Scientists develop new method to track individual bird species migrating at night
Researchers have combined eBird observations, weather radar and tracking data to move migration monitoring beyond counting birds in the sky and towards identifying which species are on the move.
New research reported by Phys.org could mark a major step forward in the way bird migration is monitored, by helping scientists estimate not only how many birds are moving, but which species are involved.
Weather radar has transformed the study of nocturnal migration by revealing when and where large numbers of birds are travelling. Until now, however, one of its biggest limitations has been that radar can detect movement but cannot identify individual species.
Researchers at the Cornell Lab of Ornithology, the University of Massachusetts and the University of Illinois have developed new methods that combine radar data with participatory science records and tracking information. The work draws on more than two billion bird observations submitted to eBird, alongside data from GPS tags, Motus radio telemetry and bird banding.
The approach is part of BirdFlow, a project using artificial intelligence models to predict the movement of bird populations during migration. One new method, known as BirdFlow Migration Traffic Rate, provides weekly, species-specific estimates of migration across North America.
The research team says the model can identify major flyway patterns and fill gaps in areas where weather radar coverage is limited. It also allows scientists to estimate the most likely species behind the movements detected by radar.
Adriaan Dokter, project leader for BirdCast and BirdFlow at the Cornell Lab, said the work “opens up exciting new directions for monitoring and forecasting bird migration in real time”. He said the new metrics make it possible to estimate the species most likely responsible for movements recorded by radar, which can detect numbers of birds aloft but not their identity.
A second approach used BirdFlow models to bring together data from individually tracked birds, including GPS, Motus radio telemetry and banding information. The researchers produced migration models for 153 North American migratory bird species, spanning 14 orders and 39 families.
To test the methods, the team compared BirdFlow estimates with 28 years of data from 152 weather surveillance radars across North America. They also compared the models with real GPS-tracked birds, finding that the predicted routes were biologically realistic.
Dokter said incorporating individual and species-specific differences, as captured by tracking and banding data, “greatly improves” population-level movement models. He described BirdFlow as a way of bringing together different sources of information on the movements of individual species.
The potential conservation uses are wide-ranging. More detailed knowledge of when particular species are moving could help reduce collision risks with buildings and other hazards during peak migration periods. Yuting Deng, a postdoctoral associate at the Cornell Lab, noted that different bird groups face different levels of risk from window collisions.
The models could also help with disease surveillance, including efforts to understand the movement of avian influenza through waterfowl populations. The research team says BirdFlow could support migration ecology, conservation planning, aviation risk assessment and public outreach.
The work could eventually be integrated with existing tools such as BirdCast, which already provides migration forecasts but does not currently offer species-specific information. If the approach can be expanded further, it could give conservationists a much clearer picture of migration across whole species ranges and throughout the full annual cycle.
June 2026
Get Breaking Birdnews First
Get all the latest breaking bird news as it happens, download BirdAlertPRO for a 30-day free trial. No payment details required and get exclusive first-time subscriber offers.
Share this story
