IDENTIFICATION OF SOURCES
CONTRIBUTING TO THE CONTAMINATION
OF THE GREAT WATERS BY TOXIC COMPOUNDS
A report prepared for:
Mellissa McCuIlough
Great Waters Program Coordinator
Pollution Assessment Branch, ESD (MD-13)
Office of Air Quality Planning and Standards
U.S. Environmental Protection Agency
Durham, N.C. 27701
By
Gerald J. Keeler and Jozef M. Pacyna
Air Quality Laboratory
The University of Michigan
Ann Arbor, Michigan 48109-2029
Terry F. Bidleman
Atmospheric Environment Service
4905 Dufferin Street
Downs view, Ontario M3H 5T4
Jerome O. Nriagu
Environment Canada
National Water Research Institute
Burlington, Ontario L7R 4A6
Revision Date: 17 March, 1993
-------
DISCLAIMER
This document was prepared by researchers in Great Waters-
related scientific disciplines, and a draft of this report was
reviewed by an expanded group of scientists at a workshop held in
November 1992 in Chapel Hill, North Carolina. Other workshop
participants included representatives from the U.S. Environmental
Protection Agency, the National Oceanic and Atmospheric
Administration, the International Joint Commission, and the
affected States.
This report has been reviewed by the Office of Air Quality
Planning and Standards, Pollutant Assessment Branch, U.S.
Environmental Protection Agency, and has been approved for
distribution as received from the team of authors. Approval does
not signify that the contents reflect the views and policies of
the U.S. Environmental Protection Agency, nor does mention of
trade names or commercial products constitute endorsement or
recommendation for use.
-------
TABLE OF CONTENTS
EXECUTIVE SUMMARY i
ACKNOWLEDGEMENTS xv
1. INTRODUCTION 1
2. ATMOSPHERJC DEPOSITION AS A MAJOR PATHWAY 2
3. GENERAL PROCEDURES FOR SOURCE IDENTIFICATION 4
3.1. Source characterization 5
3.2. Source apportionment techniques 6
3.3 Applications of source apportionment techniques for metals 8
3.4 Problems in the application of source apportionment techniques for
organic compounds: changes in chemical profiles 12
3.4.1 Alteration of PAHs in the Environment 20
3.4.2 Alteration of PCDDFs in the Environment 22
3.5 Applications of source apportionment techniques for orgamcs 24
3.5.1 Volatile organic compounds (VOCs) 24
3.5.2 Semivolatile orgamic compounds (SOCs) 28
A. Pol\ cyclic Aromatic Hydrocarbons (PAHs) 28
B. PCDDFs 33
C. Pesticides 35
D. PCBs 42
-------
4 IDENTIFICATION OF LOCAL SOURCES 45
4.1. Major source categories for toxic compound emissions o
4.2. Emission profiles for major source categories 87
4.3. Emission profiles for diffuse sources of organics 92
4.3.2 Air-surface exchange processes 92
a) Volatilization from Soils 92
b) Air-Water Gas Exchange 98
c) Air - Plant Exchange 101
4.4. Evaluation of emission inventories. Comparison with European studies 102
4.5. Application of source-receptor techniques to study the origin of pollution 104
5. IDENTIFICATION OF DISTANT SOURCES 105
5.1. Emissions from North America outside the Great Waters Regions 107
5.2. Emissions from sources outside of North America 109
5.3. Application of source apportionment techniques for identification of
impacts of emissions on the Great Waters from distant source regions 111
6. BENEFITS FROM EMISSION REDUCTION 115
7. CONCLUSIONS 119
8. RECOMMENDATIONS FOR FUTURE RESEARCH ACTIVITY 121
8.1. Measurement programs 121
8.2. Modeling estimates 122
9. REFERENCES 124
-------
EXECUTIVE SUMMARY
The Great Waters Program as written in section 112(m) of the Clean Air Act
Amendments (CAAA) specifies that U.S. EPA, in co-operation with Commerce/NOAA, shall
conduct a program to identify and assess the extent of atmospheric deposition of hazardous air
pollutants (HAPs) to the Great Waters which include the Great Lakes, Lake Champlain,
Chesapeake Bay, and other coastal waters. In order to meet the ambitious goals set forth in
the 1990 CAAA, EPA identified the need to determine the state-of-knowledge of atmospheric
deposition of HAPs as a major pathway for loadings to the Great Waters.
This report evaluates the available data on the sources of HAPs within and outside the
watersheds of the Great Waters (Great Lakes, Lake Champlain, and Chesapeake Bay), and
discusses the difficulties in deriving such quantitative information. An effort is made to reconcile
the emission data discussed with measured and estimated loading rates. Detailed inventories of
the sources and emission intensities of HAPs has become an indispensable tool in environmental
management. The quantitative targets are amenable to legislative controls, and the emission of
the HAPs is easier to regulate than the resulting atmospheric deposition or food chain effects.
Any attempt to set guidelines on deposition rates for HAPs, in fact, requires that the sources of
the pollutants be known, and that emission rates be determinable.
Atmospheric deposition is one of the major sources of lead, arsenic, cadmium, mercury.
PAHs, several organochlorine pesticides, (e.g., lindane, DDT, chlordane, dieldrin, toxaphene),
PCBs, PCDDs. and PCDFs measured in the Great Waters. The other inputs include non-point
sources, e.g., agriculture practices in the region, urban runoff, leaching from landfills, etc., direct
industrial discharges, tributary inputs, and direct dumping of wastes. Specific examples of
loadings estimates are given in the "Relative Loadings" section of the report. While
approximately half of the lead is of atmospheric origin, it is noted that tributaries account for a
substantial fraction as well. However, designating the loadings as being from tributary inputs
maybe somewhat misleading as much of the lead and other contaminants in the tributary waters
are the results of atmospheric deposition and subsequent runoff. There are many regions where
high quality data are needed to fill-in the mass balance estimates for many HAPs. Even with the
high-quality data in some areas it is very difficult to get a reasonable mass balance for a large
majority of the critical pollutants.
Evidence of atmospheric deposition as a source of persistent semi-volatile organic
compounds (SOCs) to water bodies is provided by their accumulation in soils, sediments, and
-------
peat bogs in the Great Waters region and other locations. The trends of contaminant
concentrations with depth in dated layers of sediments or peat cores track '".:ir known
production/release history. These trends show peak accumulation of PAHs in the 1950s, PCBs
in the mid-1970s, and organochlorine pesticides in the late 1960s to mid-1970s, depending on
the chemical. Rapid increases of PCDDs/PCDFs in Great Lakes sediments after 1940
paralleled the production of chlorinated aromatic compounds, suggesting that incineration of
chlorine-containing waste was the most significant contributor. Other indicators of
atmospheric sources include direct measurements of SOCs in air, rain, and snow from the
Great Waters and remote regions; and accumulation of persistent organochlorine compounds in
biota from small inland lakes in the Northwest Territories, the high Arctic, and Antarctica.
Identifying the specific sources or source types emitting the pollutants into the atmosphere
which ultimately are deposited is another matter. Identification of the major sources and the
deposition pathways of-the critical pollutants should be made for the individual compounds
separately as their sources and behavior in the environment differ substantially. Volatile organic
compounds (VOCs) and SOCs are emitted by both point and area sources. Examples of the
former are stack and fugitive emissions from industrial processes and incinerators. Sources
that emit pollutants over broad areas include vehicle exhaust and evaporation of pesticides and
PCBs.
In general, both local and distant sources contribute to the pollution load at a given
receptor. There are various definitions concerning the meaning of local and distant sources. In
this work local sources are those in the states adjacent to the Great Waters. For the Great Lakes
these states are Illinois, Indiana, Michigan, Minnesota, New York, Ohio, Pennsylvania,
Wisconsin, and Ontario in Canada. For Lake Champlain the emission sources in the states of
New York, Vermont and the Province of Quebec are considered local while local sources for the
Chesapeake Bay are located in the states of Virginia and Maryland. Sources outside the above
defined regions are regarded here as distant, regardless of how much they may contribute to the
total loading of the Great Waters. This somewhat artificial division can be justified when
considering various policy measures to reduce the pollution load. Local and distant sources can
be of either anthropogenic or natural origin. It is believed that natural sources are more important
when discussing the impact of distant source emissions on the atmospheric deposition of
pollutants to the Great Waters.
Although a number of source-receptor techniques are available for estimating the
contributions, it is still premature to conclude what part of pollution load originates within the
study region and what part results from long range transport. The major reason for the present
-------
uncertainty is the lack of reliable input data for application of these techniques including properly
reported emission data. The present lack of monitoring data as well as emissions information is
also problematic for other regions, eg. the North Sea and the Baltic Sea, the two most
extensively studied regions in Europe with respect to the environmental behavior of hazardous air
pollutants.
In general, local sources are characterized on the basis of emission measurements and/or
emission estimates. Representative measurements are considered the best information for
accurately describing emissions. These measured data can be used for source characterization
directly or indirectly. Direct use is when stack tests are performed using both continuous
monitoring or representative grab sampling during short measurement campaigns. Measured data
can also be utilized indirectly by transformation of the measurements into an emission factor or
inclusion in a special calculation procedure.
Source characterization through measurements is often very expensive, and on some
occasions extremely difficult to perform. In these cases other methods are often used, and in most
circumstances are based on emission factors and /or mass balance calculations. The transparency
and comparability of the data used to elaborate or select an emission factor from a handbook of
emission factors are of great importance. The transparency of the data refers to the level of detail
specified for the methods used to prepare a set of emissions factors, e.g., measurements of
emissions rates or concentrations in exhaust gases together with the technological and
meteorological conditions at the time that the measurements were obtained. Comparability of the
data refers to emission factor verification/validation which includes comparison of the factors for
a given source category obtained by various estimation methods.
Material or mass balance calculations can be applied to characterize emissions sources
through the assessment of their emission quantities. The input quantities of the raw materials or
fuel and the output rates of specific pollutants are determined. These rates are used to assess
what fraction of a given pollutant is released in the gas phase while the balance is made for the
amount of the pollutant associated with particles.
Emission sources can also be characterized using the concentrations of a given pollutant
measured in ambient air at a receptor site. Measured concentrations are then compared with
emission source profiles for major source categories likely to contribute to pollution at the site.
This method, referred to as receptor modeling, is useful when emissions from a given source
region originate from one dominant source or group of sources which have well denned emission
profiles.
iii
-------
An important pan of source characterization is to assess the accuracy of the methods used
to prepare emission profiles and to verify source data. The accuracy of source char: ;-rization of
emissions depends on whether the characterization has been made on the basis of actual
measurements at the specific source or through estimation procedures. The methods based on
measurements are considered as more accurate than those using assumptions and calculations,
e.g., methods based on emission factors or material balances. Unfortunately, most verification
procedures focus on activity data (statistical information) and emission factors.
Dispersion models have been the traditional work horse for calculating source-receptor
relationships for air pollutants. These models require detailed emissions inventories for various
sources for the pollutants of interest, e.g., SO2, NOx, etc. Even if the dispersion models were
accurate it is very unlikely that the emissions inventories would be adequate. Emissions
inventories for the criteria pollutants have many short-comings, as discussed earlier in this
document, and these inadequacies are even more severe for HAPs or for pollutants which have
large contributions from fugitive process emissions, natural sources, and dusts. The limitations of
the dispersion oriented methods have led to the development of receptor models. Receptor
models assess contributions from various sources based on observations at sampling or receptor
sites.
Several methods are currently available to assess sources and source regions for various
air pollutants based on the chemical composition of the air at a given receptor. Both the statistical
methods and modeling are used together with meteorological data in order to obtain this
assessment. The origin of the pollution measured at a given receptor can be studied using
information on the che^r^l composition of aerosols and/or mixture of gaseous pollutants.
Statistical methods have been developed which use information on the chemical
composition of aerosols to study contribution of sources or even source regions to the
contamination at a given receptor. The applicability of multivariate techniques for resolving
sources and source regions for aerosols measured at several locations remote from major emission
regions has been tested using absolute principal component analysis (APCA) and the chemical
mass balance (CMB) methods. APCA methods indicate the composition of major components.
such as pollution, crust, and sea-salt components which contribute to the measured concentrations
at receptors. In the past; the APCA methods were applied to air concentrations of total (both fine
and coarse fractions) aerosols. Further improvement of this receptor modeling method was
obtained by applying APCA to aerosol elemental concentration measurements in separate particle
size fractions. The results of this application of APCA gives the basis for interpretation of
IV
-------
coupled chemical reactions and physical processes in remote locations, as well as giving
information concerning atmospheric aging processes and, therefore, the history of the compounds.
Distinguishing sources through chemical mass balance (CMB) models and statistical
methods (e.g., factor analysis, principal component analysis) is more problematic for organic
compounds than metals. This is because VOCs and SOCs are transformed by chemical
reactions in the atmosphere, and different rates of reactivity lead to changes in ratios among
compounds during transport from source to receptci. In addition, some SOCs are associated
to a greater extent with atmospheric particles than others. Preferential removal of paniculate
species by precipitation and dry deposition can alter the relative proportion of SOCs and other
pollutants in ambient air. These changes are reflected in the chemical profiles of atmospheric
deposition and sediments. For example, ambient air contains light and heavy PCBs, PCDDs,
and PCDFs that are distributed between the vapor and particle phases. The heavier, more
particle-bound compounds predominate in rain and sediments.
With a total VOC emission rate of 19.5 megatons (metric) per year, the United States
leads the world in the release of most VOC types. Emissions are highest in the eastern third of
the country. The relative contribution of various industries and vehicles to ambient air VOCs
varies with location. Examples are presented in this report showing the use of CMB modeling
and factor analysis to estimate the contribution of these point and area sources. The agreement
between CMB results and emission inventories is encouraging in many cases.
Less work has been done with CMB and statistical methods for reconciliation of SOC
sources. This is largely due to: a) difficulties in sampling and analytical techniques for SOCs,
and b) alteration of chemical profiles by selective reactivity and physical removal of certain
SOCs, as mentioned above. In a few cases CMB models have been applied to estimating the
contribution of PAH sources in urban air. Principal component analysis has been used to
distinguish patterns of PCDDs and PCDFs from different combustion sources in urban air, to
examine changes in compound profiles that occur as a result of selective atmospheric
deposition, and to differentiate PCDDs/PCDFs from combustion and pulp mills.
Several "marker" compounds have been proposed to help distinguish PAHs from
different sources: wood combustion, spark vs. diesel engines, and unbumed vs. burned
petroleum products (combustion vs. street runoff). The main problem with these markers is
that few are unique to a particular source type, and they are best applied in combination with
other organic and inorganic tracers and with multivariate methods such as factor and principal
component analysis.
-------
Identification of emission sources in the Great Waters regions and their characterization
with respect to atmospheric emissions has been carried out for some time in botn the United
States and Canada. As a result, major source categories have been defined for all of the
combustion-related HAPs, and include these sources: production of electricity and heat,
combustion of fuels in industrial, commercial, and residential units, including wood combustion,
manufacturing and use of various industrial goods, mobile source emissions, incineration of
municipal and industrial wastes, and incineration of sewage sludge. Figure 1 displays the mercury
emissions from various source categories in the Great Lakes Basin. There are, however,
differences in quantitative assessment of the fluxes from the above sources, reported by various
research groups in the United States and Canada. These differences should be resolved through
thorough examination of the available data using verification techniques for emission data and
joint supplementary research programs in both countries.
A recent emissions inventory for Ontario and eastern North America has been prepared
under contract for the Ontario Ministry of the Environment (MOE). Inventories for PAHs and
PCDDs/PCDFs were derived after compilation of emission factors for a large number of
source types, including industrial processes, vehicles, residential combustion (oil, gas, wood),
power plants, incinerators, open burning, and forest fires. Estimated annual releases in eastern
North America were 9397 metric tons PAHs and 414 kg PCDDs/PCDFs. The breakdown of
PAH emissions by source type was: residential wood combustion 31%, other stationary fuel
combustion (including power plants) 17%, industrial processes 29%, transportation 12%, other
11%. In the case of PCDDs/PCDFs, stationary fuel combustion and solid waste incineration
each accounted for 46% of total releases.
The uncertainties in these estimates are major and difficult to quantify. This is largely
due to the quality of emissions data, which are often incomplete and highly variable among
sources, even within the same class. For example, reported emission factors for
PCDDs/PCDFs from incinerators span more than an order of magnitude.
Emissions data currently reported by EPA and the IJC reveal that a large portion of the
HAPs in the United States are generated outside the Great Lakes region. This is particularly true
for emissions from point sources. The states neighboring the Great Lakes Basin, particularly
Missouri, generate large quantities of these emissions in electricity and heat producing power
VI
-------
"rcDcervc Sources cf Mercury
- me Greet Lakes Bcsir.
ossn
6.
C
0
^
IL
IN
Ml
MN
NY
OH PA
Wl ONT
10
C
C
o
Waste Incineration
'9.
1
9, m% MM ™™
IL
IN
Mi
MN
NY
OH
PA
Wl
ONT
10
5
v.
a:
9 ma
$%''\ V/////A
1
'9>,
Wk mm Wfa
IL
IN
MN
NY
OH
PA
Wl
ONT
c i J
ceMcneojs Sources
IN
Mi
MN
NY
OH
PA
Wl
ONT
Total Emission By State
5 - With % of Total Emission
| :oJ 3.5* 4-
2. . I
6.4* 6.6X
3.3*
2.3*
5.6*
2.0* 16.1*
IL
IN
M!
MN
NY
OH
PA
Wi
ONT
Figure 1 Contribution of ttmospheric emissions of mercury from various source
categories to the state emissions in the Great Likes region.
Vll
-------
plants, primary and secondary non-ferrous smelters, steel and iron manufacturing plants, and
waste incinerators. Sources in the St. Louis/Granite City area were found to be a major
contributor to HAPs transported over Lake Michigan during a summertime pollution episode.
Emissions in regions outside the Great Lakes, Lake Champlain Basin, and the Chesapeake
Bay, heavy metals and persistent organic compounds can and do reach the surface of these
waters. While entering the atmosphere, these pollutants are subject to long range transport,
transformations, and deposition processes "en route". The extent to which these processes occur
depends on stack parameters, temperature and velocity of exhaust gases, meteorological
conditions, and the physical and chemical forms of pollutants. Recent studies provide the basis
for estimating what fraction of the HAPs emitted from major point sources is deposited in the
vicinity of the emission sources (local deposition) and what part is transported and deposited
outside the emission region.
Although PCBs are no longer sold in the United States and Canada, large reservoirs
still remain. Of the 640,000 metric tons of PCBs produced in the U.S., 85% are estimated to
remain in service (largely in transformers and capacitors), buried in landfills, and circulating
in the environment. Releases of PCBs in eastern North America from leaking transformers
and landfill gases were estimated at a few hundred kilograms per year. PCBs are also released
during combustion, but emission factors for incineration of various types of waste are highly
variable.
It has been difficult to obtain reliable figures for the production and use of pesticides in
the United States and Canada because of proprietary restrictions which protect their release.
However a recent survey has provided such information for herbicides on a state-by-state
basis. Annual usage of the top ten herbicides in the U.S. totaled over 150,000 metric tons in
1987-89. Four chemicals - atrazine, alachlor, metolachlor, and EPTC — accounted for 62%
of this total. The bulk of these pesticides (80% or more) were used on corn and soybeans.
The National Oceanic and Atmospheric Administration (NOAA) reported application figures
for 35 herbicides, insecticides, and fungicides in coastal drainage areas in the U.S., and
provided details of use by crop and season. Use of the 35 chemicals in 1987 amounted to over
13,000 metric tons. Three estuarine drainage basins in the mid-Atlantic and southeast states
ranked highest in pesticide use: Chesapeake Bay, Albemarle/Pamlico Sound, and Winyah Bay.
Organochlorine insecticides such as DDT, aldrin, endrin, dieldrin, chlordane,
heptachlor, lindane, and toxaphene have been of most concern in the Great Waters because of
their persistence and tendency to accumulate in biota. Most of these insecticides have been
viii
-------
banned or severely restricted in the United States and Canada, but are still used in Mexico.
Central and South America, Africa, and Asia. Reliable statistics on worldwide use that are
needed to estimate the contribution of these foreign sources are largely unavailable.
In the case of emissions from point sources with a stack height of > 150 meters (e.g. large
power plants, primary non-ferrous smelters, cement kilns, steel and iron plants and waste
incinerators), all of which employ high temperature processes, only 15 to 20 % of toxic emissions
were deposited locally. The majority of the pollutants were transported out of the emission
region. Research in the Great Waters Regions is needed to quantify local deposition.
Particularly, information is needed on the importance of urban area emissions on deposition to
nearby water bodies. It is certain that we must consider emission sources both within the Great
Waters region and outside the watersheds in order to assess the origin of atmospheric toxic
compounds deposited on the water surface in the region.
The quantity of emissions for the heavy metals and persistent organic compounds of
concern in the states around the Great Lakes, Lake Champlain, and the Chesapeake Bay is
difficult to assess due to diversity of emission numbers reported by various research groups. For
most of the heavy metals considered in this work emission estimates differ by one order of
magnitude and, therefore, are presently under revision.
Emissions from other source regions in North America may also affect the amount of
pollution load deposited to the Great Waters although no definitive evidence has been provided by
measurements and assessment for heavy metals and persistent organic pollutants. Compelling
evidence suggests that pesticides applied in the southern portion of the U.S. are subsequently
deposited into the Great Lakes and other bodies of water. Furthermore, the NAPAP concluded
that such an impact exists for deposition of sulfate. The source receptor relationships calculated
for wet sulfate deposition are shown in Figure 2. This figure suggests that the contribution from
sources beyond 1000 km dominates the sulfate deposition to Lake Champlain and the Adirondack
Mountain region. As the sulfates are transported on particles, as are many of the metals and
organic compounds discussed in this report, one would hypothesize that the metals and organic
compounds emitted from sources outside the study region would also be deposited to the Great
Waters, especially those compounds that are emitted from the same sources as the sulfur. At
present, regional models of long range transport of air pollutants are available and with some
modifications they can be used to assess the contribution of emissions from outside source regions
to the Great waters. An emission inventory with the appropriate spatial distribution of the data
needs to be prepared. The experience gained through NAPAP emissions inventory development
can be used as a starting point for this purpose. Collaboration with the Mexican authorities on
ix
-------
environment protection is highly recommended as the Mexican emissions of heavy metals and
persistent organic pollutants are expected to contribute to the contamination of the Great Waters.
* ' Source apportionment of >vet sulfate deposition (percent per state or subprovincei at
Wnueface Mountain, NY
Input of organochlorines to the Great Waters continues, as indicated by their presence
in air and precipitation. Likely sources are long-range transport from countries where
organochlorines are still applied, and volatilization from North American farmland treated
years ago. Distinguishing "new" from "old" sources of these compounds is an essential and
ongoing area of research. Tools for this purpose include air mass trajectories which can
provide the history of the air mass on days with elevated levels of pesticides in the ambient air,
often revealing a path from heavy use areas to the Great Waters, and examination of ratios
among isomers and breakdown products of certain pesticides for clues to their source. Models
have demonstrated that facile movement of organochlorine pesticides from the southern states
to the Great Lakes can take place, and elevated concentrations of these compounds in Ontario
air have been traced to air masses arriving from the south.
-------
Evaporation rates of pesticides from treated fields have been measured and related to
properties of the chemical, the soil, and the evaporation rate of water. The factors affecting
pesticide volatilization are known, and models have been developed to predict fluxes from
soils. There is a need to apply these measurement techniques and models to "old source"
situations (fields treated in the past) and on a larger scale to obtain estimates of loadings to the
atmosphere.
In the report to MOE, estimates of herbicide and insecticide releases to the atmosphere
in eastern North America were 30,795 and 3,091 metric tons per year, respectively. These
were based on usage information and a soil volatilization model. The insecticide figure is
probably low, because it does not include a large number of current-use chemicals. At the
present time it is not possible to attach a degree of uncertainty to these estimates, other than to
say that it is probably large.
Gaseous SOCs, particularly PCBs and organochlorine pesticides, exchange freely
between the atmosphere and the earth's surface -- water, soil, and vegetation. Seasonal high
and low cycles of these compounds in ambient air are observed, which are correlated to
temperature and thus volatilization rates. The Great Waters can act as either a sink or source
of vapor-phase PCBs and pesticides, and gas exchange forms a large (and poorly understood)
part of the contaminant budget. Long-range transport of persistent SOCs may not occur in a
single event, but in steps as the compounds are continually deposited and revolatilized from
land, water, and vegetation. Emissions from new sources must be cast in light of this global
background of "recycled" material.
Pesticides, such as hexachlorocyclohexane (HCH), have been found to be global air
pollutants measured in remote areas around the world. As the major application of this pesticide
is in Asia and the wind patterns at various altitudes do not exclude the air mass transport form the
Asian continent to North America, HCH deposited in the Great Waters regions may originate
from as far away as India, China, or the former Soviet Union. This hypothesis can be tested by
the application of global models, or at least hemispheric models. At present, such models are used
to study the transport of green-house gases in the atmosphere and the transport of sulfur in the
Northern Hemisphere.
Several methods can be applied to reduce emissions of toxic heavy metals and persistent
organic pollutants and eventually reduce the atmospheric deposition of these pollutants to the
Great Watefs. Technological solutions presented in a form of Best Available Technology (BAT)
package or Best Practicable Technology (BPT) package as well as non-conventional methods
xi
-------
offer emission reduction possibilities for point sources within all major source categories
contributing to the contamination on the Great Waters. Previous experience gained in past efforts
can be used in recommending emission reduction scenarios in the Great Waters region. Cost
estimates and benefits from the implementation of the control techniques should be carefully
studied.
New research initiatives are necessary in order to meet the requirements outlined in the
conclusions of this report as well as to test the important hypotheses proposed here. These
activities would include both measurement programs and modeling estimates.
Source inventories for combustion-related SOCs such as PAHs and PCDDs/PCDFs
need to be improved. The wide variability in emission factors from various industrial and
incineration processes will probably be reduced as emission controls are put into place.
Great improvements in sampling and analytical methods for SOCs have been made over
the last decade, making the determination of source profiles more reliable. These
improvements should lead to increased use of CMB models and statistical methods such as
factor and principal component analysis for source reconciliation of SOCs. Additional
"marker" compounds should be sought which can be used as source tracers.
Restrictions which protect the release of production and use statistics for pesticides in
the United States and Canada must be removed to allow free exchange of this essential
information. Further, an international effort is needed to identify types and quantities of
pesticides used in foreign countries.
New measurement programs are needed in order to improve the quality of source-
receptor techniques which are used to assess the magnitude and origin of deposited pollutants.
These measurements are needed at both the receptor, the Great Waters themselves, as well as at
the sources of the emissions. The following are recommended:
- emission rates and emission factors for toxic heavy metals and persistent organic pollutants from
large point sources in the study region should be evaluated on the basis of measurements of their
concentrations in exhaust gases;
-soil volatilization models for pesticides and mercury need to be improved and to the point
where they can be applied regionally. These models should be validated by experimental
measurements of pesticide and mercury fluxes from "old source" areas (previously treated
land) as well as from freshly treated or contaminated fields and areas.
Xll
-------
a program should be undertaken to identify and quantify current sources of banned
organochlorine pesticides to the Great Waters. In particular, contributions from old sources
within the U.S. and Canada should be assessed, and weighed against long-range transport of
organochlorines from foreign countries.
new emissions of persistent SOCs must be evaluated relative to the global background of
recycled material. To do this, we must determine the quantities of SOCs currently residing in
the air, land, water, and vegetation reservoirs and the rates of exchange among these
reservoirs.
- physical and chemical forms of the most volatile compounds should be established through
measurements carried out in major sources in the study area; and
- emission rates for the most volatile metals, and particularly mercury, as well as SOCs should be
derived on the basis of measurements over the water surface in the Great Waters and the
surrounding soils. The results should be representative for the meteorological conditions as in the
Great Waters and exemplify seasonal changes.
Measurements at receptors should provide information which is needed in order to
improve the accuracy of source-receptor relationship analysis. The following is recommended:
- size-differentiated chemical composition of aerosols should be measured at receptors which can
represent conditions over the water surface in the study area; and
- simultaneous measurements of the gaseous and particle phases of the studied pollutants with the
help of newly developed techniques (e.g., denuder methods) should be undertaken in order to
provide information on gas-to-particle conversions ( and particle-to-gas conversions) for the most
volatile pollutants under study.
Improvement is needed within the three groups of estimates: emission estimates.
dispersion modeling, and receptor modeling. The following is recommended for the improvement
of emission estimates in order to assure better understanding of source identification in the Great
waters region:
- gridded emission inventory for the studied pollutants should be approached for the whole
territory of the United States and Canada;
- seasonal changes of mercury and volatile organic compound emissions need to be quantified and
techniques developed to estimate these emissions; and
- an approach should be defined to assess emissions of pesticides in the Northern Hemisphere with
particular emphasis on Mexico and the Asian countries.
Improvements in source identification through the further development of dispersion
modeling for toxic air pollutants is needed. The following are recommended:
xiii
-------
- continue to modify and improve the existing long-range transport models so they can be used to
study the contribution of emissions from sources in North America, both within ard outside the
study region to the hazardous pollution load deposited to the Great Waters; and
- an approach should be made to apply the existing global scale models to investigate the
possibility of pesticides, e g., lindane, used in Asia to be transported within air masses to North
America and deposited also in the Great Waters region.
From the above discussion it is clear that there is a considerable amount of uncertainty in
our estimates of the sources of the HAPS measured in the Great Water areas. The level of
research activity must increase and cooperative programs must be implemented before our
understanding of the sources of the critical pollutants found in the Great Waters will be complete
enough for policy measures to be developed and implemented.
xiv
-------
ACKNOWLEDGMENTS
The authors of this section of the report would like to acknowledge the assistance of the
many people who supplied many of the data, tables, and figures used in this report. Much of the
best information on emissions of toxic compounds is only now becoming available and the
following people shared their work with us: Bill Benjey (EPA-AREAL), Terry Clark (EPA-
AREAL), Ann Pope (U.S. E.P.A.), P.K. Mishra (OME), John O'Connor (Radian), Jorg Munch
(Dornier GmBH),, Trevor Scholtz (ORTECH), Bob Stevens (EPA-AREAL), Carmen Benkowitz
(Brookhaven National Laboratory), Chris Veldt (TNO), Eva Voldner (AES). The authors would
also like to thank Mila Simmons (University of Michigan) for her input and information, and
Marion Hoyer for her assistance on many aspects of the report preparation. We would like to
thank the reviewers of this report, Tom Holsen and Ken Noll for their valuable insights and
suggestions. Lastly, we thank the participants of the Great Waters Study Report Workshop for
their helpful suggestions in preparing this revision.
xv
-------
1. INTRODUCTION
The Great Waters Program as written in section 112(m) of the Clean Air Act
Amendments (CAAA) specifies that U.S. EPA, in co-operation with Commerce/NOAA, shall
conduct a program to identify and assess the extent of atmospheric deposition of hazardous air
pollutants (HAPs) to the Great Waters which include the Great Lakes, Lake Champlain,
Chesapeake Bay, and other coastal waters. In order to meet the ambitious goals set forth in
the 1990 CAAA, EPA identified the need to determine the state-of-knowledge of atmospheric
deposition of HAPs as a major pathway for loadings to the Great Waters.
Numerous reports document the fact that the atmosphere is a major source of toxic
contaminants found in many aquatic ecosystems. This report includes an evaluation of the
available data on the sources of HAPs within and outside the watersheds of the Great Waters
(Great Lakes, Lake Champlain, and Chesapeake Bay), and will discuss the difficulties in deriving
such quantitative information. Detailed emissions inventories of the criteria pollutants, e.g. SO2,
NO2, etc, as well as HAPs have become an indispensable tool in environmental management. The
quantitative targets are amenable to legislative controls and the emission of HAPs is easier to
regulate than atmospheric deposition or food chain effects. Any attempt to set guidelines on
deposition rates for HAPs, in fact, requires that the sources of the pollutants be known and the
emission rates be determinable.
There is another cogent reason for concern about air emissions of toxics in this country.
Of the 12 billion kilograms of the material flows reported in the Toxic Release Inventory (TRI) of
1988, about 39% (or 5 billion kg) went into the atmosphere, while only 6% and 9% were
discharged directly to the surface waters and soils, respectively (EPA, 1990)1 In many aquatic
ecosystems, however, large quantities of toxic contaminants can also be derived from industrial
and municipal wastewater discharges, storm water and urban run-off, leachates from landfills and
dump sites, etc. Furthermore, the transfer from one compartment to another becomes quite
ambiguous as is in the case of combined storm-sewer systems in which contaminants are both
directly discharged into the sewers as well as being deposited from the atmosphere with the
precipitation. The wastewater eventually reaches a wastewater treatment plant which separates
the liquids and solids, discharges the processed effluents into surface waters, and then often
incinerates the solids re'moved in the process. The incineration of the sludge results in air
1 The quantity of toxic pollutants emitted is definitely underestimated in the TRI data as this inventors' includes
only 13% of the nations manufacturing facilities.
-------
emissions which continues the cycle by dispersing the hazardous pollutants into the atmosphere
where they are eventually deposited in a downwind environment by dry and wet processes. So,
while the focus of this report is on air toxics, the need for an integrated or whole ecosystem
approach to managing toxic contaminants in the Great Lakes should be emphasized.
