WACCM/CARMA MA run for ISA-MIP BG run

[zywshoon]

I will conduct a model run repeating 2000 for 20 years for background aerosol. The model is ok using CMIP6 emission. @tilmes could you point me to a 2000 case I can clone or branch? And how could I run 2000 repeatedly for 20 years? I will work on the output list and let you know if we need more variables that are not in the default.

[tilmes]

@zywshoon The case to clone is: /glade/p/cesm/chwg_dev/tilmes/cases/carma/FWmaCARMAnudged.f19_f19_mg17.carma_trop_strat06.1990_2000 However, this is a nudged model run. Is ISA-MIP model runs free running, which I assume if it is a repeated year 2000 case? If so, it may be best first to run Issue #1 to see how the model is doing without nudging?

[zywshoon]

@tilmes I think you know better than I am. The description doesn't say if I nudge or free run, it says "The simulations prescribed for this experiment are time-slice simulations for the year 2000 with prescribed SST including all sources of aerosols and aerosol precursors except for explosive volcanic eruptions. Modeling groups should run this simulation with varying QBO, either internally generated or nudged to the 1981–2000 period."

What do you think I should do first?

[tilmes]

@zywshoon we would use a nudged QBO from 1980-2000, but we would need to run the model as a cyclical simulation, I am not sure how to set this up. Maybe we could ask if we could also just run with an averaged QBO over two cycles to make this easier? Also, it seems like they want SSTs, GHGs and emissions at the year 2000, correct?

[zywshoon]

@tilmes I think as long as the QBO varies, it is ok. My understanding is that it needs prescribed SST and required sulfur emissions (as we confirmed, CMIP6 emission is ok). GHG needs to be reasonable for 2000 but not restricted to any emissions. quote: "modelling groups should prescribe for these species their usual emission database for the year 2000. Each group can specify solar forcing for year-2000 conditions according to their usual data set."

Do you think we can run the nudged model from 1980 to 2000 with cyclical emissions of 2000 and 2000 SST? Is it doable?

[tilmes]

@zywshoon I think, they want a free running (SST 2000) simulation, this is usually done. I am currently setting up a free running version (Issue #1 ). We don't have a compset for the FWmaCARMA2000 case yet, maybe @fvitt could set this up? We have one for FCARMA2000climo.

[zywshoon]

@tilmes Ok. Please let me know if I can help with validating #1

[tilmes]

@zywshoon We discussed on Wednesday, that we can run a transient FHIST run but set everything to CYCLICAL for the year 2000 and let the QBO be nudged for 1980-2000. The model would therefore run from 1980-2000 (but still be a cyclical run) with fixed SST for the year 2000. Do you want to try setting up the run or want me to do it? You can setup the run like Issue #1 but then change it to the year 2000-climo emissions and the fixed SSTs.

[zywshoon]

@tilmes Thank you very much. This is how I set it up. Would you mind taking a look? I use the strat09 tag and copy the create_case from #1.

cd /glade/work/fvitt/camdev/carma_trop_strat09/cime/scripts/

./create_newcase --compset FWmaCARMAHIST --res f19_f19_mg17 --project P19010000 --run-unsupported --case /glade/u/home/yunqian/cesm/case/ISAMIP_CESM2CARMA_WACCMMA2000climo --project P19010000 --driver mct

cp /glade/p/cesm/chwg_dev/tilmes/cases/carma/FWmaCARMAHIST.f19_f19_mg17.carma_trop_strat09.1980_2014/user_nl_cam .

Question: should I spin up for 5 years to get to equilibrium? Which setting is for QBO nudged for 1980-2000? HHow do I change it to 2000-climo emissions and the fixed SSTs? My case directory: /glade/u/home/yunqian/cesm/case/ISAMIP_CESM2CARMA_WACCMMA2000climo

[zywshoon]

@tilmes @fvitt Do you know if we have any of these passive tracers in the model? I see "AOA1" and "AOA2", but not the others. What's the difference between AOA1 and AOA2?

