README.md

In bstaton1/SimSR: Tools for Simulating Pacific Salmon Spawner-Recruit Dynamics

SimSR

This is an R package written to facilitate simulating Ricker spawner-recruit dynamics, particularly in the case of age-structured, mixed-stock fisheries. It was designed to generate input information for use by the functions in the FitSR package. The most current version may be installed using:

devtools::install_github("bstaton1/SimSR")

Specific versions can be installed using:

devtools::install_github("bstaton1/SimSR@v0.1.3")

Example Usage

The functions are all reasonably-well documented, see the help pages for more details on what each one does, or the source code for exact details.

The general workflow is laid out as follows:

1. Generate the parameters to drive the populations
2. Simulate the true dynamics of the populations, and store the states at each time step
3. Simulate the sampling of the populations

Step 1: Generate the true parameters

Use the init_sim() function to create a list of dimensional variables and driving parameters for use in many of the other functions in this package. Only the arguments U_msy and S_msy do not have defaults:

params = init_sim(
  U_msy = c(0.64, 0.46, 0.55, 0.37, 0.68, 0.5, 0.43, 0.5, 0.71, 0.61, 0.43, 0.53),
  S_msy = c(8000, 3300, 10800, 1100, 2600, 6300, 7300, 600, 2500, 1700, 300, 1100)
)
names(params)

The params object is a list storing the information needed to simulate the true and observed states.

Step 2: Simulate population dynamics

Use the ricker_sim() function to drive the populations forward in time. A simple harvest control rule is used: constant exploitation rate (with implementation error) equal among populations and set at the level that will maximize yield without overfishing more than p*100% of the populations, where p is either 0.1, 0.3, or 0.5.

true = ricker_sim(params, Ustrategy = "Ustar_0.5")

The true object is a list storing the true states, including recruitment by brood year and stock (R_ys), the realized explotiation rate (U_real), total run, escapement, and age composition by calendar year and stock (N_ts, S_ts, and q_tas, respectively).

Step 3: Simulate sampling the populations

Sampling proceeds in two steps: first, introduce measurement errors (the state that would be observed if that quantity was monitored for that stock and year), and second, impose a frequency of sampling. To introduce measurement errors in (i) population-specific escapement, (ii) population-specific age composition, and (iii) aggregate harvest, use the obs_sim() function:

obs = obs_sim(params, true)

The obs object contains the states that are possibly observable, with inserted measurement errors.

To impose a sampling frequency, use the obs_filter() function:

obs = obs_filter(params, obs, mimic = NULL, minSRobs = 3, p_age = 1)

This will impose a sampling frequency where each population has escapement monitored for 60% of the years, 100% of the populations are sampled for age composition, and a minimum of 3 observed pairs of spawners and recruits are available for each population. The sampling frequency can be structured to mimic a specified schedule, see ?obs_filter for more details.

Notes

The functions are not set up to handle age structures with fewer than three possible ages of maturity, but can handle scenarios with any number of ages greater than this.

The functions used for observing populations are not set up for fewer than two populations. The simulation of true dynamics can handle only one population.

Version v0.1.3 was used in the analyses for B. Staton's dissertation and the subsequent manuscript.

Bug reports or feature requests are invited as "Issues"; any contributions are invited through the pull request system.

bstaton1/SimSR documentation built on Aug. 20, 2019, 9:43 a.m.