2. ATMOSPHERIC DEPOSITION AS A MAJOR PATHWAY
Atmospheric deposition is one of the major sources of lead, arsenic, cadmium, mercury,
PAHs, several organochlorine pesticides, (e.g. lindane, DDT, chlordane, dieldrin, toxaphene),
PCBs, PCDDs, and PCDFs measured in the Great Waters. The other inputs include non-point
sources, e.g. agriculture practices in the region, urban runoff, leaching from landfills, etc., direct
industrial discharges, tributary inputs, and direct dumping of wastes. Specific examples of
loadings estimates are given in the "Relative Loadings" section of the report. While
approximately half of the Pb is of atmospheric origin, it is noted that tributaries account for a
substantial fraction as well. However, designating the loadings as being from tributary inputs
maybe somewhat misleading as much of the Pb and other contaminants in the tributary waters are
the results of atmospheric deposition and subsequent runoff. There are many areas where high
quality data are needed to fill-in the mass balance estimates for many pollutants and it is very
difficult to get reasonable numbers for a large majority of the critical pollutants.
Evidence of atmospheric deposition as a source of persistent semi-volatile organic
compounds (SOCs) to water bodies is provided by their accumulation in soils, sediments, and
peat bogs in the Great Waters region and other locations. The trends of contaminant
concentrations with depth in dated layers of sediments or peat cores track their known
production/release history. These trends show peak accumulation of PAHs in the 1950s, PCBs
in the mid-1970s, and organochlorine pesticides in the late 1960s to mid-1970s, depending on
the chemical. Rapid increases of PCDDs/PCDFs in Great Lakes sediments after 1940
paralleled the production of chlorinated aromatic compounds, suggesting that incineration of
chlorine-containing waste was the most significant contributor. Other indicators of
atmospheric sources include direct measurements of SOCs in air, rain, and snow from the
Great Waters and remote regions; and accumulation of persistent organochlorine compounds in
biota from small inland lakes in the Northwest Territories, the high Arctic, and Antarctica.
Compared to the other Great Waters considerably more research has been performed on
the deposition of HAPs to the Great Lakes. Using the best "available" data quantitative estimates
have recently been made on the fluxes of many toxic substances into these lakes (Eisenreich and
Strachan, 1992). The relative importance of Great Waters contamination by atmospheric
-------
deposition is the focus of another part of this report. However, the general conclusion can be
drawn on the basis of a recent review (ICF, 1992 for EPA), that between 35 and 50% of the
annual load of the HAPS of interest enter the Great Lakes waters through atmospheric deposition
It was concluded, for example, that the atmospheric lead input was responsible for 35 to 47% of
all inputs to the Great Lakes during the period from 1977 to 1981.
The deposition of airborne toxic substances has also been studied in Europe. The results
of long-term measurements on pollution of the North Sea and Baltic Sea waters lead to the
conclusion that as much as 50% of lead and mercury and between 30 and 50% of arsenic,
cadmium, chromium, copper, nickel, and zinc enter these waters through atmospheric deposition.
The picture in Europe for the persistent organic compounds, discussed in this report, is very
unclear due to lack of reliable measured data
The good agreement between the results obtained for the European Studies on the North
Sea and the Baltic Sea, and the findings reported for the Great Lakes indicates that atmospheric
deposition is a very important pathway for HAPs measured in these waters. This indication,
although regarded as preliminary, due to limited data for most of the organic toxins, argues for
research on sources and source regions generating atmospheric emissions of the studied pollutants
to identify the major sources of the Great Water's contamination. Once identified, the major
sources (often called "hot spots") can be targeted and emissions reduction programs can be
implemented leading to a decrease in the pollution load to the Great Waters.
Identifying the specific sources or source types emitting the pollutants into the atmosphere
which ultimately are deposited is another matter. Identification of the major sources and the
deposition pathways of the critical pollutants should be made for the individual compounds
separately as their sources and behavior in the environment differ substantially. Volatile organic
compounds (VOCs) and SOCs are emitted by both point and area sources. Examples of the
former are stack and fugitive emissions from industrial processes and incinerators. Sources
that emit pollutants over broad areas include vehicle exhaust and evaporation of pesticides and
PCBs.
In general, both local and distant sources contribute to the pollution load at a given
receptor. There are various definitions concerning the meaning of local and distant sources. In
this work local sources are those in the states adjacent to the Great Waters. For the Great Lakes
these states are Illinois, Indiana, Michigan, Minnesota, New York, Ohio, Pennsylvania.
Wisconsin, and Ontario in Canada. For Lake Champlain the emission sources in states of New
York, Vermont and the Province of Quebec are considered as local while for the Chesapeake Bay,
-------
the states of Virginia and Maryland. Sources outside the above defined region are regarded here
as distant, regardless of how much they may contribute to the total loading cf the f r\.Mt Waters.
This somewhat artificial division can be justified when introducing various policy measures to
reduce the pollution load. Local and distant sources can be of either anthropogenic or natural
origin. It is believed that natural sources are more important when discussing the impact of
distant source emissions on the atmospheric deposition of pollutants to the Great Waters.
Although a number of source-receptor techniques are available for estimating the
contributions, it is still premature to conclude what part of pollution load originates within the
study region and what part results from long range transport within air masses. The major reason
for the present uncertainty is the lack of reliable input data for application of these techniques
including properly reported emission data. The present lack of-monitoring and emissions data is
also problematic for other water bodies currently being studied with respect to the environmental
behavior of hazardous air pollutants, e.g. the North Sea and the Baltic Sea.
3. GENERAL PROCEDURES FOR SOURCE IDENTIFICATION
There are a number of methods which can be used to identify sources or source regions
contributing to the pollution load measured at a given receptor. These can be categorized into
statistical methods, chemical and isotopic trace methods, meteorological methods including
trajectory techniques, and various combinations of these approaches. Emissions of the
compounds discussed in this report to the atmosphere are multitudinous. Most of the compound
classes discussed here stem from human activities, although natural sources make a contribution
in some cases. For example, PAHs are released both by anthropogenic combustion and forest
fires and Hg is emitted during coal combustion and is released from mines and soils. Source
characterization, when performed accurately, is a powerftil tool for the assessment of emissions
from local sources. To assess atmospheric deposition to a receptor such as the Great Waters and
recommend remedial action, it is necessary to know the relative ontribution of various sources to
the total mass loading. These are deduced from a combination of source characterization and
source apportionment techniques such as emissions inventories, dispersion modeling, multi-variate
(statistical) methods such as factor analysis, and chemical mass balance (CMB) models. Some of
these methods are briefly described by Zweidinger et al. (1990), and Sweet and Vermette (1992)
in their article on VOCs In Illinois and St. Louis, and expanded upon in their references, and in
Gordon (1988), Henry et al. (1990) and Hopke (1991).
Source apportionment methods, often referred to as receptor models are based on
statistical techniques using measurements and information on sources for an assessment of the
-------
local source contributions. These techniques used together with meteorological information can
also be utilized for assessing the contributions of distant sources. Receptor models that implicitly
incorporate meteorological data are often referred to as hybrid models (Keeler, 1987).
3.1. Source characterization
In general, local sources are characterized on the basis of emission measurements and/or
emission estimates. Representative measurements are considered the best source of information
for accurately describing emissions. Measured data can be used for source characterization
directly or indirectly. Direct use is when stack tests are performed using both continuous
monitoring or representative grab sampling during short measurement campaigns. Examples of
inventories developed for semi-volatile organic compounds (SOCs) are given in Johnson et al.
(1992). There are several regulations in force requesting measurements of emissions rather then
estimation for certain types of sources or source categories, e.g. the Commission of European
Communities requests emission measurements for power plants with capacity higher than 300
MW In the U.S., the TRI program also requests any industry with 10 or more employees to
report the release of any of the 322 compounds on the TRI list.
Indirect use of measured data for source characterization is whenever information
obtained from measurements is transformed to an emission factor or included in a special
calculation procedures. Emission factors for PAHs in mass/vehicle-km are developed from auto
exhaust sampling or measurements in traffic tunnels. Sampling in stacks yields emission factors
for industries, power plants, and incinerators. Often these can be related to fuel usage; e.g. mg
PAHs released per ton coal burned. Aerial pesticide losses are estimated from models relating
spray drift and volatilization to application rates, soil properties, and meteorological factors.
Emissions are usually apportioned to a grid network for dispersion modeling and impact
assessment.
Source characterization through measurements is often very expensive, however, and on
some occasions extremely difficult to perform. In these cases other methods are often used, and
in most circumstances methods based on emission factors and /or mass balance calculations are
employed. The transparency and comparability of the data used to elaborate or select an emission
factor from a handbook of emission factors are of great importance.
Mass balance calculations can be applied to characterize emission sources through the
assessment of their emission quantities. The input and output rates of a given pollutant are
determined and are used to assess what portion of a given pollutant is released in the gas phase
-------
while the whole balance is made for the amounts of the pollutant on panicles. For example, to
calculate the mercury mass balance for the combustion of coal, the amount of the mercury in the
coal is compared to the amounts of mercury in the fly ash, bottom ash, and stack dust. The
imbalance is explained as the amount of mercury leaving the stack in the gas phase.
Emission sources can also be characterized using the concentrations of a given pollutant
measured in ambient air at a receptor site. Measured concentrations are then compared with
emission source profiles for major source categories likely to contribute to pollution at the site. It
is clear that this method is useful when emissions from a given source region originate from one
dominant source or group of sources which have well defined emission profiles. A large electric
power plant in a region with small industry can be a good example of such a case. In the case of
complex source region more statistical work is needed to specify the contribution of various
sources to a profile which can be then compared with a profile obtained from ambient
measurements. Principal component analysis (PCA) is often used in such case, where the
contributions from several sources can be de-coupled in the ambient data. PCA has been used to
define source profiles in cases where there are no source profiles available from source sampling
(Tuncele/a/., 1985).
An important part of source characterization is to assess how accurate are the methods
used to prepare source profiles are and what are the means of verification of source data.
Accuracy of source characterization in terms of emissions depends on whether the
characterization has been made on the basis of measurements or estimates. The methods based on
measurements are considered as more accurate than those using assumptions and calculations,
e.g. methods based on emission factors or chemical mass balances. Unfortunately, most of the
verification procedures focus on activity data (statistical information) and emission factors.
3.2. Source apportionment techniques
Historically, dispersion models have been the traditional work horse for calculating
source-receptor relationships for air pollutants. These models require detailed emissions
inventories for various sources for the pollutants of interest, e.g. TSP, SO2, etc. Even if the
dispersion models were accurate it is very unlikely the source emissions inventories for the
pollutants of interest would be adequate. Emissions inventories for the criteria pollutants have
many short-comings, as discussed earlier in this document, and these inadequacies are even more
severe for hazardous pollutants or for pollutants which have large contributions from fugitive
process emissions, natural sources, and dusts. The limitations of the dispersion oriented methods
have led to the development of receptor oriented models. Receptor models assess contributions
-------
from various sources based on observations at sampling or receptor sites. Gordon (1980, 1888)
reviews the development of receptor methods and provides a concise overview of the various
types of receptor models and their applications. Here, we will briefly describe a few of these
techniques and discuss how they can be applied along with dispersion models to define source-
receptor relationships for hazardous air pollutants.
Receptor models were developed in the early 1970s in an attempt to identify the source(s)
of paniculate matter in large urban areas and to quantify the amount of paniculate matter emitted
from the source(s) (Miller et a/., 1972; Gordon, 1980; Winchester and Nifong, 1971). The
chemical element balance, or chemical mass balance (CMB) method as it is now referred, are
based upon the premise that the emissions characteristics, in terms of chemical and elemental
composition as well as physical size and morphology, of various source types are different enough
that one can identify their contributions by measuring the characteristics in samples collected at a
receptor site. Thus, an important first step in the application of CMB model to apportion the
sources of air pollutants measured in a specific urban area is to define an emissions inventory of
the number and source types of the important sources of air pollution. These models assume that
the composition of all contributing source types are known. This is often not the case either
because the sources are not easily sampled or because the source classes have widely varying
compositions (Henry, 1991). This is an important limitation of the CMB method in that the lack
of specific source profile information for the pollutants of interest prevent this approach from
being applied. Emissions data have been sparse in the past, particularly for SOCs, but the
situation is improving.
While the CMB method has primarily applied to urban scale data, Rahn and Lowenthal
(1984, 1985) also applied this technique to their "regional signatures" to apportion the sulfate and
trace metals observed on paniculate. The application of receptor models to regional and global
scale problems has been controversial and has yet to be developed, in some peoples opinions, to
the level necessary for it to be thought of as definitive in nature. However, an independent
verification for the appropriateness of the trace element ratio approach was performed and
indicated that this technique can be quite powerful (Keeler, 1987; Keeler and Samson, 1989).
Several methods are currently available to assess sources and source regions for various
air pollutants on the basis of the data on the chemical composition of the air at a given receptor.
Both the statistical methods and modeling are used together with meteorological data in order to
obtain this assessment. The origin to the pollution measured at a given receptor can be studied
using information on the chemical composition of aerosols and/or mixture of gaseous pollutants.
-------
3.3 Applications of source apportionment techniques for metals
Statistical methods have been developed which use information on the chemical
composition of aerosols to study contribution of sources or even source regions to the
contamination at a given receptor. The applicability of multivariate techniques for resolving
sources and source regions for aerosols measured at several remote locations far from major
emission regions has been tested with the use of the absolute principal component analysis
(APCA) and the chemical mass balance (CMB) methods. The APCA method determines the
composition of the major source components, such as coal combustion related, crustal, or sea-salt
related which contributed to the measured concentrations at the receptors. In the past, the APCA
method was applied to total suspended paniculate concentrations (particles measured in both the
fine and coarse fractions). APCA has been utilized to study the origin of the Arctic aerosol. The
results of these studies can be summarized as follows. The anthropogenic component contained
many toxic compounds, however, some crustal material was also found with this component. A
second component containing the crustal elements was observed with proportions similar to those
found in the average crustal rock. However, most of the elements in the soil component were
highly enriched and this fact suggests that the second component also contained some material
from anthropogenic sources. The composition of the third component found in the Arctic aerosol
was fairly similar to that of bulk seawater, indicating that this component is essentially sea-salt.
The three-component solution for the Arctic winter aerosol was confirmed by several studies
(Barriee/a/., 1992).
Further improvement of this receptor modeling method was obtained by applying APCA
to aerosol elemental concentration measurements in separate particle size fractions (e.g. Li and
Winchester, 199C) • - .- -esalts of this application of APCA gives the basis for interpretation of
coupled chemical reactions and physical processes in remote locations, as well as giving
information concerning atmospheric aging processes and, therefore, the history of the aerosols.
For example, it was noticed that some crustal particle components were found in all size fractions.
Carbonaceous fuel combustion pollutants were indicated by the presence of Si and Cl and the
absence of Al in all size fractions, and they were usually rich in sulfur. The combustion of coal
with high ash content may release volatile SiO from reduction of silicon dioxide by carbon, which
then forms a fine aerosol after SiO oxidation back to silicon dioxide. Since Al is non-volatile
during coal combustion its absence when Si and other metals were present indicates the aerosol
was generated from burning high ash content coal.
-------
Principal component analysis (PCA) has also been applied to assess the origin of the
remote aerosols on the basis of results from automated microanalysis of individual particles. For
example, Anderson el al. (1992) employed a modified scanning electron microscope (SEM)
analysis of particles to study the origin of the Arctic aerosol. They concluded that the collected
aerosol contained a mixture of altered and unaltered particles types from a variety of sources,
such as coal combustion, sea salt, and crustal dust. The silicate types appeared as relatively
unaltered particles, as well as particles with sulfate coatings, and also as particles that have
reacted with Br. Most silicate particles are probably crustal in origin. Many compositional types
of metal-rich particles were of anthropogenic origin, and most types had temporal variation
patterns that are individually distinct. Other major particle types were of marine origin, but
extensive fractionation and reactions of the marine aerosol components was suggested. The
authors concluded that further research is needed on the nature and timing of these reactions, the
mechanism for fractionation of the marine aerosol, and the sources of some particle types. It was
also underlined that the complexity of the remote aerosols during a period presumed to be
relatively free of pollutants is striking. Two important findings were that (1) the Arctic aerosol
has pollution products from human activity even in a normal period of spring; and (2) many of the
apparent pollutant particles in the fine fraction, S-rich species and perhaps Br-rich species, are of
natural origins.
Single or Individual Particle Analysis (IPA) has been applied in aerosol research to
investigate the sources and morphology of the collected atmospheric paniculate matter (Dzubay
and Mamane, 1989; Mamane, 1990; Sheridan, 1989). In the review by Sheridan (1989) he
observed that particles emitted by anthropogenic sources, such as carbon soot and coal
combustion spheres, occurred simultaneously with the highest concentrations of H2SO4 droplets.
Mamane (1990) utilized scanning electron microscopy (SEM) to estimate the contribution of
refuse incinerators to Philadelphia. Thus, EPA can be used to estimate the source apportionment
but also physical and chemical processes occurring during the long range transport of air
pollutants. Single particle analysis provides critical size distribution information that can be
directly used to calculate the deposition of pollutants as a function of size.
One important limitation of the PCA methods is, however, that their results do not allow
one to obtain a fine resolution of the contributions from various distant source regions to the
chemical composition 'of the remote aerosol. To attempt this resolution, CMB source
apportionment must be performed using either a set of emission source profiles or a set of
elemental signatures. The emission source profiles will be discussed later in this report. One set
of elemental signatures were developed by Rahn and Lowenthal (e.g. Lowenthal and Rahn, 1985).
-------
The CNffl source apportionment proved to be a good technique to assess contributions from
major source regions but was often limited to provide fine resolution of the contributions from
more defined areas due to high collinearity of signatures or source profiles used. For example, the
use of three European signatures in the Rahn and Lowenthal studies was supposed to provide a
fine resolution of the contributions from various parts of Europe to the contamination of the
Arctic aerosol. This attempt failed due to high collinearity of the European signatures.
Therefore, both the PCA and CMB methods are with limitations when comes to fine resolutions
of emission source regions contributing to the remote aerosol.
The origins of aerosols in remote regions and their source apportionment have been
diagnosed with the help of not only heavy metals but also of isotopes, halogens, graphitic carbon,
and organic compounds. Stable lead isotope ratios have been used to assess the contribution of
emissions from various source categories, anthropogenic vs. natural sources, or to distinguish
various sources within the same source category, e.g. combustion of gasoline with lead additives
from various manufacturing plants, to the lead concentration or atmospheric deposition at a given
receptor (Sturges and Barrie, 1989; Graney el a/., 1992). The ratios between the lead isotopes in
aerosols in the United States and Western Canada were found to be much higher than those in the
Arctic aerosol, while for Europe they were lower. The ratios in Eastern Canada were similar to
the Arctic data, but meteorology argues against this region being a major contributor to the Arctic
air pollution. The stable isotope analysis was taken a step further by Graney and colleagues in
that they linked sediments from three Great Lakes to air concentrations in the basin to ascertain
the extent of regional anthropogenic lead pollutiion, and to investigate the extent to which the
sediment cores could be used as indicators of historical atmospheric deposition. One limitation of
the above described method is that it requires a detailed information on the isotope ratios at
emission sources, which is not always available. In addition, the isotope ratios at emission
sources are often not sufficiently different to permit the use of multivariate statistical models to
resolve the input ratios.
Concentrations of chlorine, bromine, and iodine have been used to assess the contribution
of emissions from marine, automotive, and crustal sources to the contamination of the air at a
given receptor. Other sources, such as coal combustion have been also identified using the
halogen concentrations and their ratios. However, an important limitation of this method is that
halogens are quite reactive in the polluted troposphere and their behavior during long range
transport may be affected by the chemistry of the atmosphere. Therefore, the application of
halogens as source tracers is probably limited to pristine regions, such as the Arctic rather than in
the polluted regions, such as the Great Waters area.
10
-------
High concentrations of graphitic carbon panicles have been used to estimate the
contribution of combustion emissions at several receptors (eg. Rosen and Hansen, 1984)
Further work on this subject has been focused on assessing the vertical distribution of graphitic
carbon particles and their associated absorption coefficients. Although this research is very
important for modeling the effects of air contamination on the solar radiation balance, it seems to
be less useful in assessing the origin of pollutants at receptors.
A number of receptor-oriented transport models have been developed to identify the
emission source regions contributing to the contamination of certain areas by trace metals. In
general, the models have proven useful in calculating trace element concentrations at various
remote receptors and trace element inputs. A Lagrangian-type of model was employed to
determine the origin of air pollutants measured during the transport episodes to the Arctic (e.g.
Pacyna el al, 1985) as well as to the remote locations in Scandinavia (e.g. Pacyna el a/., 1989).
It was concluded that concentrations calculated by models agreed with measurements within a
factor of two. Of course, the model results depend on the quality of emission inventories used as
an input data. The model performance is also sensitive to the dry and wet deposition processes.
Fixed values of the dry depostion velocities are often used in models. This is certainly a
simplification of the problem raising the inaccuracy of the modeling. The mixing height also
affects the performance of the model, but far less than the emission estimates and the wet and dry
deposition processes.
The variational formulation of Eulerian dispersion models also allows for the application
of both source-oriented and receptor-oriented modeling as complementary tools in identification
of sources contributing to the contamination of the environment in a given region. Uliasz and
Pielke (1990) concluded that applicability of the receptor oriented option is limited to linear
dispersion models and integrals describing air quality at the receptor. It is necessary to assume
that all chemical reactions of pollutants are linear and that pollutants do not affect the atmospheric
dynamics. Uliasz and Pielke (1990) suggest, however, that their variational approach may be still
useful to perform sensitivity analysis of dispersion models with nonlinear chemistry.
An improved climatological-type model on Trace Toxic Air Concentrations in Europe
(TRACE) has been developed by Alcamo el al. (1993). The model addresses some of the
drawbacks to typical climatological-type air pollution models by (1) computing time of travel
from an empirical function of geographic distance; (2) maintaining mass conservation by splitting
the computation of decay coefficients spatially, and deriving concentration equations from mass
considerations, (3) dividing calculations into two steps and calculating deposition based on local
meteorological variables, thereby avoiding unreasonably "smooth" spatial deposition patterns; and
11
-------
(4) deriving all parameters objectively, that is, without calibration to observations. TRACE has
been used to compute levels of heavy metals for 1978 to 1985 throughout Europ? Calculations
agreed with As and Pb observations within a factor of two, and underestimated Cd and Zn
observations. Using the model it was estimated that wet deposition exceeds dry deposition in
most of Central Europe. The mean residence time of the mass of heavy metals in Europe's lower
atmosphere was estimated between 64 and 96 hours.
3.4 Problems in the application of source apportionment techniques for organic
compounds: changes in chemical profiles from source to receptor
It might seem that if source chemical profiles were known, the same apportionment could
be applied to atmospheric deposition. However the problem is much more complex. Because of
their reactivity and exchange between different phases in the atmosphere (gas-to-particle
distribution) organic compounds behave less conservatively than elemental tracers. Thus, changes
in the relative abundance of individual compounds occur in transit from source to receptor due to
differential reactivity and rates of atmospheric deposition. The limitations imposed by alteration
of profiles are not well understood. This section gives an overview of factors responsible for
profile changes; problems with individual compounds or compound classes are dealt with in the
appropriate section.
Gaseous and paniculate organic compounds can be transformed by photolysis, and are
also more or less reactive toward a number of atmospheric species, including radicals (e.g.
hydroxyl, peroxyl), ozone, and nitrogen oxides (Atkinson, 1990; Ballschmiter, 1991; Bunce and
Nakai, 1989; Bunce et a/., 1989). Differential reactivity is a factor which limits source
identifications based on chemical profiles obtained at a distance. Calculated atmospheric lifetimes
for gas-phase reactions of several VOCs and SOCs with OH radicals range from a few hours to
nearly half a year (Table 1), and thus ratios of compounds change in traveling from source to
receptor. Scheff and Wadden (1991) considered the effect of reactivity on CMB models for
VOCs and concluded that for most species a 2-3 hour transit time from source to receptor would
result in little change in composition ratios. Nevertheless, diurnal cycles of several gaseous 2- and
3-ring PAHs were observed in Glendora, California during a photochemical air pollution episode
(Arey et a/., 1989). Daytime concentrations were about 2-3 times lower than at night due to
reaction with the higher Jevels of OH radicals during the day. Elevated nighttime concentrations
of nitronaphthalenes were found as a consequence of naphthalene reaction with ^05.
Transformation of paniculate PAHs may also be a problem, and is discussed in that section.
12
-------
TABLE 1 Estimated Atmospheric Lifetimes Due to Gas-Phase Reaction with OH Radicals.
Compound Lifetime Compound Lifetime
Aliphatics
Alkanes, €4 - C \ 3
Aromatics
Benzene
Toluene
Xylenes
Trimethylbenzenes
Ethylbenzene
Ethyltoluenes
Naphthalene
Biphenyl
Fluorene
Anthracene
Pyrene
Chlorinated VOCs
Methyl chloride
Dichloromethane
Chloroform
1 ,2-Dichloroethene
Trichloroethene
Tetrachloroethene
0.5-6 days
5 - 13 days
1-2.5 days
6-10 hours
4 hours
20 hours
7-11 hours
8 hours
2 - 3 days
1.2 days
1.4 hours
4 hours
1 year
110 days
1 50 days
70 days
6 days
90 days
PCBs
Monochloro
DichJoro
Trichloro
Tetrachloro
Pentachloro
PCDDs & PCDFs
2,3,7,8-TCDD
2,3,7,8-TCDF
Pesticides
DDT, DDE
Dieldrin
Chlordane
Hexachlorobenzene
Hexachlorocyclo-
hexane
EPTC
Cycloate
5-11 days
8- 17 days
14-30 days
25-60 days
60- 120 days
2-3 days
7 days
2-4 days
1 day
8 days
80 days
15 days
5.8 hours
5.2 hours
Sources: Altschuller et al., 1991; Atkinson (1987), Bidleman et al. (1990); Kwok et al. (1992).
13
-------
SOCs. differ from VOCs in that they are associated to a greater or lesser extent with
atmospheric particles. This phase distribution profoundly affects their atmospheric chemistry and
physical removal, and thus deposition into the Great Waters (Ballschmiter, 1991, Bidleman, 1988;
Mackay et a/., 1986; Schroeder and Lane, 1988; Swackhamer and Eisenreich, 1991). Factors
influencing the extent of particle association include compound volatility (vapor pressure),
temperature, and particle surface area available for adsorption. The Junge-Pankow equation
(Pankow, 1987) is the most common model for estimating the phase distribution of SOCs in
ambient air:
is the fraction of the total atmospheric concentration associated with aerosols,
P T is the liquid-phase vapor pressure of the compound, and c is a parameter which depends on
thermodynamic properties of the compound and surface properties of the aerosol. The parameter
0 is the available particle surface area per unit volume of air (cm2/cm3); typical values for urban,
rural, and clean background air are given by Bidleman (1988).
Experimental estimates of the phase distribution of SOCs in air are usually made by
drawing air through a filter to retain particles followed by an adsorbent trap to collect gaseous
SOCs. Such samplers are prone to a number of artifacts which may cause the measured
paniculate fraction on filters to differ from the phase distribution in the free atmosphere (Pankow
and Bidleman, 1992). Evaluation of these artifacts (Gotham and Bidleman, 1992; deRaat et a/.,
1990; McDow and Huntzicker, 1990; Zhang and McMurray, 1991) and development of improved
techniques to speciate SOCs (e.g. denuder samplers, Appel et a/., 1989; Coutant et a/., 1992.
1989; Krieger and Hites. 1992; Lane el al., 1988) is an active area of research.
Paniculate fractions measured with filter-adsorbent samplers and predicted by the Junge-
Pankow model have been made in a few cities and rural areas. These comparisons, summarized
by Bidleman (1991), and Pankow and Bidleman (1992) show good agreement between the model
and experimental results for PAHs, whereas the sampling methods yield O-values that are lower
by about a factor or two for organochlorine pesticides, PCBs, PCDDs, and PCDFs (Figure 1).
The gas-to-particje distribution of SOCs governs the mechanism and rate of their removal
from the atmosphere. Deposition of SOCs can occur by wet and dry removal of particles, and gas
exchange with soil, plant foliage, and water bodies (Figure 2). An understanding of gas-particle
relationships in the atmosphere is therefore necessary to predict losses of SOCs during transport
14
-------
PAHs
Figure 1. Paniculate percentages of PAHs and organochlorines (PCBs, pesticides.
PCDDFs) in urban air, as determined by high volume air sampling (solid
curves) and-predicted by the Junge-Pankow model (Equation 1) as a function
of the liquid-phase vapor pressure (p°L, Pa) of the chemical. Letters indicate
individual investigations. Source: Gotham (1990) and Bidleman (1991).
15
-------
AERIAL DISTRIBUTION AND REMOVAL OF SOC
o *° • o °
VAPORS
PARTICLES
DRY DEPOSITION
(GAS EXCHANGE)
RAIN/SNOW
SCAVENGING
RAIN/SNOW
SCAVENGING
DRY
DEPOSITION
Figure 2. Gas-particle partitioning and aerial removal processes for SOCs. Source:
Bidleman (1988, 1991).
16
-------
and deposition. An equilibrium model describing the relationship between wet scavenging and the
fraction of particle-bound SOCs is (Ligocki and Pankow, 1985)
W = WpO + Wg(l-(J>) (2)
W is the overall scavenging ratio (concentration in rain divided by concentration in air, on a
volume-volume basis), Wp and Wg are scavenging ratios for paniculate and gaseous SOCs. Wp
is obtained from field observations for species only in the paniculate phase (such as certain trace
metals). Wg is the equilibrium water/air partition coefficient, calculated from RT/H where H is
the Henry's law constant (Pa-m3/mol) for the particular species and R is the gas constant (8.3 Pa-
m3/deg-mol). Thus the relative importance of particle and gas scavenging for an individual
compound depends on its phase distribution (O) and the Henry's law constant. These must be
known as a function of temperature to estimate atmospheric removal rates.
The actual situation is more complex, because of the variation in particle size distribution,
re-equilibration of SOCs between the gas and particle phases during rain events, emission of
SOCs into the atmosphere during rain events, and differences in meteorological conditions.
Recent kinetic models have been formulated to take these factors into account (Tsai et al., 1991;
Seinfeld et al., 1991); however these require a much more extensive list of input variables,
including the particle size distribution of the SOCs (which is often not available).
In rural air at moderate temperatures the simple Equation 2 model predicts that gas
scavenging dominates for 2-3 ring PAHs, hexachlorocyclohexanes (HCHs), and dieldrin. Particle
scavenging is more important for higher-ring PAHs, PCBs, n-alkanes, chlordane, and DDT
(Bidleman, 1988, 1991; Swackhamer and Eisenreich, 1991; Figure 3). Rainfall removal of
PCDDs and PCDFs changes from a gas to a particle-dominated process as molecular weight
increases (Eitzer and Kites. 1989b; Koester and Kites, 1992).
Deposition of SOCs is a highly temperature-dependent process. Lowering of compound
vapor pressure in winter increases the paniculate fraction according to Equation 1. This can be
clearly seen in Figure 4, where proportions of PAHs in the gas and particle phases during winter
and summer are compared. Henry's law constants also decrease at lower temperatures, resulting
in greater gas solubility in rain and cloud droplets. The overall effect is more efficient deposition
of SOCs in winter than in summer.
Differential removal by physical processes is one reason why profiles of SOCs at receptor
sites differ from those at sources. The case of PCDDs and PCDFs is a good example. Profiles of
PCDDs and PCDFs from municipal incinerators and industrial waste effluents (Czuczwa and
17
-------
1500C;
iZOOCr
QOOOh
800C
3000r
GflS-SCRVENGED SOC
C PflRTICLE
-------
A. Winter
100
D Gaseous
B Paniculate
PHEN WVWFUJR PVR BePM CcdP B«A CHPY BNTH B«P BbF BkF BaP Ban* Bjfvo -23P AN'HN COR
B: Summer
100
O Gaseous
D Paniculate
PMEN ANTVW FuuR PVP BcPM CeaP 6aA
B" -je-'. SB-xA 5;-,o -13P A\--.N CCR
Figure 4. Relative proportions of PAHs in the gaseous and paniculate phases during
winter and summer, as determined by high volume air sampling. Source: Back
eial (1992).
19
-------
Hites, 1986) and ambient air (Broman el a/., 199la; Eitzer and Hites, 1989a,b; Smith et a/., 1990)
show the presence of compounds containing 4-8 chlorines. Because paniculate fractions are
greater for the more highly chlorinated homologs, rain and dry flux distributions are weighted
toward compounds containing 7-8 chlorines (Eitzer and Hites, 1989b; Koester and Hites, 1992).
Atmospheric degradation of gas-phase PCDDs and PCDFs may also contribute to profile
alteration (Koester and Hites, 1992). The result is that PCDDs and PCDFs in Great Lakes
sediments are dominated by 7- and 8-chlorinated compounds (Czuczwa and Hites, 1986; Eitzer
and Hites, 1989b; Koestc: . 1 Hites, 1992; Figure 5).
3.4.1 Alteration of PAHs in the Environment
The problem of selective physical removal of SOCs from the atmosphere and the effect on
chemical profiles has been discussed earlier. In addition, PAHs can undergo chemical and
photochemical reactions in the atmosphere than lead to changes in the relative proportion of
compounds between source and receptor.
Many studies have been done to determine reactivities of PAHs in the gas and particle
phases, with sometimes contradictory results Light, mixtures of oxidant gases, and the particle
composition all influence the rate of PAH degradation (Daisey et al., 1986; Greenburg et al,
1985). The latter authors found good correspondence between experimental PAH stabilities anc
those predicted from molecular properties.
Laboratory studies show that PAHs are more stable to photolysis when adsorbed to fly
ash than on artificial substrates like alumina, silica gel, or glass (Yokley et a/., 1986). Among the
fly ashes, those with high carbon content and black or gray were most effective in stabilizing
adsorbed PAHs (Behymer and Hites, 1988; Dunstan et a/, 1989; Yokley et al., 1986).
