Detecting small changes in populations at landscape scales: a bioacoustic site-occupancy framework

Abstract Occupancy modeling based on detection/non-detection data has become a common approach for monitoring changes in the populations of both sensitive and invasive species, with emerging bioacoustic technology enhancing opportunities for implementing such programs at landscape-scales. Statistical power, however, to detect small but biologically meaningful changes in site occupancy as part of landscape-scale monitoring is typically low, with large – yet hereto unknown – sampling efforts likely required for rigorous inference. Therefore, we (i) assessed sampling levels and detection probabilities needed to detect small changes in site occupancy driven by both intrinsic trends and management effects, and (ii) evaluated the feasibility of using bioacoustics to simultaneously monitor a common but declining species and a rare but increasing invasive competitor within a site occupancy framework. Simulation-based power analyses indicated that detection/non-detection data collected at large numbers of sites (500–1500) can yield high statistical power (>80%) to detect ≥2% annual declines in site occupancy within 10 years, but depended on the number of visits per site, initial occupancy rates, and detection probabilities. Statistical power to detect ≥30% declines in local survival rates in 10 years was also high. Based on ∼6-night passive-acoustic surveys, site occupancy and detection probabilities were 0.43 and 0.50, respectively, for the common but declining species (the spotted owl), and 0.09 and 0.67, respectively, for the rare but increasing competitor (the barred owl). Simulations parameterized with these empirically-derived rates indicated that 2% annual declines in spotted owl site occupancy could be detected with high statistical power in 10 years with 1,000 sites surveyed three times per season (year) or 1500 sites surveyed two times per season. Statistical power to detect 4% annual increases in site occupancy for expanding barred owl populations with this sampling scheme was also high. Thus, our study yielded the novel finding that passive-acoustic monitoring can be used to detect small but potentially biologically meaningful changes in site occupancy for multiple species with very different population dynamics with high confidence. More broadly, as computational improvements bring acoustic-based whole-community identification into the realm of possibility, our approach will allow managers to rapidly assess the statistical power attainable for each species: systematic and statistically robust monitoring of entire faunal communities within a unified framework at a landscape scale may become a reality.