Swapping ORCA025 for ORCA1

Introduction

The standard resolution of ocean used in Met Office model nows seems to be ORCA025 but for the UKESM model we want ORCA1. Hence we need to be able to replace the ORCA025 ocean for our UKESM configurations. The Met Office use to use ORCA1, but - unfortunately for us - their switch to ORCA025 as standard has coincided with NEMO extending their grid, which means that extended ORCA1 input files are not readily available within the Met Office.

Till has already done much of the work in replacing the ORCA025 ocean in UM10.3 GC3 in his job u-ac043. This page outlines how I've taken Till's ORCA1 ocean and inserted it into Yongmin's pseudo UKESM 0.4 job, u-ab330.

I've also done the same for the UM10.4 jobs

• Yongming's UM10.4 pseudo UKESM0.4 is u-ac710
• Till's UM10.4 GC3 with ORCA1 is u-ac926
• My mixture of the two is u-ac847

Script used to swap ocean for pseudo UKESM0.4

My personally preference is typically to write a script to create a new configuration, because it's easy to make big changes, it keeps track of the changes which have been made and hopefully a few small alterations to the script will allow it to be useful for swapping the ocean for another configuration. The downside is that is can be faffy to make one simple change, so it's unlikely to be how most people will do it.

The script I've written to insert Till's ORCA1 ocean into Yongming's job is ~mstringe/bin/rose/swapOceanResol.scr. The steps this script takes are

1. Copies every file and directory, except rose-suite-info, from Yongming's job, u-ab330, to my build job, u-ac466
2. For ~/roses/u-ac466/rose-suite.conf, change the following variables to what Till has (u-ac407 is my copy of Till's job, u-ac043)
   • CICE_BLKX, CICE_BLKY, CICE_COL, CICE_MAXBK, CICE_ROW, NEMO_IPROC, NEMO_JPROC, NEMO_LAND_SUPPRESS (set to false) and XIOS_NPROC. And comment out NEMO_NPROC.
3. Replace the following apps with those used by Till
   • fcm_make_ocean
   • install_ancil
4. In directory app/coupled/file, copy the following files from Till's equivalent directory
   • domain_def.xml
   • iodef.xml
   • namcouple
   • xmlio_server.def
5. For app/coupled/rose-app.conf
   1. Change the following variables for what Till has
      • NEMO_VERSION, RMP_DIR, land_field, meta, oasis_couple_freq_ao and oasis_couple_freq_oa. Note: NEMO_VERSION seems need to be 304, because um-coupled contains the lines

                # Prior to 3.4 NEMO dump date is always the day before the 
        restart date. 
                # We want the restart date so increment the date by 1.
                if [[ $NEMO_VERSION != "304" ]]; then
                    (( NEMO_DAYS_TO_DUMP = NEMO_DAYS_TO_DUMP + 1 ))
                fi
        

   2. Swap any namelists which appear in the lists for either [file:namelist_cfg] (the NEMO namelists) or [file:ice_in] (the CICE namelists) for the versions that Till has.
   3. Swap both UM_ANCIL_LSH_TINDMN_FILE and UM_ANCIL_LSH_TINDSD_FILE for UM_ANCIL_LSH_TIND_FILE (Yongming's job uses two separate ancillaries for these two fields, but Till's has one file containing both.)
   4. Add in the following file declarations
      • [file:ahmcoef.nc]
      • [file:geothermal_weights.nc]
   5. Remove [file:iceshelf_melting.nc] (I believe that another file is being used in namsbc_isf, probably runoff_1m_nomask, and removing this means that ICE_MELT does not have to be defined.)
6. For app/fcm_make_um/rose-app.conf, add in Richard's branch, branches/dev/richardhill/vn10.3_N96_eORCA1_cpl@17638 or branches/dev/richardhill/vn10.4_N96_eORCA1_cpl@20048. (In um-coupled there's a line which removes the RMP weights files with something like rm *weights*. Unfortunately, this also removes geothermal_weights.nc, and so Richard's branch is needed to prevent this.)