GOSI11: Vertical mixing, OSMOSIS

oceanaguiar commented 7 months ago

Test simplified OSMOSIS boundary layer mixing scheme (and Fox-Kemper) to replace TKE + nn_etau. AG, GN, CG for testing.
Documentation paper for simplified OSMOSIS scheme. AG
Ongoing work to improve full OSMOSIS scheme. AG, GN
Improvement of interior/abyssal mixing – incorporation of de Lavergne et al IWM (instead of KPP) scheme to replace Simmons et al scheme. Test again when OSMOSIS ready. Coordinate with SE-NEMO. Adam has implemented the de Lavergne scheme in his NEMO v4.2 runs under the Near Present Day project, but I don’t think anyone has compared it directly with the Simmons et al. scheme.

ukmo-cguiavarch commented 5 months ago

eORCA1 GOSI10p0 with OSMOSIS: eORCA1 configuration based on GOSI9 branch at 4.2.1 https://forge.nemo-ocean.eu/devs/ukmo/gosi/gosi/-/tree/2aa791c7a0ca1a361e5660aca4bd9f2c7e56b8b8 and the OSMOSIS branch at 4.2.1 https://forge.nemo-ocean.eu/nemo/nemo/-/tree/281-update-osmosis-branch. Note that this configuration is based on NEMO 4.2.1 but includes the implicit sea ice drag bug fix introduced at 4.2.2

30y integrations from GOSI9 1 year restart (GOSI9 froced by CORE, initialised from EN4).

u-db619: GOSI10p0 TKE forced with COREII (equivalent to GOSI9)
u-db902: GOSI10p0 TKE forced with ERA5
u-dc107: GOSI10p0 OSMOSIS forced with ERA5
u-dc244: GOSI10p0 OSMOSIS including waves forced with ERA5

Validation notes for the last decade of the integrations (y20-30):

OSMOSIS vs. TKE (u-dc107 vs. u-db902): https://gws-access.jasmin.ac.uk/public/jmmp/valid_ocean/u-dc107_nemo_vs_u-db902_nemo/assess.html

OSMOSIS with waves vs. TKE (u-dc244 vs. u-db902): https://gws-access.jasmin.ac.uk/public/jmmp/valid_ocean/u-dc244_nemo_vs_u-db902_nemo/assess.html

OSMOSIS with waves vs OSMOSIS (u-dc244 vs. u-dc107): https://gws-access.jasmin.ac.uk/public/jmmp/valid_ocean/u-dc244_nemo_vs_u-dc107_nemo/assess.html

ERA5 vs. COREII (u-db902 vs. u-db619 ): https://gws-access.jasmin.ac.uk/public/jmmp/valid_ocean/u-db902_nemo_vs_u-db619_nemo/assess.html

ukmo-cguiavarch commented 4 months ago

eORCA025 GOSI10p0 with OSMOSIS: eORCA025 configuration based on GOSI9 branch at 4.2.1 https://forge.nemo-ocean.eu/devs/ukmo/gosi/gosi/-/tree/2aa791c7a0ca1a361e5660aca4bd9f2c7e56b8b8 and the OSMOSIS branch at 4.2.1 https://forge.nemo-ocean.eu/nemo/nemo/-/tree/281-update-osmosis-branch. Note that this configuration is based on NEMO 4.2.1 but includes the implicit sea ice drag bug fix introduced at 4.2.2.

The OSMOSIS eORCA025 integration was unstable (Hollingsworth instability near the coast of Borneo), to be able to run with OSMOSIS we change the momentum advection from vector form to flux form (UBS).

_Note:The impact of the UBS scheme has been assessed with the TKE scheme but further investigation will be needed if we want to use UBS as standard (also not that UBS is removed at NEMO5)_

30y integrations from GOSI9 1 year restart (GOSI9 froced by CORE, initialised from EN4).

u-dc743: GOSI10p0 TKE vector form forced with ERA5
u-dc819: GOSI10p0 TKE flux form (UBS) forced with ERA5
u-dc811: GOSI10p0 OSMOSIS forced with ERA5

Validation notes for the last decade of the integrations (y20-30):

OSMOSIS vs. TKE (u-dc811 vs. u-dc819): https://gws-access.jasmin.ac.uk/public/jmmp/valid_ocean/u-dc811_nemo_vs_u-dc819_nemo/assess.html

TKE UBS vs. TKE VEC (u-dc819 vs. u-dc743): https://gws-access.jasmin.ac.uk/public/jmmp/valid_ocean/u-dc819_nemo_vs_u-dc743_nemo/assess.html

ukmo-cguiavarch commented 3 months ago

I have rerun the last 10y of eORCA025 OSMOSIS and TKE (u-dc811 and udc819) with the extra diagnostics. All available from MASS.

OSMOSIS: moo:/crum/u-dc811
TKE: moo:/crum/u-dc819

ukmo-cguiavarch commented 3 months ago

Impact of Fox-Kemper / OMOSIS interaction:

u-de992: 10y integration with interaction turned off (ln_osm_mle=false ) but Fox-Kemper still activated (ln_mle=true). The integration was started from u-dc811 20y restart (19951201).
u-dc811: integration with OSMOSIS and FK interaction (30y: 1976-2005)

Validation notes for the last 5y of the integrations (2000-2005): OSMOSIS with FK interaction vs. OSMOSIS: https://gws-access.jasmin.ac.uk/public/jmmp/valid_ocean/u-dc811_nemo_vs_u-de992_nemo/assess.html

With FK interaction: shallowing of the MLD in winter, warming below the subsurface at high latitude. Impact on SST localised in the Labrador Sea in winter/spring.

Update: I have now added the validation for the last 10 years of integration (1995-2005:) OSMOSIS with FK interaction vs. OSMOSIS: https://gws-access.jasmin.ac.uk/public/jmmp/valid_ocean/u-dc811_nemo_vs_u-de992_nemo_10y/assess.html

ukmo-cguiavarch commented 2 months ago

Extending OSMOSIS and TKE runs

u-dc811 (OSMOSIS) and u-dc819 (TKE control) extended to 2015 (1976-2015).
validation notes performed for 5 years periods from 1995 to 2015.

1995-2000: https://gws-access.jasmin.ac.uk/public/jmmp/valid_ocean/u-dc811_nemo_vs_u-dc819_nemo_1995-2000/assess.html 2000:2005: https://gws-access.jasmin.ac.uk/public/jmmp/valid_ocean/u-dc811_nemo_vs_u-dc819_nemo_2000-2005/assess.html 2005-2010: https://gws-access.jasmin.ac.uk/public/jmmp/valid_ocean/u-dc811_nemo_vs_u-dc819_nemo_2005-2010/assess.html 2010:2015: https://gws-access.jasmin.ac.uk/public/jmmp/valid_ocean/u-dc811_nemo_vs_u-dc819_nemo_2010-2015/assess.html

JMMP-Group / GO_coordination

GOSI11: Vertical mixing, OSMOSIS #7