Modelling over complex terrain at high altitude

[cross]

I want to set up a model over a complex terrain at high altitude (between 3000m and 5000m). The main problem here is that sometimes the surface wind field is forced by the model to make a wind direction daily cycle, instead of one steady direction (mainly from the north).

So, I most of the time use the YSU PBL scheme with wind_topo enabled. Is this ok?, considering the complex terrain and high altitude, or the wind_topo option would overestimate the wind speed?

Also, I’m thinking of trying the Shin-Hong PBL or the SMS-3dTKE, do you think this will improve the results? Thinking the main focus is on surface wind, mostly over the top of the mountains.

The resolution of the domain will be 9km, 3km and 1 km, and maybe 0.5 km.

I will be doing some tests, but I’d like to know your opinion or know if there are any other parameter to change that I might be missing.

 

[Ming Chen]

For a nested case with resolution of 9-3-1-0.5km, I would recommend using the SMS-3DTKE scheme, which is scale-aware and can better handle PBL physics across grids of different resolution.
topo_wind might yield better simulation of surface winds. However, it can only be used with YSU PBL scheme.
I would suggest you try various options and determine which options can yield better results. These schemes are often case-dependent, and it is hard to tell which options are better for a specific case.

 

[cross]

Thanks! that's what I was thinking about the SMS-3DTKE scheme.
I'll do some tests to check the different options.

 

[Sarath Guttikunda]

Thanks! that's what I was thinking about the SMS-3DTKE scheme.
I'll do some tests to check the different options.

For the high terrain simulations, did the results improve with SMS-3DTKE scheme?

Was there any (substantial) change in the precipitation patterns?

 

[Ming Chen]

@Sarath,
We haven't done multiple tests over high topography area to evaluate the performance of SMS-3DTKE. Sorry I have no answer to your question. If you have some results that demonstrate the performance of this scheme, would you please post in the Forum just in case other people may have the same question? thanks in advance.