Kamens et al. (1988-90) found rapid degradation of PAHs on wood smoke particles when
exposed to natural sunlight. A. moderate temperatures and humidities, PAHs decayed within an
hour. At low temperature and humidity and reduced light intensity the time scale for PAH loss
increased to days. Oxygenated PAHs were stable to sunlight alone, but labile in the presence of
0.2 parts-per-million (ppm) ozone and sunlight. Sunlight was the most important factor causing
loss of benzo(a)pyrene (BaP) from wood soot, followed by ozone and nitrogen dioxide. Guo and
Kamens (1991) estimated a BaP half-life of 80 h on wood smoke particles for reaction with 02
parts-per-million atmospheric NC»2 PAHs on diesel paniculate matter were converted by
exposure to part-per-million levels of ozone over a few hours (Van Vaeck and Van
Cauwenberghe, 1984).
20
-------
590 ff/
AVKKAGE AIR
F4 F5 F6 F7 F8 D4 D5 D6 D7 D8
54 p,/L
J i
AVEHAGE RAIN
F4 F5 F6 F7 F8 D4 D5 D6 D7 D8
1100
GREAT UKES SEDIMENT
F4 F5 F6 F7 F8 D4 D5 D6 D7 D8
Figure 5. Homolog profiles of PCDFs (F4 - F8) and PCDDs (D4 - D8) containing 4-8
chlorines in air and rain from Bloomington, Indiana, and in Great Lakes
sediments. Source: Eitzer and Kites (1989b).
21
-------
Losses of parent compounds does not necessarily represent a gain for the environment.
Indeed, reaction of PAHs with atmospheric oxidants leads to formation of PAH <..;.:r>ones and
PAHs substituted with hydroxy- and nitro-groups, some of which are highly mutagemc (Arey et
a/., 1987; Kamens et a/., 1985; Kleindienst et a/., 1986; Nishioka et al., 1988; Nishioka and
Lewtas, 1992; RamdahJ etal., 1986).
A different picture of PAH reactivity is seen from field studies. Freeman and Cattell
(1990) found that although diurnal variations in total PAH occurred in urban air, ratios among the
PAHs were fairly constant. Gibson et al. (1986) examined ratios of BaP to lead and elemental
carbon (EC) in coastal Delaware and Bermuda. The expected reduction in BaP following long-
range transport did not occur; BaP/Pb and BaP/EC were approximately the same in the two
locations. However the BaP/EC ratio in Detroit was much higher than in Delaware or Bermuda.
To explain the difference between the inner city and Delaware/Bermuda ratios, the authors
suggested that BaP is rapidly reduced to a low level as paniculate matter ages, after which further
losses are slow. Evidence of atmospheric transformations of other PAHs during transport were
found from elevated proportions of nitro- and hydroxynitropyrenes to inert markers in Bermuda
compared to Delaware. The ratio of reactive/unr.eactive PAHs on aerosols was about the same in
stationary southern Norway air compared to air transported from the United Kingdom and France
(Bjorseth et a/., 1979).
A method for correcting PAH profiles for degradation and/or physical removal effects
during transit was suggested by Masclet (1986). A "relative decay index" (RDI) was established
based on the diurnal variation of each PAH concentration in urban air. This RDI concept was
used in a CMB model for PAHs in Paris air (Pistikopoulos et al, 1990). PAH profiles in ambient
air were matched against profiles of six 5-6 ring PAHs from three sources: spark-ignition and
diesel engines, and domestic heating. Losses of PAHs due to differential reactivity were
accounted for by summer and winter RDIs. Spark-ignition vehicles accounted for 40-70% of
these PAHs, and diesels for 20-40%. The contribution from domestic heating rose from a few
percent in summer to 20-40% in winter (Figure 6).
3.4.2 Alteration of PCDDFs in the Environment
Degradation of PCDDFs during transport has not been given the attention received by
PAHs. Gaseous species are photosensitive (Podoll et a/., 1986) and reactive toward hydroxyl
radicals (Table 1). PCDDFs in natural waters (Friesen et a/., 1990) and on fly ash (Tysklind and
Rappe, 1991) can be photolyzed in solution. Penta- and hexa-CDDs were photo-reactive when
exposed to natural sunlight in distilled water - acetonitrile solution, and the photolysis rate was
22
-------
Domestic heating
I 1 Diesel vehicles
Spark -ignition engines
'OOr-m
80
60
a E3 C
SL.
P
1
1
:h/b/8b
M»,
M VI 111
. I'.VH '
Figure 6. Contributions to total PAHs in Paris air, calculated by CMB.
Pistikopoulos el a/. (1990).
Source:
23
-------
markedly increased in natural water -acetonitrile. Projected half-lives in sunlit surface waters
were 27-81 days. The sensitized photolysis may have been due to hurruc materials (Friesen et al.,
1990). Less chlorinated PCDDs are also photolyzed in water (Dunlin et al., 1986). Koester and
Hites (1992) noted that profiles of PCCDFs in urban and rural areas were different, and suggested
that a combination of phase distribution effects and gas-phase reactivity might be responsible.
3.5 Applications of source apportionment techniques for organics
3.5.1 Volatile organic compounds (VOCs)
Comparative ambient air concentrations and profiles of individual VOCs in major U.S.
cities, taken from a national database for VOCs (Shah and Heyerdahl, 1988), were presented by
Edgerton et al. (1989a). Similarities between aromatic hydrocarbon profiles in several cities and
their correspondence to an average auto exhaust profile suggested that auto exhaust was probably
the dominant source. Profiles of industrially related chlorinated and fluorinated hydrocarbons
showed distinct differences among the cities, depending on the source types.
Source profiles and CMB models have been used to apportion VOCs in several U.S. cities
and urban areas from around the world (Aronian et al., 1989; Doskey et al., 1992; Kenski et al.,
1991; Scheffe/ al, 1989; Scheff and Wadden, 1991; Sweet and Vermette, 1992). These papers
describe in great detail the characteristic "fingerprints" of different VOCs in sources such as
vehicle exhaust, gasoline vapors, industrial emissions (refineries, coke ovens, chemical plants),
architectural coatings (paints, thinners, and cleanup solvents), dry-cleaning, degreasing, and
graphic arts (printing). Examples of CMB apportionment of total hydrocarbons are given in
Table 2 for three cities
A check on the accuracy of CMB results is to compare them with independent emissions
inventories. This was done by Kenski et al. (1991) and Scheffe/ al. (1989) for Chicago, Detroit,
and Beaumont, Texas (see Table 2). In Detroit, CMB calculations and emissions inventories
agreed well for vehicles, gasoline vapor, coke ovens, and architectural coating contributions. The
emission from coke ovens was confirmed by close correlation between CMB modeling
coefficients and a source-based score derived from Gaussian plume dispersion models. Detroit
CMB results were high for refineries and graphic arts sources. In Chicago, agreement between
CMB and inventory results was good for all source categories except refineries, for which the
CMB model again gave high values. The opposite was the case in Beaumont, where inventoried
emissions from refineries and gasoline vapor exceeded those modeled by CMB. Very good
agreement between CMB and inventory was found for a polyethylene manufacturing plant.
24
-------
TABLE 2. Percent Contributions of Various Sources to Total Hydrocarbons in Three Cities:
Comparison of Emission Inventories with Chemical Mass Balance Results2
Source Detroit Chicago Beaumont
INV CMB INV CMB INV CMB
Vehicles 32.9 28.2 35.1 41.2 11.9 13.6
Gasoline Vapor 6.9 9.4 7.6 6.4 4.2 19.5
Point Sources
Refineries 0.7 16.5 1.3 13.5 23.1 9.1
Polyethylene Mfg. 7.3 7.0
Other Chem. Plants 35.6
Rubber Mfg. 13.4
Wood Pulp & Paper 1.3
Coke Ovens 2.0 3.7
Coatings
Architectural
Industrial
Graphic Arts
Other
3.8 2.5 5.5 6.2
14.0
0.7 4.7 9.8 11.0
38.8 34.5 40.6 16.5
0.6
1.4
a) Source: Kenski et al., 1991, Tables 6 and 7.
CMB models, factor analysis, and wind trajectories were used to investigate VOC sources
in southeast Chicago and East St. Louis by Sweet and Vermette (1992). The three techniques, if
applied appropriately, are complimentary in nature. Factor analysis requires no a priori
knowledge of sources and is useful for confirming the importance of known emissions or
suggesting sources not inventoried (Sweet and Vermette, 1992). In southeast Chicago, factor
25
-------
analysis indicated that four sources types accounted for 78.4% of the variance in the observed
data: The first and largest factor was attributed to urban air sources (perhaps a combination of
vehicle emissions, gasoline evaporation, and solvent emissions). A second factor identified only
with benzene was assigned to coke oven combustion. The third factor comprising carbon
tetrachloride and chloroform was attributed to a regional source outside the city. A fourth and
unknown factor was characterized by chlorobenzene.
CMB results showed that under average conditions most aromatics resulted from vehicle
exhaust. Trichloroethane, trichloroethylene, and tetrachloroethylene came from degreasing
solvents and dry-cleaning. In east St. Louis, a factor associated with several aromatics and light
chlorinated compounds was identified with a local chemical plant on the basis of wind trajectory
analysis.
Sweet and Vermette also showed how the situation in south-eastern Chicago can deviate
from the average during a pollution episode. During one event, 72% of the benzene and 17% of
the ethylbenzene were released by coke ovens (Table 3). A high proportion of benzene in coke
emissions was also found by Kenski el al. (1991). By apportioning total VOCs with assumed
factors, Blakley and Klevs (1990) assigned highest releases of benzene to counties at the south
end of the lake which had large coke emissions.
Pollution sources in the Kanawha River Valley, West Virginia were examined using factor
analysis of VOC and trace element data (Cohen et al., 1991a,b). The valley contains numerous
sources of VOCs, including eleven large industrial complexes that use, store, produce, and
transport organic chemicals. Only at one site were aromatic VOCs identified with an
"automobile" factor, which was defined by lead, bromine, and paniculate carbon. At other
locations in the valley, VOCs could not be apportioned to any one factor, but instead were
ascribed to a factor called "general VOCs". The "general VOC" factor differed little among sites
and was not associated with total particles or element-speciated paniculate matter.
The authors (1991b) discussed the difference between their source apportionment of
VOCs in Kanawha River Valley and the receptor modeling results of Stevens et al. (1989) in
Boise, Idaho. In the latter study several aromatic hydrocarbons showed excellent correlations to
fine paniculate lead (a gasoline source that is rapidly decreasing). These tracers were useful as a
lead replacement for estimating the contribution of vehicles to ambient concentrations of fine
paniculate extractable organic matter (Zweidinger et a/., 1990).
26
-------
TABLE 3. Contribution of Sources to Individual VOC Concentrations for a South Chicago
Sample During a Pollution Episode3
Benzene
Toluene
Ethyl-benzene
m,p-xylene
o-Xylene
1,1,1 -Tri-chloroethane
TrichJoro-ethylene
Tetrachloro-ethylene
Coke Vehicle
Ovens Exhaust
71.6 26.3
0.8 45.6
17.0 68.3
8.3 71.8
1.4 66.6
0 0
0 0
0 0
Paint
0.6
52.0
9.4
18.4
30.3
0
0
0
Gasoline
Vapor
2.3
1.8
5.3
1.8
1.4
0
0
0
Degreasing
Solvent
0
0
0
0
0
100
100
12.9
Dry
Clean.
0
0
0
0
0
0
0
87.1
a) Source: Sweet and Vermette, 1992; CMB results.
According to Cohen, the failure of VOCs to show a clear relationship to lead in Kanawha
Valley might have been because: a) there were similar VOC sources at all sites, b) there was good
mixing of pollutants in the valley, and/or c) the sources of VOCs were regional. These studies
point out the value of using elemental and paniculate information in addition to VOCs to speciate
source types.
Some of the longer-lived halocarbons (e.g. methylchloroform, tri- and tetrachloroethene)
have been used as "tracers of opportunity" to document pollutant transport out of the Los
Angeles basin to the Nevada - Arizona desert (Bastable et a/., 1990; Pryor and Hoffer, 1992;
Miller et a/., 19.90; White et a/., 1990). Methyl-chloroform has a distinctly weekly cycle, being
high during the week and dropping to near-baseline values on weekends. This cycle has been
27
-------
attributed to releases of methylchloroform from metal fabrication and electronics industries in the
basin (White el a!., 1990). Haze episodes leading to reduced visibility were correlated with
elevated concentrations of methylchloroform (Miller el al., 1990).
The concentrations of trace gases, such as chlorofluorocarbons, chJorocarbons and carbon
monoxide were also applied in the statistical methods to discuss the sources and their
contributions to the contamination at remote receptors (e.g. Khalil and Rasmussen, 1984). It was
concluded that the ratios between these gases can be used as global tracers due to their distinct
and different application pattern.
The concentration ratios of light hydrocarbons, such as ethane and propane to chlorinated
ethenes have been used as signatures of emissions from natural gas exploitation regions as
compared with emissions from regions with extended application of industrial solvents (e.g. Hov
el a/., 1984). It was indicated that the use of chlorinated ethenes is primarily confined to
industrialized countries.
3.5.2 Semivolatile orgamic compounds (SOCs)
A. Polycyclic Aromatic Hydrocarbons (PAHs)
Unlike VOCs, very little work has been done to relate PAHs in ambient air to their
sources through CMB or factor analysis modeling. This is because emission factors have been in
short supply, and even within source types the quantities and ratios of PAHs emitted have been
highly variable. One can see this from the examples of emission factors given in Johnson el al.,
(1992), and also from a review of PAH profiles by Daisey el al. (1986). These authors critically
examined literature data from the 1960s through early 1980s for PAHs in several source types.
All PAH concentrations were expressed relative to benzo(e)pyrene (BeP) because this compound
is fairly stable and is found almost exclusively in the paniculate phase. PAH profiles were
examined from auto exhaust (gasoline and diesel), residential coal and wood combustion, oil- and
coal-fired power plants, and industrial coal-fired boilers.
Interpretation of PAH data is plagued by uncertainties due to sampling and analytical
artifacts. PAHs occur in paniculate and gaseous forms, and atmospheric concentrations will be
inaccurate unless both species are collected. This has generally been done for source sampling,
but until the early 1980s most ambient air PAHs were collected with filters only and
concentrations of 2-4 ring compounds were seriously underestimated. Both the particle-and gas-
phase compounds are mutagenic (Tuominen el al., 1988; Westerholm el al., 1991; Lewis el al,
1988), providing another reason for collecting the two fractions PAHs can undergo reactions
28
-------
with oxidant gases during sampling, thus changing their proportions. Sample extraction and
analytical methods vary among laboratories and quality assurance information is not always
provided. According to the Daisey review, the three most critical factors limiting
intercomparability of the organic source emission profiles reviewed by them were: a) The
incompatibility of source emissions and ambient aerosol sampling methods and intervals, b) The
general lack of emission profiles which represent an average for a given source type, and c)
Differences in organic profiles due to variations among sampling and analytical methods rather
than in source emissions.
The conclusions of Daisey were that existing data provided some useful information with
respect to receptor modeling. PAH profiles from two source types that had been repeatedly
sampled and analyzed by the same investigator appeared to be fairly reproducible: coke ovens and
coal-fired boilers. Profiles from wood combustion varied widely and depended on combustion
conditions. Cyclopenta(cd)pyrene was especially enriched in exhaust from spark-ignition engines
and was suggested as a useful marker compound. Coronene and benzo(ghi)perylene were also
relatively high in vehicle exhaust compared to other sources. Other organics having potential as
marker compounds were retene and levoglucosan for wood combustion, and
benzo(b)naphtho(2,1 -d)-thiophene (BNT) for diesel engines and fuels containing sulfur.
Since the time of the literature covered in the Daisey review, sampling and analytical
methods for PAHs have improved greatly. Filters followed by adsorbent traps to catch gas-phase
PAHs are now in routine use for ambient air sampling. The availability of standard reference
materials for urban air paniculate matter and diesel particles has improved quality assurance (Wise
el a/., 1986; 1988). PAH emission factors and profiles in a large number of source types have
been compiled (Johnson el ai, 1992). Nevertheless, it is interesting to note that many of the
reservations expressed in the Daisey review about data quality and availability were echoed in this
1992 report.
The three PAHs cited as vehicle exhaust markers by Daisey et al. (1986) and Baek et al.
(1991a,b) are hardly unique. High levels of coronene were also produced by burning certain types
of vegetation (Freeman and Cattell, 1990). The Johnson survey found a high ratio of
benzo(ghi)perylene to BeP in wood stove emissions. Proportions of cyclopenta(cd)pyrene to BaP
in vehicle exhaust and industrial coal combustion effluents were similar, and the former compound
was elevated in wood smoke (Table 4). The Bghip/BeP ratio was much lower in wood soot than
in gasoline soot for profiles presented by Kang and Kamens (1992), in contrast to the information
given in the Johnson et al. report. Thus, ratios among the parent PAH compounds are not
foolproof source indicators.
29
-------
Daisey ei a/. (1986) felt that PAH data were best used in combination with other tracers
for source apportionment^ g. trace elements) Some information on PAHs and er?r organic
compounds that may serve as "marker compounds" of a particular source type or process is
discussed below and summarized in Table 4.
The proportion of alkyl homologs to unsubstituted compounds is an indicator of
combustion source temperature. Low temperature combustion yields soots that are more
abundant in alkylated PAHs, whereas unsubstituted (parent) PAHs predominate at high
temperatures (LaFlamme and Hites, 1978). Initially it was thought that the alkyl homolog
distribution in sediment cores could be used to distinguish "natural" (low-temperature) from
"anthropogenic" (high temperature) combustion sources. However LaFlamme and Hites pointed
out that coal combustion occurs at moderate temperatures and yields alkylated PAHs.
Furthermore, they felt that differential water solubility of alkylated and parent PAHs could alter
the distribution in sediments from what was initially deposited. Tan ei al. (1992) recently
reported that while wood burning produces mainly parent PAHs, low-intensity fires in forest floor
waste ("duff1) yields relatively high levels of alkylated phenanthrenes,
cyclopenta(def)phenanthrene, and dodecahydrochrysene.
The ratio of methylphenamhrenes to phenanthrene (MP/P) is low in combustion effluents,
but high in unburned petroleum products Takada et al. (1990) found that MP/P in auto exhausts
and asphalt were high relative to combustion products from a steam generator. Using this
information, along with sulfur heterocyclic compounds, the authors concluded that PAHs in
Tokyo street dust were strongly affected by automobile exhausts. Dusts from residential areas
had a somewhat greater contribution from stationary source combustion products. The authors
pointed out that the MP/P ratio in auto exhaust is highly variable, depending on engine load and
cylinder exhaust temperature, and suggested that extensive collection of auto exhausts should be
conducted. Runoff samples from urban and coastal South Carolina were depleted in MP relative
to crankcase oils and diese, fuel, and showed PAH profiles similar to the atmospheric particles.
The conclusion was that PAHs in street dust came mainly from atmospheric deposition and not
from dripping crankcase oils (Ngabe, 1992).
Alkylphenanthrenes were also high in samples from the Baltimore Harbor Tunnel (Benner
et a/., 1989). A comparison of tunnel air samples to standard reference diesel and urban air
paniculate matter showed that tunnel and diesel particles were enriched in MP Factor analysis
suggested that contributions from diesel and gasoline-powered vehicles might be separated by
alkvlated PAH content
30
-------
TABLE 4. Marker Information for Identifying Sources of PAHs and Other Organic Compounds.
Source
Spark-source engines
Diesel engines
Motor vehicles
High vs. low temperature
combustion &
combustion vs. unbumed
petroleum products
Wood combustion
Wood combustion
Vegetation fires
Vegetation fires
and other biogenic
sources, biogenic
vs. anthropogenic
hydrocarbons unresolved
complex mixture
Marker Compounds
cyclopenta(cd)pyrene
benzo(b)naphtho(2,1 -d)-
thiophene
benzo(ghi)perylene,
coronene
parent/alkylated PAH
ratios
methoxyphenols & other lignin
pyrolysis prod.
retene, levoglucosan
carbon-14, potassium
alkylated phenanthrenes
cyclopenta(def)phenanthrene
dodecahydrochrysene
carbon preference index
(CPI: odd for n-alkanes &
alkanones, even for acids
& alcohols), presence of UCM
for (petroleum products),
biogenic markers:
retene, di- & tri-terpen
phytosterol
Tire wear
Benzthiazoles
Examples
Daisey era/., 1986.
Daisey era/., 1986;
Alsbergera/., 1989.
Daisey era/., 1986
Alsbergera/., 1989.
LaFlamme & Hites,1978;
Takadaera/., 1991;
Ngabe, 1992.
Hawthorne et a/., 1988, 1989;
Edye & Richards, 1991.
Daisey era/., 1986
Tan era/., 1992.
Standley & Simoneit, 1987,
1990.
Simoneit, 1989;
Mazurek & Simoneit, 1984;
Kawamura & Leuenberger et al.
1988;
Gagosian & Pelzer, 1986;
Farmer & Wade, 1988;
Foreman &Bidleman, 1990
Greaves era/.,1987.
Spies era/., 1987.
31
-------
As indicated by Daisey et al. (1986) other marker compounds may be used to advantage
in differentiating PAH source contributions. BNT was used as a diesel exh. -st indicator by
Alsberg et al. (1989), who employed principal component analysis to examine profiles of PAHs
and aromatic VOCs collected in Gothenburg, Sweden. Diesel and gasoline emissions were
separated by enrichment of BNT and particle-associated light PAHs in the former, and aromatic
VOCs and heavy PAHs (particularly benzo(ghi)perylene and coronene) in the latter. Cautions
about using these two PAHs have been mentioned earlier.
Methoxylated phenols have been suggested as tracers for wood smoke pollution, which is
a major contributor of PAHs (Hawthorne et a/., 1988, 1989). More than 70 species, arising from
pyrolysis of lignin, have been identified in soot from residential wood stoves. Guaiacols in
hardwood and softwood soots were nearly the same, whereas syringols were much higher in the
hardwood soots. It was suggested that guaiacols could be used as markers for wood combustion
in general and syringols could serve to differentiate wood type.
Simoneit (1989) reviewed the wide variety of biomarker compounds available for use in
source reconciliation. These include n-alkanes and similar compounds (acids, alcohols, ketones),
phytosterols, terpenoids, and terpenols. Some of these may be useful for investigating regional
sources in long-range transport. For example, the signatures of C27-29 phytosterols were different
in aerosols from the western U.S., Nigeria, and southeastern Australia. Further work is needed to
evaluate the stability of marker compounds in the atmosphere and during sampling.
Air samples from slash-burns along the Oregon coast were analyzed for a wide variety of
plant wax components and terpenoids (Standley and Simoneit, 1987). PAHs accounted for only a
minor part of the hexane-soluble material, most of which was plant waxes, resins, and thermally
matured compounds. Smoke from the burns contained straight-chain homologs with a strong
plant wax signature: n-alkanes and n-alkanones showed an odd-carbon preference, peaking at
€27. Straight-chain alkanoic acids and alcohols showed an even-carbon preference. The acids
peaked at C22, C24> or ^30> whereas the maximum for the alcohols occurred at C22 A number of
di- and triterpenoids, as well as retene, were suggested as molecular markers for this type of bum.
In a subsequent study (Standley and Simoneit, 1990), polar cyclic di- and triterpenoids were
analyzed in extracts of residential wood combustion aerosols. Distinct signatures were found that
could be used to trace the input from coniferous, alder, and oak combustion products.
Profiles of organic compounds in atmospheric paniculate matter collected in Colorado
was examined by Greaves et al. (1987) using factor analysis. One factor associated with ozone
and oxygenated compounds such as acids, furans, aldehydes, ketones, and lactones was
32
-------
interpreted as arising mainly from photochemical processes. Factors 2 and 3 which contained
terpenoids and odd-chain hydrocarbons were ascribed to biogenic sources. This hypothesis was
strengthened by noting that factor 2 was strongly dependent on wind direction, being largest when
the wind was from a national forest to the west. Factor 4, associated with CO and several long-
chain alkanes, was identified with motor vehicles.
A proposed method for source characterization was based on the high-resolution gas
chromatographic profile of neutral-fraction components of fine organic aerosol (particle diameter
< 2 jim) (Hildemann et a/., 199la). A large number of source types were examined: wood
combustion; automobile and truck exhausts; fuel oils; natural gas appliances; vegetation detritus;
tire, brake, and road dusts; cigarette smoke; roofing tar; and charbroiled meats. The similarities
or differences among these aerosol signatures were determined by hierarchical cluster analysis.
Tire wear can contribute PAHs to road dust, since carbon black is used in the rubber
(Voldner and Smith, 1989). Spies et al. (1987) suggested using benzthiazoles, found in estuarine
sediments of San Francisco Bay, as indicators of street runoff. The benzthiazoles are breakdown
products of antioxidants added to tire rubber.
B. PCDDFs
A considerable amount of work has been done to determine differences in congener
profiles from various sources and relate them to environmental matrices. The use of principal
component analysis (PCA) and other multivariate techniques seems to be more frequent than in
PAH investigations.
Czuczwa and Hites (1986) applied PCA to differentiate sources of PCDDFs in Great
Lakes sediments. Profiles of PCDDF homologs (4-8 chlorines) were compared for sediments,
paniculate ambient air samples, incinerator combustion products, and different PCP products.
Sediments from all lakes except Ontario formed a tight cluster with ambient air particles that was
separate and distinct from the incinerator or PCP patterns. Clustering of a Lake Ontario sediment
core with the PCP data pointed to contamination from suspected sources on the Niagara River.
Changes in the relative proportions of PCDDFs in moving from source to receptor were
investigated by Eitzer and Hites (1989b) using PCA. Homolog profiles from different urban
combustion sources were widely scattered and showed no grouping tendency. Urban and
industrial air samples clustered in the middle of the sources, indicating that a mixture of sources
produces a fairly consistent homolog profile in ambient air. Differences could be noted between
urban and suburban air groups.
33
-------
Of special interest was the comparison of paniculate vs. gaseous homolog profiles in air,
and paniculate vs. dissolved profiles in rain. Examination of homolog distributors for these
individual phases by PCA clearly showed that panicle-bound PCDDFs are preferentially
transported to sediments, giving rise to the pattern alterations shown in Figure 5. Relative to
whole (paniculate + gaseous) air samples, the homolog profile of the paniculate fraction moved
toward the cluster for sediments, whereas the gaseous fraction moved away. The same trend was
noted for paniculate and dissolved components of rain samples.
PCA revealed differences between urban (Indianapolis, IN) and rural (Trout Lake, WI) air
samples, indicating that changes occur during atmospheric transport. The cluster of wet and dry
deposition samples was shifted away from those of whole air samples and toward the sediments
(Koester and Kites,-1992).
Smith el al. (1990) were able to distinguish three patterns of ambient air PCDDFs in
upstate New York cities by PCA, which they designated as "source-related", "common
background", and "enhanced lower-chlorinated compounds" The latter samples were especially
high in TCDFs. Some air samples from Niagara Falls, NY showed a "source" profile, typical of
municipal waste incinerators; others showed the "common background" pattern. A background
pattern was also yielded by air sample from a parking garage. The "lower chlorinated"
distribution was found on consecutive hot summer days when air masses were transported from
New York City and New Jersey to Albany, where the samples were taken. The authors
speculated that enhancement of less chlorinated species might result from selective deposition of
heavier, more panicle-bound compounds, but also mentioned the possibility of other, unidentified
source types.
Clues to the origin of these IcJwer-chlorinated PCDDFs were provided by Bacher el al.
(1992), who examined fingerprints of PCDDFs and their brominated analogs in chimney deposits
from wood burning and in auto exhaust. Compounds containing 1-4 chlorines were abundant,
especially for the PCDFs. Auto exhaust yielded the highest levels of brominated dibenzodioxins
and dibenzofurans, most containing 1-2 bromines. The authors felt that the full homolog
spectrum (1-8 chlorines) should be considered when deducing sources from environmental
patterns. Rappe el al. (1989) also commented on the similarity of the isomeric pattern for tetra-
CDDFs in air paniculate.samples, car exhaust, and municipal incinerator products.
Emissions data from incinerators in different countries were compared using PCA
(Edgerton el al., 1989). Of twelve municipal incinerators, homolog profiles for seven were tightly
grouped, whereas the other five were widely scattered. Those five included three incinerators that
34
-------
were burning under unsteady conditions and two that had PCDDF levels near the detection limit.
The seven grouped incinerators were clearly distinguished from the well-separated clusters of
three sewage incinerators and three Kraft mill boilers. Ambient air samples from Columbus, OH
which were taken near both a sewage sludge and a municipal incinerator were similar to both
source types. Air samples from Akron, OH were grouped in the overlapping clusters formed by
municipal incinerators and traffic tunnels. Based on a survey of 400 publications on PCDDF
sources, Pitea et al. (1989) used PCA and cluster analysis to differentiate PCDDF patterns in
various types of incinerators.
Class separation of PCDDFs in human milk samples from Sweden was achieved using
partial least squares analysis (Lindstrom et al., 1989). One set of milk samples characterized by a
relative high proportion of PCDFs (especially hexachloro-DFs) came from an area near a
magnesium plant which had high emissions of these compounds.
Multivariate analysis of atmospheric data from Stockholm and Baltic areas indicated a
change in PCDDF proportions in moving from the city to open coastal areas (Broman et ai,
199la). Three air samples showed similarities to settling paniculate matter collected in sediment
traps from the Baltic. A fourth air sample more closely resembled particles from filtered water.
Homolog profiles from combustion sources and in ambient air are substantially different
from those in pulp mill effluents. Combustion patterns of 4-8 chlorinated compounds, exemplified
in Figures show a wide range of PCDDFs. Effluents from Kraft bleaching are enriched in tetra-
CDDFs, especially in toxic compounds substituted in the 2,3,7,8-positions (Amendola et a/.,
1989; Axegard and Renberg, 1989; Clement et a/., 1989; Fouquet et a/., 1990; Swanson et al.,
1988). Homolog fingerprints in sediments from large Swedish lakes showed that mill-related
compounds were distributed throughout the lake. These patterns could be distinguished by PCA
from those in three small lakes that were atmospherically influenced (Kjeller et al., 1990).
Kang and Kamens (1992) used results from laboratory chamber degradation experiments
to correct PAH profiles in urban air for transformation effects. Corrected profiles were used in a
CMB model to differentiate three source types: residential wood combustion, gasoline spark
ignition emissions, and diesel engine emissions.
C. Pesticides
The application of source apportionment techniques to investigate the sources of
pesticides has been meagre. This is largely because of the physio-chemical nature of pesticides,
how they are used, and where they are utilized. The earth's surface is a vast reservoir for
35
-------
pesticides and other organic susbstances as well as metals such as mercury. The exchange of
pollutants is a two way process with soils, plants, water bodies, etc. acting as t • rources and
sinks of airborne chemicals. Understanding the two-way nature of air-surface exchange is
necessary if we are to interpret observed annual cycles of pesticides in the atmosphere, and
eventually to understand the sources. Several investigations have shown that ambient air
concentrations of OC pesticides and PCBs are higher in summer and lower in winter (Hermanson
and Kites, 1989; Hoff et a/., 199la; Lane et al, 1992a; Larsson and Okla, 1989; Manchester-
Neesvig and Andren, 1989)(Figures 4, 7). Differences in summer-winter fluxes of PCBs in Green
Bay (Achman et al., 1992) and HCHs in the Great Lakes (McConnell et al., 1992) have been
observed.
The temperature effect on these processes can be simply described by the relationship:
LogH(orP) = m/T + b (7)
where H or P are the Henry's law constant or vapor pressure of the compound and T is the
ambient temperature (Kelvin). The seasonal cycle of pesticides in air at Egbert, Ontario are
reasonably well described by Equation 7 (Hoff et al., 1991b, Figure 7). Similar trends were found
for HCHs in the Great Lakes region (Lane et al., 1992b) and chlordane in Columbia, South
Carolina (Bidleman, unpublished data). The plant-air BCF is related to H through Equation 6.
The implication is that concentrations of pesticides and PCBs in ambient air are controlled
by take-up and degassing from soil and plant surfaces. New emissions of pesticides are thus
super-imposed on this background, the magnitude of which varies seasonally and regionally. A
significant problem is distinguishing local surface exchange phenomena (Equations 6 and 7) from
transport- and usage-related events. For example, did the high endosulfan concentrations
observed at Egbert (Figure 7) result from volatilization or local summertime use of the chemical?
Were elevated DDT and chlordane concentrations at Egbert due to temperature controlled local
volatilization, or the fact that warm air masses were transported from the southern U.S.?
Trajectories shown by Hoff et al. (1991b) do indeed show that high concentrations of DDT,
chlordane, and other OCs were associated with southerly airflow.
The answer is that all of these phenomena are related, and at present we cannot decouple
local surface exchange effects from transport. The ability to do this is important to understanding
current sources of OC pesticides and PCBs to the Great Waters.
36
-------
(CO •
350-
SCO-
JCO-
_^ ISO-
'S 100'
g so-'
~ 0-
100-
90-
0 I0i
Q 70'
to-
SO-
30 J
20-1
10-.
0 —
I
PC*
•Djrttnn
ZOfll*
ISO 4
ioc'-
M-
per
»08i
iwoj (d)
iwoj
I4M-J
IJW' TBftlB»llll
IOMJ
KXT
HC1
400-
Jl
Mv Hi; U; Sc»x.
