Advancing Methods for Sampling and Analysis of Microplastics in the Marine Environment

Abstract

Plastic pollution is one of many growing anthropogenic stressors in the marine environment, and its proliferation and effects are not fully understood. Research reporting abundance of plastic debris in the environment is critical to our understanding of exposure and will help inform trends in pollution over time. Current methods for sampling and analyzing microplastics (MP) require significant infrastructure and have long analysis times. Reproducibility is also poor for analysis of MP > 500 µm, as methods require manual sorting.

The thesis takes a multidisciplinary approach to the subject of plastic pollution, combing insights from workshops with development of new analytical techniques and novel robotic platforms for sampling MP. We explore the importance of collaborative projects in ocean plastic research, with a focus on integration of technology to advance research. The technology in focus is near infrared hyperspectral imaging (NIR-HSI), which we evaluate in greater detail through a systematic literature review to assess the current level of validation and gaps in knowledge with regards to NIR-HSI as an MP analysis method. The studies discussed in the review demonstrate proof of concept as they report good performance data for identification of up to 11 different polymer types. We conclude that NIR-HSI is a promising method for increasing throughput for MP analysis, however further method validation is required.

Our conclusions inform our further work, where we develop an analytical framework for using NIR-HSI for MP analysis. The framework includes development of a diverse spectral dataset used for training a chemometric model for spectral classification. The analytical method is tested to assess performance for different sizes and polymers of MP, establishing working limits of the method. We determine that NIR-HSI performs well for internal and external validation sets (including weathered MP) and provides accurate particle size estimates. In parallel to high-throughput analysis techniques, we also developed a robotic platform for surface sampling of MP in aquatic environments. We conceptualize, prototype and test an uncrewed surface vehicle (USV) similar to a manta trawl for collection of MP samples. We combine USV sampling with NIR-HSI analysis to demonstrate the capabilities of these two technologies when used in tandem, and compare their performance and efficiency to current best practices. We conclude that USV sampling can help overcome logistical barriers, time constraints, improve sample quality, and increase sample throughput. When used in tandem with NIR-HSI, a rapid RAP can be envisioned for future MP studies.