= Available options for the different DA methods (PDAF3) = [[PageOutline(2-3,Contents of this page)]] || 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. || There 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. The 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. == Overview of options == === Options for ensemble Kalman filters === Overview 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. Parameters 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. ||= iparam =||= SEIK =||= LSEIK =||= EnKF =||= LEnKF =||= ETKF =||= LETKF =||= ESTKF =||= LESTKF =||= //ENSRF// =|| ||= 1 =|| dim_p || dim_p || dim_p || dim_p || dim_p || dim_p || dim_p ||dim_p || dim_p ||= 2 =|| dim_ens || dim_ens || dim_ens || dim_ens || dim_ens || dim_ens || dim_ens || dim_ens || dim_ens || ||= 3 =|| || || //dim_lag(5)// || || dim_lag || dim_lag || dim_lag || dim_lag || || ||= 4 =|| || || //rank_ana(3)// || //rank_ana(3)// || || || || || || ||= 5 =|| type_forget || type_forget || || || type_forget || type_forget || type_forget || type_forget || || ||= 6 =|| type_trans || type_trans || || || type_trans || type_trans || type_trans || type_trans || || ||= 7 =|| type_sqrt || type_sqrt || || || || || type_sqrt || type_sqrt || || ||= 8 =|| obs_ens || obs_ens || obs_ens || obs_ens || obs_ens || obs_ens || obs_ens || obs_ens || obs_ens || ||= 9 =|| type_obs || type_obs || type_obs || type_obs || type_obs || type_obs || type_obs || type_obs || type_obs || Overview 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 ||= rparam =||= SEIK =||= LSEIK =||= EnKF =||= LEnKF =||= ETKF =||= LETKF =||= ESTKF =||= LESTKF =||= //ENSRF// =|| ||= 1 =|| forget ||forget || forget || forget || forget || forget || forget || forget || forget || === Options for nonlinear DA methods === Overview 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. Parameters 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. ||= iparam =||= NETF =||= LNETF =||= LKNETF =||= PF =|| ||= 1 =|| dim_p || dim_p || dim_p || dim_p || ||= 2 =|| dim_ens || dim_ens || dim_ens || dim_ens || ||= 3 =|| dim_lag || dim_lag || || ||= 4 =|| type_noise || type_noise || //type_hyb(7)// || type_noise ||= 5 =|| type_forget || type_forget || type_forget || || ||= 6 =|| type_trans || type_trans || type_trans || //type_resample(3)// || ||= 7 =|| type_winf || type_winf || || //type_winf(6)// || ||= 8 =|| obs_ens || obs_ens || obs_ens || obs_ens || ||= 9 =|| type_obs || type_obs || type_obs || type_obs || Overview of real-valued options: ||= rparam =||= NETF =||= LNETF =||= LKNETF =||= PF =|| ||= 1 =|| forget ||forget || forget || //forget(2)// || ||= 2 =|| limit_winf || limit_winf || hyb_g || //limit_winf(3)// || ||= 3 =|| noise_amp || noise_amp || hyb_k || //noise_amp(1)// || === Options for 3D-Var methods === Overview 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. Parameters 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. ||= iparam =||= 3dvar =||= en3dvar =||= hyb3dvar =|| ||= 1 =|| dim_p || dim_p || dim_p || ||= 2 =|| dim_ens || dim_ens || dim_ens || ||= 3 =|| type_opt || type_opt || type_opt || ||= 4 =|| dim_cvec || || dim_cvec || ||= 5 =|| || dim_cvec_ens || dim_cvec_ens || ||= 6 =|| solver param1 || solver param1 || solver param1 || ||= 7 =|| solver param2 || solver param2 || solver param2 || ||= 8 =|| obs_ens || obs_ens || obs_ens || ||= 9 =|| type_obs || type_obs || type_obs || ||= 10 =|| || || || ||= 11 =|| || type_forget || type_forget || ||= 12 =|| || type_trans || type_trans || ||= 13 =|| || type_sqrt || type_sqrt || Overview of real-valued options: ||= rparam =||= 3dvar =||= en3dvar =||= hyb3dvar =|| ||= 1 =|| forget || forget || forget || ||= 2 =|| || || beta || ||= 3 =|| solver param3 || solver param3 || solver param3 || ||= 4 =|| solver param4 || solver param4 || solver param4 || In the overview 'solver param1' to 'solver param4' denote different integer and real-valued parameters that control the different solvers used in the 3D-Var methods. == Option outputs using subtype=-1 - Local Ensemble Kalman filters == === LESTKF (filtertype=PDAF_DA_LESTKF=7) === {{{ PDAF Available options for LESTKF: PDAF --- Sub-types (Parameter subtype) --- PDAF 0: Standard implementation with ensemble integration PDAF 10: Fixed error space basis PDAF 11: Fixed state covariance matrix PDAF --- Integer parameters (Array param_int) --- PDAF param_int(1): dim_p PDAF Dimension of state vector (>0), required PDAF param_int(2): dim_ens PDAF Ensemble size (>0), required PDAF param_int(3): dim_lag PDAF Size of smoothing lag (>=0), optional PDAF 0: no smoothing (default) PDAF >0: apply smoother up to specified lag PDAF param_int(4): not used PDAF param_int(5) type_forget PDAF Type of forgetting factor; optional PDAF 0: fixed forgetting factor (default) PDAF 1: adaptive forgetting factor for full domain PDAF 2: locally adaptive forgetting factor PDAF