E3SM icon indicating copy to clipboard operation
E3SM copied to clipboard

Published 20 hours ago • verified publisher icon E3SM-Project

Fix a bug in VBS SOA gas-aerosol partition calculation

Open mingxuanwupnnl opened this issue 1 year ago • 1 comments

This bug fix corrects issue #6474

In subroutine mam_soaexch_vbs_1subarea in components/eam/src/chemistry/modal_aero/modal_aero_amicphys.F90

The calculation of p0_soa_298 is not correct. Temperature of 298 K should be used instead of ambient temperature

This fix has very small impact on global mean climate

See e3sm_diags: Default tag v3.0.0: https://web.lcrc.anl.gov/public/e3sm/diagnostic_output/ac.mwu/E3SMv3/20240520.v3.LR.F2010.chrysalis/e3sm_diags/atm_monthly_180x360_aave/model_vs_obs_0002-0011/viewer/ VBSfix: https://web.lcrc.anl.gov/public/e3sm/diagnostic_output/ac.mwu/E3SMv3/20240520.v3.LR.F2010.VBSfix.chrysalis/e3sm_diags/atm_monthly_180x360_aave/model_vs_obs_0002-0011/viewer/

and IICE comparison: https://portal.nersc.gov/project/m2136/bin/iice/iice.cgi?url1=https://web.lcrc.anl.gov/public/e3sm/diagnostic_output/ac.mwu/E3SMv3/20240520.v3.LR.F2010.chrysalis/e3sm_diags/atm_monthly_180x360_aave/model_vs_obs_0002-0011/viewer&label1=default%20v3&url2=https://web.lcrc.anl.gov/public/e3sm/diagnostic_output/ac.mwu/E3SMv3/20240520.v3.LR.F2010.VBSfix.chrysalis/e3sm_diags/atm_monthly_180x360_aave/model_vs_obs_0002-0011/viewer&label2=VBSfix&category=&plot=&diff=0&dconfig=vsObs&

mingxuanwupnnl avatar Jul 07 '24 18:07 mingxuanwupnnl

PR Preview Action v1.4.7 :---: :rocket: Deployed preview to https://E3SM-Project.github.io/E3SM/pr-preview/pr-6502/ on branch gh-pages at 2024-07-07 18:53 UTC

github-actions[bot] avatar Jul 07 '24 18:07 github-actions[bot]

@wlin7 is this ready to go? Could go on next today.

rljacob avatar Oct 29 '24 20:10 rljacob

Merged to next. The fix is straightforward; it will impact all configs that use vbs soa (standard v3atm).

wlin7 avatar Oct 29 '24 20:10 wlin7

Noting that 2 of the non-BFB tests flagged this PR. https://web.lcrc.anl.gov/public/e3sm/e3smtest/evv/MVK_PS.ne4pg2_oQU480.F2010.chrysalis_intel.C.JNextNbfb20241030_010350/index.html

Only variable rejected was soa_a1_SRF

Also https://web.lcrc.anl.gov/public/e3sm/e3smtest/evv/TSC_PS.ne4pg2_oQU480.F2010.chrysalis_intel.C.JNextNbfb20241030_010350/index.html only variable num_a1

rljacob avatar Oct 31 '24 02:10 rljacob

@mingxuanwupnnl would you expect this to be climate changing? TSC tests indicate it is.

rljacob avatar Oct 31 '24 17:10 rljacob

@mingxuanwupnnl would you expect this to be climate changing? TSC tests indicate it is.

I don't think this bug fix will change the climate significantly. Not sure if I understand the test correctly. Does it mean that only soa_a1_SRF and num_a1 are climate changing?

mingxuanwupnnl avatar Oct 31 '24 17:10 mingxuanwupnnl

This may not be an indication of climate changing. The test is for time step convergence, with a very small time step size (2 seconds) for an ne4pg2 grid. num_a1 is deemed non-convergent during much of the 10-minute simulation but not always . NUMLIQ was also detected but only for the 1st step. Look like the nonconvergence in num_a1 does not propagate to other fields.

Is it possible that the value of num_a1 in the reference run is small in magnitude, and the relative fluctuation is more sensitive to the changes introduced by the fix?

wlin7 avatar Oct 31 '24 18:10 wlin7

The MVK test passed last night so it must be on the borderline. But TSC is still indicating a fail.

rljacob avatar Oct 31 '24 18:10 rljacob

TSC is consistently diffing from this PR. @wlin7 please make a bless request.

rljacob avatar Nov 04 '24 17:11 rljacob