The following table describes all elements contained within
input_list$pars, which are generated using the
SPoRC::Setup_x functions. Note that when
n_sexes > 1, the first dimension will always be females
and the second dimension will always be males.
label_dims(input_list$par, input_list$data) returns a
copy of the parameter list with these dimensions named and labeled with
calendar years and ages; see the model report article for the
details.
Fishery Removals
| Parameter | Dimension | Description | Default |
|---|---|---|---|
ln_sigmaC |
n_regions × n_years × n_seas × n_fish_fleets |
Log-scale standard deviation for aggregated (regional) catch likelihood | 0.01 |
ln_sigmaC_pop |
n_pop × n_regions × n_years × n_seas × n_fish_fleets |
Log-scale standard deviation for population-specific catch
likelihood. Only used when
UseCatch_pop[p,r,y,seas,f] == 1. |
0.01 |
ln_sigmaF |
n_regions × n_seas × n_fish_fleets |
Log-scale standard deviation for fishing mortality process error
(used under all three Fdev_model structures:
"iid", "rw", "ar1") |
0 |
Fdev_rho |
n_regions × n_seas × n_fish_fleets |
Unconstrained AR1 correlation parameter for fishing mortality
deviations, transformed to (-1, 1) via
2 / (1 + exp(-2x)) - 1. Only used when
Fdev_model = "ar1"; mapped to NA (unused)
otherwise regardless of Fdev_rho_spec
|
0 |
ln_F_mean |
n_regions × n_seas × n_fish_fleets |
Log-scale mean fishing mortality rate | 0.1 |
init_F_par |
n_regions × n_seas × n_fish_fleets |
Initial fishing mortality applied to the equilibrium age structure.
Under init_F_form = "prop" (default) it is a logit-scale
proportion of exp(ln_F_mean); under "abs" it
is the log of the initial F itself. Estimated only under
init_F_spec = "est"
|
Effectively zero: qlogis(1e-10) under
"prop", log(1e-100) under
"abs"
|
ln_F_devs |
n_regions × n_years × n_seas × n_fish_fleets |
Log-scale annual fishing mortality deviations (can be specified as
random effects). Process error structure set via
Fdev_model: "iid" (default),
"rw", or "ar1"; catch-active years need not be
contiguous for "rw"/"ar1" (see
vignette("c_model_equations")) |
0 |
ln_sigmaD |
n_regions × n_years × n_seas × n_fish_fleets |
Log-scale standard deviation for aggregated (regional) discard likelihood | 0.01 |
ln_sigmaD_pop |
n_pop × n_regions × n_years × n_seas × n_fish_fleets |
Log-scale standard deviation for population-specific discard likelihood | 0.01 |
ln_sigma_dmr |
n_regions × n_seas × n_fish_fleets |
Log-scale standard deviation for discard mortality rate process error | 0 |
logit_dmr_mean |
n_regions × n_seas × n_fish_fleets |
Logit-scale mean discard mortality rate | 0 |
logit_dmr_devs |
n_regions × n_years × n_seas × n_fish_fleets |
Logit-scale annual discard mortality rate deviations (can be
specified as random effects). Fixed at zero by default;
dmr_dev_spec = "est_all" estimates a deviation in every
fished cell, whether or not discard data are present there (see
vignette("t_model_options")) |
0 |
Age-Disaggregated Observations
Estimated only when a fleet fits catch, discards or an index at age.
Each is keyed by age and fleet through an integer matrix in which equal
entries share a parameter and NA excludes one, so one
structure gives a standard deviation per age, per age group, or per
fleet. See vignette("t_model_options"). The age dimension
is the observed ages, n_obs_ages, the columns of the
fleet’s ageing error matrix, which equals n_ages under the
default identity matrix.
| Parameter | Dimension | Description | Default |
|---|---|---|---|
ln_sigmaCAA |
n_obs_ages × n_sexes × n_fish_fleets |
Log-scale standard deviation for the catch-at-age likelihood. Keyed
by sigmaCAA_key
|
log(0.5) |
ln_sigmaCAA_pop |
n_pop × n_obs_ages × n_sexes × n_fish_fleets |
Population-specific counterpart, keyed by
sigmaCAA_pop_key
|
log(0.5) |
ln_sigmaDAA |
n_obs_ages × n_sexes × n_fish_fleets |
Log-scale standard deviation for the discard-at-age likelihood,
keyed by sigmaDAA_key
|
log(0.5) |
ln_sigmaDAA_pop |
n_pop × n_obs_ages × n_sexes × n_fish_fleets |
Population-specific counterpart, keyed by
sigmaDAA_pop_key
|
log(0.5) |
ln_sigmaSrvIdxAA |
n_obs_ages × n_sexes × n_srv_fleets |
Log-scale standard deviation for the survey index at age, keyed by
sigmaSrvIdxAA_key
|
log(0.5) |
ln_sigmaSrvIdxAA_pop |
n_pop × n_obs_ages × n_sexes × n_srv_fleets |
Population-specific counterpart | log(0.5) |
trans_rho_catch |
n_regions × n_sexes × n_fish_fleets |
Unconstrained AR(1) correlation across ages for catch at age.