Hv. Ih;
ipt NX
or
s
u
— IOOC
C
O
lo( C = -3560 T * 13.7
cu-Chlordaue
' loi C " -3890'/T * U.8
.. 4.I-DDT
loj C - -3JOO/T + 12.6
t-HCH
lot C « -8980 'T - 33 3
| MO JJOO
1000 T(°K)-'
Figure 7. Annul cycles of pesticides at Egbert, Ontario, 1988-1989 (Hoff et aL
1991a,b). Top: Monthly trends of chlorobenzenes (CBz) and several pesticides
(PCC=toxaphene). Bottom: Equation 6 plots of atmospheric concentration
(log) vs. reciprocal temperature.
37
-------
Isomeric Ratios of Pesticides
Many technical pesticides, especially OCs, are mixtures of several isor.. .1. ai;d related
compounds. Relationships among these species provide clues to sample history and source,
however ambiguities result from the use of products having different isomeric contents.
Differential reactivity of isomers also results in changes in their relative proportions which are not
well understood. Examples of compound ratios as source indicators are given below.
Hexachlorocyclohexanes. Technical HCH (formerly called "BHC") consists of five
isomers: a, (3, y, £• The isomer in highest proportion is a-HCH, accounting for 55-70% of the
mixture. The only insecticidally active species, y-HCH, is present at only about 8-15%. Technical
HCH is heavily used 'in Asia; India alone accounts for over 20,000 metric tons/y. Mexico also
uses technical HCH (FAO, 1986-89). Ratios of a-HCH/y-HCH in various technical HCH
products range from 0.6-15; 4-5 has been found in mixtures from two heavy-use countries, India
and China (Bidleman et al, 1992b). Technical HCH has been banned in Canada (1971) and the
U.S. (1978), and replaced by much smaller quantities of lindane (pure y-HCH). Lindane is also
used exclusively in western Europe.
Atmospheric samples in the Northern Hemisphere show a global background of a-HCH
with spikes of y-HCH from regional use of lindane. Air samples from India show a-HCH/y-HCH
in the 2-8 range, whereas the same ratio is 0.2-1 in Europe (references in Bidleman et al.,
1992a,b). In the Great Lakes region a-HCH/y-HCH ranges from near unity to 20 (Lane et al.,
1992b). A springtime minimum in this ratio at Egbert (Hoff et al., 199la) and Green Bay
(McConnell, 1992) may reflect regional use of lindane as a seed treatment. Lane et al. (1992b)
observed a three-fold increase in the concentration of a-HCH at Point Petre accompanying
airflow from the S-SW As a user of technical HCH, Mexico may be the source.
A confounding factor in interpreting a-HCH/y-HCH ratios is possible interconversion or
selective removal of the isomers. Several studies (referenced in Bidleman et al., 1992a) have
demonstrated that y-HCH is slowly transformed to a-HCH in soils and sediments by microbial
action. Suggestions have been made that this transformation also takes place in the atmosphere
and may account for the exceptionally high a-HCH/y-HCH ratios observed in some Arctic air
samples (Pacyna and Oehme, 1988). No experimental evidence has been found to confirm or
deny this hypothesis. Ballschmiter (1991) called into question the isomerization of y-HCH in the
atmosphere by noting that a-HCH/y-HCH ratios are typically much lower in the Southern
Hemisphere, where use of lindane is more prevalent. If atmospheric chemistry is responsible for
38
-------
the high proportion of a-HCH in the Northern Hemisphere, these processes should take place in
the Southern Hemisphere as well.
DDT and Related Compounds. Technical DDT contains about 70-80% of p,p'-DDT,
the insecticidal isomer, 20-30% o,p'-DDT, and minor percentages of other isomers and impurities.
In the environment, the DDTs are transformed to a greater or lesser extent into DDEs (p,p'-DDE
and o,p'-DDE) and DDDs (p,p'-DDD and o,p'-DDD). Polar metabolites are also formed, but
their use as markers has so far not been exploited.
DDT has a very long lifetime in soils. A 1985 survey of California soils ~ 13 years after
the ban of DDT ~ revealed detectable residues at every one of the 99 sites tested (Mischke et al.,
1985). During this time substantial changes had taken place in the relative proportions of DDT
compounds. The average ratio of DDTs (sum of p,p'-DDT and o,p'-DDT) to total DDT residues
(DDTs + compounds in the above paragraph) in soils averaged 0.49. Thus, about half the DDT
was broken down to DDE and DDD compounds.
Since DDE has a higher vapor pressure than either DDT or DDD, residues volatilizing
from soils treated long ago ("old" residues) would be expected to contain predominantly DDE.
The composition of vapors from technical DDT that is freshly applied or evaporated from soils
treated relatively recently should be enriched in the parent compounds, p,p'-DDT and o,p'-DDT.
This information allows judgments to be made concerning "old" and "new" sources of the
chemical.
Peat cores from the Great Lakes region and eastern Canada showed peak DDT
accumulations in the late 1960s, but surface samples representing the 1980s still contained DDT
residues. Moreover, the composition of this freshly deposited DDT was largely p,p'-DDT and
o,p'-DDT. These findings led Rapaport et al. (1985) to the hypothesis that fresh DDT was being
transported to the Great Lakes region, possibly from use in Mexico and Central America.
The composition of DDT residues in peat cores may not be reflective of what is in the
atmosphere because of differential deposition by precipitation and fallout. Since the DDTs have
lower vapor pressures, they are present to a greater extent on atmospheric particles and more
amenable to deposition. Examination of DDT/DDE ratios in atmospheric samples from different
areas of the world shows the effect of age on residue composition (Table 5). High levels of total
airborne DDT have been found in India, Congo, and Irkutsk (Siberia), indicating recent usage
(McConnell, 1992). In all of these locations the proportion of DDT/DDE was high. The ratio
was lower, but still above unity at Lake Baikal, which was probably the recipient of long-range
39
-------
atmospheric transport. In North America, the proportion of DDE has increased with time. This
is exemplified by the decrease in DDT/DDE in Denver from 2.0 in 1980 to 1.2 in /SO A nine-
state survey of airborne pesticides in 1970-71 showed average DDT/DDE = 2.6 (referenced in
Bidleman el ai, 1976). Larsson and Okla (1989) noted that DDT/DDE in atmospheric fallout in
Sweden had decreased significantly, from 4.1 in 1972-73 to 2.6 in 1984-85.
TABLE 5. Ratios of DDT/DDE in Ambient Air.
Location
Lake Baikal, Siberia
Irkutsk, Siberia
Brazzaville, Congo
India (Several towns)
Porto Novo, India
Denver, Colorado
Egbert, Ontario
Great Lakes
Alert, N.W.T.
Year
1991
1991
1989
1980-82
1987-89
1980, 1986
1988-1989
1990
1988
DDT/DDE
1.2
2.5
3.2
1.8
"mostly DDT"
2.0, 1.2
0.4
0.8
0.6
Reference
McConnel, 1992
McConnel, 1992
Ngabe & Bidleman, 1992
Kaushikera/., 1987
Rameshefa/., 1989
Foreman & Bidleman, 1990
Hoffetal., 1991a
McConnel, 1992
Panonetal., 1991
Current DDT/DDE ratios in air from the Great Lakes region are typically <1.
Nevertheless there are some situations in which DDT/DDE > 1 are observed in North America.
McConnell (1992) gives ratios of DDT/DDE for samples collected in 1989-90 in the Great Lakes
region. One sample taken in Green Bay during June, 1989 showed DDT/DDE = 1.8. This
sample was also enriched in toxaphene and chlordane, and was associated with airflow from the
S-SW Another air sample from Lake St. Clair in August, 1990 contained unusually high levels of
DDTs and DDT/DDE = 2.4. Trajectory information has not yet been obtained for this sample.
Although preliminary, these observations suggest that pulses of "new" DDT are being
superimposed on a North American background of atmospheric residues that are largely DDE.
40
-------
The source of this "new" DDT is unclear. Usage in Mexico and Central America is often
assumed to be responsible, because of the DDT/DDE relationships for "old" soils discussed
above. However a recent report documented long-term persistence of parent DDT -p,p'-DDT --
in soils from New Mexico and Texas (Hitch and Day, 1992). Some of these soils contained five
times as much DDT as DDE when sampled in 1985. Investigation of the DDT/DDE ratio in
combination with air trajectory information should be useful for locating such regional sources.
Chlordane Isomers. The ratio of trans-/cis-chlordane (TC/CC) in technical chlordane is
about 1.3. Because TC is slightly more volatile, the predicted vapor-phase ratio is 1.7. Average
TC/CC ratios in air samples from a source region (Columbia, South Carolina) are close the latter
value (Bidleman et a/., 1990). As chlordane undergoes atmospheric transport, the TC is depleted
and TC/CC decreases. Ratios from various locations removed from North America show this
effect: Sweden 1.3, Sable Island 0.9, summer Arctic 0.5-0.6, winter Arctic 1.0, Lake Baikal
(Siberia, summer) 0.6 (Bidleman et al., 1990, 1992c; McConnell, 1992). The situation at Sable
Island was interesting because it suggested differences between regional transport and the global
background. Air samples showing 80% transport from the S-SW showed TC/CC =1.1-1.3. By
comparison, when air masses arrived mainly from the N-NW, TC/CC dropped below unity.
TC/CC in air samples at Egbert, Ontario showed an annual cycle, about 0.8-1.0 in summer and
1.2-1.5 in winter (Hoff era/., 1991a).
Reasons for these isomeric changes have not been determined. TC is more labile in soils,
and thus old residues might be expected to contain higher proportions of CC. Changes in this
ratio, as well as those of DDT/DDE might occur if these pesticides move by the "grasshopper
effect", continually being deposited and revolatilized from soils. It is unknown if chlordane
isomers are decomposed in the atmosphere.
Recently a new technique has been developed which may help distinguish "old" from
"new" sources - the ability to separate optical isomers by gas chromatography. Like amino acids,
certain pesticides have "right" and "left" handed isomers. These enantiomers have identical
chemical and physical properties and are thus not distinguishable by normal analytical methods.
The development of chiral-phase capillary gas chromatography columns allows their separation.
The technique allowed Muller et al. (1992) to determine that, whereas the two enantiomers of a-
HCH were in a 1:1 ratio-in the technical pesticide, alterations in this ratio occurred in soil, rain,
and tissue samples. Because only enzymatic processes can change the enantiomeric ratio, this is
evidence for biological alteration of residues which are reflected in atmospheric samples (rain).
41
-------
D. PCBs
Little work has been done to link PCB residues in receptors to their sources through
examination of congener profiles. This is probably because of: a) Lack of routine use of high-
resolution analytical methods, b) Poor interlaboratory comparability because of different numbers
of congeners included in the analysis, c) Poor information on PCB profiles from different sources,
and d) Changes in PCB profiles in the environment.
In the 1960s and most of the 1970s, analysis of environmental samples for PCBs was done
by packed-column gas chromatography (GC). This is a low-resolution technique which allows
the different Aroclors to be distinguished, but does not separate individual congeners. More
recent analytical data has been obtained by high resolution capillary GC, which has far greater
separating power (Figure 8). GC-MS is also becoming more common for determining PCBs as
homolog groups (Slivon et al., 1985; Alford-Stevens el al, 1986).
The ability to carry out an analysis for individual PCBs is important. Chemical, physical,
and lexicological properties vary substantially among the PCBs. Differences in volatility, water
solubility, and reactivity often lead to PCB profiles in environmental samples that are markedly
altered from those of the Aroclor fluids responsible for the contamination. Improvements in
analytical techniques within the last 10-15 years has been critical to our current understanding of
PCB environmental chemistry. Still, it is common for laboratories to include different numbers of
congeners in their analytical scheme, making comparisons among research groups difficult. As is
the case with PAHs, there has been a lack of standard reference materials (SRMs) for PCBs.
Recently marine sediment SRMs for PCBs have become available through NIST and the National
Research Council of Canada (Schantz et a/., 1990).
As is the case of PCDDFs, PCB profiles are greatly altered between source and receptor.
This is largely due to the selective phase partitioning that is a consequence of the wide variation in
physicochemical properties of individual PCB congeners. An example of the difference between
PCB profiles in air and rain is shown in Figure 9 (Duinker and Bouchertall, 1989). PCBs in rain
obviously originate from washout of the panicle-bound, rather than the vapor-phase, PCBs in air
Further changes in profiles accompany sedimentation and bioaccumulation processes.
Schwartz el al. (1987) discussed the problem of quantifying PCBs as Aroclor mixture
equivalents vs. as the sum of individual congeners. The conclusion was that samples should be
analyzed for individual congeners, and total PCBs should be reported as their sum.
42
-------
GC sepantion of Aroctof 1260
I00
50
{») PiMtod cohimn
I J 1
0 3 6 9 12 15 16 21 24 27 30 33 36 39 42 45 « 51 54
Retention Dm* (mm)
(b) dpAlary column
100
~ 75
I
A
2
B
*
| »
B
1
25
1
J
u
JL
,
Ll
t
U
LJL>
: i i i i
16:40 20:00 23:20 28:40 30:00 33:20
RM0ntoo bme (mn)
Figure 8. Comparison of packed and capillary-column gas chromatography for
seperating components of Arolocfor 1260. Source Alford-Stevens (1986).
43
-------
Vapor phase
136
Filter
Ram
Figure 9. Profiles of PCBs in the paniculate (filter) and vapor phases of ambient air.
and in rain. Source: Duinker and Bouchertall (1992).
44
-------
Little is known about the relative proportion of PCBs from different sources. Aroclor
fluid compositions have been well established (Schulz et al., 1989), but seldom have emissions
profiles been established. Thus it is difficult to determine the origin of PCBs in an ambient air
sample.
Profiles of PCBs in ambient air have been used to suggest sources in South Chicago.
Murphy et al. (1985) noted that PCB patterns from most incinerators tested were similar to
Aroclors 1248 or 1254. However one sewage sludge incinerator showed a predominance of
heavy PCBs, similar to Aroclor 1260. Holsen et al. (1991) found that PCBs in two spring air
samples from Chicago matched Aroclor 1260, while a third was similar to Aroclor 1242, a lighter
PCB mixture. A series of air samples from South Chicago in February, 1989 showed great
differences in absolute concentrations of PCBs and their profiles (Gotham, 1990). Profiles of
most samples were skewed toward the low molecular weight PCBs, as is typical of ambient air.
For these, total PCBs ranged from 300-2500 pg/m3. Two samples contained total PCBs of 4700
and 10,000 pg/m3, and were dominated by heavy components, as is Aroclor 1260. Thus three
separate observations have documented that certain events in Chicago lead to distinct profiles of
heavy PCBs in ambient air.
Several investigations have shown that PCBs volatilize from the Great Lakes and are a
significant source to the atmosphere above the water (Achman and Eisenreich, 1992; Baker and
Eisenreich, 1990; Hornbuckle et al., 1992). Differences in PCB profiles over land and over the
lake may be valuable in distinguishing freshly transported PCBs from those being recycled into the
atmosphere from the lake (Hornbuckle et al., 1992).
4. IDENTIFICATION OF LOCAL SOURCES
The first step in defining the contribution of emissions from a given source or source
region to the pollution load at a given receptor is to prepare chemical and physical
characterization of this source/source region. This information is being prepared for the Great
Lakes region through several research projects sponsored by the EPA, the Ontario Ministry of the
Environment, and the International Joint Commission. A number of institutions have been
involved in these projects and as a result several technical reports are now becoming available
The above described emission inventorying has been done mostly for precursors of atmospheric
acid compounds, such as sulfur and nitrogen compounds, as well as some photochemical
oxidants, such as selected volatile organic compounds. The hazardous pollutants, as defined in
the 1990 Clean Air Act Amendments (CAAA), have been less well defined with respect to their
emissions but this trend may be changing.
45
-------
Two major conclusions can be drawn on the basis of current work on emission
inventorying for toxic compounds. First, a list of source categories and processes generating
emissions of hazardous air pollutants can and should be defined for the regions of interest. The
above can be done for point and area sources, separately. The second conclusion is that ongoing
work on inventorying emissions of criteria pollutants and VOCs, particularly those efforts that are
based upon emissions inventories prepared under the National Acid Precipitation Assessment
Program (NAPAP) should utilize point sources listing to estimate the emissions of toxic
compounds as well as the criteria pollutants, e.g. electric power plants, smelters, incinerators,
cement kilns, and chemical plants. This information is needed when preparing spatial distribution
of toxic compound emissions within the study area.
4.1. Major source categories for toxic compound emissions
As pursuant to the requirements of the CAAA of 1990 an interim toxic emission inventory
has been developed for the continental United States. Preliminary results of this work which
includes the geographical distribution and source type analysis has recently been presented by
Benjey and Coventry (1992). Several heavy metals were inventoried including arsenic, cadmium,
lead and mercury which are of primary interest in this study. Altogether emissions of 28
compounds, both heavy metals and persistent organic pollutants have been inventoried based on
the 1985 NAPAP inventory.
Toxic emissions in the Great Lakes region were also inventoried within other programs.
such the EPA Region V Project on Air Toxics Emission Inventories for the Lake Michigan
Region, and other organizations, such as the International Joint Commission (IJC). The IJC
report provides data on the production, usage and atmospheric emissions of 14 toxic chemicals,
including the four heavy metals studied in this work, and other priority compounds: polynuclear
aromatic hydrocarbons (PAH), dioxins, furans, polychlorinated biphenyls (PCBs), and pesticides,
with a focus on lindane. More recently, an emission inventory for toxic compounds has been
prepared within a project from the Ontario Ministry of the Environment.
An emission inventory of toxic air contaminants for the Great Lakes states is now being
developed at the Michigan Department of Environmental Protection (Vial, 1992). When this
work is completed it wilj be one of the most important sources of information on emissions in the
Great Lake Region to date.
EPA is currently developing the Air Toxics Emission Inventory Protocol for the Great
Lakes States (e.g. Radian. 1992). The Emission Inventory Branch in the EPA Office of Air
46
-------
Quality Planning and Standards has recently prepared the Air Clearinghouse for Inventories and
Emission Factors (AIR CHIEF) providing information on estimating air emissions of criteria and
toxic pollutants from selected sources
Preliminary results of the emission inventory for heavy metals in the United States,
summarized by Benjey and Coventry (1992) suggest that the toxic metal emission sources are
heavily influenced by primary and secondary metal production (93% of the arsenic emissions and
more than 83% of the cadmium emissions), gasoline combustion (about 59% of the lead
emissions) and waste incineration (more than 57% of the mercury emissions). The above
emissions inventory is presently being revised and the relative contributions from the various
source categories may change, particularly the importance of lead from gasoline combustion.
Most of the emission sources for the above compounds are located outside the Great Lakes
region according to this inventory, and mainly in Arizona for arsenic, and Missouri for cadmium.
Emissions of lead and mercury are more evenly distributed. It should be admitted, however, that
the above suggested source category contribution to the total emissions in the United States is
somewhat surprising. Taking into account similarities in production technologies and efficiency
of control equipment, as well as the chemical composition of wastes to be incinerated in the
United States and Western Europe (e.g. Pacyna and Munch, 1988) one should modify a list of
major source categories contributing to the atmospheric emissions of toxic metals in the United
States.
Major source categories for emissions of toxic heavy metals and persistent organic
compounds in the Great Waters regions include:
- combustion of bituminous coal, mostly in pulverized coal dry boilers, lignite, distillate oil .
residual oil. and natural gas to produce electricity (emissions of all heavy metals of concern
and PAH., dioxins, and furans),
- combustion of bituminous coal, distillate and residual oil, and natural gas in industrial
boilers ( emissions of pollutants as above),
- cement production in both dry and wet process kilns ( emissions of heavy metals and PAH),
- production of chJoro-alkali using Hg-cell ( emissions of mercury),
- coke production as by-product in primary iron and steel manufacturing ( emissions of all
compounds except lindane),
- secondary non-ferrous metal (and mostly lead) production (emissions of heavy metals and
PCBs),
- petroleum refineries (PAH),
- refineries and chemical industry (and particularly production of chJorine and caustic soda,
47
-------
production of batteries, production of pigments, use of paint),
- paper and pulp production ( mostly emissions of PCBs),
- waste incineration ( emissions of all compounds except lindane),
- glass industry,
- production of fertilizers,
- crematories,
- combustion of fuels in internal engines ( emissions of lead, PAH, and dioxins),
- use of lindane in livestock treatment ( emissions of lindane), and
- use of lindane in wood and seed applications ( emissions of lindane).
Other source categories which emit large amounts of almost all of the compounds of
interest are so-called diffuse sources. These include:
- combustion of gasoline and other fuels in motor vehicles,
- volatilization of compounds from landfills, both flared and unflared
- re-emission from terrestrial and aquatic environments (mostly mercury), and
- combustion of wood and other fuels to produce heat.
Impurities found in such products as pesticides, rubber tires, pigments and coatings can also
become import local sources of toxic substances (Ayers, 1987).
Emissions of heavy metals from natural sources may also be quite important. However, as
suggested by Lindqvist and Rodhe (1985) for mercury, it is perhaps misleading to categorize
present day fluxes of mercury from soils, bodies of water, and biota as being "natural emissions"
For example, past anthropogenic Hg emissions have been dispersed so thoroughly through the
environment that this distinction is probably no longer meaningful. The flux of mercury to and
from land and water surfaces has only recently been studied in any meaningful way (Schroeder et
a/., 1992; Vandal et al., 1991). These recent studies have emphasized the importance of the air-
water exchange of mercury and its importance in the behavior and fate of mercury in the
environment. In addition, we have a very poor understanding of the forms of mercury emitted
from these "natural systems", but this is the focus of ongoing research (Fitzgerald et a/., 1991;
Pacyna et a/., 1992). Natural sources which are of major importance include:
- re-emission of volatilized heavy metals from soil and surface waters,
- re-suspension of soil panicles,
- forest fires, and
- volcanic eruptions.
48
-------
Potential sources for the critical organic compounds studied in this work are vast. Some
of the major source categories include:
- application of pesticides,
- combustion of fossil fuels in electric power, co-generation and heat production plants,
- combustion of fossil fuels in commercial, industrial, and residential units,
- mobile sources,
- manufacturing and use of basic organic chemicals, and
- waste incineration.
Emission data currently reported by EPA and LJC, reveal that large portion of the
emissions of toxic pollutants to the air in the United States can be generated outside the Great
Lakes region. This is particularly true for emissions from point sources. However, the
neighboring states, and particularly Missouri, generate large quantities of these emissions in
electricity and heat producing power plants, primary and secondary non-ferrous smelters, steel
and iron manufacturing plants, and waste incinerators. The contribution of atmospheric emissions
of toxic compounds generated in various regions of the United States is presented in Figure 10,
together with the emissions from 12 regions in Canada. The data were prepared on the basis of
research carried out for IJC by Voldner and Smith (1989). This report covers two U.S. regions,
specified in Figure 10 as East North Central (ENC) and Middle Atlantic (MAL) and Ontario (ON)
in Canada as far as the Great Lakes are concerned. The contributions of toxic air compounds
studied in this work from sources within ENC and MAL to the total U.S. emissions are varying
from about 10% for mercury to 27% for lead. The emissions in ON contribute about 25% to the
total Canadian emissions with bigger contribution for arsenic (more than 40%). The MAL and
New England (NEG) should be considered for the Lake Champlain. Emissions in NEG do not
contribute significantly to the total U.S. emissions of the studied compounds except for benzo-a-
pyrene (BaP - included in PAHs). However, emissions sources in Quebec generate the largest
amounts of toxic pollutants in Canada, in addition to Ontario.
The emissions in SAL, having an impact on atmospheric deposition of toxic compounds
onto the Chesapeake Bay waters, contribute between 10% and 15% to the total U.S. emissions of
heavy metals, which is considered to be important. Even more significant is the contribution of
persistent organic compounds reaching as much as 50% for BaP.
Although -emitted in regions outside the Great Lakes, Lake Champlain, and the
Chesapeake Bay, heavy metals and persistent organic compounds can reach the surface of these
waters. Once emitted into the atmosphere, these pollutants are subject to long range transport,
49
-------
Umtea S;c;es
~ec orci Percentcce Distribution of 5 Compounas
3 —
'C3 'cr.nes Total
NEC MAL ENC WNC SAL ESC WSC MTN PAC NDF
25 -!
20 -i
15 -i
10 -I
5 -i
Mercury
i .668 Tonnes Total
(61* Natural Sources)
NEC MAL ENC WNC SAL ESC WSC MTN PAC NDF
7,- ___
% 2C -•
21 1 Tifires Total
NEC MAL ENC WNC SAL ESC WSC MTN PAC NDF
45.2
-rsenic
:.232 Tonnes Total
i " _
t _
r
"-'. ENC WNC SAL ESC WSC MTN PAC NDF
50.5
c55 Tonnes Total
NEC MAL ENC WNC SAL ESC WSC MTN PAC NDF
Region in the U.S.
FigurelOa. Contribution of atmospheric emissions of selected toxic compounds from
various regions to the total emissions in the United States.
50
-------
Peg:or.3i Pe-centcge Distribution of 5 Compouncs
25 -
,20 -
!15 -
10 -
5 -
Lead
1 1,466 Tonnes Total
NF PEI NS NB PQ ON MB SA AB BC YUK NWT
30
25 -
20 -
15 -
10
5
Mercury
3.530 Tonnes Total
(99* Natural Sources)
30
NF PEl NS NB PQ ON MB SA AB BC YUK NWT
50.9
-r J Cadmium
I 322 Tonnes TC'.QI
10 -!
H
o 4-
NF
NS N= PQ ON MB SA AB BC YUK NWT
34.0 40.5
; Arsenic
~ 471 Tonnes
NF PEi NS NB PQ ON MB SA AB BC YUK NWT
Region in Canada
FigurelOb. Contribution of atmospheric emissions of selected toxic compounds from
various regions to the total emissions in Canada.
51
-------
during which transformations and deposition processes are occurring. The extent to which these
processes take place depends upon stack parameters such as temperature and velocity of exhaust
gases, meteorological conditions, and the physical and chemical forms of pollutants. The results
from recent studies in Europe provide some basis for estimating what part of the heavy metals and
organic compounds emitted from major point sources is deposited in the area of their emission
sources (local deposition) and what part is transported and deposited outside the emission region.
In the case of emissions from point sources with a stack height of > 150 m (e.g. large
power plants, primary non-ferrous smelters, cement kilns, steel and iron plants and waste
incinerators), all of which employ high temperature processes, only 15 to 20% of toxic emissions
were deposited locally. The rest was transported out of the emission region. Less information on
this subject is available from research in North America, however development of regional models
in the area of the Great waters (e.g. Clark, 1992) would require data on local deposition. It is
certain that we must consider emission sources both within the Great Waters region and outside
the watersheds in order to assess the origin of atmospheric toxic compounds deposited on the
water surface in the region.
The quantity of emissions for the heavy metals and persistent organic compounds of
concern in the states around the Great Lakes, Lake Champlain, and the Chesapeake Bay is
difficult to assess due to diversity of emission numbers reported by various research groups. The
emissions estimates reviewed for this report differed by one order of magnitude, for most of the
heavy metals considered in this, and are presentlt under revision and modification.
In order to revise the emission data more information is required on emissions within
major source categories. Most of the work in this respect has been done for mercury. The
emission quantities of mercury within major source categories in the Great Lakes region are given
in Figure 11. Emissions of mercury during combustion of coal are clearly the highest, followed by
emissions from waste incineration. Therefore, coal-fired power plants and waste incinerators in
New York, Ohio, and Pennsylvania dominate emissions in the region. To prove this hypothesis,
emissions of SC»2 and total suspended particles (TSP) have been studied and results are shown in
Figure 12. It can be noted that the S02 emissions in Ohio alone contribute more than 11% to the
total U.S. emissions of this compound and are followed by emissions in Pennsylvania. More even
in the contribution of TSP emissions in the states around the Great Lakes to the total U.S.
emissions of TSP It should be noted that total suspended particulates are the major carrier of
atmospheric heavy metals. The difference between the distribution pattern of SO2 and TSP
emissions as shown in Figure 12 is mainly due to larger variety of sources emitting TSP compared
to SO2- The sources of the latter compound are more homogeneous. Even distribution of TSP
52
-------
Sources cf Mercury
•n f.e Great Lakes Bcsm
FOSS
Co-r.D-StiOn
IL IN Ml MN NY OH PA Wl ONT
8 -\
Woste incineration
£ 6 H
! * H
2 H
V/'M'M VMWA
V//<'/Zi
IN M! MN NY OH PA Wl C\"!
,
v i
I M
~ 2 -J
•L
'/,y7/'*
Ml
MN NY OH PA
Wl
IN Ml MN NY OH PA
Wl
20
I/)
1CH
Tc;c Emission B,, State
Witn * of Totol Emission
2 3*
6.4* 6.6*
5.6X
20* 16'*
IN
Ml MN NY OH PA Wl ONT
Figure 11. Contribution of atmospheric emissions of mercury from various source
categories to the state emissions in the Great Likes region.
53
-------
PERCENTAGE CONTRIBUTION OF THE GREAT LAKE STATES TO
THE TOTAL EMISSION OF S02 & TSP
10
11.5
30.1
8 -
6 -
4 -
Total U.S. Emission: 21.0 x 10 tonnes
0 -r
!L IN Ml MN
i
NY
OH PA Wl TOTAL
13.2
5 -|
I TSP
! 6
ictc 'J.S. Emission: 1.83 x 10 tonnes
/c
I I
IL IN Ml MN NY OH PA Wl TOTAL
Figure 12. Contribution of emissions in the Great Lakes states to the total emissions of
SO2 and total suspended particles (TSP) in the United States.
54
-------
emissions indicates that sources, such as waste incinerators, ferrous and non-ferrous smelters, and
cement kilns located all-around the U.S., as well as area sources such as heat production and paint
application (see even distribution of the Hg emissions from the latter source in Figure 11) can
substantially contribute to the atmospheric deposition of the studied pollutants to the Great Lakes,
Lake Champlain, and the Chesapeake Bay. It is, however, difficult to substantiate quantitatively
the above suggestion due to a lack of information on source receptor relationships for the studied
compounds, and particularly for the Chesapeake Bay and Lake Champlain. Although the
measurements at receptors are sometime available, there is a chronic lack of emission data to
study these relationships.
In the evaluation of the impact of emissions from local sources on deposition of pollutants
into the Great Lakes, Chesapeake Bay and Lake Champlain, the locations of the largest point
sources should be considered. If we first look at the location of the 200 largest power plant
emitters of sulfur oxides (presented in Figure 13, EPA, 1991), most of the 50 largest plants are
found in a belt from Missouri through Illinois, Indiana, Michigan, Ohio, West Virginia, to
Pennsylvania. The contribution of the power plants in these states relative to the total power
plant emissions for As, Cd, and Pb are shown in Figure 14. The plant design, and particularly the
burner configuration, influences the emissions of trace metals (Pacyna, 1989). Wet bottom boilers
generate the highest emissions among the coal-fired utility boilers because of the need to operate
above the ash-melting temperature. At a typical peak temperature of about 1550 °C the volatile
trace elements in the coal ash evaporate (Pacyna, 1980). Later they condense as submicron
aerosol particles, or on the surface of ash particles as the flue gas cools to 370 - 450 °C in the
convective heat transfer sections of the power plant. The emission rates from other types of
boilers, such as wall-fired and tangential units seem to be lower due to the lower temperature
involved. For mercury, however, the emissions do not differ as this element is volatile at low
temperatures and passes the control equipment of electric-power stations almost entirely in the
gas form.
The remainder of the largest power plants in the United States are located mostly in the
southern states; Texas, Tennessee. Alabama. Florida, and Georgia, as well as Kentucky. The next
150 largest power plants are more evenly located in the eastern and southern parts of the country
with a higher density of plants in a few regions, including the region surrounding the Chesapeake
Bay area. A fraction of the arsenic, cadmium, mercury, and lead emissions from the largest 200
power plants, and particularly those within the power plant belt as defined above, are probably
deposited to the Great Lakes, the Chesapeake Bay waters, and also to the Lake Champlain basin.
The latter region is probably also affected by emissions from combustion of fossil fuels in the
55
-------
Quebec power plants. The degree of this impact is difficult to quantify on the basis of existing
information on heavy metal emissions. However, results given in the N'AKA. :port by
Venkatram et al. (1990) indicate that nearly half of the sulfate wet deposition is uue to sulfur
emissions along the Ohio River Valley/Midwest region discussed above (See Figure 15).
Figurel3. Location of 200 major power plants in the United States.
56
-------
0)
c
c
o
100
80
60
40
20
0
Emission of Toxic Metals frcm Power Plants
(1987)
As
IL
346
IN
Ml
I""1
MO OH
PA
T
WV TOTAL *TOT
(C
~j
o
w
0)
c
c
o
E-
80 -
60 -
40 -
20 -
Cd
"ftffff^
ys/7s$A
w 9
u i i i i i i i i i i
IL
IN
Ml
OH
PA
wv TOTAL rroT
o
CO
(D
C
C
O
100
80 -
60 -
40 -
20 -
0
Pb
IL
IN
220
W////A
Ml
MO OH
PA
WV TOTAL KTOT
o
Figure 14. Contribution of the As, Cd, and Pb emissions from power plants in
Illinois,Indiana, Michigan, Missouri, Ohio, Pennsylvania, and West Virginia to
the total emissions from power plants in the United States.
57
-------
The location of major primary and secondary non-ferrous metal smelters was also
reviewed because of the large emissions of atmospheric heavy metals from those plants,
particularly those employing pyrometallurgical processes (Metal Bulletin Books - MBB, 1986).
