Changes between Initial Version and Version 1 of InsertAnalysisStep


Ignore:
Timestamp:
Apr 29, 2014, 3:55:56 PM (6 years ago)
Author:
lnerger
Comment:

--

Legend:

Unmodified
Added
Removed
Modified
  • InsertAnalysisStep

    v1 v1  
     1= Inserting the analysis step into the time stepping for fully parallel data assimialtion =
     2
     3{{{
     4#!html
     5<div class="wiki-toc">
     6<h4>Implementation Guide</h4>
     7<ol><li><a href="ImplementationGuide">Main page</a></li>
     8<li><a href="AdaptParallelization">Adaptation of the parallelization</a></li>
     9<li><a href="InitPdaf">Initialization of PDAF</a></li>
     10<li><a href="ModifyModelforEnsemble">Modifications for ensemble integration</a></li>
     11  <li>&nbsp;&nbsp;&nbsp;Fully parallel implementation variant</li>
     12  <li>&nbsp;&nbsp;&nbsp;<a href="ModifyModelforEnsembleIntegration">Flexible implementation variant</a></li>
     13<li><a href="ImplementationofAnalysisStep">Implementation of the analysis step</a></li>
     14<li><a href="AddingMemoryandTimingInformation">Memory and timing information</a></li>
     15</ol>
     16</div>
     17}}}
     18
     19[[PageOutline(2-3,Contents of this page)]]
     20
     21== Overview ==
     22
     23For the fully-parallel variant of the assimilation system, only a single routine has to added to the time stepper of the numerical model. For each filter algorithm there is a specific routine `PDAF_assimilate_X`, with `X` being the name of the filter. This routine counts the time steps during the model integration and computes the analysis step of the filter when the observation time is reached.  It can be convenient to not directly insert `PDAF_assimilate_X` into the code but to use an interface routine `assimilate_pdaf` that does not need any arguments.
     24
     25In addition, the initialization routine `init_pdaf` discussed before on the [InitPdaf page on initializing PDAF], is extended by one additional subroutine call to the PDAF routine `PDAF_get_state`.
     26
     27An example for this implementation variant can be found in the tutorial implementations in the directory `tutorial/`. Here `online_2D_serialmodel` shows the implementation with a serial (i.e. non-parallel) model, while `online_2D_parallelmodel` shows the implementation with a parallel model using domain-decomposition.
     28
     29== `PDAF_get_state` ==
     30
     31The routine `PDAF_get_state` has to be called once at the end of the initialization of PDAF. Usually, the call will be added to the routine `init_pdaf` that was discussed on the [InitPdaf page on initializing PDAF].
     32
     33The routine `PDAF_get_state` has the purpose to initialize the model fields to be propagates from the array holding the ensemble states. In addition, the routine can initialized the information, whether further model integrations have to be computed and how many time steps have to be performed. For the fully-parallel implementation variant, the number of time steps is used inside PDAF, while the flag on further model integrations is not used.
     34
     35The interface of `PDAF_get_state` is the following:
     36{{{
     37  SUBROUTINE PDAF_get_state(nsteps, timenow, doexit, U_next_observation, U_distribute_state, &
     38                            U_prepoststep, status)
     39}}}
     40with the following arguments:
     41 * `nsteps`: An integer specifying upon exit the number of time steps to be performed
     42 * `timenow`: A real specifying upon exit the current model time. (This value is usually not used in the fully-parallel implemenation variant)
     43 * `doexit`: An integer variable defining whether the assimilation process is completed and the program should exit the while loop. For compatibility 1 should be used for exit, 0 for continuing in the loop. (This value is not used in the fully-parallel implemenation variant)
     44 * [#U_next_observationnext_observation.F90 U_next_observation]: The name of a user supplied routine that initializes the variables `nsteps`, `timenow`, and `doexit`
     45 * [#U_distribute_statedistribute_state.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
     46 * [#U_prepoststepprepoststep_seik.F90 U_prepoststep]: The name of a user supplied routine that is called before and after the analysis step. Here the user has the possibility to access the state ensemble and can e.g. compute estimated variances or can write the ensemble states the state estimate into files.
     47 * `status`: The integer status flag. It is zero, if `PDAF_get_state` is exited without errors.