Transformed to (-1, 1) by rho_trans(). Shared
according to rho_catch_spec, one per fleet by default, and
mapped to NA unless AgeObsCorr_catch is
"1dar1" or "2dar1"
|
0 |
trans_rho_catch_year |
n_regions × n_sexes × n_fish_fleets |
Correlation across years, read only under
AgeObsCorr_catch = "2dar1"
|
0 |
trans_rho_catch_us |
n_pairs × n_regions × n_sexes × n_fish_fleets |
Unconstrained entries of the unstructured across-age correlation,
n_pairs = n_obs_ages(n_obs_ages - 1)/2. Read only under
AgeObsCorr_catch = "us"
|
0 |
trans_rho_discard, trans_rho_discard_year,
trans_rho_discard_us
|
as above | As above, for discards at age | 0 |
trans_rho_srv_idx, trans_rho_srv_idx_year,
trans_rho_srv_idx_us
|
n_regions × n_sexes × n_srv_fleets (and
n_pairs × …) |
As above, for the survey index at age | 0 |
trans_rho_catch_pop,
trans_rho_discard_pop,
trans_rho_srv_idx_pop
|
as above with a leading n_pop
|
Population-specific counterparts, with their own _year
and _us forms. The _pop data sources have
their own correlation rather than borrowing the aggregated one |
0 |
Fishery Selectivity
| Parameter | Dimension | Description | Default |
|---|---|---|---|
fish_fixed_sel_pars |
n_regions × max_fish_pars × max_fishsel_blks × n_sexes × n_fish_fleets |
Log-scale selectivity curve parameters | 0 |
fishsel_pe_pars |
n_regions × max(4, n_sel_pars) × n_sexes × n_fish_fleets |
Process error parameters for time-varying selectivity: sigma
(iid/rw), or partial correlations + log-variance (semi-parametric
forms), see [Fishery and Survey Selectivity] in
vignette("c_model_equations"). Second dimension is 4 unless
the selectivity form itself needs more parameters (e.g. double-normal,
non-parametric) |
0 |
ln_fishsel_devs |
n_regions × (n_years + n_proj_yrs_devs) × n_bins × n_sexes × n_fish_fleets |
Log-scale selectivity deviations w/ projection years (n_bins = n_ages or n_lens depending on selectivity type) (can be specified as random effects) | 0 |
ln_fishsel_bin_devs |
n_regions × (n_years + n_proj_yrs_devs) × n_bins × n_sexes × n_fish_fleets |
Log-scale bin-override selectivity deviations. Bins named in
fish_sel_bin_dev_bins take exp(deviation) as
their selectivity outright, replacing the functional form for that bin;
all other cells are fixed at zero and unused (can be specified as random
effects) |
0 |
fishsel_bin_devs_pe_pars |
n_regions × n_bins × n_sexes × n_fish_fleets |
Log-scale process error standard deviations for the bin-override
deviations. Only estimated for overridden bins whose fleet requests iid
or random walk structure via cont_tv_fishsel_bin_devs
|
0 |
ln_fishsel_sex_scale |
n_regions × n_blocks × n_sexes × n_fish_fleets |
Log-scale constant multiplying the realized selectivity curve of
every sex beyond the first (fish_sel_sex_offset = "scale"
or "par_scale"). The first sex is the reference and its
slots are never estimated; only the blocks a fleet has are |
0 |
ret_fixed_sel_pars |
n_regions × max_ret_pars × max_retsel_blks × n_sexes × n_fish_fleets |
Retention selectivity curve parameters | 0 |
retsel_pe_pars |
n_regions × max(4, n_sel_pars) × n_sexes × n_fish_fleets |
Process error parameters for time-varying retention selectivity (see
fishsel_pe_pars above) |
0 |
ln_retsel_devs |
n_regions × (n_years + n_proj_yrs_devs) × n_bins × n_sexes × n_fish_fleets |
Log-scale retention selectivity deviations w/ projection years (can be specified as random effects) | 0 |
ln_retsel_bin_devs |
n_regions × (n_years + n_proj_yrs_devs) × n_bins × n_sexes × n_fish_fleets |
Log-scale bin-override retention selectivity deviations (see
ln_fishsel_bin_devs above) |
0 |
ln_retsel_sex_scale |
n_regions × n_blocks × n_sexes × n_fish_fleets |
Log-scale curve multiplier for sexes beyond the first on retention
selectivity (see ln_fishsel_sex_scale above;
ret_sel_sex_offset) |
0 |
retsel_bin_devs_pe_pars |
n_regions × n_bins × n_sexes × n_fish_fleets |
Log-scale process error standard deviations for the retention bin-override deviations | 0 |
Note that fish_fixed_sel_pars holds parameters on the
scale of the chosen functional form: log scale for most forms, logit
scale for the "nonpar" form and for asymptote parameters,
and log scale per bin for the "nonparlog" form (where each
year’s curve is standardized to average one over the bins named in
fish_sel_norm_bins, every bin by default).
