Assessing the size selectivity of capture gears for reef fishes using paired stereo-baited remote underwater video

Abstract

All methods used to sample fish populations exhibit some degree of size selectivity, whether due to the physical properties of the gear, behavioral responses to the gear, or the spatial distribution of the targeted population. In most stock assessments, selectivity functions are assumed and their parameters estimated due to a paucity of empirical data necessary to estimate selectivity. However, incorrect assumptions about selectivity functions can impart significant bias, ultimately affecting assessment outcomes. In 2016, a study was conducted to assess the size selectivity of three fishery-independent survey methods for reef fish in the U.S. South Atlantic: (1) a chevron trap survey, (2) a standardized repetitive timed drop hooked-gear survey, and (3) an unstandardized hooked-gear survey designed to mimic industry fishing practices. Each survey method was paired with a stereo-baited remote underwater video (S-BRUV) camera, which was treated as a comparative reference, considering that it is the least selective of the four gears used. A total of 93 stations were sampled with all gear types, and size data were analyzed for three managed reef fishes: red snapper, Lutjanus campechanus (all gears), black sea bass, Centropristis striata (all gears), and vermilion snapper, Rhomboplites aurorubens (all gears except chevron traps). For red snapper, all three capture gears showed dome-shaped selectivity, the degree of doming being more pronounced for chevron traps as evidenced by a decreasing capture probability for individuals over ~500 mm FL. Despite the largest red snapper (> 800 mm FL) being observed more often on S-BRUV, mean size of red snapper was generally smaller on S-BRUV than for either hooked-gear survey, indicating that hooked-gears capture proportionally fewer smaller red snapper. For black sea bass, there was evidence of flat-topped selectivity for all gear types, but chevron traps captured smaller individuals on average than did hooked gears. For vermilion snapper, both hooked gears exhibited dome-shaped selectivity. Given these results it’s clear that improper assessment of selectivity can lead to poor assumptions about the sampled population. Therefore, when the selectivity function of a particular gear type is not explicitly known, studies that directly estimate selectivity should be conducted to reduce assumptions applied to assessments. This study demonstrates the utility of non-destructive video surveys to complement and evaluate the selectivity functions of standard fisheries assessment methods. Furthermore, efforts to broaden the spatial scale of this study’s survey methods, expand the species to be examined, validate collected life history data, and determine the potential influence of interannual recruitment variability on observed selectivity patterns are necessary to better understand selectivity processes for all reef fishes assessed in the region.