high-resolution dataset and nesting

[lrw]

I am wondering about the preservation of high-resolution features in a nested domain. If I run a 12/4km nested domain and use a 4km LAPS analysis for the model initial conditions I realize that the 4km analysis will be smoothed out for the 12km domain. However, will those 4km features be preserved in the 4km nested domain?

Thanks!

[jimmyc]

A. Why would you want to degrade the 4 km data back to a 12 km grid? I understand if its for computational reasons, but why not just run a 4 km grid over the area of interest?

B. I think there is an option to create a 2nd domain with the specified input data, rather than re-interpolating from the 12 km grid.
This requires that both domain 1 and 2 start at the initial time. Otherwise, in order for the solution and LBC's to be consistent you have to interpolate from the coarse grid.

So if everything is equal, you might as well just run a 4 km grid only. Besides, the use of cumulus scheme on the 12 km grid has some potential negative impacts on the inner nest.