pyPDAF.PDAFomi.set_localize_covar_noniso_locweights

pyPDAF.PDAFomi.set_localize_covar_noniso_locweights(i_obs: int, dim: int, ncoords: int, coords: np.ndarray, locweights: np.ndarray, cradius: np.ndarray, sradius: np.ndarray) → None

Initialise local observation information for non-isotropic covariance localisation.

This is used in stochastic EnKF/EAKF/EnSRF. This is called in user-supplied functions pyPDAF.c__init_dim_obs_pdaf().

Here, both weighting function and localisation radii differ for each spatial dimension.

Parameters:

  • i_obs (int) – index into observation arrays

  • dim (int) – State dimension

  • ncoords (int) – number of coordinate directions

  • coords (ndarray[np.float64, ndim=2]) – Coordinates of state vector elements Array shape: (:,:)

  • locweights (ndarray[np.intc, ndim=1]) – Types of localization function Array shape: (:)

  • cradius (ndarray[np.float64, ndim=1]) – Vector of localization cut-off radii Array shape: (:)

  • sradius (ndarray[np.float64, ndim=1]) – Vector of support radii of localization function Array shape: (:)