Changes between Version 2 and Version 3 of ImplementAnalysislenkfOmi
 Timestamp:
 Nov 24, 2020, 2:58:07 PM (3 months ago)
Legend:
 Unmodified
 Added
 Removed
 Modified

ImplementAnalysislenkfOmi
v2 v3 24 24 [[PageOutline(23,Contents of this page)]] 25 25 26 With Version 1.16 of PDAF we introduced PDAFOMI (observation module infrastructure). With OMI, a smaller number of routines needs to be supplied by the user than in the previous implementation approach. 26 27 27 28 28 == Overview == 29 29 30 For the analysis step of the LEnKF different operations related to the observations are needed. These operations are requested by PDAF by calling usersupplied routines. Intentionally, the operations are split into separate routines in order to keep the operations rather elementary. This procedure should simplify the implementation. The names of the required routines are specified in the call to the routine `PDAFomi_put_state_lenkf` for the fullyparallel implementation (or `PDAFomi_put_state_lenkf` for the 'flexible' implementation). With regard to the parallelization, all these routines are executed by the filter processes (`filterpe=.true.`) only. 30 With Version 1.16 of PDAF we introduced PDAFOMI (observation module infrastructure). With OMI, a smaller number of routines needs to be supplied by the user than in the previous implementation approach. This page described the implementation of the analysis step for the local EnKF (LEnKF). 31 32 For the analysis step of the global filters we need different operations related to the observations. These operations are requested by PDAF by callback routines supplied by the user and provided in the OMI structure. The names of the required routines are specified in the call to the routine `PDAFomi_put_state_lenkf` for the fullyparallel implementation (or `PDAFomi_put_state_lenkf` for the 'flexible' implementation). With regard to the parallelization, all these routines (except `U_collect_state`, `U_distribute_state`, and `U_next_observation`) are executed by the filter processes (`filterpe=.true.`) only. 31 33 32 34 For completeness we discuss here all usersupplied routines that are specified in the interface to `PDAFomi_put_state_lenkf`. Thus, some of the usersupplied routines that are explained on the page explaining the modification of the model code for the ensemble integration are repeated here. … … 34 36 The LEnKF implemented in PDAF follows the original LEnKF by Evensen (1994) including the correction for perturbed observations (Burgers et al. 1998). The LEnKF implemented in PDAF is reviewed by Nerger et al (2005) and described in more detail by Nerger (2004). The localization is covariance lozalization of PH^T and HPH^T as described in Houtekamer & Mitchell (2001) (See the [PublicationsandPresentations page on publications and presentations] for publications and presenations involving and about PDAF) 35 37 36 In our studies (Nerger et al. 2005, Nerger et al. 2007), the EnKF showed performance deficiencies compared to the SEIK filter. Due to this, we focused more on the SEIK filter and the ETKF andESTKF after these comparison studies. For real applications, we generally recommend using ESTKF or ETKF, or their local variants LESTKF or LETKF. However, the EnKF/LEnKF might have a good performance if very large ensemble can be used as this reduces the sampling errors.38 In our studies (Nerger et al. 2005, Nerger et al. 2007), the EnKF showed performance deficiencies compared to the SEIK filter. Due to this, we focused more on the SEIK filter, the ETKF and the ESTKF after these comparison studies. For real applications, we generally recommend using ESTKF or ETKF, or their local variants LESTKF or LETKF. However, the EnKF/LEnKF might have a good performance if very large ensemble can be used as this reduces the sampling errors. 