Measuring the Transport of Floating Plastic Debris Using Vessel-Based Optical Data and Artificial Intelligence

The North Pacific Garbage Patch (NPGP) is known for accumulating floating plastic debris, but little is known on the dominating mechanisms that form its spatial heterogeneity in concentration. Submesoscale processes are likely to be the main drivers of such heterogeneity, especially if their effect on transport is object-specific. Dynamics at these spatial scales remain largely unresolved to date in ocean circulation models, therefore, current studies have to rely on in-situ measurements. The authors present a new method that measures floating plastic debris’ horizontal transport over small scales along vessels’ trajectories. The method applies particle tracking velocimetry on objects detected by an optical artificial intelligence algorithm during The Ocean Cleanup’s campaigns. Given the method’s sensitivity to the vessel’s movement, a Monte Carlo simulation is conducted to estimate object position errors with and without the presence of waves. The same method is applied to overlapping samples of drone-based optical data and the results are compared across measuring devices. Measurement accuracy depends on factors such as sea state, object distance from the vessel, and tracking duration. A first application on a subset of manually classified objects is presented. The ability to estimate floating plastic debris’ transport from in-situ measurements, combined with the collection of meteorological and oceanographic data, will likely gather insightful information on object-specific small scale dynamics in the region of interest. This is not only valuable for research purposes, but essential to assess and improve clean-up efforts.