Ambient noise is an integral component of natural environments, but it also creates challenges for avian monitoring programs. Ambient noise can mask bird vocalizations from observers during point counts, which may lead to systematic undercounting of birds in noisy environments. Here, we estimate detection probability and population size in models that either account for or omit the influence of ambient noise. We used data for four bird species, from 2228 point counts that were conducted during the 2019 Mountain Birdwatch field season. Community scientists assessed ambient noise using a simple scale. Despite relatively quiet conditions at sampling locations (x = 2.48), our information theoretical approach favored N-mixture models that incorporated ambient noise into the detection function for all four species. At the noisiest sampling locations, our models predicted detection probabilities that were as low as 10% for some species. Accounting for ambient noise resulted in a modest, mean increase of ≤3.29% in the total population size for each species. Following our approach, other researchers can easily incorporate ambient noise assessments into their field protocols and analyses, with minimal costs or added complexity, to increase the comparability of studies conducted within different acoustic environments.

No abstract