Changes between Initial Version and Version 1 of AvailableOptionsforInitPDAFinPDAF3


Ignore:
Timestamp:
Mar 24, 2025, 11:04:16 AM (9 days ago)
Author:
lnerger
Comment:

--

Legend:

Unmodified
Added
Removed
Modified

  • AvailableOptionsforInitPDAFinPDAF3

    v1 v1  
     1= Available options for the different DA methods (PDAF3) =
     2
     3
     4[[PageOutline(2-3,Contents of this page)]]
     5
     6|| This page documents the options for PDAF 3. [[BR]]See the [wiki:AvailableOptionsforInitPDAFuntilPDAF231 page on available options in PDAF 2.31] for the options valid for the ealier releases. ||
     7
     8There are different options for each of the filter algorithms that need to be specified in the call to `PDAF_init` or that canbe set using `PDAF_set_iparam` and `PDAF_set_rparam`. To display the available options in a compiled assimilation program, one can use run with the specification `subtype=-1`. In this case `PDAF_init` wil display the available options for the selected filter algorithm and exit with an error status of -1.
     9
     10The options are mainly unified over the different DA methods. Here, the nonlinear ensemble methods have somewhaat different options from those for the ensemble Kalman filter variants. Also the 3D-Var methods have a different set of options.
     11
     12== Overview of options ==
     13
     14=== Options for ensemble Kalman filters ===
     15
     16Overview of integer options for ensemble Kalman filters. The parameters are explained in the list of options as they are displayed using `subtype=-1`, which are shown further below.
     17
     18Parameters shown in italic font have changed in PDAF V3 compared to PDAF V2.3.1. The index position in PDAF V2.3.1 is shown in parenthesis.
     19
     20||= iparam =||= SEIK =||= LSEIK =||= EnKF =||= LEnKF =||= ETKF =||= LETKF =||= ESTKF =||= LESTKF =||= //ENSRF// =||
     21||= 1 =|| dim_p || dim_p || dim_p || dim_p || dim_p || dim_p || dim_p ||dim_p || dim_p
     22||= 2 =|| dim_ens || dim_ens || dim_ens || dim_ens || dim_ens || dim_ens || dim_ens || dim_ens || dim_ens ||
     23||= 3 =||       ||       || //dim_lag(5)// ||   || dim_lag || dim_lag || dim_lag || dim_lag || ||
     24||= 4 =||       ||       || //rank_ana(3)// || //rank_ana(3)// || || || || || ||
     25||= 5 =|| type_forget || type_forget ||   ||   || type_forget || type_forget || type_forget || type_forget || ||
     26||= 6 =|| type_trans  || type_trans  ||   ||   || type_trans || type_trans || type_trans || type_trans || ||
     27||= 7 =|| type_sqrt   || type_sqrt   ||   ||   ||   ||   || type_sqrt || type_sqrt       || ||
     28||= 8 =|| obs_ens || obs_ens || obs_ens || obs_ens || obs_ens || obs_ens || obs_ens || obs_ens || obs_ens ||
     29||= 9 =|| type_obs || type_obs || type_obs || type_obs || type_obs || type_obs || type_obs || type_obs || type_obs ||
     30
     31Overview of real-valued options for ensemble Kalman filters. The parameters are explained in the list of options as they are displayed using `subtype=-1`, which are shown further below
     32                                                               
     33||= rparam =||= SEIK =||= LSEIK =||= EnKF =||= LEnKF =||= ETKF =||= LETKF =||= ESTKF =||= LESTKF =||= //ENSRF// =||
     34||= 1 =|| forget ||forget || forget || forget || forget || forget || forget || forget || forget ||
     35
     36
     37
     38=== Options for nonlinear DA methods ===
     39
     40Overview of integer options for nonlinear DA methods. The parameters are explained in the list of options as they are displayed using `subtype=-1`, which are shown further below.
     41
     42Parameters shown in italic font have changed in PDAF V3 compared to PDAF V2.3.1. The index position in PDAF V2.3.1 is shown in parenthesis.
     43
     44||= iparam =||= NETF =||= LNETF =||= LKNETF =||= PF =||
     45||= 1 =|| dim_p || dim_p || dim_p || dim_p ||
     46||= 2 =|| dim_ens || dim_ens || dim_ens || dim_ens ||
     47||= 3 =|| dim_lag || dim_lag || ||
     48||= 4 =|| type_noise || type_noise || //type_hyb(7)// || type_noise
     49||= 5 =|| type_forget || type_forget || type_forget || ||
     50||= 6 =|| type_trans || type_trans || type_trans || //type_resample(3)// ||
     51||= 7 =|| type_winf  || type_winf  ||  || //type_winf(6)// ||
     52||= 8 =|| obs_ens    || obs_ens || obs_ens || obs_ens ||
     53||= 9 =|| type_obs   || type_obs || type_obs || type_obs ||
     54
     55
     56Overview of real-valued options:
     57                                       
     58||= rparam =||= NETF =||= LNETF =||= LKNETF =||= PF =||
     59||= 1 =|| forget ||forget || forget || //forget(2)// ||
     60||= 2 =|| limit_winf || limit_winf || hyb_g || //limit_winf(3)// ||
     61||= 3 =|| noise_amp || noise_amp  || hyb_k || //noise_amp(1)// ||
     62
     63
     64=== Options for 3D-Var methods ===
     65
     66Overview of integer options for 3D-Var methods. The parameters are explained in the list of options as they are displayed using `subtype=-1`, which are shown further below.
     67
     68Parameters shown in italic font have changed in PDAF V3 compared to PDAF V2.3.1. The index position in PDAF V2.3.1 is shown in parenthesis.
