[MITgcm-support] 回复： GMRedi parameterization

[钱钰坤]

Hi Ryan,


Thanks very much for your clear reply.


Redi and GM is now combined together (as GMRedi package) and leave two diffusivities GM_isopycK and GM_background_K which parameterize respectively the diffusive and advective effects of unresolved geostrophic eddies.


For an eddying run, there are also eddy flux (probably at submesoscale or even smaller) that need to be parameterized.  So both diffusive and advective effects of these small eddies also exist in such situation.  However, the advective effect may not act to flatten isopycnals so that GM's scheme does not apply in eddying case.  In contrast, diffusive effect should be also along isopycnals so that we still need Redi scheme.


In short, in eddying run, we need Redi (GM_isopycK) but not GM (GM_background_K=0).  Is it right?

------------------
 Best regards 
 
Yu-Kun Qian (钱钰坤) 
Center for Monsoon and Environment Research 
 Department of Atmospheric Sciences
School of Environmental Science and Engineering 
 Sun Yat-sen University 
No. 135 Xingang West Road, Haizhu District 
Guangzhou, 510275, P.R. China 
Tel; 020-84115227 
Email: qianyk at mail3.sysu.edu.cn      


 
 
 
------------------ 原始邮件 ------------------
发件de人: "Ryan Abernathey"<ryan.abernathey at gmail.com>;
发送时间: 2015年8月26日(星期三) 晚上8:43
收件人: "mitgcm-support at mitgcm.org"<mitgcm-support at mitgcm.org>; 

主题: Re: [MITgcm-support] GMRedi parameterization

 
I'm not sure about the best combination of diffusivities defined between data and data.gmredi files.


Don't worry, no one is. First, make sure you have read the manual:
http://mitgcm.org/public/r2_manual/latest/online_documents/node240.html

 
For climate simulation, in z-coordinate model (such as MITgcm), diffusion should be primarily along isopycnal rather than horizontal.  Therefore, I guess diffusivities defined in data file (e.g., diffKh and diffKr) should be set to zero to prevent spurious diapycnal mixing.


You are correct about diffKh. However diffKr is the the vertical diffusivity, which is effectively a diapycnal diffusivity. So you put your desired diapycnal diffusivity into diffKr.
 
GMRedi package is used in MITgcm to diffuse tracer in local isopycnal coordinates through some kind of rotation.  However, I cannot find definitions of isopycnal and diapycnal diffusivities in data.gmredi file.  In the test case of tutorial_global_oce_latlon, there is a GM_isopycK, which is commented, that may be isopycnal diffusivity but no diapycnal definitions (or even a ratio between them).  There is also a GM_background_K in data.gmredi which I don't understand why we need it.


You are correct that GM_isopycK is the isopycnal (Redi) diffusivity. GM_background_K is the "Gent McWilliams coefficient" (K_GM in the documentation), which regulates the strength of the advective part of the eddy parameterization. It is the single most important parameter in the package.
 
Besides, does the GMRedi scheme apply to the viscosities for momentum?


No. There are other options for viscosity though.
http://mitgcm.org/public/r2_manual/latest/online_documents/node86.html



GM sheme is proposed to parameterize effects of geostrophic eddies in non-eddy resolving model.  If the model is eddy resolving, do we still need to include this scheme?


If the model is truly eddy resolving, you should not use GM. If the model is "eddy permitting," it is a tough call. You should read this paper by Bob Hallberg.
http://www.sciencedirect.com/science/article/pii/S1463500313001601

Unfortunately MITgcm does not implement any of the fixes proposed by Hallberg.
 



------------------
 Best regards 
 
Yu-Kun Qian (钱钰坤) 
Center for Monsoon and Environment Research 
 Department of Atmospheric Sciences
School of Environmental Science and Engineering 
 Sun Yat-sen University 
No. 135 Xingang West Road, Haizhu District 
Guangzhou, 510275, P.R. China 
Tel; 020-84115227 
Email: qianyk at mail3.sysu.edu.cn      


 


_______________________________________________
 MITgcm-support mailing list
 MITgcm-support at mitgcm.org
 http://mitgcm.org/mailman/listinfo/mitgcm-support
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mitgcm.org/pipermail/mitgcm-support/attachments/20150827/4d527117/attachment-0001.htm>