Update glc coupler budgets for active Greenland ice sheet

[stephenprice]

This PR adds functionality to coupler budget code so that x2g_ and l2x_ fields associated with Greenland ice sheet surface mass balance are accounted for in the coupler budget tables. Also added, but currently passing fields of zeroes, is support for g2x_ fields associated with iceberg fluxes when running with a dynamic Greenland ice sheet. Additional information for testing this PR is included in the notes below.

[BFB]

[stephenprice]

Status summary and instructions for testing:
For fluxes related to surface mass balance and for simulations of 1 day duration only, l2x_ and x2g_ fields in the coupler budget tables "close" to within approx. 5-10%. A hypothesis is that the remaining 5-10% difference is due to the fact that fluxes on a sphere (from lnd / clm) are being mapped to a planer mesh (to glc / mali). Below is an example from a 1 day BG test case (details for recreating this included below):

Screen grab of budget table for net water budget with Greenland SMB-related fluxes in the "wgsmb" row:

The values in the table can be compared with values collated from the mali and coupler history files. The grab below shows output when applying an NCO script to do that (see below) to the cpl hist files from this 1 day BG case. This script pulls out the relevant coupler hist. vector, multiplies it by the relevant areas, sums the product, and then normalizes by earth surface area (approximately) to provide a value that can be compared to budget table values.

Output from NCO script to sum relevant fields from coupler history files:

The first three values are on the x2g_ side and represent (in order) 1) the summed flux from mali history files, 2) the summed flux from cpl hist files for x2g_Flgl using "model areas" and 3) the summed flux from cpl hist files for x2g_Flgl using "mapping file areas".

The last two values are on the l2x_ side and represent (in order) 1) the summed flux from l2x_Flgl using "model areas" and 2) the summed flux from l2x_Flgl using "mapping file areas" (identical).

Note that the latter two, when multiplied by the SMB renomalization factor listed in the coupler budget table (here, 0.8823), give a value (0.0125e-6) very close to the value provided in the wgsmb row and lnd column of the coupler budget table (0.0129e-6).

Remaining issues

1. Currently, the x2g_ values in the coupler history file only work correctly for a run of 1 day. This is because the local coupler vector these fluxes are collected in represent an average value rather than a summed value (even though the 'acc' in the vector name, x2gacc_Flgl, implies 'accumulation' or summation). Thus, to recover the summed value, this vector must be multiplied by some number of time steps (here, the no. of lnd model steps in a day, 48). This value can be recovered using a counter (hack in current code) but something more robust is needed to account for runs of multiple days or longer. Similar sections of the budget code don't appear to have this problem and/or it is accounted for automatically.
2. The discrepancy between the l2x_ (from lnd to coupler) and x2g_ (from cpl to glc) values in the table may be due to the spherical vs. planer mapping issue noted above. This may be treatable by the use of a specially generated mapping file (?) or something other correction factor applied within the code. This requires more discussion to decide on a solution.

Scripts to recreate run and cpl hist file analysis
A copy of the script to set up the relevant 1 day BG case can be found here:
/home/ac.sprice/e3sm_cases/BGelm.chrys.bash

It requires updating of a few shell / path variables at the top.

A copy of the script to sum and normalize fluxes from mali and cpl history files can be found here:
/lcrc/group/e3sm/ac.sprice/scratch/chrys/20240920.BGWCYCL1850.ne30pg2_r05_EC30to60E2r2_gis20.chrysalis.gnu.CplBudgetDevel/run/calcBudgetsFromHistFiles.sh

If placed in the relevant "run" directory, it requires only that the name of the relevant 1 day mali and cpl hist files be updated at the top of the script.

[rljacob]

Notes: Jon has been meeting with Steve about this. Will test more for impact when no glc.

[stephenprice]

@jonbob -- This all looks great to me. Thanks for figuring out how to get it working properly.

One somewhat unrelated question - in line 4800, are we leaving "do_g2x=.true." unspecified because we are (currently) not sending any glc to coupler fields? But, once we start sending calving flux to coupler (for example) this would be explicitly included and set to true?

