Extend historical solar forcing from 1700 to 2300

[penguian]

See

gadi:/g/data/vk83/configurations/inputs/access-esm1p5/modern/historical/atmosphere/forcing/resolution_independent/2021.06.22/TSI_CMIP6_ESM_v2

https://github.com/ACCESS-NRI/UM7/blob/dev-access-esm1.6/umbase_hg3/src/atmosphere/short_wave_radiation/scvary_Mod.F90 line 38:

! SCVARY: Array of variable solar constant, from 1700 to 2300.

and https://github.com/ACCESS-NRI/UM7/blob/dev-access-esm1.6/umbase_hg3/src/control/top_level/setcona.F90 lines 2033 to 2041:

      if (L_SCVARY) then
      write(6,*) 'SETCONA: Reading in solar forcing from file: ',    &
     &            FILE_SCVARY
      OPEN(UNIT=56,file=FILE_SCVARY,status='old')
      do i=1,601
        read(56,*) ysol,scvary(i)
      enddo
      CLOSE(56)
      endif

The ACCESS-ESM1.6 code therefore expects the file accessed via FILE_SCVARY (currently TSI_CMIP7_ESM) to contain 601 lines of the form

1700 1365.444521
...
2300 1365.759766

representing total solar irradiance values from 1700 to 2300.

[penguian]

The UM7 source file https://github.com/ACCESS-NRI/UM7/blob/dev-access-esm1.6/umbase_hg3/src/atmosphere/short_wave_radiation/solvar.F90 contains the code

!     define solar constants for year 2301 and after to follow 11 years solar cycle
      REAL,PARAMETER :: sc_cycle(1:11) =(/1366.53954,1366.30442,1366.02501,   &
          1365.78996,1365.67390,1365.71370,1365.89665,1366.16474,&
          1366.43289,1366.61586,1366.65563/)
...
!          Calculate this year's solar constant & thence FCSC and BAND1D
!          IF YOU WANT TO INCREASE THE SIZE OF THE SOLAR CONSTANT VARIATION
!          DO NOT DO IT HERE AS IT WILL GIVE SILLY SPECTRAL STUFF.
           IF ( YEAR .GT. 2300 ) THEN
             SOV = sc_cycle(mod(YEAR-2301,11)+1)
           ELSE
             SOV = SCVARY(YEAR-1699)
           ENDIF

!          Set SC as the mean over 1700-2003 (original Met Office code scaled
!          this to 1365).
           SC = sum(SCVARY(1:304))/304.
           FCSC = SOV / SC - 1.

Would it be best to set the SCVARY value used before 1850 to 1361.617 to match the pre-industrial control, and use the values for the 11 year cycle above for the years from 2022 to 2300?

[penguian]

Also, note that the atmosphere/input_atm.nml SC value of 1361.617 is noticeably lower than the sc_cycle values quoted above. See also https://www.solarisheppa.kit.edu/75.php and https://gmd.copernicus.org/articles/17/1217/2024/ . These indicate that the values in sc_cycle are too large and need to be revised downwards.

[penguian]

See also https://www.solarisheppa.kit.edu/img/CMIP7_metadata_description_4.6.pdf

[penguian]

See also the following code in https://github.com/ACCESS-NRI/UM7/blob/dev-access-esm1.6/umbase_hg3/src/include/constant/c_mdi.h

      ! New real missing data indicator (-2**30)
      Real, Parameter    :: RMDI     = -32768.0*32768.0

Since RMDI is defined in terms of powers of 2, as -32768.0*32768.0 == -1073741824.0, this value can be unambiguously used for the SCVARY values after 2022, as long as the historical experiment does not read them. It is not clear whether the code in https://github.com/ACCESS-NRI/UM7/blob/dev-access-esm1.6/umbase_hg3/src/atmosphere/short_wave_radiation/solvar.F90 can cope with RMDI values.

[penguian]

Possibly the simplest and most valid thing to do for the SCVARY values before 1850 would be to set the values for every year to the average of the 1850-2022 values.