Running the multiple connected domain set-up with ROMS

[lanerolle]

If I run ROMS in a multiple connected domain set-up by providing a sequence of connected grid files and the corresponding initial and forcing files, does ROMS know how to put all the domains/grids together and perform the run as if on a single domain/grid? Or

(1) do I have to tell ROMS how these multiple domains are connected? and/or
(2) are the computations for each domain carried our separately? If so, is there any
information exchange between the different domains?

Also, do the model outputs (eg. history files) for each domain get written out separately?

Thank you.

[kate]

I haven't tried it so don't know the answer to all of the questions, but they do run as independent domains. There is no grid- coupling code in ROMS at this time, although we'd obviously like to have it.

[arango]

ROMS has already a nesting/composed grid infrastructure. However, there are a couple of routines missing to perform this task. We are working on it. Please be patient. The issue here is that we needed to have all the adjoint structure first before doing the full coupling. The good news is that we are done with the adjoint codes. Now, we are working on the sea-ice coupled model and wetting and drying capabalities.

There are some conservation issues that we need to take care for the full coupling. This is the tricky part.

[debcox]

I realise that there is a lot of work going on, so not being impatient but just wanted to clarify - the comment above refers to nested as well as coupled grids? ie. Nested grids are not yet functional?

Deborah

[kate]

We do a one-way nesting with separate ROMS runs. Run the largest domain first, create boundary condition files for an inner domain from the history files of the outer domain, then run the inner domain. Quite a few do this, so the boundary file creation tools exist.