A first estimation of uncertainties related to microplastic sampling in rivers

Many studies have been conducted to quantify microplastic contamination, but only a few of them have actually the sampling methodology and associated uncertainties. This study seeks to examine the influence of sampling strategy on the confidence interval of river microplastic estimates. 16 samples are collected in the Gave de Pau River (southwestern France) during a three-hour window with a 330-mu m mesh size net. Three different exposure times (3. 5 and 7 min) allow for a respective filtration rate by the net of 35.6 m(3) (3 samples), 59.4 m(3) (10 samples), and 832 m(3) (3 samples) of water. Organic matter contained in samples is removed by hydrogen peroxide oxidation. The plastic particles are then counted and classified under a binocular microscope. The microplastic concentrations vary between 2.64 and 424 microplastic/m(3), with a median value of 3.26 microplastics/m(3). Statistical analysis does not show differences in microplastic concentrations for the three exposure times. This result seems to demonstrate that a filtration of approx. 35 in , of water is sufficient under similar conditions (similar flow condition and degree of microplastic contamination) and can help reduce sampling and sample processing time. Other analyses, based on 10 filtrations of 59.4 m(3), show that the higher the number of samples, the lower the confidence interval. For triplicates, the mean confidence interval reaches 15% of the median value. Thus, collecting triplicates would seem to offer a reasonable optimum, in combining an acceptable error percentage and time efficiency. These results might depend on the microplastic load of the river, therefore making it necessary to conduct similar analyses on other rivers. This study reports for the first time uncertainties related to microplastic sampling in rivers. Such findings will serve to set up long term monitoring, highlight spatial differ- ences between sites and improve the accuracy of annual microplastic fluxes in rivers. (C) 2020 Elsevier B.V. All rights reserved.