Diffusion coefficients

General scientific issues regarding ROMS

Moderators: arango, robertson

Post Reply
Message
Author
chaos4u2

Diffusion coefficients

#1 Unread post by chaos4u2 »

I've looking for how does ROMS to calculate the diffusion coefficients. For the vertical one I've read in https://www.myroms.org/wiki/index.php/V ... erizations that there are 3 methods: KPP, GLS and Mellor-Yamada.

How can I choose in ROMS which method I want to use and which one is best? I've haven't seen anything in the input file.

For the horizontal, there's also this: https://www.myroms.org/wiki/index.php/Horizontal_Mixing; but it doesn't say how to calculate the horizontal diffusion coeficient. How does ROMS it?

The third question, is that I want in my results: the horizontal and vertical diffusion coefficients. I suppose that I must activate the switch in the input file for these 2 parameters. I write down in the history and average part:
Hout(idKhor)==T
Hout(idKver)==T

But after compiling and executing, my netcdf file of results, the Khor and Kver are not there. Am I missing anything?

Thank you in advance.

User avatar
kate
Posts: 4091
Joined: Wed Jul 02, 2003 5:29 pm
Location: CFOS/UAF, USA

Re: Diffusion coefficients

#2 Unread post by kate »

This is one of the choices that get made in your cppdefs.h file. Look through files in ROMS/Include for examples. Which to use? It depends... I believe the ROMS community feels that the Mellor-Yamada within GLS is coded better than the stand-alone one. I usually use KPP while others usually use GLS. Try them and see. Some GLS options behave better than others in my experience.
For the horizontal, there's also this: https://www.myroms.org/wiki/index.php/H ... al_Mixing; but it doesn't say how to calculate the horizontal diffusion coeficient. How does ROMS it?
ROMS reads the value from the input file. You have to provide that value based on the guidelines at the bottom of that link.
The third question, is that I want in my results: the horizontal and vertical diffusion coefficients.
The ones you list are:

Code: Select all

        integer  :: idKhor        ! convolution horizontal diffusion
        integer  :: idKver        ! convolution vertical diffusion
You instead want:

Code: Select all

        integer  :: idSdif        ! vertical S-diffusion coefficient
        integer  :: idTdif        ! vertical T-diffusion coefficient
        integer  :: idVvis        ! vertical viscosity coefficient
I don't know that it ever writes the horizontal values.

chaos4u2

Re: Diffusion coefficients

#3 Unread post by chaos4u2 »

Thank you very much Kate!! You're the best, fast and clear. You really solved my problem!

Post Reply