ocean

There is a verification called global_ocean.cs32x15 that uses the same grid as aim.5l_cs. So, let’s get used to it before trying the atmosphere-ocean coupled model.

The description of this example is provided in the README file.

global_ocean.cs32x15

global ocean using the cubed-sphere grid 32x32x32 with 15 levels

=================================================================

Specific option:

• Use Non-Linear Free surface formulation with z* coordinate with real fresh-water flux.
• Oceanic set-up on the cubed-sphere grid using the vector-invariant formulation.

Forcing : use Monthly mean climatological forcing (except P-E-R, annual mean). same data set as global-ocean lat-long experiments but interpolated on CS-32 grid.

Comments:

• bathymetry : designed to be coupled to Atmospheric model, therefore includes most of the semi-enclosed sea (Mediterranean, Black-Sea, Red-Sea, Hudson Bay …) bathy_cs32.bin: initial bathymetry h < 0 is meant to stay wet-point whatever delZ(1) is ; Consequently the global ocean area is not affected by the vertical resolution. bathy_Hmin50.bin: bathymetry file used in the current set-up generated from bathy_cs32.bin using matlab script mk_bathy4gcm.m mk_bathy4gcm.m matlab script that deepen all shallow point up to 50m.

• global integral of E-P-R and annual mean net Q flux are zero.

• package thSIce and bulk_forc are included but not used in the standard set-up.

• additional forcing fields and parameter files are provided (in input.thsice) in order to illustrate the use of thSIce pkg. the output of a short run (20.iter) is given in results/output_thsice.txt

October 1rst, 2005:

• input.viscA4/data has been added to test biharmonic viscosity on CS-grid with side-drag. However, this set of parameters has only be used for short tests and is not recommended to begin with.

It is similar to what we did for the atmospheric model. First, open the terminal and go to the example directory cd /MITgcm/verification/global_ocean.cs32x15

We also want to use more than 1 cpu, we use SIZE.h_mpi.

• go to code directory
• rename SIZE.h as SIZE.h_single
• rename SIZE.h_mpi as SIZE.h \(\rightarrow\) It specifies the number of cpus as 4.
• go to build directory
• do ../../../tools/genmake2 -mods ../code -optfile ../../../tools/build_options/linux_amd64_gfortran -mpi
• if it is finished without a severe error, do make depend
• if it is finished without a severe error, do make
• if you have mitgcmuv, then the compliation is successful

We first try to run the model with the default setting.

• go to run directory
• copy all input files to here: cp ../input/* .
• create a sympolic link of the input files to here: ./prepare_run
• create a sympolic link of the executable file to here: ln -s ../build/mitgcmuv .

Now, let’s run the model!

• execute the run: mpirun -np 4 ./mitgcmuv

See the example jupyter lab file.

We can measure the age of the water mass. The age represents how long it has passed since a particular water mass was at the surface. This shows the ocean circulation: you may find “young” water mass at depth where there is sinking and “old” water mass where there is upwelling.

To do this, we need to activate ptracers package that allows us to add passive tracers in the simulation.

• go to code directory: cd code_age
• add a line in packages.conf file: echo 'ptracers' >> packages.conf
• copy a couple of files that compute the age of water mass: cp ../../tutorial_global_oce_latlon/code/ptracers_* .
• then, let’s compile the code in the build directory:
  • go to build directory
  • do make CLEAN
  • compile the code again as in compile
• go to the run directory
• copy an input parameter file for a tracers: cp ../../tutorial_global_oce_latlon/input/data.ptracers .
• activate ptracer package: add usePTRACERS=.TRUE., in data.pkg
• we need to tell the model that the initial time step for tracer is 72000: add PTRACERS_Iter0=72000, in data.ptracers
• run the model as in model integration

This example provides a couple of variations from the default setting. So, let’s try the one with sea ice using the settingin input.seaice.

• create the run directory: mkdir run.seaice
• go to this directory: cd run.seaice
• copy all input files to here: cp ../input/* .
• create a sympolic link of the input files to here: ./prepare_run
• Then, overwrite some of the input files with those related to sea ice experiment: cp ../input.seaice/* .
• create a sympolic link of the input files associated with this sea ice experiment: ./prepare_run
• create a sympolic link of the executable file to here: ln -s ../build/mitgcmuv .
• run the model as in model integration