The major primary smelters are located outside the Great Waters study regions, with the
exception of a copper smelter in the Upper Peninsula of Michigan, and their emissions impact on
the Great waters is discussed later in this report. The type of technology employed in smelters,
refineries, and other operations, such as roasting, has a major influence on the trace element
emissions (Pacyna, 1989). Secondary smelters (International Lead and Zinc Study Group -
ILZSG, 1984; 1985) are located in many areas in the U.S. However, their contribution to the
atmospheric deposition of heavy metals is much lower than the impact of primary smelters due to
the difference in the raw materials used. On the other hand, secondary smelters are considered
mainly as local emitters as they release exhaust gases through rather small stacks (20 to 50 m) in
comparison with tall stacks (over 100 m) in primary smelters. Several secondary smelters in
Indiana, Illinois, Pennsylvania, Ohio, Michigan, New York, and Ontario generate emissions of
heavy metals to the atmosphere which are probably deposited in the same region, the Great Lakes
Basin. The same would apply to the contamination of the Lake Champlain waters by atmospheric
emissions from secondary smelters in New York and Quebec, as well as the contamination of the
Chesapeake Bay waters by emissions from smelters located in Maryland and West Virginia.
Another major point source for many of the compounds discussed in this report is sewage
sludge incinerators. The density of sewage sludge incinerators is greatest in the eastern United
States as shown in Figure 16.
VOCs
A great deal of work has been put into developing inventories and characterizing the
source profiles of different VOC emitters. In the U.S. VOCs are mainly anthropogenic; an
inventory of biogenic non-methane hydrocarbons revealed that quantities released were 20 times
lower than those from human activities (Lamb el a/., 1987). The 1985 NAPAP survey (Saeger el
ai, 1989) identified over 3000 individual point and area source types for VOCs and included over
600 individual compounds. Piccot et al. (1992) expanded this to a global inventory of VOCs
from anthropogenic sources by assuming that NAPAP data were representative of similar source
types from around the world. Results showed that the U.S. leads the world in emissions of
paraffins, aromatics, formaldehyde and other a.dehydes, and marginally reactive compounds;
accounting for 23 - 38% of the global total in each of these categories (Table 6). Emissions of
most VOCs are highest in the eastern third of the country, including the Great Lakes and mid-
58
-------
(a)
01
0.1
Figure 15. Source apportionment of wet sulfate deposition ( per cent per state or
subprovince) at a) Whiteface Mountain, NY, and b) Mt. Mithchell, NC after
Fayetal. (1985).
59
-------
TABLE 6. Global and United States Emissions of VOCs,
Paraffins3
Olefins
BTX Aromaticsb
Other Aromatics
Formaldehyde
Other Aldehydes
Marginally Reactive0
Total
World
50258
38264
14041
4666
1019
307
910
109465
U.S.A.
11785
2772
2534
1750
304
102
264
19511
X jf World
23.4
7.2
18.0
37.5
29.8
33.2
29.0
17.8
Source: Piccotetal., 1992.
a) Includes straight- and branched-chain alkanes, alcohols, esters, and ketones.
b) Benzene, toluene, xylenes.
c) Includes most chlorinated and fluorinated compounds.
60
-------
Figure 16. Location of major sewage sludge incinerators in the United States.
Source: EPA Document 450/2-90-009 (1990).
61
-------
Atlantic regions (Figure 17) The heaviest contributors of olefins are tropical Africa and Central
and South America, the result of continental-scale biomass burning.
The Lake Michigan emissions survey by Blakley and KJevs (1990) speciated 1985
NEDs/NAPAP total VOCs data and assigned releases of individual compounds on a gridded
network. Over 100,000 tons of air toxics were emitted from point and mobile sources within a
21-county area surrounding the Lake Michigan Basin. Approximately 15,000 tons were emitted
from point sources, of which over 90% were located in Illinois and Indiana. Releases from point
sources are listed by facility and county. Benzene and light chlorinated compounds were major
contributors to total VOCs from point sources, amounting to 5097 and 5548 tons/y. Highest
releases of benzene occurred in four counties near the southern end of the lake: Will and Cook
(Illinois), Lake and Porter (Indiana). The latter three counties were also listed as having
substantial coke oven emissions. Following the point source presentation is a list of emissions
from all sources (point + area) by grid within the 21 counties.
Prototype VOCs emission inventories were made for ten counties in southeast Virginia
(Emmim et al'., 1989) and later expanded to eleven midwestem and mid-Atlantic states (Wind and
Burke, 1990). The estimates were made from NEDS point source and area source data for
VOCs, and in the case of the 11-state study, also from 1988 mileage data from the Federal
Fiighway Administration. Total VOCs were speciated using apportionment factors provided by
Radian Corporation. Atmospheric release of carcinogenic VOCs in the Virginia counties was
estimated to be 1987 tons/y, broken down into: gasoline vapor 38.6%, benzene 21.7%,
formaldehyde 21.0%, chlorinated solvents 10.0%, and acrylonitrile 8.6%. Vehicles accounted for
about half the benzene released; the remainder was from point sources. For the eleven states.
VOCs emitted totaled 813,400 tons/y. Vehicles accounted for 66.3% of the 333,200 tons of
benzene + toluene released.
PAHs
For PAHs, the relative importance of different sources changes seasonally and as society
achieves greater control measures (Back et a/., 1991a,b). For example, Harkov and Greenburg
(1985) calculated that 183 kg BaP (98% of total) was released by motor vehicles in New Jersey
during the non-heating season, but 6135 kg BaP (98% of total) reached the atmosphere through
residential wood combustion during the heating season. A 5-10 fold drop in PAH levels in the
Baltimore Harbor Tunnel between the mid-1970s and 1985-86 was attributed to installation of
catalytic conveners in the later years (Benner et a/., 1989). The drop in aerial BaP concentrations
in the Great Lakes region over the last two decades has been accompanied by a decline in the BaP
content of surface sediments in the lakes (Eadie, 1984; Eadie et al., 1990).
62
-------
GRID CF\ I EMISSION «606 3
Figure 17. Global distribution of total VOCs, 109 g/y. Source: Piccot era/.. 1992.
63
-------
One of the most detailed studies on emissions of these pollutants and their spatial
distribution around the Great Lakes has been carried out for the Ontario Ministry of the
Environment (Johnson ei a/., 1990) Spatial distributions of the annual emissions of PAH, PCDD,
PCDF, and insecticides in Eastern North America within 127 km x 127 km polar stereographic
grid system are presented in Figures 18-21, respectively.
A great deal of effort was made to compile or estimate PAH emission factors for a large
number of source types, including industrial processes, vehicles, residential combustion (wood,
coal and oil), power plants, incinerators, open prescribed burning, and forest fires.
One can note the large difference in emission factors among source types. For example
those from industrial and utility coal combustion are given in mg/metric ton coal bumed, whereas
ferroalloy and aluminum production and residential wood stoves are in g/metric ton metal
produced or fuel burned. Large differences in emissions and ratios among PAH compounds can
be seen for vehicles with two types of catalysts.
The source-to-source variability within each of these categories is also quite large,
typically an order of magnitude or more. Even within a particular industry, difference' :n a
process can greatly change quantities of PAH released. For example, Johnson et al. (19£
total PAHs released by aluminum reduction facilities from 330 to 2,430 g/ metric ton aluminum
produced, depending on whether the process is a pre-bake anode Soderberg or a horizontal
Soderberg. PAHs released from metallurgical coke production were raised from 450 to 1300
mg/metric ton coal charged when partly contaminated water was used for coke quenching instead
of clean water.
Annual releases of PAHs to the atmosphere totaled 272 metric tons in Ontario (ONT) and
9397 metric tons in eastern North America (ENA). These ranged in molecular weight from
acenaphthylene (152) tr coronene (300) Of total PAHs, 1.9% wa estimated to be BaP in ONT
and 3.6% in ENA. Th. difference was attributed to the greater variety of sources in ENA and
their relative contribution to the total Within Ontario, the largest emissions took place along the
northern shores of lakes Ontario and Erie. Over ENA, PAH releases were largest in a wide
diagonal extending across the midwestern states from southern Illinois to the mid-Atlantic and
southern New England states. In this corridor, emission fluxes were 1.6 - 6.4 kg/km2-y, with an
occasional grid showing higher values. Fluxes throughout the Appalachian region and the
southeast were typically lower: 0.4 - 3.2 kg/km2-y, with "hot spots" found near southern cities.
64
-------
1985 Toxics Emission Inventory ,
PAH Combined Fluxes (kg/sqkm/yr)[
>64.0 |
6.4 to64C
3.2 10 6.4
1.6 to 3.2
0.6 to 1.6
0.4 to 0.3 |
0.2 tc 0.4 j
O.i ;o 0.2
O.Ol to 0.1
<0.01
Figure 18. Spatial .distributions of annual emissions of PAH in Eastern North America
within 127 km x 127 km polar stereographic grid system.
65
-------
1985 Toxics Emission inventory
PCDD Combined Fluxes (g/sqkm/yr)
>O.S4
0.064 10 0.64
0.0064 to 0.064
0.0032100.0064
0.0016100.0032
0.0008100.0016
0.0004 to 0.0006
O.OOC2 to 0.0004
0.0001 to 0.0002
£
-------
1985 Toxics Emission Inventory i
PCDF Combined Fluxes (g/sqkm/yr)'
0.064 to 0.54
O.OOW to 0064
0.0032 to 0 0064
0 CO16 to 0.0032
C.OOCfi to 000^6
0.30C4 to 00006
0 3002 to C 0004
0 OOC- 10 C 0002
< C 30C:
Figure 20. Spatial distributions of annual emissions of PCDF in Eastern North America
within 127 km x 127 km polar stereographic grid system.
67
-------
15 Toxics Emission Inventory
Insecticide Fluxes (kg/sqkm/yr)
>256
1.2810256
0 64 10 1 26
03210064
016IOC.32
0 08 10 C 16
004 1C 0.08
0.02 :c C 04
0 3- :o 0 02
< 0 01
Figure 21. Spatial distributions of annual emissions of insecticides in Eastern North
America within 127 km x 127 km polar stereographic grid system.
68
-------
Table 7 gives PAH emissions by source class, and also compares the recent estimates of
Johnson el al. (1992) with those from mid-1970s data by Peters (1981) and RamdahJ el al.
(1983). A much more extensive breakdown of ONT and ENA emissions is given in the original
Johnson report pp. 322, 323, and 330-333. For example, 10 types of production facilities are
listed under the "Industrial" category. "Stationary Fuel Combustion" is divided into commercial
and residential fuel and wood combustion, and electric power generation. Contributions of on-
and off-road gasoline and diesel vehicles, aircraft, railroad, and marine vessels are listed under
"Transportation" Residential wood combustion dominated emissions, accounting for 51% of
total PAHs in ONT and 31% in ENA.
TABLE 7. Atmospheric Emissions of PAHs by Source Class, Tonnes/y
Industrial
Stationary fuel
combustion
Solid waste
incineration
Transportation
Open Sources
Total
Ontario
82.3
155.0
0.7
24.5
9.2
271.7
Eastern
N. America
2,704
4,545
48.5
1,174
926
9,397
Eastern
U.S.A.
1,831
3,882
<73
1,099
440
7,325
Total
A
640
4,044
56
2,266
4,025
11,031
U.S.A.
B
3,497
1,781
50
2,170
4,100
8.598
Sources: Johnson et al. (1992, p. 321) for all but A and B. Based on 1985 data.
A = Peters et al., 1981; B = RamdahJ et a/., 1983, summarized by Baek et al.
(1991, p.283). Based on mid-1970s data.
Emissions of BaP were estimated in 10 counties surrounding Lake Michigan by speciating
1985 NAPAP paniculate matter data Blakley and Klevs (1990). Releases totaled 0.96 ton/y, of
which 0.82 tons/y came from Porter County, Indiana.
Atmospheric emissions of BaP were estimated at 654 tons/y for the entire U.S. by Wilber
et al. (1992) and Voldner and Smith (1989). The breakdown by source type was: 85.6% wood
combustion, 6.1% agricultural burning, 31% wildfire, 2.0% vehicles, 1.4% coal combustion,
1.3% coke production, and 0.5% other sources. By comparison, the Johnson report estimated
release to the atmosphere of 5.1 metric tons/y BaP in ONT and 340 metric tons/y in ENA.
69
-------
A review of major power plants and aluminum smelters has been performed TO investigate
the impact of these plants on the contamination of the Great waters by PAHs. 1 hi ,,\\er plants
within the belt, as described earlier, emit large amounts of PAHs and their impact on the
contamination of the Great Lakes and Chesapeake Bay waters is thought to be significant for
heavy metals.
Major aluminum smelters in the U.S. are located in many locations in the country,
including facilities in Indiana, Pennsylvania, Illinois, Ohio, and New York (MBB, 1986). Their
impact on the contamination of the Great Lakes waters by PAHs is probably high. However, the
impact is difficult to quantify due to the lack of emission data to be used in source-receptor
studies. Smelters in Maryland should also be of concern when discussing the origin of PAHs in
the Chesapeake Bay waters.
PCDDFs
Voldner and Smith (1989) identified primary sources of PCDDs and PCDFs in Canada as
being combustion of municipal and industrial waste, accidental fires of treated wood products,
production spills during transportation and aerial spraying of herbicides, wood stoves, and PCB
fires. Erickson (1989) concluded that PCB fires produce PCDFs, but not PCDDs. Secondary
sources listed by Voldner and Smith were volatilization and erosion of dust from landfill sites and
from areas where PCDDFs were present as impurities in herbicides. From the recent review by
Johnson et al. (1992) the most important contributors of PCDDFs in ENA are incineration of
industrial and municipal waste, residential and industrial wood combustion, and electric power
generation. Secondary copper production (wire reclaiming) was listed as an important source in
ONT.
Emission factors for some of the major source types are given in the Johnson et al. (1992)
report. The authors stressed the high degree of variability and/or lack of emission factors for a
number of processes.
Differences in PCDDF releases by incinerators are large, and depend on many variables.
For example, emission factors for the municipal incinerators surveyed by Johnson ranged from 1 -
36 mg PCDDFs per metric ton of refuse burned. Edgerton et al. (1989b) reported total PCDD
emissions from municipal incinerators ranging from 1 - 4259 ng per dry standard cubic meter air.
Highest releases were from facilities operating under unsteady burning conditions.
A great number of incinerator surveys have been done, and relationships of PCDDF
output to operating conditions have been evaluated. Many of these studies are published in the
70
-------
annual symposia on "Chlorinated Dioxins and Related Compounds", Chemosphere, Vols. 19-23
(1989-91). Some of the important variables are: a) precursor content of the feed (e.g. presence
of chlorine-containing species), b) type and extent of feed processing (eg. removal of non-
combustible material, shredding or pelletizing refuse), c) combustion and operating conditions
(temperature, oxygen, residence time), d) type of incinerator and supplementary fuel, and f) type
and efficiency of emission control devices (Johnson el al, 1992).
PCDDF emissions in ONT and ENA as estimated by Johnson et al. (1992) are listed in
Table 8 for different source categories. Stationary fuel combustion and solid waste incineration
(municipal and industrial) were by far the dominant emitters. The fact that industrial emissions for
ONT and ENA were nearly identical was explained by the fact that emission factors were
available for only two industries -- secondary copper production and Kraft pulping, and that
emissions for areas other than ONT could be underestimated for lack of point sources
identification and base quantity information.
On the gridded map of ENA, emission fluxes of PCDDFs were remarkably uniform, falling
between 6.4 - 64 mg/km2-y in most regions. Higher fluxes were generally associated with urban
areas. North of the Great Lakes, fluxes dropped to <0.1 - 0.8 mg/km2-y, with an occasional hot
spot.
TABLE 8. Atmospheric Emissions of PCDDs/PCDFs by Source Class, kg/y
Industrial
Stationary fuel
combustion
Solid waste
incineration
Transportation
Open Sources
Total
Ontario
15.6
8.62
10.5
0.1
0.016
34.8
Eastern
N. America
17.6
190
191
8.46
6.57
414
Sources: Johnson et al. (1992, p. 321) Based on 1985 data
71
-------
Pesticides
It has been very difficult to obtain reliable figures on the production and use of pesticides
because of proprietary restrictions. Most OC insecticides have been banned in industrialized
countries (exceptions in the U.S. and Canada are lindane, endo-sulfan, methoxychlor, and
dicofol). Some historical usage data are available, and have been compiled by Voldner and Smith
(1989). These and other figures taken from a U.S. EPA (1987c) report on termiticides are given
in Table 9.
Many OC pesticides which have been banned in the U.S. and Canada continue to be used
in Mexico, Central and South America, Africa, and Asia. Foreign production and usage
information is extremely important for long-lived OC pesticides. A portion of these chemicals
entering the Great Waters today could come from sources outside the U.S., or even North
America.
Reliable statistics are very difficult to obtain, however reports to the Food and
Agricultural Organization of the United Nations (FAO, 1986-89) show that Mexico used the
following OCs in recent years (metric tons/y): DDT = 200-300, technical HCH = 180-250,
lindane = 15-45, toxaphene = 600-1200. Toxaphene is manufactured in Nicaragua, and local
contamination of human milk and foods has been found (Muller et a/., 1988). The FAO also
reports heavy usage of OC pesticides in India, in excess of 20,000 metric tons technical HCH and
200-900 metric tons DDT per year. Information from the India Dept. of Chemicals &
Petrochemicals (IDCP) (Spencer, 1991) indicates annual production of 25,000 - 28,000 metric
tons of technical HCH between 1986-90, in agreement with FAO data. The IDCP gives far
higher figures for DDT: 6700 - 8600 metric tons.
In 1989/90 Resources for the Future, Inc. (RFF) conducted a mail survey of U.S. Dept. of
Agriculture Extension Service weed scientists to determine herbicide use for major crops. Replies
were integrated with data from other surveys to form a national herbicide use database. The
report (Gianessi and Puffer, 1990) summarizes applications of 50 chemicals on a state-by-state
basis and by crop treated for 1987.
Table 10, taken from the RFF report summaries, shows the top ten herbicides used in the
U.S. These accounted for 73% of herbicide use in 1987. Two-thirds of all herbicide in crop
production were applied to corn and soybeans. Non-crop usage of these 10 chemicals was minor.
with the exception of 2,4-D. The breakdown of totals herbicides by state and crop is given in
72
-------
TABLE 9. Production, Sales or Usage of Some Organochlonne Pesticides in the
U.S.A. and Canada, Metric tons/y.
Toxaphene. U.S.A.
DDT. L
1950-54
1955-59
1960-64
1965-69
1970-72
'S.A DDT
139,000 1968
160,000 1969
141,000 1970
90,000 1971
11,000
Canada
831
894
287
137
Aldrin + Dieldrin. USA
1950-54
1955-59
1960-64
1965-69
1970-73
1981-85
5,750
17,640
30,100
37,730
20,200
2,270-3,400b
1964-68
1969-73
1974-78
1979-83
1984-85a
80,700
100,700
110,900
24,730
4,000
Aldrin + Dieldrin. Canada
1968
1969
1970
86
58
20
Chlordane. U S.A.b
1980 4,300
1985 1,450
1986 1,800
Lindane. U.S.A.C
Heptachlor. U.S.A.b
1980 910
1985 340
1986 340
Lindane. Canada
1964
1966
1971
1974
1975
1976
1977
617
312
269
1,600
2,900
13,100
152,000
1968
1969
1970
1971
1984
16
3
3
3
250
Mirex-Dechlorane. U.S.A.
1959-63
1964-68
1969-73
1974-75
187
934
344
61
Mirex-Dechlorane. Canada
1963
1964
1965
1966
1967
1968
9.2
25.3
46.1
37.8
23.2
3.9
Sources: Voldner and Smith (1988), unless stated otherwise.
a) Production banned in 1982, but remaining stocks used until 1986.
b) U.S. EPA (1987b), 1981-85 aldrin/dieldrin figures are imports.
c) Imports from 1974-77. Large increase in 1977 probably due to ban of technical HCH in 1976.
73
-------
Table 11. Com- and wheat-growing states in the Midwest and Plains were the major targets for
herbicides; Texas and California also ranked among the top ten users. Heaviest ;onnages of
herbicides were applied to corn, followed by soybeans.
TABLE 10. Top Ten Herbicides Used in 1987-89.
Active Ingredients
Atrazine
Alachlor
Metolachlo
EPTC
2,4-D
Trifluralin
Cyanazine
Butylate
Pendimethalin
Glyphosate
Total Usage
Metric Tons/Y
29,090
25,000
22,727
17,818
15,000
12,273
10,455
8,636
5,909
5,454
% Share
Corn/Sovbeans
84
91
87
79
12
63
90
99
70
44
Source: Gianessi and Puffer (1990, p. 7).
The National Oceanic and Atmospheric Administration (NOAA) maintains a data base for
35 herbicides, insecticides, and fungicides commonly applied in coastal watersheds. A report
which is in the final stages of publication (Pait el al, 1992) summarizes pesticide usage in
estuarine drainage areas for 1987. This information was obtained from the RFF report and also
from state data bases. The seasonality of pesticide application was assessed by surveys of
representative counties within each coastal state. The information that follows is taken from the
pre-publication version of the Pait report.
Over 13,363 metric tons of the 35 pesticides were applied in estuarine drainage areas. Of
this, 69% were herbicides, 24% insecticides, and 7% fungicides. Alachlor and atrazine accounted
for almost 45% of the total herbicides. By region, total pesticide use was divided as follows:
North Atlantic 0.8%
Middle Atlantic 27.0%
74
-------
South Atlantic
Gulf of Mexico
Pacific
33.5%
34.3%
4.4%
The breakdown of total pesticides by drainage area is shown in Figure 22 (Pait et al,
1992). Chesapeake Bay ranked first, followed Albemarle/Pamlico Sound in North Carolina and
Winyah Bay in South Carolina.
TABLE 11. States and Crops with Largest Annual Use of Herbicides.
State
Iowa
Illinois
Minnesota
Texas
Indiana
Nebraska
Ohio
Kansas
Missouri
California
Crop
Corn
Soybeans
Pasture
Cotton
Sorghum
Wheat
Rice
Alfalfa
Peanuts
Barley
Total Usage
Metric Tons/Y
20,909
20,455
13,636
12,272
11,364
10,000
8,636
8,182
7,727
6,364
Total Usage
Metric Tons/Y
93,636
37,273
12,727
11,818
10,000
8,182
6,818
3,182
2,273
2,272
% Share
Corn/Soybeans
96
95
86
18
97
84
84
43
71
6
Source: Gianessi and Puffer (1990, pp. ^8 and 9).
75
-------
Quantities and relative amounts of various pesticides changed with location. Figure 23
compares tonnages of 15 pesticides used in four of the above estuarine regions. '/;: ppendix of
the Pait report contains quantities of each pesticides applied in the 43 drainage basir.o considered.
Data for Chesapeake Bay are given in Table 12. The seasonality of pesticide use in each region is
also given. For example, atrazine is applied in April-July in Maryland, April- May in Delaware.
and March-May in Virginia.
For a study of pesticide drainage into Lake Erie, Baker and Richards (1990) compiled a
list of the most used herbicides and insecticides in the Lake Erie basin (Table 13). Herbicides
TABLE 12. Pesticide Use in the Chesapeake Bay Drainage Basin (Metric Tons/Year).
Herbicides
Alachlor
Atrazine
Metolachlor
Cyanazine
Linuron
Simazine
Butylate
2,4-D
Trifluralin
Vernolate
Actifluorfer
Bensulide
546
483
279
141
114
98.3
60.0
42.6
24.6
2* 1
9.1
Insecticides
Carbofuran
Chlorpyrifbs
Ethoprop
Malathion
Carbaryl
Terbufos
Disolfoton
Phorate
Permethrin
Endosulfan
Methyl
Parathion
Fenvalerate
Diazinon
Parathion
Methamidophos
Fungicides
128 Chlorthalonil 17.3
45.5 Metiram 3.47
38.1
31.2
29.7
19.9
12.0
8.69
7.69
6.97
5.20
2.36
1.75
0.84
0.64
Source: Pait ei al., (1992, pp. 96-97).
76
-------
Top Ten Estuarine Drainage Areas
Inl.ll IVxIlllill- I
IIKIII Hi- , ir.irl
Pacific
South Atlantic
Gulf of Mexico
FlncUl
Figure 22. Agricultural pesticide use in estuarine drainage areas. Source: Pait et al. (1992).
-------
Table 13. Pesticide Use in the Lake Erie Basin for 1986 and Ranking by Use in
the State of Ohio for 1982.
Pesticide
Alachlor
Metolachlor
Atrazine
Cyanazine
Metribuzin
Chloramben
Linuron
Terbufos
Trifluralin
Butylate
Dicamba
Pendimethalin
Bentazon
Carbofuran
2,4-D
Chlorpyrifos
EPIC
Phorate
Fonofos
Simazine
Total herbicide use in
herbicides listed above
herbicide use in the Lake
Brand
^^^M^VMBHW
Lasso
Dual
Aatrex
Bladex
Lexone, Sencor
Amiben
Lorox, Linurex
Counter
Treflan
Sutan. Genate plus
Banvel
Prowl
Basagran
Furadan
2,4-D
Dursban
Eradicane, Eptam
Thimet
Dyfonate
Princep
the Lake Erie Basin.
make up 97.3% of
Erie Basin.
Typea
H
H
H
H
H
H
H
I
H
H
H
H
H
I
H
I
H
I
I
H
The 15
the total
Total insecticide use in Lake Erie Basin. The 5 insecticides
Quantity
Used
(metric
tons)
1319
897.2
783.9
273.4
255.5
155.5
133.0
74.67
65.7
58.19
54.39
49.49
49.36
48.23
44.44
36.33
34.67
31.51
25.10
24.82
4315.5
245.6
1986
Rank,
bvUse
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
1982
Rank
bv Ohio
Useb
1
3
2
4
7
6
8
10
12
5
19
17
15
11
14
MR
13
NR
9
16
listed above make up 87.9% of the total insecticide use in
the Lake Erie Basin.
Source: Baker and Richards, 1990 (p. 246).
78
-------
accounted for 92% of all pesticide tonnages. Corn and soybeans received 99.7% of herbicide and
90.1% of insecticide applications. In all, 4561 metric tons of pesticides were used in the basin in
1986.
Johnson et al. (1992) estimated usage of several insecticides, herbicides, fungicides, and
nematocides in Ontario, Quebec, and the Atlantic Provinces for use in their Toxic Chemicals
Emissions Inventory. According to the authors, actual sales figures for individual chemicals were
very difficult to obtain because of regulations protecting confidentiality. In some cases estimates
were made by speciating reported sales of a pesticide class. For example, "triazine herbicides"
were broken down into separate quantities of atrazine, simazine, cyanazine, and metribuzine
according to their reported usage on particular crops. Other inventorying methods are described
in their report.
Pesticide emissions for ONT and ENA were made by Johnson et al. (1992), by
considering: a) volatilization from soils, b) volatilization from vegetation, c) wind erosion of
soil, and d) spray drift losses during application. Pesticide application rates in ONT were based
on estimates of usage in Ontario, Quebec, and the Atlantic provinces. The tonnages of herbicides
reported by Gianessi and Puffer (1990) for the U.S. were used to calculate ENA emissions.
The Jury model was used to predict soil volatilization. Climatological estimates of soil
water flux were made by difference between monthly evaporation and precipitation. Monthly
values were combined to provide seasonal averages for the model. A single set of soil properties
"typical of agricultural cultivated soils" was used..
Volatilization from foliage was estimated by assuming that 70% of the pesticide impacting
vegetation evaporates in a span of 5-10 days and the other 30% is washed off by rain and returns
to the soil. These are based on a study of persistence of pesticides on vegetation by Willis and
McDowell (1987). Because the smallest grid size considered was 5 km, spray drift was
considered to deposit, then re-evaporate. Emissions were thus included in the volatilization
algorithm.
Windblown dust is most significant during tilling and in the early growing season.
Johnson et al. (1992) noted that emissions on dust will only be a problem for persistent pesticides,
most of which have been 'discontinued. They also felt that there were insufficient data on levels of
these chemicals in soils to make a meaningful estimate of windblown dust releases.
Atmospheric emissions of pesticides for ONT and ENA are shown in Table 14. Johnson
et al. (1992) note that approximately 2,700 metric tons of pesticides per year were released in
79
-------
ONT, a quantity far exceeding emissions of PAHs. PCDDFs, and PCBs. In ONT the highest
ranking chemicals were the dichloropropenes/propanes -nematocides used on tobnc.o Atrazine.
alachlor, metolachlor, and 2,4-D were the four most heavily emitted herbicides. The former three
are used on field crops and soybeans, whereas 2,4-D is mainly used on grains. The picture is
different for ENA, where the top five herbicides (in rank) were: 2,4-D, alachlor, atrazine.
metolachlor, and trifluralin. Because of the usage patterns and properties of the pesticides, this
rank of emissions is slightly different from the list of tonnages used (Table 10). The asterisks
shown in Table 14, ENA column, refer to compounds for which U.S. inventory data were not
available. For these, only emissions from the Canadian portion of the grid (ONT, Quebec,
Atlantic Provinces) were used.
Figures for linda'ne and chlordane apply only to estimated releases from agricultural
applications and do not reflect their use for seed treatment (lindane) and as a termiticide
(chlordane). In the U.S. chlordane has been banned from agricultural usage for many years, and
its use in termite control was stopped in 1988 (EPA, 1987c, Federal Register, 1988). Before this,
large tonnages of chlordane and heptachlor were applied as termiticides (Table 9). Endosulfan
releases are only for ONT and eastern Canada, since Johnson et al (1992) had no data from the
U.S. However from the Pait et al. (1992) report, it is clear that very large quantities of this
insecticide are used in estuarine drainage areas, including Chesapeake Bay (Table 12, Figure 23).
Pait et al. noted that endosulfan runoff was one of the major causes of pesticide-related fish kills
in estuaries. Because of the above difficulties, emissions of OC insecticides from ENA are
probably grossly underestimated.
The spatial distribution of herbicide emissions in ENA shows highest releases in the "corn
and grain belts" western Ohio, Indiana, Illinois, Kansas. Minnesota, eastern Dakotas and
Nebraska (the survey did not include western portions of these states). Fluxes in these regions
were typically 19-38 kg/km2-y, with pockets exceeding the upper value. Fluxes in the south.
southeast, and mid-Atlantic states were 0.6 - 4.8 kg/km2-y. Occasional areas with fluxes in the
4.8 - 9.6 kg/km2-y were located. A "hot spot" with fluxes of 9.6 - 19 kg/km2-y was found near
Chesapeake Bay. This is in accordance with the high reported usage of herbicides in this drainage
basin (Table 12, Figure 22). Fluxes in the north Atlantic states ranged from <0.03 -0.6 kg/km2-y.
The pattern of insecticide emissions in ENA was different. Highest releases were estimated
for the southeast and mid-Atlantic states, certain areas of the Midwest (Ohio, Michigan, eastern
Illinois, western Pennsylvania, southern Ontario), and western New York. In these regions fluxes
were 0.64 - 2.56 kg/km2-y, with a few areas >2.56 kg/km2-y (central Michigan, southeastern
North Carolina). These correspond largely with fruit and vegetable growing areas. Fluxes
80
-------
TABLE 14. Estimated Speciated Pesticide Emissions in Ontario and Eastern North
America, Kg/Y.
Estimated Emissions Cke)*
Herbicides
Alachlor
Atrazine
Butylate
Cyanazine
2,4-D
Linuron
Metolachlor
Metrabuzine
MCPA
Trifluralin
Total
Insecticides
Carbaryl
Carbofuran
Chlordane
Diazinon
Endosulfan
Lindane
Methoxychlor
Total
Fungicides
Captan
Nematocides
Chloropicrin
Dichloropropenes/propanes
Methylisothiocyanate
Total
Ontario
195,200
220,700
8,950
60,000
184,800
99,900
156,900
51,500
52,400
9,520
1,039,900
19,650
7,410
354
8,580
7,300
2
1,390
44,680
78,030
58,400
1,296,300
185,800
1,540,500
ENA
7,243,000
4,874,000
13,200**
1,534,000
10,691,000
108,700**
3,646,000
58,900**
568,200**
2,058,000
30,795,000
1,872,000
982,700
853**
219.600
11,300**
2,570**
2,110**
3,091,000
132,300**
59,800**
1,330,100**
185,800**
1,575.700
* Values after rounding
** Data not available in the U.S.
Source: Johnson er al. (1992, p. 340).
81
-------
Figure 23. Application of selected pesticides (1000 pounds/y) b drainage basins of the
middle Atlantic, south Atlantic, Gulf of Mexico, and Fudiic rc..io::s. Source:
Paite/a/. (1992).
''"'ion of Mated PotnJo m tfir Afcttlr Mmnhc. 1JI7
ofSdstd falidda m On Sou* AOntk. 1ST
{ JOOO-
§ isco-
ill
Illl
, o/Uta* Potato in tkr Gulf e( Mam. 1317
Afptiatio* efSdtcta] fatidda in Ike Pxifu. 19*7
82
-------
throughout the "cotton belt", the Appalachians, and southern New England were lower: 0.04 -
0.64 kg/km2-y. Values for upper New England and the Atlantic Provinces were <0.01 - 0-08
kg/km2-y. Insecticide figures for ENA do not include fungicides and nematocides, and as noted
above, releases of several other chemicals are for ONT and eastern Canada only. Emissions of
insecticides are thus underestimated for many chemicals.
A separate inventory and atmospheric transport/deposition model was constructed for
toxaphene by Voldner and Schroeder (1989). Toxaphene was very heavily used in the U.S.,
especially in the "cotton belt" during the late 1960s through mid-70s (Table 9). Usage declined
thereafter, and production was banned in 1982. Remaining stocks were allowed to be applied
until 1986. Voldner and Schroeder speciated total toxaphene use according to state and crop,
using information from several reports referenced in their paper, surveys of U.S.D. A. Cooperative
Extension programs at various universities, and an agricultural census of crop types.
