How to run CAM6 with CLM4 in CESM2

[kezhoulumelody@gatech_edu]

What version of the code are you using?
CESM2.1.5


Have you made any changes to files in the source tree?
Not yet

Describe your problem or question:

Dear community,

I plan to run a set of nudging experiments based on compset FHIST_BGC in CESM2.1.5. However, instead of using the default version of land component (i.e., CLM5), I want to use CLM4 (the version from CESM1). Besides the land, other components (such as SST and atmospheric radiative forcing) will keep the same as FHIST_BGC. On top of that, U and V will be nudged towards ERA5 reanalysis. Could any experts provide some insights of how to make the CAM6_CLM4 combination work?

I am thinking of creating a new compet looks like the following: "HIST_CAM60_CLM40%BGC-CROP_CICE%PRES_DOCN%DOM_MOSART_CISM2%NOEVOLVE_SWAV" or "HIST_CAM60_CLM40%SP_CICE%PRES_DOCN%DOM_MOSART_CISM2%NOEVOLVE_SWAV". I am not sure if I should turn off BGC.

The experiment setup will be somewhat similar to this literature: "Simpson, I. R., Bacmeister, J., Neale, R. B., Hannay, C., Gettelman, A., Garcia, R. R., et al. (2020). An evaluation of the large-scale atmospheric circulation and its variability in CESM2 and other CMIP models. Journal of Geophysical Research: Atmospheres, 125, e2020JD032835. https://doi.org/10.1029/2020JD032835"

Thanks!

Best,
Melody

 

[islas]

Hi Melody,

I don't recall using simulations in which CAM6 was coupled to CLM4 in that paper. Are you thinking of this one instead? In that paper, I ran a simulation in which CESM2 with CAM6 was run with CLM4.5. But I was also running with satellite phenology. Do you need to have the BGC option? What I did to run CESM2 with CLM4.5 was to set up a case using the FHIST compset (which uses satellite phenology as opposed to interactive BGC in the land in the FHIST_BGC case). I think I then used the following xlmchange commands:

./xmlchange CLM_CONFIG_OPTS="clm4_5"
./xmlchange LND_TUNING_MODE="clm4_5_cam6.0"

We had simulations in that paper that used CAM6 with CLM4 as well, but @hannay ran those, so she may be able to provide more information.

Isla

 

[kezhoulumelody@gatech_edu]

Hi Melody,

I don't recall using simulations in which CAM6 was coupled to CLM4 in that paper. Are you thinking of this one instead? In that paper, I ran a simulation in which CESM2 with CAM6 was run with CLM4.5. But I was also running with satellite phenology. Do you need to have the BGC option? What I did to run CESM2 with CLM4.5 was to set up a case using the FHIST compset (which uses satellite phenology as opposed to interactive BGC in the land in the FHIST_BGC case). I think I then used the following xlmchange commands:

./xmlchange CLM_CONFIG_OPTS="clm4_5"
./xmlchange LND_TUNING_MODE="clm4_5_cam6.0"

We had simulations in that paper that used CAM6 with CLM4 as well, but @hannay ran those, so she may be able to provide more information.

Isla

Hi Isla,

Oh I am very sorry, I put the wrong paper citation here. Yes I meant to attach the surface temperature variability paper you linked.

Here is the reason we wanted to swap the CLM version: we found that T2m winter variability in CESM1 wasn't improved a lot even though circulations were nudged towards ERA5. However, if we do similar wind nudging experiments in CESM2, the magnitude of T2m winter variability is much closer to ERA5. We think different CLMs might explain why T2m variance are different when doing nudging experiments with different versions of CESM, and this should be consistent with your paper. However, we still wanted to understand the exact role of circulation, by switching CLM to an old version that matches CESM1, and perform the wind nudging experiments again. Our CESM1 wind nudging experiment was done with compset FAMIPC5, which uses clm4.0. I think it would be better to use clm4.0, but I don't know how different is clm4.0 compared to clm4.5.

We have a base run that uses FHIST_BGC with winds nudged towards ERA5, but I think I can always do one with FHIST. But by the end of the day, we still need to compare our experiment results with GOGA2 and AMIP-CMIP6 which were conducted with FHIST_BGC. I don't know if interactive BGC is important in this case.

Thanks you so much Isla!

Best,

Melody

 

[islas]

Hi Melody,

OK, if you want to run with CLM4.0, I'd have to defer to @hannay as I didn't run that. You could try doing something similar to what I suggest above replacing "4_5" with "4" but I have no idea whether that actually works. Hopefully Cecile will have a better suggestion.

Isla