search

NCAR / amwg_dev

Repo to store model sandboxes and cases used for CAM development
9 stars 2 forks source link

b.cesm3_cam058_mom.B1850WscMOM.ne30_L58_t061.007 #129

Open cecilehannay opened 2 years ago

cecilehannay commented 2 years ago

Description:

The purpose of this run is to test the salt flux limiter implemented in CICE6 (based on CICE5 salt flux limiter).

The run will use the models: MOM6 + CICE6 + cam_dev.

We moved to a new sandbox that will include Adam's branch (https://github.com/adamrher/CAM.git, branch cam6_3_058.dtsens) for the atmosphere model instead of cam6_3_041.

Case directory: Locally (if still available): /glade/p/cesmdata/cseg/runs/cesm2_0/b.cesm3_cam058_mom.B1850WscMOM.ne30_L58_t061.007

On github: https://github.com/NCAR/amwg_dev/tree/b.cesm3_cam058_mom.B1850WscMOM.ne30_L58_t061.007

Sandbox: Locally (if still available): /glade/work/hannay/cesm_tags/cesm3_cam6_3_058_MOM

On github: https://github.com/NCAR/amwg_dev/tree/cesm3_cam6_3_058_MOM hash: [73d5bb3])

Diagnostics: AMWG diags (if available) https://webext.cgd.ucar.edu/BWsc1850MOM/b.cesm3_cam058_mom.B1850WscMOM.ne30_L58_t061.007/atm/

Contacts: @cecilehannay and @gustavo-marques

jedwards4b commented 2 years ago

The github location of this sandbox is https://github.com/ESCOMP/CESM branch cesm3_coupled_test commit 73d5bb35d9

justin-richling commented 2 years ago

ADF diagnostics have been uploaded here.

gustavo-marques commented 2 years ago

The simulation completed 20 years. Salt is conserved within the ocean model. As in run 008 (https://github.com/NCAR/amwg_dev/issues/130), convection in the Lab Sea shuts off after ~ 6 years (see plot below). image

In terms of salt conservation in the ice model, I noticed that the "arwt salt flx error" is always (I only looked at the last log file) much larger in the Arctic when compared to the Antarctic (see an example below). Perhaps this is because sea ice is accumulating in the Arctic. @dabail10, any thoughts? 

                       Arctic                 Antarctic

arwt salt flx error = -4.34887567972258312E-07 2.77767027356864325E-17

dabail10 commented 2 years ago

This is concerning. I wonder if it is related to the 4psu "fix". I haven't seen this difference in the errors previously. The Lab Sea freezing over is another issue which is going to be trickier.

adamrher commented 2 years ago

What has RESTOM looked like in all these b-case runs using this new code base cesm3_cam6_3_058_MOM? I just found out that the atmosphere is colder in this code base, compared to earlier ones. I found out why, and can fix it so it recovers a climate we have been using. Don't get your hopes up that this will stop the Lab sea from freezing, but it's moving in the right direction. Thoughts?

gustavo-marques commented 2 years ago

My hopes that making the atmosphere warmer will stop the Lab sea from freezing are very high.

cecilehannay commented 2 years ago

@adamrher: It looks like RESTOM is missing form the ADF diags. I will post the AMWG diags so we can look at RESTOM.

cecilehannay commented 2 years ago

But from the corresponding F cases (#112 and #132), I expect the RESTOM to be lower. It is 1 W/m2 colder in the B case.

JulioTBacmeister commented 2 years ago

The numbers I posted are from some IDL scripts I have run for 10 years from the h0 files. The numbers from my scripts agreed in the past with those from AMWG. FLNT and FSNT agree with ADF for 007. I think the global numbers for FSNT and FLNT should be written with more digits in ADF - they are currently format NNN maybe should be NNN.nn

adamrher commented 2 years ago

Julio showed here that this cesm3_cam6_3_058_MOM code base seems to result in about a 0.5 W/m2 reduction in RESTOM, compared to the code we were using before.

I have updated my CAM branch to restore this previously warmer climate. @jedwards4b @cecilehannay the new CAM hash to add to your Externals.cfg is 7e6a845.

jedwards4b commented 2 years ago

I have updated https://github.com/ESCOMP/CESM/ branch cesm3_cam6_3_058_MOM with the new cam tag.

cecilehannay commented 2 years ago

I can do a F case with the new tag.

