In CefasRepRes/FishRAM: Fisheries Resource Allocation Model

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)

library(FishRAM)

Introduction

The FishRAM model implements the bioeconomic model of @Tidbury2021, which represents biological and economic components of a simplified single-species fishery exploited by commercial and recreational fleets. The stock is split into two disjoint stocks: a juvenile stock affected by natural processes (mortality and recruitment), but unaffected by fishing, and an adult stock on which fishing acts in addition to natural processes. The commercial fleet generates revenue from the landed catch of the stock and accrues losses from the costs associated with the trips. The profitability of the commercial fleet determines if commercial activity in the fishery increases or decreases over time, up to some maximum threshold capping the total number of commercial trips. The recreational fleet is driven by the catch per unit effort (CPUE) of recreational fishers. If the CPUE is above some threshold, then recreational activity in the region increases, otherwise it decreases. The economic impact of each fleet is also described in the model. This impact is proportional to the total revenue generated for the commercial fleet, and proportional to the total expenditure of anglers for the recreational fleet.

In this vignette, we describe how to setup parameters, run models, and explore results in the FishRAM package.

Setting up the model

In FishRAM, model parameters are contained in objects of class BioeconomicParams. The simplest way to construct a custom BioeconomicParams object is reading in a csv file using the BioeconomicParams constructor. The csv file should contain the parameter name in the first column and the value in the second column, like this: ``` {r params_csv_dem, echo = FALSE} knitr::kable(data.frame("Parameter" = c("g", "p", "b", "qC", "qR", "..."), "Value" = c(0.1, 0.05, 1, 12.4, 1.1, "...")))

Each parameter value should be specified in this file, with no additional lines in the file. To create a `BioeconomicParams` object from this file, call the constructor as follows:
``` {r create_params_csv, eval = FALSE}
params <- BioeconomicParams(file = "params.csv")

In BioeconomicParams objects, each parameter required for the model is a slot in the object, which can be manually overwritten later. The following code changes the natural mortality on the juvenile stock. ``` {r override_mort_params, eval = FALSE} params@muJ <- 0.3

Parameters from the Sea Bass fishery studied in @Tidbury2021 is included in the package and can be loaded from the `"seabass"`dataset.
```r
data("seabass")

Running the model

Models can be run by calling the project() function on a BioeconomicParams object with a specified recruitment scenario and duration. This function outputs a BioeconomicSim object that contains the model results.

Specifying recruitment

In FishRAM, recruitment can be described either by explicitly defined recruitment values for each year, or as a function of the adult stock size to be calculated dynamically at each time step. Different values of the R parameter in the project() functions help to specify this.

Constant recruitment

To run a simulation with constant recruitment, the R argument should be a numeric giving the recruitment across the full time period. The t_start and t_end arguments specify how long the model should be run for.

``` {r recruitment_constant} sim <- project(params, R = 7e5, t_start = 2020, t_end = 2045)

#### Explicit recruitment values
For varying, but explicitly described recruitment, the `R` argument should be a `numeric` vector of length `t_end - t_start + 1`, with each element giving the recruitment in each year.

``` {r recruitment_vector}
set.seed(123)
# Generate log-normally distributed recruitment values.
recruitment <- exp(rnorm(26, log(7e5), 0.5))
sim <- project(params, R = recruitment, t_start = 2020, t_end = 2045)

Implicit recruitment functions

To specify dynamic recruitment functions, R should be a functiontaking exactly one numeric argument. At each time step, recruitment is calculated using this function, passing through the adult stock size from the previous time step (i.e the recruitment at time $t$ is calculated as $R(SA(t-1))$ where $SA$ is the adult stock size). If R is a function, then the R_init argument must also be specified. R_init gives the initial recruitment at the start of the simulation.

# Use a type II functional response for recruitment
rec_func <- function(stock){
  a <- 1.5e-3; b <- 7e-4; h <- 1.3e-4
  return(
    1000 *a*stock/(b + a*h*stock)
  )
}
sim <- project(params, R = rec_func, t_start = 2020, t_end = 2045, R_init = 7e5)

Model outputs

The project() function returns an object of class BioeconomicSim. The outputs of the simulation are stored in the states slot of this object. The states slot is a $T \times 17$ matrix where $T$ is the duration of the simulation. Each row in the matrix gives the states of the system at a given time step.

Plotting functions

FishRAM contains 3 different plotting functions to explore different elements of the model outputs. The stock size over the simulation can be plotted using the plotStock() function

plotStock(sim)

The number of trips for each fleet can be plotted using the plotActvity() function

#Specify the fleet using the "fleet" argument.
plotActivity(sim, fleet = "R") # The recreational fleet
plotActivity(sim, fleet = "C") # The commercial fleet

The economic impact of each fleet can be plotted using the plotEconomicImpact() function

plotEconomicImpact(sim)

Other model outputs

Other components of the simulator outputs can be extracted from the relevant column in the states matrix of the BioeconomicSim object.

``` {r manual_plots, fig.dim = c(7, 4)}

Plot the profit of the commercial fleet over time

plot.ts(sim@states$tau)

Plot the mortality of the adult stock over time

plot.ts(sim@states$ZA) ```

References

CefasRepRes/FishRAM documentation built on Feb. 1, 2023, 1:15 a.m.