Factors associated with automated detection of Northern Spotted Owl (Strix occidentalis caurina) four-note location calls

Hane, Matthew E.; Thornton-Frost, Jamie E.; Springford, Aaron; Kroll, Andrew J.

doi:10.5751/ace-02105-170126

Cited by 1 publication

(2 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Here, low quality refers to recordings with a low signal‐to‐noise ratio, which can result from owls vocalizing far away from the ARU, low call volume, high background noise, or topographic interference. There is a distance threshold at which calls become unidentifiable (Hane et al, 2022; Maegawa et al, 2021). Nevertheless, the Dale et al (2022) model performed well for calls that were clear enough to generate predictions and offers a valuable step that can be integrated into a workflow for processing acoustic recordings.…”

Section: Discussionmentioning

confidence: 99%

“…This covariate was represented by an index (0–1) of how similar a given cell (30 × 30 m) was to other areas with known spotted owl NR sites based on tree species composition and structural factors. For the survey‐level covariate on detection probability, we used the mean suitability value in cells within a 200‐m buffer of each station to approximate a likely listening distance of the ARUs (Hane et al, 2022; Maegawa et al, 2021; Wood et al, 2021) while minimizing the amount of overlap between station buffers (Appendix S2: Figure S1). For the hexagon‐level covariate on occupancy, we used the mean suitability value within a 500‐m buffer around each ARU station, unionized to create a single footprint per hexagon (Appendix S2: Figure S1).…”

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Using passive acoustic monitoring to estimate northern spotted owl landscape use and pair occupancy

et al. 2023

View full text Add to dashboard Cite

Managing forests for biodiversity conservation while maintaining economic output is a major challenge globally and requires accurate and timely monitoring of imperiled species. In the Pacific Northwest, USA, forest management is heavily influenced by the status of northern spotted owls (Strix occidentalis caurina), which have been in continued population decline for the past four decades. The monitoring program for northern spotted owls is transitioning from mark-resight surveys to a passive acoustic framework, requiring development of alternative analysis approaches. To maintain relevance for conservation and management, these analyses must accurately track underlying population changes, identify responses to disturbance, and estimate occupancy of owl pairs. We randomly selected and surveyed 5-km 2 hexagons for 6 weeks using passive acoustic monitoring in the Olympic Peninsula of Washington and the Oregon Coast Range during the 2018 spotted owl breeding season. We used a convolutional neural network to identify spotted owl calls, followed by logistic regression to determine the sex of vocalizing owls to assign pair status. We implemented multistate occupancy models to estimate probabilities of detection, species-level landscape use, and pair occupancy of spotted owls.We also quantified detections of barred owls (Strix varia), a congeneric competitor and important driver of spotted owl population declines. The overall rate of hexagon use by spotted owls was estimated at 0.21 (SD 0.04) after adjusting for imperfect detection, and pair occupancy was 0.07 (SD 0.02). The probability of detecting a pair (i.e., both female and male) during a weekly occasion was relatively low (0.03, SD 0.01), indicating that true pair occupancy was between 1.3 and 4.1 times greater than the proportion of hexagons with observed pair detections. Barred owls were ubiquitous, with a naïve occupancy rate of 0.97. The intensity of calling by barred owls had a weak, negative effect on the probability of spotted owls being paired when present but had little measurable effect on their detectability. This work establishes a framework that may be effective for spotted owl population monitoring and illustrates that

show abstract

Section: Discussionmentioning

confidence: 99%