HR CAM7 simulations - version 2

[cecilehannay]

Discussion #585 has become quite messy, so I am locking and start a new discussion summarizing where we stand.

---

We are planning a series of CAM7 simulations varying both horizontal and vertical resolutions. Specifically, we will perform AMIP simulations with observed sea surface temperature (SST) data and SST+4K at 1° and 0.25° resolutions, with both the CAM6 and CAM7 grids (i.e, with 32, 58, 93 vertical levels).

These simulations will be used to investigate the impact of both vertical and horizontal resolution on climate sensitivity, regional precipitation patterns, gravity waves, boundary layer aerosol-cloud interactions and hurricane activity

---

Configuration

- Run type: AMIP type for years 2000-2010
- Tag: cam6_4_036
- Output: We collected the request for output here
Output will be found:

/glade/campaign/cesm/community/amwg/cam_hr

---

List of runs

10 yr = 2000-2010
5yr = 2000-2005

Physics	Res	Levels	Experiment	Period	Comments	Issue
cam7	1 deg	L32	SST	10 yr	Done	#596
			SST+4K	5 yr	Done	#597
		L58	SST	10 yr	Done	#598
			SST+4K	5 yr	Done	#599
		L93	SST	10 yr	Done	#600
			SST+4K	5 yr	Done	#601
	0.25 deg	L32	SST	10 yr	Done	#593
			SST+4K	5 yr	Done	#595
		L58	SST	10 yr	Done	#591
		L93	SST	10 yr	Done	#590
			SST+4K	5 yr	Done	#594
prog mom off	0.25 deg	L93	SST	5 yr	Done	#592
cam5	1 deg	L32	SST	5 yr	checking cam5	#603
		L58	SST	5 yr	tuning
		L93	SST	5 yr	does cam5 L93 work?
	0.25 deg	L58	SST	5 yr	5 or 10 yrs ? L58 or L93?
			SST+4K	5 yr

@JulioTBacmeister, @adamrher, @islasimpson, @swrneale, @PeterHjortLauritzen, @cmcclus, @bstephens82, @megandevlan, @jedwards4b, @billsack, @wwieder, @ekluzek, @shieldsca, @slevis-lmwg,

[cecilehannay]

Estimate of the cost and output size

---

fincl

The variable and fincl are discussed here:
https://docs.google.com/document/d/1T9T8Axz1L58G0YB__QNXsx5u4rpUHEhXdFFcVQP6uJY/edit

---

Cost estimate

• Model Cost: 305201.25 pe-hrs/simulated_year
• Model Throughput: 0.48 simulated_years/day

---

Output estimate per month

On interpolated ne120 grid

• monthly - average on native grid for ADF => 2.7G
• daily - average on native grid for MDTF => 1.2G

On native grid

• monthly - average => 21G
  • our usual monthly output (15G)
  • short list from Christina (6.6G)
• monthly - average => 556K
  • the usual zm variables (does this work?)
• daily - average => 1.4 G
  • our usual daily average
• 6 hourly - instantaneous => 374G
  • Julio’s list (202G)
  • short list from Ben (172G)
• 3 hourly - average => 1.5G
  • Precip
• 3 hourly instantaneous => 15G
  • TC and AR request (from Ben, Julio Christine)

Output estimate per year:

• without the 6 hourly instantaneous: 43.3G*12 = 520G per year
• plus the 6 hourly instantaneous: 374G*12 = 4368G per year
• Land = 15G*12 = 180G per year
• TOTAL = about 5T per year

[cecilehannay]

We are ready to start the simulation.

• I did a cost estimate

Model Cost: 305201.25 pe-hrs/simulated_year
Model Throughput: 0.48 simulated_years/day

• and an output estimate

about 5T per year

---

Please check the output to ensure that the variables you need are being saved and at the correct frequency..

You can either:

• review the fincl in the googledoc here
• or directly loolk at the output on glade:

/glade/derecho/scratch/hannay/archive/f.e30_cam6_4_036.FMTHIST.ne120_L93.001_t4

[cecilehannay]

Land Setting

@adamrher created land files that I will use for this run:

flanduse_timeseries = '/glade/derecho/scratch/aherring/restart/ctsm5.2.028/tools/mksurfdata_esmf/landuse.timeseries_ne120pg3_SSP3-7.0_1979-2026_78pfts_c240925.nc' 
fsurdat = '/glade/derecho/scratch/aherring/restart/ctsm5.2.028/tools/mksurfdata_esmf/surfdata_ne120pg3_SSP3-7.0_1979_78pfts_c240925.nc'

[cecilehannay]

for the ne30 equivalent, @adamrher created a fsurdat to fix soil data.

