Unexpectedly obtaining Axisymmetric Climatology in F09 Finite Volume dynamical | HS94 & PK02

[ag4680@nyu_edu]

Hi,I am running the finite volume dynamical core with simple physics and analytical initial conditions (HS94 and Polvani-Kushner Stratosphere) using the F19 and F09 horizontal grid and 40 vertical levels. While the f19 grid generates the expected climatology with expected macroturbulence in the midlatitudes, I get an axisymmetric climatology for the f09 grid. In terms of configuration and parameters, all the parameters I am using in the user_nl_cam namelists are the same. Is there some perturbation parameters that should have been there but is not?From what I understand, in some dynamical cores(like the pseudospectral one), axisymmetric forcings (like a flat surface boundary) can generate axisymmetric circulations and some small perturbation is needed. So, I would suspect that might have been the culprit, however then I should have obtained a similar axisymmetric circulation for the f19 grid too.Besides, I noticed that the finite volume is the only dycore for which the temperature perturbation definitions use the analytic_ic_active() logical argument in dyn_comp.F90; the other dycores dont have this clause :if ((pertlim /= 0._r8) .and. (.not. analytic_ic_active()))Why is that? I am not sure if this is causing the problem, because again, it would have created a problem for the f19 resolution as well.

Please find the attached dynamical variables U and T and surface pressure PS from the generated climatology.Any help on this matter is deeply appreciated. Regards,Aman

 

[ag4680@nyu_edu]

Hi,I ran some quick tests on f19, f09(with and without the analytic_ic clause mentioned above) and ne16 with 40 levels and analytic ICs. I ran each model for 50 days. I was able to replicate the results. Just to be sure, I repeated the tests 2-3 times.Instability develops, as it should, for the ne16 and f19 runs for all the times I ran the model.However, Instabilities never developed for f09 run with the analytic_ic_active() clause in the if-condition in fv/dyn_comp.F90.(which makes me wonder why this does not happen with f19)And, Instabilities developed 1 out of 3 times for f09 runs without the analytic_ic_active() clause.Regards,Aman

 

[ag4680@nyu_edu]

Hi,I ran some quick tests on f19, f09(with and without the analytic_ic clause mentioned above) and ne16 with 40 levels and analytic ICs. I ran each model for 50 days. I was able to replicate the results. Just to be sure, I repeated the tests 2-3 times.Instability develops, as it should, for the ne16 and f19 runs for all the times I ran the model.However, Instabilities never developed for f09 run with the analytic_ic_active() clause in the if-condition in fv/dyn_comp.F90.(which makes me wonder why this does not happen with f19)And, Instabilities developed 1 out of 3 times for f09 runs without the analytic_ic_active() clause.Regards,Aman

 

[ag4680@nyu_edu]

Hi,I think I have realised the reason; This might be because, due to using analytic ICs, the model is not reading the pertlim vaue of 1.0D-5 from the held suarez test case .xml file. I think the spectral element model does not use the pertlim value to perturb the temperature(the control never went to the perturbation function even though , however the finite volume dycore still does. I removed the analytic_ic_active clause() from the if-condition and added the pertlim parameter to my user namelist and now the finite volume f09 model behaves fine and develops instability within first 10 days.Regards,Aman 

 

[ag4680@nyu_edu]

Hi,I think I have realised the reason; This might be because, due to using analytic ICs, the model is not reading the pertlim vaue of 1.0D-5 from the held suarez test case .xml file. I think the spectral element model does not use the pertlim value to perturb the temperature(the control never went to the perturbation function even though , however the finite volume dycore still does. I removed the analytic_ic_active clause() from the if-condition and added the pertlim parameter to my user namelist and now the finite volume f09 model behaves fine and develops instability within first 10 days.Regards,Aman