What should we add?
The trick with global models is that a subset could cross the dateline (or prime meridian, depending on the coordinate system) -- darn that round earth!
For instance, if a model is using 0-360 longitude coordinates, then a subset that crosses 0 degrees can be wonky.
But: Do we have any examples of a global model that uses a curvilinear grid (i.e. quads and 2D lat and lon) ?
If not then maybe we don't have to do anything, but it would good if it raised an error someone tried to subset a model across the 'backside" of its coordinates.
Reference
No response
What should we add?
The trick with global models is that a subset could cross the dateline (or prime meridian, depending on the coordinate system) -- darn that round earth!
For instance, if a model is using 0-360 longitude coordinates, then a subset that crosses 0 degrees can be wonky.
But: Do we have any examples of a global model that uses a curvilinear grid (i.e. quads and 2D lat and lon) ?
If not then maybe we don't have to do anything, but it would good if it raised an error someone tried to subset a model across the 'backside" of its coordinates.
Reference
No response