Fishery Catchability
| Parameter | Dimension | Description | Default |
|---|---|---|---|
ln_fish_q |
n_regions × max_fishq_blks × n_fish_fleets |
Log-scale fishery catchability coefficient | 0 |
ln_fish_q_devs |
n_regions × n_years × n_fish_fleets |
Log-scale annual deviations about the mean fishery catchability,
under fish_q_model
|
0 |
ln_sigma_fish_q |
n_regions × n_fish_fleets |
Log-scale standard deviation of the fishery catchability deviations | log(0.1) |
fish_q_rho |
n_regions × n_fish_fleets |
Unconstrained ar1 correlation of the fishery catchability
deviations, read only under fish_q_model = "ar1"
|
0 |
ln_sigmaFishIdx |
n_fish_fleets |
Log-scale estimated component of the region-aggregated fishery index
standard deviation. It is added to the reported standard errors, added
in quadrature, or replaces them under
sigmaFishIdx_spec = "est_additive",
"est_quadrature" or "est_replace", and is not
used under "fix" (default) |
log(0.01) |
ln_sigmaFishIdx_pop |
n_fish_fleets |
Population-specific counterpart, under
sigmaFishIdx_pop_spec
|
log(0.01) |
Fishery Composition Likelihoods
Region-aggregated and population-specific composition likelihoods
have separate overdispersion and correlation parameters. The
_pop_ variants mirror the structure of their
region-aggregated counterparts with an added leading n_pop
dimension.
| Parameter | Dimension | Description | Default |
|---|---|---|---|
ln_FishAge_theta |
n_regions × n_sexes × n_fish_fleets |
Log-scale overdispersion parameter for region-aggregated fishery age compositions | 0 |
ln_FishAge_theta_agg |
n_fish_fleets |
Log-scale overdispersion for region- and sex-aggregated fishery age compositions | 0 |
FishAge_corr_pars |
n_regions × n_sexes × n_fish_fleets × 2 |
Correlation parameters for logistic-normal region-aggregated fishery age composition likelihood | 0.01 |
FishAge_corr_pars_agg |
n_fish_fleets |
Correlation parameters for region- and sex-aggregated logistic-normal fishery age compositions | 0.01 |
ln_FishAge_pop_theta |
n_pop × n_regions × n_sexes × n_fish_fleets |
Log-scale overdispersion parameter for population-specific fishery age compositions | 0 |
ln_FishAge_pop_theta_agg |
n_pop × n_fish_fleets |
Log-scale overdispersion for population-specific sex-aggregated fishery age compositions | 0 |
FishAge_pop_corr_pars |
n_pop × n_regions × n_sexes × n_fish_fleets × 2 |
Correlation parameters for logistic-normal population-specific fishery age composition likelihood | 0.01 |
FishAge_pop_corr_pars_agg |
n_pop × n_fish_fleets |
Correlation parameters for population-specific sex-aggregated logistic-normal fishery age compositions | 0.01 |
ln_FishLen_theta |
n_regions × n_sexes × n_fish_fleets |
Log-scale overdispersion parameter for region-aggregated fishery length compositions | 0 |
ln_FishLen_theta_agg |
n_fish_fleets |
Log-scale overdispersion for region- and sex-aggregated fishery length compositions | 0 |
FishLen_corr_pars |
n_regions × n_sexes × n_fish_fleets × 2 |
Correlation parameters for logistic-normal region-aggregated fishery length composition likelihood. The first parameter represents bin correlations, while the second represents sex correlations (if specified) | 0.01 |
FishLen_corr_pars_agg |
n_fish_fleets |
Correlation parameters for region- and sex-aggregated logistic-normal fishery length compositions. The correlation represents bin correlations | 0.01 |
ln_FishLen_pop_theta |
n_pop × n_regions × n_sexes × n_fish_fleets |
Log-scale overdispersion parameter for population-specific fishery length compositions | 0 |
ln_FishLen_pop_theta_agg |
n_pop × n_fish_fleets |
Log-scale overdispersion for population-specific sex-aggregated fishery length compositions | 0 |
FishLen_pop_corr_pars |
n_pop × n_regions × n_sexes × n_fish_fleets × 2 |
Correlation parameters for logistic-normal population-specific fishery length composition likelihood. The first parameter represents bin correlations, while the second represents sex correlations (if specified) | 0.01 |
FishLen_pop_corr_pars_agg |
n_pop × n_fish_fleets |