Apportioned toxaphene tonnages in eastern and midwestem North America for 1980 are
shown in Figure 24.
PCBs
Originally PCBs were used for a wide variety of purposes (Voldner and Smith, 1989).
These were: closed system electrical and heat transfer fluids (approximately 60%), plasticizers
(25%), hydraulic fluids and lubricants (10%), and miscellaneous uses (5%) which included flame
retardants, additives to paints, inks, seal-ants, and carbonless copy paper. After 1971 uses were
restricted almost entirely to closed electrical systems such as transformers and large capacitors.
According to Voldner and Smith (1989), 640,000 metric tons of PCBs were produced in
the U.S., of which 70% were sold to manufacturers of transformers and capacitors.
Approximately 40,000 metric tons were imported into Canada and another 3,000 metric tons
entered Canada in manufactured fluorescent light ballasts and high intensity discharge lighting
fixtures. As of 1982, the status of the 640,000 metric tons of PCBs in the U.S. was estimated to
be:
• Destroyed 3%
• Exported 11%
• Buried in landfills 21%
• Still in service 54%
• Circulating in the environment 11%
83
-------
Source: Voldner an
84
-------
PCBs are released into the atmosphere from both point and area sources. The former
include municipal and hazardous waste land-fills, which contribute PCBs through volatilization
and release with landfill gases (Murphy et a/., 1985; Lewis et ai, 1985). PCBs are also emitted
by refuse and sewage sludge incinerators (Murphy et al., 1985). Occasional leakage of
transformers and capacitors, and emissions from other in-service equipment (e.g. old fluorescent
light ballasts) releases PCBs on an area scale.
Like OC pesticides, PCBs are continually cycling through the environment and air-surface
exchange is important in controlling their atmospheric concentrations. Several studies have
shown that PCB levels in ambient air are temperature-dependent, much like the pesticides in
Figure 16 (Hermanson and Kites, 1989; Hoff et ai, 1991b; Manchester-Neesvig and Andren,
1989; Larsson and Okla, 1989). Revolatilization of PCBs from surface water must be considered
in the PCB balance of the Great Lakes (Achman and Eisenreich, 1992; Baker and Eisenreich,
1990; HornbuckJe et al., 1992, Strachan and Eisenreich, 1988).
Because of the sporadic nature of PCB emissions, it is extremely difficult to predict
quantities released to the atmosphere. Murphy et al. (1985) estimated that between 10 and 100
kg/y were emitted from sanitary landfills in the U.S., based on a survey of six sites in the Great
Lakes area. At these landfills 70-500 ng of PCBs were expelled per m3 methane. Differences
among hazardous waste disposal facilities are very large. Lewis et al. (1985) found that PCB
concentrations over "hot spot" chemical waste landfills greatly exceeded background
concentrations, whereas emissions from a well-controlled facility were negligible. Based on other
reports, Voldner and Smith (1989) estimated that PCBs released to the environment through
accidental spills and municipal waste incineration were 50,000 and 5000 kg/y, respectively.
Johnson et al. (1992) identified the most important sources of PCBs to the atmosphere as:
1. Transformer leakage
2. Electric power generation
3. Industrial fuel combustion
4. Landfills
5. Sewage sludge combustion
6. Waste oil combustion
Transformer leakage/spillage estimates were based on a report for utility industry closed-
system equipment in the U.S. These losses as percentages of total PCBs contained in the
equipment were: large transformers (0.027%), large capacitors (0.42%). A loss rate of 0.24% for
85
-------
the industry was assumed. Of the spilled PCBs, 0.3% was estimated to be evaporated before
cleanup. The overall atmospheric emission factor was 7.2 mg for each kg PCBs remained in
transformers and capacitors (0.0024 x 0.003 = 7.2 x 10'6 = 0.00072%).
The information for (2) and (3) was based on very limited data on PCB emission factors
from bituminous coal combustion. These were 0.65 mg/metric ton coal for utility boilers and 3-
26 mg/metric ton for industrial coal-fired stokers. It is not known why coal combustion should be
a source of PCBs. Emission factors (mg/metric ton) for incineration sources were given by
Johnson etal as follows: industrial liquid waste 1.18, commercial waste: 2.5, hospital: 29, waste
oil 390, sewage sludge (0.69-14, 5.4 selected), municipal solid waste 1.3-4.5.
The PCB release estimates of Johnson et al from open sources (including transformer and
capacitor leakage and landfills), incineration, and stationary fuel combustion are summarized in
Table 15. Only 38 kg/y was attributed to solid waste incineration in ENA. This differs markedly
from a U.S. municipal waste combustion study (quoted by Voldner and Smith, 1989) which
estimated about 5000 kg/y from this source. The atmospheric release of PCBs to ENA from
transformer and capacitor leakage was 370 kg/y. Voldner and Smith quote a report which
estimated that 50,000 kg/y PCBs were spilled in the Great Lakes region. Applying the
evaporation factor of Johnson el al. (0.3%) yields 150 kg/y PCBs entering the atmosphere.
No inventory has considered re-emission of PCBs from soil, plants, and water; and this
may be a significant part of the mass balance. Also, as Voldner and Smith point out, the
availability and data quality for PCB emissions "leaves a lot to be desired"
When displayed on a gridded scale (Johnson et a/., 1992), highest emissions of PCBs
occurred in a broad belt extending from Illinois eastward to the mid-Atlantic and southern New
England states. In this region fluxes were generally 0.32 - 3.2 g/km2-y (= ug/m2-y). This is
remarkably similar to the 1.5-1.8 ng/m2-y precipitation and dry deposition fluxes out of the
atmosphere estimated for the Great Lakes (Eisenreich and Strachan, 1992).
Other areas with atmospheric emissions of the same magnitude were Chicago -
Milwaukee, southeastern Michigan, a band around southern and western Lake Ontario, and some
populated areas in the south. A few pockets with fluxes of 3.2 to >32 g/km2/y were identified.
Elsewhere fluxes were usually in the 0.032 - 0.32 g/km2-y range. Fluxes north of the Great Lakes
and in eastern Canada were <0.001 -0.032 g/km2-y.
86
-------
4.2. Emission profiles for major source categories
Emission source profiles provide a useful tool for quantifying the contributions of
pollutants from various sources in a given region. This is particularly true in the cases when
actual emission measurements are not available. Chemical mass balance (CMB) analysis is then
frequently used which requires information about source profiles. It is desirable to prepare a set
of emission source profiles for fine and course particles separately. The CMB receptor modeling
has been described in the literature (Miller et a/., 1972; Gordon, 1980, 1988; Winchester and
Nifong, 1971;Watsone/a/., 1984).
An important step of the CMB analysis is construction of emission source profiles. A
detailed description of emission sources and emission generating processes is needed first. This
information has been created for the Great Lakes region as mentioned above. Then the emission
source profiles have been constructed using information on emission factors for several trace
elements and persistent organic compounds, such as PAH, polychlorinated di-benzo dioxins
(PCDD), polychlorinated di-benzo furans (PCDF), and PCBs. The profiles were constructed on
the basis of emission factors calculated for the global emission survey for trace elements by
Nriagu and Pacyna (1988) and emission factors from the Ministry of the Environment (MOE)
Toxic Chemical Emission Inventory for Ontario and Eastern North America by Johnson et al
(1992).
The emission source profiles for the combustion of coal in utility, industrial, and residential
boilers, the combustion of residual oil in utility and industrial boilers, the production of copper,
lead, and zinc in primary and secondary plants, the production of iron and steel, the incineration of
municipal wastes and sewage sludge, the production of cement as well as phosphate fertilizers, the
combustion of wood in stoves, and fireplaces, and the combustion of gasoline and diesel oil in
mobile transportation are presented in Figures 25a-c.
The emission source profiles as shown in Figures 25a-c can then be used to construct
regional emission profiles for the Great Waters area by combining information on the source
profiles with information on contribution of a given emission source category to a total emission
of a studied compound in the region. Regional emission profiles are used to assess emission
contribution from various regions to the air concentration or deposition at a given receptor.
87
-------
TABLE 15. Atmospheric Emission of PCBs in Ontario and Eastern North America, kg/y.
Ontario
Open sources3 36.9
.., —-—-ration 1.3
Su... ._. -'bustion 12.2
Eastern
North America
453.2
38.2
434.9
a) Includes transfonv ?- & capacitor leakage (80%) and emissions from lanfills (20%).
Source: Johnson et il. <:?92, pp. 263, 266, 333).
-------
Coo: Combustion
(Power Plant)
1/5 _ _
< O O
100
TO -
Coal Combustion
(Industrial Units'
Combustior
Figure 25a. Emission source profiles for major source categories estimated with
cadmium as a reference element.
89
-------
183
Primary Cu-Ni Production
c.:::
a
as a reference element.
maj°r SOUrce
with
90
-------
MoDne Transportation
Leoded Gasoline
10
u
1 Unleaded Gasoline
(SI
I
o
o
o
CL
o
u
CL
10
10"
1C'
<
a.
a
a
o
a
o
u
a.
o
o
a
Q.
o
u
CL
Figure 25c. Emission source profiles for major source categories estimated with
cadmium as a reference element.
91
-------
4.3. Emission profiles for diffuse sources of organics
The earth's surface is a vast reservoir for pesticides and other organic chemicals.
Exchange of gaseous SOCs is a two-way process; and plants, soils, lakes, and oceans act as both
sources and sinks of airborne chemicals. Examples are volatilization of pesticides from sprayed
foliage vs. uptake of pesticides from the air by pine needles. Since these exchanges occur,
transport of pesticides is not a simple source-to-sink relationship. Deposition and re-evaporation
of the chemical may occur many times during transit, in what is becoming known as the
"grasshopper effect"
Understanding the processes underlying these cycles is critical to developing transport and
deposition models for pesticides and determining sources to-the Great Waters. Below three
processes are discussed, volatilization from soils, gas exchange with water, and gas exchange with
plants. The two-way nature is emphasized for the latter two. In the case of soil, uptake of
gaseous pesticides in non-target areas is poorly understood. The intent is to present an overview
of the processes and their significance. Mathematical details are given in the references.
4.3.2 Air -surface exchange processes
a) Volatilization from Soils
Although spray drift can have a local impact, post-application volatilization is a more
important pathway of pesticide loss from fields. Many early studies are referenced in work by
Nash and Hill (1990) and Spencer and Cliath (1990). These investigations have been done by
following pesticide disappearance from soils under field conditions, and also through controlled
experiments in agroecosystem chambers. Major losses can take place within a few days if the
chemical is applied to soil surfaces. Incorporation into the soil layer greatly reduces volatilization;
but even so, long-lived pesticides continue to evaporate. Spencer and Cliath (1990) state that:
"Even in areas where DDT use has been discontinued, the persistence of DDT residues is
sufficiently great that they will continue to be redistributed for many years. In these areas.
volatilization from the soil probably will be the main source of DDT components moving into the
atmosphere ..."
If not washed off by rain or irrigation, pesticides are quickly evaporated from foliage. In
the absence of precipitation, Willis el al. (1985) found half-lives for toxaphene, methyl parathion.
and fenvalerate on cotton plants of 0.7, 0.1, and 3.3 days. In another study (Seiber et al., 1979).
92
-------
toxaphene which was aerially sprayed on a cotton field disappeared rapidly from the foliage, with
59% lost in 28 days. Loss of toxaphene in the aerated top soil was slower, 51% in 58 days.
The fate of diazinon sprayed on a dormant peach orchard was followed by accounting for
the amount of pesticide in air, soil, and tree parts (Glotfelty et al., 1990). Spray drift and
volatilization during application were small, and were exceeded by evaporation losses over the
next two days. After the first 24 h, diazinon in the soil dissipated by a first-order process with a
19-d half life. The conclusion was that most of the diazinon in the atmosphere of California's
Central Valley comes from from volatilization.
A mass balance for a DCPA-treated onion field yielded a half-life of 40.5 days for the
herbicide. Flux measurements based on air samples taken above the plot showed that in that time
29% of the DCPA was lost by volatilization (Ross et a/., 1990).
In the tropics OC pesticides are heavily used and volatilization is the main loss process. A
mass balance of HCH in the Vellar River estuary (India) showed that of 42,000 kg HCH applied,
41,830 kg (99.5%) evaporated and only 170 kg was transported by the river (Takeoka et al.,
1991). Yeadon and Perfect (1981) found that DDT applied to soils in Nigeria evaporated with a
half-life of 9 days.
Field measurements of pesticide volatilization are time-consuming and costly. A classic
paper by Parmele et al. (1972) described several micrometeorological techniques and used them
to determine losses of heptachlor and dieldrin from bare soil and corn fields. Common methods
are briefly described below For details, see papers by Parmele et al. (1972) and Majewski et al.
(1989, 1990). All of these flux methods except the last (TPS) require measurement of the vertical
gradient in pesticide concentration.
Typically this is accomplished by collecting air samples at five or more heights above the
soil up to a few meters.
•Thornthwaite-Holzman, or Aerodynamic (AD): This is the most frequently used field technique,
and is based on accurate measurements of pesticide and wind speed gradients in the turbulent
boundary layer. Requirements are a large, uniformly surfaced area with similar land surrounding
it, and a long fetch (unobstructed upwind distance).
•Lysimeter and Energy Balance (EB): These methods describe pesticide fluxes (F, mass/area-
time) by the general equation:
93
-------
F = K2(dC/dz) (3)
where Kz is the vertical diffusivity and dC/dz is the pesticide concentration gradient above the
surface. Flux estimation methods assume that Kz of pesticide vapor and water vapor or heat are
the same. Kz is determined by direct measurement of water loss from a weighing Lysimeter, or
(more commonly) by energy balance. In the latter technique, vertical gradients of temperature
and humidity above the surface are used with measured radiation and soil heat fluxes to deduce
Kz. Field size requirements are the same as those for the AD method.
The method is simple, but the EB instrumentation is sophisticated.
• Integrated Horizontal Flux (HF): Fluxes from the field can be calculated from the vertically
integrated pesticide concentration and wind profiles. This was used by Glotfelty et al. (1990) to
determine transport of volatilized diazinon from a treated orchard:
F = (1/R) |0u«dz (4)
In Equation (4), C and u are average pesticide concentrations and wind speeds at a particular
height, z. R is the distance the wind has to travel over the treated surface. In practice, the
integral is estimated by summing over finite height intervals. The method requires a uniform
surface and a uniform source strength, but is independent of assumptions regarding the
equivalence of pesticide, water vapor, and heat Kz values. Also, fetch requirements are not as
critical. However, IHF gives no information above the highest and lowest sampling point.
• Eddy Correlation (EC): In the EC method pesticide fluxes are determined by measuring short-
term fluctuations in vertical heat flux. The technique requires meteorological equipment that can
respond rapidly to small changes in temperature, wind speed, and wind direction. As in the EB
method, it is assumed that Kz are the same for heat and pesticide vapor. Disadvantages are
complexity of instrumentation, need for fast-response sensors, and precise alignment
requirements.
• Theoretical Profile Shape (TPS): A trajectory simulation model predicts that there is a point
above the center of a circular-shaped source where the gaseous horizontal flux to the atmosphere
can be determined from measurements of wind speed and mean pesticide air concentration at that
point. In contrast to the above methods, TPS requires measurement of wind speed and pesticide
concentration at only one height. A drawback is that field plots must be circular.
94
-------
AD, EHF, EB, EC, and IPS methods were compared in a field test of volatilization from
fallow soil of four pesticides: lindane, diazinon, nitrapyrin, and chJorpyrifos (Majewski et a/.,
1990). No statistical differences were found at the 95% confidence level for all methods and
compounds. However the authors cautioned that it is not clear whether any of the methods are
accurate ~ that is, if they describe what is really occurring. Careful mass balance experiments
were suggested as a way to calibrate flux estimation methods.
The expense and time required to do these studies have greatly limited the number of field
experiments that have been carried out over the years. Moreover, results are likely to be highly
dependent on local soil properties, meteorological conditions, and pesticide application methods.
Concurrent with the development of methods for flux estimation have been experiments to
identify the factors that influence pesticide volatilization. Much work has been done in field plots
and environmental chambers and to follow pesticide dissipation from soils. Recent studies and
previous work are described by Nash and Hill, 1990; Spencer and Cliath, 1990; Clendening et al.,
1990; Jenkins et al., 1990; and Woodrow etal., 1990.
In a series of four papers, Jury et al. (1983, 1984a-c) described the development and
application of a model for assessing the behavior of organic compounds in soil. Losses of
chemicals from soil are the result of leaching, volatilization, and degradation. The Jury model
incorporates all three processes. To apply the model, it is necessary to know:
• Properties of the chemical. Henry's law constant (H), soil organic carbon - water partition
coefficient (Koc),
Diffusion coefficients in water and air, reaction rate coefficients (parameters of first-order
loss equations that describe chemical and microbial reactivity). The latter are difficult to assess.
• Soil properties, porosity, bulk density, volumetric water content.
In addition, fluxes are affected by the amount of chemical applied, depth of incorporation
into soil, temperature, and relative humidity.
The Jury model views pesticide evaporation as being controlled by two resistances, the
soil layer and the air boundary layer.
That is, emissions of gaseous pesticides are related to: a) the supply of pesticides from
below to the soil surface, and b) the volatilization rate from the soil surface into the atmosphere.
95
-------
The former process occurs by vapor-phase diffusion in soil pores and movement of dissolved
chemical upward with water flow.
For some chemicals water evaporation from the soil is critical, because it causes an
upward flow of water which trans-ports chemical to the surface. If water flow carries an
insignificant amount of chemical relative to vapor diffusion, water evaporation has less effect.
Jury el al. (1984a) and Spencer and Cliath (1990) show how, based on their physicochemical
properties, substances can be classified by the effect of water evaporation on their volatility from
soils (Figure 26):
• Class I. High Henry's law constants (OCs, EPTC, triallate, trifluralin). These chemicals
volatilize from the surface faster than mass transport can replenish them from below. Control
thus resides in the soil. Volatilization rates are increased by evaporation of water for some of
these compounds, but not others. The influence of water flux is less than for Class II and III
compounds. Volatilization rates decrease with time.
• Class III. Low Henry's law constants (atrazine, prometryn, napropamide, 2,4-D). These move
to the surface in water flow faster than they can be lost to the atmosphere through the boundary
layer. Without water evaporation their volatilization rates are very low. As water evaporates,
pesticide concentrations build up at the soil surface and volatilization rates increase with time.
• Class II: Intermediate in behavior (methyl parathion, parathion, ethoprophos)
From Figure 26 and the further description of chemical classifications (Jury et al., 1984a),
it is fair to say that for most pesticides water evaporation will play an important role in loss from
soils.
In the final paper of the series (Jury et al., 1984c) the authors compare results of the
behavior assessment model to volatilization data from laboratory chamber experiments. Very
good agreement was obtained, and the effects of water evaporation for chemicals in the different
volatility classes were correctly predicted. Despite this, Jury et al. (1984b) stressed that the
model is to screen chemicals for their behavior in soils, not to simulate field results. In field mass
balance studies (Clendening et al., 1990) volatilization behavior of five Class I and III compounds
was in accordance with the Jury model, however deep migration by leaching was difficult to
predict.
Several field models have been developed to simulate pesticide volatilization. A six-
compartment model was formulated by Nash and Hill (1990) to take into account changing
96
-------
10'*
o
•••*
o>
fsi
O
CATEGORY I
E :2.5mrr\/
-------
mechanisms of loss over time. When first applied to a soil surface, pesticides volatilize quickly.
Later the loss rate slows down, in pan due to adsorption of chemical to the soils.
Woodrow et al. (1990) investigated volatilization of three herbicides and methyl parathion
from flooded rice fields in California's Sacramento Valley. Good agreement was found between
measured loss rates and those predicted by EPA's EXAMS model. A Gaussian plume dispersion
model (ALOHA) was used to estimate transport of the chemicals out of target areas. Measurable
air concentrations of pesticides in residential and business areas were due to drift during
application, post-application volatilization, and photooxidation of pesticides during transport
(conversion of methyl parathion to its oxon).
A volatility model that takes into account changing meteorological conditions, soil
temperature and moisture, and water evaporation rates over the period of interest was developed
by Scholtz and Voldner (1992), who then applied it to the evaporation of three pesticides
(lindane, chlordane, and 2,4-D) from sandy loam soil. Modeled daily emission rates and the
cumulative fraction of pesticide lost over a 6.5-month period are shown in Figure 27. Although
the Henry's law constant of the pesticides change with temperature, including this temperature
dependence had little effect on the results. Differences in water evaporation rates had the
strongest influence on daily fluxes.
b)._ Air-Water Gas Exchange
In the frequently used "two-film" model gases are exchanged by diffusion through thin air and
water films on either side of the air-water interface. The rate at which molecules diffuse through
these films is slow compared to convective mixing in the bulk air and water, and thus the interface
acts as a resistance to transfer. The equations for describing flux (F = mass/area-time) of material
have the form:
F = KAC (5)
where AC is the concentration gradient (that is, difference in concentration from the top to the
bottom of the air or water film). The overall mass transfer coefficient (K) takes into account the
resistance of both films. Algebraic expressions for K include transfer constants for the individual
air and water phases and the Henry's law constant of the chemical. Often the resistance of either
the air or water will dominate, and the chemical is said to be "gas phase" or "liquid-phase"
controlled. Which situation applies is largely related to the Henry's law constant.
98
-------
DAILY EMISSION RATE. SANDY-LOAM
SPAAY APPUCATON: 1 i-D; Henry's U* coefficient s function of
), independent of icmpeniurc (dotied).
CUMULATIVE EMISSION: SANDY-LOAM
SPRAYED APPUCA.TlON-.1kg/ha.06h April 15
2.4-D
MOTTH
Figure 27. Modeled daily emission rates and cumulative fractions emitted for pesticides
sprayed on sandy loam soil. Top: lindane; Bottom: lindane, chlordane, and 2,
4-D. Source: Scholtzand Voldner (1992).
99
-------
The detailed equations for two-film gas exchange of pesticides, PCBs, and PAHs have
been presented by many authors, including Baker and Eisenreich (1990), Corlidi?! and Bidleman
(1991), WMO (1989), Hinckley el al. (1991), and McConnell et al. (1992).
As a way of quantitatively describing water bodies as sources or sinks, the model many
shortcomings. Major unknowns are:
• Mass transfer coefficients as a function of wind speed.
• The role of bubbles and breaking waves in the exchange.
• Henry's law constants of SOCs as functions of temperature and salinity.
• Physical states of SOCs in air and water: To employ the model, it is necessary to know the
fraction of gaseous and dissolved compound. Further method development is needed to
speciate SOCs in air and water into gaseous, paniculate, colloidal, and dissolved
components.
Calculations based on the two-film model indicate that gas exchange of SOCs is an
extremely important factor in the Great Lakes budget. Murphy (1983) proposed evasion of PCBs
from the Great Lakes, and a subsequent mass balance study concluded that the net flux of PCBs
and DDT was out of the lake on an annual basis (Strachan and Eisenreich, 1988). At that time
these conclusions were based on limited air and water data. Mass balance models of Siskewit
Lake also show the importance of volatilization (McVeety and Kites, 1988; Swackhamer et al.,
1988; Swackhamer and Kites, 1988).
Recent paired air and water measurements have provided much more insight to this
process. Data from Lake Superior indicated volatilization of PCBs during late summer (Baker
and Eisenreich, 1990). Intensive studies in Green Bay showed that PCBs volatilized from the
surface water at rates ranging from 13 - 1300 ng/m2-d. Moreover, the profile of individual PCB
congeners was different in the air over Green Bay than over land (Achman et a/., 1992;
Hornbucklee/a/., 1992).
Concurrent sampling of air and water in Green Bay and the lower four Great Lakes
allowed McConnell et al. (1992) to determine the direction of HCH exchange. Transfer of HCHs
to Green Bay was air-to-water (deposition) in early June. In August when surface waters were
warmer, the flux direction was reversed (volatilization) in Michigan, Huron, and Erie, but
remained depositional in Ontario. Based on the annual cycle of HCHs in the atmosphere (Hoff et
al., 199la), the authors predicted that transfer of gaseous HCHs was into the lakes for most of
the year, with short periods of volatilization in late summer. The loading of HCHs to lakes
Michigan, Huron, Erie, and Ontario by gas deposition was estimated to be 370 kg/y. By
100
-------
comparison, 970 kg/y entered the four lakes by precipitation and dry panicle deposition
(Eisenreich and Strachan, 1992).
Gas exchange contributes significant quantities of OCs to the oceans. The GESAMP
report (WMO, 1989) concluded that air-to-water gas transfer of OCs accounted for 24-72% of
total atmospheric deposition. Gotham and Bidleman (1991) estimated that about 110 metric
tons/y of HCHs entered the Arctic, about two-thirds of which came from gas uptake by the
surface ocean. Gas exchange formed 20-60% of the total atmospheric budget for other OC
pesticides.
c) Air - Plant Exchange
Plants accumulate OCs from the atmosphere, and many investigators have suggested their
use to survey atmospheric contamination in non-agricultural and remote regions. The case of pine
needles as an indicator of DDT transport from eastern Europe was mentioned earlier (Section V-
C-lc). Examination of different plant species from around the world revealed especially high
levels of DDT, HCHs, and PCBs in China. DDT was also high in plants from several African
countries, and in Russia near Moscow. Lowest concentrations of most pesticides and PCBs were
found in lichens from the Antarctic Peninsula (Bacci et al., 1986, 1988; Gaggi et ai, 1985;
Villeneuve el ai, 1988).
Accumulation of pesticides and PCBs by plants is important for several reasons. The
potential for using plants as a monitoring tool was mentioned above. Plants are also at the base of
the terrestrial food chain. Transfer of OCs from lichen to reindeer to man was investigated in
northern Sweden (Villeneuve et a/., 1985). The percentage of transfer from lichen to reindeer was
positively correlated with the octanol-water partition coefficient (Kow) of the compound.
Finally, the earth's plant biomass is a large reservoir for persistent organic compounds.
Accumulation from the atmosphere into plants is related to the proximity of sources (i.e.,
concentration of chemical in air) and temperature. The strong role of temperature was shown by
Calamari el al. (1991), who surveyed plants from around the world for OC pesticides and PCBs.
HCB concentrations in plant tissue were highest in the Arctic and Antarctic and lowest in
equatorial regions. For this compound the "source factor" can be eliminated, because
concentrations of HCB in the troposphere are uniform, varying by no more than a factor of 2-3.
Levels of HCHs and DDT were far higher in temperate and tropical zones than at the poles,
showing the strong dependence on regional use of these pesticides.
101
-------
Recent work has focused on the rate and mechanism of atmospheric uptake and
depuration of OCs by plants. Like fish in water, plants take up chemical from a polluted
atmosphere and release it when placed in clean air (Bacci et al., 1990a,b). The steady-state
bioconcentration factor (BCF = concentration of chemical in the plant divided by that in air) is
related to two common physicochemical properties, Kow and H:
BCF = constant • K^RT/H (6)
Reviews of accumulation processes and comparison of different models have been written by
Paterson et al. (1990) and Schonherr and Reiderer (1989).
The idea that large areas of "green space" on the planet can take up and release pesticides
to the atmosphere has important consequences for long-range transport and seasonal cycles of
pesticides in the atmosphere.
4.4. Evaluation of emission inventories. Comparison with European studies
In general, emission inventories have only recently been compiled for toxic heavy metals
and persistent organic compounds, with the exception of lead. Emission inventories were first
used to evaluate the environmental impacts of the emissions from single point sources, thus were
of local importance. In the late 1970's, it was recognized that heavy metals can be transported for
distances of up to a few thousand kilometers. This suggested the need for regional, and even
global, emission inventories to be formulated for these persistent pollutants. When the parameters
affecting the emission quantities were defined for this study, it became clear that a regional
emission inventory for heavy metals should be prepared using so-called bottom-up approach. In
other words, major emission sources should be identified and emission quantities assessed in sub-
regions, e.g. administrative units such as county or city, and then added in order to obtain
emissions within a region. In this way a detailed assessment of emission sources can be collected
at a county or city level.
Emission measurements are often carried out at specific sources and their results are then
reported at a sub-region level. Indeed, measured data have been used to assess emissions of toxic
pollutants in various states in the study region. This applies particularly to emissions from single
point sources, and mainly waste incinerators. The emission data obtained through such
approaches are usually accurate providing that the measurements are representative with respect
to sample collection and analytical methods used. Thus, the above approach is recommended for
collection of emission data from major point sources, including large power plants (e.g. over 1000
102
-------
MW electricity), heat boilers (e.g. 200 GJ/h capacity), primary and secondary non-ferrous metal
smelters, cement kilns and already mentioned waste incinerators. In Europe, there are already
regulations requiring emission data to be measured for large power plants within the parties of the
Commission of the European Communities. Standardization of measurement methods and
harmonization of reporting procedures are important factors when the bottom-up approach is
used to obtained emission data. These factors need to be employed in order to assure the
comparability of the emission data to be used to construct the emission inventory for a given
region.
There are currently very few measurements, on the emissions of hazardous air pollutants
from sources within the Great Waters regions. One reason for this could be that on the national
scale in the U.S. there is no formal obligation to undertake an emission survey for toxics, the TRJ
program being voluntary. Only some states have taken the initiative to prepare emission surveys,
mostly using emission factors from the literature and statistical information on the production of
industrial goods and the consumption of raw materials. The measured data reported for the study
region are more accurate than the estimated values. The measured source profiles provide
information which may be used when estimating emissions for regions with no measurements.
The second approach used to prepare an emission inventory is called a top-down approach
and is used when the inventory is based on emission estimates. The emission inventories prepared
by EPA, IJC, and the Canadian authorities, and used in this report, are based on the top-down
approach. Major emission source categories for a larger region, e.g. a state are defined in the first
step, emission factors are selected, and emission quantities calculated. Further division of these
estimates in order to obtain spatial distribution of emissions is carried out using either surrogate
parameters, such as population density maps for distribution of area source emissions or a list of
point sources with information on geographical location and emission data for other compounds,
such as sulfur and nitrogen oxides, volatile organic compounds, and total suspended panicles.
The NAPAP database provides aJarge body of information in this respect for the sources in U.S.
and Canada. Similar approach is used in Europe to assess sources and fluxes of atmospheric
heavy metals and persistent organic pollutants within various activities of the UN ECE Task
Force on Heavy Metals Emissions and Task Force on Persistent Organic Pollutants, as well as
within PARCOM and RELCOM.
Several parameters affect the accuracy of emission data prepared using emission factors.
The emission factors need to be representative for a given source category, sub-category, and
industrial activity. Therefore, the quality of emission data are dependent upon selection and
utilization of emission factors which are transparent with respect to the conditions, both technical
103
-------
and meteorological, at the time the factors were determined. Various sets of emission factors
were used to calculate emission quantities for the study area by EPA and the Canadi - i authorities.
Emission factors used in the preliminary version of the EPA emission inventory for neavy metals
in the United States (Benjey and Coventry, 1992) seem to overestimate the emissions of mercury,
arsenic, cadmium, and particularly lead. They are higher than the emission factors used in Canada
(Voldner and Smith, 1989; Johnson et a/., 1990) and Europe (Axenfeld et al., 1992; Pacyna,
1986). The preliminary EPA emission factors are now being revised. They need to be validated
and cross-checked. For example, methods proposed recently within the UN ECE Task Force on
Emission Inventories could be used for verification of emission data for heavy metals and
persistent organic compounds (Mobley, 1992).
In summary, the process of collecting information on emissions for heavy metals and
persistent organic pollutants from important sources in the Great Waters region has begun.
However, it will take some time before an accurate emission inventory becomes available. This
conclusion is made on the basis of the existing data for the region and research plans, including
those within the international organizations dealing with the emission inventorying for
atmospheric heavy metals and persistent organic compounds. Therefore, the analyses and
particularly conclusions based on presently available emission data for the study region should be
considered with some caution. This situation is similar in Europe in terms of atmospheric input of
heavy metals and persistent organic compounds to the North Sea and the Baltic Sea. The overall
conclusion from the research in Europe is that the current state of knowledge on emission fluxes
of the above pollutants needs further improvement in order to satisfy requirements posed by
policy makers and the scientific community.
4.5. Application of source-receptor techniques to study the origin of pollution
Although the source apportionment techniques are most widely used to assess the sources
of air contamination at distant locations, chemical mass balance model can be applied to identify
contributions from local sources, such as in an urban area (Scheffand Wadden, 1991) . There
have been attempts to construct both static and dynamic mass balances for various pollutants in
the study area in order to quantitatively account for the pollutant loadings and the flows through
the environmental system. The mass balancing method has been applied in the Great Lakes basin
to manage phosphorus load reductions to Lake Erie, and to elucidate behavior in smaller regions,
such as Saginaw Bay, as reported by Mackay (1992) for the Virtual Elimination Task Force.
The current level of knowledge of loadings, water and sediment concentrations, and biota
concentrations is not satisfactory and several assumptions need to be made when preparing the
104
-------
chemical mass balance for the study area. We are also limited by the available knowledge on the
behavior of pollutants in the environment. The physical and chemical processes that are important
depends upon the pollutant being investigated, and varies drastically from organics to metals, and
even more for mercury.
