Data from: Global marine flyways identified for long-distance migrating seabirds from tracking data
Data files
Mar 28, 2025 version files 7.47 MB
-
anonymised_sample_data_IO.csv
6.61 MB
-
Polygons.7z
854.38 KB
-
README.md
3.90 KB
Abstract
Aim: To identify the broad-scale oceanic migration routes (‘marine flyways’) used by multiple pelagic, long-distance migratory seabirds based on a global compilation of tracking data.
Location: Global
Time period: 1989 – 2023
Major taxa studied: Seabirds (Families: Phaethontidae, Hydrobatidae, Diomedeidae, Procellariidae, Laridae and Stercorariidae)
Methods: We collated a comprehensive global tracking dataset that included the migratory routes of 48 pelagic and long-distance migrating seabird species across the Atlantic, Indian, Pacific and Southern Oceans. We grouped individuals that followed similar routes, independent of species or timings of migration, using a dynamic time warping clustering approach. We visualized the routes of each cluster using a line density analysis and used knowledge of seabird spatial ecology to combine the clusters to identify the broad-scale flyways followed by most pelagic migratory seabirds tracked to-date at an ocean-basin scale.
Results: Six marine flyways were identified across the world’s oceans: the Atlantic Ocean Flyway, North Indian Ocean Flyway, East Indian Ocean Flyway, West Pacific Ocean Flyway, Pacific Ocean Flyway, and Southern Ocean Flyway. Generally, the flyways were used bi-directionally, and individuals either followed sections of a flyway, a complete flyway, or their movements linked to two or more flyways. Transhemispheric figure-of-eight routes in the Atlantic and Pacific oceans, and a circumnavigation flyway in the Southern Ocean correspond with major wind-driven ocean currents.
Main conclusions: The marine flyways identified demonstrate that pelagic seabirds have similar and repeatable migration routes across ocean-basin scales. Our study highlights the need to account for connectivity in seabird conservation and provides a framework for international cooperation.
https://doi.org/10.5061/dryad.59zw3r2jc
Description of the data and file structure
Files shared
Since data used in the analysis are accessible upon request in the Seabird Tracking Database or directly to researchers, these can not be openly shared. To ensure the reproducibility of the workflow, a processed, filtered, and anonymized dataset for one ocean basin (Indian Ocean) is provided with the codes to carry out the Dynamic-Time Warp analysis that is then used in ArcGIS. Intermediate outputs, including polygons of the clusters that form the final flyway in each ocean basin and the scripts to reproduce all the figures in the manuscript, are also provided along with a schematic description of all the steps performed during the original analysis (SF2_Data processing schematic.pdf).
Files and variables
File: anonymised_sample_data_IO.csv
Description: processed, filtered, and anonymized dataset of seabird tracking for the Indian Ocean
Variables
- dataset_id: categorical variable (A, B, C, D or E)
- scientific_name: categorical variable (Sp1, Sp2 or Sp3)
- common_name: categorical variable (Sp1, Sp2 or Sp3)
- site_name: NA (categorical variable when downloaded from the STDB - removed the identifying information from anonymised dataset here)
- colony_name: NA (categorical variable when downloaded from the STDB - removed the identifying information from anonymised dataset here)
- lat_colony: NA (decimal degrees of the colony latitude when downloaded from the STDB - removed the identifying information from anonymised dataset here)
- lon_colony: NA (decimal degrees of the colony longitude when downloaded from the STDB - removed the identifying information from anonymised dataset here)
- device: categorical variable of the logging device when known
- bird_id: id of the bird
- track_id: id of the track
- original_track_id:
- age: age of the bird (only adults for this processed dataset)
- sex: sex of the bird
- breed_stage: breeding stage when known
- breed_status: breeding status when known
- date_gmt: date in gmt
- time_gmt: time in gmt
- latitude: latitude of track in decimal degrees
- longitude: longitude of track in decimal degrees
- argos_quality: (only for PTT data) quality of the position
- equinox: (only for GLS) whether the position correspond to an equinox period or not (yes or no)
File: Polygons.7z
Description: 7z file containing 6 gpkg files of polygons
- Marine_Flyways: boundaries of all the marine flyways
- Atlantic_Ocean_clusters.gpkg: clusters for the Atlantic Ocean Flyway
- Indian_Ocean_clusters.gpkg: clusters for the Indian Ocean Flyways (North and East Indian Ocean Flyways)
- Pacific_Ocean_clusters.gpkg: clusters for the Pacific Ocean Flyways (Pacific and West Pacific Ocean Flyways)
- Southern_Ocean_clusters.gpkg: clusters for the Southern Ocean Flyway
- Oceans.gpkg: (Uploaded to Zenodo) basemap ocean layer [adapted from: Flanders Marine Institute (2021). Global Oceans and Seas, version 1. Available online at https://www.marineregions.org/. https://doi.org/10.14284/542]
Code/software
All of the files and scripts provided can be run in R using the following packages:
- library(dplyr)
- library(ggplot2)
- library(gridExtra)
- library(sf)
- library(rnaturalearth)
- library(dtwclust)
-
library(amt)
- library(tidyverse)
- library(ggspatial)
- library(metR)
- library(cowplot)
Part of the analysis includes using the line density tool in ArcGIS. Since this is not an open software, we shared the resulting polygons in the form of .gpkg files for all the Marine Flyways and for the clusters in each ocean basin as described before in the “Files and variables”.
We collated and standardised seabird tracking data for pelagic, long-distance migratory species according to previously developed protocols (Carneiro et al., 2020). We assigned the annual life cycle stage to each location, and only those recorded during migration were used in our analyses. We processed and filtered tracking data to exclude individuals that did not meet a set of criteria. Individuals were then grouped based on the shape of the migratory route, irrespective of species identity or the timing of migration. We estimated line densities for each cluster of individuals and the results were smoothed and combined where appropriate to delineate a network of marine flyways.