search

E3SM-Project / v3atm

Fork of E3SM for testing v3 atm changes
Other
0 stars 5 forks source link

Add VBS SOA and couple it with chemUCI #22

Closed mingxuanwupnnl closed 1 year ago

mingxuanwupnnl commented 1 year ago

This PR provides the capability of running VBS SOA treatment with chemUCI. This PR is a stealth feature which does not turn on VBS by default.

One conditional compilation if statement: VBS_SOA (add -vbs in CAM_CONFIG_OPTS)

The code modifications are mainly for:

  1. replace default SOA gas-aerosol partition subroutine mam_soaexch_1subarea with new VBS partition subroutine mam_soaexch_vbs_1subarea

  2. Now there are 8 SOAG species, and 7 (SOAG15, 24, 35, 34, 33, 32, 31) of them will condense to form SOA. SOAG0 is the gas precursor emitted. Many changes in modal_aero_amicphys.F90 are made for the index, as now nsoa == 1 but nsoag == 7 (default is nsoa ==1 and nsoag == 1)

  3. A new chemistry mechanism file is added, pp_chemUCI_linozv3_mam4_resus_mom_vbs_tag.in.

It is BFB (stealth) for chemUCI only simulations. test suites results compared with NGD_v3atm baseline (./create_test $TESTNAMES --compare -b NGD_v3atm): /lcrc/group/e3sm/ac.mwu/scratch/chrys/xxx.C.20221019_212747_l7r75l

PASS ERS_Ln11_P512x1.ne30pg2_EC30to60E2r2.F20TR_chemUCI-Linozv3.chrysalis_intel.eam-20tr_v3atm_rtmoff DIFF ERS_Ln9.ne4pg2_oQU480.F20TR_chemUCI-Linozv3.chrysalis_intel.eam-20tr_v3atm_rtmoff FAIL PEM_Ln5.ne4pg2_oQU480.F20TR_chemUCI-Linozv3.chrysalis_intel.eam-20tr_v3atm_rtmoff FAIL PET_Ln5.ne4pg2_oQU480.F20TR_chemUCI-Linozv3.chrysalis_intel.eam-20tr_v3atm_rtmoff DIFF REP_Ln9.ne4pg2_oQU480.F20TR_chemUCI-Linozv3.chrysalis_intel.eam-20tr_v3atm DIFF SMS_Ln5.ne30pg2_EC30to60E2r2.F20TR_chemUCI-Linozv3.chrysalis_intel.eam-20tr_v3atm_cosplite DIFF SMS_Ln5.ne4pg2_oQU480.F20TR_chemUCI-Linozv3.chrysalis_intel.eam-20tr_v3atm_cosplite

FAIL ERP_Ld3.ne30pg2_EC30to60E2r2.F20TR_chemUCI-Linozv3.chrysalis_intel.eam-20tr_v3atm DIFF ERS_Ld3.ne4pg2_oQU480.F20TR_chemUCI-Linozv3.chrysalis_intel.eam-20tr_v3atm DIFF ERS_Ln11.ne30pg2_EC30to60E2r2.F20TR_chemUCI-Linozv3.chrysalis_intel.eam-20tr_v3atm_rtmoff FAIL PEM_Ln9.ne30pg2_EC30to60E2r2.F20TR_chemUCI-Linozv3.chrysalis_intel.eam-20tr_v3atm FAIL PET_Ln5.ne30pg2_EC30to60E2r2.F20TR_chemUCI-Linozv3.chrysalis_intel.eam-20tr_v3atm FAIL PET_Ln5_P256x2.ne30pg2_EC30to60E2r2.F20TR_chemUCI-Linozv3.chrysalis_intel.eam-20tr_v3atm

Except P3_input_dim in eam.h0 and P3_output_dim in eam.h1, other variables are BFB for tests showing DIFF.

Minimal tests passed and show BFB for custom-10_1x10, custom-10_2x5, custom-30_1x10, and custom-30_2x5.

Two-year amip run show a global SOA burden of 3.59 Tg (2005). The spatial pattern looks reasonable.

wlin7 commented 1 year ago

Merged to NGD_v3atm.