Spatial capture-recapture for acoustic surveys

Spatial capture-recapture (SCR) is one method to estimate the animal density of a wildlife population, where animal density describes the number of animals per unit area within the survey region. Since SCR’s inception, it has seen increasing use in estimating animal density from data collected by non-invasive sampling methods such as passive acoustics, camera trapping, and genetic sampling. In this thesis, I focus on SCR methods for passive acoustic surveys, where animals are observed by sampling their calls using audio devices like microphones. Passive acoustic surveys are suitable for wildlife populations that are physically or visually elusive but are known to produce vocalisations actively. Wildlife populations that meet these criteria include birds, cetaceans, and primates. One caveat is that animal vocalisations are what an acoustic survey observes, and we cannot necessarily attribute the vocalisations to individuals. Therefore, standard acoustic SCR models estimate call density, the number of calls per unit area per unit time, instead of animal density. Call density can be a suitable proxy of animal density for vocally active wildlife populations, and an independently estimated call rate estimate can convert call density to animal density. With this thesis, I aim to expand the repertoire of SCR literature for passive acoustic survey data. It begins with investigating whether SCR’s density estimator is robust to misspecification of the detection function for typical acoustic survey designs. The results establish that SCR’s density estimator is sensitive to the choice of detection function when applied to acoustic survey data. It then assesses the performance of SCR models when applied to acoustic surveys of cetaceans as SCR oversimplifies the detection process for underwater sound propagation. The applications involve data collected by sonobuoys and hydrophones. In light of the oversimplification of the detection process, the inference made with SCR was comparable to other models and previous studies. Lastly, a recent SCR model has been developed to estimate animal density from a single acoustic survey without requiring an independent call rate estimate. However, the current model assumes that animals remain stationary throughout the survey. In this thesis, I propose a new SCR model that allows animals to call from different locations within their territories.

Name thesis:degree_name

PhD

Level thesis:degree_level

Doctoral

Discipline thesis:degree_discipline

Statistics

Grantor dc:publisher

ResearchSpace@Auckland

Year dc:date.issued

2022

Author dc:creator

• Chan, David Ka En

Advisors dc:contributor.advisor

• Stevenson, Ben C.
• Fewster, Rachel M.

Statement dc:rights

• Items in ResearchSpace are protected by copyright, with all rights reserved, unless otherwise indicated.

Licence dc:rights.uri

• https://researchspace.auckland.ac.nz/docs/uoa-docs/rights.htm

Handle dc:identifier.uri

https://hdl.handle.net/2292/64188

OAI identifier oai:identifier

oai:researchspace.auckland.ac.nz:2292/64188