Scheduled Downtime
On Tuesday 24 October 2023 @ 5pm MT the forums will be in read only mode in preparation for the downtime. On Wednesday 25 October 2023 @ 5am MT, this website will be down for maintenance and expected to return online later in the morning.
Normal Operations
The forums are back online with normal operations. If you notice any issues or errors related to the forums, please reach out to help@ucar.edu

FAQ: Data ocean slab mode (DOCN-SOM)

dbailey

CSEG and Liaisons
Staff member
For mixed-layer depth, I use the variable HBLT which is the turbulent boundary layer depth. That might explain the difference.
 

cpatrizio

Casey Patrizio
New Member
OK thanks, I will check that.

The plot on the right does seem more realistic to me, and agrees better with some observational estimates of the mixed-layer depth. So it makes me wonder if the mixed-layer depth is underestimated in the Q-flux forcing file, or if there was another reason for that choice...
 

dbailey

CSEG and Liaisons
Staff member
The actual choice of mixed-layer depth is not that important as long as the Q-flux represents the heat divergence/convergence into that mixed-layer.
 

cpatrizio

Casey Patrizio
New Member
For the mean climate I agree with the above, but I think it may be important for climate variability. For example, if the mixed-layer is too shallow the SST variance will be unrealistically high.
 

weihaoguo

New Member
Hi, Forums!
I am trying to add the SST anomalies in the tropical Pacific to check the responses of other areas by slab ocean model. Therefore, i prefer that the tropical Pacific just plays as the forcing role, but the coupling is carried out in other areas. This step is necessary in "pacemaker" experiments in a full-couple model. How can i achieve it in the slab ocean model?
 

edu061

Etienne Dunn-Sigouin
New Member
Hi,

I would like to run an E compset simulation using a SOM forcing file calculated from a fully coupled simulation similar to the preindustrial slab ocean control run from LENS (E1850C5CN.f09_g16). My end goal is to modify the SOM forcing file with idealized perturbations and run new experiments.

If I understand correctly, one way would be to run the compset E1850C5CN forced with a SOM file posted in the link in the FAQ (pop_frc.b.e11.B1850C5CN.f09_g16.005.082914.nc). This approach means I don't need to run my own fully coupled simulation to generate the SOM forcing file and it allows me to check if I set up everything correctly by comparing with the slab ocean control run from LENS.

I have a few questions about this:

1) Is my reasoning correct ?

2) I can't find the 'recommended' SOM file forcing ending with '150217' derived from ~1000 years of data from the FAQ link. It seems you can get it from an NCAR machine at /glade/p/cesmdata/cseg/inputdata/ocn/docn7/SOM/, but how do I gain access to this ?

3) My set up is CESM2.1.0. The most similar compset to E1850C5CN seems to be E1850TEST which uses CAM6 not CAM5. I tried to generate the E1850C5CN case with the long name 1850_CAM50_CLM40%CN_CICE_DOCN%SOM_RTM_SGLC_SWAV but I get an error when building: the use case '1850_cam5' is invalid. Is there any way to generate the E1850C5CN case with CESM2 ?

4) This is likely not a good idea but can I use the available SOM forcing files for CAM5 E1850C5CN to force CAM6 E1850TEST ?

5) Finally, is there an available SOM forcing file that I can use to force the CAM6 E1850TEST case ?

Thanks for your help,
Etienne
 

dbailey

CSEG and Liaisons
Staff member
This is basically correct. We can provide the SOM forcing file. This is from the CESM1-LENS control, so not quite the same as for CESM2. However, we should probably produce another SOM forcing file from the CESM2 piControl run. In terms of compsets, the short names have gone away for CESM2 for SOM cases. It is probably better to use the long name. I think the problem above is that it should be CLM50. I will find the appropriate long name. Do you actually want to use CAM5 instead of CAM6 here?

Dave
 

edu061

Etienne Dunn-Sigouin
New Member
Hi Dave,

Thanks for the rapid response. Ideally I would like to use CAM5. However, if a SOM forcing file is available for the CESM2 piControl run, then that could work also.

Could it just be easier to set up CESM1 and run 1850_CAM50_CLM40%CN_CICE_DOCN%SOM_RTM_SGLC_SWAV with the available forcing files ? Otherwise, if there is an equivalent long name in CESM2, that would be great.

Thanks,
Etienne
 

dbailey

CSEG and Liaisons
Staff member
If you really want to use CAM5, then sticking with CESM1 is the best idea. However, the problem is that CESM1 is not necessarily supported on a lot of machines.
 

minminfu

Member
Hi Dave,

Would the b.e11.B1850C5CN.f09_g16.005 file work for both E1850C5 and E1850C5CN? Also was that file computed from a run using CLM5.0? I noticed the E compsets in CESM1 use CLM4 by default. More generally, how critical is it that all components match exactly between the b compset that the file comes from and the e compset the file is being used to force?

