S6-5 Tue Jan 5 14:00 – 02:30 Sexual selection, natural selection, and artificial intelligence: Implementing technological advances to understand variation in signaling behavior Symes, LB*; Madhusudhana, S; Martinson, SJ; Kernan, CE; ter Hofstede, HM; Center for Conservation Bioacoustics, Cornell Lab of Ornithology, Cornell University; Center for Conservation Bioacoustics, Cornell Lab of Ornithology, Cornell University; Center for Conservation Bioacoustics, Cornell Lab of Ornithology, Cornell University, Dartmouth College; Dartmouth College; Dartmouth College symes@cornell.edu
One of the most fundamental features of communication is when and how often a signal is repeated. Historically, limitations on observation time and monitoring technology have made it difficult to quantify and compare differences in signal activity across a community of species to generate and test hypotheses about how often signals are produced (signaling rate) and the daily duration of activity (signaling window). Advances in monitoring equipment and data processing are opening new avenues for research on signaling investment. Here, we assess potential drivers of signaling rate and signaling window using two complementary types of data: 1) automatic detections (using machine learning) of calls in soundscape recordings and 2) 24 hour recordings of focal individuals.
We focus on a community of Neotropical forest katydids on Barro Colorado Island in Panama. This community contains at least 80 species that vary dramatically in signaling rate and signaling window. Some species produce more than 10,000 calls per day. Other species produce less than 30 calls/day with each call lasting less than 50 ms, generating a total of less than two seconds of sound per day. Some species signal 24 hours a day, while other species signal during a window of a few hours. Using machine learning and focal recording data, we assess correlations and patterns in signaling investment, signal design, morphology, anti-predator defenses, and seasonality in this community.