     69
     70||= iparam =||= 3dvar =||= en3dvar =||= hyb3dvar =||
     71||= 1 =|| dim_p || dim_p || dim_p ||
     72||= 2 =|| dim_ens || dim_ens || dim_ens ||
     73||= 3 =|| type_opt || type_opt || type_opt ||
     74||= 4 =|| dim_cvec ||   || dim_cvec ||
     75||= 5 =||    || dim_cvec_ens || dim_cvec_ens ||
     76||= 6 =|| solver param1 || solver param1 || solver param1 ||
     77||= 7 =|| solver param2 || solver param2 || solver param2 ||
     78||= 8 =|| obs_ens || obs_ens || obs_ens ||
     79||= 9 =|| type_obs || type_obs || type_obs ||
     80||= 10 =|| || || ||
     81||= 11 =|| || type_forget || type_forget ||
     82||= 12 =|| || type_trans  || type_trans ||
     83||= 13 =|| || type_sqrt   || type_sqrt ||
     84
     85Overview of real-valued options:
     86||= rparam =||= 3dvar =||= en3dvar =||= hyb3dvar =||
     87||= 1 =|| forget || forget || forget ||
     88||= 2 =||  ||  || beta ||
     89||= 3 =|| solver param3 || solver param3 || solver param3 ||
     90||= 4 =|| solver param4 || solver param4 || solver param4 ||
     91
     92
     93== Option outputs using subtype=-1 ==
     94
     95=== SEIK (filtertype=PDAF_DA_SEIK=1) ===
     96
     97{{{
     98PDAF     Available options for SEIK:
     99PDAF     --- Sub-types (Parameter subtype) ---
     100PDAF       0: full ensemble integration; left-sided application of T
     101PDAF       1: full ensemble integration; right-sided application of T
     102PDAF       2: full ensemble integration; explicit ensemble transformation
     103PDAF       10: Fixed error space basis
     104PDAF       11: Fixed state covariance matrix
     105PDAF     --- Integer parameters (Array param_int) ---
     106PDAF       param_int(1): Dimension of state vector (>0), required
     107PDAF       param_int(2): Ensemble size (>0), required
     108PDAF       param_int(3): not used
     109PDAF       param_int(4): not used
     110PDAF       param_int(5) type_forget
     111PDAF           Type of forgetting factor; optional
     112PDAF            0: fixed forgetting factor (default)
     113PDAF            1: adaptive forgetting factor (experimental)
     114PDAF       param_int(6) type_trans
     115PDAF           Type of ensemble transformation matrix; optional
     116PDAF            0: deterministic Omega (default)
     117PDAF            1: random orthonormal Omega orthogonal to (1,...,1)^T
     118PDAF            2: use product of 0 with random orthonomal matrix with eigenvector (1,...,1)^T
     119PDAF              (experimental; for random transformations, 1 is recommended)
     120PDAF       param_int(7) type_sqrt
     121PDAF           Type of transformation matrix square root; optional
     122PDAF            (Only relevant for subtype/=11)
     123PDAF            0: symmetric square root (default)
     124PDAF            1: Cholesky decomposition
     125PDAF       param_int(8): observe_ens
     126PDAF           Application of observation operator H, optional
     127PDAF            0: Apply H to ensemble mean to compute innovation
     128PDAF            1: Apply H to ensemble states; then compute innovation from their mean (default)
     129PDAF               param_int(8)=1 is the recomended choice for nonlinear H
     130PDAF       param_int(9): type_obs_init
     131PDAF           Initialize observations before or after call to prepoststep_pdaf
     132PDAF            0: Initialize observations before call to prepoststep_pdaf
     133PDAF            1: Initialize observations after call to prepoststep_pdaf (default)
     134PDAF     --- Floating point parameters (Array param_real) ---
     135PDAF       param_real(1): forget
     136PDAF           Forgetting factor (usually >0 and <=1), required
     137PDAF     --- Further parameters ---
     138PDAF       n_modeltasks: Number of parallel model integration tasks
     139PDAF           >=1 for subtypes 0, 1 and 2; not larger than total number of processors
     140PDAF           =1 required for subtypes 10 and 11
     141PDAF       screen: Control verbosity of PDAF
     142PDAF           0: no outputs
     143PDAF           1: basic output (default)
     144PDAF           2: 1 plus timing output
     145PDAF           3: 2 plus debug output
     146PDAF     --- Internal parameter (defined inside PDAF) ---
     147PDAF       Nm1vsN: Normalization of covariance matrix; default: 1
     148PDAF           0: normalization with 1/(Ensemble size)
     149PDAF              (original SEIK, mainly for compatibility with older studies)
     150PDAF           1: normalization with 1/(Ensemble size - 1)
     151PDAF           (sample covariance matrix consistent with other EnKFs)
     152PDAF     +++++++++ End of option overview for the SEIK filter ++++++++++
     153}}}
     154
     155
     156=== EnKF (filtertype=PDAF_DA_ENKF=2) ===
     157
     158{{{
     159PDAF     Available options for EnKF:
     160PDAF     --- Sub-types (Parameter subtype) ---
     161PDAF       0: Full ensemble integration; analysis for 2*dim_obs>dim_ens
     162PDAF       1: Full ensemble integration; analysis for 2*dim_obs<=dim_ens
     163PDAF     --- Integer parameters (Array param_int) ---
     164PDAF       param_int(1): Dimension of state vector (>0), required
     165PDAF       param_int(2): Ensemble size (>0), required
     166PDAF       param_int(3): dim_lag
     167PDAF           Size of smoothing lag (>=0), optional
     168PDAF           0: no smoothing (default)
     169PDAF           >0: apply smoother up to specified lag
     170PDAF       param_int(4): rank_ana_enkf
     171PDAF           maximum rank for inversion of HPH^T, optional, default=0
     172PDAF            for =0, HPH is inverted by solving the representer equation
     173PDAF            allowed range is 0 to ensemble size - 1
     174PDAF       param_int(5): not used
     175PDAF       param_int(6): not used
     176PDAF       param_int(7): not used
     177PDAF       param_int(8): observe_ens
     178PDAF           Application of observation operator H, optional
     179PDAF            0: Apply H to ensemble mean to compute innovation
     180PDAF            1: Apply H to ensemble states; then compute innovation from their mean (default)
     181PDAF               param_int(8)=1 is the recomended choice for nonlinear H
     182PDAF       param_int(9): type_obs_init
     183PDAF           Initialize observations before or after call to prepoststep_pdaf
     184PDAF            0: Initialize observations before call to prepoststep_pdaf
     185PDAF            1: Initialize observations after call to prepoststep_pdaf (default)
     186PDAF     --- Floating point parameters (Array param_real) ---
     187PDAF       param_real(1): forget
     188PDAF           Forgetting factor (usually >0 and <=1), required
     189PDAF     --- Further parameters ---
