Lisa Watkins , Patrick J. Sullivan, M. Todd Walter

Microplastic pollution is measured with a variety of sampling methods. Field experiments indicate that commonly used sampling methods, including net, pump and grab samples, do not always result in equivalent measured concentration. We investigate the comparability of these methods through a meta-analysis of over one hundred surface water microplastic studies. We find systematic relationships between measured concentration and sampled volume, method of collection, mesh size used for filtration, and water body sampled. Most significantly, a strong log-linear relationship exists between sample volume and measured concentration, with small-volume grab samples measuring up to 10^4 particles/L higher concentrations than larger volume net samples, even when sampled concurrently. Potential biases explored included filtration size (±10^2 particles/L), net volume overestimation (±10^1 particles/L), fiber loss through net mesh (unknown magnitude), and intersample variability (±10^1 particles/L). Contamination is the one potential bias with an effect large enough (±10^3 particles/L) to explain the observed differences. Based on these results, we caution the practice of comparing concentrations across multiple studies or combining multiple study results to identify regional patterns. Additionally, we reiterate previous recommendations emphasizing the importance of contamination reduction strategies, namely that blank samples be collected, tested, and reported as a matter of course for such studies.