     48
     49PDAF also has a [PdafSimplifiedInterface Simplified Interface] providing the routine `PDAF_get_state_si`. In the simplified interface, the names of all user-supplied call back routines are predefined such that they not appear in the call to `PDAF_get_state_si`. More information on the pre-defined names is provided in the [PdafSimplifiedInterface documentation of PDAF's simplified interface].
     50
     51== Inserting `assimilate_pdaf` ==
     52
     53The right place to insert the interface routine `assimilate_pdaf` into the model code, is at the end of the time stepping loop. Thus, usually this is directly before the '''END DO''' of the time stepping loop. One has to ensure that the rounte is called at each time step, so that PDAF can count the time steps until the next analysis time.
     54
     55The purpose of `assimilate_pdaf` is to call the filter-specific PDAF-core routine `PDAF_assimilate_X`, with `X` being the name of the filter method. It is also possible to insert the call to `PDAF_assimilate_X` directly into the model code. However, using the additional interface routine yield usually cleaner source code. This is because of subroutine name that are specified in the call the `PDAF_assimilate_X` or when more than one filter are implemented.
     56
     57== `PDAF_assimilate_X` ==
     58
     59There is a separate routine `PDAF_assimilate_X` for each of the filter algorithms. The name of the routine includes the name of the filter at its end (instead of `X`). The purpose of the `PDAF_assimilate_X` routines is to count the time steps. When the forecast phase is complete the routine writes back the forecast model fields into the array holding the ensemble of model state vectors and executes the analysis step of the chosen filter algorithm. The interface to each 'assimilate' routine is specific for each filter algorithm, because the names of several user-supplied routines have to be specified, which are specific for each filter algorithm. However, at the stage of implementing the ensemble integration only the first and last arguments of the routines are relevant.
     60
     61For example, the interface when using the ESTKF filter is the following:
     62{{{
     63  SUBROUTINE PDAF_assimilate_estkf(U_distribute_state, U_collect_state, &
     64                                 U_init_dim_obs, U_obs_op, &
     65                                 U_init_obs, U_prepoststep, U_prodRinvA, &
     66                                 U_init_obsvar, next_observation_pdaf, status)
     67}}}
     68At this state of the implementation only these arguments are relevant:
     69 * [#U_distribute_statedistribute_state.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. (This routine is also used in `PDAF_get_state`)
     70 * [#U_collect_statecollect_state_pdaf.F90 U_distribute_state]: The name of the user-supplied 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_dist_state`.
     71 * [#U_next_observationnext_observation.F90 U_next_observation]: The name of a user supplied routine that initializes the variables `nsteps`, `timenow`, and `doexit`. (This routine is also used in `PDAF_get_state`)
     72 * `status`: The integer status flag. It is zero, if PDAF_get_state is exited without errors.
     73
     74The other arguments are names of user-supplied subroutines that are only executed if the analysis step is executed (See the section [#Compilationandtesting 'Compilation and testing'] for how to provide these routines for compilation at this stage). These routines are explained in the next section of the implementation guide ([ImplementationofAnalysisStep Implementation of the Analysis step]) separately for each available filter algorithm.
     75
     76PDAF also has a [PdafSimplifiedInterface Simplified Interface] providing the routine `PDAF_assimilate_X_si`. In the simplified interface, the names of all user-supplied call back routines are predefined such that they not appear in the call to `PDAF_assimilate_X_si`. More information on the pre-defined names is provided in the [PdafSimplifiedInterface documentation of PDAF's simplified interface].
     77
     78== User-supplied routines ==
     79
     80Here, only the user-supplied routines are discussed that are required at this stage of the implementation (that is, the ensemble integration). For testing (see [#Compilationandtesting 'Compilation and testing']), all routines need to exist, but only those described here in detail need to be implemented with functionality.
     81
     82To indicate user-supplied 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.
     83
     84=== `U_next_observation` (next_observation_pdaf.F90) ===
     85
     86The interface for this routine is
     87{{{
     88SUBROUTINE next_observation(stepnow, nsteps, doexit, timenow)
     89
     90  INTEGER, INTENT(in)  :: stepnow  ! Number of the current time step
     91  INTEGER, INTENT(out) :: nsteps   ! Number of time steps until next obs
     92  INTEGER, INTENT(out) :: doexit   ! Whether to exit forecasting (1 for exit)
     93  REAL, INTENT(out)    :: timenow  ! Current model (physical) time
     94}}}
     95
     96The routine is called once at the beginning of each forecast phase. It is executed by all processes that participate in the model integrations.