Correlation parameters for population-specific sex-aggregated logistic-normal fishery length compositions. The correlation represents bin correlations | 0.01 |
ln_FishAge_discard_theta |
n_regions × n_sexes × n_fish_fleets |
Log-scale overdispersion parameter for region-aggregated discard age compositions | 0 |
ln_FishAge_discard_theta_agg |
n_fish_fleets |
Log-scale overdispersion for region- and sex-aggregated discard age compositions | 0 |
FishAge_discard_corr_pars |
n_regions × n_sexes × n_fish_fleets × 2 |
Correlation parameters for logistic-normal region-aggregated discard age composition likelihood | 0.01 |
FishAge_discard_corr_pars_agg |
n_fish_fleets |
Correlation parameters for region- and sex-aggregated logistic-normal discard age compositions | 0.01 |
ln_FishAge_discard_pop_theta |
n_pop × n_regions × n_sexes × n_fish_fleets |
Log-scale overdispersion parameter for population-specific discard age compositions | 0 |
ln_FishAge_discard_pop_theta_agg |
n_pop × n_fish_fleets |
Log-scale overdispersion for population-specific sex-aggregated discard age compositions | 0 |
FishAge_discard_pop_corr_pars |
n_pop × n_regions × n_sexes × n_fish_fleets × 2 |
Correlation parameters for logistic-normal population-specific discard age composition likelihood | 0.01 |
FishAge_discard_pop_corr_pars_agg |
n_pop × n_fish_fleets |
Correlation parameters for population-specific sex-aggregated logistic-normal discard age compositions | 0.01 |
ln_FishLen_discard_theta |
n_regions × n_sexes × n_fish_fleets |
Log-scale overdispersion parameter for region-aggregated discard length compositions | 0 |
ln_FishLen_discard_theta_agg |
n_fish_fleets |
Log-scale overdispersion for region- and sex-aggregated discard length compositions | 0 |
FishLen_discard_corr_pars |
n_regions × n_sexes × n_fish_fleets × 2 |
Correlation parameters for logistic-normal region-aggregated discard length composition likelihood. The first parameter represents bin correlations, while the second represents sex correlations (if specified) | 0.01 |
FishLen_discard_corr_pars_agg |
n_fish_fleets |
Correlation parameters for region- and sex-aggregated logistic-normal discard length compositions. The correlation represents bin correlations | 0.01 |
ln_FishLen_discard_pop_theta |
n_pop × n_regions × n_sexes × n_fish_fleets |
Log-scale overdispersion parameter for population-specific discard length compositions | 0 |
ln_FishLen_discard_pop_theta_agg |
n_pop × n_fish_fleets |
Log-scale overdispersion for population-specific sex-aggregated discard length compositions | 0 |
FishLen_discard_pop_corr_pars |
n_pop × n_regions × n_sexes × n_fish_fleets × 2 |
Correlation parameters for logistic-normal population-specific discard length composition likelihood. The first parameter represents bin correlations, while the second represents sex correlations (if specified) | 0.01 |
FishLen_discard_pop_corr_pars_agg |
n_pop × n_fish_fleets |
Correlation parameters for population-specific sex-aggregated logistic-normal discard length compositions. The correlation represents bin correlations | 0.01 |
ln_Fish_caal_theta |
n_regions × n_sexes × n_fish_fleets |
Log-scale overdispersion for fishery conditional age-at-length data,
one per region and sex and shared across length bins. Estimated only
under Fish_caal_LikeType = "Dirichlet-Multinomial"
|
0 |
ln_Fish_caal_theta_agg |
n_fish_fleets |
Log-scale overdispersion for region- and sex-aggregated fishery conditional age-at-length data | 0 |
Survey Selectivity
| Parameter | Dimension | Description | Default |
|---|---|---|---|
srv_fixed_sel_pars |
n_regions × max_srv_pars × max_srv_blks × n_sexes × n_srv_fleets |
Log-scale selectivity curve parameters | 0 |
srvsel_pe_pars |
n_regions × max(4, n_sel_pars) × n_sexes × n_srv_fleets |
Process error parameters for time-varying survey selectivity (see
fishsel_pe_pars in Fishery
Selectivity) |
0 |
ln_srvsel_devs |
n_regions × (n_years + n_proj_yrs_devs) × n_bins × n_sexes × n_srv_fleets |
Log-scale selectivity deviations w/ projection years (n_bins = n_ages or n_lens depending on selectivity type) (can be specified as random effects) | 0 |
ln_srvsel_bin_devs |
