European Union chemical legislation requires the calculation of local and regional Predicted Environmental Concentrations (PEC) for the assessment of the exposure of new and existing chemicals to aquatic and terrestrial ecosystems. Current methods use local models for air, water and soil to estimate chemical concentrations close to the source and a generic multimedia 'unit world' approach to estimate regional PECs. These models assume generic environmental scenarios representing typical situations in European countries and do not account for the spatial heterogeneity and temporal variability in ecosystem characteristics, soil properties, river flow rates, chemical emissions, etc. The environmental and ecological complexity can best be represented in a Geographic Information System (GIS). By coupling a GIS with a fate simulation model the concentrations of substances in a specific environment are predicted more realistically. The GREAT-ER project (Geography-referenced Regional Exposure Assessment Tool for European Rivers) was launched to refine regional and local exposure assessments for down-the-drain chemicals by applying real, spatial-referenced datasets instead of generic or average values. A modular approach was developed consisting of a hydrological model and a waste-flow, river quality and fate model which are linked to a regional GIS-database. For the calibration and validation in two European study areas representative detergent chemicals (LAS, boron) are used. In a parallel study, high-volume intermediates discharged into the river Rhine are simulated.