37 39 38 40 == `PDAFomi_assimilate_lenkf` == 39 41 40 The general aspects of the filter specific routines `PDAF_assimilate_*` have been described on the page [ModifyModelforEnsembleIntegration Modification of the model code for the ensemble integration] and its subpage on [InsertAnalysisStep inserting the analysis step]. The routine is used in the fullyparallel implementation variant of the data assimilation system. When the 'flexible' implementation variant is used, the routines `PDAF_put_state_*' is used as described further below. Here, we list once more the full interface of the routine. Subsequently, the full set of usersupplied routines specified in the call to `PDAFomi_assimilate_lenkf`is explained.42 The general aspects of the filter specific routines `PDAF_assimilate_*` have been described on the page [ModifyModelforEnsembleIntegration Modification of the model code for the ensemble integration] and its subpage on [InsertAnalysisStep inserting the analysis step]. The routine is used in the fullyparallel implementation variant of the data assimilation system. When the 'flexible' implementation variant is used, the routines `PDAF_put_state_*' is used as described further below. Here, we list the full interface of the routine. Subsequently, the usersupplied routines specified in the call is explained. 41 43 42 44 The interface when using the LEnKF is the following: 43 45 {{{ 44 46 SUBROUTINE PDAFomi_assimilate_lenkf(U_collect_state, U_distribute_state, & 45 U_init_dim_obs , U_obs_op, &46 U_prepoststep, U_localize_covar , &47 U_init_dim_obs_pdafomi, U_obs_op_pdafomi, & 48 U_prepoststep, U_localize_covar_pdafomi, & 47 49 U_next_observation, status) 48 50 }}} … … 50 52 * [#U_collect_statecollect_state_pdaf.F90 U_collect_state]: The name of the usersupplied routine that initializes a state vector from the array holding the ensemble of model states from the model fields. This is basically the inverse operation to `U_distribute_state` used in `PDAF_get_state` as well as here. 51 53 * [#U_distribute_statedistribute_state_pdaf.F90 U_distribute_state]: The name of a user supplied routine that initializes the model fields from the array holding the ensemble of model state vectors. 52 * [#U_init_dim_obs callback_obs_pdafomi.F90 U_init_dim_obs]: The name of the usersupplied routine thatprovides the size of observation vector53 * [#U_obs_op callback_obs_pdafomi.F90 U_obs_op]: The name of the usersupplied routine that acts as the observation operator on some state vector54 * [#U_init_dim_obs_pdafomicallback_obs_pdafomi.F90 U_init_dim_obs_pdafomi]: The name of the usersupplied routine that initializes the observation information and provides the size of observation vector 55 * [#U_obs_op_pdafomicallback_obs_pdafomi.F90 U_obs_op_pdafomi]: The name of the usersupplied routine that acts as the observation operator on some state vector 54 56 * [#U_prepoststepprepoststep_ens_pdaf.F90 U_prepoststep]: The name of the pre/poststep routine as in `PDAF_get_state` 55 * [#U_localize_covar callback_obs_pdafomi.F90 U_localize_covar]: Apply covariance localization to the matrices HP and HPH^T^57 * [#U_localize_covar_pdafomicallback_obs_pdafomi.F90 U_localize_covar]: Apply covariance localization to the matrices HP and HPH^T^ 56 58 * [#U_next_observationnext_observation_pdaf.F90 U_next_observation]: The name of a user supplied routine that initializes the variables `nsteps`, `timenow`, and `doexit`. The same routine is also used in `PDAF_get_state`. 57 59 * `status`: The integer status flag. It is zero, if `PDAFomi_assimilate_lenkf` is exited without errors. … … 64 66 {{{ 65 67 SUBROUTINE PDAFomi_put_state_lenkf(U_collect_state, & 66 U_init_dim_obs , U_obs_op, &67 U_prepoststep, U_localize , &68 U_init_dim_obs_pdafomi, U_obs_op_pdafomi, & 69 U_prepoststep, U_localize_covar_pdafomi, & 68 70 status) 69 71 }}} … … 74 76 Here all usersupplied routines are described that are required in the call to `PDAFomi_assimilate_lenkf`. For some of the generic routines, we link to the page on [ModifyModelforEnsembleIntegration modifying the model code for the ensemble integration]. 