b.cesm3_cam058_mom_e.B1850MOM.ne30_L32_t061.cam6_cice5.019b

[cecilehannay]

Description: Same as 017 (Issue: https://github.com/NCAR/amwg_dev/issues/145) but with

 clubb_c8               =  4.6

The goal being to bring increase the cloud and bring the RESTOM closer to zero.

This similar to #146 but using another set of parameters to retune RESTOM.

This is based on Adam's comment: _If you recall all the way back Jiang Zhu's talk on the Eocene simulations in April (here), C8 is an alternative way to impact the skewness. clubbgamma hits it really hard because it controls a PDF parameter - the width of the w' distribution. But C8 does it through a more natural way, in that it is the damping coefficient on the prognostic w'^3 equation. Increasing C8 will reduce skewness, which will increase cloud thickness.

Case directory: Locally (if still available): /glade/p/cesmdata/cseg/runs/cesm2_0/b.cesm3_cam058_mom_e.B1850MOM.ne30_L32_t061.cam6_cice5.019b

On github: https://github.com/NCAR/amwg_dev/tree/b.cesm3_cam058_mom_e.B1850MOM.ne30_L32_t061.cam6_cice5.019b

Sandbox: Locally (if still available): /glade/work/hannay/cesm_tags/cesm3_cam6_3_058_MOM_e

On github: hash: https://github.com/NCAR/amwg_dev/commit/8f70c084f85987a02a88df6570c8499a62fbebd9

Diagnostics: AMWG diags (if available) https://webext.cgd.ucar.edu/B1850MOM/b.cesm3_cam058_mom_e.B1850MOM.ne30_L32_t061.cam6_cice5.019b/atm/

Contacts: @cecilehannay, @gustavo-marques, @JulioTBacmeister, @adamrher

[JulioTBacmeister]

I think both 18b and 19b are looking good. 18b is a little on the warm side - as expected. I'd say let them both run another day, and we can reevaluate tomorrow.

[JulioTBacmeister]

I'm interested in how c19b evolves. I agree it may be to switch the default c8

[JulioTBacmeister]

Unless Gustavo's Lab Sea convection diagnostics look bad for 018b or 019b I'd allow them to keep going one more day. Then we decide on 020.

I am not inclined to do more tuning of either 018b or 019b even though they appear to have RESTOM ~0.7 and ~0.4 respectively.

[gustavo-marques]

Case 18b is looking better than 19b in terms of Lab Sea convection.

[JulioTBacmeister]

We should think about the next steps here. Looks like convection in 019b died pretty quickly. Correct? How is convection in 017?

One more timeseries plot:

[swrneale]

I think if 017 is still freezing then we need to take stock. @gustavo-marques remind me. Convection in/near the lab. sea was too strong in the g-cases right, and so there would be no real justification for upping any of the contributions to cvmix?

[JulioTBacmeister]

I don't think we have seen that 017 is freezing. It isn't in Gustavo's plot.

[gustavo-marques]

Sorry, I forgot to include 017 in the time series. Below is the updated plot.

As a reminder, the above time series are area-weighted mixed layer depths (MLDs). The plot below shows see Lab Sea area:

Focusing on case 017, let's look at the March MLD averaged between years 20-35.

017 is underestimating MLDs in the Lab Sea, but it's still convecting. Notice that the most vigorous convection has moved eastward, just east of the southern tip of Greenland. This is also the case for 018b and 019b (see plots below and notice that averages were done between years 15-25 in these cases).

@swrneale: yes, convection in the Lab Sea tend to be on the stronger side in G cases and we do not believe there is a justification for tunning cvmix. However, remember that we do not have the Estuary Box Model parameterization in MOM at the moment, and this missing phsyics might be contributing to this issue.

[swrneale]

So really only 18b has any appreciable deep convection left and that is a bit anaemic.

[gustavo-marques]

Updating the MLD plots.

Area-weighted (Lab Sea) mixed layer time series

March mean MLD (years 25-35)

case 009

case 016 (years 20-40)

case 017

case 018b

case 019b

Lab Sea temperature bias (model - WOA18)

case 009

case 016

case 017

case 018b

case 019b

Lab Sea salinity bias (model - WOA18)

case 009

case 016

case 017

case 018b

case 019b

[gustavo-marques]

The Lab Sea in 017 is very fresh near the surface (just like 009), but it's not as cold as 009. It's still convecting a bit, but the Lab Sea might freeze soon.

018b is looking good in terms of T&S biases and MLDs.

[adamrher]

@jiang-zhu I'm wondering if you wouldn't mind commenting on the differences between increasing c8 and decreasing gamma. Instead of cranking up c8, could you have instead lowered clubb_gamma to prevent the 6XCO2 runaway greenhouse? I've found that comparing 18b to 17 (clubb_gamma=0.298->0.280) and 19b to 17 (c8=4.2->4.6), the change in SWCF looks very similar, spatially.

They both look to be doing the same thing, which is expected since they both target skewness, but the patterns are still remarkably similar given the very different ways these two methods effect skewness.

[jiang-zhu]

Hi @adamrher clubb_gamma_coef is already low in the 2deg model (0.275). We may need a very low clubb_gamma_coef to tune the Eocene high CO2 simulation, which I didn't do, partly because you once said you usually don't push clubb_gamma_coef to lower than 0.26.

I did test the gamma_coefb (default value=0.32 in parameters_tunable.F90), though. I tried setting gamma_coefb = clubb_gamma_coef = 0.275. The simulated Eocene temperature response to CO2 increase (from 1× to 6×) is not impacted. In contrast, increasing clubb_c8 drastically impacts the temperature sensitivity.