spawning-dataset

Code for the generation and processing of fish spawning regions

Organization

• code/ includes R scripts for both preparation and display.
• inputs/ contains data drawn from AquaMaps/FishBase or constructed by hand.
• outputs/ contains the shapefile dataset and other results.

Reproduction process

First, clone the repository, and set the working directory in R to the root of the repository. All R code assumes that this is the working directory.

The R code relies on a number of libraries. These should be installed before running the code: maps, dplyr, PBSmapping, ggplot2, viridis, sf, scales, RColorBrewer, rgeos, rgdal, ncdf4, raster, stringi.

To reproduce the Spawning ProCreator spreadsheet

The Spawning ProCreator spreadsheet, Master Spawning ProCreator.csv, describes how each spawning region should be constructed. Improvements to these constructions can be made by editing the spreadsheet, which does not require reproducing it. However, if the underlying spawning information from FishBase and SCRFA is extended, the spreadsheet should be recreated and new rows should be merged with the existing file.

To reproduce the spreadsheet, as it was prior to adding the information that describes how regions should be constructed, follow these steps:

1. You may optionally regenerate the mapping of EEZs to FAO regions. This is produced by prelim/fao2eez/mapping.R. After running it, move the resulting outputs/fao2eez.csv to inputs/fao2eez.csv.

2. Regenerate the input/specieseez.csv and input/speciespid.csv files if the inputs/Region FAO EEZ matching-DO NOT EDIT IN EXCEL.csv has changed. These files describes the FAO regions corresponding to multinational descriptions in the spawning dataset. To regenerate them, run prelim/fao2eez/species2eez.R, which produces output/specieseez.csv and output/speciespid.csv and move them to the inputs directory.

3. Merge the FishBase and SCRFA spawning records: Run the code/prelim/spawning-merge.R script. This generates a file outputs/spawning-records.csv which should be moved to inputs/spawning-records.csv for the next step.

4. Geocode spawning region names: Set source = 'arcgis' in code/prelim/geocode.py and run the script from the prelim directory. You will need to have the geocoder python package installed. Then set source = 'geonames' and run the script again. This script produces geocoded result files named localities-arcgis.csv and localities-geonames.csv. Move these to the inputs/ directory.

5. Run the code/prelim/spawning-geoprep.R script, which constructs the raw Spawning ProCreator spreadsheet into outputs/master.csv. The code includes logic for generating maps of the geocoded regions, for choosing between them, and saving these to an accessible dropbox folder, but dropbox.path and dropbox.url need to be provided for this to work. The resulting master.csv file can then be imported into Excel or Google Sheets for filling out the Verdict column.

6. When the Spawning ProCreator spreadsheet is prepared (the Verdict and other columns are manually entered), save the result as a CSV file at inputs/Master Spawning ProCreator.csv.

Spawning ProCreator metadata

• ID: Unique ID number assigned to each unique [Locality, Country] pair before manual verdict assignment (see verdict variable description below). During verdict assignment, if the researcher believed the location needed to be split into two entries, then the ID number was duplicated. 
• Country: Spawning country listed in Fishbase.org’s spawning dataset
• Localities: Locality listed in Fishbase.org’s spawning dataset.
• Total Catch: Indicative catch associated with this spawning region, calculated by dividing the 2014 catch for each species associated with the spawning locality by the total number of spawning localities for that species, and summing these values across all of the species associated with the given locality. Catch values from Sea Around Us.
• Total Value: Indicative value associated with this spawning region, calculated as in the Total Catch, but using landed values from Sea Around Us.
• species: All species listed by scientific name that had the same [Locality, Country] from Fishbase.org’s spawning dataset. Percentages in parentheses describe the % of catch that each species contributed to the Total Catch value.
• Geocoded: Maps produced by GeoNames Search Webservice and ArcGIS World Geocoding Service (documentation for these services can be found at https://www.geonames.org/export/web-services.html and https://developers.arcgis.com/documentation/mapping-apis-and-services/geocoding/, respectively). GeoNames were in red and ArcGIS were in green
  • Example of map: https://www.dropbox.com/home/Spawning%20ProCreator/maps?preview=map346.pdf
• Verdict: Method researcher used to geocode the spawning location. See article’s method section for more details.
• New Country: When there was no country in the description and the boundaries of a country’s EEZ would help with the accuracy of the geocoding, a country was used. An example is “Seamounts off southern part of Africa.” A seamount product was used, but South Africa’s EEZ helped to bound the seamounts. It was also used to override incorrect countries listed. When checked by a second researcher, if the original verdict was incorrect and should have been EEZ, this new decision overrode the original verdict by filling in “EEZ” here. Finally, if the description matched the native range better than any EEZ, “any” was written to indicate the use of the native range. 
• (1) Southwest Coordinate, (1) Northeast Coordinate, (2) Southwest Coordinate, (2) Northeast Coordinate, (3) Southwest Coordinate, (3) Northeast Coordinate: There were up to 3 boxes used for these manual geocoding entries described in the methods of the paper. When the researcher manually created a box using Google Maps to bound the spawning grounds, the researcher recorded the two diagonal coordinates to draw the box. Coordinates with the same (#) are pairs. 
• Latitude range: When the description included latitudes or areas that commonly use latitudes, we noted the latitude range here. We used 35°S - 35°N for the subtropics and tropics, 23.5°S - 23.5°N for the tropics, 20°S - 20°N for the equatorial region, and 30°S - 30°N for lower latitudes.
• Notes: The researcher made notes of difficult entries, primary sources when they changed the description, or reasons why an entry was dropped. These notes are not comprehensive of the full discussion of all the challenging entries. 
• Notes (4/13/23): During a last check of the data on April 13, 2023, there were some rows that were dropped in the data merging often due to punctuation differences. These notes corrected those issues. These notes were also used to replace any verdicts with a more accurate shapefile of a region using marineregions.org shapefiles. The most common shapefiles were seas, gulfs, and bays. 

