Reconstructing Velocities of Migrating Birds from Weather Radar - A Case Study in Computational Sustainability

SUMMER 2014 31 Ecological processes that occur at global scales are among the most impressive natural phenomena on Earth; they link environments across the globe and capture the imagination of observers. For example, billions of North American birds, representing more than 400 species, migrate annually, often thousands of miles, frequently under the cover of darkness, between breeding and nonbreeding areas. There is an urgent need to understand global processes in order to mitigate threats from anthropogenic change. For example, human-made collision hazards are killing hundreds of millions of birds annually (Crawford and Engstrom 2001; Longcore, Rich, and Gauthreaux 2008), and direct and indirect habitat alteration is reducing stopover habitats vital for birds to refuel during migration (Bonter, Donovan, and Brooks 2009; Veltri and Klem 2005). Our ability to mitigate this anthropogenic mortality relies on a detailed understanding of the magnitude and composition of migration in time and space so we can take actions such as protecting critical stopover sites and turning off wind-turbines and lights on man-made structures at key places and times during migration. However, because of their scale and complexity, global phenomena are also very difficult to study. Bird migration is