param_int(6) type_trans PDAF Type of ensemble transformation matrix; optional PDAF 0: deterministic Omega (default) PDAF 1: random orthonormal Omega orthogonal to (1,...,1)^T PDAF 2: use product of 0 with random orthonomal matrix with eigenvector (1,...,1)^T PDAF (experimental; for random transformations, 1 is recommended) PDAF param_int(7) type_sqrt PDAF Type of transformation matrix square root; optional PDAF 0: symmetric square root (default) PDAF 1: Cholesky decomposition PDAF param_int(8): observe_ens PDAF Application of observation operator H, optional PDAF 0: Apply H to ensemble mean to compute innovation PDAF 1: Apply H to ensemble states; then compute innovation from their mean (default) PDAF param_int(8)=1 is the recomended choice for nonlinear H PDAF param_int(9): type_obs_init PDAF Initialize observations before or after call to prepoststep_pdaf PDAF 0: Initialize observations before call to prepoststep_pdaf PDAF 1: Initialize observations after call to prepoststep_pdaf (default) PDAF --- Floating point parameters (Array param_real) --- PDAF param_real(1): forget PDAF Forgetting factor (usually >0 and <=1), required PDAF --- Further parameters --- PDAF n_modeltasks: Number of parallel model integration tasks PDAF >=1 for subtype 0; not larger than total number of processors PDAF =1 required for subtypes 10 and 11 PDAF screen: Control verbosity of PDAF PDAF 0: no outputs PDAF 1: basic output (default) PDAF 2: 1 plus timing output PDAF 3: 2 plus debug output PDAF +++++++++ End of option overview for the LESTKF ++++++++++ }}} === LETKF (filtertype=PDAF_DA_LETKF=5) === {{{ PDAF Available options for LETKF: PDAF --- Sub-types (Parameter subtype) --- PDAF 0: full ensemble integration; apply T-matrix analogously to SEIK PDAF 1: full ensemble integration; formulation cf. Hunt et al. (2007) without T matrix PDAF 10: Fixed error space basis; analysis with T-matrix PDAF 11: Fixed state covariance matrix; analysis with T-matrix PDAF --- Integer parameters (Array param_int) --- PDAF param_int(1): dim_p PDAF Dimension of state vector (>0), required PDAF param_int(2): dim_ens PDAF Ensemble size (>0), required PDAF param_int(3): dim_lag PDAF Size of smoothing lag (>=0), optional PDAF 0: no smoothing (default) PDAF >0: apply smoother up to specified lag PDAF param_int(4): not used PDAF param_int(5) type_forget PDAF Type of forgetting factor; optional PDAF 0: fixed forgetting factor (default) PDAF 1: adaptive forgetting factor for full domain PDAF 2: locally adaptive forgetting factor PDAF param_int(6) type_trans PDAF Type of ensemble transformation matrix; optional PDAF 0: deterministic transformation (default) PDAF 2: use product of 0 with random orthonomal matrix with eigenvector (1,...,1)^T PDAF param_int(8): observe_ens PDAF Application of observation operator H, optional PDAF 0: Apply H to ensemble mean to compute innovation PDAF 1: Apply H to ensemble states; then compute innovation from their mean (default) PDAF param_int(8)=1 is the recomended choice for nonlinear H PDAF param_int(9): type_obs_init PDAF Initialize observations before or after call to prepoststep_pdaf PDAF 0: Initialize observations before call to prepoststep_pdaf PDAF 1: Initialize observations after call to prepoststep_pdaf (default) PDAF --- Floating point parameters (Array param_real) --- PDAF param_real(1): forget PDAF Forgetting factor (usually >0 and <=1), required PDAF --- Further parameters --- PDAF n_modeltasks: Number of parallel model integration tasks PDAF >=1 for subtypes 0 and 1; not larger than total number of processors PDAF =1 required for subtypes 10 and 11 PDAF screen: Control verbosity of PDAF PDAF 0: no outputs PDAF 1: basic output (default) PDAF 2: 1 plus timing output PDAF 3: 2 plus debug output PDAF +++++++++ End of option overview for the LETKF ++++++++++ }}} === LEnKF (filtertype=PDAF_DA_LENKF=8) === {{{ PDAF Available options for LEnKF: PDAF --- Sub-types (Parameter subtype) --- PDAF 0: Standard EnKF analysis with covariance localization PDAF --- Integer parameters (Array param_int) --- PDAF param_int(1): dim_p PDAF Dimension of state vector (>0), required PDAF param_int(2): dim_ens PDAF Ensemble size (>0), required PDAF param_int(3): not used PDAF param_int(4): rank_ana_enkf PDAF maximum rank for inversion of HPH^T, optional, default=0 PDAF for =0, HPH is inverted by solving the representer equation PDAF allowed range is 0 to ensemble size - 1 PDAF param_int(5): not used PDAF param_int(6): not used PDAF param_int(7): not used PDAF param_int(8): observe_ens PDAF Application of observation operator H, optional PDAF 0: Apply H to ensemble mean to compute innovation PDAF 1: Apply H to ensemble states; then compute innovation from their mean (default) PDAF param_int(8)=1 is the recomended choice for nonlinear H PDAF param_int(9): type_obs_init PDAF Initialize observations before or after call to