Differences between CSIRO MOM5 and harmonised MOM5 outputs

mom-ocean / MOM5

The Modular Ocean Model

https://mom-ocean.github.io/

GNU Lesser General Public License v3.0

82 stars 95 forks source link

Differences between CSIRO MOM5 and harmonised MOM5 outputs #258

Open peterdobro opened 5 years ago

peterdobro commented 5 years ago

We are seeing, differences in the ocean outputs between the existing CSIRO MOM5 code and the new harmonised MOM5 in ACCESS-CM2 tests. For example, the potential temperature differs regionally by several degrees in the first month.

Harmonised MOM5 outputs at /short/p66/pbd562/archive/av630/history/ocn can be compared with the existing CM2 tests with CSIRO MOM5 at /g/data1a/p66/dhb599/ACCESS-CM2/archive/av630/history/ocn/

The corresponding namelists can be found in /short/p66/pbd562/cylc-run/u-av630/work (new harmonised) and /short/p66/dhb599/cylc-run/u-av630/work (existing CSIRO).

I will be away until Friday next week, 16 Nov 2018, but please contact Dave or Arnold if you need read permissions on anything. Data are readable by p66.

Thanks, Peter

russfiedler commented 5 years ago

@peterdobro You need to set do_langmuir_cvmix=.true. in the kpp namelist. It's using the Smythe version instead.

peterdobro commented 5 years ago

Thanks @russfiedler - I will give it a go.

peterdobro commented 5 years ago

Hi @russfiedler, there is a small improvement with this setting but still seeing several degrees difference regionally in monthly average potential temperature after one month:

This is /short/p66/pbd562/archive/av630.v3/history/ocn/ocean_month.nc-02010630 minus /g/data1a/p66/dhb599/ACCESS-CM2/archive/av630/history/ocn/ocean_month.nc-02010630

aidanheerdegen commented 5 years ago

The verification plots etc for the code changes were done in the pull request:

https://github.com/mom-ocean/MOM5/pull/229

I note that the colour scale is not entered on zeros, so the large changes are all two patches off the coast of Greenland, which is where some of the greatest effects did show up in the plots from https://github.com/mom-ocean/MOM5/pull/229.

Which is the original run and which is the one with the new code? I guess av630.v3 is harmonised and av630 is not? So the anomalies as plotted show a warming in those two patches off Greenland?

peterdobro commented 5 years ago

Yes that's right: av630.v3 is harmonised and av630 is not. And the namelists can be found in /short/p66/pbd562/cylc-run/u-av630.v3/work/... and /short/p66/dhb599/cylc-run/u-av630/work/... respectively.

ofa001 commented 5 years ago

Looking at the Greenland anomaly region I would have thought the sea ice was playing a role amplifying the change, its January and at 5m near the surface but the SO has widespread anomalies well away from the ice edge so it ahs to be something more.