Opened 9 years ago

Closed 9 years ago

Last modified 4 years ago

#989 closed Task (fixed)

Create a branch for Smagorinsky model

Reported by: hliu Owned by: Maria Luneva
Priority: low Milestone:
Component: OCE Version: release-3.4
Severity: Keywords: Smagorinsky mixing
Cc: Branch review:
MP ready?: Task progress:


Create a branch for the implementation of Smagorinsky turbulence model based on r3452

Commit History (0)

(No commits)

Change History (7)

comment:1 Changed 9 years ago by hliu

  • Keywords Smagorinsky lateral mixing added

Smagorinsky type diffusivity/viscosity for lateral mixing has been introduced.
When key_traldf_smag or key_dynldf_smag are defined, Smagorinsky 3D diffusivity or viscosity are calculated. These keys should be used together with key_traldf_c3d or key_dynldf_c3d respectively. Griffies and Hallberg, 2000 diffusivity-viscosity formulation has been used with the following modifications:
a) Coded stability criteria are twice strict in comparison with derived from linear stability theory to provide positiveness of solutions;
b) Numerical test demonstrated that even these stability criteria were too weak, as in this case equations can be strongly nonlinear and linear theory does not work. Additional constrains has been added (maximum value of diffusivity/viscosity), defined by rn_aht_m, rn_ahm_m_blp ,rn_ahm_m_lap
c) It has been found, that when tides are present, the input of tangential derivatives of velocity is too strong , as tidal velocity is roughly inversely proportional to the depth. Option, when only shear strain is used for calculations available : set rn_cmsh=0 , rn_smsh=0 in namelist.

  1. Bilaplacian operator with Smagorinsky diffusion for tracer is unstable or non-conservative and has been removed from code.
  2. For momentum the order of operator in Smagorinsky case has been changed
    • delta(ahm*delta(u)) —> - ahm*delta(delta(u))
  3. Use much smaller constants rn_chsmag, rn_cmsmag_1,_2 in comparison with recommended rn_chsmag=2.2 , rn_cmsmag_1,_2=3-4 for eddy-resolving resolving simulations (0.7 : 1).

Code needs a small modification to introduce new variable in namelist - background diffusivity/viscosity. Now rn_aht_0, rn_ahm_0_lap/bilap are background values, which can be confusing in other cases

comment:2 Changed 9 years ago by acc

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

Branch included into 2012 merge ready for v3.5 release. Documentation still in preparation but closing ticket to discourage further changes on this branch.

comment:3 Changed 5 years ago by nicolasmartin

  • Milestone 2012 Stream 3: New features deleted

Milestone 2012 Stream 3: New features deleted

comment:4 Changed 5 years ago by nicolasmartin

  • Keywords lateral_mixing added; lateral removed

comment:5 Changed 5 years ago by nicolasmartin

  • Keywords mixing removed

comment:6 Changed 4 years ago by nemo

  • Keywords mixing added; lateral_mixing removed

comment:7 Changed 4 years ago by nemo

  • Type changed from Development to Task

Remove 'Development' type

Note: See TracTickets for help on using tickets.