f.cam6_3_112.FLTHIST_v0c.ne30.non-ogw-ubcT-effgw0.3-rdg_beta0.5-vtrmi1.5-hb-tt.001

[cecilehannay]

Description: A new implementation of "HB when CLUBB is not active" where HB is only doing free atmosphere mixing (Richardson number based) and not PBL mixing - thanks to @tto061 (Thomas Toniazzo) for draft code https://github.com/NCAR/amwg_dev/discussions/284#discussioncomment-6307477.

See science discussion here: https://github.com/NCAR/amwg_dev/discussions/284#discussioncomment-6307477

Proposed run: Same configuration as https://github.com/NCAR/amwg_dev/issues/316 but with new HB source mods:

/glade/u/home/pel/src/for-hb-nohack-freeatm_hb-cam6_3_110/clubb_intr.F90
/glade/u/home/pel/src/for-hb-nohack-freeatm_hb-cam6_3_110/vertical_diffusion.F90
/glade/u/home/pel/src/for-hb-nohack-freeatm_hb-cam6_3_110/hb_diff.F90

Case directory: Locally (if still available): /glade/p/cesmdata/cseg/runs/cesm2_0/f.cam6_3_112.FLTHIST_v0c.ne30.non-ogw-ubcT-effgw0.3-rdg_beta0.5-vtrmi1.5-hb-tt.001

On github: https://github.com/NCAR/amwg_dev/tree/f.cam6_3_112.FLTHIST_v0c.ne30.non-ogw-ubcT-effgw0.3-rdg_beta0.5-vtrmi1.5-hb-tt.001 Sandbox: Locally (if still available): /glade/work/hannay/cesm_tags/cam6_3_112

On github: (https://github.com/ESCOMP/CAM/tree/cam6_3_112)

Diagnostics: ADF diags (if available) https://webext.cgd.ucar.edu/FLTHIST/f.cam6_3_112.FLTHIST_v0c.ne30.non-ogw-ubcT-effgw0.3-rdg_beta0.5-vtrmi1.5-hb-tt.001/atm/

Contacts: @tto061 @PeterHjortLauritzen @adamrher @JulioTBacmeister @cecilehannay

[justin-richling]

ADF diags vs f.cam6_3_112.FLTHIST_v0c.ne30.non-ogw-ubcT-effgw0.3-rdg_beta0.5-vtrmi1.5.001

[PeterHjortLauritzen]

See discussion here https://github.com/NCAR/amwg_dev/discussions/284#discussioncomment-6347191

[adamrher]

@PeterHjortLauritzen

ADF diags vs f.cam6_3_112.FLTHIST_v0c.ne30.non-ogw-ubcT-effgw0.3-rdg_beta0.5-vtrmi1.5.001

Is the differences with the control here both: (a) no PBL, counter-gradient term, AND (b) only apply where clubb is not active? Or is the only difference (a), and the control also does not apply HB where clubb is active?

[PeterHjortLauritzen]

Control: HB free atmosphere and HB PBL where CLUBB is not active. New run: HB free atmosphere only where CLUBB is not active.

Hope that clarifies?

[adamrher]

It does. In terms of code, it looks like cgh/cgs coefficients are allowed to be non-zero above clubbtop in the Control. Whereas in the New run, cgh/cgs are always zero everywhere.

If I've got that right, then these cgh/cgs terms are actually doing a lot of vertical mixing in the Control. See the difference in zonal-height Q here. The New run here moistens the PBL and dries the free-troposphere. But if you look the impact of the hb mods in the control (which we can't directly discern because we don't a vtrmi=1.5 run w/o hb mods) ... looking instead at our first hb run vs. w/o hb here. You can see the Q changes are the ~equal but opposite; it dries the PBL and moistens the free troposphere. That is, the cgd/cgh terms being active above clubbtop are still mixing a large amount of moisture into the free troposphere, and they explain ~all of the changes in free troposphere Q, compared with turning off the hb mods.

I'm glad we are zero'ing out these cgd/cgh terms. Additional Ri mixing above clubbtop I can stomach, but these counter-gradient terms are like having another convection scheme in the model -- and we've already got too many of those. Plus zeroing out only part of the vertical structure of the non-local, counter-gradient term is a recipe for disaster, if you get my drift.

[PeterHjortLauritzen]

Thanks for your analysis!