75 77 76 To indicate usersupplied routines we use the prefix `U_`. In the template directory `templates/` as well as in the example implementation in `testsuite/src/dummymodel_1D` these routines exist without the prefix, but with the extension `_pdaf.F90`. In the section titles below we provide the name of the template file in parentheses.78 To indicate usersupplied routines we use the prefix `U_`. In the template directory `templates/` as well as in the example implementation in `testsuite/src/dummymodel_1D` these routines exist without the prefix, but with the extension `_pdaf.F90`. The userroutines relating to OMI are collected in the file `callback_obs_pdafomi.F90`. In the section titles below we provide the name of the template file in parentheses. 77 79 78 80 In the subroutine interfaces some variables appear with the suffix `_p`. This suffix indicates that the variable is particular to a model subdomain, if a domain decomposed model is used. Thus, the value(s) in the variable will be different for different model subdomains. … … 82 84 83 85 This routine is independent of the filter algorithm used. 86 84 87 See the page on [InsertAnalysisStep#U_collect_statecollect_state_pdaf.F90 inserting the analysis step] for the description of this routine. 85 88 … … 88 91 89 92 This routine is independent of the filter algorithm used. 93 90 94 See the page on [InsertAnalysisStep#U_distribute_statedistribute_state_pdaf.F90 inserting the analysis step] for the description of this routine. 91 95 92 96 93 === `U_init_dim_obs ` (callback_obs_pdafomi.F90) ===97 === `U_init_dim_obs_pdafomi` (callback_obs_pdafomi.F90) === 94 98 95 This routine is used by all global filter algorithms (SEEK, SEIK, EnKF, ETKF, NETF, PF) and by the LEnKF.99 This is a callback routine for PDAFOMI initializing the observation information. The routine just calls a routine from the observation module for each observation type. 96 100 97 The interface for this routine is: 98 {{{ 99 SUBROUTINE init_dim_obs_pdafomi(step, dim_obs_p) 100 101 INTEGER, INTENT(in) :: step ! Current time step 102 INTEGER, INTENT(out) :: dim_obs_p ! Dimension of observation vector 103 }}} 104 105 The routine is called at the beginning of each analysis step. It has to initialize the size `dim_obs_p` of the observation vector according to the current time step. Without parallelization `dim_obs_p` will be the size for the full model domain. When a domaindecomposed model is used, `dim_obs_p` will be the size of the observation vector for the subdomain of the calling process. 106 107 With PDAFOMI, the routine just calls a routine from the observation module for each observation type. 101 See the [wiki:OMI_Callback_obs_pdafomi documentation on callback_obs_pdafomi.F90] for more information. 108 102 109 103 110 === `U_obs_op` (calllback_obs_pdafomi.F90) ===111 104 112 This routine is used by all global filter algorithms (SEEK, SEIK, EnKF, ETKF, NETF, PF) and the LEnKF. 105 === `U_obs_op_pdafomi` (callback_obs_pdafomi.F90) === 113 106 114 The interface for this routine is: 115 {{{ 116 SUBROUTINE obs_op_pdafomi(step, dim_p, dim_obs_p, state_p, m_state_p) 107 This is a callback routine for PDAFOMI applying the observation operator to the state vector. The routine calls a routine from the observation module for each observation type. 117 108 118 INTEGER, INTENT(in) :: step ! Currrent time step 119 INTEGER, INTENT(in) :: dim_p ! PElocal dimension of state 120 INTEGER, INTENT(in) :: dim_obs_p ! Dimension of observed state 121 REAL, INTENT(in) :: state_p(dim_p) ! PElocal model state 122 REAL, INTENT(out) :: m_state_p(dim_obs_p) ! PElocal observed state 123 }}} 124 125 The routine is called during the analysis step. It has to perform the operation of the observation operator acting on a state vector that is provided as `state_p`. The observed state has to be returned in `m_state_p`. 126 127 For a model using domain decomposition, the operation is on the PElocal subdomain of the model and has to provide the observed substate for the PElocal domain. 