pycsa.core.delaunay

Delaunay decomposition and land-cell selection.

Partitions a lat-lon topography domain into Delaunay triangles whose vertices sit on a coarse regular grid (get_decomposition()), and selects the subset of triangles that contain enough above-threshold topography to be treated as land cells (get_land_cells()). The scipy.spatial.Delaunay object returned by get_decomposition() is augmented in place with .tri_lat_verts / .tri_lon_verts (and centroid arrays) that downstream consumers, including get_land_cells(), rely on.

Functions

get_decomposition(topo[, xnp, ynp, padding])

Partitions a lat-lon domain into a number of coarser but regularly spaced points that comprises the vertices of the Delaunay triangles.

get_land_cells(tri, topo[, height_tol, ...])

Land cell selector based on how much of a grid cell contains topography of a certain elevation.
pycsa.core.delaunay.get_decomposition(topo, xnp=11, ynp=6, padding=0)

Partitions a lat-lon domain into a number of coarser but regularly spaced points that comprises the vertices of the Delaunay triangles.

Parameters:

  • topo (array-like) – 2D topography data

  • xnp (int, optional) – number of points in the first horizontal direction, by default 11

  • ynp (int, optional) – number of points in the second horizontal direction, by default 6

  • padding (int, optional) – number of grid points to include as a boundary (padded) region, by default 0

Returns:

scipy Delaunay triangulation instance, augmented in place with tri_lat_verts/tri_lon_verts (per-triangle vertex coordinates) and tri_clats/tri_clons (centroids).

Return type:

scipy.spatial.Delaunay instance

pycsa.core.delaunay.get_land_cells(tri, topo, height_tol=0.5, percent_tol=0.95)

Land cell selector based on how much of a grid cell contains topography of a certain elevation.

Parameters:

  • tri (scipy.spatial.Delaunay) – Triangulation as returned by get_decomposition(). Must carry the .tri_lat_verts / .tri_lon_verts per-triangle vertex-coordinate arrays that get_decomposition() attaches.

  • topo (array-like) – 2D topographic data

  • height_tol (float, optional) – elevation above height_tol is considered land, by default 0.5 [m]

  • percent_tol (float, optional) – Maximum fraction of a cell that may sit at or below height_tol before the cell is rejected as ocean, by default 0.95. The cell is dropped when more than percent_tol of its grid points are at or below height_tol ((topo <= height_tol).sum() / size > percent_tol) and kept as a land cell otherwise.

Returns:

list of land cell indices. Only even-indexed triangles (range(n_tri)[::2]) are evaluated; odd-indexed triangles are skipped because each grid quad is split into two triangles and the pair shares the same topography footprint.

Return type:

list