The tests will involve mesh to mesh regridding, ie unstructured representations of nodes, edges and faces with face to node connectivity and edge to node connectivity. The tests will concentrate on vector to vector field regridding only, since this will exercise the extensive field to extensive field regridding and the conversion of vector fields to extensive fields and vice versa.
We can use analytical solutions for the tests provided there is some trivial lon-lat dependency -- u and v need to vary in non-trivial ways. Emphasis is on low resolution, mostly.
[ ] 2d: lon-lat rectilinear mesh to rectilinear mesh. W1 and W2h vector fields.
[x] 2d: lon-lat to cubed-sphere. W1 and W2h vector fields.
[x] 2d: cubed-sphere to lon-lat. W1 and W2h vector fields.
The tests will involve mesh to mesh regridding, ie unstructured representations of nodes, edges and faces with face to node connectivity and edge to node connectivity. The tests will concentrate on vector to vector field regridding only, since this will exercise the extensive field to extensive field regridding and the conversion of vector fields to extensive fields and vice versa. We can use analytical solutions for the tests provided there is some trivial lon-lat dependency -- u and v need to vary in non-trivial ways. Emphasis is on low resolution, mostly.