Abstract
The reduction of pollutant loads in rivers requires a good comprehension of the contribution of all major (point and diffuse) sources in each sub‐basin. Due to the sudden change of a centrally planned economy to a free‐market economy in 1989, the River Elbe is one of the fastest changing transboundary river basins in Central Europe. The quantification of fluxes of heavy metals in the Elbe basin requires information on the spatial and temporal characteristics of emissions and on physical boundary conditions of the entire river basin. In this paper, a GIS‐based model METALPOL will be presented, that describes the load of five heavy metals (Cd, Cu, Hg, Pb, Zn) in the Elbe basin over the period 1985–1999 as a function of (1) emissions from point and diffuse sources, (2) runoff, (3) soil types, (4) land use and (5) hydrogeology. The results from the model were validated with the measured heavy metal loads and showed reasonable agreement. Between 64% (for Hg) and 84% (for Pb) of the total heavy metal emissions in the Elbe basin is supplied by inputs from diffuse sources. The diffuse hydrological pathways with the highest share are erosion and urban areas. In most cases the measured heavy metal load is lower than the sum of the heavy metal emissions. Retention processes (e.g. sedimentation) largely explain this behaviour and are found to be dependent on the specific runoff of a catchment. About 81–97% of variance for the different heavy metals can be explained by the dependency on specific runoff. The uncertainty analysis presented in this paper has also illustrated that the reliability of the model simulations is directly affected by the quality of the input data (especially point source and urban emissions) and the distribution of the input data like rainfall data, emission data, soil data and land cover data. Finally, scenario analysis indicated that the current proposed measures are not stringent enough to achieve large reductions in heavy metal loads in the Elbe basin. Copyright © 2003 John Wiley & Sons, Ltd.