n_regions × (n_years + n_proj_yrs_devs) × n_bins × n_sexes × n_srv_fleets |
Log-scale bin-override selectivity deviations (see
ln_fishsel_bin_devs in Fishery Selectivity) |
0 |
ln_srvsel_sex_scale |
n_regions × n_blocks × n_sexes × n_srv_fleets |
Log-scale curve multiplier for sexes beyond the first on survey
selectivity (see ln_fishsel_sex_scale in Fishery Selectivity;
srv_sel_sex_offset) |
0 |
srvsel_bin_devs_pe_pars |
n_regions × n_bins × n_sexes × n_srv_fleets |
Log-scale process error standard deviations for the survey bin-override deviations | 0 |
Survey Catchability
| Parameter | Dimension | Description | Default |
|---|---|---|---|
ln_srv_q |
n_regions × max_srvq_blks × n_srv_fleets |
Log-scale survey catchability coefficient. Automatically fixed (not
estimated) for fleets whose catchability is solved analytically via
srv_q_type
|
0 |
ln_srv_q_devs |
n_regions × n_years × n_srv_fleets |
Log-scale annual deviations about the mean survey catchability,
under srv_q_model
|
0 |
ln_sigma_srv_q |
n_regions × n_srv_fleets |
Log-scale standard deviation of the survey catchability deviations | log(0.1) |
srv_q_rho |
n_regions × n_srv_fleets |
Unconstrained ar1 correlation of the survey catchability deviations,
read only under srv_q_model = "ar1"
|
0 |
ln_sigmaSrvIdx |
n_srv_fleets |
Log-scale estimated component of the region-aggregated survey index
standard deviation. It is added to the reported standard errors, added
in quadrature, or replaces them under
sigmaSrvIdx_spec = "est_additive",
"est_quadrature" or "est_replace", and is not
used under "fix" (default) |
log(0.01) |
ln_sigmaSrvIdx_pop |
n_srv_fleets |
Population-specific counterpart, under
sigmaSrvIdx_pop_spec
|
log(0.01) |
Survey Composition Likelihoods
Region-aggregated and population-specific composition likelihoods
have separate overdispersion and correlation parameters. The
_pop_ variants mirror the structure of their
region-aggregated counterparts with an added leading n_pop
dimension.
| Parameter | Dimension | Description | Default |
|---|---|---|---|
ln_SrvAge_theta |
n_regions × n_sexes × n_srv_fleets |
Log-scale overdispersion parameter for region-aggregated survey age compositions | 0 |
ln_SrvAge_theta_agg |
n_srv_fleets |
Log-scale overdispersion for region- and sex-aggregated survey age compositions | 0 |
SrvAge_corr_pars |
n_regions × n_sexes × n_srv_fleets × 2 |
Correlation parameters for logistic-normal region-aggregated survey age composition likelihood | 0.01 |
SrvAge_corr_pars_agg |
n_srv_fleets |
Correlation parameters for region-aggregated logistic-normal survey age compositions | 0.01 |
ln_SrvAge_pop_theta |
n_pop × n_regions × n_sexes × n_srv_fleets |
Log-scale overdispersion parameter for population-specific survey age compositions | 0 |
ln_SrvAge_pop_theta_agg |
n_pop × n_srv_fleets |
Log-scale overdispersion for population-specific sex-aggregated survey age compositions | 0 |
SrvAge_pop_corr_pars |
n_pop × n_regions × n_sexes × n_srv_fleets × 2 |
Correlation parameters for logistic-normal population-specific survey age composition likelihood | 0.01 |
SrvAge_pop_corr_pars_agg |
n_pop × n_srv_fleets |
Correlation parameters for population-specific sex-aggregated logistic-normal survey age compositions | 0.01 |
ln_SrvLen_theta |
n_regions × n_sexes × n_srv_fleets |
Log-scale overdispersion parameter for region-aggregated survey length compositions | 0 |
ln_SrvLen_theta_agg |
n_srv_fleets |
Log-scale overdispersion for region- and sex-aggregated survey length compositions | 0 |
SrvLen_corr_pars |
n_regions × n_sexes × n_srv_fleets × 2 |
Correlation parameters for logistic-normal region-aggregated survey length composition likelihood. The first parameter represents bin correlations, while the second represents sex correlations (if specified) | 0.01 |
SrvLen_corr_pars_agg |
n_srv_fleets |
Correlation parameters for region- and sex-aggregated logistic-normal survey length compositions. The correlation represents bin correlations | 0.01 |
ln_SrvLen_pop_theta |
n_pop × n_regions × n_sexes × n_srv_fleets |
Log-scale overdispersion parameter for population-specific survey length compositions | 0 |
