Option to turn on HB where CLUBB is not active

[PeterHjortLauritzen]

What is the feature/what would you like to discuss?

To address stability issues with non-FV dynamical cores we would like to turn on Holtslag-Boville PBL scheme where CLUBB is not active. This code change will help address

#772

Science discussion here:

NCAR/amwg_dev#284

Have code ready and will start PR when time permits ... branch with code mods here:

https://github.com/PeterHjortLauritzen/CAM/tree/clubb_hb

Is there anyone in particular you want to be part of this conversation?

@adamrher @JulioTBacmeister @fvitt

Will this change (regression test) answers?

Yes

Will you be implementing this enhancement yourself?

Yes

[tto061]

Hi Peter, here's my proposed version for vertical_diffusion.F90 where only the free-atmosphere diffusivity is calculated (and applied).

I also highlight the tuning paramter of the (square) mixing length, ml2 (sticking to the original code, where it's unnecessarily a vector of constants set every time, but just for clarity).

I would also point out that the mixing length for scalars may be much smaller than that for momentum, so in principle khq=khm *small_factor , which would help keeping inversions/cloud tops sharp while diffusing momentum (e.g. at the top of Sc the shear scale height teds to be much larger than that of q,T).

Note (from the declarations in vertical_diffusion_tend) that the subroutine turbintd should be made public in hb_diff.F90, or, as I think perhaps would be better, it could be moved as an additional public module to pbl_utils (since it is generic and not tied to the Holtslag-Boville scheme).

Please finally note that I did not test this yet, for now it's just an outline.

vertical_diffusion.F90.gz