spatial_gradient#

rojak.core.derivatives.spatial_gradient(array, units, gradient_mode, dimension=None, geod=None, crs=None)[source]#

Spatial gradient of a scalar field on a latitude/longitude grid

Parameters:

  • array (DataArray) – Field to compute the gradient of. Must have "longitude" and "latitude" dimensions.

  • units (LatLonUnits) – Units array’s latitude/longitude coordinates are expressed in

  • gradient_mode (GradientMode) – Whether to correct for map projection distortion (GradientMode.GEOSPATIAL) or not (GradientMode.CARTESIAN)

  • dimension (CartesianDimension | None) – If provided, only compute the gradient along this dimension. Defaults to both x and y.

  • geod (Geod | None) – Geodesic used to compute the nominal grid spacing (see nominal_grid_spacing()). Defaults to a WGS84 ellipsoid.

  • crs (CRS | None) – Coordinate reference system used to compute projection correction factors when gradient_mode is GradientMode.GEOSPATIAL. Defaults to plain latitude/longitude.

Returns:

Mapping with "dfdx" and/or "dfdy" keys (depending on dimension) to the corresponding component of the gradient

Return type:

dict[SpatialGradientKeys, DataArray]