Association of Field Ornithologists

While phenotypes in a population are constantly undergoing minor changes as individuals have varied success from year to year, major shifts in population phenotypes can happen far more rapidly under extreme conditions. I wondered, could a similar phenomenon occur in a culturally transmitted phenotype such as song? Specifically, if a stressor was effectively applied across an entire breeding songbird population simultaneously, such that virtually all juveniles experienced impaired learning, would there be a detectable perturbation in population song in the next breeding season? If these juveniles’ songs became the models for future generations, could such a punctuated stressor ultimately induce a lasting change in a population’s—or a species’—song?

Since we knew that detecting potential changes in songs across a population would require many recordings over an extended period, we identified two main ways to test this hypothesis: with a captive population or a natural experiment. For the latter, an environmental stressor would need to occur in an area with ample prior song recordings, either from a long-term field site or public recordings from community scientists.

We didn’t have access to a captive population or data from a long-term field site, but our lab was starting to work with repositories of field recordings like Macaulay Library and Xeno-canto, so I looked for regions with an extraordinarily high density of recordings. In Macaulay Library in 2016, there was one area that stood out – the region around Tompkins County, New York, the home of the Cornell Lab of Ornithology. The next step was figuring out if an extreme environmental stressor had ever occurred there, and whether it was recent enough to have enough recordings before and after the event to detect possible changes in song features. It was a moment of surreal serendipity when a simple Google search for “New York drought” came back with multiple news articles only a few months old that revealed that a large area of New York state, including Tompkins County, had just that summer experienced a drought that was unprecedented in recorded history for the region.

We preferentially considered species for study that (a) had a large number of publicly available song recordings, (b) had some evidence of site fidelity, and (c) have diets that would likely be impacted by drought. One potential challenge is that severe environmental stress often results in members of a population simply abandoning reproductive behaviors. In the context of this study, this would mean that there would be fewer juveniles developing during the height of the drought, and thus a minimal effect on the represented songs the following years. However, the timing of the drought might have mitigated this challenge: since the drought did not begin until mid-summer, many species would likely have already bred and raised chicks to at least fledgling age, increasing the likelihood that enough juveniles were exposed to drought stress to significantly affect population song.

The species that best fit were the dark-eyed junco (with relatively simple songs) and the song sparrow (with more complex songs). We requested the available recordings from the region from the previous decade and I (and my dog, Moki) spent a week in the summer of 2017 making targeted recordings in central-western New York with a homemade parabolic recorder to bolster our sample size in the year following the drought. We enlisted the help of an undergraduate researcher, Maria Sellers (now pursuing a PhD in biological anthropology), who meticulously quality-checked over 750 individual recordings, extracted bouts of song, parsed the individual syllables using our lab’s software, “Chipper,” and learned to analyze the data in R over the course of her undergraduate career.

This study was also a lesson about both the importance of community science data and the challenges of acquiring experimental and control data from a repository after an environmental stressor has occurred. Our main challenges were that there were relatively few recordings from before the 2016 drought that we could use to compare to post-drought recordings, and there were very few regions outside of the extreme drought zone where there were enough recordings spanning the same time period to act as a control. However, we are optimistic about the future in this regard; the number of recordings deposited to publicly available repositories seems to have skyrocketed in the past several years, with better and better audio recording technology finding its way into our pockets via smartphones. Future studies should thus be much better equipped to further probe the link between stressful climatic events and cultural evolution in songbirds.