Identification of Bird Migration Events in NEXRAD Data

Listen to the Podcast:  Radar is for the Birds

Click on images to see and hear video of snow geese at Freezout Lake, Montana.

Click on image to view what snow geese look like on NEXRAD.

Management agencies often do not have adequate data to delineate time periods having the greatest number of birds migrating, nor the quantitative data to map the critical migration areas and landscape corridors. Such information is especially important to the management of shorebirds, waterfowl, and other wetland birds. Such historic information can be used for habitat acquisition programs, design of agricultural conservation programs, evaluating energy development projects, and designing inventory and monitoring programs. Real time migration knowledge could be used in such emerging issues as the possible tweaking of wind turbine operation and the understanding the potential spread of diseases such as avian influenza.

When NEXRAD weather radar data is collected, it not only contains information about weather events, but it also captures the presence of birds in the air column. A few expert ornithologists can visually examine such data and make determinations about the presence or absence of birds. However, because the NEXRAD data archive is so voluminous, it is virtually impossible for humans to examine all the data collected since the late 80s. Machine learning methods, however, are capable of such data mining activity. Our goal at NOROCK is to conduct research that allows us to distinguish between precipitation and birds in NEXRAD weather radar data leading to the development of a landscape model of bird migration that incorporates spatial and temporal components. We are currently developing an algorithm based on artificial intelligence methodologies that is quite effective at sorting through existing NEXRAD data and identifying birds in the data. Eventually we would like to build algorithms that can efficiently process real time NEXRAD data and provide associated decision support about ongoing bird migration to resource managers.

Collaborators within the field of artificial intelligence include Dr. John Paxton and graduate student, Reggie Mead, from Montana State University. Additional collaborators include the Department of Biological Sciences at University of Southern Mississippi and the National Renewable Energy Laboratory (in association with the U.S. Department of Energy). To follow the progress of our project, see http://crane.cs.montana.edu/nexrad.

Scientists from around the USGS have formed a collaborative to research in application of radar ornithology, similar to that being done at the Northern Rocky Mountain Science Center. Links to work around the USGS can be found at: http://www.fort.usgs.gov/Radar/.

In a collaborative project with Ducks Unlimited, there are “goose chasers” travelling throughout the Midwest who locate flocks of birds and record their precise latitude and longitude, measure their height above ground level using a digital rangefinder, and record the observation time to the nearest second. NOROCK scientists then download that specific data from NOAA’s National Climate Data Center, knowing that it, in fact, has geese in it. Erv Klaas, a retired leader of the USGS’ Iowa Cooperative Fish and Wildlife Research Unit, has been working tirelessly as one of these goose chasers.

Return to Rick Sojda's Research Projects Return to Birds of the Northern Rockies Return to Science