[jonbob]

That's right -- we actually already have some fields in that part of the budget calculation, even if they are 0

[stephenprice]

That's right -- we actually already have some fields in that part of the budget calculation, even if they are 0

... but because they are currently 0 we are just avoiding that part of the subroutine call for now (?).

[jonbob]

test merge passes:

• SMS_D_Ld1.ne30pg2_r05_IcoswISC30E3r5.WCYCL1850.chrysalis_intel.allactive-wcprod
• e3sm_landice_developer using gnu on chrysalis

verified the new budget lines are all 0 when smb values are not being passed

[jonbob]

@stephenprice -- I pushed some commits to your branch to get the budget balanced. This is what I get now after a month from a SMS.ne30pg2_r05_IcoswISC30E3r5_gis20.BGWCYCL1850.chrysalis_gnu.allactive-gis20km run:

(seq_diag_print_mct) NET WATER BUDGET (kg/m2s*1e6): period = all_time: date =     10201     0
                           atm            lnd            rof            ocn         ice nh         ice sh            glc        *SUM*
         wfreeze     0.00000000     0.00000000     0.00000000    -0.18610602     0.18322969     0.00287633     0.00000000    -0.00000000
           wmelt     0.00000000     0.00000000     0.00000000    13.43122439     1.59396420   -15.03046056     0.00000000    -0.00527197
           wrain   -31.13715598     5.35477771     0.00000000    25.56288221     0.00394081     0.21047538     0.00000000    -0.00507987
           wsnow    -2.47463165     1.12544865     0.00000000     0.76868060     0.16655182     0.41349834     0.00000000    -0.00045224
          wpolar     0.00000000     0.00000000     0.00000000     0.00000000     0.00000000     0.00000000     0.00000000     0.00000000
           wgsmb     0.00000000    -0.01882795     0.00000000     0.00000000     0.00000000     0.00000000     0.01882795     0.00000000
           wevap    33.37446619    -2.81504334     0.00000000   -30.46925396    -0.00902799    -0.08131534     0.00000000    -0.00017445
         wrunoff     0.00000000    -0.34837619     0.18919146     0.15878451     0.00000000     0.00000000     0.00000000    -0.00040021
         wfrzrof     0.00000000    -0.09188720     0.00000000     0.09168766     0.00000000     0.00000000     0.00000000    -0.00019954
          wirrig     0.00000000     0.00000000     0.00000000     0.00000000     0.00000000     0.00000000     0.00000000     0.00000000
           *SUM*    -0.23732144     3.20609168     0.18919146     9.35789939     1.93865853   -14.48492585     0.01882795    -0.01157828

[stephenprice]

@jonbob -- This branch has now been synched with master (as of ~5pm, 12-5-24). Let me know if you need anything else from me, and thanks again for finishing this off.

[jonbob]

@rljacob - I think this is ready for you to look at. I'll test again, but I did make sure there is no impact when glc is not in the configuration other than a couple of new lines in the budget that are populated with 0's

[rljacob]

please remove commented out code.

[stephenprice]

Done(ish). Note that I left the commented out "call route_ice_runoff" in because that is something we are likely going to need to activate with further work on this section of the budgets.

[rljacob]

Approving but not I have not carefully checked the logic. In my exploration of land ice coupling I didn't look at the budget diags.

[jonbob]

verified the budget is not impacted by this PR for non-MALI runs other than adding water and heat rows for smb that are populated with 0's, by running ne30pg2_r05_IcoswISC30E3r5 WCYCL1850 configurations for a month

[jonbob]

A one-month run of SMS.ne30pg2_r05_IcoswISC30E3r5_gis20.BGWCYCL1850.chrysalis_gnu.allactive-gis20km has this for the one-month all_time budget:

(seq_diag_print_mct) NET AREA BUDGET (m2/m2): period = all_time: date =     10201     0
                           atm            lnd            ocn         ice nh         ice sh        *SUM*
            area    -1.00000000     0.29340918     0.64585715     0.02727751     0.03345216    -0.00000401

