MClarkson
EF5
Some of you must have more experience with diffusion than I do... so I'd appreciate some input. I have this 2d gridded model that I am creating, the aim of which is to model thunderstorms. It has a 20km domain, uses a C grid, fully compressible and non-hydrostatic etc etc. After a large amount of effort to make it computationally stable, it is finally working well and now I am adding diffusion. At a grid spacing in the range of 100-20m, both horizontally and vertically, what percentage of a grid square should be "shared" with its neighbors each second? .1%? 1%? 10%? This is a fairly high shear (especially in the updraft region), conditionally unstable environment. My exact diffusion coefficient is based on shear and stability, I am just looking for ballpark figures for now, to make sure I am at least thinking on the right scale.
Thanks!
Thanks!