_nh_50: Passive tracer with fix surface concentration equal to 100 ppb between 30˚N and 50˚N and equal to 0 outside of this latitudinal band, e-folding decay time of 50 days

tr_50: Passive tracer with fix surface concentration equal to 100 ppb between 20˚S and 20˚N and equal to 0 outside of this latitudinal band, e-folding decay time of 50 days;

sh_50: Passive tracer with fix surface concentration equal to 100 ppb between 50˚S and 30˚S and equal to 0 outside of this latitudinal band, e-folding decay time of 50 days.

AOA: Passive tracer for the stratospheric mean age-of-air. Modelling groups can use their existing implementation or implement a tracer with a global fixed surface layer mixing ratio of 0 ppbv and a uniform unspecified fixed source (at all levels) everywhere else, which must be constant in space and time.

ST80_25: Passive tracer to estimate the exchange from the stratosphere to the troposphere. This is achieved by fixing the mixing ratio above 80hPa (200ppbv) to a constant value, and imposing a uniform fixed 25-day exponential decay in the troposphere only.

Volc: Passive volcanic tracer for the HerSEA experiments. The tracer is initialized in the same way as the volcanic SO2 emission, with an initial value of 1._

[zywshoon]

@tilmes I briefly checked the Extinction output. Looks good compared to GloSSAC. So I plan to start the run for 20 years. The organizer recommended I start from 1990 to 2010 for 20 years to nudge the QBO. Is it easy to change the start date? Since we plan to start from 1980 right now. I don't think it's a big deal if you think we just started in 1980.

I am copying your initial file for 2000-01-01. I think that's the only thing we need to change. Let me know if you think of others things we talked but I forgot.

[zywshoon]

@tilmes ISAMIP CESM2/CARMA run directory:

/glade/u/home/yunqian/cesm/case/ISAMIP_CESM2CARMA_WACCMMA2000climo

[zywshoon]

@tilmes I diff the current ISAMIP CARMA run with the FWma2000climo. There's many difference but I select several that I cannot understand and want to check with you to make sure they are ok:

FWma2000climo: lght_no_prd_factor = 1.00D0 CARMA: lght_no_prd_factor = 5.0D0

FWma2000climo: history_scwaccm_forcing = .true. CARMA: history_scwaccm_forcing = .false.

FWma2000climo: effgw_beres_dp = 0.7D0 CARMA: effgw_beres_dp = 0.1D0

FWma2000climo: epp_all_filepath = '/glade/p/cesmdata/cseg/inputdata/atm/cam/solar/SolarForcing1995-2005avg_c160929.nc' FWma2000climo: epp_all_varname = 'epp_ion_rates' CARMA: Nothing

FWma2000climo: water_refindex_file = '/glade/p/cesmdata/cseg/inputdata/atm/cam/physprops/water_refindex_rrtmg_c080 CARMA: Nothing

FWma2000climo: cam_chempkg = 'waccm_ma_mam4' CARMA: cam_chempkg = 'waccm_ma_noaero'

FWma2000climo: history_scwaccm_forcing = .true. CARMA: history_scwaccm_forcing = .false.

[tilmes]

@zywshoon I checked the run and all looks good, and has to be equal to what is set in the CARMA run that I have run before.

[tilmes]

@zywshoon I noticed your run is done. We need to move it into time series, so it will use less space. Have you done this before? We have instructions on the CAMchem wiki page how to do this. This is better than doing it by hand, because it applies some compression that makes the files smaller.

[zywshoon]

@tilmes Oh, I don't know that. Could you point me to these. because I just did them by hand.

[tilmes]

https://wiki.ucar.edu/display/camchem/CAM-chem+automated+diagnostics
You want to do Step 1 and 2a (so the postprocessing scripts go in your case directory of your run. Then you can just do Step 5 for timeseries, you really don't have to edit anything but your "timeseires" to add an account number and submit it, that should work. Please let me know if you need help.

[zywshoon]

@tilmes I tried this instruction, thanks. I see postprocess/ created, and then I submit the timeseries, but I don't see anything created. Where it suppose to appear?

