Page 1 of 1

Concurrent vs Separate Run One-way Nesting

PostPosted: Thu Aug 23, 2018 3:00 am
by kelucas
What are the differences (in concept and results) between concurrent one-way nesting (setting the feedback to 0) and separate one-way nesting (using ndown)? If the idea on one-way nesting is how the parent domain does not interact with the nested domain, theoretically, shouldn't their results be the same? Or is my understanding of one-way nesting incorrect? Thanks! :)

Re: Concurrent vs Separate Run One-way Nesting

PostPosted: Mon Aug 27, 2018 1:29 am
by dcvz
The main difference in the two approaches is that in the concurrent method lateral boundary conditions are available to the nest at every coarse-grid time step. With ndown, the LBCs are computed for every coarse-grid history_interval. This can be a big deal depending on the configuration of the grids and phenomenon being studied.

Ndown does have some advantages, such as if you want to do a reanalysis on the nest or want to change some namelist switches in your nested re-run.