prepoststep_pdaf PDAF 0: Initialize observations before call to prepoststep_pdaf PDAF 1: Initialize observations after call to prepoststep_pdaf (default) PDAF --- Floating point parameters (Array param_real) --- PDAF param_real(1): forget PDAF Forgetting factor (usually >0 and <=1), required PDAF --- Further parameters --- PDAF n_modeltasks: Number of parallel model integration tasks PDAF (>=1; not larger than total number of processors) PDAF screen: Control verbosity of PDAF PDAF 0: no outputs PDAF 1: basic output (default) PDAF 2: 1 plus timing output PDAF 3: 2 plus debug output PDAF +++++++++ End of option overview for the LEnKF ++++++++++ }}} === LSEIK (filtertype=PDAF_DA_LSEIK=3) === {{{ PDAF Available options for LSEIK: PDAF --- Sub-types (Parameter subtype) --- PDAF 0: full ensemble integration; left-sided application of T PDAF 1: full ensemble integration; explicit ensemble transformation PDAF 10: Fixed error space basis PDAF 11: Fixed state covariance matrix PDAF --- Integer parameters (Array param_int) --- PDAF param_int(1): dim_p PDAF Dimension of state vector (>0), required PDAF param_int(2): dim_ens PDAF Ensemble size (>0), required PDAF param_int(3): not used PDAF param_int(4): not used PDAF param_int(5) type_forget PDAF Type of forgetting factor; optional PDAF 0: fixed forgetting factor (default) PDAF 1: adaptive forgetting factor PDAF 2: locally adaptive forgetting factor PDAF param_int(6) type_trans PDAF Type of ensemble transformation matrix; optional PDAF 0: deterministic Omega (default) PDAF 1: random orthonormal Omega orthogonal to (1,...,1)^T PDAF 2: use product of 0 with random orthonomal matrix with eigenvector (1,...,1)^T PDAF (experimental; for random transformations, 1 is recommended) PDAF param_int(7) type_sqrt PDAF Type of transformation matrix square root; optional PDAF (Only relevant for subtype/=11) PDAF 0: symmetric square root (default) PDAF 1: Cholesky decomposition PDAF param_int(8): observe_ens PDAF Application of observation operator H, optional PDAF 0: Apply H to ensemble mean to compute innovation PDAF 1: Apply H to ensemble states; then compute innovation from their mean (default) PDAF param_int(8)=1 is the recomended choice for nonlinear H PDAF param_int(9): type_obs_init PDAF Initialize observations before or after call to prepoststep_pdaf PDAF 0: Initialize observations before call to prepoststep_pdaf PDAF 1: Initialize observations after call to prepoststep_pdaf (default) PDAF --- Floating point parameters (Array param_real) --- PDAF param_real(1): forget PDAF Forgetting factor (usually >0 and <=1), required PDAF --- Further parameters --- PDAF n_modeltasks: Number of parallel model integration tasks PDAF >=1 for subtypes 0 and 1; not larger than total number of processors PDAF =1 required for subtypes 10 and 11 PDAF screen: Control verbosity of PDAF PDAF 0: no outputs PDAF 1: basic output (default) PDAF 2: 1 plus timing output PDAF 3: 2 plus debug output PDAF --- Internal parameter (defined inside PDAF) --- PDAF Nm1vsN: Normalization of covariance matrix; default: 1 PDAF 0: normalization with 1/(Ensemble size) PDAF (original SEIK, mainly for compatibility with older studies) PDAF 1: normalization with 1/(Ensemble size - 1) PDAF (sample covariance matrix consistent with other EnKFs) PDAF +++++++++ End of option overview for the LSEIK filter ++++++++++ }}} === ENSRF/EAKF (filtertype=PDAF_DA_ENSRF=13) === {{{ PDAF Available options for ENSRF: PDAF --- Sub-types (Parameter subtype) --- PDAF 0: ENSRF with serial observation processing (cf. Houtekamer/Mitchell, 2002) PDAF 1: EAKF/2-step local least squares filter (cf. Anderson, 2003) PDAF --- Integer parameters (Array param_int) --- PDAF param_int(1): dim_p PDAF Dimension of state vector (>0), required PDAF param_int(2): dim_ens PDAF Ensemble size (>0), required PDAF param_int(3): not used PDAF param_int(4): not used PDAF param_int(5): not used PDAF param_int(6): not used PDAF param_int(7): not used PDAF param_int(8): observe_ens PDAF Application of observation operator H, optional PDAF 0: Apply H to ensemble mean to compute innovation PDAF 1: Apply H to ensemble states; then compute innovation from their mean (default) PDAF param_int(8)=1 is the recomended choice for nonlinear H PDAF param_int(9): type_obs_init PDAF Initialize observations before or after call to prepoststep_pdaf PDAF 0: Initialize observations before call to prepoststep_pdaf PDAF 1: Initialize observations after call to prepoststep_pdaf (default) PDAF --- Floating point parameters (Array param_real) --- PDAF param_real(1): forget PDAF Forgetting factor (usually >0 and <=1), required PDAF --- Further parameters --- PDAF n_modeltasks: Number of parallel model integration tasks PDAF (>=1; not larger than total number of processors) PDAF screen: Control verbosity of PDAF PDAF 0: no outputs PDAF 1: basic output (default) PDAF 2: 1 plus