- py__prodrinva_pdaf()¶
Provide \(\mathbf{R}^{-1} \times \mathbf{A}\).
Here, one should compute \(\mathbf{R}^{-1} \times \mathbf{A}\) where \(\mathbf{R}\) is observation error covariance matrix. The matrix \(\mathbf{A}\) depends on the filter algorithm. In ESTKF, \(\mathbf{R}\) can is ensemble perturbation in observation space.
Parameters¶
- stepint
Current time step
- dim_obs_pint
Dimension of observation vector
- rank: int
Rank of the matrix A (second dimension of A) This is ensemble size for ETKF and ensemble size - 1 for ESTKF.
- obs_p: np.ndarray[np.float, dim=1]
Observation vector. shape: (dim_obs_p,)
- a_p: np.ndarray[np.float, dim=2]
Input matrix A. shape: (dim_obs_p, rank)
- c_p: np.ndarray[np.float, dim=2]
Output matrix \(\mathbf{C} = \mathbf{R}^{-1} \times \mathbf{A}\). shape: (dim_obs_p, rank)
Returns¶
- c_p: np.ndarray[np.float64, dim=2]
Output matrix \(\mathbf{C} = \mathbf{R}^{-1} \times \mathbf{A}\). shape: (dim_obs_p, rank)