-
Notifications
You must be signed in to change notification settings - Fork 4
Home
Since the 1980s, fisheries economists have employed spatial models in a variety of fisheries to better understand and explain the factors that influence the spatial behavior and fishery participation choices that fishers make when fishing. This is important for predicting how fishers may respond to, for example, marine protection areas (MPAs), climate-related species range shifts, changes in fishing costs or fish prices, fish size differences, or the implementation of various management actions such as catch share policies.
The Spatial Economics Toolbox for Fisheries (FishSET) is a set of statistical programming and data management tools developed to achieve the following goals:
- Standardize data management and organization.
- Provide easily accessible tools to enable location choice models to provide input to the management of key fisheries.
- Organize statistical code so that predictions of fisher behavior developed by the field’s leading innovators can be incorporated and transparent to all users.
Modeling fisher behavior is important for designing effective fishery management policy. Failing to account for fisher behavior can lead to unanticipated consequences or the failure of the management policy (Abbott and Haynie 2012, Hilborn 2007, Peterson 2000). Discrete choice random utility maximization (RUM) models, which utilize explanatory variables at the level of individual vessels to predict aggregate effort levels in alternative fishing locations, have become more widely used in fisheries (Eales and Wilen 1986, Raphael 2017, Depalle, Thebaud, and Sanchirico 2020). These models assume that fishers choose, from a discrete set of fishing locations, a location to maximize expected utility, which is modeled as a function of expected revenue, distance from the fisher’s current location, other information about the vessel or fisher, and/or other information about the potential fishing locations (e.g. depth, temperature, spatial quota limits). Commonly used RUM modeling techniques have been conditional and multinomial logit models, nested logit models, and the mixed logit model (Train 2002). Many model variants and formulations have been developed by economists in fisheries and other fields, with innovations motivated by the policy question being analyzed, the nature and extent of available data, and the statistical background and computing power of the analysts which has increased greatly with time.
FishSET is intended for all users, regardless of experience with R or any other programming language.
FishSET includes an easy-to-use graphical user interface that guides users through the steps from loading data to evaluation models without writing any code.
For users with experience in R who need more flexibility than the graphical user interface, functions can be run directly in the R console.
Regardless of whether FishSET functions are run in the graphical user interface or the R console, users must have R and, optimally, RStudio installed.
Instructions to install R and RStudio and the FishSET package (a collection of tools or functions) are provided in this manual along with details of using the package.
The intent of FishSET is to create tools which improve both the quality of management analyses and the ease of conducting analyses. The FishSET R Package provides a wide range of tools, including functions for:
- Data management
- Data analysis
- Visualizing data
- Statistical modeling of fisher behavior
- Policy comparison
- Reproducibility
Good data management can save time over the duration of a project, prevent errors, increase quality of analyses, and allow for validation and replication of findings. In FishSET, all data files are automatically saved to a single location, a SQLite database called the FishSET database. Changes to the data file are saved in the database as a new table which is then updated; the original loaded data is never modified.
FishSET provides functions, grouped into steps, which guide users through important steps in checking and addressing data quality, understanding the variance, distribution, and availability of data through data exploration, and the steps to generate derived variables (e.g., zonal averages, variable interactions, or calculated variables such as CPUE) that may be needed in analyses.
Creating figures from fisheries data is a critical part of examining and summarizing data. FishSET provides numerous tools to view the spatial distribution of vessels, hauls, or other variables of interest, including hotspot analysis and haul paths, as well as to disaggregate data by time, location, fishery, vessel, or other characteristic.
FishSET includes functions to run location choice models. Using FishSET allows users with knowledge of the models to use location choice models developed by different researchers. FishSET enables comparisons of numerous model specifications, model assumptions, and/or model types.
After models are developed for a fishery, FishSET allows comparisons of actual or simulated fisheries policies, such as the creation of spatial closures.
Function calls in FishSET are logged and saved in a dated file within the FishSET R package directory. The log function calls save the function name, the parameter values, including input data, and output. Function messages, such as the presence of missing values, are also stored. Storing messages provides a record of the data that informs choices, such as removing rows from the data that contain missing values. This means that all steps in the analysis are automatically saved and the analysis can be quickly and easily reproduced at any time. It also allows for the analysis to be rerun with updated data.