timing output PDAF 3: 2 plus debug output PDAF +++++++++ End of option overview for the ENSRF ++++++++++ }}} == Option outputs using subtype=-1 - Global Ensemble Kalman filters == === ESTKF (filtertype=PDAF_DA_ESTKF=6) === {{{ PDAF Available options for ESTKF: PDAF --- Sub-types (Parameter subtype) --- PDAF 0: Standard implementation with ensemble integration PDAF 10: Fixed error space basis PDAF 11: Fixed state covariance matrix PDAF --- Integer parameters (Array param_int) --- PDAF param_int(1): dim_p PDAF Dimension of state vector (>0), required PDAF param_int(2): dim_ens PDAF Ensemble size (>0), required PDAF param_int(3): dim_lag PDAF Size of smoothing lag (>=0), optional PDAF 0: no smoothing (default) PDAF >0: apply smoother up to specified lag PDAF param_int(4): not used PDAF param_int(5) type_forget PDAF Type of forgetting factor; optional PDAF 0: fixed forgetting factor (default) PDAF 1: adaptive forgetting factor (experimental) PDAF param_int(6) type_trans PDAF Type of ensemble transformation matrix; optional PDAF 0: deterministic Omega (default) PDAF 1: random orthonormal Omega orthogonal to (1,...,1)^T PDAF 2: use product of 0 with random orthonomal matrix with eigenvector (1,...,1)^T PDAF (experimental; for random transformations, 0 or 1 are recommended) PDAF param_int(7) type_sqrt PDAF Type of transformation matrix square root; optional PDAF 0: symmetric square root (default) PDAF 1: Cholesky decomposition PDAF param_int(8): observe_ens PDAF Application of observation operator H, optional PDAF 0: Apply H to ensemble mean to compute innovation PDAF 1: Apply H to ensemble states; then compute innovation from their mean (default) PDAF param_int(8)=1 is the recomended choice for nonlinear H PDAF param_int(9): type_obs_init PDAF Initialize observations before or after call to prepoststep_pdaf PDAF 0: Initialize observations before call to prepoststep_pdaf PDAF 1: Initialize observations after call to prepoststep_pdaf (default) PDAF --- Floating point parameters (Array param_real) --- PDAF param_real(1): forget PDAF Forgetting factor (usually >0 and <=1), required PDAF --- Further parameters --- PDAF n_modeltasks: Number of parallel model integration tasks PDAF >=1 for subtype 0; not larger than total number of processors PDAF =1 required for subtypes 10 and 11 PDAF screen: Control verbosity of PDAF PDAF 0: no outputs PDAF 1: basic output (default) PDAF 2: 1 plus timing output PDAF 3: 2 plus debug output PDAF +++++++++ End of option overview for the ESTKF ++++++++++ }}} === ETKF (filtertype=PDAF_DA_ETKF=4) === {{{ PDAF Available options for ETKF: PDAF --- Sub-types (Parameter subtype) --- PDAF 0: full ensemble integration; apply T-matrix analogously to SEIK PDAF 1: full ensemble integration; formulation cf. Hunt et al. (2007) without T matrix PDAF 10: Fixed error space basis; analysis with T-matrix PDAF 11: Fixed state covariance matrix; analysis with T-matrix PDAF --- Integer parameters (Array param_int) --- PDAF param_int(1): dim_p PDAF Dimension of state vector (>0), required PDAF param_int(2): dim_ens PDAF Ensemble size (>0), required PDAF param_int(3): dim_lag PDAF Size of smoothing lag (>=0), optional PDAF 0: no smoothing (default) PDAF >0: apply smoother up to specified lag PDAF param_int(4): not used PDAF param_int(5) type_forget PDAF Type of forgetting factor; optional PDAF 0: fixed forgetting factor (default) PDAF 1: adaptive forgetting factor (experimental) PDAF param_int(6) type_trans PDAF Type of ensemble transformation matrix; optional PDAF 0: deterministic transformation (default) PDAF 2: use product of 0 with random orthonomal matrix with eigenvector (1,...,1)^T PDAF param_int(7): not used PDAF param_int(8): observe_ens PDAF Application of observation operator H, optional PDAF 0: Apply H to ensemble mean to compute innovation PDAF 1: Apply H to ensemble states; then compute innovation from their mean (default) PDAF param_int(8)=1 is the recomended choice for nonlinear H PDAF param_int(9): type_obs_init PDAF Initialize observations before or after call to prepoststep_pdaf PDAF 0: Initialize observations before call to prepoststep_pdaf PDAF 1: Initialize observations after call to prepoststep_pdaf (default) PDAF --- Floating point parameters (Array param_real) --- PDAF param_real(1): forget PDAF Forgetting factor (usually >0 and <=1), required PDAF --- Further parameters --- PDAF n_modeltasks: Number of parallel model integration tasks PDAF >=1 for subtypes 0 and 1; not larger than total number of processors PDAF =1 required for subtypes 10 and 11 PDAF screen: Control verbosity of PDAF PDAF 0: no outputs PDAF 1: basic output (default) PDAF 2: 1 plus timing output PDAF 3: 2 plus debug output PDAF +++++++++ End of option overview for the ETKF ++++++++++ }}} === EnKF (filtertype=PDAF_DA_ENKF=2) === {{{ PDAF Available options for EnKF: PDAF --- Sub-types (Parameter