Changes between Version 6 and Version 7 of PDAF3_interface
- Timestamp:
- Mar 30, 2025, 8:42:41 AM (3 days ago)
Legend:
- Unmodified
- Added
- Removed
- Modified
-
PDAF3_interface
v6 v7 7 7 8 8 The PDAF3 interface was introduced with PDAF V3.0 to provide a consistent interface for all DA methods. 9 The PDAF3 interfaces utilizes the functionality of PDAF-OMI (Obervation Module Interface) and PDAFlocal for easy state vector localization. 9 The PDAF3 interfaces utilizes the functionality of PDAF-OMI (Obervation Module Interface) and PDAFlocal for easy state vector localization. Using the PDAF3 interface is recommended for new implementations. 10 10 11 == Routines to perform the analysis step == 11 For existing implementations from PDAF V2.x, there are cases where the PDAF3 interface can be used without much changes. Generally the routines to be called for the analysis step are analogous to those present in PDAF2.3. Here PDAFomi_* routines used the functionality of PDAFomi and PDAFlocalomi_* routines used PDAF-OMI and PDAFlocal for local filter methods. We describe the necessary porting steps in the section on [wiki:PDAF3_interface?PortingtoPDAF3 Porting to the PDAF3 interface]. 12 12 13 Generally the routines to be called for the analysis step are analogous to those present in PDAF2.3. Here PDAFomi_* routines used the functionality of PDAFomi and PDAFlocalomi_* routines used PDAF-OMI and PDAFlocal for local filter methods. If one used such routines one an easilty port to the PDAF3 interface. For new implemenetation we recommend to use the PDAF3 interfaces right away.13 Here we provide an overview of the routine names and the links to the interface descriptions. We distinguish the cases that the observation error covariance matrix **R** is diagonal from those that **R** is non-diagonal, i.e. the case of correlated observation errors. 14 14 15 Here we provide an overview of the routine names and the links to the interface descriptions. 15 == Analysis step with diagonal R matrix == 16 16 17 === Ensemble filters === 17 Assuming that the observation error covariance matrix **R** is diagonal is the more common case. The assumption is that observation errors are uncorrelated. In this case, only the observation error variances need to be considered. 18 18 19 **For diagonal R matrix** 19 === Ensemble filters and smoothers === 20 20 21 For the more typical assuming that the observation error covariance matrix **R** is diagonal there are only two routines.22 21 23 Generally there is only one routine that can execute all filter method: 24 ||= Universal Routine [[BR]]for all filters =|| [wiki:PDAF3_assimilate] || 25 This routine can be used for both the fully parallel and flexible parallization variants. 22 For the more typical assuming that the observation error covariance matrix **R** is diagonal there is one universal routine for online-coupled programs and one for offline-coupled (file-based) programs: 26 23 27 To enable an easier transition for existing PDAF2 implementations of the flexible parallelization mode to PDAF3, there is also the corresponding routine 24 ||= =||= online coupled =||= offline coupled =|| 25 ||= Universal Routine [[BR]]for all filters =|| [wiki:PDAF3_assimilate] || [wiki:PDAF3_put_state] || 26 For the online-coupled case, `PDAF3_assimilate` can be used for both the fully parallel and flexible parallization variants. 27 28 **Routines for particular cases** 29 30 To enable an easier transition for existing PDAF2 implementations of the flexible parallelization mode to the PDAF3interface, one can use the routine 28 31 ||= Universal Routine [[BR]]for flexible parallel[[BR]](backward compatibility) =|| [wiki:PDAF3_put_state] || 29 32 30 The universal routines include in their interface 3 routines that are only used for local filters. For the particualr case, that one only uses global filters or the LEnKF one can use oe of the routines33 The universal routines include in their interface three routines related to localization. These