MAM run for H2SO4 injections

dan-visioni commented 1 year ago

Issue for the MAM runs that requires a different baseline case.

GHGs and ozone-depleting substances (ODSs) use the 2040 projection values from the SSP5-8.5 scenario (O’Neill et al., 2016)
monthly mean climatological sea surface temperatures (SSTs) and sea ice boundary conditions derived from an average of the years 1988-2007 of the CMIP5 PCMDI-AMIP-1.1.0 SST/Sea Ice dataset (Taylor et al., 2000) are annually repeated.

So, things I am unclear about:

can you use 1º emissions on the 2º or do we need to regrid them? How does it work?
Sadly, I didn't retain the CMIP5 SST, so need to reproduce them again: do I need to use a 2º version of the SST?

tilmes commented 1 year ago

@dan-visioni SSTs: what SSTs have you used for the original MAM4 injection runs? I assume those would be from CMIP6? We can produce new SST averages if needed, but usually it is fine to use the 1deg file for 2deg (the model can interpolate them). Regarding emissions, I am producing 2deg emissions.

dan-visioni commented 1 year ago

The file is in /glade/work/visioni/SST under sstice_yaga_1988_2007_avg_0.9x1.5_1950-2100.nc

This is a file we produced with Yaga that has 1988-2007 averaged conditions, that was used for AMIP style runs. This is the same one that Ulrike also used, and if I remember correctly, it's from QBOi runs.

tilmes commented 1 year ago

That file should also work for the 2degree model run.

tilmes commented 1 year ago

I created 2 deg future emissions: /glade/scratch/tilmes/Emissions_CMIP6_ssp/regrid_mol/ScenarioMIP_IAMC-REMIND-MAGPIE-ssp585-1-1/emis We still have to compare them to the 1deg to make sure the totals are OK.

tilmes commented 1 year ago

@dan-visioni I compared the emissions and they are OK. Can you please share the namelist and case of the run when you set it up, so we can check the emissions?

dan-visioni commented 1 year ago

Starting the case: /glade/work/visioni/cases/FW2000climoMA.f19_f19_mg17.Control.001 Cloning from /glade/p/cesm/chwg_dev/tilmes/cases/carma/FWmaMAM4HIST.f19_f19_mg17.carma_trop_strat09.1980_2014

Added new emissions in user_nl_cam (I copied them in /glade/work/visioni/inputdata/emis/ScenarioMIP_IAMC-REMIND-MAGPIE-ssp585-1-1/)

I'm using also the /glade/work/visioni/cases/f.e21.FW2000climoMA.Control.001 (the 1Deg equivalent) as guide for the namelist

Not sure what to change here:

flbc_cycle_yr = 2040 flbc_file = '/glade/p/cesmdata/cseg/inputdata/atm/waccm/lb/LBC_20140116-25001216_CMIP6_SSP585_0p5degLat_c180905.nc'

@tilmes can you please have a check?

Also, should I start this as a startup eventually? Since we don't have a restart file for 2040.

tilmes commented 1 year ago

@dan-visioni
Thanks, I think, you want to setup the run the same as the example case, so you need to use the same

SST settings in env_run.xml: SSTICE_DATA_FILENAME, SSTICE_GRID_FILENAME, SSTICE_YEAR_ALIGN, SSTICE_YEAR_START, SSTICE_YEAR_END, MOSART_SIMYR
running cyclical for the emissions and lbc for 2040 as in your other namelist, so you can keep the lbc namelist variables as they are
I think, we want to do a hybrid run, like your other case, but I think, we could use the IC file and other restart files from the end of our ISAMIP run, since this was a year 2000 case.
we want to change csw_time_type = 'SERIAL' to CYCICAL and csw_time_cycle_yr = '2000' (I think, this is right, but please check)
for surface emissions, we need to add SOAG emissions. I guess, I have not processed them, so I will look into this. let me know when you have updated it, and I check again.

