Do street lights affect flight behavior of nocturnally migrating birds?

Methods

Summary

The Chester River Field Research Station in Chestertown, MD, is a banding station collaborating with our project. According to satellite imagery, it is sited in one of the darkest areas of the Mid-Atlantic region. They allowed us to mount three portable LED downcast floodlights (24,000 lumens each) on top of 25-feet tall wooden poles. These lights illuminate a portion of their mist-netting array from sunset to sunrise during a 3-day on and 5-day off sampling scheme.  That is our experimental site.

We have deployed (Spring and Fall 2016) and will deploy (Spring and Fall 2017) a tracking radar approximately 500 meters away from the experimental site. This tool can track individual targets aloft (birds, insects, and bats) for up to 3km, collecting one set of polar coordinates every second detailing the position of the target with respect to the radar (distance, orientation and altitude). These data translate into continuous 3 dimensional track of targets within the radar domain. We used this radar to sample nocturnal bird migration occurring in the surroundings of our experimental site, collecting data during nights with the lights on and off. Before sampling, we confirm that nocturnal bird migration was going on by consulting the network of weather radars. During post-processing of collected tracks we filter out non-avian targets with a high degree of confidence by considering factors like their flight speeds and characteristics of flight trajectories. For future seasons, we will be able to differentiate bird, insect, and bat targets with complete certainty based on their wingbeat patterns; the appropriate software is in development.

Challenges

We hope to improve the process of data collection by implementing a software that will show the identity of targets detected by the radar (e.g. bird, bat, insect). But even if we do not get this software up and running, it is still possible to differentiate target types by considering their flight speeds and characteristics of the flight trajectory; this is how we have filtered our data from 2016.

Pre Analysis Plan

With our data we will compare behavioral patterns between nights with lights-on and lights-off, such as attraction or avoidance to the experimentally lit site, which we will identify due to changes in 1) flight altitude, for instance do birds descend towards the lights and ascend away from them? Are their flights completely level when the lights are off?; 2) changes flight trajectory, e.g. heading of flight trajectories shifting towards or away the site when the lights are on vs. completely linear trajectories when the lights are off; and 3) changes in flight speeds, for example slowing down when flying close and above the lit site and accelerating again when flying away from the site when the lights are on, vs. no changes in speed when the lights are off.