(seq_diag_print_mct) NET HEAT BUDGET (W/m2): period = all_time: date =     10201     0
                           atm            lnd            rof            ocn         ice nh         ice sh            glc        *SUM*
         hfreeze     0.00000000     0.00000000     0.00000000     0.06210358    -0.06114375    -0.00095983     0.00000000    -0.00000000
           hmelt     0.00000000     0.00000000     0.00000000    -5.34931396     0.79747421     4.55405801     0.00000000     0.00221825
          hnetsw  -166.28141845    35.04787271     0.00000000   129.02308230     0.03069174     2.15035241     0.00000000    -0.02941929
           hlwdn  -333.66523522    81.36596963     0.00000000   238.20803346     4.51562084     9.58010606     0.00000000     0.00449475
           hlwup   388.27658736   -99.48369701     0.00000000  -272.84638414    -5.65989063   -10.28543284     0.00000000     0.00118275
         hlatvap    83.51294851    -7.05493404     0.00000000   -76.20360416    -0.02520697    -0.22963963     0.00000000    -0.00043629
         hlatfus     0.82578458    -0.37556221     0.00000000    -0.25650872    -0.05557834    -0.13798440     0.00000000     0.00015091
          hiroff     0.00000000     0.03066276     0.00000000    -0.03059617     0.00000000     0.00000000     0.00000000     0.00006659
            hsen    19.75596745    -8.09200998     0.00000000   -11.78016796     0.01281887     0.10244754     0.00000000    -0.00094408
          hpolar     0.00000000     0.00000000     0.00000000     0.00000000     0.00000000     0.00000000     0.00000000     0.00000000
        hh2otemp     0.58867312     0.00000000     0.00000000    -0.58904713     0.00000000     0.00000000     0.00000000    -0.00037401
           hgsmb     0.00000000    -0.00628289     0.00000000     0.00000000     0.00000000     0.00000000     0.00628289     0.00000000
           *SUM*    -6.98669265     1.43201896     0.00000000     0.23759709    -0.44521404     5.73294732     0.00628289    -0.02306042

(seq_diag_print_mct) NET WATER BUDGET (kg/m2s*1e6): period = all_time: date =     10201     0
                           atm            lnd            rof            ocn         ice nh         ice sh            glc        *SUM*
         wfreeze     0.00000000     0.00000000     0.00000000    -0.18610602     0.18322969     0.00287633     0.00000000    -0.00000000
           wmelt     0.00000000     0.00000000     0.00000000    13.43122439     1.59396420   -15.03046056     0.00000000    -0.00527197
           wrain   -31.13715598     5.35477771     0.00000000    25.56288221     0.00394081     0.21047538     0.00000000    -0.00507987
           wsnow    -2.47463165     1.12544865     0.00000000     0.76868060     0.16655182     0.41349834     0.00000000    -0.00045224
          wpolar     0.00000000     0.00000000     0.00000000     0.00000000     0.00000000     0.00000000     0.00000000     0.00000000
           wgsmb     0.00000000    -0.01882795     0.00000000     0.00000000     0.00000000     0.00000000     0.01882795     0.00000000
           wevap    33.37446619    -2.81504334     0.00000000   -30.46925396    -0.00902799    -0.08131534     0.00000000    -0.00017445
         wrunoff     0.00000000    -0.34837619     0.18919146     0.15878451     0.00000000     0.00000000     0.00000000    -0.00040021
         wfrzrof     0.00000000    -0.09188720     0.00000000     0.09168766     0.00000000     0.00000000     0.00000000    -0.00019954
          wirrig     0.00000000     0.00000000     0.00000000     0.00000000     0.00000000     0.00000000     0.00000000     0.00000000
           *SUM*    -0.23732144     3.20609168     0.18919146     9.35789939     1.93865853   -14.48492585     0.01882795    -0.01157828

[jonbob]

passes:

• SMS_D_Ld1.ne30pg2_r05_IcoswISC30E3r5.WCYCL1850.chrysalis_intel.allactive-wcprod

but shows DIFFs running e3sm_landice_developer. @stephenprice -- the change I made to the renormalization to use lfrin instead of lfrac will change the results. Do you or @chloewhicker want to make a longer run to make sure this impact s OK before merging this PR?