dan-visioni commented 1 year ago

Thanks Simone! Yes, ok, so I changed all SST to those I had. You're right on the hybrid run of course, I wasn't thinking it's fixed SSTs. So I got a 2000 restart, we'll just have to wait a bit for Cl concentrations to go to 2040 levels like we did last time. And yes, a few surface emission files were missing. Aside from that, everything else seems to be ok?

tilmes commented 1 year ago

What about you use these emissions for now: 'SOAG -> /glade/p/cesmdata/cseg/inputdata/atm/cam/chem/emis/emissions_ssp585/emissions-cmip6-ScenarioMIP_IAMC-REMIND-MAGPIE-ssp585-1-1_SOAGx1.5_anthro_surface_mol_175001-210101_0.9x1.25_c20190224.nc', 'SOAG -> /glade/p/cesmdata/cseg/inputdata/atm/cam/chem/emis/emissions_ssp585/emissions-cmip6-ScenarioMIP_IAMC-REMIND-MAGPIE-ssp585-1-1_SOAGx1.5_bb_surface_mol_175001-210101_0.9x1.25_c20190224.nc', 'SOAG -> /glade/p/cesmdata/cseg/inputdata/atm/cam/chem/emis/emissions_ssp585/emissions-cmip6-ScenarioMIP_IAMC-REMIND-MAGPIE-ssp585-1-1_SOAGx1.5_biogenic_surface_mol_175001-210101_0.9x1.25_c20190226.nc'

tilmes commented 1 year ago

@dan-visioni I am noticing that you don't have all the emissions included yet. It will be important to have the same species in the emission files that you have in your earlier case you set up. I will check in more detail later today.

tilmes commented 1 year ago

@dan-visioni here are the updated emissions we can use for this run:

For external forcings we need to also add the continuously outgassing volcanoes for SO2, so4_a1, so4_a2, and num_a1, num_a2, as done in your earlier case. So you have to add:

'num_a1 -> /glade/u/home/visioni/emis/emissions_ssp585/emissions-cmip6_num_a1_so4_contvolcano_vertical_2040climo_0.9x1.25_c20190523.nc', 'num_a2 -> /glade/u/home/visioni/emis/emissions_ssp585/emissions-cmip6_num_a2_so4_contvolcano_vertical_2040climo_0.9x1.25_c20190523.nc', 'SO2 -> /glade/u/home/visioni/emis/emissions_ssp585/emissions-cmip6_SO2_contvolcano_vertical_2040climo_0.9x1.25_c20190523.nc', 'so4_a1 -> /glade/u/home/visioni/emis/emissions_ssp585/emissions-cmip6_so4_a1_contvolcano_vertical_2040climo_0.9x1.25_c20190523.nc', 'so4_a2 -> /glade/u/home/visioni/emis/emissions_ssp585/emissions-cmip6_so4_a2_contvolcano_vertical_2040climo_0.9x1.25_c20190523.nc'

For surface emissions: we need to add the "other" emissions for CO, NO, not DMS Here, I think, 1deg emissions are also fine.

We also can use the 2deg SOAG emissions I just created for you for antho and biomass burning, but keep the 1deg biogenic emissions.

dan-visioni commented 1 year ago

Ok, added. Just to clarify and for future reference:

we added continuously outgassing volcanoes because WACCM has them also for the future (unlike most models?)
Why not DMS?
Why keep the 1deg biogenic emissions even if I can see that file for 2deg in the emissions you created?

tilmes commented 1 year ago

@dan-visioni here are some comments: Continuous outgasing volcanoes are added because we assume they continue to contribute to the tropospheric sulfate burden. The other DMS emissions are now replaced by the new OASIS emission scheme that uses IA to derive the emissions amount. However, I just noticed, that we still need to change the year. It is currently set to the year 2000 but needs to be changed to 2040.