     190PDAF       n_modeltasks: Number of parallel model integration tasks
     191PDAF           (>=1; not larger than total number of processors)
     192PDAF       screen: Control verbosity of PDAF
     193PDAF           0: no outputs
     194PDAF           1: basic output (default)
     195PDAF           2: 1 plus timing output
     196PDAF           3: 2 plus debug output
     197PDAF     +++++++++ End of option overview for the EnKF ++++++++++
     198}}}
     199
     200
     201=== LSEIK (filtertype=PDAF_DA_LSEIK=3) ===
     202
     203{{{
     204PDAF     Available options for LSEIK:
     205PDAF     --- Sub-types (Parameter subtype) ---
     206PDAF       0: full ensemble integration; left-sided application of T
     207PDAF       1: full ensemble integration; explicit ensemble transformation
     208PDAF       10: Fixed error space basis
     209PDAF       11: Fixed state covariance matrix
     210PDAF     --- Integer parameters (Array param_int) ---
     211PDAF       param_int(1): Dimension of state vector (>0), required
     212PDAF       param_int(2): Ensemble size (>0), required
     213PDAF       param_int(3): not used
     214PDAF       param_int(4): not used
     215PDAF       param_int(5) type_forget
     216PDAF           Type of forgetting factor; optional
     217PDAF            0: fixed forgetting factor (default)
     218PDAF            1: adaptive forgetting factor
     219PDAF            2: locally adaptive forgetting factor
     220PDAF       param_int(6) type_trans
     221PDAF           Type of ensemble transformation matrix; optional
     222PDAF            0: deterministic Omega (default)
     223PDAF            1: random orthonormal Omega orthogonal to (1,...,1)^T
     224PDAF            2: use product of 0 with random orthonomal matrix with eigenvector (1,...,1)^T
     225PDAF              (experimental; for random transformations, 1 is recommended)
     226PDAF       param_int(7) type_sqrt
     227PDAF           Type of transformation matrix square root; optional
     228PDAF            (Only relevant for subtype/=11)
     229PDAF            0: symmetric square root (default)
     230PDAF            1: Cholesky decomposition
     231PDAF       param_int(8): observe_ens
     232PDAF           Application of observation operator H, optional
     233PDAF            0: Apply H to ensemble mean to compute innovation
     234PDAF            1: Apply H to ensemble states; then compute innovation from their mean (default)
     235PDAF               param_int(8)=1 is the recomended choice for nonlinear H
     236PDAF       param_int(9): type_obs_init
     237PDAF           Initialize observations before or after call to prepoststep_pdaf
     238PDAF            0: Initialize observations before call to prepoststep_pdaf
     239PDAF            1: Initialize observations after call to prepoststep_pdaf (default)
     240PDAF     --- Floating point parameters (Array param_real) ---
     241PDAF       param_real(1): forget
     242PDAF           Forgetting factor (usually >0 and <=1), required
     243PDAF     --- Further parameters ---
     244PDAF       n_modeltasks: Number of parallel model integration tasks
     245PDAF           >=1 for subtypes 0 and 1; not larger than total number of processors
     246PDAF           =1 required for subtypes 10 and 11
     247PDAF       screen: Control verbosity of PDAF
     248PDAF           0: no outputs
     249PDAF           1: basic output (default)
     250PDAF           2: 1 plus timing output
     251PDAF           3: 2 plus debug output
     252PDAF     --- Internal parameter (defined inside PDAF) ---
     253PDAF       Nm1vsN: Normalization of covariance matrix; default: 1
     254PDAF           0: normalization with 1/(Ensemble size)
     255PDAF              (original SEIK, mainly for compatibility with older studies)
     256PDAF           1: normalization with 1/(Ensemble size - 1)
     257PDAF              (sample covariance matrix consistent with other EnKFs)
     258PDAF     +++++++++ End of option overview for the LSEIK filter ++++++++++
     259}}}
     260
     261
     262=== ETKF (filtertype=PDAF_DA_ETKF=4) ===
     263
     264{{{
     265PDAF     Available options for ETKF:
     266PDAF     --- Sub-types (Parameter subtype) ---
     267PDAF       0: full ensemble integration; apply T-matrix analogously to SEIK
     268PDAF       1: full ensemble integration; formulation cf. Hunt et al. (2007) without T matrix
     269PDAF       10: Fixed error space basis; analysis with T-matrix
     270PDAF       11: Fixed state covariance matrix; analysis with T-matrix
     271PDAF     --- Integer parameters (Array param_int) ---
     272PDAF       param_int(1): Dimension of state vector (>0), required
     273PDAF       param_int(2): Ensemble size (>0), required
     274PDAF       param_int(3): dim_lag
     275PDAF           Size of smoothing lag (>=0), optional
     276PDAF            0: no smoothing (default)
     277PDAF            >0: apply smoother up to specified lag
     278PDAF       param_int(4): not used
     279PDAF       param_int(5) type_forget
     280PDAF           Type of forgetting factor; optional
     281PDAF            0: fixed forgetting factor (default)
     282PDAF            1: adaptive forgetting factor (experimental)
     283PDAF       param_int(6) type_trans
     284PDAF           Type of ensemble transformation matrix; optional
     285PDAF            0: deterministic transformation (default)
     286PDAF            2: use product of 0 with random orthonomal matrix with eigenvector (1,...,1)^T
     287PDAF       param_int(7): not used
     288PDAF       param_int(8): observe_ens
     289PDAF           Application of observation operator H, optional
     290PDAF            0: Apply H to ensemble mean to compute innovation
     291PDAF            1: Apply H to ensemble states; then compute innovation from their mean (default)
     292PDAF               param_int(8)=1 is the recomended choice for nonlinear H
     293PDAF       param_int(9): type_obs_init
     294PDAF           Initialize observations before or after call to prepoststep_pdaf
     295PDAF           0: Initialize observations before call to prepoststep_pdaf
     296PDAF           1: Initialize observations after call to prepoststep_pdaf (default)
     297PDAF     --- Floating point parameters (Array param_real) ---
     298PDAF       param_real(1): forget
     299PDAF           Forgetting factor (usually >0 and <=1), required
     300PDAF     --- Further parameters ---
     301PDAF       n_modeltasks: Number of parallel model integration tasks
     302PDAF           >=1 for subtypes 0 and 1; not larger than total number of processors
     303PDAF           =1 required for subtypes 10 and 11
     304PDAF       screen: Control verbosity of PDAF
