Bird Migration Explorer Supporters Partners Data Providers Species Migration Connections Conservation Challenges Learning Resources What's New

Species Migration Maps

Migratory birds make impressive journeys each year, often traveling thousands of miles between summer and winter ranges to complete their annual cycle. From flocks of Tundra Swans crossing the Alaska Range on their way to the Arctic, to Tree Swallows that weigh less than an ounce fighting headwinds while crossing the Gulf of Mexico, these heroic journeys bring wonder to bird enthusiasts. The Species Migration Maps show the movements of 458 species of migratory birds that regularly occur in the United States and Canada as they travel throughout the hemisphere to complete their annual cycle.

The driving force of migration is a need for resources—such as food, warmer weather, safe places to nest, or safe places to molt new feathers. The Bird Migration Explorer illustrates a variety of migration strategies used by different species, including tracking data that document the unique pathways of individual birds. Some long-distance migrants like the American Golden-Plover cross two continents, while short-distance migrants like Phainopeplas may move only counties away. Golden Eagles exhibit both migratory and non-migratory behavior—many individuals migrate thousands of miles while others reside locally year-round. Some species like the Whooping Crane migrate in spring and fall along the same route, while others like Brant use a looping strategy—birds head north across mountainous inland terrain and return by following the coast before making a direct flight across open ocean. Brown-capped Rosy-Finches are altitudinal migrants that concentrate on the tops of mountains in summer then expand into the lowlands in winter. Many waterfowl, like the King Eider, migrate first to a safe molting location before continuing the journey weeks later to their wintering grounds. Migration can be fast or slow; Blackpoll Warblers complete migration in a few days of non-stop flight while Pectoral Sandpipers spend a few months traveling to their destination.

Technologies for Tracking Birds

Over decades tracking devices have become lighter, smaller, and longer-lasting, allowing migration researchers to their expand efforts. Currently, about half of the migratory species in the US and Canada have been tracked. The amount of tracking data varies among species, from a single tracked individual to hundreds, or from a few weeks of data to many years. Larger birds such as waterfowl, cranes, pelicans, and raptors tend to be better-tracked because they can bear heavier devices that transmit frequent, high-accuracy locations. These larger devices have been in use since the 1980s. Smaller-sized birds, including songbirds such as warblers and thrushes, require lightweight tracking devices that tend to have more locational error and require recapturing the bird to download the data. These smaller devices have only been available for the last two decades, and with each year the devices become increasingly smaller, making it possible to collect data on smaller birds.

Global positioning system (GPS) satellite telemetry. GPS units use satellites to triangulate their location. Locations can be transmitted or stored. Most GPS units transmit their locations in near-real-time using satellites or cellular towers; the units are relatively heavy and suitable for larger birds. Archival units store data onboard and are among the smallest tracking devices available; these collect a very limited number of locations and require that the bird be recaptured and the unit recovered to retrieve the stored data. GPS data have the highest accuracy—up to 30 m, fine enough to track a bird to a specific backyard.

Doppler satellite telemetry. Often called PTT devices, Platform Transmitting Terminals use passing satellites to estimate their location using the Doppler Effect. These units transmit their locations to Argos, a wildlife tracking system. Doppler data have good accuracy—often around 1 km, which is good enough to track a bird to an area of town.

Light-level geolocator. These tiny devices can be used on a wide variety of species. They record light levels throughout the day. By using these light data to estimate sunrise and sunset times, researchers generate daily locations of migratory birds. Geolocators are too small to transmit data, so the data are stored on the unit and the bird must be recaptured and the tag recovered to retrieve the data. Because daylength is equal around the globe during equinox periods, geolocators cannot estimate locations for several weeks during spring and fall. Compared to other tracking devices, geolocators have poor accuracy—error can sometimes exceed 200 km, which allows researchers to track a bird to the state or country level. Despite these limitations, geolocator data have been incredibly valuable for describing migration routes for many small species.

Automated radio telemetry. Radio telemetry uses unique radio frequencies to locate and identify a radio-tagged individual. Radio telemetry can be automated through an established array of receiving towers that automatically record radio signals when a bird passes by. The Motus wildlife tracking system is the world’s largest collaborative automated radio telemetry array, including over 750 receiving stations across the Western Hemisphere. An advantage of radio tags is their relatively low cost and very small size—Motus tags are small enough to deploy on any species of bird.

