Machine learning may solve the mystery of bird movements

Many bird species migrate large distances across the globe and this movement is integral to their behavior, survival and reproduction. It is important to understand where they go, and when, so that effective conservation actions can be implemented, but it is not possible to track millions of birds as they travel from and to locations all over the world. Migratory birds rely on a predictable series of seasonally and regionally varying resources and are thus susceptible to the global changes in climate and weather patterns that are becoming more common. 

Although the movement patterns of individual birds can be deduced from the recapture or recovery of ringed or tagged birds, these techniques are labor intensive and cannot be applied to large numbers of birds. In addition, the weight and shape of tracking devices make them unsuitable for most birds to carry, and they are often too costly to be used on more than a handful of individuals. 

However, the world is full of avid birdwatchers who enjoy monitoring and reporting the presence of migratory birds. For example, citizen scientists contribute more than 200 million annual bird sightings through eBird, a project managed by the Cornell Lab of Ornithology and international partners. Datasets such as this provide enormous quantities of data on bird presence on a weekly basis, but do not track individuals birds to the end points of their migrations. 

In collaboration with computer scientists at the University of Massachusetts Amherst, the Cornell team has now developed a new predictive model using these two different types of data on the movements of migratory birds. In their paper, published in the journal Methods in Ecology and Evolution, the scientists state that their model, called BirdFlow, is capable of forecasting accurately where a migratory bird will go next, one of the most difficult tasks in biology. 

“Humans have been trying to figure out bird migration for a really long time,” said study senior author Professor Dan Sheldon, who is a passionate amateur birder. “But, it’s incredibly difficult to get precise, real-time information on which birds are where, let alone where, exactly, they are going,” noted study last author Miguel Fuentes.

There have been many efforts, both previous and ongoing, to tag and track individual birds and, although data from this approach has produced important insights, it cannot give a complete picture that can be used to predict bird movements. 

“It’s really hard to understand how an entire species moves across the continent with tracking approaches, because they tell you the routes that some birds caught in specific locations followed, but not how birds in completely different locations might move,” said Professor Sheldon.

The new model combines data from tagged and ringed birds with the millions of sightings made by bird enthusiasts around the world through eBird. This is one of the largest biodiversity-related science projects in existence. The database comprises over one billion global bird observations, provided by hundreds of thousands of users. It facilitates state-of-the-art species distribution modeling through the Lab’s eBird Status & Trends project. 

“eBird data is amazing because it shows where birds of a given species are every week across their entire range, but it doesn’t track individuals, so we need to infer what routes individual birds follow to best explain the species-level patterns,” explained Professor Sheldon.

The BirdFlow model draws on eBird’s Status & Trends database and its estimates of relative bird abundance, and then runs that information through a probabilistic machine-learning model. This model is tuned with real-time GPS and satellite tracking data so that it can “learn” to predict where individual birds will move next as they migrate. The model is still being perfected, but should be available to scientists within the year, and will eventually also be accessible to the general public.

The researchers tested BirdFlow on 11 species of North American birds – including the American woodcock, wood thrush and Swainson’s hawk – and found that not only did BirdFlow outperform other models for tracking bird migration, bit it accurately predicted migration flows without the real-time GPS and satellite tracking data, which makes BirdFlow a valuable tool for tracking species that may literally fly under the radar.

The new model allows scientists to monitor bird movements and the ways in which these respond to global changes in climate. In North America alone, an estimated three billion birds have been lost in the last half-century, representing nearly a third of the continent’s avifauna. According to the researchers, the BirdFlow framework has the potential to “boost insights gained from direct tracking studies and serve a number of applied functions in conservation, disease surveillance, aviation and public outreach.”

“Birds today are experiencing rapid environmental change, and many species are declining,” said study co-author Benjamin Van Doren. “Using BirdFlow, we can unite different data sources and paint a more complete picture of bird movements, with exciting applications for guiding conservation action.”

—

By Alison Bosman, Earth.com Staff Writer

Check us out on EarthSnap, a free app brought to you by Eric Ralls and Earth.com.