Skip to contents

Compute Fishing Mortality Deviation Process Error Log-Likelihood (Negative Scale)

Source: R/model_priors_penalties.R
Get_Fdev_PE_loglik.Rd

The negative log-likelihood of ln_F_devs under an iid, random walk or AR1 process. Unlike Get_PE_loglik and Get_move_PE_loglik, which return one positive scalar for the caller to negate, this returns an already-negated array shaped like ln_F_devs, zero where catch is not used, matching the Fmort_nLL reporting.

Usage

Get_Fdev_PE_loglik(
  PE_model,
  ln_sigmaF,
  Fdev_rho,
  ln_F_devs,
  map_ln_F_devs,
  Fdev_pen_center = 0
)

Arguments

PE_model

Integer process error structure: 1 iid, 2 random walk with the first active year on a diffuse \(N(0, 5)\), 3 AR1 with the first active year drawn from its stationary marginal.

ln_sigmaF

Array [n_regions x n_seas x n_fish_fleets] of log-scale process error sd.

Fdev_rho

Array [n_regions x n_seas x n_fish_fleets] of unconstrained AR1 partial correlation, transformed to \((-1, 1)\) here and read under PE_model == 3 only.

ln_F_devs

Array [n_regions x n_years x n_seas x n_fish_fleets] of log-scale fishing mortality deviations.

map_ln_F_devs

Array shaped like ln_F_devs mirroring $map$ln_F_devs: an estimation index where a deviation is estimated and NA where it is fixed. Only estimated deviations are penalized and they alone form the active sequence, so a fixed cell is skipped and widens the gap between the deviations either side of it. do_Fmort_mapping builds it from the catch use indicators, estimating a deviation wherever aggregated or population-specific catch is used, or where ObsCatch is NA, which is a missing value rather than a real closure; a recorded zero is a closure and is excluded.

Fdev_pen_center

Integer. 1 centers on the deviations' own mean, 0 on zero.

Value

Array shaped like ln_F_devs of the negative log-likelihood per cell.

Details

The walk and the AR1 do not need catch-active years to be contiguous: the transition between two active years is taken over the elapsed gap between them, which is the same marginal as estimating deviations for the closed years and integrating them out, and reduces to the single-step transition when the gap is one year.