&ocean_emis_nl bubble_mediated_transfer = .FALSE. csw_cycle_yr = 2000 csw_specifier = 'DMS -> /glade/work/siyuan/Csw_DMS_Lana2011_f09f09_1750_2100_20200717a.nc' csw_time_type = 'CYCLICAL' ocean_salinity_file = '/glade/work/siyuan/SSS_recooked_0-360.nc' /

1deg biogenic emissions: I created 1deg biomass burning and anthro emissions, but not biogenic emissions, but if I miss something, can you point me to them? Your namelist still has to be updated to include the 2deg SOAG emissions.

dan-visioni commented 1 year ago

Updated the ocean to 2040. I can see the SOAG emissions in the last three rows of srf_emis_specifier - I double checked with the 1deg case and they are there, I added them yesterday. Or am I missing something else?

Both cases (1deg and 2deg) now have:

2 CH2O, 3 CO, 3 NO, 1 DMS (because of what you explained, this is different), 2 bc_a4, 2 num_a1 from so4, 1 num_a2 from so4, 4 num_a4 from bc and pom, 2 pom_a4, 3 SO₂, 3 SOAG, 2 so4_a1 (one of them was missing, the bb) and 1 so4_a2. I can't see anything else.

tilmes commented 1 year ago

OK great, did you use the SOAG 2deg emissions I produced? If so then it should be all ready to go!

dan-visioni commented 1 year ago

I did, yes. I'll give it a go!

dan-visioni commented 1 year ago

Running into a problem already when building the case: I get this error Variable name csw_time_cycle_yr not found in /glade/work/fvitt/camdev/carma_trop_strat/bld/namelist_files/namelist_definition.xml

But I only have

csw_specifier = 'DMS -> /glade/work/siyuan/Csw_DMS_Lana2011_f09f09_1750_2100_20200717a.nc' csw_time_type = 'CYCLICAL' csw_cycle_yr = 2040

which are correct. What am I doing wrong?

tilmes commented 1 year ago

@fvitt could you help with Daniele's problem, why is this not building: csw_specifier = 'DMS -> /glade/work/siyuan/Csw_DMS_Lana2011_f09f09_1750_2100_20200717a.nc' csw_time_type = 'CYCLICAL' csw_cycle_yr = 2040 this is the case: /glade/work/visioni/cases/FW2000climoMA.f19_f19_mg17.Control.001

fvitt commented 1 year ago

@dan-visioni I was able to clone and build your case without problems.
Maybe you can try the build again:

  case.build --clean-all
  qcmd -- case.build

dan-visioni commented 1 year ago

Thans @fvitt, I had tried that already and re-tried but I always get the same error. While running the build, this is what it says

ERROR: /glade/work/fvitt/camdev/carma_trop_strat/cime/src/build_scripts/buildlib.mct FAILED

Could it be a compiler problem? Currently I have:

1) ncarenv/1.2 (H) 2) intel/17.0.1 3) ncarcompilers/0.4.1 (H) 4) mpt/2.18 (H) 9) python/3.7.12

fvitt commented 1 year ago

I am using python/3.7.9. Don't know if that makes a difference.

You didn't create the case or try to build from casper did you?

Do you have plenty of disk space on /glade/scratch? Use gladequota to check.

dan-visioni commented 1 year ago

I have 40 TB of storage left and no, I'm from Cheyenne. I tried loading your version of Python but hasn't changed the results. The glade/work/visioni/cases/FW2000climoMA.f19_f19_mg17.Control.001/CaseStatus always gives me the same erro /glade/scratch/visioni/FW2000climoMA.f19_f19_mg17.Control.001/bld/mct.bldlog.230425-111124 also not helpful..!

tilmes commented 1 year ago

@dan-visioni can you try to setup a completely new case, and try again?

dan-visioni commented 1 year ago

I have tried already (fe21.FW2000climoMA.f19_f19_mg17.Control.001) but I got the same problem. I have been cloning your case, so now I'll try setting up a new one...

