Population dynamics — do_dynamics • opal

Runs the core age- and season-structured population dynamics loop for bigeye tuna. Starts from initial equilibrium numbers (derived from B0 and h), applies seasonal harvest, natural mortality, spawning, and recruitment (Beverton-Holt with log-normal deviates), and computes predicted catches and harvest rates.

Usage

do_dynamics(
  data,
  parameters,
  B0,
  R0,
  alpha,
  beta,
  h = 0.95,
  sigma_r = 0.6,
  M_a,
  spawning_potential_a,
  weight_fya,
  init_number_a,
  sel_fya
)

Arguments

data: A list of model data. Must contain at minimum: first_yr, first_yr_catch, n_year, n_season, n_fishery, n_age, catch_obs_ysf, catch_units_f.
parameters: A list of model parameters. Must contain at minimum: rdev_y.
B0: Numeric. Unfished equilibrium spawning biomass.
R0: Numeric. Unfished equilibrium recruitment.
alpha: Numeric. Beverton-Holt stock-recruitment alpha parameter.
beta: Numeric. Beverton-Holt stock-recruitment beta parameter.
h: Numeric (0.2–1). Steepness of the Beverton-Holt stock-recruitment relationship.
sigma_r: Numeric > 0. Standard deviation of log recruitment deviations.
M_a: Numeric vector of length n_age. Natural mortality at age. Passed explicitly so AD gradients propagate if M is ever estimated.
spawning_potential_a: Numeric vector of length n_age. Spawning potential at age (maturity × fecundity). Passed explicitly so AD gradients propagate if growth is ever estimated.
weight_fya: Numeric array [n_fishery, n_year, n_age]. Mean weight at age by fishery and year. Passed explicitly so AD gradients propagate if growth is ever estimated.
init_number_a: Numeric vector of length n_age. Initial equilibrium numbers-at-age (from get_initial_numbers).
sel_fya: Numeric array [n_fishery, n_year, n_age]. Fishery-specific selectivity at age by year (from get_selectivity).

Value

A named list with:

number_ysa: Numbers-at-age array [n_year+1, n_season, n_age].
lp_penalty: Total penalty from posfun (harvest rate constraints).
catch_pred_fya: Predicted catch-at-age array [n_fishery, n_year, n_age].

Details

All derived biology arrays (M_a, spawning_potential_a, weight_fya) are passed as explicit arguments rather than read from data. This ensures that AD gradients propagate correctly if any of these quantities carry estimated parameters in the future (e.g., growth parameters estimated via the PLA, or natural mortality via the Lorenzen equation).