'/glade/derecho/scratch/aherring/restart/ctsm5.2.028/tools/mksurfdata_esmf/surfdata_ne30pg3_SSP3-7.0_1979_78pfts_c241014.nc'
'/glade/derecho/scratch/aherring/restart/ctsm5.2.028/tools/mksurfdata_esmf/landuse.timeseries_ne30pg3_SSP3-7.0_1979-2026_78pfts_c241014.nc'

As a result, the ne30 runs use the following datasets:

finidat = '/glade/campaign/cesm/cesmdata/inputdata/lnd/clm2/initdata_esmf/ctsm5.3/ctsm52018_f09_xsiceON_acON_pSASU.clm2.r.0201-01-01-00000.nc'
fsurdat = '/glade/derecho/scratch/aherring/restart/ctsm5.2.028/tools/mksurfdata_esmf/surfdata_ne30pg3_SSP3-7.0_1979_78pfts_c241014.nc'
flanduse_timeseries = '/glade/derecho/scratch/aherring/restart/ctsm5.2.028/tools/mksurfdata_esmf/landuse.timeseries_ne30pg3_SSP3-7.0_1979-2026_78pfts_c241014.nc'

[slevis-lmwg]

I want to point out that @olyson has generated a new spun up finidat file for ne30, specific to 1979:
/glade/campaign/cesm/cesmdata/inputdata/lnd/clm2/initdata_esmf/ctsm5.3/ctsm53n04ctsm52028_ne30pg3t232_BgcCrop_exice_hist.clm60.r.1979-01-01.nc
This and other new finidat files are slated to come in with the beta04 tag.

[cecilehannay]

Thanks, @slevis-lmwg. Do you recommend I use that file? My run starts in 2000 but I assume it should be good enough.

[slevis-lmwg]

In the same directory we have a new 2000 file
ctsm53n04ctsm52028_ne30pg3t232_BgcCrop_exice_hist.clm60.r.2000-01-01.nc

[slevis-lmwg]

I suggest that you try it. If for some reason it causes trouble, go back to your original plan for now.

[JulioTBacmeister]

List of runs

Don't think we need L93 CAM7 SST+4K for 10 years
Don't need prognostic mom flux OFF SST+4K
Would like 1 year of CAM5 L93 ... sometime.

General comment: Comparing L93 CAM7 to L58 CAM5 should be OK for examing how high-res impacts tropospheric climate (@islasimpson ?)

[cecilehannay]

I revised the above table List of Runs following discussion with @islasimpson and @JulioTBacmeister.

• We added cam5/1deg experiments to tune cam5 before moving to ne120.
• We will check whether we can do cam5 at L93.
• We might want to drop all L58 experiments and only do L93 ?

[swrneale]

SSTs

While I understand the +4K SST experiments and how they can reveal climate sensitivities, the patterns of observed SST anomalies are much more complex and subtle that this. I'm wondering whether there's space here to do something a bit more nuanced such as maybe adding 2x or 3x the observed spatial distribution of SST increases from pre-industrial to present day? The pattern of recent SST change is quite detailed and the structure of change in the N. Atlantic. likely affects tropical cyclones and the stratocumulus regions have experienced SST change of opposite sign. Just a thought.

[cecilehannay]

Another unrelated comment about SSTs:
The 2000-2010 runs are using the HadSST 1 degree:

/glade/campaign/cesm/cesmdata/cseg/inputdata/atm/cam/sst/sst_HadOIBl_bc_1x1_1850_2021_c120422.nc

I wanted to make sure we are happy with using that dataset.

[swrneale]

Are you saying we need a higher resolution dataset or a different dataset source? Do we want daily as well? sst_reynolds_daily_0.9x1.25_yrs1982-2018_timeseries_c20190822.nc
But it is time limited.

[cecilehannay]

All the runs completed (although we will extent prog mom off to 10 years).
I am very close to my quota and I am moving the data to campaign in:

/glade/campaign/cesm/community/amwg/cam_hr

although it will take a few to complete the process on globus.