Darienne Lancaster

Topic

Investigating non-destructive acoustic methods for monitoring rocky reef fishes

Department of Biology

Date & location

• Tuesday, August 11, 2026
• 1:00 P.M.
• Clearihue Building, Room B007

Examining Committee

Supervisory Committee

• Dr. Francis Juanes, Department of Biology, University of Victoria (Co-Supervisor)
• Dr. Dana Haggarty, Department of Biology, UVic (Co-Supervisor)
• Dr. Stéphane Gauthier, Department of Biology, UVic (Member)
• Dr. Xavier Mouy, Assistant Scientist, Woods Hole Oceanographic Institution (Outside Member)

External Examiner

• Dr. Nick Tolimieri, Northwest Fisheries Science Center, National Oceanic and Atmospheric Administration

Chair of Oral Examination

• Dr. Nancy Shackelford, School of Environmental Studies, UVic

Abstract

Comprehensive monitoring of marine species is required to effectively manage and conserve vulnerable fish stocks and threatened species. Common monitoring methods for rocky reef fish species in BC, Canada, include fisheries surveys and visual assessment techniques. These tools are often effective, but fisheries surveys are extractive and may not be suitable for monitoring in protected marine reserves or for vulnerable species, and visual assessment methods are often time consuming, struggle to achieve wide survey coverage, and perform poorly in low visibility conditions. Passive and active acoustic monitoring techniques may be able to address some of these monitoring challenges. Passive acoustic monitoring (PAM) uses hydrophones to record animal sounds and has been successfully used to monitor abundance, density, and species richness for birds, bats, anurans, and many marine mammals. However, quantitative PAM techniques for monitoring most fish species have not yet been developed. Active acoustic monitoring uses echosounders to create images of marine environments and is commonly used to monitor pelagic fish species and zooplankton. However, there are currently no active acoustic techniques for monitoring benthic fish species like rockfishes (Sebastes spp.) due to issues ensonifying near bottom areas with sonar beams. Despite these challenges, the speed of echosounder surveys and their ability to quickly process large volumes of data using standardized analysis templates make it worthwhile to explore methods for detecting benthic rockfish hotspots using active acoustics. The aim of this dissertation was to explore and develop new, acoustic methods for non-invasively monitoring rocky reef fishes. For the passive acoustic component of this research, we used an acoustic localization array and paired video to collect data on rocky reef fish sounds and behaviour at three sites in Barkley Sound, BC. We identified sounds for eight species of rocky reef fish, five of which had never been documented making sounds in the wild. We then used random forest models to compare 47 sound features (e.g. peak frequency, sound duration), to determine if rocky reef fish sounds were unique enough to be identified to the species level. Our models showed fish sounds could be identified to the correct species with 80% accuracy for knock sounds and 88% accuracy for grunt sounds. We also used partial dependence probability plots to identify the most important sound features for differentiating species calls. We then used stereo camera footage to examine the impact of size on fish sound characteristics. We found most fish sound frequency characteristics decreased as fish lengths increased, with larger fish making lower frequency sounds. We also identified four active behaviours associated with fish calling, including fleeing and chasing. We determined that most behaviours had minimal impact on fish sound features and calling patterns. However, Copper Rockfish (Sebastes caurinus) and Quillback Rockfish (Sebastes maliger) made significantly more calling sounds and more grunting type sounds while fleeing. To evaluate the strengths and weaknesses of passive acoustic monitoring for studying rocky reef fish we then compared our passive acoustic data to paired video and SCUBA monitoring data. We found acoustics detected most soniferous species at each site but had lower overall species richness values than video and SCUBA. Acoustics also required more monitoring hours to detect each species than the other methods but provided more usable monitoring hours than video. We also detected strong evidence of species-specific differences in calling rates but rates were variable across sites so more monitoring is required to develop stable estimates of species-specific call rates. We detected diel patterns in fish calling and vessel noise with increased fish calling and decreased vessel noise between dusk and dawn at most sites. These passive acoustic results provide crucial information for the future development of quantitative passive acoustic monitoring for rocky reef fish. We also conducted a collaborative research project with Ha’oom Fisheries Society and Mowachaht/Muchalaht First Nation using active acoustics to detect rocky reef fish hotspots. We performed 39 paired echosounder and remote operated vehicle (ROV) surveys in Nootka Sound, BC. Using generalized linear models, we identified a strong significant correlation between near-bottom fish backscatter and ROV detected benthic fish density with a deviance explained of 52%. This correlation can be used to detect high density areas of benthic rocky reef fish in near-shore coastal habitats using rapid and standardized echosounder surveys. We also developed a step-by-step echosounder and ROV survey guide to facilitate knowledge sharing and increase the accessibility of these techniques. This dissertation provides new, non-invasive methods for monitoring vulnerable rocky reef fish populations which will increase our ability to rapidly survey large areas of the coast and perform long term monitoring in low visibility or remote areas.