Quantifying the Influence of Uncertainty and Variability on Groundwater Risk Assessment for Trace Elements

Abstract

The application of leaching models to predict field‐scale heavy metal transport has been successful at several sites. Past studies involved site‐specific sorption experiments to quantify the leaching process and to investigate its spatial variability. Uncertainty due to lack of knowledge was frequently ignored. Here we present a leaching model that is based on an extended Freundlich equation to describe sorption. The equation is derived on a nationwide scale and is applicable to arbitrary sites in Germany. Instead of relying on site‐specific sorption experiments, it only requires information on the spatial variability of pH, organic carbon (C org ), and clay content. In addition to accounting for spatial variability, the model also considers the major sources of uncertainty influencing a leaching prediction. These sources involve uncertainty in the spatial distribution of pH, C org , and clay content, as well as uncertainty of the extended Freundlich equation. The modeling strategy is based on a parallel soil column approach, in which a deterministic transport model is combined with a two‐dimensional Monte Carlo method. The model's ability to predict downward movement is tested on two sites in Germany—a wastewater irrigation area and an area in the vicinity of a metal smelter—where cadmium leaching has been modeled previously using extended Freundlich equations derived from local sorption data. Without any parameter fitting involved, the predicted field‐averaged cadmium profiles agree quite well with measured profiles in both cases.