Prognosis of environmental concentrations by geo-referenced and generic models: a comparison of GREAT-ER and EUSES exposure simulations for some consumer-product ingredients in the Itter

Abstract

The aim of this study was the comparison between predicted environmental concentrations (PEC) derived using a generic aspacial model, European Union System for the Evaluation of Substances (EUSES), and a geo-referenced model, the Geo-referenced Regional Environmental Assessment Tool for European Rivers (GREAT-ER). The PECs of some consumer-product ingredients (boron, LAS) and professional uses (EDTA, NTA and Triclosan) were calculated for the river catchment of the Itter, a small tributary to the river Rhine. The PEC local and PEC regional for the water compartment generated by EUSES (default scenario) were subsequently refined with data that realistically reflects the region of North Rhine-Westphalia (NRW scenario) and the Itter catchment (Itter scenario). The results of the three scenarios were then compared with the PEC initial and PEC catchment calculated by GREAT-ER, that was designed as a higher-tiered exposure assessment tool, and with concrete concentrations in the Itter, measured as 24-h composite samples. While the PEC regional of all scenarios was close to the lower end of the measured concentrations, the geo-referenced PECs described equally well the real spacial situation. The measured environmental concentrations confirmed the built-in conservatism of the PEC local calculations by EUSES showing for all investigated chemicals an unrealistically high PEC local (default). The refinement in the more realistic scenarios could not provide a straight forward general improvement of the PEC local. In conclusion, when the EUSES prognosis is refined using more detailed substance and regional specific data, it may provide a fairly accurate modelling especially of substances that are not eliminated in the environment. However, in the case of eliminable substances, it does not match the accuracy of higher-tiered geo-referenced exposure models like GREAT-ER.