Longwave cloud radiative forcing, cloud liquid and cloud ice biases

[adamrher]

All of our CESM3 runs have systemically less LWCF then CESM2 and observations. CESM3 also has significantly less cloud liquid and cloud ice compared to CESM2. We believe this is a consequence of resolving the ice nucleation limiter issue in CESM2 using the latest version of PUMAS (See plot from Gettelman et al. 2023 below). We have carried out a set of tuning experiments to increase LWCF and cloud ice (by changing hetfrz_dust, vtrmi and wsubi_factor). While we are successful in reducing the LWCF bias, it is due to increasing cloud ice in the upper troposphere, and we are uncertain whether this is a more realistic cloud field.

LWCF bias and parameter sensitivity:

CLDICE differences from CESM2 and parameter sensitivity:

Figure from Gettelman et al. 2023 (Control = CESM2; PUMAS = CESM3):

The next steps are to evaluate these parameter sensitivity experiments against cloud observations. Three possibilities, none which address the issues directly (where in the vertical should we increase cloud ice?) are (1) evaluate LWP vs latitude, as in the Gettelman plot above, (2) cloud regime analysis after Davis and Medeiros 2024, (3) or a comparison to aircraft observations of Southern Ocean low clouds after Gettelman et al. 2020.

@PeterHjortLauritzen @JulioTBacmeister @brianpm @cmcclus @islasimpson @swrneale

[brianpm]

Ice water path is difficult to constrain, but we can probably get some feeling from this paper:

Duncan, D. I. and Eriksson, P.: An update on global atmospheric ice estimates from satellite observations and reanalyses, Atmos. Chem. Phys., 18, 11205–11219, https://doi.org/10.5194/acp-18-11205-2018, 2018.

Their Figure 3:

[adamrher]

Thanks for pointing out this paper. I understand most of these estimates include the ice condensates, whereas the TGCLDIWP variable in CAM is just cldice. They don't include any satellite estimates of this variable, but we could compare TGCLDICE to reanalysis 'CIWP.' We could also compute a total ice water path including condensates, but the spread in the obs is massive and so perhaps not too useful.

[adamrher]

Hi All,

I've put together two slides suggesting that we increase micro_mg_dcs = 500->600 microns. I am planning on adding these to the slides for the CESM3 project mtg tmrw morning. Please feel free to provide feedback, thanks!

ADF diags for the dcs run here: https://webext.cgd.ucar.edu/FLTHIST/f.e30_beta04.FLTHIST.ne30.dcs600.001/atm/

[swrneale]

Are there any impacts on the tape recorder given the upper tropospheric Q changes?

[adamrher]

It does cause some marginal drying of the tape recorder. See special plots in the ADF link.