jonrh / techinteractions

This program allows implementing a management strategy evaluation (MSE) for a multispecies fishery with catch constraints

techinteractions

Description

techinteractions is a program that allows implementing a management strategy evaluation (MSE) for a multispecies fishery with catch constraints e.g. bycatch constraints, cap in total species catch, mixed-species cap. The model combines a biological module, a quota allocation module and a vessel dynamics module to mimic the fishery. Quota allocation and vessel dynamics modules are both implemented using linear programming (a constrained optimization approach). This simple yet flexible approach is ideal for MSE as it allows including much complexity of both quota and vessel dynamics while saving time for running the MSE (generally takes few seconds to run these modules). Currently, the code is based on the Bering Sea Aleutian Islands groundfish fishery example.

Contents

• Structure of the techinteractions program
• Computer setup for running the MSE
• techinteractions MSE setup
• Output files
• File description

Structure of the techinteractions program

• doc contains documentations for understanding, learning, and running program, in addition to the manuscript
• R contains all the R codes and data needed to run various pieces of program
• runs contains all model configuration files (where we specify scenarios). Simulation results will be stored in a subfolder named results
• src contains all source codes (both fortran and ADMB)

Computer setup for running the MSE

Programs to install

The main body of techinteractions is written in fortran but the program calls for ADMD and R to run various parts of the model. Therefore, users are required to have the following software in order to run this program:

• install ADMB [here] (http://www.admb-project.org/downloads/) and follow their documentations/instructions to check whether the installation was performed correctly. P.S: Any of the above ADMB installation should come with fortran compiler to run the program so no need to install a separate fortran compiler.
• install R [here] (https://www.r-project.org/)

Setup the path

The next step is to set the path.

• path to ADMB: Follow the instructions written [here] (http://www.admb-project.org/downloads/) to check whether the path was correctly set. If not, please set the path to ADMB in your environment variables i.e. it should be the location where you installed AMDB e.g. C:\ADMB\bin
• path to fortran
• path to R: Go to the directory where you cloned techinteractions. shift + right mouse click above the techinteractions folder and select the option Open command window here. Run the command script

rscript R/test.r

If it prints the message path to R is already setup, you are all good. If not, please set the path to R in your environment variables i.e. it should be the location where you installed R e.g. C:\Program Files\R\R-3.2.2\bin

techinteractions MSE setup

Below is the illustration of the full MSE cycle ![MSE cycle] (https://github.com/Kotkot/techinteractions/blob/trial/doc/MSE_fig.png)

Creating a scenario for the multispecies MSE

Creating a scenario requires changes to many nodes in the program:

• the operating model: this specifies the underlying truth about the population dynamics (e.g. how they grow, recruit, fished)
• the assessment model: this both defines the data that are collected to perform the assessment and the assessment model configuration itself
• the management model: this defines/mimics the behavior of quota allocation process by fishery managers
• the vessel dynamics model: this defines/mimics fishers' behavior

ALL these configurations are controlled by several important files in the program:

• [runs/OM.dat] (https://github.com/Kotkot/techinteractions/blob/trial/runs/OM.DAT) : This file controls the operating model i.e. the underlying biological truth of the model (this includes growth, death, recruitment, but also fishing)
• [runs/EM.dat] (https://github.com/Kotkot/techinteractions/blob/trial/runs/EM.DAT) : This file controls the estimation/assessment model (quantity and quality of data that goes into the assessment and what is estimated and what is not)
• [runs/CR.dat] (https://github.com/Kotkot/techinteractions/blob/trial/runs/CR.dat) : This file controls the harvest control rule (option between the AFSC tier 3-like control rule of a P* approach)
• [runs/Random_seed_OM.dat] (https://github.com/Kotkot/techinteractions/blob/trial/runs/Random_seed_OM.dat) : This file controls the random seed that is used in the vessel dynamics model. More specifically, the seeds controls the catch composition in each metier. Important for reproducibility.
• [runs/Random_seed_EM.dat] (https://github.com/Kotkot/techinteractions/blob/trial/runs/Random_seed_EM.dat) : This file controls the random seed that is used in the quota allocation model. More specifically, the seeds controls the catch composition in each metier. Possibility are to make i/ the exact copy of Random_seed_OM.dat i.e. fishery managers know what vessel will catch in the future, ii/ lag it by one year so that fishery managers set quota based on what fishers did in the previous year (the default), and iii/ random so that fishery managers set quota assuming they know nothing about fishers behavior.
• [R/fishing strategy.r] (https://github.com/Kotkot/techinteractions/blob/trial/R/fishing_strategy.r) : This file controls in essence both the quota allocation and the vessel dynamics models. The above two files i.e. Random_seed_OM.dat and Random_seed_EM.dat are input values for this file. This R file runs linear programming to determine either the quota or the realized catch for a year. This file also controls the temporal dynamics of the linear programming i.e. how species' net prices (users' relative preference between species) change over time.
• [runs/Bycatch.DAT] (https://github.com/Kotkot/techinteractions/blob/trial/runs/Bycatch.DAT) : This is where we write down the fixed bycatch PSC limit for the scenario.

Step by step detail for setting an MSE scenario are provided in [specifying-scenario.Rmd] (https://github.com/Kotkot/techinteractions/blob/trial/doc/specifying-scenario.Rmd)

Compiling and running the scenario

Once you have set-up the scenario, now it is time to run it. This is done in two steps:

1. Compile the source codes. Open the command line within the src folder and type in make (this will run the Makefile). This will compile all necessary files to run the model
2. Run the model. Either open the command line within the runs folder and type run (this will run the run batch file) or double-click the run batch file. This will run the models

Output files

There are many files that are produced by the program that could be used to debug, explore, analyse, and interpret the results. [R/Codes_plots.r] (https://github.com/Kotkot/techinteractions/blob/trial/R/Codes_plots.r) is one utility function that allows to do the above.

File description

If one want to have a better understanding of what each file does, please refer to [this document] (https://github.com/Kotkot/techinteractions/tree/trial/doc/File-description.Rmd)