Figure 1.
Details of the flying array equipment and operation underwater.
1a. The flying array with High-Definition (HD) video camera attached (labelled “a”). Highlighted on the photo are lights (b), laser pointers (c). The frame dimensions are approximately 1×1×0.5 m (L×W×H). The buoyancy tubes extend approximately 0.25 m beyond the frame fore and aft. 1b. The flying array in operation underwater, illustrating the buoyancy tubes supporting the array attached to the top of the frame.
Figure 2.
Arrangement for deployment of the video array.
Stylised diagram illustrating the full arrangement of equipment during deployment of the video array (not to scale).
Figure 3.
Example frame grabs from the High Definition video.
Example frame grabs from the HD video used for quantitative assessment of habitat characteristics or organism assemblage composition, from Wave Hub site, Cornwall UK, at 60 m. 3a. Two Red Gurnard (Trigla lucerna) and a lesser octopus (Eledone cirrhosa, centre back). 3b. A large Ross Coral (Pentapora fascialis, a bryozoan). In both images, the distance between laser points is 50 cm. 3c. One of 100 random frame grabs/transect for detailed assemblage composition assessment illustrating high resolution enabling identification of smaller organisms.
Figure 4.
Sampling design at the Wave Hub site, Cornwall, UK.
Sampling design at the Wave Hub site, Cornwall, UK, for the High Definition video survey, indicating replicate sample points within the Wave Hub safety zone, plus two control areas. The mixed-nature of the seabed is also highlighted, meaning such sites are impossible to adequately sample using most standard methodology.
Figure 5.
Example data analysis from HD video transects and frame grabs.
Analysis of example data for four taxa from the Lyme Bay MPA study highlighting the potential for the video methodology to provide robust data for testing hypotheses. Here treatments are: existing voluntary closures to fishing (CC), new enforced closures (NC) and near or far controls where fishing continues (NOC, FOC). Data represent baseline conditions in September 2008. 5a, b: data from analysing whole replicate transects; 5c,d: data from detailed random frame grabs (n = 30/transect).