Baited rather than unbaited stereo-video provides robust metrics to assess demersal fish assemblages across deeper coastal shelf marine parks

Coastal shelf ecosystems are essential for the productivity of global fisheries and provide a range of ecosystem services. Cost effective methods for collecting essential information on habitats and species are needed to effectively monitor and manage these key ecosystems across large spatial scales. Demersal fish assemblages can be suitable indicators for understanding the effects of human activity in coastal ecosystems including the implementation of no-take zones and fisheries management strategies. In this study we compare demersal fish assemblage and body-size distribution data collected using an established baited remote underwater stereo-video method (n = 158) and a novel unbaited, wide combined field of view (∼270°) stereo-camera method (n = 341) in one tropical and one subtropical marine park in western Australia. We use multivariate principal coordinate ordination analysis to identify differences in fish assemblages sampled by the two methods and generalised additive models to examine differences in community temperature index (CTI) and the abundance and length distribution of the large-bodied generalist carnivore Lethrinus miniatus. Assemblage composition was significantly different between each method at both Ningaloo (pseudo-F = 70.52, p = 0.0001) and the Abrolhos (pseudo-F = 9.8, p = 0.0001) marine parks. No significant difference was found between the CTI recorded by the two methods at either location. The use of bait significantly increased the number of larger mature L. miniatus observed at both the Ningaloo (BOSS = 0.13 ± 0.05, stereo-BRUV = 5.18 ± 0.56) and Abrolhos marine parks (BOSS = 0.22 ± 0.08, stereo-BRUV = 1.97 ± 0.50) by an order of magnitude. The unbaited method observed a significantly greater number of juvenile L. miniatus at the Abrolhos (BOSS = 0.28 ± 0.19, stereo-BRUV = 0.10 ± 0.06) although this difference was slight. This study demonstrates that the novel unbaited method used here can provide useful information on components of demersal and benthic fish assemblages, however baited methods are likely more suitable for monitoring demersal fish assemblages in offshore marine parks.