Set up catchability deviations for the operating model

Gives each fleet a process error on annual catchability, matching srv_q_model and fish_q_model in the estimation setup. Catchability in replicate \(i\) becomes \(q_{r,y,f} \exp(\epsilon_{r,y,f,i})\), where \(q\) is the mean Setup_Sim_Fishery and Setup_Sim_Survey supplied.

Usage

Setup_Sim_q_devs(
  sim_list,
  fish_q_model = NULL,
  srv_q_model = NULL,
  sigma_fish_q = 0,
  sigma_srv_q = 0,
  fish_q_rho = 0,
  srv_q_rho = 0,
  fish_q_rw_init_sigma = NA,
  srv_q_rw_init_sigma = NA
)

Arguments

sim_list

List of operating model inputs, with fish_q and srv_q already set.

fish_q_model, srv_q_model

Character vectors, one per fleet, of "none" (default), "iid", "rw", "ar1" or "dsem". A "dsem" fleet draws nothing here and takes its series from Setup_Sim_DSEM.

sigma_fish_q, sigma_srv_q

Numeric arrays [n_regions, n_fleets] of deviation standard deviations on the log scale, or a single value used for every region and fleet. Default zero. An array [n_regions, n_fleets, n_sims] gives each replicate its own.

fish_q_rho, srv_q_rho

Numeric arrays [n_regions, n_fleets] of ar1 correlations on the natural scale, read only under "ar1". Default zero. An array [n_regions, n_fleets, n_sims] gives each replicate its own.

fish_q_rw_init_sigma, srv_q_rw_init_sigma

The sd the estimation model gives a walk's first estimated year (*_q_rw_init_sigma there). Under a value that year keeps each replicate's own deviation, since a diffuse start leaves the level to the data; under NA (default) it is drawn at the walk's own sd from zero, as the penalty reads it.

Value

sim_list with the deviation settings stored.

Details

Conditioning years read their deviations from the fit and draw nothing, so a random walk or ar1 in the projection continues from the fit's last deviation rather than restarting at zero.