jhan-dev-20250225-1 intended UM7 update to Pearse's ~600 year ocean spinup

[JhanSrbinovsky]

Branch name: jhan-dev-20250225-1

Description of experiment

spack.yaml in ACCESS-ESM1.6 build, PR27-5 suggests that the he UM executable currently running in Pearse's ocean spin up, dates back to 17th December. At this time UM7 was basically the ESM15 model, updated to use CABLE3. We ported the fix to convective momentum transport (CMT) at that time.

My recollection of the events prior to Xmas are that we had several things we wanted to get into the model prior to starting a long run. However, our fixes were not quite ready so Pearse began his simulation with the version of UM7 as it existed at that time.

Therefore, the modifications in this UM executable are essentially representative of all the development work done in the UM7 repo, less that early CMT fix.

The purpose of the experiment is to evaluate the enhanced atmospere/land executable. Our hope is to then continue the 600+ year spinup, substituting this improved UM executable. The improvements in this UM executable are limited to the land-surface and the land-surface interaction with the atmosphere. They include but are not limited to:

• Assuming spatial maps for soil texture parameters
• Input CABLE's PFT and soil parameters via namelists
• Numerous bug fixes that vastly improve performance
• Improved surface dust emission
• Improved energy and water balances
• Improved treatment of ICE component in soil
• Fix wood harvesting (thinning) scheme
• Adjust water accumulation in lakes
• Improve parametrisation of root fractions
• Test other parametrzations
• Test model configurations such as use Medlyn stomatal model, Soil Thermal Fix
• Adjust correction terms
• Uses a revised LUH3 based PFT distribution

Note: As this simulation predates the implementation of this platform, the name of the simulation does not match the title of this branch. The simulation is base-70b173e - where 70b173e is the SHA of the src code from where the executable was built. The output is at /scratch/p66/jxs599/access-esm/archive/base-70b173e-dev-preindustrial+concentrations-253fc085, currently at ~136 years.

tag: @har917 @rml599gh @tiloz

[JhanSrbinovsky]

transient error on NCI? so far at year 372 and I think 256 there was file system failure to write ocean restart and so following year submission fell over whilst looking for it. Peeled back a layer and restarted both times @tiloz @har917 @rml599gh @blimlim

[MartinDix]

Figures from @har917

[MartinDix]

Here splitting Mediterrean at 15 E

[rml599gh]

Arnold ran an ESM1.5 historical with the CMT fix included. It would be useful to see whether there are any changes in East African rainfall in that run and what the impact on Mediterranean salinity was i.e. will it get even drier/worse or wetter and improve.

[MartinDix]

Current MOM configuration https://github.com/ACCESS-NRI/access-esm1.6-configs/blob/2a33443c4a1c50f3f2b175fd7190bd2cf6fe8338/ocean/input.nml#L465 has

s_max = 55.0
s_max_limit =42.0

The model will stop if salinity gets above s_max. We had to increase this limit in ACCESS-CM2 with the SSP5-8.5 extension to 2300 useing s_max=70.

MOM documentation says that s_max_limit is

!  Maximum salinity below which will employ upwind advection instead
!  of quicker, and horizontal diffusion instead of neutral physics. 

[MartinDix]

Updated salinity graphs suggest it might be slowing down? Proably don't have to worry about hitting the 55 limit for another ~ 1000 years.

[ofa001]

Thanks @MartinDix , its rate of increase is getting slower fractionally and its mixing to the whole depth and isn't mixing into the Western Mediterranean at all which is interesting. I guess we could increase the 55 ppt limit to 60 if it got close in any run in the future, if we cant find a better solution, (local rain!).

[har917]

@rml599gh @JhanSrbinovsky @pearseb @MartinDix update of main global metrics to year 650 of the new run (ps I may have got number of years of CMT overlap run given the addition of 101 between payu/output).

main ocean metrics:

global near-surface air temperature:

land carbon:



Note that soil respiration is trending upwards which indicates that carbon cycle is not quite in statistical equilibrium - possibly because climate is also still adjusting to the solar constant, but perhaps also because of a slow accumulation of carbon.

[har917]

An additional confirming comment - near-surface air temperature is now sitting around (just above) 287K. This is consistent with the CMIP6 multi-model ensemble for the historical period - see figure and only slightly too warm when compared to observations (1960-1990).

[MartinDix]

E Med increase is definitely slowing. Perhaps more of a sign of leakage to the W now.