Reliable estimates of population abundance are fundamental to the effective management and conservation of any species. However, estimating the abundance of cetaceans, which spend the large majority of their lives concealed underwater, can be difficult. Approaches have been developed, and are being refined for using the underwater sounds of cetaceans, to estimate their numbers. Passive acoustic sensors mounted on the seafloor are often the only viable solution in contexts where abundance estimation involving visual observation methods or towed arrays of hydrophones are impractical (for example, due to rough seas, darkness). My recently completed dissertation research sought to evaluate current passive acoustic abundance estimation methods using recorded signature whistles of Sarasota Bay bottlenose dolphins, through comparisons to true abundance obtained via a census. Because of the limited study area and extensive background data and recordings available for the individually identifiable resident dolphins, Sarasota Bay was well-suited for this kind of ground-truthing research.
Dolphin whistles were recorded by five bottom-mounted acoustic recorders deployed during June-August 2013, and moved periodically to sample a total of 66 point locations within the study area, Palma Sola Bay, as part of a systematic survey design. The methods assessed included conventional distance sampling (CDS), which involves the estimation of distances to animals or their cues (for example, the onset of a signature whistle), conventional photo-ID capture-recapture (CR), which relies on animal photographic capture histories, and combinations or extensions of these two approaches. Of these methods, snapshot mark-recapture distance sampling (SMRDS), a hybrid method that combines elements of distance sampling and capture-recapture with an approach that limits the monitoring period at each point, produced the most accurate estimate. This method relied on auxiliary data in the form of focal animal behavioral follows. However, spatially-explicit capture-recapture (SECR), an extension of CR methods that incorporates spatial data into the estimation process, demonstrated the greatest potential for broad applicability to other species and locations, with minimal to no auxiliary data. Furthermore, when sound transmission loss over distance was considered, replacing traditional straight-line distance in a novel approach to SECR, estimation accuracy improved. In contrast, conventional CR greatly underestimated abundance, despite attempts to account for major sources of variability. In addition to performing similarly to CR, the CDS method requirement of detecting the same whistle on at least 3 sensors (to obtain distances via localization) made this method inefficient and produced small sizes. Thus, it is not recommended for studies involving only a few sensors. The most appropriate passive acoustic abundance estimation method for a given context will depend on that specific context. It is hoped that the insights provided by this study serve to inform future studies, as well as management decisions, with the ultimate goal of more effective species conservation at other sites, where less is known about the animals.
Support for this project was largely provided by the Chicago Zoological Society, Loggerhead Instruments, SDRP staff and numerous SDRP interns.
This article appeared on pages 22-23 of the 2017 SDRP Annual Report, Nicks n Notches.