Removing ocean from a coupled

I've been asked to remove the ocean from u-am008 so that we're left with an AMIP job. Here's what I've done to create u-am614.

The major changes

The submission scripts, postproc and few other bits we want to be for a AMIP job, so I think it's easiest to begin with a GA7.1 job. I temporarily forgot that u-am008 is UM10.6 and not UM10.7 so I've started with GA7.1 UM10.7, u-al613, which I've copied to u-am614.

From Yongming's u-am008 I've taken

  • Copy app/um/rose-app.conf
  • Copy app/fcm_make_um/rose-app.conf. I needed to change COUPLER from `oasis3_mct' to `none'.
  • Copy app/install_ancil/rose-app.conf, although I only want the atmosphere ancil.

These leaves the coupling options to be removed from u-am008's app/um/rose-app.conf

Remove the coupling flags

To do this, I just need to change

  • ->um
    • ->env
      • ->Coupled Settings
        • ->COUPLER: OASIS_MCT -> none

And this automatically takes out all the l_oasis flags

Removing the coupling from JULES

I basically just need to reverse the instructions on Add coupling options to JULES namelists which has meant doing the following

  • ->um
    • ->namelist
      • ->JULES Science Settings
        • ->JULES radiation options
          • ->l_albedo_obs: false -> true
        • ->JULES sea and sea-ice options
          • ->l_cice_alb: true -> false
          • ->l_saldep_freeze: true -> false
          • ->l_sice_meltponds: true -> false
          • ->l_sice_meltponds_cice: true -> false
          • ->l_sice_multilayers: true -> false
          • ->l_sice_scattering: true -> false
          • ->l_ssice_albedo: true -> false
            • dtice: 2.0
          • ->l_tstar_sice_new: false -> true
          • ->nice: 5 -> 1
          • ->nice_use: 5 -> 1
          • ->kappai_snow: 0.255 -> 0.50

STASH requests

Under ->um ->namelist ->Model Input and Output ->STASH Requests and Profiles ->STASH Requests, I've removed all the fields with `use_name' of `UPCOUP'. These are fields are used to calculate the 3 hour averaged fields which are read into stash codes between 171-197 for passing from ATMOS to ocean.

STASH items

The STASH items can be found in ->um ->namelist ->Reconfigure and Ancillary Control ->Configure ancils and initialise dump fields.

From u-al613 I've added STASH

  • 24 (SST)
  • 31, 32 (Fraction of sea ice in sea and sea ice depth (mean over ice)

and to get this to work I've also had to

  • ->um
    • ->namelist
      • ->Reconfiguration and Ancillary Control
        • Ancil options
          • ->l_amipii_ice_processing: CHANGE false -> true

which implements the AMIP II method where we only need ice concentrations (i.e. we don't seem to need stash 32 in $UM_ANCIL_SEAICE_DIR/$UM_ANCIL_SEAICE_FILE even though we request it from there).

I've ignored the following fields (which were in coupled job)

  • Coupling fields: stash 171-189 and 191-197
  • CO2 flux from ocean: stash 250
  • Sea ice fields: stash 415, 428, 429, 440, 441 and 508.

Choosing a date

The starting date used in u-al613 was 1988-09-01 and the starting date used in u-am008 was 1978-12-01, and I've had to pick a different to either of these. The ainitial file for u-am008 is ah882a.da20771201_00, which means I'm tied to December, and the sea ice data for stash 31 is only available between September 1981 and end of 2012, so I've chosen 1988-12-01. I've changed the following, but maybe I didn't need to change all of them

  • in rose-suite.conf
    • ->BASIS: 19880901T0000Z -> 19881201T0000Z
  • in app/um/rose-app.conf
    • new_date_time: 1978 -> 1988
    • ancil_reftime: 1981,12,1,0,0,0 -> 1988,12,1,0,0,0 (uneccessary to change, I think)
    • mean_reftime: 1981,12,1,0,0,0 -> 1988,12,1,0,0,0 (uneccessary to change, I think)
    • (I also needed to change model_basis_time: $MODELBASIS -> $BASIS_UM_PT, because of suite style change)

Additional change (probably uneccessary)

For stash 31, I've replaced $UM_ANCIL_SEAICE_DIR/$UM_ANCIL_SEAICE_FILE' with /projects/um1/ancil/atmos/n96e/orca1/seaice/reynolds/1981_2012_360/v3/$UM_ANCIL_SEAICE_FILE' to be consistent with u-al613 because I was having some problems with this (before finding the l_amipii_ice_processing change), but it probably isn't needed. Without this change the v2, rather than the v3, directory is used.

Follow-up from Jane

Jane says that use_lookup_dates_anc_time_interp should also be changed from false to true, to be consistent with GA jobs.