Baited remote underwater stereo-video outperforms baited downward-facing single-video for assessments of fish diversity, abundance and size composition

Rapid changes in video technology have allowed for the development of sophisticated, efficient methods for surveying fish communities, including systems that use single or stereo video cameras, which are baited or unbaited and used remotely, by divers or on Remote Operated Vehicles. Video methods are non-extractive and their deployment can be standardised. As a result of the spatial and temporal repeatability of video techniques they are often used to monitor the biodiversity, assemblage composition and size structure of marine fishes. Because of the biases and sampling efficiencies of different configurations, consideration is required as to which is the most appropriate design for the objectives of a particular study. Baited remote underwater stereo-video systems (stereo-BRUVs), which record across the seascape, and downward-facing single camera baited underwater video systems (downward-BUVs) were deployed in temperate reef habitats on the west coast of Australia to compare the numbers of species and individuals, the assemblage composition, and the relative abundances and size distributions of focal species recorded by both techniques. Stereo-BRUVs sampled a different assemblage composition of fishes than downward-BUVs, observing significantly more species (84 vs 63) and individuals (7321 vs 4490). In general, stereo-BRUVs sampled a greater range of trophic groups than downward BUVs, including species not directly attracted to the bait (e.g. herbivores). Some carnivores that were recorded on the stereo-BRUVs were rarely, or never observed by downward-BUVs. This is attributed to the increased numbers of fish and species recorded in the broader field of view of the stereo-BRUVs. The power to detect a 20, 50 or 100% change (at α = 0.05) in numbers of species and individuals was comparable between methods, but typically greater for stereo-BRUVs for some of the focal species. Length distributions of focal species differed significantly between methods in most cases, with stereo-BRUVs providing accurate and precise measurements, while downward-BUVs often over-estimated lengths. We conclude that forward-facing stereo-BRUVs were superior to downward-facing single camera BUVs in virtually all aspects tested