- py__prodrinva_hyb_l_pdaf()¶
Provide \(\mathbf{R}^{-1}_l \times \mathbf{A}_l\) with weighting.
Here, one should do \(\mathbf{R}^{-1}_l \times \mathbf{A}_l\). The matrix \(\mathbf{A}_l\) depends on the filter algorithm.
This function is used in LKNETF where gamma is multipled with c_l for weighting between LETKF and LNETF.
One can also perform observation localisation. This can be helped by the function
pyPDAF.PDAF.local_weight()
to get the observation weight.Parameters¶
- domain_p: int
Current local domain index.
- step: int
Current time step
- dim_obs_l: int
Dimension of observation vector in local analysis domain
- rank: int
Rank of the local analysis domain The size of it dpends on the filter algorithms.
- obs_l: np.ndarray[np.float, dim=1]
Observation vector in local analysis domain. shape: (dim_obs_l, )
- gamma: float
Hybrid weight provided by PDAF
- a_l: np.ndarray[np.float, dim=2]
Matrix A in local analysis domain. shape: (dim_obs_l, rank)
- c_l: np.ndarray[np.float, dim=2]
\(\mathbf{R}^{-1}_l \times \mathbf{A}_l\) in local analysis domain. shape: (dim_obs_l, rank)
Returns¶
- c_l: np.ndarray[np.float, dim=2]
\(\mathbf{R}^{-1}_l \times \mathbf{A}_l\) in local analysis domain. shape: (dim_obs_l, rank)