128 129 With PDAFOMI, the routine just calls a routine from the observation module for each observation type. 109 See the [wiki:OMI_Callback_obs_pdafomi documentation on callback_obs_pdafomi.F90] for more information. 130 110 131 111 132 112 === `U_prepoststep` (prepoststep_ens_pdaf.F90) === 133 113 134 The general aspects of this routines have already been described on the [ModifyModelforEnsembleIntegration#U_prepoststepprepoststep_ens_pdaf.F90 page on modifying the model code for the ensemble integration] for the SEIK filter. For completeness, the description is repeated specifically for the EnKF:114 The routine has already been described for modifying the model for the ensemble integration and for inserting the analysis step. 135 115 136 The interface of the routine is identical for all filters, but sizes can vary. Also, the particular operations that are performed in the routine can be specific for each filter algorithm. 137 138 The interface for this routine is for the LEnKF 139 {{{ 140 SUBROUTINE prepoststep(step, dim_p, dim_ens, dim_ens_p, dim_obs_p, & 141 state_p, Uinv, ens_p, flag) 142 143 INTEGER, INTENT(in) :: step ! Current time step 144 ! (When the routine is called before the analysis step is provided.) 145 INTEGER, INTENT(in) :: dim_p ! PElocal state dimension 146 INTEGER, INTENT(in) :: dim_ens ! Size of state ensemble 147 INTEGER, INTENT(in) :: dim_ens_p ! PElocal size of ensemble 148 INTEGER, INTENT(in) :: dim_obs_p ! PElocal dimension of observation vector 149 REAL, INTENT(inout) :: state_p(dim_p) ! PElocal forecast/analysis state 150 ! The array 'state_p' is not generally not initialized in the case of SEIK/EnKF/ETKF. 151 ! It can be used freely in this routine. 152 REAL, INTENT(inout) :: Uinv(1, 1) ! Not used not LEnKF 153 REAL, INTENT(inout) :: ens_p(dim_p, dim_ens) ! PElocal state ensemble 154 INTEGER, INTENT(in) :: flag ! PDAF status flag 155 }}} 156 157 The routine `U_prepoststep` is called once at the beginning of the assimilation process. In addition, it is called during the assimilation cycles before the analysis step and after the ensemble transformation. The routine is called by all filter processes (that is `filterpe=1`). 158 159 The routine provides for the user the full access to the ensemble of model states. Thus, usercontrolled pre and poststep operations can be performed. For example the forecast and the analysis states and ensemble covariance matrix can be analyzed, e.g. by computing the estimated variances. In addition, the estimates can be written to disk. 160 161 Hint: 162 * If a user considers to perform adjustments to the estimates (e.g. for balances), this routine is the right place for it. 163 * Only for the SEEK filter the state vector (`state_p`) is initialized. For all other filters, the array is allocated, but it can be used freely during the execution of `U_prepoststep`. 164 * The array `Uinv` is not used in the EnKF. Internally to PDAF, it is allocated to be of size (1,1). 165 * Apart from the size of the array `Uinv`, the interface is identical for all ensemble filters (SEIK/ETKF/EnKF/LSEIK/LETKF/LEnKF). In general it should be possible to use an identical pre/poststep routine for all these filters. 166 * The interface through which `U_prepoststep` is called does not include the array of smoothed ensembles. In order to access the smoother ensemble array one has to set a pointer to it using a call to the routine `PDAF_get_smootherens` (see page on [AuxiliaryRoutines auxiliary routines]) 116 See the page on [InsertAnalysisStep#U_prepoststepprepoststep_ens_pdaf.F90 inserting the analysis step] for the description of this routine. 167 117 168 118 169 119 170 === `U_localize ` (callback_obs_pdafomi.F90) ===120 === `U_localize_covar_pdafomi` (callback_obs_pdafomi.F90) === 171 121 172 122 This routine is only used for the LEnKF. … … 190 140 191 141 This routine is independent of the filter algorithm used. 142 192 143 See the page on [InsertAnalysisStep#U_next_observationnext_observation_pdaf.F90 inserting the analysis step] for the description of this routine. 193 144