Which GLS scheme is the best to model wind on surface water?

[antoinemarie]

Hi ROMS users,

I have a question about the turbulence closure models. I try to model a small stratified basin (approximatly 1km long with 5m deep), and I want to observe the impact of wind on the mixture of the basin.
I try all GLS closur models.
With k-kl, k-epsilon and k-omega model, the water of basin is mixed without wind, and my stratification disappear, whereas my initial conditions are stable (hot water in the upper layer).
With the generic model, the stratification doesn't disappear, but in the article of JC WARNER about the GLS turbulence closure parameters, I see that the k-kl model should be better??

I want to know if there is a way to use the k-kl model without losing my stratification, and if the generic model is correct in my application?

I hope I am clear and you understand my questions.

Thanks in advance,

Antoine

[staalstrom]

Hi

I'm not an experienced turbulence closeure modeller.
I have always used k-kl, but I have plans to try the others.
I work in a fjord in Norway, the Oslofjord, with max depth of 200m and sill depth of 20m.
It is clear that what kind of topography and vertical coordinates I use is important for the mixing.

It think, but is not sure, that too steep s-surfaces crossing your thermocline will cause too much mixing.
I fixed a little on the vertical transform and stretching, so that I get almost horizontal s-surfaces in the pycnocline (at a fixed depth, 20m).
This helps a lot, because the s-coordinates mimic z-coordinates where the stratification is strongest.
Do your bathymetry vary a lot? What is your minimum depth, and is there any sills?

Maybe it helps to increase the number of layers. Try many more layers.
The only advantage of using terrain following vertical coordinates compared to z-coordinates is that you can have more layers.

Best regards
Andre Staalstrom

[antoinemarie]

Thank you Andre for your reply,

I try to increase the number of layers, but it doesn't work. The maximum deep of my bathymetry is equal to 8 meter, it doesn't vary a lot, and it is a basin, so it is surrounded by seep.
The problem with the k-kl model is that my vertical mixing coefficient for momentum AKv increase with the time despite of the absence of wind, and with a stable stratification!?!?

Eventually, I choose to use the generic model wich is the most relevant in my case, in fact, parameters AKt and AKv are closed to 0 without wind.

Best regards,

Antoine

[carniel]

Dear all,

the "correct" modeling of the upper surface layers is a delicate issues, involving many aspects (including the choice of the TCM).
Some (hopefully) useful hints in this direction may be gained from the paper in attachment.

Cheers

Sandro