search

marbl-ecosys / MARBL

Marine Biogeochemistry Library
https://marbl-ecosys.github.io
Other
14 stars 25 forks source link

initial commit for 4p2z settings #423

Closed kristenkrumhardt closed 1 year ago

kristenkrumhardt commented 1 year ago

This is my first attempt at creating the yaml file for the 4p2z settings.

mnlevy1981 commented 1 year ago

Progress so far

  1. The marbl.settings file generated by MARBL_tools/MARBL_generate_settings_file.py pointing at the new 4p2z JSON file (and using CESM_x1 for the grid) produces bit-for-bit identical results in the call_compute_subroutines test when compared to the same test reading the marbl_in file @kristenkrumhardt has been using for her 4p2z runs, e.g. /glade/work/kristenk/cesm_work/cesm2.2_scope_cases/g.e22.GOMIPECOIAF_JRA-1p4-2018.TL319_g17.4p2z.001branch/CaseDocs/marbl_in (note that I needed to generate an updated IC file to include cocco tracers and split zooC into microzooC and mesozooC; I also updated both settings files to turn off tracer restoring)
  2. I've merged the update_zoo_diags branch into this branch; there were some diagnostics that were named *_zoo* that were effectively *_zootot*, but now that we can run with multiple zooplankton we should really be providing them as per-zoo fields.

The next step will be to merge development into this branch, and run the same tests. There are also two points we should discuss before accepting this PR...

Discussion Topics

  1. update_zoo_diags also includes a change to picpoc that is not BFB
diff --git a/src/marbl_interior_tendency_mod.F90 b/src/marbl_interior_tendency_mod.F90
index e6017ef..398b42c 100644
--- a/src/marbl_interior_tendency_mod.F90
+++ b/src/marbl_interior_tendency_mod.F90
@@ -1452,7 +1452,8 @@ subroutine compute_autotroph_calcification(km, autotroph_local, temperature, car
           picpoc = max(0.0_r8, -0.0136_r8 * CO2(:) + picpoc(:) + 0.21_r8)

           !P-limitation effect (CaCO2/organicC ratio increases when P limitation term is low)
-          picpoc = max(0.0_r8, -0.48_r8 * Plim(auto_ind,:) + picpoc(:) + 0.48_r8)
+          ! max out picpoc at one to prevent crashing due to "ballast exceeds POC" error
+          picpoc = min(1., max(0.0_r8, -0.48_r8 * Plim(auto_ind,:) + picpoc(:) + 0.48_r8))

           !multiply cocco growth rate by picpoc to get CaCO3 formation
           CaCO3_form(auto_ind,:) = picpoc(:) * photoC(auto_ind,:)

I'd like to know if we would expect to see answer changes in the current 3p1z setup, or if picpoc is only used in explicit calcification so it will only effect runs with cocco

  1. The current way we handle grazing requires the pred_ind in grazing_relationship_settings(prey_ind, pred_ind) to match the index of zooplankton_settings(:) referring to the predator. This can be seen in marbl_interior_tendency_mod.F90, where we need the biomass of the predator and use zooplankton_local%C(pred_ind,k). Addressing #320 would remove this problem, but in the meantime we need grazing_relationship_settings(4,1) = "null" and grazing_relationship_settings(4,2) = "Grazing of microzooplankton by mesozooplankton" because grazing_relationship_settings(4,1) = "Grazing of microzooplankton by mesozooplankton" implicitly uses microzoo biomass instead of mesozoo biomass.

I am happy waiting until we fix #320 and letting this take care of itself, but an alternative approach would be to add pred_ind to the grazing_relationship_settings_type; we could then collapse grazing_relationship_settings to a 1D array (there are a few changes in this PR that are meant to handle the fact that there are 7 grazing relationships in the 4p2z setup but we need to allocate a 4x2 array to store them all)

Next Steps

Next up, I'll merge development into this branch; it looks like there are conflicts in a couple of files to resolve, and I'll also need to add the following new variables to the 4p2z settings file:

diff --git a/defaults/settings_latest.yaml b/defaults/settings_latest.yaml
index 87e50c0..2f105ce 100644
--- a/defaults/settings_latest.yaml
+++ b/defaults/settings_latest.yaml
@@ -232,6 +232,12 @@ general_parms :
       units : unitless
       datatype : logical
       default_value : .true.
+   lcheck_forcing :
+      longname : Control whether consistency checks are performed on forcing input (useful for debugging)
+      subcategory : 2. config flags
+      units : unitless
+      datatype : logical
+      default_value : .false.
    lecovars_full_depth_tavg :
       longname : If .false., MARBL will recommend truncating the column for some diagnostics
       subcategory : 2. config flags
@@ -283,6 +289,12 @@ general_parms :
       units : unitless
       datatype : logical
       default_value : .false.
+   bftt_dz_sum_thres:
+      longname : MARBL will abort if abs(1 - sum(bot_flux_to_tend)) exceeds this threshold
+      subcategory : 4. general parameters
+      units : unitless
+      datatype : real
+      default_value : 1.0e-14
    Jint_Ctot_thres_molpm2pyr :
       longname : MARBL will abort if abs(Jint_Ctot) exceeds this threshold
       subcategory : 4. general parameters
@@ -1008,6 +1020,14 @@ PFT_derived_types :
             default_value :
                default : 1e34
                ((zooplankton_sname)) == "zoo" : 0.4
+         basal_respiration_rate_per_day :
+            longname : Basal respiration rate
+            subcategory : 11. zooplankton
+            units : 1/day
+            datatype : real
+            default_value :
+               default : 1e34
+               ((zooplankton_sname)) == "zoo" : 0.0
          loss_thres :
             longname : Zoo concentration where losses go to zero
             subcategory : 11. zooplankton