Testing the impact of Remotely Operated Vehicle (ROVs) camera angle on community metrics of temperate mesophotic organisms: A 3D model-based approach

Abstract

While Remotely Operated Vehicle (ROV) technologies have rapidly increased our understanding of deeper water mesophotic ecosystems, there are some inherent problems in their use when estimating the abundance of sessile benthic organisms. Specifically, for rocky and coral reefs there has been recent debate about the importance of the angle of the ROV camera to the substrate, which is difficult to standardise, and how variation in this angle might impact abundance estimates and therefore our ability to detect change in mesophotic ecosystems. Here we simulated a series of 3D rocky temperate mesophotic communities based on real data with three different levels of habitat complexity (high, medium, and low). We ‘sampled’ these communities in a virtual space varying the angle of the camera to the substrate, to assess the influence of camera angle on percent cover estimates and overall benthic community composition. We found that increases in the ROV to substrate angle overestimated the percent cover (per degree of camera angle) of gorgonians (large: R2 = 0.16; small: R2 = 0.18), tubular sponges (R2 = 0.11), and ascidians (R2 = 0.05–0.13), and underestimated the abundance of Crustose Coraline Algae (R2 = 0.1, likely due to shading) in the most complex habitats. While we also found some significant differences in the overall benthic community and data dispersion for the more complex habitats, angle explained only a small amount of total variation (<6%) and significant dispersion was only found between 0- and 50-degree angles in high and medium complexity habitats. We conclude that while ROV to surface angle should be a consideration when sampling deep water reefs, its influence on standard measurements that we estimate with photography is likely to be low, and restricted to errors in the abundance of organisms with a high ratio of height to planar projection area (e.g., gorgonians).