are only executed for local filters. For the particualr case, that one only uses global filters or the LEnKF one can use one of the routines 31 34 ||= **Filter** =||= **fully parallel &[[BR]]flexible parallel** =||= **flexible parallel[[BR]](backward-compatibility)** =|| 32 35 ||= global filters[[BR]]LEnKF =|| [wiki:PDAF3_assimilate_global] || [wiki:PDAF3_put_state_global] || … … 34 37 To enable an easier transition for existing PDAF2 implementations using the LEnKF to PDAF3, there is also the corresponding routine 35 38 ||= Specific LEnKF routine[[BR]](backward compatibility) =|| [wiki:PDAF3_assimilate_lenkf] || 39 36 40 There is analogously the routine [wiki:PDAF3_put_state_lenkf]. However, using the new routine [wiki:PDAFomi_set_localize_covar] in `init_dim_obs_pdafomi` in each observation module one can switch to using the new universal routine. 37 41 38 42 **Notes:** 39 * The routines `PDAF3_put_state` and `PDAF3_put_state_global` exist for backward-compatibility of the flexible parallelization variant. Starting with PDAF V3.0, we recommend to use `PDAF3_assimilate`. See the [wiki:ExternalModelLoop instructions for implementing the flexible parallelization variant] for the updated instructions. 40 * One can let PDAF select which of the routines to call. The function [wiki:PDAF_localfilter] allows to select PDAF3_assimilate_global for global filters (`PDAF_localfilter()=0`). 43 * The 'put_state' routines exist for backward-compatibility of the flexible parallelization variant. Starting with PDAF V3.0, we recommend to use `PDAF3_assimilate` also for the flexible parallelization variant. See the [wiki:ExternalModelLoop instructions for implementing the flexible parallelization variant] for the updated instructions. 44 * One can let PDAF select which of the routines to call. The function [wiki:PDAF_localfilter] allows to select `PDAF3_assimilate_global` for global filters (`PDAF_localfilter()=0`). 45 46 === 3D-Var === 47 48 Only En3DVar and hybrid 3D-Var use a local filter. The routines are the following. 49 50 For 3D-Var we distinguish routines for parameterized 3D-Var, ensemble 3D-Var and hybrid 3D-Var. 51 52 ||= **Method** =||= online coupled =||= offline coupled =|| 53 ||= 3DVar =|| [wiki:PDAF3_assimilate_3dvar] || [wiki:PDAF3_put_state_3dvar] || 54 ||= En3DVar with ESTKF or LESTKF=|| [wiki:PDAF3_assimilate_en3dvar] || [wiki:PDAF3_put_state_en3dvar] || 55 ||= hyb3DVar with ESTKF or LESTKF =|| [wiki:PDAF3_assimilate_hyb3dvar] || [wiki:PDAF3_put_state_hyb3dvar] || 56 57 The routines for En3D-Var and hybrid 3D-Var are universal an can be used to with either the LESTKF or the global ESTKF depending on the chosen subtype. 58 59 ||= En3DVar with LESTKF=|| [wiki:PDAF3_assimilate_en3dvar_lestkf] || [wiki:PDAF3_put_state_en3dvar_lestkf] || 60 ||= Hyb3DVar with ESTKF =|| [wiki:PDAF3_assimilate_hyb3dvar_lestkf] || [wiki:PDAF3_put_state_hyb3dvar_lestkf] || 61 62 63 64 == Analysis step with non-diagonal R matrix == 65 66 === Ensemble filters === 41 67 42 68 **For non-diagonal R matrix** … … 56 82 * There is no routine for the ENSRF/EAKF for nondiagonal **R**, because these filters assume that a diagonal **R** matrix exists. 57 83 84 85 86 58 87 === 3D-Var === 59 60 Only En3DVar and hybrid 3D-Var use a local filter. The routines are the following.61 62 **for diagonal R matrix**63 64 ||= **Method** =||= **fully parallel &[[BR]]flexible parallel** =||= **flexible parallel[[BR]](backward-compatibility)** =||65 ||= 3DVar =|| [wiki:PDAF3_assimilate_3dvar] || [wiki:PDAF3_put_state_3dvar] ||66 ||= En3DVar with ESTKF=|| [wiki:PDAF3_assimilate_en3dvar_estkf] || [wiki:PDAF3_put_state_en3dvar_estkf] ||67 ||= Hyb3DVar with ESTKF =|| [wiki:PDAF3_assimilate_hyb3dvar_estkf] || [wiki:PDAF3_put_state_hyb3dvar_estkf] ||68 ||= En3DVar with LESTKF=|| [wiki:PDAF3_assimilate_en3dvar_lestkf] || [wiki:PDAF3_put_state_en3dvar_lestkf] ||69 ||= Hyb3DVar with ESTKF =|| [wiki:PDAF3_assimilate_hyb3dvar_lestkf] || [wiki:PDAF3_put_state_hyb3dvar_lestkf] ||70 88 71 89