On Wed, Jun 15, 2022 at 3:49 PM Adam Herrington @.***> wrote:

Julio showed here https://github.com/NCAR/amwg_dev/issues/132#issuecomment-1156912581 that this cesm3_cam6_3_058_MOM https://github.com/NCAR/amwg_dev/tree/cesm3_cam6_3_058_MOM code base seems to result in about a 0.5 W/m2 reduction in RESTOM, compared to the code we were using before.

I have updated my CAM branch to restore this previously warmer climate. @jedwards4b https://github.com/jedwards4b @cecilehannay https://github.com/cecilehannay the new CAM hash to add to your Externals.cfg is 7e6a845 https://github.com/adamrher/CAM/commit/7e6a845779cf72fc01399d8f602fcbf4ef262b1c .

— Reply to this email directly, view it on GitHub https://github.com/NCAR/amwg_dev/issues/129#issuecomment-1156980669, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACKF2VAM7CKXXD7F5GRZXKLVPJFXBANCNFSM5YA7TEAA . You are receiving this because you were mentioned.Message ID: @.***>

adamrher commented 2 years ago

Should we do a B case? To see impacts on Lab sea? Or do we want to understand the salt issues a bit more first?

cecilehannay commented 2 years ago

I was suggesting a F case to make sure the climate is the same as in cam6_3_041 but maybe you check this already.

JulioTBacmeister commented 2 years ago

Adam could extend his F-case with revice, while Cecile starts a B-case.

cecilehannay commented 2 years ago

Yes, but one thing is that I was going to do the F case with the same sandbox as the one we will use for the B case. I know in theory, cam should be the same in Adam's sandbox and mine, but...

I am going to meet with Gustavo to talk about the B case.

adamrher commented 2 years ago

I'd say lets forgo the F case until we want to do another F case experiment, at which point we can redo the F case control.

Maybe just do a new B case, and check after 10 years whether the climate is similar to the older B case runs with the older sandbox.

cecilehannay commented 2 years ago

I just talked with @gustavo-marques and the salinity issues are not solved yet. They need to look at this more.

JulioTBacmeister commented 2 years ago

Agree. If the new B case is ready to go let’s start it.

I say use the old topography for this.

Then let’s swap topo if it looks ok

On Wed, Jun 15, 2022 at 5:00 PM Adam Herrington @.***> wrote:

I'd say lets forgo the F case until we want to do another F case experiment, at which point we can redo the F case control.

Maybe just do a new B case, and check after 10 years whether the climate is similar to the older B case runs with the older sandbox.

— Reply to this email directly, view it on GitHub https://github.com/NCAR/amwg_dev/issues/129#issuecomment-1157028968, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACGLMTQ4ADS3BOIWJF7MNUTVPJOBPANCNFSM5YA7TEAA . You are receiving this because you commented.Message ID: @.***>

gustavo-marques commented 2 years ago

I suggest running an F case using the updated sandbox (cesm3_cam6_3_058_MOM) to make sure we are getting the same climate as before. In the meantime, I will look into the salinity issue in CICE with @dabail10. We need to understand and solve this issue before running a new B case.

JulioTBacmeister commented 2 years ago

Is this an issue with CICE6?

If it is I’d consider a run with Adam’s cam058+revice+cice5

On Wed, Jun 15, 2022 at 5:12 PM Gustavo Marques @.***> wrote:

I suggest running an F case using the updated sandbox (cesm3_cam6_3_058_MOM) to make sure we are getting the same climate as before. In the meantime, I will look into the salinity issue in CICE with @dabail10 https://github.com/dabail10. We need to understand and solve this issue before running a new B case.

— Reply to this email directly, view it on GitHub https://github.com/NCAR/amwg_dev/issues/129#issuecomment-1157061381, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACGLMTQCGFOTAQYQEM6XJOTVPJPMNANCNFSM5YA7TEAA . You are receiving this because you commented.Message ID: @.***>

gustavo-marques commented 2 years ago

It's an issue with both CICE5 and CICE6. Both versions use the hack that avoids CICE from taking more salt that's available in the ocean model.