A gift and a grant totaling $3.45 million will help the Cornell Lab of Ornithology develop new computer technologies to better understand the movements and behaviors of birds and other species.

A $1.25 million gift from the Kenneth L. Harder Trust will be used by the lab's Acoustic Monitoring Project (AMP) to develop a new, faster system of automated acoustic data analysis for the bevy of natural sounds that the lab collects globally. Such recordings hold massive amounts of data that must be analyzed -- an incredibly time-consuming endeavor -- to uncover relevant sounds for study.

The Harder gift will be used to develop algorithms to "train" computers to identify specific sounds -- such as those of nocturnal-migrating birds, forest elephants in the wild or long-distance whale moans -- by their acoustic parameters.

"Acoustic monitoring generates terabytes of data to manage and analyze," says Paul Allen, AMP project leader. "The infrastructure and tools developed and integrated by AMP will expand our capacity for monitoring, which is increasingly in demand, and extend our ability to automatically classify what species created the sound."

The gift comes courtesy of Cornell trustee Scott Harder, Karen (Tillman) Harder '81, Liv Harder '11 and Donald Harder. Kenneth Harder, Scott's uncle, was a lifelong birder, and the Harders are longtime supporters of the Cornell Lab of Ornithology.

Also, the Lab of Ornithology and Oregon State University received a $2.2 million National Science Foundation grant to fund BirdCast, a project focused on forecasting bird migrations.

In the spring and fall, many millions of neotropical migrants stream across North America flying as far north as the boreal forests of the Yukon in Alaska from March through May, and as far as South America between July and November. Since many of these birds fly at night when they can't be seen, BirdCast will combine night flight calls from acoustic monitoring stations to distinguish individual species, weather radar station data to detect clouds of night migrating birds, and eBird data -- bird observations made by a massive network of volunteers who submit records of bird sightings to a Cornell database.

The $1.2 million that has been apportioned to Cornell will be used to combine those data with weather and terrain data to develop a new statistical model to allow researchers to predict how, when and where birds are moving through the landscape.

"We believe our models are going to be very accurate to make predictions of where a bird is flying," said Steve Kelling, the Lab of Ornithology's director of information science and a principal investigator for the NSF-funded project. "This is really a computer science project using new computational and analytic technologies to forecast the tsunami of bird migrations, which represent a huge amount of biomass moving around," he added.

Such accurate models will allow researchers to understand behavioral aspects of migration, such as how migration timings and pathways respond to climate change. Also, information on migratory behavior will aid decisions for where wind turbines are placed, and when building lights should be turned off at night, as birds are confused by and attracted to such lights and collide with buildings. Novel Web-based data visualizations of migration paths will also have strong educational and outreach applications.

The collaboration with Oregon State University stems from work by Carla Gomes, Cornell professor of computing and information science, who has worked to develop computer technologies to solve issues in the environment, material sciences and biology.