Stereo-imaging AUV detects trends in sea urchin abundance on deep overgrazed reefs

Remote underwater vehicles are cutting-edge tools for mapping benthic habitats, yet their reliability for detecting patterns in abundance of benthic species remains largely unexplored. Here, we use a stereo-imaging AUV to investigate changes in abundance of the overgrazing sea urchin, Centrostephanus rodgersii, which has undergone climate driven range-extension to Tasmania. As part of ongoing bi-yearly monitoring of urchin impacts (June 2009–2013), the benthic stereo-imaging AUV (Sirius) surveyed fixed geo-referenced 25 × 25 m plots (625 m²) on deep “urchin barrens” (25–30 m depth) and shallow barren/kelp transition zones (8–16 m) at two sites at St. Helens, northeast Tasmania (∼ 41.25; 148.34). Coincident with initial AUV deployments, urchin abundance was also estimated in the same reef plots using conventional SCUBA diver belt-transects; with comparison of AUV and diver sampling showing AUV-derived estimates to be ∼ 40% lower; while additional AUV sampling at night (high risk for divers in deep water) detected abundances only ∼ 16% lower than that measured by daytime divers, demonstrating strong nocturnal emergence of C. rodgersii. Importantly, patterns in C. rodgersii abundance across reefs and depths were similar between methods; and long-term population trends were concordant between diver and AUV methods. At finer-scales, AUV detections were compromised where remnant kelp canopies obscured urchins, indicating divers to be superior for detecting early-warning of population increases within intact kelp beds. Comparison of C. rodgersii with two other macro-invertebrates (sea cucumber Australostichopus mollis and sea urchin Heliocidaris erythrogramma) revealed that while stereo-imaging AUVs can detect space/time variability in macro-invertebrate abundance, detectability is highly dependent on local ecologies and species-specific behaviours.