This is my directory: /glade/scratch/yunqian/archive/ISAMIP_CESM2CARMA_WACCMMA2000climo/atm/hist/modelresults

[zywshoon]

@tilmes It seems it is working. The log file shows something. But seems it is for lnd model output, not atm. Also I don't know where the timeseries output it stored.

/glade/u/home/yunqian/cesm/case/ISAMIP_CESM2CARMA_WACCMMA2000climo/postprocess/logs/timeseries.log.20230105-114649

[tilmes]

try setting this to true in env_postrocess.xml

This should create all the different time series streams.

[zywshoon]

@tilmes it's working! Thank you!

@tilmes @dan-visioni I am trying to figure out how to upload to their folder. It is so complicated... Maybe we can do it together when the MAM run is done.

[zywshoon]

Oh I actually figured it out! I will show you when you are done with yours. @dan-visioni

[tilmes]

@zywshoon just noticed that your case uses SSTs from 2010 not 2000:

We should use 2000 instead I think: sst_HadOIBl_bc_1.9x2.5_2000climo_c180511.nc As for the MAM4, I am thinking we should rerun the CARMA run with the suggested bug fixes. I am going to run it for a couple years to see if we get in balance.. first and do the budgets...

[dan-visioni]

I'm preparing some comparisons between CARMA and MAM based on current simulations @tilmes and I can share them at tomorrow's meeting so maybe we can get a better understanding of the differences... maybe let's wait before rerunning CARMA?

[zywshoon]

@tilmes @dan-visioni good idea. Let's discuss tomorrow and see what are all the changes I should make before I rerun the model

[tilmes]

I performed one year of a run with the updated CARMA code, including all the fixes we discussed. I just produced my budget from the diagnostic output (not Chuck's since I did not add them properly) and we are getting lower values, as expected. I will run my case a couple of years first and then we can make a new sandbox and redo this run, but we need to first see if we have sufficient computer time. I need to redo all the runs for the paper as well. For the entire atmosphere:

For the troposphere:

[tilmes]

@zywshoon I just set up a new test run for you, that should include everything we need for starting the new ISAMIP run with CARMA. Please see my new Case: /glade/p/cesm/chwg_dev/tilmes/cases/carma/ISAMIP_CESM2CARMA_WACCMMA2000climo.carma_trop_stratv10.001 You can clone my case, and check that everything is OK (I will briefly check the diagnostics when the run is done).

For information to the new run, please see: https://github.com/NCAR/CARMA_dev/discussions/22

1. Please use /glade/work/fvitt/camdev/carma_trop_strat10 as the code base.
2. If possible leave all the diagnostics and namelist settings as they are, but please add any variables that you need for ISAMIP in additions
3. Please let me know if you have any problems or questions

[tilmes]

@zywshoon you can start your CARMA run now. Please use this code /glade/work/fvitt/camdev/carma_trop_strat11 and clone this case: /glade/p/cesm/chwg_dev/tilmes/cases/carma/ISAMIP_CESM2CARMA_WACCMMA2000climo.carma_trop_stratv10.test

It should include all the right settings and variables. I would keep them all and just add what is missing for ISAMIP output on top of it!!! Thanks!

[zywshoon]

@tilmes thank you this is great.

There's one difference in the env_run.xml: yours: entry id="MOSART_SIM_YEAR" value="1850"

mine: entry id="MOSART_SIM_YEAR" value="2000"

I want to confirm I should use 2000, right?

[tilmes]

Oh yes, we want 2000 for that, thanks for checking!

[tilmes]

The new simulations look much better, that is great!

[tilmes]

@zywshoon have you saved the ISAMIP year 2000 run somewhere on campaign storage? I am trying to find an IC file for this case. Thanks Simone

[zywshoon]

@tilmes /glade/campaign/acom/acom-climate/yunqian/ISAMIP_CESM2CARMA_WACCMMA2000climo_carmatag11/ISAMIP_CESM2CARMA_WACCMMA2000climo_carmatag11/