Passive Acoustic Monitoring Power Analysis: A Tool for Designing an Acoustic Monitoring Program
Springer Nature Reference
DOI: Springer Nature Living Reference 10.1007/978-3-031-10417-6_140-1
ABSTRACT Monitoring trends of marine mammal populations are notoriously challenging, as density and abundance estimates are frequently characterized by high levels of uncertainty. Statistical power is described as correctly detecting a trend when one is present and is influenced by the uncertainty surrounding the input parameters. The “AVADECAF” tool evaluates the power of a fixed Passive Acoustic Monitoring (PAM) array to detect changes in animal density, via simulation. The simulations can be based on existing surveys or can be used to design monitoring programs. The tool allows for density estimation via distance sampling where ranging of acoustic detections is possible, or Spatial Capture Recapture (SCR), when an animal’s sounds can be detected on more than one of the acoustic sensors. The tool provides an opportunity to evaluate how the power to detect population declines might be improved by comparing alternate survey designs. Improving statistical power increases the probability of detecting a change in the population and can reduce the required duration of a monitoring program. The AVADECAF tool was applied to simulated Blainville’s beaked whale clicks detected by hydrophones at the US Navy’s Atlantic Undersea Test and Evaluation Center (AUTEC), and the potential future applications of this tool are discussed.
KEYWORDS
Cite this article as: Ryder, M., Booth, C., Oedekoven, C., Marques, T., Joy, R., Harris, D. (2023). Passive Acoustic Monitoring Power Analysis: A Tool for Designing an Acoustic Monitoring Program. In: Popper, A.N., Sisneros, J., Hawkins, A.D., Thomsen, F. (eds) The Effects of Noise on Aquatic Life. Springer, Cham. https://doi.org/10.1007/978-3-031-10417-6_140-1
Using a sequence deep learning model to increase the acoustic context of a killer whale detector
The Journal of the Acoustical Society of America
DOI: J. Acoust. Soc. Am. 155, A87 (2024) 10.1121/10.0026903
ABSTRACT Although Killer whales (Orcinus orca) produce many stereotypical vocalizations, their sounds can be difficult to identify in isolation. Experts often rely on acoustic context to accurately identify these animals acoustically. Automated detectors and classifiers, on the other hand, frequently rely on short clips that capture individual vocalizations, not leveraging information regarding previous sounds. We developed deep learning models that used 1-minute inputs containing from 0 to 50 calls, with the average clip having 18. We tested three artificial neural network architectures that used recurrent layers to take the sequence of acoustic events into account. As a baseline, we used a convolutional neural network that only took 3-s clips at a time, without considering sequences of events. Here, we present preliminary evaluations on a dataset containing 360 min with Southern Resident killer whale activity in the Salish Sea, and an equal amount of data without killer whale sounds. The best model used a combination of temporal convolutional layers and gated recurrent units to achieve a recall of 0.95 at the maximum precision of 0.98. The models will be applied to near real-time monitoring efforts and will be open-sourced in the future.
KEYWORDS
Cite this article as: Fabio Frazao, Oliver S. Kirsebom, April Houweling, Jennifer Wladichuk, Jasper Kanes, Ruth Joy, Mike Dowd; Using a sequence deep learning model to increase the acoustic context of a killer whale detector. J. Acoust. Soc. Am. 1 March 2024; 155 (3_Supplement): A87. https://doi.org/10.1121/10.0026903
Functional data analysis to describe and classify southern resident killer whale calls
Ecological Informatics
DOI: 10.1016/j.ecoinf.2024.102841
ABSTRACT The Southern Resident Killer Whale (SRKW) is an endangered population of whales found in the northeast Pacific. They have a vocal dialect unique from other killer whales, having a repertoire of distinct stereotyped calls. A framework for distinguishing SRKW call types using the frequency traces of the amplitude ridges from their spectrograms (termed frequency ridges) is proposed. The first step is the extraction of these ridges of SRKW calls using an Sequential Monte Carlo approach. Next, they are converted into functional data using B-spline functions. They are analyzed with a functional principal component (FPC) analysis to characterise the intrinsic variability of frequency ridges within a call type. The FPCs are able to capture the general patterns in the frequency ridges of the different SRKW call types. The FPCs are also used as the basis for call classification. Using a cross-validation procedure to assess the robustness of the classification, this framework proves to be successful for classification with some call types having an F1-score , but other calls less well discriminated. On balance, this approach showed reasonable performance given the small sample size available, and provides a useful contribution towards the development of a universal tool for call classification.
