A low-cost methodology for estimating beach-stranded plastic pellets using drone-based aerial photographs and marine natural wracks as proxy

Abstract

We established a relationship between natural wrack debris and plastic pellet density on a sandy beach using Unmanned Aerial Vehicles (UAV), a low-cost quadcopter drone. Floating marine debris (natural or anthropogenic), including plastic pellets, tend to coalesce, drift, and strand together on sandy beaches. Drones are a cheap and widely available technology. Here, we verified whether it is plausible to indirectly estimate plastic pellet densities using drone aerial photos by identifying the local coverage of natural debris. Pellets and natural debris were manually sampled in situ on two beaches on the coast of São Paulo (Brazil) to verify the relationship between them. Once a model was established, drone aerial photos were taken on the largest beach (Itaguaré - 2 km in length), and heat maps of the distribution of natural debris from both methods were compared. The regression models from the field method predicted pellet densities with 45 % accuracy. Other factors not considered in this study (wave energy and sediment composition) may have influenced this accuracy. However, when comparing the pellet dispersion heat maps obtained from both methods (manual sampling and drone-based aerial photos), it can be conclude that the drone-based method was effective, especially in identifying hot spots for pellet entries onto beaches. The drone-based method allows a single person to evaluate the occurrence of plastic pellets in 35 thousand times more beach area than the conventional field methodology in just over 1 h of flight. This method is fast, easy, scalable, and accessible.