During peak passage, when air temperatures warm in early May, more than 500 million migratory birds take flight each night, a new study finds. The report presents a new way to forecast continental-scale bird migrations using weather forecasting radar systems, which also happen to detect flocks of passing birds. Each year, as billions of birds navigate the skies on content-spanning migrations, hundreds of millions die along the way due to collisions with human-made structures and vehicles, like wind turbines and airplanes. The ability to predict the timing and movement of migrating birds could help to greatly reduce this mortality. While previous research has shown that environmental drivers, such as winds, temperature, atmospheric pressure and precipitation play important roles in determining when large groups of birds take flight, the complex interactions between these conditions and the diverse behaviors over many species have made forecasting bird migrations a challenge. Using observations from the continent-wide metrological Next-Generation Radar (NEXRAD) network, Benjamin Van Doren and Kyle Horton created a forecasting model that captures the migratory patterns of birds with high spatial accuracy. Using the model, which combines 23 years of Doppler radar observations of spring nocturnal migration with atmospheric condition data, Van Doren and Horton were able to explain up to 81% of the variation in migration timing and intensity across the U.S. and forecast movements as far as a week, as well as estimate the total numbers of birds making the journey. Since bird migration is characterized by pulses of marked travel separated with periods of low activity, the ability to forecast when they occur could inform decisions to take temporary actions to prevent human and avian mortality, re-routing air traffic, reducing light pollution or slowing down wind-turbines, the authors suggest.