Airborne point counts: a method for estimating songbird abundance with drones

Abstract Using drones to conduct airborne bioacoustic surveys is a potentially useful new way to estimate the abundance of vocal bird species. Here we show that by using two recording devices suspended from a quadcopter drone it is possible to estimate distances to birds with precision. In an experimental test, the mean error of our estimated distances to a broadcast song across 11 points between 0 and 100 m away was just 3.47 m. In field tests we compared 1-minute airborne counts with 5-minute terrestrial counts at 34 count locations. We found that the airborne counts yielded similar data to the terrestrial point counts for most of the 10 the songbirds included in our analysis, and that the effective detection radii were also similar. However, airborne counts significantly under-detected the Northern Cardinal (χ 2 9 = 22.8, post-hoc test P = 0.007), which we attribute to a behavioral response to the drone. Airborne counts work best for species that vocalize close to the ground and have high frequency-range songs. Under those circumstances, airborne bioacoustics could have several advantages over ground-based surveys, including increased precision, increased repeatability, and easier access in difficult terrain. Further, we show that it is possible to do rapid surveys using airborne techniques, which could lead to the development of much more efficient survey protocols than are possible using traditional survey techniques. Lay Summary We show that it is possible to estimate the distance of singing birds from a drone, which then allows bird counts to be converted to true abundance or population densities. Using drones to count birds allows researchers to survey areas that may be difficult or dangerous to access on foot. Airborne counts are potentially a highly efficient and highly repeatable way to estimate populations of vocal bird species.