Study Purpose

The purpose of the mercury modeling is to provide state policy makers with insights as to the potential outcome of proposed NR446 mercury emission reduction regulations. The proposed mercury control actions are intended to reduce in-state mercury deposition, and ultimately mercury levels in the state's fish. Because information about potential rule impacts was not available, Wisconsin utilities contracted for this work to be conducted as part of the public involvement process for the proposed rule-making.

The study was developed in cooperation with the Electric Power Research Institute (EPRI) and conducted by Atmospheric and Environmental Research, Inc.

Objectives

The objective of this work was to estimate the effect of various emission sources on the atmospheric deposition of mercury in Wisconsin. This project focused on the upper Midwest, and in particular the state of Wisconsin. The project relied on well-characterized inventories of mercury atmospheric emission sources in Wisconsin and the other states in the upper Midwest to calculate deposition in the region.

Model Specification and Performance

For this effort, EPRI simulated the global atmospheric cycling of mercury as well as its deposition on a finer continental and regional scale. The modeling system used incorporated models at three spatial scales: a global scale model for initial conditions, with 8º x 10º grid cells encompassing a global source inventory; and the TEAM (Trace Element Analysis Model), consisting of both a continental-scale dispersion model, with 100 kilometer (km) grid spacing; and a sub-continental simulation at 20 km spacing.

Model performance was evaluated by comparing model results at the finest 20 km scale with actual monitored 1998 deposition data from the Mercury Deposition Network (MDN) for Wisconsin and surrounding states. The MDN database includes 27 sites in the United States (including four in Wisconsin) and three sites in Canada. Compared with these values, the model tends to slightly over-predict deposition at the Wisconsin monitoring stations. The normalized error ranges from four percent to 36 percent at the four Wisconsin sites. Overall, the normalized gross bias (and error) is 22 percent, which is well within reasonable performance parameters and considered a reputable modeling effort.

Model bias is believed to be due to inexact mercury emission inventories (and their unknown mercury chemistries) and to uncertain power plant plume chemical reactions involving mercury. In the latter case, both laboratory and field measurements suggest that the ionic form of mercury in power plant plumes undergoes reactions that rapidly convert it into the elemental form, which, in turn, does not significantly deposit locally or regionally. Thus, the mercury deposition simulations conducted in this study are likely to represent an upper bound on the contribution of local and regional sources to mercury deposition in Wisconsin.

The study used a national mercury inventory developed by EPRI for the entire United States, but modified for Wisconsin to incorporate the DNR's 1997 mercury emissions inventory. The study mapped the estimated inventory of both point sources (e.g., coal-fueled power plants) and area (e.g., mobile) sources within the state. Area sources of mercury were distributed in the inventory according to population density.
Modeling Results

In addition to a base case, three emission reduction simulations were conducted. In the first, all anthropogenic (man-made) mercury emissions from Wisconsin were set to zero. In the second, all anthropogenic mercury emissions from Wisconsin plus all coal utility boiler emissions from Minnesota, Iowa, Illinois, Indiana, Michigan, Missouri, and Ohio were set to zero. In the third, all Wisconsin coal utility boiler emissions were set to zero. The modeling generally concluded that one to four percent comes from Wisconsin power plant emissions, four to 10 percent of mercury deposited in-state comes from the combined total of Wisconsin sources, and that 6-18% comes from Wisconsin sources plus regional power plant emissions:

Estimated Reductions in Mercury Deposition at Model Locations Corresponding to Wisconsin MDN Measurement Stations

MDN Site	No Wisconsin mercury emissions	No Wisconsin mercury emissions and no regional power plant mercury emissions	No power plant mercury emissions in Wisconsin
WI08 – Brule River	-6%	-9%	-1%
WI09 – Popple River	-5%	-8%	-1%
WI36 – Trout Lake	-4%	-6%	-1%
WI99 – Lake Geneva	-10%	-18%	-4%

These estimates are consistent with the results of two other studies using U.S. Environmental Protection Agency (EPA) mercury models that have been conducted by state and federal regulatory agencies. The first study was conducted by EPA in the mid-1990's as part of the comprehensive Mercury Study Report to Congress, required as part of the 1990 Clean Air Act Amendments. This study estimated that less than seven percent of mercury emissions from large coal-fueled utility boilers is deposited within 50 km of the facility.

The second study was funded by the Lake Michigan Air Directors Consortium (LADCO), and was released in January of this year. It estimated that utility sources in Wisconsin contribute one to five percent of the simulated wet deposition as measured at the four Wisconsin MDN monitors.

A key overall finding can be reached by comparing the major modeling efforts completed to date by both regulatory agencies and EPRI, including this study. Even when acknowledging all the differences in model structure and years simulated, the models are in general agreement in their attribution of coal-fueled utility boiler mercury sources to deposition. A common finding across the three modeling simulations – the Wisconsin Mercury Deposition Case Study, the LADCO study and the EPA study - is that over most of the state all of the models attribute less than 10 percent of local or regional deposition to utility sources.

Conclusions

The Wisconsin Mercury Deposition Case Study estimated the effect of various emission sources on the atmospheric deposition of mercury in Wisconsin. Applying this study to the proposed NR446 mercury emission, the most important finding is the low (one to four percent) estimate of mercury deposited in-state that is attributed to Wisconsin's coal-fueled utility boilers. This is key given that the proposed mercury control actions are intended to reduce in-state mercury deposition, and ultimately mercury levels in the state's fish. While this study did not specifically evaluate the direct impact that reducing in-state mercury reductions would have on state fish advisories, some general conclusions can be drawn. The study's low estimates of mercury deposition that would be accomplished from in-state mercury control actions strongly suggest that the resulting impact on mercury levels in fish is also low, and that no reduction in fish advisories can be expected.

References

EPRI, 2002. Deposition of Atmospheric Mercury in Wisconsin, Report prepared by AER for EPRI, Palo Alto, CA

EPA, 1997. Mercury Study Report to Congress, Volume III: Fate and Transport of Mercury in the Environment, EPA-452/R-97-005, U.S. Environmental Protection Agency, Washington, D.C.

ICF Consulting, 2002. Application of the REMSAD Modeling System to the Midwest, Memorandum to LADCO, San Rafael, California.