     305PDAF           0: no outputs
     306PDAF           1: basic output (default)
     307PDAF           2: 1 plus timing output
     308PDAF           3: 2 plus debug output
     309PDAF     +++++++++ End of option overview for the ETKF ++++++++++
     310}}}
     311
     312
     313
     314
     315=== LETKF (filtertype=PDAF_DA_LETKF=5) ===
     316
     317{{{
     318PDAF     Available options for LETKF:
     319PDAF     --- Sub-types (Parameter subtype) ---
     320PDAF       0: full ensemble integration;  apply T-matrix analogously to SEIK
     321PDAF       1: full ensemble integration; formulation cf. Hunt et al. (2007) without T matrix
     322PDAF       10: Fixed error space basis; analysis with T-matrix
     323PDAF       11: Fixed state covariance matrix; analysis with T-matrix
     324PDAF     --- Integer parameters (Array param_int) ---
     325PDAF       param_int(1): Dimension of state vector (>0), required
     326PDAF       param_int(2): Ensemble size (>0), required
     327PDAF       param_int(3): dim_lag
     328PDAF           Size of smoothing lag (>=0), optional
     329PDAF            0: no smoothing (default)
     330PDAF            >0: apply smoother up to specified lag
     331PDAF       param_int(4): not used
     332PDAF       param_int(5) type_forget
     333PDAF           Type of forgetting factor; optional
     334PDAF            0: fixed forgetting factor (default)
     335PDAF            1: adaptive forgetting factor for full domain
     336PDAF            2: locally adaptive forgetting factor
     337PDAF       param_int(6) type_trans
     338PDAF           Type of ensemble transformation matrix; optional
     339PDAF            0: deterministic transformation (default)
     340PDAF            2: use product of 0 with random orthonomal matrix with eigenvector (1,...,1)^T
     341PDAF       param_int(8): observe_ens
     342PDAF           Application of observation operator H, optional
     343PDAF            0: Apply H to ensemble mean to compute innovation
     344PDAF            1: Apply H to ensemble states; then compute innovation from their mean (default)
     345PDAF               param_int(8)=1 is the recomended choice for nonlinear H
     346PDAF       param_int(9): type_obs_init
     347PDAF           Initialize observations before or after call to prepoststep_pdaf
     348PDAF            0: Initialize observations before call to prepoststep_pdaf
     349PDAF            1: Initialize observations after call to prepoststep_pdaf (default)
     350PDAF     --- Floating point parameters (Array param_real) ---
     351PDAF       param_real(1): forget
     352PDAF           Forgetting factor (usually >0 and <=1), required
     353PDAF     --- Further parameters ---
     354PDAF       n_modeltasks: Number of parallel model integration tasks
     355PDAF           >=1 for subtypes 0 and 1; not larger than total number of processors
     356PDAF           =1 required for subtypes 10 and 11
     357PDAF       screen: Control verbosity of PDAF
     358PDAF           0: no outputs
     359PDAF           1: basic output (default)
     360PDAF           2: 1 plus timing output
     361PDAF           3: 2 plus debug output
     362PDAF     +++++++++ End of option overview for the LETKF ++++++++++
     363}}}
     364
     365
     366
     367
     368=== ESTKF (filtertype=PDAF_DA_ESTKF=6) ===
     369
     370{{{
     371PDAF     Available options for ESTKF:
     372PDAF     --- Sub-types (Parameter subtype) ---
     373PDAF       0: Standard implementation with ensemble integration
     374PDAF       10: Fixed error space basis
     375PDAF       11: Fixed state covariance matrix
     376PDAF     --- Integer parameters (Array param_int) ---
     377PDAF       param_int(1): Dimension of state vector (>0), required
     378PDAF       param_int(2): Ensemble size (>0), required
     379PDAF       param_int(3): dim_lag
     380PDAF           Size of smoothing lag (>=0), optional
     381PDAF            0: no smoothing (default)
     382PDAF            >0: apply smoother up to specified lag
     383PDAF       param_int(4): not used
     384PDAF       param_int(5) type_forget
     385PDAF           Type of forgetting factor; optional
     386PDAF            0: fixed forgetting factor (default)
     387PDAF            1: adaptive forgetting factor (experimental)
     388PDAF       param_int(6) type_trans
     389PDAF           Type of ensemble transformation matrix; optional
     390PDAF            0: deterministic Omega (default)
     391PDAF            1: random orthonormal Omega orthogonal to (1,...,1)^T
     392PDAF            2: use product of 0 with random orthonomal matrix with eigenvector (1,...,1)^T
     393PDAF              (experimental; for random transformations, 0 or 1 are recommended)
     394PDAF       param_int(7) type_sqrt
     395PDAF           Type of transformation matrix square root; optional
     396PDAF            0: symmetric square root (default)
     397PDAF            1: Cholesky decomposition
     398PDAF       param_int(8): observe_ens
     399PDAF           Application of observation operator H, optional
     400PDAF            0: Apply H to ensemble mean to compute innovation
     401PDAF            1: Apply H to ensemble states; then compute innovation from their mean (default)
     402PDAF               param_int(8)=1 is the recomended choice for nonlinear H
     403PDAF       param_int(9): type_obs_init
     404PDAF           Initialize observations before or after call to prepoststep_pdaf
     405PDAF           0: Initialize observations before call to prepoststep_pdaf
     406PDAF           1: Initialize observations after call to prepoststep_pdaf (default)
     407PDAF     --- Floating point parameters (Array param_real) ---
     408PDAF       param_real(1): forget
     409PDAF           Forgetting factor (usually >0 and <=1), required
     410PDAF     --- Further parameters ---
     411PDAF       n_modeltasks: Number of parallel model integration tasks
     412PDAF           >=1 for subtype 0; not larger than total number of processors
     413PDAF           =1 required for subtypes 10 and 11
     414PDAF       screen: Control verbosity of PDAF
     415PDAF           0: no outputs
     416PDAF           1: basic output (default)
     417PDAF           2: 1 plus timing output
     418PDAF           3: 2 plus debug output
     419PDAF     +++++++++ End of option overview for the ESTKF  ++++++++++
     420}}}
     421
     422
     423
     424
     425=== LESTKF (filtertype=PDAF_DA_LESTKF=7) ===
     426
     427{{{
     428PDAF     Available options for LESTKF:
     429PDAF     --- Sub-types (Parameter subtype) ---
     430PDAF       0: Standard implementation with ensemble integration
     431PDAF       10: Fixed error space basis
     432PDAF       11: Fixed state covariance matrix
