How to judge a model if it is an Eddy-resolving model

[zhouwei]

Dear all,

I ofen see ' ROMS is an Eddy-resolving model ' from various of paper , but I don't know how to judge a model if it is a eddy-resolving model.
What unique charateristic and function does an eddy-resolving model have? Any information are greatly appreciated .

[kate]

The statement "ROMS is an eddy-resolving model" doesn't really make sense. You can run it on grids that resolve your eddies - or don't resolve them. How big are your eddies? They may not be all one size. One thing to think about is your damping timescale due to horizontal viscosity. The larger the grid cells, the larger the typical value of VISC2/VISC4 used to control the gridscale noise. For a wave of length 2 dx, there will be an associated damping time. Now think about a wave of say 10 dx - that too will be damped, but more slowly. How long do you need your eddies to survive in an "eddy-resolving" model? Rather longer than the generation time. Note that you can have "eddy-permitting" setups as well where there are eddy-like features, but you can't call them truly resolved.

There's an appendix in the ROMS manual with the formulae as taught to me by Dale Haidvogel (mumble years ago).

[zhouwei]

Hi, Kate. Thank you for your explanations ,which help me remove my misunderstanding of this problem.