How to use unstructured open boundaries?

1

Hello!

I’m having trouble using unstructured open boundaries without a coordinates.nc file in NEMO 4.2.2. I have a configuration that includes both ice and ocean modules. Despite reviewing previous discussions on BDY input data, I am still encountering several issues.

Here’s what we have tried so far:

  • We use nambdy_index to identify boundaries and then trim the data according to these identified boundaries.
  • We set a rimwidth of 10 cells.
  • For the W/E and S/N boundaries, we used different types of input data, ensuring that the dimensions in the .nc file matched the dimensions in the variables: x, y, z, time, and y, x, z, time (cross boundary and along boundary).
  • We tried both consistent ordering for all boundaries (x, y, z, time) and different orders for each boundary, but neither approach has worked.
  • We’ve tested several schemes, but nothing works.

Some problems, that we’ve noticed:

  1. After reading ice concentration boundary, nothing reading more in output file, why?

    image
    image1063×649 85.2 KB

  2. Problem with boundary produces NaN-values (T-points are good, but velocities produces NaN, so you can see on step 0 T-points are good, then on step 1 there are NaN-values like on velocities from step 0), on this picture we’ve tested only short W-bdy:

    image
    image1342×1306 147 KB

I’ll greatly appreciate if anyone could send a working file of bdy data and explain how it works.

Thank you!

2

Hi.
Yes, when you delfine coordinate file for BDY, you cannot use index for T point for U and V point. The problem is that even if T point is in the sea, U and V for the same index can end up on the land and you will get None or your solution will blow up.

2 Likes
3

There is and example of nambdy_index and nambdy_dta here. It you still have problems, could you upload your namelist_cfg ?