Harnessing the power of citizen-science data to study the spatiotemporal patterns of birdsong

Abstract

Recently the availability of hand-held recording devices has driven the immense growth of citizen-science repositories. While these repositories have the benefit of providing a well-distributed dataset, they also bring the challenge of analyzing recordings collected using devices of various quality, during different weather conditions, and with inconsistent methods. In order to take advantage of such resources, the field needs new methods to analyze recordings of numerous species at various recording qualities. Henceforth, we discuss our twofold approach: to satisfy this need for new software and to demonstrate the value of using citizen-science data to study birdsong. We developed Chipper, an open-source software to semi-automate syllable segmentation and streamline the measurement and analysis of birdsongs from various species and of varied quality. We validate the workflow of using Chipper and citizen-science data to find subtle patterns in song previously revealed by fieldwork on a “little brown bird” of North America—the chipping sparrow. We then conduct analyzes to uncover previously undetected long-range cultural structure in the chipping sparrow’s simple song. Specifically, we found significant regional song differences between Eastern and Western U.S./Canada. Interestingly, song variation was not correlated with mitochondrial genetic variation across this range. Taken together, our results suggest that there are two culturally distinct sub-populations of chipping sparrows that are genetically indistinguishable. Lastly, we explore the cultural evolution of the chipping sparrow song with particular interest in the distribution and evolution of syllable shapes. By combining genes, songs, and computational models, our analyses provide new insights into the life history of a seasonally migratory species.