Embedded Files
Vortex Shedding Test Cases with/without Coriolis force, and 120/60 resolution
Test case results at: /glade/work/kyho/vortex_shedding_results/
mpasa120.no_coriolis
mpasa120.no_coriolis
vortex_mpasa120_no_coriolis.cam.h0i.0001-01-01-00000.regrid.1x1.nc
mpasa120.with_coriolis
mpasa120.with_coriolis
vortex_mpasa120_coriolis.cam.h0i.0001-01-01-00000.regrid.1x1.nc
mpasa60.no_coriolis
mpasa60.no_coriolis
mpasa60.no_coriolis
mpasa60.with_coriolis
mpasa60.with_coriolis
vortex_mpasa60_coriolis.cam.h0i.0001-01-01-00000.regrid.0.5x0.5.nc
Vortex shedding test case, 1 degree MPAS 120, with coriolis, default parameters
Vortex shedding test case, 1 degree MPAS 120, with coriolis, default parameters
Vortex shedding test case, 0.5 degree MPAS 60, with coriolis, default parameters
Vortex shedding test case, 0.5 degree MPAS 60, with coriolis, default parameters
Vortex shedding test case, 1 degree MPAS 120, with coriolis, 2nd order time stepping
Vortex shedding test case, 1 degree MPAS 120, with coriolis, 2nd order time stepping
Very similar to default 3rd order time stepping
Temporal convergence with MPAS 120
Temporal convergence with MPAS 120
Very small differences in the eyeball norm
Look at the L^2 norm of the difference between surface pressure between dt={45, 22.5, 15, 9, 5, 3} and dt=1. No observed time convergence when varying the time step. Worth looking at this result for the native output grid, but remeshing shouldn't make it 0th order, should be at least 1st order.
Smagorinsky diffusion parameter impact
Smagorinsky diffusion parameter impact
Remarkably stable with a large range of diffusion coefficients. High parameter values lead to strong damping of shear, and damping of shear prevents the formation of vorticity.
Laplacian Eddy Diffusion impact
Laplacian Eddy Diffusion impact
Set mpas_horiz_mixing='2d_fixed', then set mpas_h_mom_eddy_visc2, mpas_v_mom_eddy_visc2, mpas_h_theta_eddy_visc2, and mpas_v_theta_eddy_visc2 all to the same value. 1e3 was on the upper range for stability which is reasonable, taking into account the Earth scaling factor 1e3 becomes 4e5. However, the fact that the model remains stable for a coeff of 1e-2 suggests that other diffusion settings are maintaining stability. Not super large differences between the various settings.
Page updated
Report abuse