MOM5 updates from ACCESS-ESM1.5 #381
Conversation
| @@ -4,3 +4,4 @@ module load intel-compiler/2019.5.281 | |||
| module load netcdf/4.7.4 | |||
| module load openmpi/4.0.2 | |||
| setenv mpirunCommand "mpirun --mca orte_base_help_aggregate 0 -np" | |||
| setenv OASIS_ROOT /g/data/p66/pbd562/test/t47-hxw/jan20/4.0.2/oasis3-mct | |||
There was a problem hiding this comment.
not sure we want to hardwire a specific path like this, which is only useable for users with membership of p66
There was a problem hiding this comment.
I'm needing this setting otherwise the compile for the ESM fails, for instance...
.../src/accesscm_coupler/mom_oasis3_interface.F90(79): error #7002: Error in opening the compiled module file. Check INCLUDE paths. [MOD_PRISM]
use mod_prism
----^
Let me know if there are alternatives to try.
| (-125.62 + t_prog(indtemp)%field(i,j,1,time%taum1) * & | ||
| (3.6276 + t_prog(indtemp)%field(i,j,1,time%taum1) * & | ||
| schmidt_temp = min(35.0, t_prog(indtemp)%field(i,j,1,time%taum1))! add max limit to temp used for Schmidt number calculation. mac, jan21. | ||
| sc_co2(i,j) = 2073.1 + schmidt_temp * & |
There was a problem hiding this comment.
will this give numerically identical results to the old sc_co2 calculation?
There was a problem hiding this comment.
Good point, the Schmidt Number and gas exchange will not be identical wherever there is a surface temperature over 35 degC.
The change was made for the ESM after some points in extended warm climate CMIP6 experiments reached about 42 degC, where the Schmidt number went negative and blew up the model, as found by Tilo Z. and Martin D. in late 2020.
Checking the reference to the Schmidt formula (Wanninkhof 1992, https://agupubs.onlinelibrary.wiley.com/doi/epdf/10.1029/92JC00188), the temperature range of the calculation should only be up to 30 degC, which we have probably been abusing forever.
Martin found that the limit of 35 degC was already being used in the equivalent component of NEMO, so we adopted the same in the ESM1.5.
This way the model is stable for these hot climate experiments.
| (-81.83 + t_prog(indtemp)%field(i,j,1,time%taum1) * & | ||
| (1.483 + t_prog(indtemp)%field(i,j,1,time%taum1) * & | ||
| schmidt_temp = min(35.0, t_prog(indtemp)%field(i,j,1,time%taum1))! add max limit to temp used for Schmidt number calculation. mac, jan21. | ||
| sc_o2(i,j) = 1638.0 + schmidt_temp * & |
There was a problem hiding this comment.
as above, will this give numerically identical results to the old sc_co2 calculation?
There was a problem hiding this comment.
Good point, the Schmidt Number and gas exchange will not be identical wherever there is a surface temperature over 35 degC.
The change was made for the ESM after some points in extended warm climate CMIP6 experiments reached about 42 degC, where the Schmidt number went negative and blew up the model, as found by Tilo Z. and Martin D. in late 2020.
Checking the reference to the Schmidt formula (Wanninkhof 1992, https://agupubs.onlinelibrary.wiley.com/doi/epdf/10.1029/92JC00188), the temperature range of the calculation should only be up to 30 degC, which we have probably been abusing forever.
Martin found that the limit of 35 degC was already being used in the equivalent component of NEMO, so we adopted the same in the ESM1.5.
This way the model is stable for these hot climate experiments.
|
@russfiedler would you have time to review this PR please? |
|
Any idea why it's not building on ubuntu-18.04 ? |
|
Not sure why it is not building. |
As you can see from the linked PR, 18.04 is now deprecated, so ignore the failures with that runner. |
Some minor changes were required to the MOM5 source code in order to run with ACCESS-ESM1.5.
There are also some updated diagnostics and commentary in the ocean BGC.
None of these affect the normal running of the model.
These changes compiled and tested OK with ESM1.5 on gadi.
In particular:
Cheers, Matt.