search

E3SM-Project / v3atm

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

Improve wet removal and add low Nc limit #34

Closed yunpengshan2014 closed 1 year ago

yunpengshan2014 commented 1 year ago

Aerosol wet removal improvement:

  1. Use vertical velocity from ZMmp scheme in the deep convection wet removal
  2. Detrain the cloud-borne aerosol into stratiform
  3. Reorder the stratiform wet removal and the deep convection wet removal
  4. Add Nc limit (default 10 per CC) in the P3 scheme
  5. Fix the hard-coded qc exponent in the enhancement factors of auto-conversion and accretion
crterai commented 1 year ago

A series of custom-10, custom-30 tests all came back BFB. 

6e6451f2bd7e8865cba6b303a7ccb3a5  atm_custom-10_1x2_nmonths.txt
6e6451f2bd7e8865cba6b303a7ccb3a5  atm_custom-10_2x1_nmonths.txt
6e6451f2bd7e8865cba6b303a7ccb3a5  atm_custom-30_1x2_nmonths.txt
crterai commented 1 year ago

@wlin7 - if it looks okay, this PR is ready to be merged to the NGD_v3atm branch.

wlin7 commented 1 year ago

Merged to NGD_v3atm. Testing underway. DIFFs in nml and results are expected.

wlin7 commented 1 year ago

Alert: Compy ERS tests on ne30pg2 failed, though ne4pg2 passed. (See the line of Compare_base_rest). 227 out of 448 fields in cpl.hi have non-zero differences. Chrysalis' ERS tests are even worse, FAIL on both ne4pg2 and ne30pg2 grids.

@yunpengshan2014 , any idea about this? When this happens, typically, it is because updates for some fields by physics are not carried to the restart file. Such fields are either not in the restart file at all, or the updating is not in sync with the restart writing (due to sub-stepping). Do you see which variables or where in the codes could be suspicious?