How to evaluate sediment fingerprinting source apportionments

Abstract Purpose Evaluating sediment fingerprinting source apportionments with artificial mixtures is crucial for supporting decision-making and advancing modeling approaches. However, are rarely incorporated into research guidelines model testing currently lacking. Here, we demonstrate how to test using laboratory virtual by comparing the results from Bayesian bootstrapped Materials methods Laboratory ( n = 79) known proportions were created soil samples two catchments in Fukushima Prefecture, Japan. Soil sieved at 63 µm analyzed colorimetric geochemical parameters. The MixSIAR framework a mixing (BMM) used estimate contributions mixtures. In addition, proposed demonstrated use of multiple evaluation metrics report on uncertainty, residual errors, performance, contingency criteria. Results discussion Overall, there negligible differences between mixtures, both models. comparison BMM illustrated trade-off accuracy precision results. more certain solutions encompassed lesser proportion values, whereas markedly less precise. Although performance declined single contributing greater than 0.75 material, models represented general trends identified their major sources. Conclusions Virtual as robust assessing if analytical errors negligible. We therefore recommend always include part process. Additionally, highlight value that consider results, importance reporting uncertainty when apportionments.