Skip to contents

Management strategy evaluation performance metric summary

Source: R/10-mse_summary.R
mse_summary.Rd

Management strategy evaluation performance metric summary

Usage

mse_summary(mse, om_only = FALSE)

Arguments

mse

MSE runs from run_mse or load_mse

om_only

If TRUE, report only operating-model (true) status and skip the estimation-model (EM) perception metrics.

Value

A named list, one element per entity dimension so each metric lives only where it applies (no NA padding):

  • species, a data.frame with one row per species (keyed by Species) of the conservation / status metrics: per-species average catch, catch inter-annual variability (IAV), and P(Closed) = the probability the fishery is closed (catch ~ 0); the relative mean-squared error of terminal and average SSB; the estimation-model- and operating-model- perceived overfishing / overfished probabilities (via build_hcr) and the probability each status is misclassified (estimation model disagrees with the operating-model truth); and terminal biomass, SSB, dynamic SB0, SSB depletion (equilibrium and dynamic), average SSB depletion, and SSB-collapse counts.

  • fleet, a data.frame with one row per fishery fleet (keyed by Fleet_code / Fleet_name) of average catch, catch IAV, and P(Closed).

  • total, a named numeric of the across-fleet total average catch and catch IAV.

  • meta, run provenance (nsim, nspp, nflts, HCR, projection-year range).

All metrics are averaged across projection years and simulations.

Metric columns hold syntactic names, so they can be typed without backticks: avg_catch, catch_iav, p_closed, ssb_rmse_avg, ssb_rmse_terminal, em_p_overfishing, em_p_overfished, om_p_overfishing, om_p_overfished, p_overfishing_false_pos, p_overfishing_false_neg, p_overfished_false_pos, p_overfished_false_neg, om_terminal_biomass, om_terminal_ssb, om_terminal_dynamic_sb0, om_terminal_depletion, om_terminal_depletion_dynamic, om_avg_depletion, om_sims_collapsed, om_no_f_sims_collapsed, om_sims_collapsed_from_f. An om_ prefix is the operating model's truth and em_ the estimation model's perception; the four *_false_pos / *_false_neg metrics are the probability the two disagree. The three *_sims_collapsed metrics are counts of simulations, not probabilities.

The three depletions divide by different references:

  • om_terminal_depletion_dynamic: terminal SSB over the OM's own DynamicSB0, its history with no fishing on the stock-recruit curve with the realized recruitment deviations and, under predation, the suitability fitted in the hindcast. Defined for single- and multispecies OMs. A species with input numbers-at-age (estDynamics > 0) keeps those numbers in the dynamic run, so its dynamic depletion is near 1 and says nothing about fishing.

  • om_terminal_depletion: terminal SSB over SB0, the equilibrium value for a single-species OM. Under msmMode > 0 it is MSSB0, SSB at the end of a no-fishing projection, which fit_mod() derives only when a harvest control rule is present; without one (HCR = "NoFishing") this is NA.

  • om_avg_depletion: the OM's own ssb_depletion averaged over the projection. Under predation with no HCR that series divides by SSB in the last projection year, not by an unfished reference.

The collapse metrics compare the OM with OM_no_F, the OM refit with no fishing after the original operating model's endyr (remove_F()).

Each frame holds a "labels" attribute mapping those names to the long display strings (e.g. om_terminal_depletion_dynamic -> "OM: Terminal SSB Depletion (Dynamic)") for plots and tables: attr(summ$species, "labels").

Examples

if (FALSE) { # \dontrun{
mse <- load_mse(dir = "mse_runs")
s <- mse_summary(mse)
s$species   # one row per species
s$fleet     # one row per fleet
s$total     # across the whole system
s$meta      # nsim, nspp, nflts, HCR
} # }