Contributions of environmental conditions and sound characteristics to differences in perceptibility: Recommendations for passive acoustic monitoring
Data files
Sep 11, 2025 version files 3.70 MB
-
habitat.csv
3.70 MB
-
README.md
2.03 KB
Abstract
Passive acoustic monitoring is increasingly used in ecological research to study a wide variety of taxa; however, properly accounting for perceptibility, a component of detection probability, has been challenging because sound attenuation is influenced by numerous variables in the surrounding environment, and the rate of sound attenuation is not typically quantified. We used sound playback experiments to investigate the effect of several environmental variables, including vegetation type and weather, on the perceptibility of a variety of sounds and species. We also investigated species-specific variables such as sound frequency, syllable rate, and bandwidth. We quantified the amount of bias in perceptibility resulting from species sound characteristics and vegetation type, and estimated perceptibility and the area surveyed for different combinations of species and habitat type. We found that distance was the strongest predictor of perceptibility, but that vegetation type significantly influenced perceptibility, particularly when comparing open versus closed vegetation classes. Sound frequency and bandwidth of species were also important predictors of perceptibility. We found that ignoring variation in perceptibility due to excess sound attenuation from vegetation could bias area surveyed during acoustic surveys by up to 400% (mean = 167.4%). We conclude that other than distance, perceptibility is influenced most by species-specific vocalization traits and by the habitat these vocalizations transmit through. We provide recommendations for incorporating these findings into future research to account for varying perceptibility and improve the accuracy of bird monitoring and research.
Dataset DOI: 10.5061/dryad.2rbnzs82p
Description of the data and file structure
Experimental data to investigate factors that may influence the detection probability of birds and other species. We broadcast vocalizations of species commonly found in the Canadian boreal forest and re-recorded them at different distances in a variety of vegetation types and under different environmental conditions. Audio recorders were listened to by observers to determine if sounds were identifiable.
Files and variables
File: habitat.csv
Description: All detection data used for ORNITH-APP-24-173
Variables
- Species: species or sound that was broadcast
- Observer: observer ID, to account for observer variability
- ID: identification of a sound or species that was broadcast by an observer listening to a recording of the sound
- Distance: distance between the sound broadcast and the audio recorder
- Comments: notes about that specific broadcast or site
- Transect: the id of the transect where the broadcast occurred
- Habitat: vegetation type that was present during the broadcast
- Wind: wind speed in km/h
- Temp: temperature in degrees Celsius
- Humidity: humidity as a percentage
- Detected: was the species or sound correctly identified
- Duration: duration of the sound or species signal, in seconds
- Syllables: the number of syllables present in a species or sound broadcast
- Srate: the rate at which syllables occurred, calculated as the number of syllables per second
- MinFreq: the lower bound of a sound or species vocalization spectrogram in Hz
- MaxFreq: the upper bound of a sound or species vocalization spectrogram in Hz
- Bandwidth: MaxFreq minus MinFreq
Missing data code: NA
Code/software
This is a .csv file which can be viewed in Excel or imported and manipulated in R or other analytical software.
