Radar and AI Help Dim Skyscrapers for Migrating Birds

Every spring and autumn, billions of songbirds migrate at night, facing the deadly hazard of brightly lit city skyscrapers. To prevent fatal collisions, avian conservationists now combine weather surveillance radar with artificial intelligence that analyzes nocturnal bird calls. This article explores how scientists measure radar reflectivity to track massive flocks, use acoustic AI to identify specific species in flight, and combine this data to trigger targeted "Lights Out" alerts that convince urban building managers to dim their lights during peak migration nights.

Tracking Flocks with Weather Radar

Meteorologists use a network of weather radars, known as NEXRAD, to track rain and storms. However, these radars are so sensitive that they also detect biological targets in the atmosphere, particularly birds.

Conservationists use this radar data to monitor nocturnal migration in real time. When millions of songbirds take flight after sunset, they appear on the radar as massive, expanding blooms. Scientists measure "radar reflectivity"—the amount of transmitted radar energy that bounces off the birds and returns to the sensor. High reflectivity values, combined with Doppler velocity data that separates slow-moving bugs from faster-flying birds, tell researchers the exact density, altitude, and direction of migrating flocks. By coupling this radar data with weather forecasts, scientists can predict precisely when massive waves of birds will pass over major metropolitan areas.

Identifying Species with Acoustic AI

While radar reveals the size and location of a flock, it cannot identify which species are flying overhead. Because different birds respond differently to artificial light, knowing the exact species is critical.

Many songbirds, such as warblers, thrushes, and sparrows, emit short, high-pitched vocalizations while flying in the dark. These "nocturnal flight calls" help flocks stay together and navigate. Conservationists place networks of specialized microphones on city rooftops and in rural flyways to record the night sky.

In the past, humans had to listen to thousands of hours of audio to identify these calls. Today, researchers use acoustic artificial intelligence. Machine learning models are trained on massive libraries of bird vocalizations. The AI converts the recorded audio into visual spectrograms—graphs of sound frequencies over time—and uses computer vision to instantly identify the exact species flying above. This process filters out urban noise like sirens and wind, providing a real-time census of the specific birds entering a city's airspace.

Triggering "Lights Out" Alerts

Artificial light at night disorients migrating birds, blinding them and drawing them away from their natural navigation cues, like the stars. Once lured into dense urban centers, exhausted birds often collide with reflective glass windows or circle buildings until they fall from fatigue.

By merging the macro-level data from weather radar with the micro-level species data from acoustic AI, conservation groups like the BirdCast project can issue highly accurate migration alerts. When radar reflectivity shows millions of birds approaching a city and acoustic AI confirms the presence of species highly vulnerable to light pollution, automated systems trigger peak migration warnings.

These alerts are sent directly to city governments, skyscraper managers, and homeowners. Instead of asking cities to keep lights off for entire seasons, this precise data allows for targeted conservation. Building operators are asked to turn off non-essential exterior lighting, dim lobby lights, and close blinds during the specific a few nights when migration traffic is at its absolute highest. This data-driven approach saves energy for building owners and creates safe, dark corridors that allow millions of songbirds to survive their journey.