BirdVox-full-season: 6672 hours of audio from migratory birds

BirdVox-full-season: 6672 hours of audio from migratory birds ================================================ Version 1.1, May 2022. The full-season dataset contains 6671 hours of audio from the Fall 2015 migration season in Tompkins County, NY. Created By --------------- Andrew Farnsworth (1), Benjamin Mark Van Doren (1), Steve Kelling (1), Vincent Lostanlen (2), Justin Salamon (3), Aurora Cramer (4), Juan Pablo Bello (4) (1): Cornell Lab of Ornithology (CLO) (2): Laboratoire des Sciences du Numérique de Nantes (LS2N), CNRS (3): Adobe Research (4): New York University https://wp.nyu.edu/birdvox Data acquisition -------------------- In 2015, we placed nine bioacoustic sensors - Recording and Observing Bird Identification Node (ROBIN) developed by the Cornell Lab of Ornithology - in residential areas of Tompkins County, NY, USA, primarily surrounding the town of Ithaca, NY, USA. All sensors had the same hardware configurationcomprising a Knowles EK23132 microphone element, an analog-to-digital converter, a Raspberry Pi Model B single-board computer, a solid-state memory card, and a battery. The microphone element is omnidirectional and has an approximately flat sensitivity of 53±5 dB between 2 and 10 kHz; that is, the frequency range of flight calls. The microphone element sits at the bottom of a small horn-shaped enclosure oriented upwards. In turn, this enclosure sits inside a hard plastic housing, whose purpose is to reject lateral sound sources, such as insects or car engines. The analog-to-digital converter encodes the monophonic signal recorded by the microphone into a linear pulse-code modulation sequence at a sample rate of 24 kHz and a sample depth of 16 bits. This sample rate corresponds to an appropriate Nyquist bound to capture the diversity of avian flight calls, which occur almost exculsively below 11 kHz in frequency. The single-board computer streams this sequence under the form of 20-second buffers, which are progressively appended to a lossless audio file in FLAC format. This acquisition procedure is repeated every night from civil twilight dawn to dusk between August 3rd, 2015 and December 8th, 2015. This temporal period corresponds to the general pattern of the timing of nocturnal bird migration, which birds usually initiate 30-45 minutes after local sunset and cease in the hours around dawn; this is not always the case for cessation. This duty cycle corresponds to roughly 1,500 hours of audio per sensor, and thus 13,500 hours for the entire sensor network. However, due to intermittent failures of sensing hardware, a common feature of many autonomous recording platforms, we retrieved only 6,651 hours successfully. We gathered FLAC files according to their location of provenance or "unit". Note that unit09 was never deployed: hence, the unit IDs are 01, 02, 03, 04, 05, 06, 07, 08, and 10. Despite hard plastic housing and deployment locations that attempted to physically facilitate avoidance of non-target signal capture, these sensors captured audio of nocturally migrating birds as well as additional features of this soundscape including human activities (anthrophony), meteorological phenomena (e.g. geophony) and non-targeted avian and other biological signals (biophony). These latter signals include non-human mammals (e.g. White-tailed Deer, flying squirrel, canines), diurnal vocalizations of resident and migrant birds that are not flight calls (e.g. Blue Jays, American Crow, American Goldfinch), anurans (e.g. spring peepers), and many insect species (e.g. Fork-tailed Bush-Katydid, Snowy Tree Cricket). Derivative Datasets ---------------------------- A representative subset of these audio recordings were selected for annotation. Ornithologist Andrew Farnsworth used the Raven software to pinpoint and label every avian flight call in time and frequency. He found 26138 sound events, of which 21546 are flight calls from Passeriformes. Of those, 13385 are identifiable in terms of family, and 8669 are identifiable in terms of both family and species. The annotation process took over 600 hours. This subset of recordings and the corresponding annotations have been released as BirdVox-296h (https://doi.org/10.5281/zenodo.4415480). The isolated flight calls and annotations have been released as BirdVox-14SD (https://doi.org/10.5281/zenodo.3667093) and its follow-up release BirdVox-25SD (https://doi.org/10.5281/zenodo.5889214). Feedback ------------- Please help us improve BirdVox-full-night by sending your feedback to: vincent.lostanlen@ls2n.fr and af27@cornell.edu In case of a problem, please include as many details as possible. Acknowledgement -------------------------- We thank the following people for their contributions to the development, construction, maintenance, deployment, and acquisition of the ROBIN units: Jessie Barry, Ian Davies, Tom Fredericks, Jeff Gerbracht, Sara Keen, Holger Klinck, Anne Klingensmith, Ray Mack, Peter Marchetto, Ed Moore, Matt Robbins, Ken Rosenberg, and Chris Tessaglia-Hymes, Chris Wood. Initial data collection activities were supported by NSF 1125098, Wolf Creek Foundation, and Leon Levy Foundation; analyses were supported largely by NSF 1633206 as well as Leon Levy Foundation and NSF 1661329. Cornell University is located on the traditional homelands of the Gayogo̱hó꞉nǫ' (Guy-yo-KO-no) (the Cayuga Nation). The Gayogo̱hó꞉nǫ' are members of the Haudenosaunee (Ho-di-no-so-ni) Confederacy, an alliance of six sovereign Nations with a historic and contemporary presence on this land. The Confederacy precedes the establishment of Cornell University, New York state, and the United States of America. We acknowledge the painful history of Gayogo̱hó꞉nǫ' dispossession, and honor the ongoing connection of Gayogo̱hó꞉nǫ' people, past and present, to these lands and waters. We acknowledge that the land on which the data was collected is the unceded territory of the Cayuga nation