Gets a list of default initial values for the stock assessment model. Initial values are hand-coded from recent fitted assessment runs where available. Selectivity starts use the legacy v1-derived selectivity surface, then overlay recent fitted starts by matching the current data dimensions.
Value
A named list of initial parameter values containing:
- par_log_B0
Numeric. Log-scale unfished biomass.
- par_log_psi
Numeric. Log-scale psi parameter.
- par_log_m0, par_log_m4, par_log_m10, par_log_m30, par_mc
Numeric. Natural-mortality parameters selected by
data$M_switch.- par_log_h
Numeric. Log-scale steepness parameter.
- par_log_sigma_r
Numeric. Log-scale recruitment standard deviation.
- par_log_cpue_q, par_cpue_creep
Numeric. CPUE catchability coefficient(s) and creep parameter.
- par_log_cpue_sigma, par_log_cpue_omega
Numeric. Log-scale CPUE observation error parameters.
- par_log_aerial_tau
Numeric. Log-scale additional aerial-survey standard deviation.
- par_log_aerial_sel
Numeric vector (length 2). Log-scale aerial selectivity parameters.
- par_log_troll_tau
Numeric. Log-scale additional troll-survey standard deviation.
- par_log_gt_q, par_log_hsp_q
Numeric. Log-scale catchability for gene-tagging and half-sibling-pair data.
- pop_od, hsp_od, gt_od
Numeric. Fixed beta-binomial concentration parameters for parent-offspring, half-sibling, and gene-tagging data.
- par_log_tag_H_factor
Numeric. Log-scale tagging mortality factor.
- par_log_af_alpha
Numeric vector (length 2). Log-scale age-frequency alpha parameters.
- par_log_lf_alpha
Numeric vector (length 5). Log-scale length-frequency alpha parameters.
- par_sel_rho_y, par_sel_rho_a
Numeric vectors (length 7). Logit-scale selectivity temporal and age correlation parameters.
- par_log_sel_sigma
Numeric vector (length 7). Log-scale selectivity standard deviations.
- par_log_sel_1 through par_log_sel_7
Arrays. Log-scale selectivity by year and age for each fishery.
- par_rdev_y
Numeric vector. Annual recruitment deviations.
Details
Current-model defaults set the additional aerial-survey standard
deviation to 0.59, copy the fixed LL3 selectivity hyperparameters to LL4,
and set the length-mortality exponent to -1 when
data$M_switch = 2. The corresponding fixed or estimated treatment
is defined by get_map().
