Growth module — Get_Growth • SPoRC

Builds the size-age transition arrays and, when asked, weight at age and maturity at age from the growth parameters, for every year. Every fleet gets its own key and weight, read at that fleet's timing within the season: each fishery fleet's at t_fish, each survey's at t_srv, and the spawning weight at t_spawn.

Usage

Get_Growth(
  ln_growth_pars,
  growth_A1,
  growth_A2,
  growth_L0,
  growth_len_lower,
  growth_cv_type,
  growth_sd_type,
  growth_dist,
  growth_plus_group,
  growth_L2_asymptote = 0,
  derive_waa,
  wt_len_pars,
  ages,
  seasdur,
  spawn_seas,
  t_spawn,
  n_pop,
  n_regions,
  n_yrs,
  n_seas,
  n_sexes,
  n_fish_fleets,
  n_srv_fleets,
  t_fish,
  t_srv,
  ln_growth_devs = NULL,
  growth_tv_model = NULL,
  growth_tv_link = 0,
  growth_par_bounds = NULL,
  growth_tv_type = 0,
  growth_cohort_styr = 1,
  years_eval = NULL,
  ln_growth_semipar_devs = NULL,
  growth_semipar = 0
)

Arguments

ln_growth_pars: Array [pop, region, sex, n_gpars] of log growth parameters in the order L1, L2, K, CV1, CV2 and, for the Richards form, rho.
growth_A1, growth_A2: Reference ages; see get_laa_curve.
growth_L0: Length at age zero.
growth_len_lower: Lower edges of the length bins.
growth_cv_type, growth_sd_type, growth_dist: Integer switches passed to get_laa_curve and get_alk.
growth_plus_group: Integer (0/1); apply plus_group_size.
derive_waa: Integer (0/1); build weight at age from the key.
wt_len_pars: Array [pop, region, sex, 2] of weight-length parameters \(a, b\) in \(W = a L^b\), read when derive_waa = 1.
ages: Numeric vector of model ages.
seasdur: Season durations as fractions of a year.
spawn_seas, t_spawn: Spawning season and fraction of it elapsed at spawning.
n_pop, n_regions, n_yrs, n_seas, n_sexes, n_fish_fleets, n_srv_fleets: Dimensions.
t_fish: Array [region, season, fish_fleet] of fishery timings, the fraction of the season elapsed at which each fishery fleet's key and weight at age are read.
t_srv: Array [region, season, srv_fleet] of survey timings, the fraction of the season elapsed when each survey is taken, at which that survey's key and weight at age are read.
ln_growth_devs: Array [pop, region, year, n_gpars, sex] of time-varying deviations, or NULL for none.
growth_tv_model: Integer vector [n_gpars], 0 constant, 1 iid, 2 random walk, per parameter.
growth_tv_link: Integer, 0 log link, 1 logit link within growth_par_bounds.
growth_par_bounds: Matrix [n_gpars x 2] of bounds for the logit link.
growth_tv_type: Integer, 0 each year on its own curve, 1 cohort propagation.
growth_cohort_styr: Year index the cohort propagation starts from.
years_eval: Integer vector of the years to evaluate; the default is every year.
ln_growth_semipar_devs: Array [pop, region, year, age, sex] of log deviations on mean length at age, or NULL for none. Multiplies the parametric mean at age, so the curve stays the parametric part and the deviations have departures from it; the spread at age follows the deviated mean, leaving the coefficient of variation at age alone.
growth_semipar: Integer process error code for those deviations, 0 for none. Only its being nonzero is read here; the structure is penalized in the objective.

Value

List with SizeAgeTrans_fish and SizeAgeTrans_srv [pop, region, year, season, len, age, sex, fleet], one key per fleet at that fleet's timing, the spawning key SizeAgeTrans_spawn [pop, region, year, len, age, sex], mean_LAA_fish, sd_LAA_fish, mean_LAA_srv, sd_LAA_srv [pop, region, year, season, age, sex, fleet], mean_LAA_spawn and sd_LAA_spawn [pop, region, year, season, age, sex], L_beg [pop, region, year, age, sex] (the start-of-year mean length), Linf [pop, region, year, sex], growth_pars_y [pop, region, year, 6, sex], and when derive_waa = 1 also WAA, WAA_fish and WAA_srv.

Time variation

With no deviations the key is built once and broadcast over the years. With deviations on any parameter and growth_tv_type = 0 ("curve") every year's sizes are read from that year's own curve. Under growth_tv_type = 1 ("cohort") the sizes are advanced cohort by cohort from growth_cohort_styr on, which needs the numbers at age of each year to blend the plus group; that form is run one year at a time from the population dynamics through Get_Growth_Year, and this function only evaluates the years before the propagation starts, which all sit on the first year's curve.

Timing within the year

A season starts at the cumulative duration of the seasons before it , and a point inside a season is that start plus the fraction of the season elapsed times the season's duration , so every evaluation is anchored at the season start rather than compounded from the previous evaluation. Mean length is the curve read at the real age, the integer age plus that elapsed fraction of a year, and the plus group grows from its adjusted size by the curve's increment over the same elapsed time. Under constant growth the increment over a season is exactly the curve's difference between the two real ages, so splitting the year into seasons of any durations changes no mean length. The spawning weight is read at the spawning fraction of the season, and each fleet's key and weight at that fleet's own timing, t_fish or t_srv, so a composition and the weight behind an index are formed at the point in the season the observation is taken. Fleets that share a timing share one evaluation. Seasonal multipliers on \(K\) are not kept.