss_DMS.nc',
             ='/data/d00/hadtr/ancil/ukca_emiss/cmip5/1860/v2/ukca_emiss_Monoterp.nc',
             ='/data/d01/tkuhlbro/ancils/ukca_emiss/ukca_emiss_OC_biofuel_1850.nc',
             ='/data/d01/tkuhlbro/ancils/ukca_emiss/ukca_emiss_OC_fossil_1850.nc',
             ='/data/d01/tkuhlbro/ancils/ukca_emiss/ukca_emiss_SO2_high_1850.nc',
             ='/data/d01/tkuhlbro/ancils/ukca_emiss/ukca_emiss_SO2_low_1850.nc',
             ='/data/d00/hadtr/ancil/ukca_emiss/andres_kasgnoc/v2/ukca_emiss_SO2_nat_r2428.nc',
             ='/data/d01/tkuhlbro/ancils/ukca_emiss/ukca_emiss_BC_biomass_high_1850.nc',
             ='/data/d01/tkuhlbro/ancils/ukca_emiss/ukca_emiss_BC_biomass_low_1850.nc',
             ='/data/d01/tkuhlbro/ancils/ukca_emiss/ukca_emiss_OC_biomass_high_1850.nc',
             ='/data/d01/tkuhlbro/ancils/ukca_emiss/ukca_emiss_OC_biomass_low_1850.nc'

There are files of the same name on JASMIN in directory CMIP6_ANCILS/users/bosley/ancils_N96e. For now this directory has also been copied to /g/data/access/TIDS/CMIP6_ANCILS.

TODO

  • Check that the Bosley files are the ones actually being used
  • Add netcdf attribute so they work with any calendar

comment:3 Changed 23 months ago by Martin Dix

As first step, try setting up an AMIP run with the PI control configuration from the Met Office u-am927 coupled run.

u-am927 is vn10.7 and includes the branch branches/dev/benjohnson/vn10.7_easyaerosol_cmip6@36305. easy_aerosol used for the volcanic forcing.

There's also a vn10.6.1_easyaerosol_cmip6 branch with the comment

Branch at vn10.6.1 necessary for CMIP6 to avoid conflicts with vn10.6_aer_rad_diag that was lodged for vn10.6.1

so it looks like we may have to use vn10.6.1 rather than vn10.6.

First set up an vn10.7 AMIP run with the PI forcing and then check that this can be matched with a vn10.6 one.

I checked that the NCI version of the standard vn10.7 AMIP run (u-ak889) matched the 10.6 version (u-aj458). Both using ga7.1 and the 360 day calendar.

u-ap790 is a vn10.7 PI control atmosphere only suite.

The intel compiler flagged a problem with passing parameters to a routine that had an intent(inout) argument specification. Temporary fix in working copy ~/mrd599/mosrs/um/vn10.7_easyaerosol_cmip6.

Now failed with error trying to unpack the time:units attribute of emissions files. Code comment says this is expected to be of form days since yyyy-01-01 00:00:00 but several files have days since 1750-01-01 0:0:0. Fixed using ncatted (from nco), e.g.,

ncatted -a units,time,o,c,'days since 1850-01-01 00:00:00' ukca_emiss_OC_biomass_low_1850.nc

Also need to change update_type attribute to 2 (periodic) for a couple of files

ncatted -a update_type,global,o,c,"2" ukca_emiss_BC_biomass_low_1850.nc
ncatted -a update_type,global,o,c,"2" ukca_emiss_BC_biomass_high_1850.nc 
ncatted -a update_type,global,o,c,"2" ukca_emiss_OC_biomass_low_1850.nc 
ncatted -a update_type,global,o,c,"2" ukca_emiss_OC_biomass_high_1850.nc 

One month run with u-ap790 was now successful.

Last edited 22 months ago by Martin Dix (previous) (diff)

comment:4 Changed 22 months ago by Martin Dix

Owner: set to Martin Dix
Status: newassigned

comment:5 Changed 22 months ago by Martin Dix

I set up a vn10.6 branch that includes all the vn10.6.1 changes and the easy_aerosol changes, branches/dev/martindix/vn10.6_nci_cmip6_merge.

A vn10.6 AMIP job using this, u-ap751, gives identical results to an vn10.7 AMIP job, u-ap790-gregorian.

comment:6 Changed 22 months ago by Martin Dix

Suite u-aq959 is a PI control coupled model configuration. Atmospheric configuration modified to match u-ap751. This is the same as the Met Office PI control configuration from u-am927, except for the 1850 vegetation (doesn't match our land mask so can't be used).

Run failed with the error

?  Error from routine: SOLVAR
?  Error message: No solar variability data for model time.

/g/data/access/TIDS/CMIP6_ANCIL/data/ancils/any/clim_1850-1873/Solar/v3.2_picontrol_cmip6_solar_244/sp_sw_ga7_var
had year set as 1850 (code would work for any model year >= that). Changed file year to 0001.

Last edited 20 months ago by Martin Dix (previous) (diff)

comment:7 Changed 21 months ago by Martin Dix

Further suite changes

  • Add restart capability from #344.
  • Use repository for scripts (#329)
  • Reduce spurious warnings (#315)
  • Start from 19890101 of u-aj458 AMIP run (#342)


TODO

  • Postprocessing
  • Water budget fix (Ice bergs?)
  • Reduce data volumes for testing
  • Size of log files - are they returned.
Last edited 21 months ago by Martin Dix (previous) (diff)
Note: See TracTickets for help on using tickets.