subtype) --- PDAF 0: Full ensemble integration; analysis for 2*dim_obs>dim_ens PDAF 1: Full ensemble integration; analysis for 2*dim_obs<=dim_ens PDAF --- Integer parameters (Array param_int) --- PDAF param_int(1): dim_p PDAF Dimension of state vector (>0), required PDAF param_int(2): dim_ens PDAF Ensemble size (>0), required PDAF param_int(3): dim_lag PDAF Size of smoothing lag (>=0), optional PDAF 0: no smoothing (default) PDAF >0: apply smoother up to specified lag PDAF param_int(4): rank_ana_enkf PDAF maximum rank for inversion of HPH^T, optional, default=0 PDAF for =0, HPH is inverted by solving the representer equation PDAF allowed range is 0 to ensemble size - 1 PDAF param_int(5): not used PDAF param_int(6): not used PDAF param_int(7): not used PDAF param_int(8): observe_ens PDAF Application of observation operator H, optional PDAF 0: Apply H to ensemble mean to compute innovation PDAF 1: Apply H to ensemble states; then compute innovation from their mean (default) PDAF param_int(8)=1 is the recomended choice for nonlinear H PDAF param_int(9): type_obs_init PDAF Initialize observations before or after call to prepoststep_pdaf PDAF 0: Initialize observations before call to prepoststep_pdaf PDAF 1: Initialize observations after call to prepoststep_pdaf (default) PDAF --- Floating point parameters (Array param_real) --- PDAF param_real(1): forget PDAF Forgetting factor (usually >0 and <=1), required PDAF --- Further parameters --- PDAF n_modeltasks: Number of parallel model integration tasks PDAF (>=1; not larger than total number of processors) PDAF screen: Control verbosity of PDAF PDAF 0: no outputs PDAF 1: basic output (default) PDAF 2: 1 plus timing output PDAF 3: 2 plus debug output PDAF +++++++++ End of option overview for the EnKF ++++++++++ }}} === SEIK (filtertype=PDAF_DA_SEIK=1) === {{{ PDAF Available options for SEIK: PDAF --- Sub-types (Parameter subtype) --- PDAF 0: full ensemble integration; left-sided application of T PDAF 1: full ensemble integration; right-sided application of T PDAF 2: full ensemble integration; explicit ensemble transformation PDAF 10: Fixed error space basis PDAF 11: Fixed state covariance matrix PDAF --- Integer parameters (Array param_int) --- PDAF param_int(1): dim_p PDAF Dimension of state vector (>0), required PDAF param_int(2): dim_ens PDAF Ensemble size (>0), required PDAF param_int(3): not used PDAF param_int(4): not used PDAF param_int(5) type_forget PDAF Type of forgetting factor; optional PDAF 0: fixed forgetting factor (default) PDAF 1: adaptive forgetting factor (experimental) PDAF param_int(6) type_trans PDAF Type of ensemble transformation matrix; optional PDAF 0: deterministic Omega (default) PDAF 1: random orthonormal Omega orthogonal to (1,...,1)^T PDAF 2: use product of 0 with random orthonomal matrix with eigenvector (1,...,1)^T PDAF (experimental; for random transformations, 1 is recommended) PDAF param_int(7) type_sqrt PDAF Type of transformation matrix square root; optional PDAF (Only relevant for subtype/=11) PDAF 0: symmetric square root (default) PDAF 1: Cholesky decomposition PDAF param_int(8): observe_ens PDAF Application of observation operator H, optional PDAF 0: Apply H to ensemble mean to compute innovation PDAF 1: Apply H to ensemble states; then compute innovation from their mean (default) PDAF param_int(8)=1 is the recomended choice for nonlinear H PDAF param_int(9): type_obs_init PDAF Initialize observations before or after call to prepoststep_pdaf PDAF 0: Initialize observations before call to prepoststep_pdaf PDAF 1: Initialize observations after call to prepoststep_pdaf (default) PDAF --- Floating point parameters (Array param_real) --- PDAF param_real(1): forget PDAF Forgetting factor (usually >0 and <=1), required PDAF --- Further parameters --- PDAF n_modeltasks: Number of parallel model integration tasks PDAF >=1 for subtypes 0, 1 and 2; not larger than total number of processors PDAF =1 required for subtypes 10 and 11 PDAF screen: Control verbosity of PDAF PDAF 0: no outputs PDAF 1: basic output (default) PDAF 2: 1 plus timing output PDAF 3: 2 plus debug output PDAF --- Internal parameter (defined inside PDAF) --- PDAF Nm1vsN: Normalization of covariance matrix; default: 1 PDAF 0: normalization with 1/(Ensemble size) PDAF (original SEIK, mainly for compatibility with older studies) PDAF 1: normalization with 1/(Ensemble size - 1) PDAF (sample covariance matrix consistent with other EnKFs) PDAF +++++++++ End of option overview for the SEIK filter ++++++++++ }}} == Option outputs using subtype=-1 - Nonlinear DA Methods == === NETF (filtertype=PDAF_DA_NETF=9) === {{{ PDAF Available options for NETF: PDAF --- Sub-types (Parameter subtype) --- PDAF 0: Standard implementation with ensemble integration PDAF --- Integer parameters (Array param_int) --- PDAF param_int(1): dim_p PDAF Dimension of state vector (>0), required PDAF param_int(2): dim_ens PDAF Ensemble size (>0), required PDAF param_int(3): dim_lag PDAF Size of smoothing lag (>=0), optional PDAF 0: no smoothing (default) PDAF >0: apply smoother up to specified lag PDAF param_int(4): type_noise PDAF Type of ensemble perturbations, optional PDAF 0: no perturbations (default) PDAF 1: constant standard deviation PDAF 2: relative to ensemble standard deviation PDAF param_int(5) type_forget PDAF Type of forgetting factor; optional PDAF 0: forgetting factor on forecast ensemble (default) PDAF 2: forgetting factor on analysis ensemble PDAF param_int(6) type_trans PDAF Type of ensemble transformation matrix; optional PDAF 0: random orthonormal matrix orthogonal to (1,...,1)^T (default) PDAF 1: deterministic transformation PDAF param_int(7): type_winf PDAF Type of weights inflation; optional PDAF 0: no weights inflation (default) PDAF 1: inflate so that N_eff/N > param_real(2) PDAF param_int(8): observe_ens PDAF Application of observation operator H, optional PDAF Note: This parameter has no influence on the NETF assimilation result PDAF 0: Apply H to ensemble mean to compute innovation PDAF 1: Apply H to ensemble states; then compute innovation from their mean (default) PDAF param_int(8)=1 is the recomended choice for nonlinear H PDAF param_int(9): type_obs_init PDAF Initialize observations before or after call to prepoststep_pdaf PDAF 0: Initialize observations before call to prepoststep_pdaf PDAF 1: Initialize observations after call to prepoststep_pdaf (default) PDAF --- Floating point parameters (Array param_real) --- PDAF param_real(1): forget PDAF Forgetting factor (usually >0 and <=1), required PDAF param_real(2): limit_winf PDAF Limit for weigts inflation N_eff/N > param_real(2), optional, default=0.0 PDAF param_real(3): noise_amp PDAF Ensemble perturbation level (>0), required, only used if param_int(4)>0 PDAF --- Further parameters --- PDAF n_modeltasks: Number of parallel model integration tasks PDAF >=1; not larger than total number of processors PDAF screen: Control verbosity of PDAF PDAF 0: no outputs PDAF 1: basic output (default) PDAF 2: 1 plus timing output PDAF 3: 2 plus debug output PDAF +++++++++ End of option overview for the NETF ++++++++++ }}} === LNETF (filtertype=PDAF_DA_LNETF=10) === {{{ PDAF Available options for LNETF: PDAF --- Sub-types (Parameter subtype) --- PDAF 0: Standard implementation with ensemble integration PDAF --- Integer parameters (Array param_int) --- PDAF param_int(1): dim_p PDAF Dimension of state vector (>0), required PDAF param_int(2): dim_ens PDAF Ensemble size (>0), required PDAF param_int(3): dim_lag PDAF Size of smoothing lag (>=0), optional PDAF 0: no smoothing (default) PDAF >0: apply smoother up to specified lag PDAF param_int(4): type_noise PDAF Type of ensemble perturbations, optional PDAF 0: no perturbations (default) PDAF 1: constant standard deviation PDAF 2: relative to ensemble standard deviation PDAF param_int(5) type_forget PDAF Type of forgetting factor; optional PDAF 0: forgetting factor on forecast ensemble (default) PDAF 1: forgetting factor on forecast ensemble only observed domains PDAF 2: forgetting factor on analysis ensemble PDAF 3: forgetting factor on analysis ensemble only observed domains PDAF param_int(6) type_trans PDAF Type of ensemble transformation matrix; optional PDAF 0: random orthonormal matrix orthogonal to (1,...,1)^T (default) PDAF 1: deterministic transformation PDAF param_int(7): type_winf PDAF Type of weights inflation; optional PDAF 0: no weights inflation (default) PDAF 1: inflate so that N_eff/N > param_real(2) PDAF param_int(8): observe_ens PDAF Application of observation operator H, optional PDAF Note: This parameter has no influence on the LNETF assimilation result PDAF 0: Apply H to ensemble mean to compute innovation PDAF 1: Apply H to ensemble states; then compute innovation from their mean (default) PDAF param_int(8)=1 is the recomended choice for nonlinear H PDAF param_int(9): type_obs_init PDAF Initialize observations before or after call to prepoststep_pdaf PDAF 0: Initialize observations before call to prepoststep_pdaf PDAF 1: Initialize observations after call to prepoststep_pdaf (default) PDAF --- Floating point parameters (Array param_real) --- PDAF param_real(1): forget PDAF Forgetting factor (usually >0 and <=1), required PDAF param_real(2): limit_winf PDAF Limit for weigts inflation N_eff/N > param_real(2), optional, default=0.0 PDAF param_real(3): noise_amp PDAF Ensemble perturbation level (>0), required, only used if param_int(4)>0 PDAF --- Further parameters --- PDAF n_modeltasks: Number of parallel model integration tasks PDAF >=1; not larger than total number of processors PDAF screen: Control verbosity of PDAF PDAF 0: no outputs PDAF 1: basic output (default) PDAF 2: 1 plus timing output PDAF 3: 2 plus debug output PDAF +++++++++ End of option overview for the LNETF ++++++++++ }}} === LKNETF (filtertype=PDAF_DA_LKNETF=11) === {{{ PDAF Available options for LKNETF: PDAF --- Sub-types (Parameter subtype) --- PDAF 0: HNK: 2-step LKNETF with NETF before LETKF PDAF 1: HKN: 2-step LKNETF with LETKF before NETF PDAF 2: HSync: LKNETF synchronous PDAF --- Integer parameters (Array param_int) --- PDAF param_int(1): dim_p PDAF Dimension of state vector (>0), required PDAF param_int(2): dim_ens PDAF Ensemble size (>0), required PDAF param_int(3): not used PDAF param_int(4): type_hyb PDAF Type of hybrid weight; optional PDAF 0: fixed value PDAF 1: gamma_lin: (1 - N_eff/N_e)*param_real(2) (default) PDAF 2: gamma_alpha: hybrid weight from N_eff/N>=param_real(2) PDAF 3: gamma_ska: 1 - min(s,k)/sqrt(param_real(3)) with N_eff/N>=param_real(2) PDAF 4: gamma_sklin: 1 - min(s,k)/sqrt(param_real(3)) >= 1-N_eff/N>=param_real(2) PDAF param_int(5): type_forget PDAF Type of forgetting factor; optional PDAF 0: inflate forecast ensemble by 1/forget (default) PDAF 1: inflate forecast ensemble by 1/forget only observed domains PDAF 2: inflate analysis ensemble by 1/forget PDAF 3: inflate analysis ensemble by 1/forget only observed domains PDAF 5: adaptive forgetting factor for full domain in LETKF part PDAF 6: locally adaptive forgetting factor in LETKF part PDAF param_int(6): type_trans PDAF Type of ensemble transformation matrix; optional PDAF 0: random orthonormal matrix orthogonal to (1,...,1)^T (default) PDAF 1: deterministic transformation PDAF param_int(7): not used PDAF --- Floating point parameters (Array param_real) --- PDAF param_real(1): forget PDAF Forgetting factor (usually >0 and <=1), required PDAF param_real(2): hyb_g PDAF prescribed hybrid weight gamma (usually >0 and <=1), optional, default=0.95 PDAF param_real(3): hyb_k PDAF hybrid norm kappa (>0), optional, default=dim_ens PDAF --- Further parameters --- PDAF n_modeltasks: Number of parallel model integration tasks PDAF >=1; not larger than total number of processors PDAF screen: Control verbosity of PDAF PDAF 0: no outputs PDAF 1: basic output (default) PDAF 2: 1 plus timing output PDAF 3: 2 plus debug output PDAF +++++++++ End of option overview for the LKNETF ++++++++++ }}} === PF (filtertype=PDAF_DA_PF=12) === {{{ PDAF Available options for PF: PDAF --- Sub-types (Parameter subtype) --- PDAF 0: Standard implementation with ensemble integration PDAF --- Integer parameters (Array param_int) --- PDAF param_int(1): dim_p PDAF Dimension of state vector (>0), required PDAF param_int(2): dim_ens PDAF Ensemble size (>0), required PDAF param_int(3): type_resample PDAF Resampling type, optional PDAF 1: probabilistic resamping (default) PDAF 2: stochastic universal resampling PDAF 3: residual resampling PDAF param_int(4): type_noise PDAF Type of ensemble perturbations, optional PDAF 0: no perturbations (default) PDAF 1: constant standard deviation PDAF 2: relative to ensemble standard deviation PDAF param_int(5) type_forget PDAF Type of forgetting factor; optional PDAF 0: forgetting factor on forecast ensemble (default) PDAF 2: forgetting factor on analysis ensemble PDAF param_int(6): not used PDAF param_int(7): type_winf PDAF Type of weights inflation; optional PDAF 0: no weights inflation (default) PDAF 1: inflate so that N_eff/N > param_real(2) PDAF param_int(8): observe_ens PDAF Application of observation operator H, optional PDAF Note: This parameter has no influence on the PF assimilation result PDAF 0: Apply H to ensemble mean to compute innovation PDAF 1: Apply H to ensemble states; then compute innovation from their mean (default) PDAF param_int(8)=1 is the recomended choice for nonlinear H PDAF param_int(9): type_obs_init PDAF Initialize observations before or after call to prepoststep_pdaf PDAF 0: Initialize observations before call to prepoststep_pdaf PDAF 1: Initialize observations after call to prepoststep_pdaf (default) PDAF --- Floating point parameters (Array param_real) --- PDAF param_real(1): forget PDAF Forgetting factor (usually >0 and <=1), required PDAF param_real(2): limit_winf PDAF Limit for weigts inflation N_eff/N > param_real(2), optional, default=0.0 PDAF param_real(3): noise_amp PDAF Ensemble perturbation level (>0), required, only used if param_int(4)>0 PDAF --- Further parameters --- PDAF n_modeltasks: Number of parallel model integration tasks PDAF >=1; not larger than total number of processors PDAF screen: Control verbosity of PDAF PDAF 0: no outputs PDAF 1: basic output (default) PDAF 2: 1 plus timing output PDAF 3: 2 plus debug output PDAF +++++++++ End of option overview for the PF ++++++++++ }}} == Option outputs using subtype=-1 - Observation Geneeration == === GENOBS (filtertype=PDAF_DA_GENOBS=100) === {{{ PDAF Available options for GENOBS: PDAF --- Sub-types (Parameter subtype) --- PDAF 0: Standard implementation with ensemble