CO9 integration in Met Office systems overview

[hadjt]

DA

James almost has the DA working (at least for one day), and has added TEOS10. He has come across a number of model minor bugs, that he has referred to the Modelling group. Next step He hopes to tidy it up and recommit his changes. James to tell Juan when recommitted so he can use the TEOS10 code Then, a long run for optimisation, and calculate the error covariances. Might await bug fix from modelling group though.

Waves

New WW3 grid produced. WW3 dev to run with WAD done. One day tests don’t blow up. Next steps Run for a few months with SSH’s from AMM15 CO9p2. Juan was to run the model to provide these, but since the meeting, realised James can provide these directly, so Juan does not need to do this run now.

REP

Juan has taken the suite form James, and added it to the RCS. He has updated to CO9p2, and compared to James’ suites 2 day comparison run.. similar, but not identical, but that’s to be expected. James to pass TEOS by end of week, so will add. Next steps Coupling waves and NEMO. Could be simple, but likely to have bugs, Already thought about dry points on the waves vs dry points in NEMO… Oasis flag can handle this. Need to recreate LBCs and ICs for TEOS10 (no change to SBC)

[mo-mattmartin]

Update on progress.

DA:

• DA experiments have run for a few months comparing CO9 with CO8 (both with DA). Similar statistics (compared to obs) for most variables. Perhaps small improvements in SST for CO9. Increase bias in SLA (SSH) and deeper temperatures being investigated, but perhaps due to initial conditions and/or need for longer spin up of CO9.
• Animations produced to zoom in on regions (German Bight mainly so far) with W&D but don't seem to show any adverse interactions between DA and W&D
• Updates to the FOAM suite are now going for review by Hedong.

REP:

• Juan tried running ocean CO9 with wave physics in uncoupled mode, using as driving data the 2-month long wave CO9 job he completed.
• The job crashes because of the boundaries, he is waiting for James to send him the final ocean configuration to try again.

[mo-mattmartin]

Update on progress. DA:

• The main updates to get DA working in CO9 are in the trunk of the FOAM R&D suite.
• James spent a long time trying to convert and interpolate a CO8 restart to use for CO9 to be able to do a fair comparison. Everything he tried led to a blow-up in CO9.
• He has been trying to understand the cause of the SLA biases. There is a large bias at the Baltic boundary and he is calculating an offset between the LBC and the MDT+SLA data being used. There is also a bias around the coasts which James is now investigating.

[jpolton]

Update on progress: NOC Dev:

0) Resolve issue with Wetting and Drying that impacted 4.2 but strangely not 4.0.4 1) As a candidate for CO10:

• Development of AMM15 at v4.2 (https://github.com/JMMP-Group/CO_AMM15/releases/tag/v10.0.0 and on branch https://github.com/JMMP-Group/CO_AMM15/tree/CO10 ). Explorations with lateral diffusion. Improved rivers. SarahW/RyanP
• Development of AMM7 at v4.2 to align with AMM15 (https://github.com/JMMP-Group/CO_AMM7/tree/CO9_AMM7_v4.2) AnthonyW:

• PRJ --> DJC
• change of equation of state from EOS80 to TEOS10 (conservative temperature and absolute salinity must now be supplied)
• Currently the surface bulk formulation is set to ECMWF (consisitent with ERA5 forcing) rather than COARE.
• The MY_SRC files have been harmonised with AMM15.

2) Momentum Trend updates - porting to 4.2, into NEMO gitlab, for release. RyanP/AnthonyW

3) Implementation of P2 for CHAMFER application (https://github.com/NOC-MSM/CO_AMM15_CHAMFER) - RyanP:

• GLOSEA6 lateral forcing for both the atlantic and baltic boundary conditions
  • MY_SRC includes Momentum Trend capability with Anthony Wise's atmospheric pressure additions

4) Write up of P2. 10 years runs of ERA5 forced CO7 and CO9. Slow. RyanP

NOC challenges/pending: 1) Haven't been able to verify new harmonic analysis code at 4.2+ 2) Haven't been able to test NEMOv5 for shelf seas 3) Haven't started to look at new AMM15 data from Jon