Mass balance modeling of 18 chemicals that are representative of Level I substances identified under the Great Lakes Binational Toxics Strategy and targeted for virtual elimination from the Great Lakes has been carried out using a suite of models. The goals of this work are to assess the potential of each substance for transport from local and distant sources and subsequent deposition to the surface of the Great Lakes, and to make estimates of the contribution to total atmospheric loading attributable to emissions at different locations in North America and globally. Models are applied to analyze the efficiency of long-range transport and deposition of Level I substances to the Great Lakes (the Great Lakes transfer efficiency, GLTE). The GLTE is the percentage of chemical released to air in a source region that is expected to be deposited from the atmosphere to the surface waters of the Great Lakes.Modeling at the North American and global scale is carried out using two models based on the Berkeley-Trent (BETR) contaminant fate modeling framework: BETR North America and BETR Global. Model-based assessments of Great Lakes transfer efficiency are used to group the substances according to the spatial scale of emission likely to impact the Lakes: (1) Local or regional scale substances: Dieldrin, Aldrin and benzo[a]pyrene, (2) Continental scale: chlordane, 2,3,7,8-tetrachlordibenzodioxin, p,p-DDT, toxaphene, octachlorostyrene and mirex, (3) Hemispheric scale: PCBs, (4) Global scale: hexachlorobenzene and a-HCH.Using available emissions estimates and the models, the contribution of emissions of Level I substances in different regions of North America and globally to the total atmospheric loading to the Lakes has been estimated. These estimates are subject to large uncertainties, most notably because of uncertainties in emission scenarios, degradation rates of the substances in environmental media. However, model uncertainties due to simplified descriptions of exchange processes between environmental media and environmental conditions also contribute to overall uncertainty in the assessment. Mass balance calculations are presented for seven PCB congeners and toxaphene at the North American spatial scale and for the PCBs and a-HCH at the global scale. Comparison of cumulative historical emissions scenarios with estimated emissions in the year 2000 indicates that relative contributions from sources outside North America are increasing as sources are curtailed in the United States and Canada. In particular, Eastern Europe appears to be becoming a relatively more important source to the Lakes. However, under all emission scenarios considered, the majority of PCB deposition to the Lakes from the atmosphere is attributable to sources in North America. The mass balance models presented in this report provide a quantitative framework for assembling the best available information about properties, sources, partitioning, degradation, transport, and the ultimate fate of persistent organic substances. The uncertainties associated with these assessments are believed to be dominated by uncertainties in emission estimates and environmental degradation rates for the Level I substances and further research should be focused on better characterization of emissions and studies of degradation reactions in various environmental media. Given these uncertainties in the overall mass balance calculations, further model-based studies should concentrate on assessing the influence of more refined descriptions of fate and transport processes within the existing model frameworks, and not on increasing the spatial and temporal resolution of the existing models. Once our understanding of the basic mechanisms resulting in deposition to the Lakes has improved sufficiently, research should focus on spatial and temporal scaling issues.