integration PDAF --- Integer parameters (Array param_int) --- PDAF param_int(1): dim_p PDAF Dimension of state vector (>0), required PDAF param_int(2): dim_ens PDAF Ensemble size (>0), required PDAF param_int(3): seedset PDAF seed set index for random number generator, optional PDAF valid are values between 1 and 20; default=1 PDAF --- Floating point parameters (Array param_real) --- PDAF param_real(1): Forgetting factor (usually >0 and <=1), required, but not used PDAF --- Further parameters --- PDAF n_modeltasks: Number of parallel model integration tasks PDAF =1 for GENOBS; not larger than total number of processors PDAF =1 required for subtypes 10 and 11 PDAF screen: Control verbosity of PDAF PDAF 0: no outputs PDAF 1: basic output (default) PDAF 2: 1 plus timing output PDAF 3: 2 plus debug output PDAF +++++++++ End of option overview for GENOBS ++++++++++ }}} == Option outputs using subtype=-1 - 3D-Var Methods == === 3D-Var (filtertype=PDAF_DA_3DVAR=200) === {{{ PDAF Available options for 3D-Var: PDAF --- Sub-types (Parameter subtype) --- PDAF 0: incremental 3D-Var with parameterized covariance matrix PDAF 1: 3D ensemble Var using LESTKF for ensemble transformation PDAF 2: 3D ensemble Var using ESTKF for ensemble transformation PDAF 3: hybrid 3D-Var using LESTKF for ensemble transformation PDAF 4: hybrid 3D-Var using ESTKF for ensemble transformation PDAF --- Integer parameters (Array param_int) --- PDAF param_int(1): dim_p PDAF Dimension of state vector (>0), required PDAF param_int(2): dim_ens PDAF Ensemble size (>0), required PDAF param_int(3): type_opt PDAF Select optimization method (solver), required PDAF 1: LBFGS (default) PDAF 2: CG+ PDAF 3: direct implementation of CG PDAF 12: CG+ with decomposed control vector PDAF 13: direct implementation of CG with decomposed control vector PDAF param_int(4): size of parameterized control vector (for 3D-Var and hybrid 3D-Var), required PDAF param_int(5): size of ensemble control vector (required for ensemble and hybrid 3D-Var), PDAF param_int(6): Solver-specific parameter, optional PDAF LBFGS: parameter m (default=5) PDAF Number of corrections used in limited memory matrix; 3<=m<=20 PDAF CG+: parameter method (default=2) PDAF (1) Fletcher-Reeves, (2) Polak-Ribiere, (3) positive Polak-Ribiere PDAF CG: maximum number of iterations (default=200) PDAF param_int(7): Solver-specific parameter, optional PDAF CG+: parameter irest (default=1) PDAF (0) no restarts; (n>0) restart every n steps PDAF param_int(8): observe_ens PDAF Application of observation operator H, optional PDAF 0: Apply H to ensemble mean to compute innovation PDAF 1: Apply H to ensemble states; then compute innovation from their mean (default) PDAF param_int(8)=1 is the recomended choice for nonlinear H PDAF param_int(9): type_obs_init PDAF Initialize observations before or after call to prepoststep_pdaf PDAF 0: Initialize observations before call to prepoststep_pdaf PDAF 1: Initialize observations after call to prepoststep_pdaf (default) PDAF param_int(10): not used PDAF ___Options for ESTKF/LESTKF for En3DVar/hyb3DVar___ PDAF param_int(11) type_forget PDAF Type of forgetting factor; optional PDAF 0: fixed forgetting factor (default) PDAF 1: adaptive forgetting factor (experimental) PDAF 2: locally adaptive forgetting factor (experimental) PDAF param_int(12) type_trans PDAF Type of ensemble transformation matrix; optional PDAF 0: deterministic Omega (default) PDAF 1: random orthonormal Omega orthogonal to (1,...,1)^T PDAF 2: use product of 0 with random orthonomal matrix with eigenvector (1,...,1)^T PDAF (experimental; for random transformations, 0 or 1 are recommended) PDAF param_int(13) type_sqrt PDAF Type of transformation matrix square root; optional PDAF 0: symmetric square root (default) PDAF 1: Cholesky decomposition PDAF --- Floating point parameters (Array param_real) --- PDAF param_real(1): Forgetting factor (usually >0 and <=1), required; PDAF (only used for ensemble and hybrid 3D-Var) PDAF param_real(2): hybrid weight beta, optional (only for hybrid 3D-Var) PDAF range >=0.0 and <=1.0, =1.0 for pure ensemble 3D-var (default=0.5) PDAF param_real(3): Solver-specific parameter, optional PDAF LBFGS: Limit for stopping iterations (pgtol, default=1.0e-5) PDAF CG+: convergence parameter eps (default=1.0e-5) PDAF CG: convergence parameter eps (default=1.0e-6) PDAF param_real(4): Solver-specific parameter, optional PDAF LBFGS: Tolerance in termination test (factr, default=1.0e+7) PDAF --- Further parameters --- PDAF n_modeltasks: Number of parallel model integration tasks PDAF >=1 for subtypes >0; not larger than total number of processors PDAF =1 required for subtype 0 PDAF screen: Control verbosity of PDAF PDAF 0: no outputs PDAF 1: basic output (default) PDAF 2: 1 plus timing output PDAF 3: 2 plus debug output PDAF +++++++++ End of option overview for 3DVAR ++++++++++ }}}