polaris.mesh.reconstruct.tangential_reconstruction

polaris.mesh.reconstruct.tangential_reconstruction(ds: Dataset, edge_normal_field: DataArray, stencil: DataArray | None = None, weights: DataArray | None = None, location: Literal['cell', 'vertex'] | None = None) → tuple[DataArray, DataArray, DataArray][source]

Reconstruct a tangential vector field from an edge-normal vector field.

Parameters:

  • ds (xr.Dataset) – MPAS mesh dataset

  • edge_normal_field (xr.DataArray (nEdges, ...)) – Scalar field defined normal to each edge (e.g. normalVelocity). May include additional dimensions besides nEdges (e.g. nVertLevels, Time); these are carried through unchanged.

  • stencil (xr.DataArray (nCells or nVertices, maxEdges2 or NINE), optional) – Precomputed edge stencil for each reconstruction point. If not provided (along with weights), it is built from ds using location.

  • weights (xr.DataArray (nCells or nVertices, R3, maxEdges2/NINE), optional) – Precomputed reconstruction weights for each reconstruction point. If not provided (along with stencil), it is built from ds using location.

  • location (str ["cell", "vertex"], optional) – Point location where the reconstruction occurs. Required if stencil and weights are not both provided.

Returns:

  • u_x (xr.DataArray (nCells or nVertices, …)) – Reconstructed x-component of the tangential vector field

  • u_y (xr.DataArray (nCells or nVertices, …)) – Reconstructed y-component of the tangential vector field

  • u_z (xr.DataArray (nCells or nVertices, …)) – Reconstructed z-component of the tangential vector field