Sightings from Bird Watchers to Aid in Study of Avian Flu

Cornell scientists tap into Avian Knowledge Network

By Lauren Cahoon and Miyoko Chu

Using millions of bird records from citizen-science participants and biologists, researchers have created new animated maps showing the flow of bird migration across the United States and Canada. A research team at Cornell University will be using these maps in computer models, along with data on the locations of poultry centers, to understand how avian flu might spread. Although the highly pathogenic H5N1 avian flu virus still hasn't been found in North America, the team hopes the model will be useful in predicting what might happen if it ever arrives.

The virus spreads primarily through infected poultry, but because wild birds can carry the virus too, modeling the potential spread of avian flu requires data on the migratory movements of hundreds of bird species. The researchers harvested about 3 million records of 147 bird species through the online Avian Knowledge Network, including data from citizen-science participants of eBird and from researchers for the USGS Arctic Shorebird Survey.

"As far as we know, this is the first time migratory information has been used in an epidemiological model," said Dexter Kozen, a professor in Cornell's Computer Science department who is leading the effort. When completed, the model will incorporate the patterns of wild bird migrations in North America, the location and density of domestic poultry, and the rate and pattern with which the virus passes from host to host.

An animated map shows the distribution of diving ducks as it changes throughout the year. The colored points represent bird sightings. Click to view the animation (7.8 Mbytes ).

Kozen combined data from multiple years to illustrate the relative abundance and movements of birds during a typical year. The model shows a map of North America with bird sightings appearing as tiny squares. The colors range from white to red to indicate the abundance of birds. As the seasons progress from winter to spring on the animated map, a wave of these tiny squares flows up across the map from South America toward Canada, showing how birds are moving across the continent.

The project is a compelling example of how millions of bird records, accessible through the Avian Knowledge Network, can be mined for new information at a moment's notice. Steve Kelling, director of Information Science at the Cornell Lab of Ornithology, said, "The ability to gather information on birds at broad geographical scales is only possible through citizen networks such as eBird."

Janis Dickinson, the Lab's director of Citizen Science, added, "One of the strengths of eBird is its flexibility, because it allows us to get large numbers of people gathering data anywhere, any time, over a broad geographic scale. This flexibility has proven crucial to acquiring unprecedented numbers of observations of birds over all the seasons—and this is exactly what is needed to understand patterns of bird migration."

Lesser Scaup

Photo by Irma Eisthen, GBBC 2006

Data from the Arctic Shorebird Survey were also available through the Avian Knowledge Network. Those data played a key role in showing the northern range of migratory birds, since few birders live in the Arctic.

Although researchers will be incorporating patterns of bird migration into the computer model, it's unclear whether wild birds will play an important role in transmitting avian flu to poultry or humans. The spread of the highly pathogenic H5N1 virus has been mostly affected by outbreaks in domestic poultry and transportation of infected birds by the poultry industry. To create a realistic model of how avian flu might spread, researchers must include information on poultry distributions and the rate of transmission of the virus among domesticated and wild birds.

Researchers from Cornell's College of Veterinary Medicine, Center for Applied Mathematics, Department of Computer Science, and Cornell Lab of Ornithology are combining their expertise to create this new model that will help make sense of the many factors that could influence the spread of this multifaceted disease.

To view some of the new maps of bird migration, visit www.birds.cornell.edu/birdflu/maps.

For permission to reprint all or part of this article, please contact Laura Erickson, editor, Cornell Lab of Ornithology, 159 Sapsucker Woods Rd., Ithaca, NY, 14850. Phone: (607) 254-1114. email: lle24@cornell.edu