7.
The evaluation and improvement of models is severely
model will be able to precisely match some or all of the
limited due to the paucity of ambient and emissions
measurements described above.  Nevertheless, preliminary
measurements of mercury, particularly of the specia-
comparisons with several of these measurements show
tion of mercury in these measurement programs, and
encouraging agreement and will be reported on in detail in
limited access to much of the existing data.
forthcoming publications.
8.
The limitations of existing emissions inventories were
seen to play a significant role in this and other compa-
1.6.5
Conclusions
rable modeling activities.
1.
The NOAA-HYSPLIT atmospheric transport model,
previously applied to the transport and deposition of
1.6.6
Recommendations
dioxin to the Great Lakes basin, has been shown to be
capable of linking local and distant sources of mercury
The IAQAB recommends the following to the IJC:
emissions to subsequent deposition into the Great
Lakes basin.
Every effort should be made to increase the detail,
particularly on speciation of mercury, and accu-
2.
Given the limited comparative ambient data available,
racy associated with emissions inventories in the
particularly estimates of dry deposition, and known
United States, Canada, and on a global basis.
weaknesses in the input data, especially those associ-
Some coordination in the production of invento-
ated with emission inventories, exact comparison with
ries between the United States and Canada,
monitoring data cannot be expected.  While compari-
resulting in largely seamless inventories appli-
sons of model outputs to ambient measurements and
cable to the same base year, would be most
the nature of the apparent biases continue to be
helpful.
explored, findings to date support the use of this
model as an effective link between sources of mercury
Additional detailed ambient monitoring activities,
and subsequent deposition in the Great Lakes.
including the completion of enhancements to the
recently established mercury measurement
3.
The model demonstrates that mercury deposition from
capabilities of the Integrated Atmospheric Deposi-
anthropogenic sources to the Great Lakes is a regional
tion Network, should be supported, and provi-
and, to a certain extent, a continental issue and further
sions should be put in place to make all com-
reductions would be necessary on a national scale to
pleted measurements available in a timely fashion.
achieve the goal of virtual elimination.
A mercury modeling intercomparison study in
4.
A small but significant portion of this deposition is
North America, similar to that currently underway
likely associated with sources outside the United States
in Europe, could stimulate improvements while
and Canada, and these contributions could be
increasing confidence in the outcomes and
examined in future work.
linkages developed by atmospheric transport
models.  Improvements to the emissions data base
5.
Modeling is generally necessary to develop source-
and assembly of a comprehensive ambient
receptor relationships; however the utility of models
measurement data set should be part of this
will depend largely on the quality of the comparison
model intercomparison.
between their outputs and available ambient measure-
ments.  Therefore, adequate support for both model-
Following model intercomparison and comparison
ing and measurement activities is essential to provision
against measured data, these models should be
of the source-receptor information necessary for
used as a predictive tool in determining the
development and demonstration of effective mercury
potential impact on deposition of proposed
emission reduction policies.
mercury emission control programs currently
under consideration in the United States and
6.
Unfortunately, gaps remain in the understanding of the
Canada.
behavior of atmospheric mercury; the most notable of
these are in the area of atmospheric chemistry.
38