Visual and camera-based detection of floating plastic debris is being used as a monitoring tool without the need for time- and resource-intensive net sampling. As reliance on these technologies increases, in part due to their large sample size, easy implementation, and cost efficiency, a need arises for cross-validation of the resulting datasets with the contents of physical trawl samples. During a recent research expedition to the North Pacific Subtropical Gyre, two experimental remote sensing technologies (UAV and timelapse camera strip transects) were deployed simultaneously with sampling trawls and visual surveys to gather a holistic and comprehensive dataset on the quantity and composition of plastic debris afloat at the sea surface. Manta (0.9m opening, 0.5mm mesh size) and Mega (6m opening, 1.5cm mesh size) surface trawl nets were deployed alongside tethered UAVs, vessel-mounted timelapse cameras, and visual surveyors in various areas of the plastic accumulation zone in the North Pacific. Physical validation of the footage collected by the UAVs’ strip scans and timelapse cameras has allowed for a side-by-side comparison of the various remote sensing methods with traditional physical trawling. A compositional analysis of the net trawls was conducted to investigate the selectivity and potential biases of the visual and camera-based observations. By comparing how and whether different debris items were registered by multiple sampling techniques, we can assess the strengths and limitations related to the use of remote sensing for detection of plastic debris afloat at sea.

Primary Presenter: Helen Wolter, The Ocean Cleanup (h.wolter@theoceancleanup.com)

Authors:

Marjolein Van Vulpen, Utrecht University  (m.vanvulpen@theoceancleanup.com)

Matthias Egger, The Ocean Cleanup  (matthias.egger@theoceancleanup.com)

Robin De Vries, The Ocean Cleanup  (robin.devries@theoceancleanup.com)

Laurent Lebreton, The Ocean Cleanup  (laurent.lebreton@theoceancleanup.com)