ln_SrvLen_pop_theta_agg |
n_pop × n_srv_fleets |
Log-scale overdispersion for population-specific sex-aggregated survey length compositions | 0 |
SrvLen_pop_corr_pars |
n_pop × n_regions × n_sexes × n_srv_fleets × 2 |
Correlation parameters for logistic-normal population-specific survey length composition likelihood. The first parameter represents bin correlations, while the second represents sex correlations (if specified) | 0.01 |
SrvLen_pop_corr_pars_agg |
n_pop × n_srv_fleets |
Correlation parameters for population-specific sex-aggregated logistic-normal survey length compositions. The correlation represents bin correlations | 0.01 |
ln_Srv_caal_theta |
n_regions × n_sexes × n_srv_fleets |
Log-scale overdispersion for survey conditional age-at-length data,
one per region and sex and shared across length bins. Estimated only
under Srv_caal_LikeType = "Dirichlet-Multinomial"
|
0 |
ln_Srv_caal_theta_agg |
n_srv_fleets |
Log-scale overdispersion for region- and sex-aggregated survey conditional age-at-length data | 0 |
Natural Mortality
| Parameter | Dimension | Description | Default |
|---|---|---|---|
ln_M |
n_pop_block × n_regions_block × n_years_block × n_seasons_block × n_ages_block × n_sexes_block |
Log-scale natural mortality rate (dimensioned by user-defined blocks) | 0.5 |
Growth
Estimated only when growth_model is
"vb_schnute" or "richards", which build the
size-age transition inside the model. n_growth_pars is 5
under "vb_schnute" (L1, L2, K, CV1, CV2) and 6 under
"richards", which adds rho.
| Parameter | Dimension | Description | Default |
|---|---|---|---|
ln_growth_pars |
n_pop × n_regions × n_sexes × n_growth_pars |
Log-scale growth parameters in the order L1, L2, K, CV1, CV2 and
rho. Shared according to growth_spec
|
The log of L1 and L2 at the smallest and largest length bins, K of 0.15, CVs of 0.1 and rho of 1 |
ln_growth_devs |
n_pop × n_regions × (n_years + n_proj_yrs_devs) × n_growth_pars × n_sexes |
Annual deviations of each growth parameter under
growth_tv_model, entering on the scale
growth_tv_link sets |
0 |
ln_growth_semipar_devs |
n_pop × n_regions × (n_years + n_proj_yrs_devs) × n_ages × n_sexes |
Year by age deviations multiplying the parametric mean length under
growth_semipar
|
0 |
growth_pe_pars |
n_pop × n_regions × max(4, n_ages, n_growth_pars) × n_sexes × 2 |
Process error for growth. The first half (last index 1) holds one
log sigma per time-varying growth parameter. The second holds the
semi-parametric surface’s correlations by age, year and cohort in slots
one to three and its log scale in slot four, or one log sigma per age
under "iid" and "rw"
|
log(0.1) in the first half and log(0.05) in the second, with correlations of 0.3 where a form reads them |
Recruitment
| Parameter | Dimension | Description | Default |
|---|---|---|---|
ln_global_R0 |
n_pop × n_R0_blocks |
Log-scale unfished/mean recruitment per population and R0 time
block. One column unless R0_blocks asks for more, so a
model without R0 blocks is unchanged. Only the recruitment computed each
year reads that year’s block; the initial age structure, the regional
apportionment, the R0 prior, the ln_rinit penalty and the
stock-recruit scale all read R0_ref_block. Also supplies
the scale of the stock-recruit penalty curve when
sr_R0_spec = "shared" (the default) |
15 |
ln_rinit |
n_pop |
Log-scale initial recruitment scalar per population. Initializes the
equilibrium age structure when use_rinit = 1, and
additionally supplies the scale of the stock-recruit penalty curve when
sr_R0_spec = "rinit", so one parameter sets both; ignored
when use_rinit = 0. |
15 |
ln_sr_R0 |
n_pop |
Log-scale asymptote of the stock-recruit curve fitted as a penalty
under rec_model = "mean_rec". Estimated only when
sr_R0_spec = "est" and sr_penalty is not
"none"; mapped off in every other configuration, so it
never enters the parameter vector by accident |
ln_global_R0 |
rec_region_prop_pars |
n_pop × (n_regions - 1) |
Logit-scale regional recruitment proportion parameters (simplex
parameterization; omitted when
rec_region_prop_spec == 1) |
