Sample size requirements for riverbank macrolitter characterization

Anthropogenic litter is omnipresent in terrestrial and freshwater systems, can have major economic ecological impacts. Monitoring modeling of anthropogenic comes with large uncertainties due to the wide variety characteristics, including size, mass, item type. It unclear as what effect sample set size on reliability representativeness statistics. Reliable statistics are needed (1) improve monitoring strategies, (2) parameterize transport models, (3) convert counts mass for stock flux calculations. In this paper, we quantify requirement riverbank characterization, using a database more than 14,000 macrolitter items (>0.5 cm), sampled 1 year at eight locations along Dutch Rhine, IJssel, Meuse rivers. We use perform Monte Carlo based bootstrap analysis statistics, determine relation between variability mean median values. Based this, present requirements, corresponding selected uncertainty confidence levels. Optima sampling effort information gain suggested (depending acceptable level), which function type heterogeneity. found that heterogeneity characteristics varies different categories, demonstrate minimum required depends category. This implies heterogeneous categories need be reach same level For example, describe category soft fragments (>2.5 cm) 90% confidence, 990 were needed, while only 39 uniform metal bottle caps. Finally, within assess requirements each river system. All data collected study freely available, may form basis an open access global used by scientists, practitioners, policymakers future strategies efforts.