Interpolate State to Gaussian Grid

[guillaumevernieres]

Duplicate of #1057

jedi-ci-test-select=intel

[guillaumevernieres]

The ci failures aren't related to this PR, the added test (converttogauss) is fine:

[Dooruk]

Looks good to me, I will wait for tests to pass before I approve it.

Does this handle land mask interpolation well?

[guillaumevernieres]

Looks good to me, I will wait for tests to pass before I approve it.

Does this handle land mask interpolation well?

no

[shlyaeva]

I'll look into the test failures, I agree they are not related, so must be happening on develop as well.

[shlyaeva]

Does this handle land mask interpolation well?

no

@guillaumevernieres what do you mean by that?

[guillaumevernieres]

Does this handle land mask interpolation well?

no

@guillaumevernieres what do you mean by that?

@shlyaeva , I guess it depends on our definition of "well". I'd say it works for most cases.
But it doesn't work for all cases: This is related to that flooding business I was trying to push to oops a while back. What we have here will work well enough for most cases but will be off when the coastline geometry is "complex" (island, peninsula, ...) and the interpolation ends up using the wrong neighbors.

[Dooruk]

@shlyaeva , I guess it depends on our definition of "well". I'd say it works for most cases. But it doesn't work for all cases: This is related to that flooding business I was trying to push to oops a while back. What we have here will work well enough for most cases but will be off when the coastline geometry is "complex" (island, peninsula, ...) and the interpolation ends up using the wrong neighbors.

To me that sounds unwell 😃 it's tough nut to crack though..

[Dooruk]

Very well 👍🏼