0 |
rec_seas_prop_pars |
n_pop × (n_seas - 1) |
Logit-scale seasonal recruitment proportion parameters (simplex
parameterization; omitted when
use_fixed_rec_seas_prop == 1) |
0 |
steepness_h |
n_pop × n_regions |
Steepness of the stock-recruitment relationship (logit scale,
bounded 0.2 to 1). Used by both "bh_rec" and
"ricker_rec", and by the curve fitted as a penalty under
rec_model = "mean_rec" with sr_penalty; mapped
off under mean recruitment with no such curve. Steepness values are not
interchangeable between the two curves (see
vignette("c_model_equations")) |
0 |
ln_sigmaR |
2 × n_pop × n_regions |
Log-scale standard deviations for early (index 1) and late (index 2) recruitment periods | 0 |
ln_InitDevs |
n_pop × n_regions × (n_ages - 1) × n_sexes |
Log-scale initial age structure deviations (can be specified as
random effects). Under InitDevs_sex_spec = "est_shared_s"
(default) every sex reads one shared curve; under "est_all"
each sex has its own |
0 |
ln_RecDevs |
n_pop × n_regions × (n_recdev_years + n_proj_yrs_devs) |
Log-scale annual recruitment deviations w/ projection years (can be specified as random effects) | 0 |
RecDevs_rho |
n_pop × n_regions |
Unconstrained AR1 correlation of the recruitment deviations,
transformed to (-1, 1) via
2 / (1 + exp(-2x)) - 1. Read only under
RecDevs_model = "ar1" and shared according to
RecDevs_rho_spec
|
0 |
ln_NAA |
n_pop × n_regions × n_years × n_seasons × n_ages × n_sexes |
Log-scale numbers at the start of a season when NAA_re
is on. A centered state, the log numbers themselves rather than a
deviation, estimated for ages two and older in years two onward
including the plus group; cells outside NAA_re_ages,
NAA_re_years and NAA_re_seasons are mapped
off. Season one is the year boundary, so the default
NAA_re_seasons = "annual" leaves the numbers deterministic
within a year. Specified as random effects. Starts from an equilibrium
decay from
at the mean
,
split over regions and sexes, with the plus group accumulated; supply
ln_NAA in starting_values to override |
equilibrium |
ln_sigmaNAA |
n_pop_block × n_regions_block × n_years_block × n_seasons_block × n_ages_block × n_sexes_block |
Log-scale process error standard deviation for ln_NAA,
dimensioned by the NAA_sigma_*blk_spec blocks as
ln_M is. Only NAA_re = "iid" admits variation
over the year and age dims; the season dim varies freely unless
NAA_re_season = "us"
|
log(0.3) |
NAA_pe_pars |
n_pop × n_regions × 3 × n_sexes |
Unconstrained correlations over the age-year grid for
ln_NAA, read as age, year and cohort. Only the slots the
chosen NAA_re form uses are estimated |
0 |
NAA_region_corr_pars |
n_pop × n_regions(n_regions - 1)/2 × n_sexes |
Unconstrained parameters for an unstructured correlation across
regions, shared according to NAA_re_region_spec
|
0 |
NAA_season_corr_pars |
n_pop × n_k(n_k - 1)/2 × n_sexes |
Unconstrained parameters for an unstructured correlation across the
active seasons, shared according to NAA_re_season_spec
|
0 |
NAA_pop_corr_pars |
n_pop(n_pop - 1)/2 |
Unconstrained parameters for an unstructured correlation across populations, global to the model | 0 |
NAA_sex_corr_pars |
n_sexes(n_sexes - 1)/2 |
Unconstrained parameters for an unstructured correlation across sexes, global to the model. One parameter with two sexes, and weakly identified unless sex-structured compositions hold the contrast | 0 |
sexratio_pars |
n_pop × n_regions × n_blocks |
Logit-scale sex ratio parameters (defaults to 50:50 when n_sexes = 1) | 0 |
stray_rate_pars |
n_pop × n_stray_blocks |
Logit-scale stray rate parameters controlling the proportion of
population p’s spawning biomass in a non-natal region that contributes
to other populations’ effective SSB at their natal regions. Only
estimated when use_fixed_stray_rate == 0. Note: stray rate
is likely unidentifiable from fishery and survey data alone and should
be fixed or given an informative prior. |
0 |
Dynamic Structural Equation Model
Only present when Setup_Mod_DSEM() was called. The grid
has one column per series, covariates first, then the linked deviation
series.
