I have been looking at the dry convective boundary layer (
Attached is a plot of the TKE budget which shows that the sum of the right-hand-side of the TKE equation (tendencies) does not match the output dk/dt which I calculate with an explicit Euler method. I know that this doesn't match the time integration scheme of CM1, but I would expect more agreement as I was able to achieve a better TKE closure for the Rayleigh-Benard (
Are there other physics terms not covered by the output budget terms that I need to incorporate? Or is the residual between dk/dt and its tendencies inherently due to additional dissipation introduced by the psolver in removing acoustics? I've also attached the namelist as well, and I am happy to provide other information as needed.
I appreciate any help or suggestions.
les_ConvBoundLayer) preconfigured example (testcase = 1). I am interested in computing the turbulent kinetic energy budget, but I am having difficulty closing it. I know previous studies have mentioned that the budget terms need to be computed with the same numerics as the solver, so I have been outputting the various momentum budget terms from the solver directly (output_ubudget = output_vbudget = output_wbudget = 1) and appropriately multiplying them with the respective fluctuating field, ui' (I define ' as the deviation from the horizontal average rather than the base state as in CM1). I have also taken care not to perform interpolation and have computed all SGS terms explicitly (doimpl = 0).Attached is a plot of the TKE budget which shows that the sum of the right-hand-side of the TKE equation (tendencies) does not match the output dk/dt which I calculate with an explicit Euler method. I know that this doesn't match the time integration scheme of CM1, but I would expect more agreement as I was able to achieve a better TKE closure for the Rayleigh-Benard (
dns_RayleighBenard) testcase (see attached plot).Are there other physics terms not covered by the output budget terms that I need to incorporate? Or is the residual between dk/dt and its tendencies inherently due to additional dissipation introduced by the psolver in removing acoustics? I've also attached the namelist as well, and I am happy to provide other information as needed.
I appreciate any help or suggestions.
Last edited: