Opened 6 months ago

Last modified 5 months ago

#164 new defect

viscosity computed at firstcall in newsedim

Reported by: emillour Owned by:
Priority: minor Milestone:
Component: GENERIC GCM Version:
Keywords: Cc:

Description (last modified by emillour)

Not really a bug, but a potential issue, in newsedim: https://trac.lmd.jussieu.fr/Planeto/browser/trunk/LMDZ.GENERIC/libf/phystd/newsedim.F , the viscosity of the atmosphere is computed at firstcall based on the gas mixture provided by gases.def, but also using the temperature (if He or H2 are involved) at that first step:

         do igas=1, ngasmx
           if(gfrac(igas).ge.0.0) then
             if(igas.eq.igas_CO2) then
               molrad = molrad + gfrac(igas)*2.2e-10                              ! CO2 
               visc(:,:) = visc(:,:) + gfrac(igas)*1.0e-5                         ! CO2
             elseif(igas.eq.igas_N2) then
               molrad = molrad + gfrac(igas)*1.8e-10                              ! N2 (Kunze et al. 2022)
               visc(:,:) = visc(:,:) + gfrac(igas)*1.0e-5                         ! N2
             elseif(igas.eq.igas_H2) then
               molrad = molrad + gfrac(igas)*1.41e-10                             ! H2 (Ackerman & Marley 2001)
               visc(:,:) = visc(:,:) + gfrac(igas)*2.0d-07*pt(:,:)**0.66          ! H2 (from Rosner 2000)
             elseif(igas.eq.igas_H2O) then 
               molrad = molrad + gfrac(igas)*2.3e-10                              ! H2O (Crifo 1989 at 300K)
               visc(:,:) = visc(:,:) + gfrac(igas)*8e-6                           ! H2O (Sengers & Kamgar-Parsi 1984)
             elseif(igas.eq.igas_He) then
               molrad = molrad + gfrac(igas)*1.1e-10                              ! He (Kunze et al. 2022)
               visc(:,:) = visc(:,:) +                                            ! He
     &                     gfrac(igas)*1.9e-5*(pt(:,:)/273.15)**0.7               ! He (Petersen 1970)
...
  • This might be a problem in cases where there is a strong (e.g. diurnal) temperature contrast during the simulation.
  • This will kill any 1+1=2 test

From ongoing discussion, a reasonable way forward would be to add a flag so the user can select if viscosity need be recomputed at every time step.

Change History (1)

comment:1 Changed 5 months ago by emillour

Description: modified (diff)
Note: See TracTickets for help on using tickets.