Interaction of mix_isotropic with gray-zone schemes (Shin-Hong and 3D-TKE)?

[Pechudin]

So, in order to run hectometer scale simulations, I used a range of options (km_opt=2, 4, 5 or the Shin-Hong scheme). I use a 9-3-1-0.2 km nested domains and this is what I used thus far:

• km_opt=2 or 4 with YSU scheme on the D01-03
• km_opt=5 on all domains (3D-TKE scheme)
• Shin-Hong scheme on all domains

However, when I read about using km_opt=2, 4 (Smagorisky or TKE scheme), the recommendation was to use mix_isotropic=1. How does mix_isotropic interact with the scale-sensitive schemes, since their diffusivity is itself scale-dependent? Which takes priority? Should I use mix_isotropic=0 or only enable it on the 200 m domain?

I will run the simulations anyway and report what happens, but it would be helpful if there was a body of knowledge on this.

 

[Pechudin]

Well, the second option, with km_opt=4,4,4,2 and mix_isotropic=0,0,0,1 failed immediately with the errors of this type in all the rsl.* files:

rsl.error.0063: med_latbound_in: error opening wrfbdy_d01 for reading. IERR = -1021

Not sure what the problem is there since real worked fine and produced input and boundary files. The case with mix_isotropic=0 in all domains is running fine it seems.

 

[kwerner]

Hi,
km_opt=3 is the 3d Smagorinsky and should be used instead of km_opt=4 when no PBL scheme is used. mix_isotropic = 1 uses the same mixing length for vertical and horizontal mixing, and 0 separately computes them from the different grid lengths.