| Parameter | Dimension | Description | Default |
|---|---|---|---|
dsem_beta |
n_paths |
Path coefficients, one per named path or covariance arrow, in the
order the arrows name them (data$dsem_model$beta_names). A
fixed arrow (NA name, value in the fourth field) has no
entry here. |
0, or the arrow’s start value |
ln_dsem_sd |
n_sd_lines |
Log innovation standard deviations, one per sd line
(data$dsem_model$ln_sd_names). |
log(1), or the arrow’s start value |
dsem_mu |
n_series |
Series means. Estimated for the covariates dsem_mu_spec
names, kept at the observed mean otherwise, and always zero for a
deviation series. |
observed mean, 0 |
dsem_x |
n_grid_years × n_series |
The dsem grid. Known covariate cells and linked deviation cells are
mapped off (the objective writes the deviations in); missing covariate
years, "normal" covariates and the rows a linked series
does not reach (the forecast) are estimated. |
the series means |
logit_dsem_tweedie_p |
n_covariates |
Tweedie power as 1 + plogis(.), estimated for
dsem_family = "tweedie" only. |
|
ln_dsem_obs_sd |
n_covariates |
Log spread of each covariate’s observations on its family’s scale:
sd ("normal"), CV ("gamma"), sd of the log
("lognormal"), dispersion ("tweedie"); mapped
off for "fixed". |
log(sd / 2) |
dsem_delta0 |
n_series |
First-year offset of each series, estimated for those
dsem_delta0_spec names. |
0 |
Movement
| Parameter | Dimension | Description | Default |
|---|---|---|---|
move_pars |
n_pop × n_regions × (n_regions - 1) × n_years × n_seas × n_ages × n_sexes |
Logit-scale movement probabilities (unstructured Markov; move_type = 0) | 0 |
log_move_diffusion_pars |
(Vector; depends on formula) | Log-scale diffusion parameters for CTMC movement (move_type = 1) | log(0.1) |
move_preference_pars |
(Vector; depends on formula) | Preference parameters for CTMC movement (move_type = 1) | 0 |
move_devs |
n_pop × n_regions × (n_regions - 1) × (n_years + n_proj_yrs_devs) × n_seas × n_ages × n_sexes,
the third dim 1 under CTMC movement |
Movement deviations w/ projection years (can be specified as random
effects); offsets on the logit of a region pair (unstructured) or on a
region’s preference (CTMC); one surface per pair whose density is the sd
times a Kronecker product over the dims, set by
move_year_re, move_age_re,
move_pop_re, move_seas_re and
move_sex_re
|
0 |
move_pe_pars |
n_regions × (n_regions - 1) × 3, the second dim
1 under CTMC movement |
Log sd, age correlation and year correlation of each movement
process error block of pairs; a correlation is estimated only where its
dim is ar1
|
0 |
move_pop_corr_pars, move_seas_corr_pars,
move_sex_corr_pars
|
n(n - 1) / 2 each, at least 1 |
Unstructured correlation of the movement deviations across
populations, seasons and sexes, shared by every pair; estimated only
under us
|
0 |
Tagging
| Parameter | Dimension | Description | Default |
|---|---|---|---|
ln_init_conv_tag_mort |
n_conv_tag_cohorts |
Log-scale initial tagging mortality rate, one per release event; a scalar is recycled | -1000 |
ln_conv_tag_shed |
n_conv_tag_cohorts |
Log-scale chronic tag shedding rate, one per release event; a scalar is recycled | -1000 |
conv_tag_fish_reporting_pars |
n_regions × n_blocks × n_fish_fleets |
Logit-scale tag reporting rates | 0 |
ln_conv_fish_tag_theta |
1 |
Log-scale overdispersion parameter for tag recapture likelihood | 0 |