     433PDAF     --- Integer parameters (Array param_int) ---
     434PDAF       param_int(1): Dimension of state vector (>0), required
     435PDAF       param_int(2): Ensemble size (>0), required
     436PDAF       param_int(3): dim_lag
     437PDAF           Size of smoothing lag (>=0), optional
     438PDAF            0: no smoothing (default)
     439PDAF            >0: apply smoother up to specified lag
     440PDAF       param_int(4): not used
     441PDAF       param_int(5) type_forget
     442PDAF           Type of forgetting factor; optional
     443PDAF            0: fixed forgetting factor (default)
     444PDAF            1: adaptive forgetting factor for full domain
     445PDAF            2: locally adaptive forgetting factor
     446PDAF       param_int(6) type_trans
     447PDAF           Type of ensemble transformation matrix; optional
     448PDAF            0: deterministic Omega (default)
     449PDAF            1: random orthonormal Omega orthogonal to (1,...,1)^T
     450PDAF            2: use product of 0 with random orthonomal matrix with eigenvector (1,...,1)^T
     451PDAF              (experimental; for random transformations, 1 is recommended)
     452PDAF       param_int(7) type_sqrt
     453PDAF           Type of transformation matrix square root; optional
     454PDAF            0: symmetric square root (default)
     455PDAF            1: Cholesky decomposition
     456PDAF       param_int(8): observe_ens
     457PDAF           Application of observation operator H, optional
     458PDAF            0: Apply H to ensemble mean to compute innovation
     459PDAF            1: Apply H to ensemble states; then compute innovation from their mean (default)
     460PDAF               param_int(8)=1 is the recomended choice for nonlinear H
     461PDAF       param_int(9): type_obs_init
     462PDAF           Initialize observations before or after call to prepoststep_pdaf
     463PDAF            0: Initialize observations before call to prepoststep_pdaf
     464PDAF            1: Initialize observations after call to prepoststep_pdaf (default)
     465PDAF     --- Floating point parameters (Array param_real) ---
     466PDAF       param_real(1): forget
     467PDAF           Forgetting factor (usually >0 and <=1), required
     468PDAF     --- Further parameters ---
     469PDAF       n_modeltasks: Number of parallel model integration tasks
     470PDAF           >=1 for subtype 0; not larger than total number of processors
     471PDAF           =1 required for subtypes 10 and 11
     472PDAF       screen: Control verbosity of PDAF
     473PDAF           0: no outputs
     474PDAF           1: basic output (default)
     475PDAF           2: 1 plus timing output
     476PDAF           3: 2 plus debug output
     477PDAF     +++++++++ End of option overview for the LESTKF ++++++++++
     478}}}
     479
     480
     481
     482
     483=== LEnKF (filtertype=PDAF_DA_LENKF=8) ===
     484
     485{{{
     486PDAF     Available options for LEnKF:
     487PDAF     --- Sub-types (Parameter subtype) ---
     488PDAF       0: Standard EnKF analysis with covariance localization
     489PDAF     --- Integer parameters (Array param_int) ---
     490PDAF       param_int(1): Dimension of state vector (>0), required
     491PDAF       param_int(2): Ensemble size (>0), required
     492PDAF       param_int(3): not used
     493PDAF       param_int(4): rank_ana_enkf
     494PDAF           maximum rank for inversion of HPH^T, optional, default=0
     495PDAF            for =0, HPH is inverted by solving the representer equation
     496PDAF            allowed range is 0 to ensemble size - 1
     497PDAF       param_int(5): not used
     498PDAF       param_int(6): not used
     499PDAF       param_int(7): not used
     500PDAF       param_int(8): observe_ens
     501PDAF           Application of observation operator H, optional
     502PDAF            0: Apply H to ensemble mean to compute innovation
     503PDAF            1: Apply H to ensemble states; then compute innovation from their mean (default)
     504PDAF               param_int(8)=1 is the recomended choice for nonlinear H
     505PDAF       param_int(9): type_obs_init
     506PDAF           Initialize observations before or after call to prepoststep_pdaf
     507PDAF            0: Initialize observations before call to prepoststep_pdaf
     508PDAF            1: Initialize observations after call to prepoststep_pdaf (default)
     509PDAF     --- Floating point parameters (Array param_real) ---
     510PDAF       param_real(1): forget
     511PDAF           Forgetting factor (usually >0 and <=1), required
     512PDAF     --- Further parameters ---
     513PDAF       n_modeltasks: Number of parallel model integration tasks
     514PDAF           (>=1; not larger than total number of processors)
     515PDAF       screen: Control verbosity of PDAF
     516PDAF           0: no outputs
     517PDAF           1: basic output (default)
     518PDAF           2: 1 plus timing output
     519PDAF           3: 2 plus debug output
     520PDAF     +++++++++ End of option overview for the LEnKF ++++++++++
     521}}}
     522
     523
     524
     525
     526=== NETF (filtertype=PDAF_DA_NETF=9) ===
     527
     528{{{
     529PDAF     Available options for NETF:
     530PDAF     --- Sub-types (Parameter subtype) ---
     531PDAF       0: Standard implementation with ensemble integration
     532PDAF     --- Integer parameters (Array param_int) ---
     533PDAF       param_int(1): Dimension of state vector (>0), required
     534PDAF       param_int(2): Ensemble size (>0), required
     535PDAF       param_int(3): dim_lag
     536PDAF           Size of smoothing lag (>=0), optional
     537PDAF            0: no smoothing (default)
     538PDAF            >0: apply smoother up to specified lag
     539PDAF       param_int(4): type_noise
     540PDAF           Type of ensemble perturbations, optional
     541PDAF            0: no perturbations (default)
     542PDAF            1: constant standard deviation
     543PDAF            2: relative to ensemble standard deviation
     544PDAF       param_int(5) type_forget
     545PDAF           Type of forgetting factor; optional
     546PDAF            0: forgetting factor on forecast ensemble (default)
     547PDAF            2: forgetting factor on analysis ensemble
     548PDAF       param_int(6) type_trans
     549PDAF           Type of ensemble transformation matrix; optional
     550PDAF            0: random orthonormal matrix orthogonal to (1,...,1)^T (default)
     551PDAF            1: deterministic transformation
     552PDAF       param_int(7): type_winf
     553PDAF           Type of weights inflation; optional
     554PDAF            0: no weights inflation (default)