5. IDENTIFICATION OF DISTANT SOURCES
Atmospheric deposition has been identified as one of the major, if not dominant, pathways
for toxic heavy metals and persistent organic pollutants measured in the Great waters. However,
these is no definitive information available to the authors of this report which indicates the relative
contributions from local and distant sources. There is presently a limited amount of information
on the transport and deposition of HAPs to the Great Lakes-(Rice et a/., 1986; Voldner and
Schroeder, 1990) which originated outside the Great Lakes Basin. However, there has recently
been a number of studies which investigated the long-range transport and deposition of
organochlorines (OCs) to remote areas such as the Arctic. Regional transport of metals, including
mercury, and other toxic compounds is also the subject of intensive studies at EPA at present
(e.g. Clark, 1992). While this compelling evidence gives indirect proof that this phenomenon is
important to the Great Waters, there is still a lack of quality data on the hazardous pollutants to
support a definitive conclusion. However, the results from two research programs: National Acid
Precipitation Assessment Program (NAPAP) and the Baltic Sea Environmental Program in
Europe can provide clues on the potential contribution of distant source emissions to the
atmospheric deposition in the Great waters region.
NAPAP focused mainly on the deposition of acidic compounds and their precursors. The
atmospheric transport and deposition of sulfur is discussed here as an example for the transport
and source-receptor relationships for pollutants that undergo chemical transformation and that are
on particles including heavy metals and persistent organic pollutants. The transport and chemical
transformations for organics and metals, such as mercury, may be quite different and direct
analogies should not be drawn at this time. NAPAP (1990) concluded that for receptors in the
eastern United States, more than 70% of the total deposition of sulfur originates from sources
within 500 km of the receptors and dry deposition of sulfur species contributes more than half to
the total deposition at the average receptor. In Canada, about two-thirds of the total deposition
at average receptors originate from source areas at distances greater than 500 km from the
receptors. Wet deposition constitutes about two-thirds of the total deposition (except receptors
close to large point sources of sulfur emissions). As the main sources of heavy metals and
persistent organic compounds are south of the Great Lakes Basin and these pollutants are more
105
-------
resistant to wet deposition, one can hypothesize that emissions from regions outside the Great
Waters are very important when performing source apportionment studies for th. Great Waters.
While the evidence to date strongly suggests that distant sources are contributing to the
contamination of the Great Lakes and other bodies, further research is needed to prove this
hypothesis and to quantify the relative loadings for specific pollutants.
There has been a great deal of work looking at similar problems facing large bodies of
water in Europe. The Baltic Sea and the Great Waters can be compared taking into account
similarities in the level of water contamination by heavy metals and persistent organic pollutants,
and emission source categories contributing to this contamination, although the meteorological
conditions, and particularly precipitation patterns are quite different. The Baltic Sea Program
concluded that major source regions contributing to the atmospheric deposition of heavy metals
and persistent organic compounds except for pesticides are located several hundred kilometers
from the Baltic Sea shore (Pacyna, 1992). Most of the major contributing point sources identified
within the study were at least 500 km far from the Baltic shore. Even so, emissions from these
sources and source regions were found to contribute the majority of the pollution load from the
atmosphere. The above results confirm that the regions with high emissions of toxic metals and
organic compounds can contribute to the contamination of the environment at remote receptors.
When diagnosing the major sources of contaminants in the Great Lakes Basin, for
example, the large emissions of arsenic, cadmium, lead, mercury, and other heavy metals in
Missouri (Benjey and Coventry, 1992) can be transported and deposited in the basin. Data
supporting the transport of toxic compounds from the St. Louis area to Lake Michigan were
obtained during the Lake Michigan Urban Air Toxics Study performed during the summer of
1991. The largest air pollution episode observed during the study was associated with mixed-
layer transport from the general area around St. Louis, MO to the southern Lake Michigan area
(Keeler el ai, 1992). Measurements of all of the toxic air pollutants discussed in this report were
made at several sites around Lake Michigan. Preliminary results suggest that the iron-steel and
other metallurgical activities in the St. Louis/Granite City area were probably the largest sources
of heavy metals measured in Chicago and South Haven, MI. Additional research is needed to
further support this hypothesis both with further measurement and through the application of long
range transport models. The major difficulty in applying dispersion models to this problem is
obtaining a set of reliable emission data within the specified grid system.
106
-------
5.1. Emissions from North America outside the Great Waters Regions.
Contribution of heavy metals and persistent organic pollutants from North American
sources located outside the Great waters study regions can be assessed with the help of source-
receptor techniques. However, the results would indicate contribution of either emissions
generated within major source categories or arriving from major source regions if the source -
receptor techniques are used together with information on meteorological conditions. Emission
data for major point and area sources within the U.S. and Canada are needed in order to verify the
results of source-receptor techniques and can be used for developing strategies to reduce the
pollution load to the Great waters. The preliminary results from the EPA and the IJC data (e.g.
Voldner and Smith, 1989) indicate that four states lead the emissions of heavy metals generated
from the production of energy and industrial goods. The states include: Arizona, Louisiana,
Missouri, and Texas. Primary copper smelters in Arizona result in high emissions of arsenic,
mercury, and cadmium. Secondary smelters, refineries, and lead alky] plants emit large quantities
of lead in Texas and Louisiana. Combustion of fossil fuels, and primarily coal in Missouri,
produce large amounts of all studied heavy metals. Industrial activities in Nevada, Montana, Utah
and New Mexico also generate substantial amounts of atmospheric heavy metals. Recent EPA
study on toxics in the community (1990) listed a ranking of states emitting toxic pollutants to the
air in 1988. The list was topped with Texas followed by Ohio, Tennessee, Louisiana, Virginia,
and Utah. The list of the top ten was completed by the states in the study area: Indiana, Illinois,
Michigan, and New York. Quebec and Manitoba in addition to Ontario generate the largest
amounts of studied heavy metals in Canada (e.g. Voldner and Smith, 1989).
Primary emission sources of PAHs to the atmosphere include combustion of fuels for heat
and power generation, transportation, solid waste incineration, industrial processes such as coal
and coke processing and petroleum refining. Therefore, their sources are similar to the sources of
mercury (except transportation) and the states of Missouri, Texas, and Louisiana generate the
largest PAH emissions.
The ubiquitous past use of PCBs as well as current generation led to widespread
geographical distribution of these pollutants in U.S. and Canada. The leaks and spills from current
use all-over U.S. and Canada appear to be the largest source of PCBs to the environment.
The major primary sources of PCDDs and PCDFs to the atmosphere include the
combustion of munitipal and industrial waste in many locations in U.S. and Canada. Therefore,
the emissions of these pollutants are quite ubiquitous. In addition, it has been speculated that the
dominant sources of PCDDs to the total environment could be manufacturing of chlorophenols
107
-------
and their derivatives and the disposal of chemicals containing these wastes (e.g. Voldner and
Smith, 1989) Although the locations of the companies which have been major p ucers and
formulators of these products is known, no data is available to speculate on the likely emission
amounts.
Information on the use of lindane in U.S. and Canada is very limited. Its application in
agriculture is banned but is probably used in livestock treatment. The use pattern of lindane in
U.S. and Canada is largely unknown.
It is postulated that large sources of atmospheric emissions of heavy metals and persistent
organic pollutants are located in Mexico, and that the effects of these sources are being felt in
Great Waters areas (Eisenreich and Strachan, 1988). Two attempts have been made within the
Global Emission Inventory Activities (GEIA) program of the International Geosphere Biosphere
Program (IGBP) to assess the sulfur dioxide and nitrogen oxides emissions in Mexico as a part of
the global emission inventory for these pollutants. The assessment has been made in co-operation
with the Mexican experts and it was concluded that the sulfur and nitrogen emissions divided by
the number of inhabitants was one of the largest in the world. As the combustion of fossil fuels is
by far the most important source of sulfur emissions, it can be expected that substantial amounts
of mercury are also emitted into the atmosphere from the Mexican power plants, particularly due
to the lack of desulfurization installations on these Mexican plants.
Large quantities of nitrogen oxides are emitted in Mexico from the combustion of
gasoline. The Mexican gasoline contains lead in amounts of 0.4 g/liter and results in emissions to
the atmosphere which are also relatively high. Recently, a preliminary study was been carried out
at the Norwegian Institute for Air Research to assess global emissions of atmospheric lead. The
study is a continuation of global emission inventory development for which preliminary results
have been published by Nriagu and Pacyna (1988). It has been estimated that at least 5000 t of
lead is emitted from gasoline combustion in Mexico each year, compared to about 15000 t in U.S.
in 1985. Current emissions of lead in U.S. are significantly lower than those in 1985 and probably
do not exceed 7000 t per year.
Mexico is also an important producer of non-ferrous metals and particularly copper, lead.
zinc, cadmium, arsenic, silver, gold, antimony, and bismuth. There are several companies in the
country producing the above metals with the two major ones being Industrial Minera Mexico SA
de CV, and Industrias Penoles SA de CV Substantial amounts of atmospheric emissions of all
heavy metals of concern for this report are expected from the concentrating facilities in Santa
Eulalia (Chih.), Taxco (Gro.), Rosario (Sin.), Velardena (Dgo.), Santa Barbara (Chih.), and
108
-------
smelters and refineries in Nueva Rosita (Coah.), San Luis Potosi (SLP), Chihuahua (Chih.),
Monterey (NL), and Torreon (Coah). In addition, there are several secondary non-ferrous metal
plants in the country generating atmospheric emissions of. heavy metals. Although quantitative
information was not available to the authors of this report, knowledge on industrial technologies
employed in the above facilities and information on emission control equipment, although very
limited, indicates that emissions from Mexican non-ferrous metal industry are higher than those in
U.S..
There are several aluminum plants in Mexico operated by major companies such as
Almexa Aluminio SA de CV, Aluminio SA de CV, Laminadora de Aluminio SA de CV - Lasa,
and Reynolds Aluminio SA. It is expected that substantial emissions of PAHs are generated in
these plants, particularly in smelters employing the Soderberg process to produce aluminum, such
as the Aluminio SA de CV smelter in Veracruz. An emission factor ranging from 500 to 5000 g
PAH / ton of Al produced was estimated for this process (e.g. Axenfeld el a/., 1992).
The emissions from the above described sources in North America but outside the study
region can be transported and deposited to the Great Waters areas. One of the methods utilized
to study the long range transport of heavy metals and organic pollutants are regional transport
models. There are two major groups of input data needed to employ these models: gridded
emissions data and meteorological input data including wind and precipitation observations from
surface and upper air meteorological sites. The emissions data are always more difficult to obtain
and as such they should be given a priority in future research plans. This task requires not only a
financial support but also international co-operation.
5.2. Emissions from sources outside of North America.
Measurements in the Great waters region show presence of pollutants which are banned in
U.S. and Canada entirely or partially. One of this pollutants is y-hexachlorocyclohexane, an
insecticide called lindane. An interesting question is what may be the origin of lindane measured
in regions where its use is banned.
There is rather limited information on the use of lindane in the literature although this
pesticide is widely used in various parts of the world. The Food and Agriculture Organization
(FAO) statistics indicate that large amounts of lindane have been used by several countries during
the last two decades,, particularly in Asia. Although it is difficult to obtain exact values , the
production and use of lindane in India and China is at a level of tens of thousand of tons per year
in each of these countries (as indicated by Semb and Pacyna, 1988). Similar amounts a*re believed
109
-------
to be used on the former Soviet Union as reported by Komarov (1980). Lindane is used in the
above mentioned countries to improve the agricultural crop yield and to reduce d.^. Lindane
has been also used in Europe and Australia (reported by Semb and Pacyna, 19Sb after other
authors).
Recent investigations of the Arctic air pollution have shown that lindane has been present
in appreciable amounts in air samples collected at the ground stations as well as with the use of
the aircraft (Pacyna and Oehme, 1988). The authors suggested the Asian continent as the
probable source of the lindane. During the spring storms in the Asian deserts fractions of dust
with small particles can become airborne and subject to long-range transport within air masses.
Spring is the period with extensive use of lindane and other pesticides in Asia. Therefore, in it
reasonable to suggest that lindane is taken up with dust particles'and transported out of the region
(Pacyna and Ortar, 1989). It is suggested here that lindane measured in the Great waters region
can originate partly in Asia through the process described above. A possibility of long range
transport of air pollutants from the Asian continent over Hawaii to the western United States and
then further east has been already proposed to study, however no support was obtained. It should
be added that our understanding of meteorological processes governing the movement of air
masses and wind patterns do not exclude the possibility of the air mass transport from Asia to
North America.
Of course, other source regions can not be excluded when discussing the origin of lindane
in the Great Waters region, e.g. the Latin American countries. It is well established that certain
organochlorine compounds, such as DDT and PCBs can be deposited to lakes far removed from
their source (Hites and Eisenreich, 1987). The concentration ratio of a- to y-
hexachlorocyclohexane (HCH) in air masses can be a helpful tool to estimate the residence time of
aerosols and then to assess source regions. The y-HCH is photochemically transformed to the cc-
isomer. Pacyna and Oehme (1988) concluded that the ratios of 50 and higher would indicate old
air masses. These values would be expected in the Great waters region if the HCHs were to be
from the sources in Asia.
It is rather difficult to expect emissions from parts of the world other than mentioned
above to contribute directly to the deposition in the study region. One theoretical possibility
could be an episodic transport of pollutants with the Arctic air masses. Emissions from various
source regions contribute to the contamination of Arctic air, especially from sources in northern
part of the former Soviet Union and Europe - so called Eurasian sources. One of the major
scientific questions still to be answered by the Arctic researchers is what happens to the pollution
load entering the Arctic region. A part of this load is deposited to the surface but meteorological
110
-------
conditions in the Arctic do not favor efficient deposition of the pollutants. Thus one theory
suggests that a part of the pollution brought to the Arctic is carried-out of the region (e.g. Pacyna
and Shaw, 1991). Could part of this pollution load reach the Great Lakes and other Great Waters
under certain meteorological conditions? If so, what is the frequency of occurrence? These
questions need to be addressed in the future research activities in both the Great Waters regions
and the Arctic.
5.3. Application of source apportionment techniques for identification of impacts of
emissions on the Great Waters from distant source regions.
There have been only a few studies to date which attempted to diagnose the sources of
toxic air pollutants in the Great Lakes basin. Mamane et al. (1992) carried out a systematic study
of the sources of pollutants and toxic compounds measured in the air over Green Bay as part of a
larger study of toxics deposition to Green Bay. There findings suggest that regional source
influences were the dominant contributor to the paniculate mass measured during the study. They
also identified incinerator emissions impacting the measurement site as the concentrations of Cl,
K, and Pb in fine particles were 2 to 3 times higher than average in these samples. Confirmation
that one or more incinerators contributed to the metals measured on particles in Green Bay was
found by SEM analysis of individual particles collected on the filters. However, this study did not
directly quantify the sources of the pollutants deposited to Green Bay.
More recently, Clark has utilized the RELMAP model to calculate the transport and
deposition. The RELMAP model is a lagrangian dispersion model which relies upon the
availability of accurate emissions data for the compounds of interest. Utilizing dispersion models
for defining source-receptor relationships offers the distinct advantage that one can directly
calculate the contributions from various sources to the actual deposition to the lakes. The total
deposition (dry + wet) of Pb to Lake Michigan was estimated by Clark (1992) and the source-
receptor relationships are given in Figure 28. Cells which had sources contributing more than 2%
of the total annual atmospheric deposition directly to Lake Michigan are indicted on the Figure.
The preliminary modeling suggests that sources of Pb outside of the Great Lakes Basin are in fact
the major contributors to the Pb deposition to Lake Michigan. Figure 29 shows the relative
contributions of sources contributing to the deposition of Cd to Lake Michigan. Figure 30 shows
the relative contributions of sources contributing to the deposition of benzo(a) pyrene to Lake
Michigan. The relative contribution pattern for this PAH compound is quite different than those
seen for Pb and Cd. The contribution of BaP is primarily from local sources in the Chicago/Gray
area.
Ill
-------
40-KM EMISSION/DEPOSITION GRID FOR MODELING TOXICS
95
PERCENTAGE CONTRIBUTION TO LAKE MICHIGAN DEPOSITION
LEAD
Figure 28. The relative contribution patterns for total annual deposition to Lake
Michigan. Preliminary modeling estimates from Clarke (1992).
112
-------
40-KM EMISSION/DEPOSITION GRID FOR MODELING TOXICS
95
PERCENTAGE CONTRIBUTION TO LAKE MICHIGAN DEPOSITION7
CADMIUM
Figure 29. The relative contribution patterns for total annual deposition to Lake
Michigan. Preliminary1 modeling estimates from Clarke (1992).
113
-------
40-KM EMISSION/DEPOSITION GRID FOR MODELING ~'XICS
95
PERCENTAGE CONTRIBUTION TO LAKE MICHIGAN DEPOSITION
BENZO(A)PYRENE
Figure 30. The relative contribution patterns for total annual deposition to Lake
Michigan. Preliminary modeling estimates from Clarke (1992).
114
-------
Rahn et al. (1989) applied their regional signature technique to the aerosols and
precipitation samples collected in Underhill Vermont. They estimated that approximately 80% of
the sulfate and selenium originated from sources located in the Midwest whereas 20% of the
sulfate and selenium deposited to the Lake Champlain basin were derived from local sources on
the East Coast. Keeler and Samson (1989) applied a hybrid receptor model to investigate the
utility of using the "regional signature" approach and concluded that variations in elemental ratios
may reflect different mixes in emissions of sources in different source regions. In their analysis
Keeler and Samson calculated the source contribution fields using Quantitative Transport Bias
Analysis (Keeler, 1987) for several heavy metals including arsenic measured in the eastern North
America. The contribution field for As contributions is given in Figure 31. This analysis would
suggest that the largest contributor to As concentrations measured during the month of August, in
the Lake Champlain basin for example, was found in the Midwest area around Pittsburgh but that
significant contributions were also observed from a large region extending up towards the
Canadian smelters in Sudbury and Noranda.
6. BENEFITS FROM EMISSION REDUCTION
A number of the abatement techniques are available for the reduction of emissions of
pollutants studied in this work and emitted from sources in the Great Waters area. Implementation
of these techniques shall result in several benefits which can be measured in a local environment,
e.g. around a given point source of emission, as well as in the whole region of the Great Waters.
Various alternatives to reduce emissions apply to anthropogenic sources. Emissions from
natural sources are very difficult to control, if impossible. In general, different alternatives are
proposed for stationary (point) and mobile sources. It is not intended here to discuss the
applicability of different alternatives to reduce emissions or even their detailed description. It is.
however, important to identify various alternatives in a view of benefits resulting from their
implementation.
Three different groups of methods can be identified for reductions of emissions from
variety of point sources: application of best available technology (so-called BAT, or best
practicable technology - BPT), methods leading to increased energy efficiency, and methods
resulting in waste minimization. The BAT concept is based on the application of the state -of-the
art control technology to remove pollutants from exhaust gases. As a majority of heavy metals
and persistent organic pollutants from industrial sources is released on particles, the BAT in this
case is concerned with the equipment to efficiently remove dust from exhaust gases leaving
primary and secondary iron and steel plants, primary and secondary non-ferrous smelters,
115
-------
Figure 31. The spatially averaged As contribution to ambient concentrations in the
northeastern United States during August, 1983.
116
-------
ferroalloy plants, waste incinerators, and fossil fuel power plants producing electricity and heat.
In general, the installation of fabric filters or equally efficient control techniques (e g. electrostatic
precipitators - ESPs) with practical removal efficiency of better than 98% for fine panicles (with
diameter smaller than 2 pirn) will assure the reduction of dust concentrations in the exhaust gases
lower than 10 mg/Nm3 It should be admitted that both, primary exhaust gas and fugitive
emissions must be controlled. The installation of BAT as described above will result in reduction
of up to 95% of heavy metals and persistent organic pollutants on particles, introduced to the
industrial processes as impurities of raw materials (Pacyna, 1992).
Removal of mercury from exhaust gases is different that the removal of other heavy
metals, as the majority of the mercury emissions occurs in a gaseous phase. Therefore, BAT for
mercury removal includes both wet scrubbers and ESPs, capable-of reducing metal concentrations
in exhaust gases to at least 50 [ig/Nm3 A presence of any flue gas desulfurization technique
results in removal of 40 to 80% of gaseous mercury in exhaust gases.
Removal of dioxins from exhaust gases of waste incinerators could obtained through the
application of various flue gas cleaning techniques. The dry flue gas cleaning with scrubbing on
fabric filters has proved to be very efficient to remove dioxins on an industrial scale in waste
incineration plants.
Reduction of emissions of heavy metals and persistent organic pollutants can also be
obtained through application of so-called pre-treatment methods, and first of all washing of raw
materials before use and switch of fuels. In fact both methods are applied primarily to fuels and
not to other raw materials. It has been recently concluded that washing of coal prior to
combustion results in removal of 10 to 30% of heavy metals contained in coal (reported by
Pacyna, 1992). Fuel switching can be used in some operations which involves replacing some of
the coal and residual oil fired in a boiler with select natural gas. The degree of emission reduction
depends on the amount of fuel to be substituted and the kind of fuel. In the best cases up to 80%
reduction can be obtained.
Appropriate design and management of the combustion processes in new incinerators
would result in considerable reductions in dioxin formation when compared with many existing
incinerators and could bring the emission back to about 1 ng TEQ (Toxic Equivalents Quantity)/
Mm3
Reduction of emissions through the application of techniques leading to increase of energy
efficiency is based on the use of so-called primary reduction measures. An example of such
117
-------
measures is modification of combustion processes leading to higher combustion efficiency
(through suitable manipulation of the stoichiometry/temperature profiles within the boiler). It is
difficult to assess to what extent the emission reduction can be achieved through the combustion
modification. For gaseous pollutants, such as nitrogen oxides a reduction range of 60 to 90% was
found technically possibly (reported by Pacyna, 1992).
Emission reductions can also be obtained through minimization of wastes. Production
technologies resulting in lower amounts of exhaust gases should be given priority over the
technologies with large quantities of waste gases. The cost of emission control installations is
usually lower for the low waste gas technologies to produce a certain industrial product.
The use of unleaded gasoline is the best option to reduce lead emissions from mobile
sources. Extended research has been carried out on the cost and benefits from using the unleaded
gasoline (e.g. CONCAWE, 1980; CEC, 1984). A production of unleaded gasoline requires about
5% increase in total energy (crude oil). A considerable reduction of dioxin input to the
atmosphere from vehicles can be achieved through the general use of catalytic converters and the
use of unleaded gasoline.
Reduction of emissions of pollutants in the Great Waters Study areas may be achieved
through decreases in vehicle miles traveled, fuel reformulation, as well as through the introduction
of alternative vehicles. Tailpipe controls result in lowering emissions from other area sources.
There are also alternatives to reduce emissions from fugitive and indirect sources although
control of releases from these sources is much more difficult that the control of emissions from
the above discussed sources. Landfills should be organized as controlled landfills connected to
waste water treatment plants if possible. Leaching potential of metals, defined as the fraction of
metal present in a solid waste that may become water - soluble under certain chemical conditions
should be controlled at least for volatile metals, such as mercury.
The phasing out the pesticides which are the most persistent, toxic and liable to
bioaccumulate is the best way to deal with the unwanted environmental effects of these chemicals.
The implementation of the emission reduction techniques as described above would
improve the quality of the environment through the reduction of atmospheric deposition of the
pollutants of interest. This is the major environmental benefit. The reduction of atmospheric
deposition will inevitably decrease the uptake of pollutants by surface waters, soils, and plants in
the vicinity of major point sources of emissions, and limit migration of these pollutants through
various environmental media.
118
-------
Installation of equipment to reduce the dust concentration in exhaust gases, or
improvement of its performance will contribute to the improvement of air visibility in the studv
region.
Reduction of atmospheric emissions of heavy metals and persistent organic pollutants will
result in lowering their intake to human body, mostly due to reducing their ingestion in the study
region. This intake has already exceeded the WHO/FAO maximum permissible values for some
pollutants in some locations within the study area.
Implementation of emission control techniques will be a substantial step towards
improving the chemical and to some extent biological recovery of the environment in the study
area. As a result, an increase of fish population, an important factor of local economy in the
region, can be expected. The most obvious and measurable costs are those that stem from
damage to fish and wildlife and the loss of commercial fisheries represents one of the first and
most easily identifiable losses of economic value. As indicated in the documents for virtually
eliminating inputs to the Great Lakes, loss income estimates, in 1990 dollars, due to toxics-related
closures or market losses could be as high as 8.5 million dollars per year including only losses due
to mercury in Lake St. Clair and toxics in Lake Ontario. Since most of the emissions deposited in
the study region originate outside the region, limitations of these emissions will reduce deposition
in a much larger region than the study area.
7. CONCLUSIONS
1. Atmospheric deposition is one of the major sources of lead, arsenic, cadmium, mercury, PAHs.
lindane, and possible PCBs, PCDDs, and PCDFs measured in the Great Waters. The other
sources include leaching from the landfills, direct industrial discharges, agriculture practices in the
region, transport with river waters, and direct dumping of wastes. However, identifying the
specific sources or source types emitting the pollutants into the atmosphere which ultimately are
deposited is another matter. Identification of the dominant pathway and the major sources of the
critical pollutants should be made for the individual compounds separately as their sources and
behavior in the environment differ substantially.
2. Emissions from sources within and outside the Great Waters regions both contribute to the
load of pollution in atmospheric deposition to the waters in the region. Although a number of
source-receptor techniques are available for estimating the contributions, it is still premature to
conclude what pan of pollution load originates within the study region and what part results from
long range transport within air masses. The major reason for the present uncertainty is the lack of
119
-------
reliable input data for application of these techniques including properly reported emission data.
The present lack of monitoring data as well as emissions information is also prob'en.-.:.;, ibr other
regions, e.g. the North Sea and the Baltic Sea, the two most extensively studied regions in Europe
with respect to the environmental behavior of toxic heavy metals and persistent organic
pollutants.
3. Identification of emission sources in the Great Waters regions and their characterization with
respect to atmospheric emissions has been carried out for some time in both the United States and
Canada. As a result, major source categories have been defined for all of the studied pollutants.
The major sources include: production of electricity and heat, combustion of fuels in industrial,
commercial, and residential units, including wood combustion, manufacturing and use of various
industrial goods, and incineration of municipal and industrial wastes, and incineration of sewage
sludge dominate in the group of local sources. There are, however, differences in quantitative
assessment of the fluxes from the above sources, reported by various research groups in the
United States and Canada. These differences should be resolved through thorough examination
of the available data using various techniques of verification of emission data and joint
supplementary research programs in both countries.
4. Emissions from other source regions in North America may also affect the amount of pollution
load deposited to the Great Waters although no evidence has been provided by measurements and
assessment for heavy metals and persistent organic pollutants. The NAPAP concluded that such
an impact exists for deposition of sulfates. As the sulfates are transported within air masses on
particles, as so do most of the metals and organic compounds discussed in this report, one would
also hypothesize that the metals and organic compounds emitted from sources outside the study
region can be deposited to the Great Waters. At present, regional models of long range transport
of air pollutants are available and with some modifications they can de used to assess the
contribution of emissions from outside source regions to the Great waters. An emission inventory
with the appropriate spatial distribution of the data needs to be prepared. The experience gained
through NAPAP emissions inventory development can be used as a starting point for this purpose
Collaboration with the Mexican authorities on environment protection is highly recommended as
the Mexican emissions of heavy metals and persistent organic pollutants are expected to
contribute to the contamination of the Great Waters.
5. Lindane was found to be a global air pollutant measured in remote areas around the world. As
the major application of this pesticide is in Asia and the wind patterns at various altitudes do not
exclude the air mass transport form the Asian continent to North America, lindane deposited in
the Great Waters regions may originate from as far away as India, China, or the former Soviet
120
-------
Union. This hypothesis can be tested by the application of global models, or at least hemispheric
models. At present, such models are used to study the transport of green-house gases in the
atmosphere and the transport of sulfur in the Northern Hemisphere. A knowledge of emission
rates or fluxes of lindane in Asia within 1 degree by 1 degree grid system should be elaborated for
the use in the models. This task can be carried out in co-operation with international programs
involved in preparation of global emission inventories, e.g. the IGBP program on Global Emission
Inventories Activity (GEIA) or the OECD program on emission of greenhouse gases.
6. Several methods can be applied to reduce emissions of toxic heavy metals and persistent
organic pollutants and eventually reduce the atmospheric deposition of these pollutants to the
Great Waters. Technological solutions presented in a form of Best Available Technology (BAT)
package or Best Practicable Technology (BPT) package offer emission reduction possibilities for
point sources within all major source categories contributing to the contamination on the Great
Waters. Experience gained in this respect in North America and Europe, and particularly in the
Baltic Sea Environmental Program can be useful in recommending emission reduction scenarios in
the Great waters region. Cost estimates and benefits from the implementation of the control
techniques should be carefully studied. Experience gained during NAPAP may prove very useful.
8. RECOMMENDATIONS ON FUTURE RESEARCH ACTIVITY TO IMPROVE
THE ACCURACY OF METHODS FOR SOURCE IDENTIFICATION
New research initiatives are necessary in order to meet the requirements outlined in the
conclusions of this report as well as to test the important hypotheses proposed here. These
activities would include both measurement programs and modeling estimates.
8.1. Measurement programs
New measurement programs are needed in order to improve the quality of source-
receptor techniques which are used to assess the magnitude and origin of deposited pollutants.
These measurements are needed at both the receptor, the Great Waters themselves, as well as at
the sources of the emissions. The following source emissions information are recommended.
- emission rates and emission factors for toxic heavy metals and persistent organic pollutants from
large point sources in the study region should be evaluated on the basis of measurements of their
concentrations in' exhaust gases. Large point sources should include electricity and heat
producing plants, ferrous and non-ferrous metal smelters, cement kilns, and waste incinerators.
121
-------
They can be defined as in the CEC directives. The emission measurements shall be representative
and their reporting transparent;
- physical and chemical forms of the most volatile compounds should be established through
measurements carried out in major sources in the study area; and
- emission rates for the most volatile metals, and particularly mercury, as well as lindane and other
pesticides should be derived on the basis of measurements over the water surface in the Great
Waters and the surrounding soils. The results should be representative for the meteorological
conditions as in the Great Waters and exemplify seasonal changes.
Measurements at receptors should provide with the information which is needed in order
to improve the accuracy of source-receptor relationship analysis. The following is recommended:
- size-differentiated chemical composition of aerosols should be measured at receptors which can
represent conditions over the water surface in the study area. The results used in various
statistical methods for source-receptor modeling, e.g. principal component analysis can be used to
improve the accuracy of identification of source categories discussed; and
- simultaneous measurements of the gaseous and particle phases of the studied pollutants with the
help of newly developed techniques (e.g. denuder methods) should be undertaken in order to
provide information on gas-to-particle conversions ( and particle-to-gas conversions) for the most
volatile pollutants under study. The results should be useful to explain the chemical behavior of
these pollutants, particularly during the episodes of their transport within air masses from source
regions to the receptors in the Great Waters areas.
8.2. Modeling estimates
Improvement is needed within the three groups of estimates: emission estimates.
dispersion modeling, and receptor modeling.
The following is recommended for the improvement of emission estimates in order to
assure better understanding of source identification in the Great waters region:
- a set of emission factors and emission rates should be prepared for all sources contributing to
the contamination of the Great waters, and particularly for sources or even whole source
categories for which measurements are not available. In general, rather limited information exists
on emission factors for heavy metals and even less for persistent organic compounds. In the past
122
-------
major research has been carried out in Europe, particularly within the UN ECE programs.
PARCOM, and RELCOM. Recently several projects has been carried out in North America.
mainly for EPA, IJC, Ontario Ministry of the Environment, and Environment Canada. The
information on emission factors, available from the above mentioned programs should be
reviewed and a set of emission factors and emission rates selected with the aim of their application
to prepare emission inventory for atmospheric heavy metals and persistent organic compounds in
North America with special emphasis on the sources within the Great waters region. Guidelines
on emission estimation and reporting should be elaborated in order to assure the data
representativeness, comparability, completeness, consistency, and accuracy. As the subject is of
broad interest, a close co-operation with other programs and international organizations is highly
recommended, particularly with the UN ECE task forces on emission of heavy metals and
persistent organic pollutants;
- gridded emission inventory for the studied pollutants should be approached for the whole
territory of the United States and Canada. A large body of information on the parameters used to
prepare spatial distribution of emission data, e.g. geographical location of point sources and
surrogate parameters to distribute emissions from area sources has been collected during NAPAP
This information should be used to distribute the emission data for pollutants under study here;
- seasonal changes of mercury and volatile organic compound emissions need to be quantified and
techniques developed to estimate these emissions; and
- an approach should be defined to assess emissions of pesticudes (e.g. lindane) in the Northern
Hemisphere with particular emphasis on Mexico and the Asian countries. This task should be
carried out in co-operation with international organizations, such as IGBP
Improvements in source identification through the further development of dispersion
modeling is needed. The following are recommended:
- continue to modify and improve the existing long-range transport models so they can be used to
study the contribution of emissions from sources in North America, both within and outside the
study region to the hazardous pollution load deposited to the Great Waters. Models developed
during NAPAP as well as in other programs in North America and Europe ( e.g. the UN ECE
EMEP model) should be taken into account; and
- an approach should be made to apply the existing global scale models to investigate the
possibility of lindane used in Asia to be transported within air masses to North America and
deposited also in the Great Waters region.
123
-------
9. REFERENCES
Achman, D.A., HombuckJe, K C. and Eisenreich, S.J. (1992). Volatilization of PCI3.S from Green
Bay, Lake Michigan. Environ. Sci. Technol. 27, 75-87.
AJcamo, J., Bartnicki, J., Olendrzynski, K. and Pacyna J.M. (1993) Computing Heavy Metals in
Europe's Atmosphere. Atmos. Environ, (in press).
Alford-Stevens, A.L. (1986). Analyzing PCBs. Basic information about PCBs and how they are
identified and measured. Environ. Sci. Technol. 20, 1194-1199.
