Synopsis: This project will support numerous modeling activities aimed at determining source regions and source types of mercury depositing within the Great Lakes region. Among the model types to be used are receptor based methods such as positive matrix factorization, back-trajectory calculation, determination of potential source contribution functions and quantitative transport bias analysis. These methods will allow determination of the likely origin of mercury concentrations and mercury deposition measured at several monitoring stations within the region. These analyses will be complemented by source-oriented modeling of mercury transport, reaction and deposition within the region using the Community Multiscale Air Quality model. Use of these complementary modeling methods will provide a robust assessment of the origin of mercury deposition within the region. In addition, this project will support continued monitoring of event-based mercury precipitation levels at six sites and speciated ambient mercury levels at two sites.

Chemicals Studied: The study will focus on mercury, a substance of high concern within the Great Lakes region due to its high toxicity and tendency to accumulate at high levels within fish. Both the monitoring and modeling approaches used will allow differentiation among the various mercury species commonly found in the atmosphere and in the region’s waters.

Geographic Areas: The modeling approaches will be applied to cover a range including the entire Great Lakes region. Both the receptor-oriented and source-oriented modeling activities will be based largely upon data collected at six monitoring stations throughout the state of Michigan, as well as data obtained at partnering sites within the Great Lakes region and northeastern United States.

Project Duration: The project will begin in the Spring of 2006 and extend through the Fall of 2007.

Methods Used: Several receptor modeling and source apportionment approaches will be used. Data that will support these assessments includes that from a previous GLAD project, as well as from other sources. Multivariate statistical techniques, such as positive matrix factorization (PMF) and UNMIX will be used to segregate several principal components of the measured mercury and trace metal concentrations. The results from this analysis will be paired with back-trajectories (as determined by HYSPLIT) to assess regional contributions of these principal source profiles. Quantitative transport bias analysis (QTBA) will also be used to determine geographic gradients in source contributions.

Source-oriented modeling will also be conducted, using a modification of the Community Multiscale Air Quality (CMAQ) model. This work will be used to predict mercury transport, transformation and deposition from know sources. These predictions can then be compared to observations to assess the validity of model outcomes. Alternate scenarios, such as the removal of a certain set of sources, can also be assessed with this methodology.

This project will also support the continued collection of event-based wet deposition samples as six locations throughout Michigan (Grand Rapids, Flint, Detroit, Dexter, Pellston and Eagle Harbor), as well as the high temporal resolution measurement of speciated atmospheric mercury at the Detroit and Dexter sites.

Potential Results and Implications: The results of the various modeling approaches will provide much-needed information regarding the sources of mercury to regional deposition and their relative contributions. In combination, these techniques will provide an important knowledge base for informing policy decisions within the region and beyond regarding the reduction of mercury deposition. Continued support of monitoring activities will further enhance the regional information base for future analysis and modeling activities.

Project Contact:
Gerald J. Keeler, Ph.D.
University of Michigan Air Quality Laboratory
Phone: 734-936-1836
Email: jkeeler@umich.edu