     555PDAF            1: inflate so that N_eff/N > param_real(2)
     556PDAF       param_int(8): observe_ens
     557PDAF           Application of observation operator H, optional
     558PDAF           Note: This parameter has no influence on the NETF assimilation result
     559PDAF            0: Apply H to ensemble mean to compute innovation
     560PDAF            1: Apply H to ensemble states; then compute innovation from their mean (default)
     561PDAF               param_int(8)=1 is the recomended choice for nonlinear H
     562PDAF       param_int(9): type_obs_init
     563PDAF           Initialize observations before or after call to prepoststep_pdaf
     564PDAF           0: Initialize observations before call to prepoststep_pdaf
     565PDAF           1: Initialize observations after call to prepoststep_pdaf (default)
     566PDAF     --- Floating point parameters (Array param_real) ---
     567PDAF       param_real(1): forget
     568PDAF           Forgetting factor (usually >0 and <=1), required
     569PDAF       param_real(2): limit_winf
     570PDAF           Limit for weigts inflation N_eff/N > param_real(2), optional, default=0.0
     571PDAF       param_real(3): noise_amp
     572PDAF           Ensemble perturbation level (>0), required, only used if param_int(4)>0
     573PDAF     --- Further parameters ---
     574PDAF       n_modeltasks: Number of parallel model integration tasks
     575PDAF           >=1; not larger than total number of processors
     576PDAF       screen: Control verbosity of PDAF
     577PDAF           0: no outputs
     578PDAF           1: basic output (default)
     579PDAF           2: 1 plus timing output
     580PDAF           3: 2 plus debug output
     581PDAF     +++++++++ End of option overview for the NETF  ++++++++++
     582}}}
     583
     584
     585
     586
     587=== LNETF (filtertype=PDAF_DA_LNETF=10) ===
     588
     589{{{
     590PDAF     Available options for LNETF:
     591PDAF     --- Sub-types (Parameter subtype) ---
     592PDAF       0: Standard implementation with ensemble integration
     593PDAF     --- Integer parameters (Array param_int) ---
     594PDAF       param_int(1): Dimension of state vector (>0), required
     595PDAF       param_int(2): Ensemble size (>0), required
     596PDAF       param_int(3): dim_lag
     597PDAF           Size of smoothing lag (>=0), optional
     598PDAF            0: no smoothing (default)
     599PDAF            >0: apply smoother up to specified lag
     600PDAF       param_int(4): type_noise
     601PDAF           Type of ensemble perturbations, optional
     602PDAF            0: no perturbations (default)
     603PDAF            1: constant standard deviation
     604PDAF            2: relative to ensemble standard deviation
     605PDAF       param_int(5) type_forget
     606PDAF           Type of forgetting factor; optional
     607PDAF            0: forgetting factor on forecast ensemble (default)
     608PDAF            1: forgetting factor on forecast ensemble only observed domains
     609PDAF            2: forgetting factor on analysis ensemble
     610PDAF            3: forgetting factor on analysis ensemble only observed domains
     611PDAF       param_int(6) type_trans
     612PDAF           Type of ensemble transformation matrix; optional
     613PDAF            0: random orthonormal matrix orthogonal to (1,...,1)^T (default)
     614PDAF            1: deterministic transformation
     615PDAF       param_int(7): type_winf
     616PDAF           Type of weights inflation; optional
     617PDAF            0: no weights inflation (default)
     618PDAF            1: inflate so that N_eff/N > param_real(2)
     619PDAF       param_int(8): observe_ens
     620PDAF           Application of observation operator H, optional
     621PDAF           Note: This parameter has no influence on the LNETF assimilation result
     622PDAF            0: Apply H to ensemble mean to compute innovation
     623PDAF            1: Apply H to ensemble states; then compute innovation from their mean (default)
     624PDAF               param_int(8)=1 is the recomended choice for nonlinear H
     625PDAF       param_int(9): type_obs_init
     626PDAF           Initialize observations before or after call to prepoststep_pdaf
     627PDAF           0: Initialize observations before call to prepoststep_pdaf
     628PDAF           1: Initialize observations after call to prepoststep_pdaf (default)
     629PDAF     --- Floating point parameters (Array param_real) ---
     630PDAF       param_real(1): forget
     631PDAF           Forgetting factor (usually >0 and <=1), required
     632PDAF       param_real(2): limit_winf
     633PDAF           Limit for weigts inflation N_eff/N > param_real(2), optional, default=0.0
     634PDAF       param_real(3): noise_amp
     635PDAF           Ensemble perturbation level (>0), required, only used if param_int(4)>0
     636PDAF     --- Further parameters ---
     637PDAF       n_modeltasks: Number of parallel model integration tasks
     638PDAF           >=1; not larger than total number of processors
     639PDAF       screen: Control verbosity of PDAF
     640PDAF           0: no outputs
     641PDAF           1: basic output (default)
     642PDAF           2: 1 plus timing output
     643PDAF           3: 2 plus debug output
     644PDAF     +++++++++ End of option overview for the LNETF  ++++++++++
     645}}}
     646
     647
     648
     649
     650=== LKNETF (filtertype=PDAF_DA_LKNETF=11) ===
     651
     652{{{
     653PDAF     Available options for LKNETF:
     654PDAF     --- Sub-types (Parameter subtype) ---
     655PDAF       0: HNK: 2-step LKNETF with NETF before LETKF
     656PDAF       1: HKN: 2-step LKNETF with LETKF before NETF
     657PDAF       2: HSync: LKNETF synchronous
     658PDAF     --- Integer parameters (Array param_int) ---
     659PDAF       param_int(1): Dimension of state vector (>0), required
     660PDAF       param_int(2): Ensemble size (>0), required
     661PDAF       param_int(3): not used
     662PDAF       param_int(4): type_hyb
     663PDAF           Type of hybrid weight; optional
     664PDAF            0: fixed value
     665PDAF            1: gamma_lin: (1 - N_eff/N_e)*param_real(2) (default)
     666PDAF            2: gamma_alpha: hybrid weight from N_eff/N>=param_real(2)
     667PDAF            3: gamma_ska: 1 - min(s,k)/sqrt(param_real(3)) with N_eff/N>=param_real(2)
     668PDAF            4: gamma_sklin: 1 - min(s,k)/sqrt(param_real(3)) >= 1-N_eff/N>=param_real(2)
     669PDAF       param_int(5): type_forget
     670PDAF           Type of forgetting factor; optional
     671PDAF            0: inflate forecast ensemble by 1/forget (default)
     672PDAF            1: inflate forecast ensemble by 1/forget only observed domains
