Optimizing coral reef monitoring: species traits and habitat factors predict visual census performance
摘要
Accurate underwater visual census (UVC) of fish community structure is essential for monitoring ecosystem health, managing fisheries, and evaluating response to stressors on coral reefs. However, comparisons between survey methods often reveal significant differences in species detection and estimated density. Here we evaluate the extent to which generalizable species traits (nocturnality, water column position, aggregation, crypsis, coloration, shape, size) alongside habitat characteristics and survey area influence the relative performance of three UVC methods: stationary visual censuses (SVCs), belt transects, and roving surveys. By comparing density estimates for 165 fish species gathered simultaneously via these methods at 42 reef sites in the Florida Keys, USA, we demonstrate that multiple fish traits and coral cover signficantly influence density differences among survey types. Belt transects capture significantly higher densities of smaller (< 10 cm), camouflaged, neutrally colored, and elongated taxa than SVC surveys. Consequently, belt transects may provide more accurate information for programs with broad or multi-faceted conservation objectives, such as coral restoration and invasion management, where species with these traits are of interest. Conversely, SVCs captured higher densities of large, silvered, and fusiform or elongated individuals compared with belt transects, while roving surveys captured slightly greater densities of large-bodied predators than SVCs across all trait levels, suggesting both SVCs and roving surveys are highly suitable for monitoring species of fisheries conservation concern. Our analyses reveal the critical importance of considering the traits of focal taxa alongside management objectives when selecting monitoring techniques, and identifies key fish traits that are important covariates in UVC calibration analyses.