To reproduce the GO-FISH shapefile

The code/generate/read.R functions translate information from the inputs/Master Spawning ProCreator.csv spreadsheet into shapefile regions.

The code/generate/publicdataset.R script produces a shapefile that includes all available spawning regions, intersected with suitability information. It produces outputs/GO-FISH.shp and outputs/GO-FISH.csv, the latter of which corresponds to the polygon attributes in the shapefile.

The code/generate/stats.R script generates spawning-species.csv which provides information about each species and spawning region provided in the spawning regions dataset; sau-species.csv which provides information about each species in the SAU dataset; and sumstats.csv which provides a summary of this information by continent and fish group.

To regenerate the public dataset, first run code/generate/publicdataset.R and then code/generate/stats.R.

To regenerate the figures

• figures/Figure1.R generates spawning maps across the whole year, by season, or across the whole year.

• figures/Figure2.R generates the bar charts of spawning fish groups by region.

License

This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.

This repository contains data extracted from public datasets, as described below:

• FishBase: The data under inputs/spawning and summarized in inputs/spawning-records.csv is derived from FishBase (CC-BY-NC 3.0).

  Froese, R. and D. Pauly. Editors. 2023. FishBase. World Wide Web electronic publication. www.fishbase.org, version (02/2023).

• Sea Around Us: The data under inputs/saudata is derived from Sea Around Us (CC-BY-NC 4.0)

  Pauly D., Zeller D., Palomares M.L.D. (Editors), 2020. Sea Around Us Concepts, Design and Data (seaaroundus.org).

• AquaMaps: The data under inputs/ranges is derived from AquaMaps (CC-BY-NC 3.0)

  Kaschner, K., Kesner-Reyes, K., Garilao, C., Segschneider, J., Rius-Barile, J. Rees, T., & Froese, R. (2019, October). AquaMaps: Predicted range maps for aquatic species. Retrieved from https://www.aquamaps.org.

• Natural Earth: The shapefiles under inputs/shapefiles/ne_10m_admin_0_countries and inputs/shapefiles/ne_50m_coastline where made by Natural Earth (public domain): https://www.naturalearthdata.com/downloads/

• SCRFA: The file inputs/scrfa.csv and the summarized dataset inputs/spawning-records.csv contains information derived from the SCRFA Aggregations Database: https://www.scrfa.org/database/

• Marine Regions: Some shapefiles in inputs/shapefiles are extracted from Marine Regions (CC-BY-4.0): https://www.marineregions.org/downloads.php

• Knolls and seamounts in the world ocean: Some shapefiles in inputs/shapefiles are extracted from Yesson et al. (2011) (CC-BY-3.0)

  Yesson, Chris; Clark, M R; Taylor, M; Rogers, A D (2011): Knolls and seamounts in the world ocean - links to shape, kml and data files. PANGAEA, https://doi.org/10.1594/PANGAEA.757563,

• Bathemetry: Some shapefiles in inputs/shapefiles 2-minute Gridded Global Relief Data (ETOPO2) v2: https://doi.org/10.7289/V5J1012Q

  NOAA National Geophysical Data Center. 2006: 2-minute Gridded Global Relief Data (ETOPO2) v2. NOAA National Centers for Environmental Information. https://doi.org/10.7289/V5J1012Q.

• FAO Major Fishing Areas: FAO regions are used to interpret international spawning regions.

  FAO 2023. FAO Major Fishing Areas. Fisheries and Aquaculture Division [online]. Rome. https://www.fao.org/fishery/en/collection/area