dan-visioni commented 1 year ago

Ok, started a new case without cloning and it seems to have worked. Case is f21.FW2000climoMA.f19_f19_mg17.Control.001 I'll submit 1 year and check it tomorrow morning.

dan-visioni commented 1 year ago

3:(shr_dmodel_readstrm) ERROR in data sizes 288 192 1812 64800 3: ERROR: (shr_dmodel_readstrm) ERROR in data sizes

I have this error I've never seen before, but it looks like it's because of one of the forcing files having the wrong size?

tilmes commented 1 year ago

@dan-visioni oh, I missed that part, you changed the SSTs: /glade/work/visioni/SST/sstice_yaga_1988_2007_avg_0.9x1.5_1950-2100.nc but this file does not seem to work with this file: entry id="SSTICE_GRID_FILENAME" value="$DIN_LOC_ROOT/ocn/docn7/domain.ocn.1x1.111007.nc even though you have used it in your other version.

You can try using this instead (in env_run.xml) entry id="SSTICE_GRID_FILENAME" value="$DIN_LOC_ROOT/share/domains/domain.ocn.fv0.9x1.25_gx1v7.151020.nc"

dan-visioni commented 1 year ago

It seems like this case is being hostile to me at every turn. Now the case runs but crashes immediately with no apparent error, but with the following warning in the CESM log:

197: imp_sol: time step 1800.000 failed to converge @ (lchnk,vctrpos,nstep) = 882 22 0 WARNING: Aerosol optical depth is unreasonably high in this layer.:

See log at /glade/scratch/visioni/f21.FW2000climoMA.f19_f19_mg17.Control.001/run/cesm.log.9811691.chadmin1.ib0.cheyenne.ucar.edu.230426-072800

tilmes commented 1 year ago

Yes, it seems like it... but this is how modeling is :)

I looked carefully again through the namelist, and noticed that you have this line (sorry I was not so careful before). It should point to the num_so4_a1 field instead of so4_a1. Maybe that is the problem. 'num_a1 -> /glade/work/visioni/inputdata/emis/ScenarioMIP_IAMC-REMIND-MAGPIE-ssp585-1-1/emissions-cmip6-ScenarioMIP_IAMC-REMIND-MAGPIE-ssp585-1-1_so4_a1_ene_vertical_anthro_surface_mol_201501-210012_1.9x2.5_c20230310.nc',

another one I found, here we need to point to so4_a1 not num_so4_a1: 'so4_a1 -> /glade/work/visioni/inputdata/emis/ScenarioMIP_IAMC-REMIND-MAGPIE-ssp585-1-1/emissions-cmip6-ScenarioMIP_IAMC-REMIND-MAGPIE-ssp585-1-1_num_so4_a1_biomassburning_surface_mol_201501-210012_1.9x2.5_c20230323.nc',

finally (not important for this but just to be more consistent): can you use: /glade/scratch/tilmes/Emissions_CMIP6_ssp/regrid_mol/ScenarioMIP_IAMC-REMIND-MAGPIE-ssp585-1-1/emis/emissions-cmip6-ScenarioMIP_IAMC-REMIND-MAGPIE-ssp585-1-1_SOAG_anthro_surface_mol_201501-210012_1.9x2.5_c20230310.nc and emissions-cmip6-ScenarioMIP_IAMC-REMIND-MAGPIE-ssp585-1-1_SOAG_biomassburning_surface_mol_201501-210012_1.9x2.5_c20230323.nc for SOAG anthro and biomass burning. I mentioned that before, but did not clarify what I meant...

fvitt commented 1 year ago

You are using a rather old tag of the CARMA development branch carma_trop_strat09. The latest tag is carma_trop_strat12. This run does not use CARMA. So why not use a CAM tag?

dan-visioni commented 1 year ago

@fvitt I was using the same tag as /glade/p/cesm/chwg_dev/tilmes/cases/carma/FWmaMAM4HIST.f19_f19_mg17.carma_trop_strat09.1980_2014 from which I was taking the restart files. Which one do you suggest I use?

