rather than the simple totals allows for a conceptual compari-
son of United States and Canadian contributions.  If the
The geographical region of significant contributions
simple total emissions are shown, Canadian values would in
most cases appear very small relative to the United States
is  somewhat  distinct  for  each  lake,  as  would  be
values.  The use of per capita values may perhaps make the
expected  given  their  different  locations  and  the
comparison between the two countries more meaningful, as
the effect of the dramatically different populations is re-
variations  in  the  extent  of  industrialization  and
moved.  Any of the per capita values can be converted to the
urbanization in each basin.  For example, there are
total emissions for a given country by multiplication by that
country’s population.
significant  contributions  to  Lake  Superior  from
regions within approximately 1000 km west of the
The United States inventory contained estimates of emissions
lake, but the relative contribution of this region to
of the three forms of mercury: elemental (Hg ), ionic (Hg(II)
and particulate (Hg(p), but the Canadian inventory did not.
the other lakes is lower, due to both the increased
Therefore, the United States data also were used to make
remoteness and the presence of adjacent relatively
estimates of the proportions of the different forms emitted
from the corresponding source categories in Canada.  Figure
large point and areal sources.  For Lake Michigan,
11 shows the average emissions profile or distribution of
the contribution from the Chicago region stands
mercury species in the emissions from individual source
categories; the combined United States / Canada annual
out, due to its significant emissions and proximity
emissions of each form of
to  the  lake;  this  particular  region  appears  to  be
mercury from each source category are shown in Figure 12.
Unfortunately, there have been relatively few measurements
somewhat  important  for  the  other  lakes  as  well.
of the proportion of the three forms of mercury emitted from
For Lake Erie and Lake Ontario, contributions from
various source categories, and so this aspect of the inventory
the Ohio River Valley appear to be very significant,
is particularly uncertain.  As is discussed throughout this
chapter, the atmospheric fate of the different emitted forms is
again, due to the high emissions in this region and
quite distinct.  Accordingly, source-receptor relationships are
the comparative proximity to these lakes.
strongly dependent on the emissions profile of the individual
source, and the accuracy of atmospheric mercury modeling
analyses would increase if more and better information
regarding this issue becomes available.
inventory, area sources in the Canadian inventory were
assumed to be located at the centroid of each grid square.
Only anthropogenic emissions from the reference year(s)
It was believed that there were very few (if any) significant
from sources in the United States and Canada have been
changes in Canadian emissions between 1995 and 1996
explicitly included in this analysis.  Other modeling exercises
(Niemi 2001) and so this 1995 inventory was assumed to be
have suggested that the contribution of global sources to
representative of 1996 Canadian emissions.
atmospheric deposition to the Great Lakes accounts for on
the order of 13 percent (Shannon and Voldner, 1995) to ~20
To summarize the inventory used in this modeling analysis,
percent (Dastoor, 2003, personal communication) of the
Figure 7 shows the geographical distribution of the total
total deposition, and inclusion of the global input in this
anthropogenic mercury emissions in the above inventory
HYSPLIT_4-based modeling methodology is planned for the
throughout the United States and Canada.  The annual
emissions in the two countries are broken down by source
category in Figure 8 and on a per capita basis in Figures 9
There are significant uncertainties in both the United States
and 10.  The relative importance of coal combustion in the
and Canadian inventories and in the application of such
United States and metallurgical processes (e.g. smelting) in
inventories in this modeling analysis.  Some potentially
Canada is apparent.  It should be noted that emissions from
prominent sources (e.g. electric arc furnaces) were not
metallurgical processes in Canada have been sharply
explicitly included, and, while emissions from some source
curtailed in more recent years.
categories (e.g. coal-fired power plants) have been measured
with some regularity, other categories have been rarely
The populations used to normalize the national emissions
measured.  As a result, the annual emissions estimates are
to a per capita basis were 265,000,000 for the United States
uncertain for many sources.  Information on the temporal
and 30,000,000 for Canada from 1996 census estimates
variation of emissions was not available, and so all emissions
made in each country.  Displaying the per capita values
sources in the inventories were assumed to have been