     97
     98Based on the information of the current time step, the routine has to define the number of time steps `nsteps` for the next forecast phase. In addition, the flag `doexit` has to be initialized to provide the information if the external ensemble loop can be exited. `timenow` is the current model time. This variable should also be initialized. It is particularly important, if an ensemble task integrates more than one model state. In this case `timenow` can be used to correctly jump back in time.
     99
     100Some hints:
     101 * If the time interval between successive observations is known, `nsteps` can be simply initialized by dividing the time interval by the size of the time step
     102 * `doexit` should be 0 to continue the assimilation process. In most cases `doexit` is set to 1, when `PDAF_get_state` is called after the last analysis for which observations are available.
     103 * At the first call to `U_next_obs` the variable `timenow` should be initialized with the current model time. At the next call a forecast phase has been completed. Thus, the new value of `timenow` follows from the timer interval for the previous forecast phase.
     104
     105=== `U_distribute_state` (distribute_state_pdaf.F90) ===
     106
     107The interface for this routine is
     108{{{
     109SUBROUTINE distribute_state(dim_p, state_p)
     110
     111  INTEGER, INTENT(in) :: dim_p           ! State dimension for PE-local model sub-domain
     112  REAL, INTENT(inout) :: state_p(dim_p)  ! State vector for PE-local model sub-domain
     113}}}
     114
     115This routine is called during the forecast phase as many times as there are states to be integrated by a model task. Again, the routine is executed by all processes that belong to model tasks.
     116
     117When the routine is called a state vector `state_p` and its size `dim_p` are provided. As the user has defined how the model fields are stored in the state vector, one can initialize the model fields from this information. If the model is not parallelized, `state_p` will contain a full state vector. If the model is parallelized using domain decomposition, `state_p` will contain the part of the state vector that corresponds to the model sub-domain for the calling process.
     118
     119Some hints:
     120 * If the state vector does not include all model fields, it can be useful to keep a separate array to store those additional fields. This array has to be kept separate from PDAF, but can be defined using a module like `mod_assimilation`.
     121
     122
     123=== `U_prepoststep` (prepoststep_ens_pdaf.F90) ===
     124
     125The interface of the routine is identical for all filters. However, the particular operations that are performed in the routine can be specific for each filter algorithm. Here, we exemplify the interface on the example of the SEIK filter.
     126
     127The interface for this routine is
     128{{{
     129SUBROUTINE prepoststep(step, dim_p, dim_ens, dim_ens_p, dim_obs_p, &
     130                       state_p, Uinv, ens_p, flag)
     131
     132  INTEGER, INTENT(in) :: step        ! Current time step
     133                         ! (When the routine is called before the analysis -step is provided.)
     134  INTEGER, INTENT(in) :: dim_p       ! PE-local state dimension
     135  INTEGER, INTENT(in) :: dim_ens     ! Size of state ensemble
     136  INTEGER, INTENT(in) :: dim_ens_p   ! PE-local size of ensemble
     137  INTEGER, INTENT(in) :: dim_obs_p   ! PE-local dimension of observation vector
     138  REAL, INTENT(inout) :: state_p(dim_p) ! PE-local forecast/analysis state
     139                                     ! The array 'state_p' is not generally not initialized in the case of SEIK/EnKF/ETKF.
     140                                     ! It can be used freely in this routine.
     141  REAL, INTENT(inout) :: Uinv(dim_ens-1, dim_ens-1) ! Inverse of matrix U
     142  REAL, INTENT(inout) :: ens_p(dim_p, dim_ens)      ! PE-local state ensemble
     143  INTEGER, INTENT(in) :: flag        ! PDAF status flag
     144}}}
     145
     146The 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`).
     147
     148The routine provides for the user the full access to the ensemble of model states. Thus, user-controlled pre- and post-step 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. If the smoother is used, also the smoothed ensembles can be analyzed. In addition, the estimates can be written to disk.
     149
     150Hint:
     151 * If a user considers to perform adjustments to the estimates (e.g. for balances), this routine is the right place for it.
     152 * 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`.
     153 * 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])
     154
     155
     156=== `U_collect_state` (collect_state_pdaf.F90) ===
     157
     158The interface for this routine is
     159{{{
     160SUBROUTINE collect_state(dim_p, state_p)
     161
     162  INTEGER, INTENT(in) :: dim_p           ! State dimension for PE-local model sub-domain
     163  REAL, INTENT(inout) :: state_p(dim_p)  ! State vector for PE-local model sub-domain
     164}}}
     165
     166This routine is called during the forecast phase as many times as there are states to be integrated by a model task. It is called at the end of the integration of a member state of the ensemble. The routine is executed by all processes that belong to model tasks.
