An experimental evaluation of the performance of acoustic recording systems for estimating avian species richness and abundance

Abstract

Comparisons between field observers and acoustic recording systems have shown great promise for sampling birds using acoustics methods. Comparisons provide information about the performance of recording systems and field observers but do not provide a robust validation of their true sampling performance—i.e., precision and accuracy relative to known population size and richness. We used a 35-speaker bird song simulation system to experimentally test the accuracy and precision of two stereo (Telinga and SS1) and one quadraphonic recording system (SRS) for estimating species richness, abundance, and total abundance (across all species) of vocalizing birds. We simulated 25 bird communities under natural field conditions by placing speakers in a wooded area at 4–119 m from the center of the survey at differing heights and orientations. We assigned recordings randomly to one of eight skilled observers. We found a significant difference among microphones in their ability to accurately estimate richness (p = 0.0019) and total bird abundance (p = < 0.0001). Our study demonstrates that acoustic recording systems can potentially estimate bird abundance and species richness accurately; however, their performance is likely to vary by its technical characteristics (recording pattern, microphone arrangement, etc.).