Don't smooth ISOMIP+ topography on base mesh

[xylar]

Checklist

• Testing comment in the PR documents testing used to verify the changes

[xylar]

Testing

I tried 3 levels of smoothing (expansion factor 1, 2, and 4) as part of work to add mbtempest support. I discovered that a change I made in #201 wasn't good. We don't want to smooth the topography before culling because it expands the ocean domain in a way we don't want.

With this change, a smoothing factor of 1 (no smoothing) and 2 lead to the same ocean domain size, as expected. I tested this with 4km, planar Ocean0 on Chrysalis.