Incorporates population-specific, seasonal, and discarding dynamics.
Major changes
Incorporated ability to simulate and estimate both population-specific (natal homing) and seasonal dynamics.
Most model and simulation dimensions now include population- and season-specific indices, following the general dimension order: population, region, year, season, age, sex, fleet.
Recoded tagging module to allow fleet-specific tag reporting rates, as well as missing attributes in tagged fish.
Reference points for local density-dependence in both meta-population and natal homing contexts.
Added ability to simulate and fit to population-specific catches, indices, compositions, and tagging data.
Added ability to simulate and fit to population-specific discards and discarded compositions.
Added in ability to internally construct OSA residuals for catch, indices, compositions, and tagging data.
Added Fdev_model option ("iid", "rw", "ar1") to Setup_Mod_Catch_and_F for fishing mortality deviations (ln_F_devs), with a new AR1 correlation parameter Fdev_rho (shared/fixed via Fdev_rho_spec). Previously only IID deviations were supported.
Random walk and AR1 fishing mortality deviations no longer require catch-active years to be contiguous. A fishery closure spanning multiple years is now handled via the exact closed-form marginal transition over the elapsed gap d (random walk: variance inflates to d*sigma^2; AR1: mean decays by rho^d, variance becomes sigma^2 * sum_{i=0}^{d-1} rho^(2i)) – mathematically identical to estimating deviations for the closed years and integrating them out, without actually estimating them. Both reduce exactly to the standard single-step transition when d = 1.
Fishing mortality deviations now distinguish a genuinely missing aggregate catch observation from a true recorded zero. If ObsCatch is NA at a cell where UseCatch == 0 (and no population-specific catch is used), fishing is assumed to have continued and Fmort/ln_F_devs are estimated as an ordinary active year; if ObsCatch holds a real value (typically 0), the cell is treated as a genuine closure (Fmort forced to zero, no deviation estimated), matching prior behavior. See @param ObsCatch in Setup_Mod_Catch_and_F and Get_Fdev_PE_loglik.
Added age-0 (rec_lag = 0) Beverton-Holt recruitment, set via rec_lag in Setup_Mod_Rec/Setup_Sim_Rec. Previously rec_lag had to be >= 1 (recruitment driven by SSB from rec_lag seasons prior, entering in any season). With rec_lag = 0, recruitment for a year is driven by that same year’s own SSB.
Minor changes
Changed parameter names of ln_srv_fixed_sel_pars and ln_fish_fixed_sel_pars to srv_fixed_sel_pars and fish_fixed_sel_pars for clarity.
Included new options to estimate non-parametric selectivity, logistic selectivity with an asymptote parameter, as well as provide fixed selectivity (fishery, retention, and survey) inputs.
Changed dimensions of init_F_prop to be region, season, and fleet-specific (as opposed to just being based on the first fishery fleet).
Coded in an argument in do_retrospective to return retrospective RTMB model objects (return_models).
Added 95% confidence intervals for SDNR based on a Chi squared test for OSA residuals.
Force non-parametric selectivity to be mean-standardized.
Added OSA residuals and oneStepPredict functionality to time-series observations (indices and catch).
Improvements
Allowed both numeric and character codes for specifying dynamics in simulator, so as to be more consistent with how estimation models are specified, while maintaining backwards compatibility.
Added a generic internal process-error sharing-spec helper (build_pe_map/build_shared_spec_map) and migrated sigmaF_spec, sigmaC_spec, sigmaR_spec, Fdev_rho_spec, and the fishing mortality deviation map (do_Fmort_mapping) onto it, replacing hand-enumerated per-combination branches with a single dimension-collapsing implementation (no change in behavior for existing models) (for developers).
Refactored movement’s continuous process-error map (do_cont_vary_move_mapping) and log-likelihood (Get_move_PE_loglik) to use the same generic sharing-spec machinery and a single dimension-aware likelihood loop, replacing 10 hand-written iid_* branches in each (no change in behavior for existing models; new dedicated tests added since this module previously had no coverage) (for developers).
Bug Fixes
Fixed bug on smoothing penalty of time-varying selectivity where it was not being applied in the first year.
Fixed inconsistent logic between the fishing mortality and discard mortality rate zeroing conditions for population-specific catch: Fmort was forced to zero if any population’s catch was unused, while the parallel discard mortality rate condition correctly required all populations to be unused. Both now consistently use the latter.
Fixed a data issue in the bundled sgl_rg_sable_data dataset where 3 cells of ObsCatch (fleet 2, years 1-3, representing a fleet that had not yet started fishing) were NA; corrected to 0 to represent a true closure under the new missing-vs-true-zero convention above (prior fitted results for this dataset are unaffected).
version 1.1.0
Release Date: 2026-3-31
Major changes
Added capability to estimate movement using continuous time Markov chains (CTMC) using preference functions.
Bug Fixes
Fixed bug to allow for size-based selectivity in the estimation model.
version 1.0.0
Release Date: 2025-11-24
Major changes
First stable public release of SPoRC.
Introduced a modular framework for estimating age-, sex-, season-, and region-structured population dynamics.
Added closed-loop simulation and management strategy evaluation (MSE) capabilities.