Previous efforts to utilize remote sensing technologies to detect birds, have been able to use AI tools to automate the interpretation of radar scans to isolate bird densities. Using a probabilistic model and an approximate inference algorithm, researchers can identify densities and velocities of migrating birds, which are not only key to understanding broad migration patterns, but also to differentiate birds from other atmospheric densities like precipitation, insects, and dust (Farnsworth et al., 2014).
While a fair amount of research has been conducted both to describe and to analyze artificial light at night, many studies have historically maintained only a broad geospatial scale when considering ALAN. Recently, Falchi et al. describe the creation of a worldwide map of artificial sky luminance that considers global regions subjected to different categories of artificial light. The study notes the importance of low-angle, upward-pointing light sources as a primary cause of light pollution with a significant impact on night luminance (Falchi et al., 2016).
More quantitative approaches to evaluate artificial light emissions at night have been performed in an urban context by Dobler et al., who utilized ALAN emissions from buildings in the New York City skyline to measure aggregate human behavior based on the identification of on/off light transitions. Brightness from each building was calculated after identifying unique buildings in the scene by employing aperture windows across a stacked image. Patterns of light intensity and sudden shifts in average brightness of a building could be discerned with these methods. It was found that while individual building lighting behavior is somewhat random on a nightly basis, the "pulse" of the urban environment is much more periodic in its aggregate behavior (Dobler et al., 2015). The goal of this work is to determine if a similar pulse is present in bird counts over time and if that is geospatially and temporally correlated to the lighting variability.