AUTHOR=Gorman Daniel , Gutiérrez Alina R. , Turra Alexander , Manzano Aruanã B. , Balthazar-Silva Danilo , Oliveira Natalia R. , Harari Joseph TITLE=Predicting the Dispersal and Accumulation of Microplastic Pellets Within the Estuarine and Coastal Waters of South-Eastern Brazil Using Integrated Rainfall Data and Lagrangian Particle Tracking Models JOURNAL=Frontiers in Environmental Science VOLUME=8 YEAR=2020 URL=https://www.frontiersin.org/journals/environmental-science/articles/10.3389/fenvs.2020.559405 DOI=10.3389/fenvs.2020.559405 ISSN=2296-665X ABSTRACT=

Understanding how microplastic particles move and accumulate within estuarine and coastal waters requires consideration of primary inputs (e.g., raw materials from industrial zones) as well as secondary inputs resulting from fluvial processes (i.e., materials carried into coastal waters by rivers and streams). This study presents a novel approach to achieve this aim, by comparing the individual and combined ability of Particle Tracking Models (PTMs) and seasonal rainfall data, to explain observed inputs of microplastic pellets to the ocean beaches of Santos City (south-eastern Brazil). A Lagrangian PTM based on high-resolution hydrodynamic models was used to simulate seasonal patterns of pellet dispersal from five release points within the Santos Estuarine System (SES) and nearshore waters which are known contributors to the regions microplastic debris problem. Model outputs suggested that the debris field is likely to be small within the estuary (ranging from 3.6 to 8.1 km2), intermediate at the river mouth (mean 34 km2) and greatest for near- and offshore sites (ranging from 34 to 40 km2). The spatial footprints were strongly modulated by season (and rainfall), with simulations alone unable to reconcile daily inputs of pellets observed on the beaches of Santos Bay (ranging from 2 to 51 particles m2 ⋅ d–1). Given this discrepancy, a Generalized Additive Modeling approach was employed to integrate the PTM outputs with rainfall data to improve predictions of beached particles. Results confirmed that considering fluvial processes, could significantly improve the ability to predict rates of pellet accumulation (raising the explained deviance in observed inputs from 41 to 93%). Thus, the study highlights the potential to couple widely used dispersion models with metrics that describe fluvial forcing (rainfall and estuarine flushing) in order to better understand the spatio-temporal dynamics of microplastic debris transport and accumulation within dynamic coastal environments.