Skip to contents

Changelog

Source: NEWS.md

marlin 1.1.1

Behavior Changes

  • The default youngest age (min_age) tracked by create_critter() / Fish$new() is now one time step (1 / seasons) rather than 0. Because recruits are deposited at the youngest age after the population is aged, the youngest cohort is already one time step removed from the spawning pool, so age 0 (“at birth”) was never really represented; starting at one time step is more consistent with that, keeps the extrapolated (below the Lorenzen fitting range) age-0 natural mortality out of the model, and makes per-recruit quantities comparable with age-1-indexed assessment frameworks. Pass min_age = 0 to recover the previous behavior. Existing simulations will see a shifted age vector and, with default Lorenzen M, somewhat different absolute per-recruit quantities (this does not move SPR- or F_msy-type reference points).

  • Lorenzen natural mortality (lorenzen_m = TRUE, the default) is now anchored so that the mean M over the adult age range equals the supplied m, rather than pinning m at the asymptotic (largest/oldest) size. A single reported m is therefore interpreted as an average adult rate — consistent with how tmax- and growth-based empirical M estimators are derived, and analogous to the age-range Lorenzen option in Stock Synthesis (Powers 2014, ICES J. Mar. Sci. 71:1629). The new lorenzen_ref_age argument to create_critter() / Fish$new() sets the lower age bound of that range (defaults to the age at maturity). Existing simulations using the default Lorenzen M will produce slightly different M-at-age and, in turn, unfished equilibria and tuned catchabilities.

Bug Fixes

  • Fish$plot_movement() no longer errors on its first call in a fresh session (e.g. on a critter restored via readRDS()). Matrix is now a declared dependency, so its S4 methods are registered when marlin loads, and the plot-time multiply densifies the (small) movement matrices to avoid relying on those methods at all.

marlin 1.1.0

marlin 1.0.1

marlin 1.0.0

Breaking Changes

  • Travel costs are now set by travel_fraction, which species the proportion of the total costs that come from travel. cost_per_distance is now deprecated and may cause breaks

  • Old open access parameters no longer work. see ?create_fleet for new parameters

Improvements

  • Movement is now dealt with via sparse matrices, vastly increasing speed and memory efficiency in most cases

  • process_marlin now works vis data.table, is orders of magnitude faster

  • fleet spatial allocation strategies are now much more stable and consistent

New Capabilities

  • Added calculation of marginal revenues and marginal profits in space and time by fleet, allowing for spatial allocation models marginal_profit and marginal_revenue, as well as fleet_model sole_owner

  • Added flexible patch areas

marlin 0.8.0

marlin 0.7.0

marlin 0.6.1

marlin 0.6.0

marlin 0.5.0

marlin 0.2.0

marlin 1.0.3

marlin 1.0.2

marlin 1.0.1

marlin 1.0.0

marlin 0.0.0.9000

  • Added a NEWS.md file to track changes to the package.