Other methods of studying bird movement between locations include banding and genetic markers. Read about Connections to learn more about collecting re-encounter data.

Interpreting the Species Migration Maps

The Species Migration Maps can be used to understand the movements of individual birds and the entire population as they migrate between seasonal ranges.

Species Range by Season

The ranges shown represent generalized areas where each species regularly occurs in their selected habitat. Species often occur inside or outside the mapped summer and winter areas at other times of the year, especially during migration. Migration ranges are not explicitly mapped and are instead conveyed via the pathway and weekly abundance datasets.

Genetic Populations

The genetic population layers, or "genoscapes", represent the locations of distinct genetic groups across the breeding range of a species. The boundaries of these regions are uncertain and are visualized using a blur effect in the Bird Migration Explorer maps. The genoscapes were determined by the Bird Genoscape Project. Learn more about this research at https://www.birdgenoscape.org/making-genoscapes/.

Individually Tracked Birds

These maps pull together current knowledge of species movement patterns, but much is yet unknown about bird migration. Pathway data on the Homepage/Migration Journeys Map and the Species Migration Maps are from numerous studies shared with Audubon, however these maps do not include all tracking studies. Individual tracks do not represent complete migratory pathways for species and about half of the included species have yet to be tracked by researchers.

Precision/spatial error in recorded tracking locations varies widely. Our data processing smoothed the recorded locations provided in the original pathway datasets, resulting in an accurate but less precise visual representation of migration pathways. Researchers interested in using these data for scientific purposes should refer back to the original studies. Throughout the year, geolocator locations tend to be more accurate in longitude than latitude; as a result, some maps appear to have tracks that “shake” as they show north-south movements that do not truly occur. These shaky pathways may also show unrealistic movements over the ocean or over the land that are not consistent with the natural history of a species.

Weekly Abundance/Occurrence

The eBird Status abundance models are generated by using eBird observations to predict bird distribution across their range. For each species/week, we grouped hexagons into five ranked distribution classes where hexagons were sorted from highest to lowest abundance, then split into five classes that each sum to 20% of the total abundance. The resulting classes (very high [top 20%], high, moderate, low, very low [bottom 20%]) show the contribution of each hexagon to the total population of the species. These models occasionally have spatial and temporal gaps in areas or weeks when observation data are too scarce to generate accurate predictions. The extent of the prediction area varies by week, and data gaps tend to affect areas that are most sparsely populated (e.g., the Canadian far north, the upper Amazon basin).

For species where eBird Status models were not available, eBird Basic Dataset observations were used. These tend to be biased toward areas where there are more observers, such as in more populated areas and along roads. These data were not corrected for spatial bias. Species mapped using the eBird Basic Dataset were Black-headed Gull, Black-legged Kittiwake, Black Rail, Bridled Tern, Long-eared Owl, Mexican Whip-poor-will, Pomarine Jaeger, Red-legged Kittiwake, Red Phalarope, Sabine's Gull, Sooty Tern, White-winged Scoter, Yellow-billed Loon, and Yellow Rail.

Mapping Methods

The Species Migration maps are made possible by the generous contribution of data and expertise from the research community and our partners. Tracking data that show individual pathways were contributed by hundreds of researchers.

Species Range by Season

Summer, winter, and year-round ranges indicate generalized areas where each species regularly occurs during each season. We produced a hemispheric range dataset using a combination of sources. Data were primarily derived from eBird Status 2018 range maps, where available. For areas outside the eBird prediction area, we merged in range data from Audubon Alaska’s 2017 Ecological Atlas of the Bering, Chukchi, and Beaufort Seas and BirdLife International’s 2020 global range dataset on a seasonal basis. These additional data filled in spatial and/or seasonal gaps in the eBird Status maps. Areas where summer and winter ranges overlapped were considered year-round range. The composite ranges were internally reviewed by experts in bird distribution and compared to the Audubon digital Bird Guide maps, eBird species occurrence maps, eBird Status 2020 range datasets, Birds of the World, and the BirdLife International Data Zone. Necessary manual adjustments were made to improve accuracy where known issues could not be resolved through geoprocessing.

