Halos wrap around y-direction in some gungho vertical slice tests

### Are there any linked Issues or Pull Requests?

_No response_

### Brief description

Some gungho vertical slice tests have a halo depth that is too large, and wraps around itself.

e.g. the Schar mountain test has 8 cells in the y-direction, but halo depth of 4. This means that the furthest cell in the +ve and -ve halos are the same. In the depths of the past this has caused some unexpected issues (and is why many vertical slice tests have 10 cells in the y-direction).

It also explains why KGOs are changed in #174.

It seems unlikely that infrastructure will ever support only having halos in the x-direction for this case (which will be used operationally). The appropriate action for the short term is to reduce `dep_pt_stencil_extent` to `1` for these tests, which would also allow the number of cells to be reduced to 4 in the y-direction.

### Further details of the issue.

_No response_