A Comparison of Multi-media Models Used in Regulatory Decision-making

Abstract

Although multi-media models are based on the same principle of mass balance, different assumptions may be invoked in the model construct that ultimately can lead to inconsistencies in international fate/exposure assessments. To determine the precision of five different models (i.e., HAZCHEM, CHEMCAN, CALTOX, SIMPLEBOX, USES), a comparison exercise was conducted. Calculations were performed for a set of five disparate nonpolar organic chemicals (benzene, methyl tertiary butyl ether, isopropanol, diethylhexyl phthalate and benzo[a]pyrene), with each model using the same set of inputs to define a ‘unit world’ environment and a chemical-specific emission scenario. Resulting steady-state predicted' environmental concentrations (PECs) in primary (air, water, soil, sediment) and secondary (plants, fish, meat) compartments were analyzed to assess inter-model precision. The precision of estimated residence and persistence times for each chemical was also examined. Results indicated primary compartment PECs for each chemical were typically within an order of magnitude between models. Estimated persistence and residence times were more precise than primary PECs and global transport potential was ranked similarly by four of the models. In contrast, the precision of secondary PECs was poor with discrepancies often spanning several orders of magnitude. Further efforts are needed to: 1) harmonize the prediction of secondary PECs before such estimates are routinely used in regulatory contexts, such as indirect exposure assessments; 2) critically assess and define applicability limits of default algorithms to avoid unrealistic predictions; and 3) validate model results with environmental monitoring data to provide the technical basis to justify scientific acceptance of these tools.