Preamble

Course objectives

This course is an introduction to using Gadget as an ecosystem simulator and stock assessment tool. It is introduced using Rgadget, an R library that simplifies and standardizes the procedure for creating the input model files needed for creating a Gadget model, as well as gather and visualize ouput files created by Gadget. Although input files can be created manually, Rgadget facilitates reproducability and aids internal consistency of Gadget model creation by allowing Gadget models to be created, run, analyzed, and shared using a single set of R scripts. Please note, however, that both Gadget and Rgadget are continuously being developed, so please remember to update your installations frequently (1x / month). In addition, Rgadget is programmed using ‘tidyverse’, so a good understanding of how to program using tidyverse is a necessity.

By the end of this course you should:

  • correctly structure and format input files for building a simple Gadget model and running a simulation via R using Rgadget
  • become familiar enough with the various options in Gadget and how to implement them using the Gadget User Guide that comes with the Gadget installation, so that a more complex ecosystem can be simulated
  • input data, generate likelihood components, and modify optimisation settings
  • understand the iterative reweighting procedure implemented in Rgadget
  • become familiar with predefined figures and data frames generated by gadget.fit

Your a priori homework

Prior to the course make sure your operating system is a unix derivative, such as linux or macosx, as the code and examples in this course is only tested against these operating systems. Windows should in theory work, but issues related to Windows line endings are known to cause random issues in Gadget. For those who do not want dual boot their computers they can install VirtualBox and install linux on a virtual machine (you can find a selection of pre-install images here where it is recommended that you choose Ubuntu). However if you insist on using Windows you should install the icesTAF R-package to handle all the file conversion issues.

To install Gadget follow the installation guide.

Please also have the most recent version of R and Rstudio installed.

In addition, if you plan to run a simple model using your own data, please:

  1. familiarize yourself with tidyverse programming

  2. think about your model structure/ aggregation level and have R scripts ready that aggregate your data into an appropriately formatted data frame (i.e., columns of : year, time step, area, age bins, length bins, or age- and length-bins - have a sneak peak at our example model data here)

Time and location

Time: 2019-09-30 09:00 - 2019-10-04 12:00

Location:

Draft workplan

Monday 9:00 - 16:00

Tuesday 9:00 - 16:00

Wednesday 9:00 - 16:00

Thursday 9:00 - 16:00

Friday 9:00 - 12:00

  • Troubleshooting
    • Known issues
    • Asking for help
  • When to use Gagdet? Some examples.
    • Ling
    • Northern shrimp
    • Atlantic wolffish
  • MSE using Gadget
    • Ling
  • Wrap up