Best,
Minmin
 

dbailey

CSEG and Liaisons
Staff member
In CESM1.1 the CLM version is CLM4. This is what was used for CESM1-LENS. In terms of CN (carbon-nitrogen) on or off, this SOM forcing should still be fine for both. Generally, the SOM forcing can work well with many differences. For example, we typically do simulations with 1850 CO2 levels and then double the CO2 levels and the SOM forcing remains fixed. There was some work out of Livermore where they did an uncertainty quantification and made many physics/parameter changes within CAM. Out of 50 or so experiments, only one did not remain stable.
 

minminfu

Member
In CESM1.1 the CLM version is CLM4. This is what was used for CESM1-LENS. In terms of CN (carbon-nitrogen) on or off, this SOM forcing should still be fine for both. Generally, the SOM forcing can work well with many differences. For example, we typically do simulations with 1850 CO2 levels and then double the CO2 levels and the SOM forcing remains fixed. There was some work out of Livermore where they did an uncertainty quantification and made many physics/parameter changes within CAM. Out of 50 or so experiments, only one did not remain stable.
Thank you for the explanation, this was very helpful.
 
Hi Dave,

I would like to conduct fixed-SST(AMIP type) and coupled (CMIP type) simulations using CESM1_0_4. For the CMIP type simulations, I plan to run a E (E_2000) compset simulation using a SOM forcing file with CAM5 physics package; and for the AMIP type simulations, I plan to run a FC5 (F_2000_CAM5) compset simulation with fixed-SST forcing.

However, I found that the E compset simulation can work with the f09_g16 resolution, but the FC5 simulation we have already finished is at f09_f09 resolution.

Here are my questions:

1) Can the simulation results using f09_g16 resolution can be analyzed and compared with the simulation results using the f09_f09 resolution?

2)Is it feasible/scientifically appropriate to use E and FC5 compsets to do CMIP and AMIP type simulations?
There's also a FAMIPC5 compset with description: 'AMIP run for CMIP5 protocol with cam5'. I'm not sure whether this compset (FAMIPC5) is appropriate
for the AMIP type simulation.

Thanks in advance for your help,

Thanks,
Alison
 

dbailey

CSEG and Liaisons
Staff member
If I understand correctly, this is kind of comparing apples to oranges. The F compset uses specified SST/ICEFRAC from observations, while the sea ice/sst in a SOM run is prognostic. If you are comparing the atmosphere in these, then they are both f09. However, I don't see what you would gain by comparing the SST/ICEFRAC between these two. What you could do instead is derive SST/ICEFRAC from a fully coupled run. This would be from the same run that the SOM forcing was derived from. Then you can actually run the F compset at f09_g16 and read in the model SST/ICEFRAC. This would be more analogous to the SOM run in my mind.
 
If I understand correctly, this is kind of comparing apples to oranges. The F compset uses specified SST/ICEFRAC from observations, while the sea ice/sst in a SOM run is prognostic. If you are comparing the atmosphere in these, then they are both f09. However, I don't see what you would gain by comparing the SST/ICEFRAC between these two. What you could do instead is derive SST/ICEFRAC from a fully coupled run. This would be from the same run that the SOM forcing was derived from. Then you can actually run the F compset at f09_g16 and read in the model SST/ICEFRAC. This would be more analogous to the SOM run in my mind.
Thanks for your reply!
What we want to do is to diagnose the fast and slow climatic responses to anthropogenic forcings through doing the F-compset (AMIP) and E-compset (CMIP) simulations.
The articles we have read indicate that the SST data used in AMIP simulation could either be a prescribed climatological data or comes from the outputs of the coupled simulation (B- or E-compset).
Moreover, if I understand it correctly, the simulation results with f09_g16 resolution(e.g. global surface temperature) can be analyzed and compared with the simulation results with f09_f09 resolution if we convert the resolution of the former into f09_f09 resoution. Am I right?

Thanks,
Alison
 

dbailey

CSEG and Liaisons
Staff member
Correct, the DOCN can read in SST/IFRAC data from a fully-coupled (B) or SOM (E) run. However, I would recommend running the F compset with the f09_g16 resolution. Then there is no need to interpolate the SST/IFRAC data. However, if you do run it with f09_f09 or f09_g16, the atmospheric variables will be comparable regardless.
 
Thanks! Dave!
I really appreciate your help and I'm thinking about rerunning the AMIP type simulation.
So there's another quick question. As I understand it, FC5 and FAMIPC5 compsets are both scientifically appropriate for the AMIP type simulation; and which one to use, however, depends on our research objectives. Am I right?
Thank you again for your reply!

Best,
Alison
 
Top