Opened 3 years ago

Closed 3 years ago

#1963 closed Defect (fixed)

CMIP6 / OMIP diagnostics of KE lateral/vertical dissipation and EKE to PE conversion when using EIV

Reported by: gm Owned by:
Priority: normal Milestone:
Component: OCE Version: release-3.6
Severity: minor Keywords: CMIP6 diagnostics
Cc:

Description

In OMIP/ CMIP6 diag., the dissipation of Kinetic Energy due to horizontal and vertical mixing is demanded.

It is not possible to simply use the build-in KE-trend diagnostics to compute them.

Indeed, in KE-trend diagnostics we simply evaluate Un*U-mixing trend. Such a term represent the combination of both diffusive and dissipation trends of KE (see NEMO doc appendix C). In case of a laplacian mixing with a coefficient K, the dissipation trends of KE is -K*[grad(U)]2, NOT u.div[K*grad(U)]

In NEMO, I modify the following modules:

M NEMO/OPA_SRC/DYN/dynldf_bilap.F90
M NEMO/OPA_SRC/DYN/dynldf_lap.F90
M NEMO/OPA_SRC/DYN/dynzdf_imp.F90

so that they output the vertical sum of KE dissipation trend as demanded by OMIP. There is no need to activate the KE-trend diag (ln_KE_trd=.true.) to obtains theses fields (which slightly speed up the code)

In addition, I also correct a misplaced mixing coefficient in Smagorinski case in dynldf_bilap.F90 (an error which concerns only the v3.6, not the trunk)

A second point concerns the conversion of eddy kinetic energy (EKE) into potential energy due to EIV parameterization (Gent & Mc Williams? param)

No error here, but misplaced diagnostics. We don't need anymore to activated ln_KE_trd=T to obtain this diag.

I also change the name as it is a conversion from EKE to PE, not form KE…

To do that I modify the following modules:

M NEMO/OPA_SRC/TRA/traadv_eiv.F90
M NEMO/OPA_SRC/TRD/trd_oce.F90
M NEMO/OPA_SRC/TRD/trdken.F90

I introduce new variables in field_def.xml file which output the required fields:

      <!-- variables always available -->
      <field id="dispkexyfo"    long_name="KE-trend: lateral  mixing induced dissipation"   standard_name="ocean_kinetic_energy_dissipation_per_unit_area_due_to_xy_friction"                   unit="W/m^2" grid_ref="grid_T_2D" />   
      <field id="dispkevfo"     long_name="KE-trend: vertical mixing induced dissipation"   standard_name="ocean_kinetic_energy_dissipation_per_unit_area_due_to_vertical_friction"             unit="W/m^2" grid_ref="grid_T_2D" />   
      <!-- variables available with key_traadv_eiv and key_diaeiv -->
      <field id="eketrd_eiv"    long_name="EKE-trend due to parameterized eddy advection"   standard_name="tendency_of_ocean_eddy_kinetic_energy_content_due_to_parameterized_eddy_advection"   unit="W/m^2" grid_ref="grid_T_2D" />   

NB this also concerns the trunk version

I will update the v3.6_stable version tomorrow (on Monday October the 16th) as well as the trunk.

Gurvan

Commit History (1)

ChangesetAuthorTimeChangeLog
8627gm2017-10-16T16:19:11+02:00

#1963 : Correct implementation of CMIP6 KE dissipation + move EKE⇔PE diag in traadv_eiv (in both cases, no more need of ln_trd_KE=T)

Change History (1)

comment:1 Changed 3 years ago by gm

  • Resolution set to fixed
  • Status changed from new to closed

committed in revision 8627 (see changes here)

The correction will not be done in the trunk version but in a development branch, dev_r7881_ENHANCE09_RK3, which will be used to create the new trunk by the end of the year.

Note: See TracTickets for help on using tickets.