Issue with the vertical mixing

[sprasad]

Hi all,

We are trying to set up  test case for ROMS to understand the vertical mixing process. The basin has a flat bottom with maximum depth 30m, walls on 4 sides of the domain (100x100x60 grid). We are providing analytical grid and initial conditions. The initial condition at T=0, for the temperature is shown in the attaced figure (Initial condition.png), all other variables are set to zero. . The forcings are provided hourly: shortwave radiation (W/m^2), downwelling long wave radiation (W/m^2), u-wind component is zero, v-wind component is a sinusoidal in the y axis (m/s), air temperature (deg C) and relative humidity(%), air pressure (mb), rain = 0, surface air pressure = 1035.5 mb. The simulation was carried out for 15 days. The result on the15th day, T=15 is shown in the attached figure (Day 15.png).
It seems as if no vertical momentum transfer and vertical mixing occurs. The following are the CPP options that were selected for ROMS. I will appreciate it if anyone give some light on what could have gone wrong.

/** Memory Management **/
# define TOY_MODEL

# define INLINE_2DIO
# undef SPHERICAL
# undef CURVGRID
# define ANA_GRID

/* Model Configuration Options */
# define MASKING
# ifdef MASKING
# define ANA_MASK
# endif
# define ANA_INITIAL

# define SOLVE3D

/* Pressure Gradient Algorithm */
#define DJ_GRADPS

/* Momentum Equation Options */
# define UV_ADV
# define UV_COR
# define UV_QDRAG

/* Momentum advection */
/** The default horizontal advection is 3rd-order upstream bias for **
* ** 3D momentum and 4th-order centered for 2D momentum. The default **
* ** vertical advection is 4th-order centered for 3D momentum. If this **
* ** is the case, no flags for momentum advection need to be activated. ** */

/* Momemtum horizontal mixing */
# define MIX_S_UV
# define UV_QDRAG

/* Tracer Equation Options */
# define SALINITY
# define NONLIN_EOS


/* Tracer advection */
# define TS_A4HADVECTION
# define TS_A4VADVECTION

/* Tracer horizontal mixing */
# define MIX_GEO_TS

# define BULK_FLUXES
# ifdef BULK_FLUXES
# define ANA_RAIN
# define ANA_PAIR
## define SOLAR_SOURCE
# define LONGWAVE_OUT
# define EMINUSP
# endif

/* turbulence mixing scheme */
# define LMD_MIXING

# ifdef LMD_MIXING
# define LMD_RIMIX
# define LMD_CONVEC
# define LMD_SKPP
# define LMD_NONLOCAL
# define RI_HORAVG
# define RI_VERAVG
# define LMD_SHAPIRO
# endif

# ifdef SALINITY
# define ANA_SSS
# endif

/* bottom surface ANA Flux */
# define ANA_BTFLUX
# define ANA_BSFLUX


Thanks
Siva

[jcwarner]

this is a power point i gave in 2002:

the text is here:
Warner, J.C., Sherwood, C., Arango, H., and Signell, R. (2005). “Performance of Four Turbulence Closure Models Implemented Using a Generic Length Scale Method.” Ocean Modelling, v. 8/1-2, p. 81-113.

check your surface stress. we did not use LMD, but were testing GLS back in those days. if you do double periodic, there will be no horiz flux divergence and thus W=0 so no vertical advection. but you can get vertical mixing with the surface stress. make the depth deep enough so the bottom does not matter.

[sonaljit.m]

The region between red and blue - is that the transition layer? The red area seems to be a mixed layer. At day 15 your plot shows minute undulations in the transition layer, which indicates that there is some activity happening. I'd suggest look into the vertical profiles of N^2 and shear^2 to get a better idea on how they are evolving.

- If your mixed layer deepens slightly, that should indicate mixing is happening (Ri > 0.25 can be considered a mixed layer boundary).
- If your wind forcing is not strong enough, mixing could be very little.
- If your initial stratification is strong, the downwelling flux may not be strong enough to generate convective mixing.

I'd suggest the best way to make sure the model is working properly, is to try with 0 heat fluxes and a basic constant wind-forcing (around 1 N/m^2) and an initial profile with a constant stratification throughout the depth, N^2 = 1.e-04. Run this config for over a week. This should create a prominent mixed layer from the surface, and that should be your proof that the model is working good.

[sprasad]

Thank you all, for the tips to correct the issue. The problem was due to the low wind stress and is resolved now.

Thanks again
Siva