Open Boundary Conditions Roms 3.8

[mainara]

Hello,

I am running ROMS 3.8 for Southwestern Atlantic, and I am having trouble with my boundaries. I have tried different settings, but I always get an odd gradient at the borders, as you can see in the figures attached.

My cppdefs.h is as follows:

Code: Select all

define SOLVE3D
# define DJ_GRADPS 
define UV_ADV            
#define UV_COR 
# define TS_U3HADVECTION 
# define TS_C4VADVECTION  
#define UV_VIS2
#define VISC_3DCOEF     
#define MIX_S_UV        
#define VISC_GRID        
#define SPONGE 
# define TS_DIF2         
# define MIX_GEO_TS      
# define DIFF_GRID
# define SOLAR_SOURCE
# define LMD_MIXING  
#  define LMD_NONLOCAL   
#  define LMD_SHAPIRO  
# define BULK_FLUXES 
# define UV_QDRAG
# define RADIATION_2D
# define ANA_BSFLUX     
# define ANA_BTFLUX 

My boundary conditions are set to:

Code: Select all

LBC(isFsur) ==   Clo     Cha     Cha        Clo    
LBC(isUbar) ==   Clo     Fla     Fla        Clo    
LBC(isVbar) ==   Clo     Fla     Fla        Clo 
LBC(isUvel) ==   Clo     RadNud  RadNud     Clo 	
LBC(isVvel) ==   Clo     RadNud  RadNud     Clo 
LBC(isMtke) ==   Clo     RadNud  RadNud     Clo 
LBC(isTvar) ==   Clo     RadNud  RadNud     Clo  	
                 Clo     RadNud  RadNud     Clo  	

Any help will be appreciated!

Thank you in advance,

Mainara

[kate]

This is getting to be an FAQ - anyone with good advice to write up on the wiki?

For the 2-D flow, the current recommendations are Che and Shc instead of Cha and Fla, but I don't think that will make any difference here.

John Wilkin advised someone to make sure the tides were on to help move things through the open boundaries, but it looks like you weren't trying for tides.

I assume you have looked at the boundary files you are providing and that they are sensible. I have better luck with model fields on a relatively fine timescale (say daily HYCOM results) compared to monthly climatologies. Perhaps you have tried different nudging timescales?

A final desperate measure would be to add a nudging to say monthly climatology fields in a band along the open boundaries.

[mainara]

Hi Kate,

Thank you for your quick response.
I've tested the boundary conditions Shc and Che, as you suggested, but did not solve my problem.
I am running my simulations without the tides. And I have done several simulations with different NUDING values, but it did not solve my problem with the edges.
I have also tested other boundary conditions with different scales for NUDING, but the edges always continue with the same problem.
I decided to test your suggestion in version 3.7 and the edges improved when compared to the same simulation with version 3.8(attached).
What has changed from one version to another that may be interfering with these rims?
I also have another problem. In this area there are two upwelling centers that can not be observed in my simulations, in any of the ROMS versions. Maybe it could have to do with the parametrizations which affect the interaction with atmosphere?

Thank you!

Best regards,

Mainara