     167
     168When the routine is called, a state vector `state_p` and its size `dim_p` are provided. The operation to be performed in this routine is inverse to that of the routine `U_distribute_state`. That is, the state vector `state_p` has to be initialized from the model fields. If the model is not parallelized, `state_p` will contain a full state vector. If the model is parallelized using domain decomposition, `state_p` will contain the part of the state vector that corresponds to the model sub-domain for the calling process.
     169
     170Some hints:
     171 * If the state vector does not include all model fields, it can be useful to keep a separate array to store those additional fields. This array has to be kept separate from PDAF, but can be defined using a module like `mod_assimilation`.
     172
     173== Simulating model errors ==
     174
     175The implementation of the filter algorithms does not support the specification of a model error covariance matrix. This was left out, because in the SEEK and SEIK filter, the handling can be extremely costly, as the model error covariance matrix has to be projected onto the ensemble space. Instead PDAF support the simulation of model errors by disturbing fields during the model integration. For this, some routine will be required that is inserted into the time stepping loop of the model. As this procedure is specific to each model, the is no routine provided by PDAF for this.
     176
     177== Compilation and testing ==
     178
     179To compile the extended model code with PDAF, one has to extend the Makefile for the model by adding the additional user-supplied routines. While all of the user-supplied routines need to exist not all of them need to be fully implemented at this time if the following procedure is used. The routines that will not be called are `U_init_dim_obs`, `U_obs_op`, `U_init_obs`, `U_prodRinvA`, `U_init_obsvar`. A simple way to provide them for the compilation could be to copy the corresponding files (i.e. named without `U_`) from the template directory `templates/` and to include these files in the compilation and linking. These templates are simple stubs without any functionality.
     180
     181At this implementation stage one can use the preprocessor definition `PDAF_NO_UPDATE` (available from Version 1.6.1). With this, the actual analysis step of the chosen filter algorithm is not executed. Accordingly, only the user-supplied routines used in `PDAF_get_state` as well as the routine `U_collect_state` need to be implemented with functionality. The other routines will not be executed, because they are only called during the analysis step. Generally with `PDAF_NO_UPDATE` the program performs just an ensemble integration. That is, PDAF is initialized by `PDAF_init`. Then a forecast is computed by using `PDAF_get_state` and the chosen `PDAF_assimilate_X` routine. At the initial time `U_prepoststep` is executed by `PDAF_get_state`. `U_next_obs` will provide the number of time steps to be computed by the model and `U_distributed_state` will initialize the model fields. Subsequently the ensemble integration is performed and the forecast fields are written back to the ensemble array by `U_collect_state`. Upon completion of the forecast phase, the routine `U_prepoststep` is executed twice. The first time is the regular call before the analysis is executed. Thus, it allows to access the forecast ensemble. If the analysis would not be deactivated, the second call to `U_prepoststep` would be after the analysis allowing access to the ensemble directly after the analysis. As the analysis is deactivated here, the ensemble will be the same as in the first call.
     182
     183This test allows to check the following:
     184 * Is `U_prepoststep` working correctly?
     185 * Does `U_next_observation` work correctly and is the information from this routine used correctly for the model integration
     186 * Are `U_distribute_state` and `U_collect_state` work correctly?
     187One could also comment out the actual time stepping part of the model. This would allow to only test the interfacing between PDAF and the model.
     188
     189It is important to ensure that the ensemble integration performs correctly. The simplest case should be a parallel configuration in which the number of model tasks equals the ensemble size as here the model tasks always compute forward in time. If the number of model tasks is smaller than the ensemble size, some model tasks will have to integrate multiple states of the ensemble. If a model task has to integrate two states, the model will have to jump back in time for the integration of the second state. It might be that some arrays of the model need to be re-initialized to ensure that the second integration is consistent. Also, one might need to check if the initialization of forcing fields (e.g. wind stress over the ocean) performs correctly for the second integration. (Sometimes model are implemented with the constraint that the model time always increases, which is the normal case for pure model simulations without assimilation.) A useful test is to initialize an ensemble in which all states are equal. If this ensemble is integrated the forecast states of the ensemble should, of course, still be equal.