     673PDAF            2: inflate analysis ensemble by 1/forget
     674PDAF            3: inflate analysis ensemble by 1/forget only observed domains
     675PDAF            5: adaptive forgetting factor for full domain in LETKF part
     676PDAF            6: locally adaptive forgetting factor in LETKF part
     677PDAF       param_int(6): type_trans
     678PDAF           Type of ensemble transformation matrix; optional
     679PDAF            0: random orthonormal matrix orthogonal to (1,...,1)^T (default)
     680PDAF            1: deterministic transformation
     681PDAF       param_int(7): not used
     682PDAF     --- Floating point parameters (Array param_real) ---
     683PDAF       param_real(1): forget
     684PDAF           Forgetting factor (usually >0 and <=1), required
     685PDAF       param_real(2): hyb_g
     686PDAF           prescribed hybrid weight gamma (usually >0 and <=1), optional, default=0.95
     687PDAF       param_real(3): hyb_k
     688PDAF           hybrid norm kappa (>0), optional, default=dim_ens
     689PDAF     --- Further parameters ---
     690PDAF       n_modeltasks: Number of parallel model integration tasks
     691PDAF           >=1; not larger than total number of processors
     692PDAF       screen: Control verbosity of PDAF
     693PDAF           0: no outputs
     694PDAF           1: basic output (default)
     695PDAF           2: 1 plus timing output
     696PDAF           3: 2 plus debug output
     697PDAF     +++++++++ End of option overview for the LKNETF  ++++++++++
     698}}}
     699
     700
     701
     702
     703=== PF (filtertype=PDAF_DA_PF=12) ===
     704
     705{{{
     706PDAF     Available options for PF:
     707PDAF     --- Sub-types (Parameter subtype) ---
     708PDAF       0: Standard implementation with ensemble integration
     709PDAF     --- Integer parameters (Array param_int) ---
     710PDAF       param_int(1): Dimension of state vector (>0), required
     711PDAF       param_int(2): Ensemble size (>0), required
     712PDAF       param_int(3): type_resample
     713PDAF           Resampling type, optional
     714PDAF            1: probabilistic resamping (default)
     715PDAF            2: stochastic universal resampling
     716PDAF            3: residual resampling
     717PDAF       param_int(4): type_noise
     718PDAF           Type of ensemble perturbations, optional
     719PDAF            0: no perturbations (default)
     720PDAF            1: constant standard deviation
     721PDAF            2: relative to ensemble standard deviation
     722PDAF       param_int(5) type_forget
     723PDAF           Type of forgetting factor; optional
     724PDAF            0: forgetting factor on forecast ensemble (default)
     725PDAF            2: forgetting factor on analysis ensemble
     726PDAF       param_int(6): not used
     727PDAF       param_int(7): type_winf
     728PDAF           Type of weights inflation; optional
     729PDAF            0: no weights inflation (default)
     730PDAF            1: inflate so that N_eff/N > param_real(2)
     731PDAF       param_int(8): observe_ens
     732PDAF           Application of observation operator H, optional
     733PDAF           Note: This parameter has no influence on the PF assimilation result
     734PDAF            0: Apply H to ensemble mean to compute innovation
     735PDAF            1: Apply H to ensemble states; then compute innovation from their mean (default)
     736PDAF               param_int(8)=1 is the recomended choice for nonlinear H
     737PDAF       param_int(9): type_obs_init
     738PDAF           Initialize observations before or after call to prepoststep_pdaf
     739PDAF           0: Initialize observations before call to prepoststep_pdaf
     740PDAF           1: Initialize observations after call to prepoststep_pdaf (default)
     741PDAF     --- Floating point parameters (Array param_real) ---
     742PDAF       param_real(1): forget
     743PDAF           Forgetting factor (usually >0 and <=1), required
     744PDAF       param_real(2): limit_winf
     745PDAF           Limit for weigts inflation N_eff/N > param_real(2), optional, default=0.0
     746PDAF       param_real(3): noise_amp
     747PDAF           Ensemble perturbation level (>0), required, only used if param_int(4)>0
     748PDAF     --- Further parameters ---
     749PDAF       n_modeltasks: Number of parallel model integration tasks
     750PDAF           >=1; not larger than total number of processors
     751PDAF       screen: Control verbosity of PDAF
     752PDAF           0: no outputs
     753PDAF           1: basic output (default)
     754PDAF           2: 1 plus timing output
     755PDAF           3: 2 plus debug output
     756PDAF     +++++++++ End of option overview for the PF  ++++++++++
     757}}}
     758
     759
     760
     761
     762=== ENSRF/EAKF (filtertype=PDAF_DA_ENSRF=13) ===
     763
     764{{{
     765PDAF     Available options for ENSRF:
     766PDAF     --- Sub-types (Parameter subtype) ---
     767PDAF       0: ENSRF with serial observation processing (cf. Houtekamer/Mitchell, 2002)
     768PDAF       1: EAKF/2-step local least squares filter (cf. Anderson, 2003)
     769PDAF     --- Integer parameters (Array param_int) ---
     770PDAF       param_int(1): Dimension of state vector (>0), required
     771PDAF       param_int(2): Ensemble size (>0), required
     772PDAF       param_int(3): not used
     773PDAF       param_int(4): not used
     774PDAF       param_int(5): not used
     775PDAF       param_int(6): not used
     776PDAF       param_int(7): not used
     777PDAF       param_int(8): observe_ens
     778PDAF           Application of observation operator H, optional
     779PDAF            0: Apply H to ensemble mean to compute innovation
     780PDAF            1: Apply H to ensemble states; then compute innovation from their mean (default)
     781PDAF               param_int(8)=1 is the recomended choice for nonlinear H
     782PDAF       param_int(9): type_obs_init
     783PDAF           Initialize observations before or after call to prepoststep_pdaf
     784PDAF            0: Initialize observations before call to prepoststep_pdaf
     785PDAF            1: Initialize observations after call to prepoststep_pdaf (default)
     786PDAF     --- Floating point parameters (Array param_real) ---
     787PDAF       param_real(1): forget
     788PDAF           Forgetting factor (usually >0 and <=1), required
     789PDAF     --- Further parameters ---
     790PDAF       n_modeltasks: Number of parallel model integration tasks
     791PDAF           (>=1; not larger than total number of processors)
     792PDAF       screen: Control verbosity of PDAF
     793PDAF           0: no outputs
     794PDAF           1: basic output (default)