tilmes commented 1 year ago

@dan-visioni I think, we want to be consistent with what we used for ISAMIP, since this is evaluated and also used for the CARMA paper. Is the tag you are using the same as for the ISAMIP runs?

dan-visioni commented 1 year ago

Yes, this is the same tag I used for ISAMIP.

tilmes commented 1 year ago

@dan-visioni I also noticed that you need to add this as well: emissions-cmip6-ScenarioMIP_IAMC-REMIND-MAGPIE-ssp585-1-1_num_so4_a1_ship_fef_anthro_surface_mol_201501-210012_1.9x2.5_c20230310.nc Number so4_a1 for anthropogenic emissions is split into 4 files, while so4_a1 only has 3 files for anthro. We split the number up in more categories, I am not sure why, but we need to add all 4 files for number.

dan-visioni commented 1 year ago

A lingering error is that I do not have so4_a1 for those emissions. Both in my emission folder and in yours I have:

emissions-cmip6-ScenarioMIP_IAMC-REMIND-MAGPIE-ssp585-1-1_num_so4_a1_ag_anthro_surface_mol_201501-210012_1.9x2.5_c20230310.nc emissions-cmip6-ScenarioMIP_IAMC-REMIND-MAGPIE-ssp585-1-1_num_so4_a1_biomassburning_surface_mol_201501-210012_1.9x2.5_c20230323.nc emissions-cmip6-ScenarioMIP_IAMC-REMIND-MAGPIE-ssp585-1-1_num_so4_a1_ene_vertical_anthro_surface_mol_201501-210012_1.9x2.5_c20230310.nc emissions-cmip6-ScenarioMIP_IAMC-REMIND-MAGPIE-ssp585-1-1_num_so4_a1_ship_fef_anthro_surface_mol_201501-210012_1.9x2.5_c20230310.nc emissions-cmip6-ScenarioMIP_IAMC-REMIND-MAGPIE-ssp585-1-1_so4_a1_ag_ship_fef_anthro_surface_mol_201501-210012_1.9x2.5_c20230310.nc emissions-cmip6-ScenarioMIP_IAMC-REMIND-MAGPIE-ssp585-1-1_so4_a1_ene_vertical_anthro_surface_mol_201501-210012_1.9x2.5_c20230310.nc

so the so4_a1 files are missing, while the num_so4_a1 are there. But I think I need both, one for the mass and one for the number. So the case can't find them.

tilmes commented 1 year ago

@dan-visioni your user_nl_cam file points to this directory for some emissions: /glade/u/home/visioni/inputdata/emis/emissions_ssp585/ But I don't think that exists, you need to point to your work directory, I think, the emissions exist there. Hope that works!

dan-visioni commented 1 year ago

Yes, those were wrong, sorry. Still, the so4 files are still missing from the correct folder.

tilmes commented 1 year ago

@dan-visioni I found the so4 emissions in your folder:

Now I see the file that is wrong is this one: it shows so4_a4, but should be called so4_a1. emissions-cmip6-ScenarioMIP_IAMC-REMIND-MAGPIE-ssp585-1-1_so4_a4_biomassburning_surface_mol_201501-210012_1.9x2.5_c20230323.nc I just created a new one for you in my directory!

Besides that, as a note, we have one more number files than mmr files, because the for ag_ship_fef number was split into two files. I am going to create an updated file.

Also, in your namelist you need to change these, that have the home location.

