Tripolar reading of MetO data.

[RicCrocker]

This is a discussion setup to encapulate an existing conversation about UK Met Office reading of Tripolar data.

We have 2 different aspects whee we hit issues when trying to use our tripolar fields.

1. when using point2grid to go from a tripolar grid to a regular Lat-lon grid, we end up with lots of striping in the output (Northern hemisphere). This affects matching to obs point locations as MET thinks the underlying forecast value is missing (which it is). Ideally we would like the gaps to be filled. I suspect this is related to the relative resolution of the grids.

2. We are tying to read in tripolar data via UGRID settings. (nominally the functionality described here Enhance vx_data_ugrid library to read in tripolar projected datasets MET#2859 When we try running through pointstat (using a ugrid_metadata_map) we end up with segmentation errors. (If we actively convert to UGRID format we can get this to work but would rather avoid that overhead). Again it is possible that we are either getting ahead of ourselves or have just badly formatted our metadata mapping file.

I will upload the relevant data and configs we are using via google forms.