     795PDAF           2: 1 plus timing output
     796PDAF           3: 2 plus debug output
     797PDAF     +++++++++ End of option overview for the ENSRF ++++++++++
     798}}}
     799
     800
     801
     802
     803=== GENOBS (filtertype=PDAF_DA_GENOBS=100) ===
     804
     805{{{
     806PDAF     Available options for GENOBS:
     807PDAF     --- Sub-types (Parameter subtype) ---
     808PDAF       0: Standard implementation with ensemble integration
     809PDAF     --- Integer parameters (Array param_int) ---
     810PDAF       param_int(1): Dimension of state vector (>0), required
     811PDAF       param_int(2): Ensemble size (>0), required
     812PDAF       param_int(3): seedset
     813PDAF           seed set index for random number generator, optional
     814PDAF           valid are values between 1 and 20; default=1
     815PDAF     --- Floating point parameters (Array param_real) ---
     816PDAF       param_real(1): Forgetting factor (usually >0 and <=1), required, but not used
     817PDAF     --- Further parameters ---
     818PDAF       n_modeltasks: Number of parallel model integration tasks
     819PDAF           =1 for GENOBS; not larger than total number of processors
     820PDAF           =1 required for subtypes 10 and 11
     821PDAF       screen: Control verbosity of PDAF
     822PDAF           0: no outputs
     823PDAF           1: basic output (default)
     824PDAF           2: 1 plus timing output
     825PDAF           3: 2 plus debug output
     826PDAF     +++++++++ End of option overview for GENOBS  ++++++++++
     827}}}
     828
     829
     830
     831
     832
     833=== 3D-Var (filtertype=PDAF_DA_3DVAR=200) ===
     834
     835{{{
     836PDAF     Available options for 3D-Var:
     837PDAF     --- Sub-types (Parameter subtype) ---
     838PDAF       0: incremental 3D-Var with parameterized covariance matrix
     839PDAF       1: 3D ensemble Var using LESTKF for ensemble transformation
     840PDAF       2: 3D ensemble Var using ESTKF for ensemble transformation
     841PDAF       3: hybrid 3D-Var using LESTKF for ensemble transformation
     842PDAF       4: hybrid 3D-Var using ESTKF for ensemble transformation
     843PDAF     --- Integer parameters (Array param_int) ---
     844PDAF       param_int(1): Dimension of state vector (>0), required
     845PDAF       param_int(2): Ensemble size (>0), required
     846PDAF       param_int(3): type_opt
     847PDAF           Select optimization method (solver), required
     848PDAF            1: LBFGS (default)
     849PDAF            2: CG+
     850PDAF            3: direct implementation of CG
     851PDAF            12: CG+ with decomposed control vector
     852PDAF            13: direct implementation of CG with decomposed control vector
     853PDAF       param_int(4): size of parameterized control vector (for 3D-Var and hybrid 3D-Var), required
     854PDAF       param_int(5): size of ensemble control vector (required for ensemble and hybrid 3D-Var),
     855PDAF       param_int(6): Solver-specific parameter, optional
     856PDAF           LBFGS: parameter m (default=5)
     857PDAF                Number of corrections used in limited memory matrix; 3<=m<=20
     858PDAF           CG+: parameter method (default=2)
     859PDAF                (1) Fletcher-Reeves, (2) Polak-Ribiere, (3) positive Polak-Ribiere
     860PDAF           CG: maximum number of iterations (default=200)
     861PDAF       param_int(7): Solver-specific parameter, optional
     862PDAF           CG+: parameter irest (default=1)
     863PDAF                (0) no restarts; (n>0) restart every n steps
     864PDAF       param_int(8): observe_ens
     865PDAF           Application of observation operator H, optional
     866PDAF            0: Apply H to ensemble mean to compute innovation
     867PDAF            1: Apply H to ensemble states; then compute innovation from their mean (default)
     868PDAF               param_int(8)=1 is the recomended choice for nonlinear H
     869PDAF       param_int(9): type_obs_init
     870PDAF           Initialize observations before or after call to prepoststep_pdaf
     871PDAF           0: Initialize observations before call to prepoststep_pdaf
     872PDAF           1: Initialize observations after call to prepoststep_pdaf (default)
     873PDAF       param_int(10): not used
     874PDAF       ___Options for ESTKF/LESTKF for En3DVar/hyb3DVar___
     875PDAF       param_int(11) type_forget
     876PDAF           Type of forgetting factor; optional
     877PDAF            0: fixed forgetting factor (default)
     878PDAF            1: adaptive forgetting factor (experimental)
     879PDAF            2: locally adaptive forgetting factor (experimental)
     880PDAF       param_int(12) type_trans
     881PDAF           Type of ensemble transformation matrix; optional
     882PDAF            0: deterministic Omega (default)
     883PDAF            1: random orthonormal Omega orthogonal to (1,...,1)^T
     884PDAF            2: use product of 0 with random orthonomal matrix with eigenvector (1,...,1)^T
     885PDAF              (experimental; for random transformations, 0 or 1 are recommended)
     886PDAF       param_int(13) type_sqrt
     887PDAF           Type of transformation matrix square root; optional
     888PDAF            0: symmetric square root (default)
     889PDAF            1: Cholesky decomposition
     890PDAF     --- Floating point parameters (Array param_real) ---
     891PDAF       param_real(1): Forgetting factor (usually >0 and <=1), required;
     892PDAF           (only used for ensemble and hybrid 3D-Var)
     893PDAF       param_real(2): hybrid weight beta, optional (only for hybrid 3D-Var)
     894PDAF           range >=0.0 and <=1.0, =1.0 for pure ensemble 3D-var  (default=0.5)
     895PDAF       param_real(3): Solver-specific parameter, optional
     896PDAF           LBFGS: Limit for stopping iterations (pgtol, default=1.0e-5)
     897PDAF           CG+: convergence parameter eps (default=1.0e-5)
     898PDAF           CG: convergence parameter eps (default=1.0e-6)
     899PDAF       param_real(4): Solver-specific parameter, optional
     900PDAF           LBFGS: Tolerance in termination test (factr, default=1.0e+7)
     901PDAF     --- Further parameters ---
     902PDAF       n_modeltasks: Number of parallel model integration tasks
     903PDAF           >=1 for subtypes >0; not larger than total number of processors
     904PDAF           =1 required for subtype 0
     905PDAF       screen: Control verbosity of PDAF
     906PDAF           0: no outputs
     907PDAF           1: basic output (default)
     908PDAF           2: 1 plus timing output
     909PDAF           3: 2 plus debug output
     910PDAF     +++++++++ End of option overview for 3DVAR ++++++++++
     911}}}