Some comparisons of model fields with different timesteps and grids.
Some comparisons of model fields with different timesteps and grids.
These compare 3 model runs:
1. as in the standard run described in Kessler et al (1997), which has a coarse grid spacing about 110 km in x and 40 km in y (concentrated near the equator and the east and west boundaries) and with a one hour timestep;
2. a run with fine grid spacing up to 20km in both x and y (concentrated north of the equator and in the east), and with a 30-minute timestep;
3. a run with the fine grid as in 2 but with 20-minute timestep.
-> The grid spacing for the coarse and fine grids.
-> CFL condition for the fine grid
-> Grid and CFL condition for fine grid (detail near Central America)
The forcing in each case is identical: smooth annual cycle winds, solar radiation and observed clouds.
All comparisons are snapshots at either day 360 or day 720.
Ideally, the fields should not change significantly among these runs.
Unfortunately, that is not the case ...
Comparisons on model levels:
Comparisons on meridional sections:
Comparison of the coarse and fine grids for the model forced by ECMWF winds during a stiff Tehuantepecker (Jan 96). Tehuantepec eddies are only generated on the fine grid. See also the Tehuantepecker page.
Another comparison shows the effect of doubling the Shapiro filter
Exp-113 is like 82 (std run) but year 4. For these purposes 4 years is sufficient. In this run nord=nshapu/h/t=8 (light filtering).
Exp-154 is exactly like 113 but has nord=nshpu/h/t=4 (doubled filter strength).
