Evaluating Microplastic Pollution Along the Dubai Coast: An Empirical Model Combining On-Site Sampling and Sentinel-2 Remote Sensing Data

Abstract

The study addresses the growing concern of microplastic pollution in environmental matrices, emphasizing the significance of monitoring for understanding their distribution, sources, and mitigation. Laboratory-based spectral reflectance analysis of water samples containing visible microplastics revealed distinctive spectral signatures. Coastal water samples collected over two campaigns were subjected to pre-treatment in order to extract microplastics and microscopic inspection followed by spectroscopic confirmation. Results indicated average microplastics concentrations of 0.633 and 0.324mg/L, along with 7.85 and 5.30 items/L in the datasets. Leveraging these findings, along with Sentinel-2 (Level-1C) data and spectral signatures, an empirical spectral microplastics model was developed to convert Sentinel-2's reflectance into microplastics concentrations. This model displayed an 87.30% R2 and ±0.015mg/L RMSE. Subsequently, the model was employed to estimate microplastics concentrations in 2018, 2019, 2020, and 2021, showcasing its potential for monitoring microplastics pollution in the study area and similar regions.

Keywords

• microplastics; remote sensing; sentinel-2; regression