Again, I hope that is it, if not, I will clone your case and make it run!

dan-visioni commented 1 year ago

All issues with the namelist seem to be fixed (in terms of where things are...) but now the problem is: FIND_TIMES: ALL data times are after 0.000000000000000E+000 All flags are set to "CYCLICAL" so unsure where this is coming from

tilmes commented 1 year ago

@dan-visioni I finally found it going through all the emissions. You added vertical SO4 emissions to the surface emissions. However, they are already in vertical emissions and should not be added there. After removing them, it worked for me, see: /glade/p/cesm/chwg_dev/tilmes/cases/carma/f21.FW2000climoMA.f19_f19_mg17.Control.test (for 5 days)!

dan-visioni commented 1 year ago

For some reason my user_nl_cam wasn't working, but copying yours did. This drove me crazy today. Now, finally running...

dan-visioni commented 1 year ago

Checking now the first 10 years of simulation. We need Chlorine levels to go down from 2000 to 2040 levels, which is happening - see plot but yes, wrong axis - (and ozone is responding accordingly). I'll run 10 more years and check again on Monday. For the 1deg version, we checked the QBO but here it's nudged so there's no point. Other suggestions on things worth checking to make sure this baseline is correct?

tilmes commented 1 year ago

It looks like you used this IC file for the earlier run: /glade/work/visioni/inputdata/b.e21.BWSSP585cmip6.f09_g17.CMIP6-SSP5-8.5-WACCM.001_SOAG.cam.i.2040-01-01-00000_c190523.nc which had warmer temperatures and therefore needed a spinup, correct? Here, we start with temperatures from the year 2000 (using an IC from 2000) but need to get the GHGs down to 2040 levels. Given that we changed other things like DMS (which needs spinup) it may probably better to continue and not use the 2040 IC file. so, yes, I agree, to continue, I would check temperature changes, and also other long-lived GHGs, CFC evolution, N2O, to make sure they reach the 2040 levels (like in the 1deg run).

dan-visioni commented 1 year ago

Created the cases for the SO₂/H2SO4 runs for MAM, based on /glade/work/visioni/cases/f21.FW2000climoMA.f19_f19_mg17.Control.001

Starting only f21.FW2000climoMA.f19_f19_mg17.SO2_5Tg_region.001 and I'll check after a couple of years

For CARMA, I'll ask again just to make sure: which case should we start the Control one from? The current ISAMIP run?

tilmes commented 1 year ago

Let's wait for the CARMA... we want to have Francis finish the new tag with Chuck's changes first.

tilmes commented 1 year ago

@fvitt please let us know when you created a new branch. @dan-visioni we probably want to add some more diagnostics, like the different optics diagnostics we have been looking at and the double radiation call.

dan-visioni commented 1 year ago

All MAM injection cases are here: /glade/scratch/visioni/archive/ (case directory in work/visioni/cases/ Control: f21.FW2000climoMA.f19_f19_mg17.Control.001 Example case: f21.FW2000climoMA.f19_f19_mg17.SO2_5Tg_region.001 Cases can be either SO₂ or H2SO4, 5Tg or 25Tg, and region (uniform injection 30N-30S, 18-20km) or points (point injections at 30N or 30S, 19km).

Jupiter Notebook checking the runs is in u/home/visioni/chek_runs-H2SO4.ipynb

H2SO4 still missing (are running now), will add next week as soon as I can.

tilmes commented 10 months ago

@dan-visioni I am trying to find your jupiter notebook, it is not in /glade/u/home/visioni could you check?

tilmes commented 10 months ago

@dan-visioni Could you remind me, do we really need 50 years for the control run? I am planning to start the CARMA run, and just want to make sure that it is needed...I could also start in 2020, 10 year spinup seems to be enough? Also, in our setup (transient) while emissions are set to 2040, land use is transient. I just wondered if that matters, but I assume it does not, do you agree?

dan-visioni commented 10 months ago

@tilmes yes, the notebook is in u/home/visioni/chek_runs-H2SO4.ipynb I think we ran 50 years just to be on the safe side, as ozone needs to balance. I think we can run 10 years and just check? At the time, I checked chlorine load, stratospheric temperatures and ozone. Good question about land use tho, that might mess up things... I don't know, we should talk about it tomorrow!

NCAR / CARMA_dev

MAM run for H2SO4 injections #24