Spinup for an ISSP585 simulation

[KeerZ]

Hi all,

I am running a land-only simulation from 2095 to 2099 driven by the coupler history file of a cmip6 simulation (b.e21.BSSP585cmip6.f09_g17.CMIP6-SSP5-8.5.102). The compset I am using is SSP585_DATM%CPLHIST_CLM50%SP_SICE_SOCN_MOSART_SGLC_SWAV. Since I changed the surface data, I need to spin up the land initial condition. I plan to run a simulation using the same compset for 10 years. I set:

./xmlchange DATM_CPLHIST_CASE=b.e21.BSSP585cmip6.f09_g17.CMIP6-SSP5-8.5.102,DATM_CPLHIST_DIR=/glade/derecho/scratch/keerzhang/archive/SSP585cpl/b.e21.BSSP585cmip6.f09_g17.CMIP6-SSP5-8.5.102/cpl/hist.mon

./xmlchange RUN_STARTDATE=2095-01-01

./xmlchange DATM_CPLHIST_YR_ALIGN=2095,DATM_CPLHIST_YR_START=2095,DATM_CPLHIST_YR_END=2099

I also set do_transient_crops=.false. and do_transient_pfts=.false.

The land initial condition of the spinning up simulation is b.e21.BSSP585cmip6.f09_g17.CMIP6-SSP5-8.5.102.clm2.r.2095-01-01-00000.nc.

I wonder if this is a good way to spin up the land condition. In particular, I noticed that the CLM_CO2_TYPE is diagnostic, and the CLM_NML_USE_CASE is 1850-2100_SSP5-8.5_transient. If I run this simulation for 10 years to spin up, are there any transient behaviors that might influence the spinning up of my land condition?

My spinup case can be found here: /glade/u/home/keerzhang/CESM2.1.4_DC/cime/scripts/ModEmi/SSP585/ISSP585Sp.f09_g17.Default_spin. Any suggestions are appreciated. Thank you!

 

[KeerZ]

Also, I focus on urban temperature and anthropogenic heat flux. In CLMU, the AC is turned off in year 2100. Here if I would like to keep the AC on, should I set the RUN_STARTDATE to be 2015-01-01 instead of 2095-01-01?

 

[oleson]

I'm looking at your case, can you run ./case.setup and ./preview_namelists so that I can look at the resolved namelists?
But, I believe the case comes out of the box with:
stream_fldfilename_urbantv = '/glade/p/cesmdata/cseg/inputdata/lnd/clm2/urbandata/CTSM52_tbuildmax_OlesonFeddema_2020_0.9x1.25_simyr1849-2106_c200605.nc'
stream_year_first_urbantv = 2000
stream_year_last_urbantv = 2000

So I think the variable that controls AC will always be from year 2000. So AC should be on for the duration of your experiment.
It really should come of the box with stream_year_last_urbantv=2106 though, although the data for years 2000 through 2099 are the same.
I would verify that it is picking up year 2000 using a short model run.

The case seems to come out of the box (at least with release-cesm2.1.4) with:
CLM_CO2_TYPE: constant
CCSM_CO2_PPMV: 284.7

But the coupler history files should be providing the CO2 through the following setting in the datm.streams.txt.CPLHISTForcing.State3hr:

a2x3h_Sa_co2diag co2diag
a2x3h_Sa_co2prog co2prog

It should be using the a2x3h_Sa_co2diag in the coupler history files but maybe do a short run to see that it is.

I'm also seeing:

CLM_NML_USE_CASE: UNSET

which is different from yours. Not sure why?

Otherwise, I don't see anything else that is transient in this simulation (except of course for the atmospheric forcing).

 

[KeerZ]

I'm looking at your case, can you run ./case.setup and ./preview_namelists so that I can look at the resolved namelists?
But, I believe the case comes out of the box with:
stream_fldfilename_urbantv = '/glade/p/cesmdata/cseg/inputdata/lnd/clm2/urbandata/CTSM52_tbuildmax_OlesonFeddema_2020_0.9x1.25_simyr1849-2106_c200605.nc'
stream_year_first_urbantv = 2000
stream_year_last_urbantv = 2000

So I think the variable that controls AC will always be from year 2000. So AC should be on for the duration of your experiment.
It really should come of the box with stream_year_last_urbantv=2106 though, although the data for years 2000 through 2099 are the same.
I would verify that it is picking up year 2000 using a short model run.

The case seems to come out of the box (at least with release-cesm2.1.4) with:
CLM_CO2_TYPE: constant
CCSM_CO2_PPMV: 284.7

But the coupler history files should be providing the CO2 through the following setting in the datm.streams.txt.CPLHISTForcing.State3hr:

a2x3h_Sa_co2diag co2diag
a2x3h_Sa_co2prog co2prog

It should be using the a2x3h_Sa_co2diag in the coupler history files but maybe do a short run to see that it is.

I'm also seeing:

CLM_NML_USE_CASE: UNSET

which is different from yours. Not sure why?

Otherwise, I don't see anything else that is transient in this simulation (except of course for the atmospheric forcing).

Thanks, Keith! I have run ./preview_namelist to generate name list files. I messed up some setups of the old case, and now I am working on this one: /glade/u/home/keerzhang/CESM2.1.4_DC/cime/scripts/ModEmi/SSP585/ISSP585Sp.f09_g17.Default_spin231109.

It looks like my case comes out of the box with:
stream_year_first_urbantv = 2015
stream_year_last_urbantv = 2106
CLM_NML_USE_CASE: 1850-2100_SSP5-8.5_transient
CLM_CO2_TYPE: diagnostic

Not sure why these setups are different from yours. Is it because I am working on Derecho? I created my case by
./create_newcase --case ~/CESM2.1.4_DC/cime/scripts/ModEmi/SSP585/ISSP585Sp.f09_g17.Default_spin231109 --compset SSP585_DATM%CPLHIST_CLM50%SP_SICE_SOCN_MOSART_SGLC_SWAV --res f09_g17 --project UYLE0022 --run-unsupported

For this spin-up simulation, perhaps I could just set stream_year_first_urbantv and stream_year_last_urbantv to be 2095 to make sure AC doesn't turn off.

Two files in my CaseDocs contain co2diag: the datm.streams.txt.co2tseries.SSP5-8.5 and datm.streams.txt.CPLHISTForcing.State3hr. I will run a short run to confirm if the a2x3h_Sa_co2diag is used. As long as the CO2 is cycled from 2095 to 2099 (instead of 2095 to 2104) during the 10 years, I think it should be fine.

 

[oleson]

I now see what happened with mine. I missed the first "S" in the compset and it didn't seem to give me an error, it just generated some non-transient case, interesting.
Yes, it seems like you could set stream_year_first_urbantv and stream_year_last_urbantv to be 2095