Genetic Populations

A genoscape is a map of genetic variation across the breeding range of a species. Locations of genetic population clusters were determined by the Bird Genoscape Project. This research involves identifying unique combinations of DNA base pairs from feather samples, which are predictors of geographically separated populations. For display in the Bird Migration Explorer maps, genetic cluster rasters from the Bird Genoscape Project were converted to vector data, clipped to the species breeding range, and smoothed. The resulting polygons were visualized using a blur effect to indicate uncertainty in the boundaries of these regions.

Individually Tracked Birds

To appropriately summarize and map individual migration pathways, the technology used to track the bird and its associated locational error were taken into account. We processed data to flag and remove points with high locational error (outliers), such as points collected during the spring or fall equinox period for geolocator data, and consecutive points that exceeded speed, distance, angle, or other thresholds.

For each individual bird, the remaining point locations were summarized into a single average point for each day and smoothed using a ±1 day running average. For birds that were tracked over multiple years, each year of data was treated as a separate track. The smoothed points were converted to pathways, with the exception of consecutive points two or more weeks apart or more than 1,000 km apart (though we removed this maximum distance filter for several species upon review). The resulting lines were smoothed with a tolerance of 100 km.

Weekly Abundance/Occurrence

We used weekly distribution data from eBird Status abundance models developed by the Cornell Lab of Ornithology to derive 150-km and 50-km hexagon abundance layers. We summarized total abundance of each species within each hexagon, for each week of the year, then grouped hexagons into five classes that each sum to 20% of the total abundance.

For a small number of species that did not have an eBird Status model available, observations from the eBird Basic Dataset were used instead. For each hexagon/week, we compared the frequency of observations of a species (i.e., the count of occurrences regardless of group size) to the frequency of observations of all species recorded in eBird, then binned these proportions into five frequency classes using an equal interval classifier.

Species Conservation Statistics

Climate Vulnerability

Climate Vulnerability is derived from a published analysis by the National Audubon Society. This analysis compiled more than 140 million observations to develop species distribution models and assess climate change vulnerability for 604 North American bird species at a continental scale in both breeding and non-breeding seasons and under a climate scenario of 3°C warming. Based on projected range loss and potential range gain for each species, climate vulnerability was categorized as neutral, low, moderate, or high. Learn more about climate vulnerability here.

Global Population Estimate

The global population estimate is derived from the Partners in Flight Avian Conservation Assessment Database, managed by Bird Conservancy of the Rockies. This database contains the most current and complete information on the biological vulnerability of North American bird species at various scales, and identifies important species and regions for conservation action. This database was developed with data from dozens of large-scale, long-term bird population surveys as well as input and review from dozens of ornithological experts from ten countries.

The global population size is an estimate of the number of breeding age individuals that occur across the entire range of a species. Learn more about the Avian Conservation Assessment Database here.

Continental Conservation Status

Continental Conservation Status is derived from the Partners in Flight Avian Conservation Assessment Database, managed by Bird Conservancy of the Rockies. This database contains the most current and complete information on the biological vulnerability of North American bird species at various scales, and identifies important species and regions for conservation action. This database was developed with data from dozens of large-scale, long-term bird population surveys as well as input from dozens of ornithological experts from ten countries.

Based on quantitative criteria, the continental (US/Canada) conservation status is assigned to species that are designated as: Highly Vulnerable and in Urgent Need of Special Attention (Red List), Range Restricted and Small Populations in Need of Constant Care (Yellow Watch List R), Steep Declines and Major Threats (Yellow Watch List D), or Common Birds in Steep Decline. Learn more about the Avian Conservation Assessment Database here.

Global IUCN Red List Category

Global IUCN Red List Category is derived from BirdLife International's assessments of extinction risk for the International Union for Conservation of Nature (IUCN) Red List of Threatened Species, which is the world's most comprehensive inventory of the extinction risk of species.

Based on quantitative criteria, the global extinction risk of each species has been assigned to one of the following levels: Critically Endangered, Endangered, Vulnerable, Near Threatened, or Least Concern. Learn more about the IUCN Red List here.

Take Action