KEYWORDS Marine bioacoustics, Functional data analysis, Frequency ridges, Sound classification
Cite this article as: Paul Nguyen Hong Duc, David A. Campbell, Michael Dowd, Ruth Joy, Functional data analysis to describe and classify southern resident killer whale calls, Ecological Informatics, Volume 83, 2024, 102841, ISSN 1574-9541, https://doi.org/10.1016/j.ecoinf.2024.102841.
Assessing the efficacy of ecological reserves: killer whale beach rubbing behaviour and vessel disturbance
Endangered Species Research
DOI: ESR 53:555-567 (2024) 10.3354/esr01317
ABSTRACT: Area-based protection is an important tool for safeguarding key habitat. Reserves that focus on mitigation of specific threats are particularly effective and are more likely to support a measurable outcome. In the marine environment, reserves that limit vessel presence have the potential to reduce disturbance to marine mammals. However, assessing the efficacy of reserves has been an ongoing challenge. Physical and acoustic disturbance from vessels is recognized as a primary threat to recovery for the northern resident killer whale (NRKW) population in Canadian Pacific waters. The Robson Bight Michael Bigg Ecological Reserve (RBMBER) was developed to support the behaviour of beach rubbing, a culturally distinct and traditionally important activity. Beach rubbing provides a rare opportunity to quantify vessel disturbance of a behaviour associated with a fixed geographic location, identifiable by visual cues, and verifiable acoustically. Observations on vessel presence, NRKW rubbing frequency, and duration were collected from a beach inside the reserve and compared to a beach in proximity to, but outside of, the RBMBER. In 2019-2022, vessel counts near the RBMBER beach were significantly lower than near the unprotected beach, and overall, rubbing occurred more frequently inside the reserve (78% of visits) than outside (35%). However, outside the reserve, concurrent vessel presence did not predict the occurrence of rubbing activity, indicating that vessel presence may negatively affect beach rubbing through long-term learned avoidance of frequently impacted areas.
KEY WORDS: Marine reserve efficacy · Vessel impacts · Killer whales · Orcinus orca · Area-based management · Rubbing beach
Cite this article as: Konrad Clarke CM, Stredulinsky E, Toews S, Rani V and others (2024) Assessing the efficacy of ecological reserves: killer whale beach rubbing behaviour and vessel disturbance. Endang Species Res 53:555-567. https://doi.org/10.3354/esr01317
Forecasting the trajectories of Southern Resident Killer Whales with stochastic continuous-time movement models
Department of Statistics and Actuarial Science
The Southern Resident Killer Whale (SRKW) is an endangered population of killer whales that is present in the Salish Sea. This fish-eating predator has been heavily impacted by human activities in the region, particularly by commercial vessels in shipping lanes that traverse federally-designated SRKW critical habitat. Forecasting the movement trajectories of these whales would help provide early warning alerts to slow down or reroute commercial vessels, and reduce the risks of ships overlapping with whale presence. In this study, we develop a stochastic animal movement model that is guided by a historical database of sighting records of SRKW. Specifically, we make use of a continuous-time Ornstein-Uhlenbeck (O- U) velocity process that provides the basis for a movement forecast system and simulates realizations of SRKW velocities and trajectories given initial conditions. However, if the forecast system were to simply rely on the O-U velocity process alone, it would steer simulated whale trajectories to areas where SRKWs are rarely found. To address this, we propose a direction blending scheme to project the simulated velocities in more realistic directions. It makes use of historical directional information along with the O-U velocity process to create more probable pathways consistent with observed SRKW movement patterns. By integrating the simulated trajectories generated from the simulated velocities, we establish a dynamic probability-based forecast scheme that demonstrates skill in forecasting SRKW trajectories on a time-scale that aligns with the time to slow and reroute commercial vessels.