Alford-Stevens, A.L., Bellar, T.A., Eichelberger, J.W. and Budde, W L. (1986). Characterization
of commercial Aroclors by automated mass spectrometric determination of polychlorinated
biphenyls by level of chlorination. Anal. Chem. 58, 2014-2022.
Alsberg, T., Hakansson, S., Strandall, M. and Westerholm, R. (1989). Profile analysis of urban air
pollution. Chemometrics and Intelligent Lab. Sys. 7, 143-152.
Amendola, G., Barna, D., Blosser, R., LaFleur, L., McBride, Al, Thomas, F., Tiernan, T. and
Whittemore, R. (1989). The occurrence and fate of PCDDs and PCDFs in five bleached kraft
paper pulp and paper mills. Chemosphere 18, 1181-1188.
Anderson, J.R., Buseck, P.R., Saucy, D.A. and Pacyna, J.M. (1992). Compositions, size
distributions, and principal component analysis of individual fine-fraction panicles from the Arctic
aerosol at Spitsbergen, May - June, 1987. Atmos. Environ. 26A, 1747 - 1762.
Appel, B.R., Cheng, W. and Salaymeh. F (1989). Sampling of carbonaceous particles in the
atmosphere II Atmos. Environ. 23, 2167-2175.
Arey, J., Zielinska, B., Atkinson, R. and Winer, A.M. (1987). PAH and nitroarene concentrations
in ambient air during a wintertime high NOX episode in the Los Angeles Basin. Atmos. Environ.
21, 1437-1444.
Arey, J., Atkinson, R., Zielinska, B. and McElroy, P. A. (1989). Diurnal concentrations of volatile
polycyclic aromatic hydrocarbons and nitroarenes during a photochemical air pollution episode in
Glendora, California. Environ. Sci. Technol. 23, 321-327.
Aronian, P.P., Scheff, P.A. and Wadden, R.A. (1989). Wintertime source reconciliation of
ambient organics. Atmos. Environ. 23, 911-920.
Atkinson, R. (1987). Estimation of OH radical reaction rate constants and atmospheric lifetimes
for polychlorobiphenyls, dibenzo-p-dioxins, and dibenzofurans. Environ. Sci. Technol. 21, 305-
307.
Atkinson, R. (1990). Gas phase tropospheric chemistry of organic compounds: a review. Atmos.
Environ. 24A, 1-41.
124
-------
Axelgard. P and Renberg, L (1989) The influence of bleaching chemicals and lignin content on
the formation of polychlorinated dioxins and dibenzofurans. Chemosphere 19, 661-668
Axenfeld. F, Munch. J., Pacyna, J.M.. Duiser, J.A. and Veldt, C. (1992). Test-
Emissionsdatenbasis der Spurenelemente As, Cd, Hg, Pb, Zn, und der Speziellen Organischen
Verbindungen y-HCH (Lindan), HCB. PCB, und PAX fur Modellrechnungen in Europa.
Forschungsbericht 104 02 588, Dornier GmbH, Friedrichshafen, Germany (in German).
Bacci, E., Calamari, D., Gaggi, C., Fanelli, R., Focardi, S. and Morosini, M. (1986). Chlorinated
hydrocarbons in lichen and moss samples from the Antarctic Peninsula. Chemosphere 15, 747-
754.
Bacci, E., Calamari, D., Gaggi, C., Focardi, S. and Morosini, M. (1988). Organochlorine
pesticide and PCB residues in plant foliage (Magnifera indicd) from West Africa. Chemosphere
17,693-702.
Bacci, E., Calamari, D., Gaggi, C. and Vighi, M. (1990a). Bioconcentration of organic chemical
vapors in plant leaves: experimental measurements and correlation. Environ. Set. Technol. 24,
885-889.
Bacci, E., Cerejeira, M.J., Gaggi, C., Chemello, G., Calamari, D. and Vighi, M. (1990b).
Bioconcentration of organic chemical vapours in plant leaves: the azalea model. Chemosphere
21,525-535.
Bacher, R., Swerev, M. and Ballschmiter. K. (1992). Profile and pattern of monochloro- through
octachlorodibenzodioxins and dibenzofurans in chimney deposits from wood burning. Environ.
Sci. Technol. 26, 1649-1658.
Back, S.O., Field, R.A., Goldstone, M.E., Kirk, P W., Lester, J.N. and Perry, R. (1991a). A
review of atmospheric polycyclic aromatic hydrocarbons: sources, fate, and behavior. Water,
Air, SoilPollut. 60, 279-300.
Back, S.O.. Goldstone. M.E., Kirk, P W.W., Lester, J.N. and Perry, R. (1991b). Phase
distribution and particle size dependency of polycyclic aromatic hydrocarbons in the urban
environment. Chemosphere 22, 503-520.
Baker, D.B and Richards, R.P (1990). Transport of soluble pesticides through drainage
networks in large agricultural basins. In. Long-Range Transport of Pesticides. Kurtz, D.A. (ed.),
pp. 241-270. Lewis Publishers, Chelsea, Michigan.
Baker, J.E. and Eisenreich, S.J. (1990). Concentrations and fluxes of polycyclic aromatic
hydrocarbons and polychlorinated biphenyls across the air- water interface of Lake Superior.
Environ. Sci. Technol. 24. 342-352.
*
Ballschmiter, K. (1991). Global distribution of organic compounds. Environ. Carcino. Ecotox.
Rev. C9, 1-46.
125
-------
Barrie, L.A., Gregor, D.J., Hargrave, BT, Lake. R, Muir, D.C.G., Shearer, R., Tracey, B. and
Bidleman, T.F (1992). Arctic contaminants: sources, occurrence, and pathos Sci. Tot.
Environ. 122, 1-74.
Bastable, H.G., Rogers, D.P and Schorran, D.E. (1990). Tracers of opportunity and pollutant
transport in southern California. Atmos. Environ. 24B, 137-151.
Behymer, T.D. and Hites, R.A. (1988). Photolysis of polycyclic aromatic hydrocarbons adsorbed
on fly ash. Environ. Sci. Techno!. 22, 1311-1319.
Benjey, W.G., and Coventry, D.H. (1992). Geographical distribution and source type analysis of
toxic metal emissions. Presented at the Int. Symp. on Measurement of Toxic and Related Air
Pollutants, Durham, NC, May 3 -8, 1992.
Benner, Jr., B.A., Gordon, G.E. and Wise, S.A. (1989). Mobile sources of atmospheric
polycyclic aromatic hydrocarbons: a roadway tunnel study. Environ. Sci. Technol. 23, 1269-
1278.
Bidleman, T.F. and Olney, C.E. (1974). Chlorinated hydrocarbons in the Sargasso Sea
atmosphere and surface water. Science 183, 516-518.
Bidleman, T.F and Olney, C.E. (1975). Long range transport of toxaphene insecticide in the
atmosphere of the western North Atlantic. Nature 257, 475-477.
Bidleman, T.F., Rice, C.P and Olney, C.E (1976). Fiigh molecular weight hydrocarbons in the
air and sea: rates and mechanisms of air-sea transfer. In: Marine Pollutant Transfer. Windom,
H.L., Duce, R.A. (eds.), pp. 323-351 Lexington Books (D.C. Heath and Co.), Lexington,
Massachusetts.
Bidleman, T.F., Wideqvist, U., Jansson, B. and Soderlund, R. (1987). Organochlorine pesticides
and polychlorinated biphenyls in the atmosphere of southern Sweden. Atmos. Environ. 21.
641-654.
Bidleman, T.F. (1988). Atmospheric processes: wet and dry deposition of organic compounds
are controlled by their vapor-particle partitioning. Environ. Sci. Technol. 22, 361-367, correction
on 726-727.
Bidieman, T.F., Zaranski, M.T. and Walla, M.D. (1988). Toxaphene: usage, aerial transport, and
deposition. In: Toxic Contamination in Large Lakes. Vol. Ill, Schmidtke, N.W. (ed.), pp. 257-
284. Lewis Publishers, Chelsea, Michigan
Bidleman. T.F., Atlas, E.L., Atkinson, R., Bonsang, B., Burns, K., Keene, W.C., Knap, A.H.,
Miller, J., Rudolph, J. and Tanabe, S. (1990). The long-range transport of organic compounds.
In: The Long-Range Atmospheric Transport of Natural and Contaminant Substances. Knap, A.H.
(ed.), pp. 259-301. NATO ASI Series C, Vol. 297, Kluwer Academic Publishers, Boston.
126
-------
Bidleman, T.F (1991). Gas-particle distribution and atmospheric deposition of semivolatile
organic compounds. In: Municipal Waste Incineration Risk Assessment. Travis, C.C. (ed.), pp.
65-86. Plenum Publishing Co., New York.
Bidleman, T.F., Walla, M.D., Roura, R., Carr, E. and Schmidt, S. (1992a). OrganochJorine
pesticides in the atmosphere of the Southern Ocean and Antarctica. Marine Pollut. Bull, (in
press).
Bidleman, T.F., Patton, G.W. and Hinckley, D.A. (1992b). Hexachlorocyclohexanes,
hexachlorobenzene, and polychlorocamphenes in the Arctic atmosphere and surface ocean. In:
Proceedings of the Eighth International Conference of the Comite Arctique International on the
Global Significance of the Transport and Accumulation of Polychlorinated Hydrocarbons in the
Arctic, Shearer, R. (ed.), Plenum Publishing Co. (in press).
Bidleman, T.F., Cotham, W.E., Addison, R.F. and Zinck, M.E. (1992c). Organic contaminants in
the northwest Atlantic atmosphere at Sable Island, Nova Scotia, 1988-89. Chemosphere 24,
1389-1412.
Bjorseth, A., Lunde, G. and Lindskog, A. (1979). Long range transport of polycyclic aromatic
hydrocarbons. Atmos. Environ. 13, 45-53.
Blakley, P. and Klevs, M. (1990). Air Toxics Emission Inventories for the Lake Michigan
Region. Contract 68-02-4398 from the U.S. Environmental Protection Agency, Region V, to
Engineering Science, Inc., Fairfax, Virginia, 192 pp.
Bobet, E., Berand, M.F and Dann, T. (1989). The measurement of PCDD and PCDF in ambient
air in southwestern Ontario. Chemosphere 20, 1439-1445.
Bodo, B.A. (1991). Trend analysis and mass discharge estimates of atrazine in southwestern
Ontario Great Lakes tributaries, 1981 -89. Environ. Toxicol. Chem. 10, 1105-1121.
Broman, D.. Naf. C. and Zebuhr, Y. (199la). Long-term high- and low-volume air sampling of
polychlorinated dibenzo-p-dioxins and dibenzofurans and polycyclic aromatic hydrocarbons along
a transect from urban to remote areas on the Swedish Baltic coast. Environ. Sci. Technol. 25.
1841-1849.
Bumb, R.R., Crumett, W.B., Cutie, S.S., Gledhill, J.R., Hummel, R.H., Kagel, R.O., Lamparski.
L.L., Luoma, E.V., Miller, D.L., Nestrick, T.J., Shadoff, L.A., Stehl, R.H. and Woods, J.S.
(1983). Trace chemistries of fire. Science 210, 385.
Bunce, N.J., Landers, J.P., Langshaw, J-A. and Nakai, J.S. (1989). An assessment of the
importance of direct solar degradation of some simple chlorinated benzenes and biphenyls in the
vapor phase. Environ. Sci. Technol. 23, 213-218.
Bunce, N.J. and Nakai, J.S. (1989). Atmospheric chemistry of chlorinated phenols. J. Air Pollut.
Cont. Assoc. 39, 820-823.
127
-------
Calamari, D., Bacci, E., Focardi, S., Gaggi, c., Morosini, M. and Vighi, M. (1991). Role of plant
biomass in the global environmental partitioning of chlorinated hydrocarbons. Environ. Sa.
Technol. 25, 1489-1495.
Christmann, W., Kloppel, K.D., Partscht, H. and Retard, W. (1989). Determination of
PCDD/PCDF in ambient air. Chemosphere 19, 521-526.
Clark, T.L. (1992). Exposure/risk assessment status report: regional analysis. Presented at the
Joint Workshop on Mercury and Multimedia Risk Assessment, Durham, NC, October 22, 1992.
Clement, R.E., Suter, S.A., Reiner, E., McCurvin, D. and Hollinger, D. (1989). Concentrations
of chlorinated dibenzo~p-dioxins and dibenzofurans in effluents and centrifuged particulates from
Ontario pulp and paper mills. Chemosphere 19, 649-654.
Clendening, L.D., Jury, W.A. and Ernst, F.F. (1990). A field mass balance study of pesticide
volatilization, leaching, and persistence. In: Long-Range Transport of Pesticides. Kurtz, D.A.
(ed.), pp. 47-60. Lewis Publishers, Chelsea, Michigan.
Cohen, J. and Pinkerton, C. (1966). Widespread translocation of pesticides by air transport and
rainout, In: Organic Pesticides in the Environment. Gould, R.F. (ed.), pp. 163-176. Adv. in
Chem. Ser., vol. 60, Amer. Chem. Soc., Washington, D.C.
Cohen, M.A., Ryan, P.B., Spengler, J.D., Ozkaynak, H. and Hayes, C. (199la). Source-receptor
study of volatile organic compounds and paniculate matter in the Kanawha Valley, West Virginia.
I. methods and descriptive statistics. Atmos. Environ. 25B, 79-93.
Cohen, M.A., Ryan, P.B., Spengler, J.D., Ozkaynak, HI, Hayes, C. (1991b). Source-receptor
study of volatile organic compounds and paniculate matter in the Kanawha Valley, West Virginia.
II. Analysis of factors contributing to VOC and particle exposures. Atmos. Environ. 25B, 79-93.
Gotham, W.E. (1990). Chemical and Physical Properties Affecting the Transport and Fate of
Semivolatile Organic Compounds in the Environment. Ph.D. dissertation, Dept. of Chemistry and
Biochemistry, University of South Carolina, Columbia, South Carolina.
Gotham, W.E. and Bidleman, T.F. (1991). Estimating the atmospheric deposition of
organochlorine contaminants to the Arctic. Chemosphere 22, 165-188.
Gotham, W.E. and Bidleman, T.F. (1992). Laboratory investigations of the partitioning of
organochlorine compounds between the gas phase and atmospheric aerosols on glass fiber filters.
Environ. Sci. Technol. 26, 469-478.
Coutant, R.W., Callahan, P.J., Chuang, J.C. and Lewis, R.G. (1992). Efficiency of silicone grease
coated denuders for collection of polycyclic aromatic hydrocarbons. Atmos. Environ. 26A,
2831-2834.
Czuczwa, J.M. and Hites, R.A. (1986). Airborne dioxins and dibenzofurans: sources and fates.
Environ. Sci. Technol. 20, 195-200.
128
-------
Daisey, J.M., Cheney, J.L. and Lioy, P J (1986). Profiles of organic paniculate emissions from
air pollution sources: status and needs for receptor source apportionment modeling. J. Air
Pollut. Com. Assoc. 36. 17-33.
Dann, T. (1990). The Detroit Incinerator Monitoring Program, Data Reports 3 and 4, PMD-89-
8, PMD-90-8, Environment Canada, Conservation and Protection, Ottawa.
deRaat, W.K., Bakker, G.L. and Meijere, F.A. (1990). Comparison of filter materials for
sampling of mutagens and polycyclic aromatic hydrocarbons in ambient airborne particles. Atmos.
Environ. 24A, 2875-2887.
Doskey, P V., Porter, J.A. and Scheff, P.A. (1992). Source fingerprints for volatile non-methane
hydrocarbons. J. Air Waste Mgmt. Assoc. (in press).
Ducret, J. and Cachier, H. (1992). Paniculate carbon content in rain at various temperate and
tropical locations. J. Atmos. Chem. 15, 55-67.
Duinker, J.C. and Bouchertall, R. (1989). On the distribution of atmospheric polychlorinated
biphenyl congeners between the vapor phase, aerosols, and rain. Environ. Sci. Technol. 23, 57-
62.
Dunstan, T.D.J., Mauldin, R.F., Jinxlan, Z., Hipps, A.D., Wehry, E.L. and Mamantov, G. (1989).
Adsorption and photodegradation of pyrene on magnetic, carbonaceous, and mineral subtractions
of coal stack ash. Environ. Sci. Technol. 23, 303-308.
Dzubay, T.G. and Mamane, Y. (1989). Use of electron microscopy data in receptor models for
PM-10. Atmos. Environ. 23, 467-476.
Eadie, B.J. (1984). Distribution of polycyclic aromatic hydrocarbons in the Great Lakes. In:
Toxic Contaminants in the Great Lakes. Nriagu, J.O. and Simmons, M.S. (eds.), pp. 195-211.
Adv. Environ. Sci. Technol. Ser. Vol. 14, J. Wiley and Sons, New York.
Eadie, B.J., Robbins, J.A., Faust, W.R. and Landrum, P.F. (1990). Polycyclic aromatic
hydrocarbons in sediments and pore waters of the lower Great Lakes: reconstruction of a
regional benzo(a)pyrene source function, pp. 171-189, In:
Edgerton, S.J., Holdren, M.W. and Smith, D.L. (1989a). Interurban comparisons of ambient
volatile organic compound concentrations in U.S. cities. J. Air Pollut. Cont. Assoc. 39, 729-732
Edgerton, S.J., Czuczwa. J.M., Rench, J.D., Hodanbosi, R.F and Koval, P.J. (1989b). Ambient
air concentrations of polychlorinated dibenzo-p-dioxins and dibenzofurans in Ohio: sources and
health risk assessment. Chemosphere 18, 1713-1730.
Eduljee, G.H. (198-7). Comment on "Airborne dioxins, sources and fates" Environ. Sci.
Technol. 21, 922-924.
129
-------
Eisenreich, S.J. and Strachan. W.M.J. (1992). Estimating Atmospheric Deposition of Toxic
Substances to the Great Lakes - An Update. Proceedings of a workshop held at the Canada
Centre for Inland Waters, Burlington, Ontario, Jan. 31- Feb. 2, 1992, sponsored by the Great
Lakes Protection Fund and Environment Canada, 59 pp.
Eitzer, B.D. and Kites, R.A. (1989a). Polychlorinated dibenzo-p-dioxins and dibenzofurans in the
ambient air of Bloomington, Indiana. Environ. Sci. Technol. 23, 1389-1395.
Eitzer, B.D. and Kites, R.A. (1989b). Atmospheric transport and deposition of polychlorinated
dibenzo-p-dioxins and dibenzofurans. Environ. Sci. Technol. 23, 1396-1401.
Emmim, J., Gregory, C. and Wind, J.J. (1989). A Prototype Toxics Loading Inventory for Ten
Counties in Southeast Virginia, Prepared for the U.S. Environmental Protection Agency, Office of
Resources Management, by American Management Systems, Inc., Contract GS-OOK-85-AFD-
2777, 47 pp.
Environmental Protection Agency (1987a). National Dioxin Study ~ Report to Congress. EP V
530/SW-87-025, Research Triangle Park, North Carolina.
Environmental Protection Agency (1987c). Chlordane, Heptachlor, Aldrin, and Dieldrin.
Technical Support Document, Office of Pesticide Programs, Washington, D.C.
Environmental Protection Agency (1988a). Compilation of Air Pollutant Emission Factors. Vol.
i. Stationary Point and Area Sources, Supplement B., 4th ed., U.S. Environmental Protection
Agency, Office of Air Quality Planning and Standards. NTIS PB89-128631.
Environmental Protection Agency (1988b). Air Emissions Species Manual, Vol. I. Volatile
Organic Compound Species Profiles. Report No. EPA-450/2-88-033.
Environmental Protection Agency (1988c). Air Emissions Species Manual, Vol. II. Paniculate
Matter Species Profiles. Report No. EPA-450/2-88-033.
Environmental Protection Agency (1990). Toxics in the Community. 1988. National and Local
Perspectives. U.S. Environmental Protection Agency, Washington, DC, Report No. EPA
560/4-90-017.
Environmental Protection Agency (1991). National Air Quality and Emissions Trends Report,
1990. U.S. Environmental Protection Agency, Research Triangle Park, NC, 1991.
Erickson, M.D. (1989). PCDFs and related compounds produced from PCB fires: a review.
Chemosphere 19, 161-165.
Federal Register (1988). Chlordane and heptachlor termiticides. notice of cancellation and
amendment of existing stocks determination. 52, 11798-11801
Food and Agricultural Organization of the United Nations (1986-89). Production Yearbook.
Vols. 39-42, Rome.
130
-------
Fouquet, A., Messier, F., Cogan, D. and Grondin, A. (1990). Investigation for presence of
chlorinated dibenzo-p-dioxins (PCDDs) and dibenzofurans (PCDFs) in bleaching processes from
pulp and paper industries. Chemosphere 20, 1707-1714.
Freeman, D.J. and Cattell, F.C.R. (1990), Woodburning as a source of atmospheric polycyclic
aromatic hydrocarbons. Environ. Sci. Technol. 24, 1581-1585.
Friesen, K.J., Muir, D.C.G. and Webster, G.R.B. (1990). Evidence of sensitized photolysis of
polychlorinated dibenzo-p-dioxins in natural waters under sunlight conditions. Environ. Sci.
Technol. 24, 1739-1744.
Gaggi, C., Bacci, E., Calamari, D. and Fanelli, R. (1985). Chlorinated hydrocarbons in plant
foliage: an indicator of tropospheric contamination level. Chemosphere 14, 1673-1686.
Gianessi, L.P. and Puffer, C. (1990, rev. 1991). Herbicide Use in the United States. Quality of
the Environment Division, Resources for the Future, Washington, D.C., 128 pp.
Gibson, T.L., Korsog, P.E. and Wolff, G.T. (1986). Evidence for the transformation of
polycyclic organic matter in the atmosphere. Atmos. Environ. 20, 1575-1578.
Glotfelty, D.E., Schomburg, C.J., McChesney, M.M., Sagebiel, J.C. and Seiber, J.N. (1990).
Studies of the distribution, drift, and volatilization of diazinon residues resulting from spray
application to a dormant peach orchard. Chemosphere 21, 1303-1314.
Graney, J.R., Halliday, A.N., Keeler, G.J., Robbins, J.A. and Nriagu, J.O (1992) Isotopic tracing
of lead pollution in the Great Lakes. Paper presented at the 1992 Fall AGU Meeting.
Greaves, R.C., Barkley, R.M., Sievers, R.E. and Meglen, R.R. (1987). Covariations in the
concentrations of organic compounds associated with springtime atmospheric aerosols. Atmos.
Environ. 21,2549-2561.
Greenburg, A., Darack, F., Harkov, R., Lioy, P and Daisey, J. (1985). Polycyclic aromatic
hydrocarbons in New Jersey, a comparison of winter and summer concentrations over a two-year
period. Atmos. Environ. 19, 1325-1339.
Gregor, D.J. (1991). Trace organic chemicals in the arctic environment: atmospheric transport
and deposition. In: Pollution of the Arctic Atmosphere. Sturges, W.T. (ed.), pp. 217-254. Elsevier
Applied Science, New York.
Guo, Z. and Kamens, R.M. (1991). An experimental technique for studying heterogeneous
reactions of polycyclic aromatic hydrocarbons on particle surfaces. J. Atmos. Chem. 12, 137-151.
Hansen, A.D.A., Conway, T.J., Steele, L.P., Bodhaine, B.A., Thoning, K.W., Tans, Z. and
Novakov, T. (1989)., Correlations among combustion effluent species at Barrow, Alaska: aerosol
black carbon, carbon dioxide, and methane. J. Atmos. Chem. 9, 283-299.
131
-------
Hanson, D.J. (1991). Dioxin toxicity: new studies prompt debate, regulatory ac.i^,. Chemical
and Engineering News, Aug. 12, 7-14
Hargrave, B.T., Vass, W.P., Erickson. P.E. and Fowler, B.R. (1988). Atmospheric uansport of
organochlorines to the Arctic Ocean. Tellus 40B, 480-493.
Harkov, R. and Greenburg, A. (1985). Benzo(a)pyrene in New Jersey -results from a twenty-
seven site study. J. Air Pollut. Cont. Assoc. 35, 238-243.
Hawthorne, S.B., Miller, D.J., Barkley, R.M. and Krieger, M.S. (1988). Identification of
methoxylated phenols as candidate tracers for atmospheric wood smoke pollution. Environ. Sci.
Technol. 22, 1191-1196.
Hawthorne, S.B., Krieger, M.S., Miller, D.J. and Mathiason, M.B. (1989). Collection and
quantitation of methoxylated phenol tracers for atmospheric pollution from residential wood
stoves. Environ. Sci. Technol. 23, 470-475.
Heffter, S.L. (1980). Air Resources Laboratories atmospheric transport and dispersion model.
NOAA Technical Memorandum ERL ARL-81, National Oceanic and Atmospheric
Administration Air Resources Laboratory, Silver Spring, Maryland, February 1980.
Henry, R.C., Lewis, C.W., Hopke, P.K. and Williamson, H.J. (1984). Review of receptor model
fundamentals. Atmos. Environ. 18, 1507-1515.
Henry, R.C. (1991). Multivariate receptor models. In: Receptor Modeling for Air Quality
Management. Hopke, P.K. (ed.), pp. 117-147. Elsevier, New York.
Hermanson, M.H. and Hites, R.A. (1989). Long-term measurement of atmospheric
polychlorinated biphenyls in the vicinity of Superfund sites. Environ. Sci. Techno!. 23, 1253-
1258.
Hildemann, L.M., Markowski, G.R. and Cass, G.R. (199la). Chemical composition of emissions
from urban sources of fine organic aerosol. Environ. Sci. Technol. 25, 744-759.
Hildemann, L.M., Mazurek, M.A., Cass, G.R. and Simoneit, B.R.T. (1991b). Quantitative
characterization of urban sources of organic aerosol by high resolution gas chromatography.
Environ. Sci. Technol. 25, 1311-1325.
Hinckley, D.A., Bidleman, T.F and Rice, C.P (1991) Atmospheric organochlorine pollutants
and air-sea exchange of hexachlorocyclo-hexanes in the Bering and Chukchi Seas. J. Geophys.
Res. 96, 7201-7213.
Hitch, R.K. and Day, HH. (1992). Unusual persistence of DDT in some western soils. Bull.
Environ. Contain. Toxicol. 48, 259-264.
Hoff, R.M. and Chan, K.W (1987). Measurement of polycyclic aromatic hydrocarbons in the air
along the Niagara River. Environ. Sci. Technol. 21, 556-561.
132
-------
Hoff, R.M., Muir, D.C.G. and Grift, N.P (199la). Annual cycle of polychlorinated biphenyls and
organohalogen pesticides in air in southern Ontario. 1 Air concentration data. Environ. Sci.
Technol. 26, 266-275.
Hoff, R.M., Muir, D.C.G. and Grift, N.P (1991b). Annual cycle of polychJorinated biphenyls and
organohalogen pesticides in air in southern Ontario. 2. Atmospheric transport and sources.
Environ. Sci. Technol. 26, 276-283.
Holsen, T.M., Noll, K.E., Liu, S-P and Lee, W-J (1991). Dry deposition of polychJorinated
biphenyls in urban areas. Environ. Sci. Technol. 25, 1075-1081.
Hopke, P.K. (1991). An introduction to receptor modeling. In: Receptor Modeling for Air
Quality Management. Hopke, P.K. (ed.), pp. 1-10. Elsevier, New York.
HornbuckJe, K.C., Achman, D.A. and Eisenreich, S.J. (1992). Over-water and over-land
concentrations of PCBs in Green Bay, Lake Michigan. Environ. Sci. Technol. 27, 87-98.
Hov, 0., Penkett, S.A., Isaksen, I.S.A. and Semb, A. (1984). Organic gases in the
Norwegian Arctic. Geophys. Res. Lett. 11, 425-428.
Husain, L., Webber, J.S., Canelli, E., Dutkiewicz, V.A. and Halstead, J.A. (1984). Mn/V
ratio as a tracer of aerosol SO42'transport. Atmos. Environ. 18, 1059-1071.
ICF (1992). Atmospheric deposition of toxic chemicals to surface waters: Identification and
summary of the recent literature. ICF Incorporated, Report, Fairfax, VA.
ILZSG (1984). World Directory: Lead and Zinc Mines and Primary Metallurgical Works.
International Lead and Zinc Study Group, London, U.K.
ILZSG (1985). World Directory: Secondary Lead Plants. International Lead and Zinc Study
Group, London, U.K.
Jenkins, J.J., Cooper, R.J. and Curtis, A.S. (1990). Comparison of pendimethalin airborne and
dislodgeable residues following application to turfgrass. In: Long-Range Transport of Pesticides.
Kurtz, D.A. (ed.), pp. 29-46. Lewis Publishers, Chelsea, Michigan.
Johnson, N.D., Scholtz, M.T., Cassaday, V. and Davidson, K. (1990). MOE Toxic Chemical
Emission Inventory for Ontario and Eastern North America. ORTECH International, Draft
Report No.P.89 - 50 - 5429/OG, Mississauga, Canada.
Johnson, N.D., Scholtz, M.T., Cassaday, V., Davidson, K. and Ord, D. (1992). MOE Toxic
Chemical Emission Inventory for Ontario and Eastern North America. Final report to the Air
Resources Branch, Ontario Ministry of the Environment, by Ortech International Report No
P92-T61 -5429/OG; 402 pp.
Jury, W.A., Spencer, W.F. and Farmer, W.J. (1983). Behavior assessment model for trace
organics in soil. I. model description. J. Environ. Qital. 12, 558-564.
133
-------
Jury, W.A., Farmer, W J and Spencer, W.F (1984a). Behavior assessment model for trace
organics in soil. II chemical classification and parameter sensitivity, J. Environ. Qua! 13, 567-
572.
Jury, W.A., Spencer, WF and Farmer, W.J. (1984b). Behavior assessment model for trace
organics in soil. III. application of screening model. J. Environ. Qual. 13, 573-579.
Jury, W.A., Spencer, W.F. and Farmer, W.J. (1984c). Behavior assessment model for trace
organics in soil. IV review of experimental evidence, J. Environ. Qual. 13, 580-586.
Kamens, R., Bell, D., Dietrich, A., Perry, J., Goodman, R., Claxton, L. and Tejada, S. (1985).
Mutagenic transformation of dilute wood smoke systems in the presence of 63 and NO2'
analysis of selected HPLC fractions from wood smoke particle extracts. Environ. Sci. Technol.
19, 63-69.
Kamens, R., Guo, Z., Fulcher, J. and Bell, D. (1988). Influence of humidity, sunlight, and
temperature on the daytime decay of PAHs on atmospheric soot particles. Environ. Sci. Technol.
22. 103-108.
Kamens, R.M., Karam, H., Guo, J., Perry, J.M. and Stockburger, L. (1989). The behavior of
oxygenated polycyclic aromatic hydrocarbons on atmospheric soot particles. Environ. Sci.
Technol. 23,801-806.
Kamens, R.M., Guo, J., Guo, Z. and McDow, S.R. (1990). Atmos. Environ. 24A, 1161-1173.
Keeler, G.J., Stevens, R. and Olmez, I (1992) Sources of toxic air pollutants measured the Lake
Michigan Urban Air Toxics Study (LMUATS). Presented at: The International Symposium on
Measurement of Toxic and Related Air Pollutants, Raleigh, N.C.
Keeler, G.J., Japar, S.M., Brachaczek, W.W., Gorse, R.A., Jr. and Pierson, W.R. (1990)
The sources of aerosol elemental carbon at Allegheny Mountain. Atmos. Environ. 24A, 2795-
2805.
Keeler, G.J., Spengler, J.D., Koutrakis, P., Allen, G.A., Raizenne, M. and Stern, B. (1990)
Transported acidic aerosols measured in southern Ontario. Amos. Environ. 24A, 2935-2950.
Keeler, G.J. and Samson, P.J. (1989) On the spatial representativeness of trace element
ratios. Environ. Sci. Technol. 23, 1358-1364.
Keeler, G.J. (1987) A hybrid approach for source apportionment of atmospheric pollutants in the
Northeastern United States. Ph. D. Thesis, The University of Michigan, Ann Arbor.
Kenski, D.M., Wadden, R.A., Scheff, P.A. and Lonneman, W.A. (1991). Receptor modeling of
VOCs in Chicago, Beaumont, and Detroit. Presented at the 84th Annual Meeting of the Air and
Waste Management Assoc., Vancouver, B.C., June, paper no. 91-82.3.
134
-------
Khalil, M.A.K. and Rasmussen, R.A. (1984). Statistical analysis of trace gases in Arctic haze.
Geophys. Res. Leu. 11, 437-40.
Kjeller, L-O., Kulp, S-E., Bergek, S., Bostrom, M, Bergqvist, P-A., Rappe, C, Jonsson, B.,
deWit, C., Jansson, B. and Olsson, M (1990). Levels and possible sources of PCDDs/PCDFs in
sediment and pike samples from Swedish lakes and rivers, part I. Chemosphere 20, 1489-1496
Kleindienst, T.E., Shepson, P.B., Edney, E.G., Claxton, L.D. and Cupitt, L.T. (1986). Wood
smoke: measurement of the mutagenic activities of its gas- and particulate-phase photooxidation
products. Environ. Sci. Technol. 20, 493-501.
Knap, A.H. and Binkley, K.S. (1991). Chlorinated organic compounds in the troposphere over
the western North Atlantic Ocean measured by aircraft. Atmos. Environ. 25 A, 1507-1516.
Koester, C.J. and Hites, R.A. (1992). Wet and dry deposition pf chlorinated dioxins and fiirans.
Environ. Sci. Technol. 26, 1375-1382.
Komarov, B. (1980). The Destruction of Nature in the Soviet Union. Sharpe Inc., New York,
NY.
Krieger, M.S. and Hites, R.A. (1992). Diffusion denuder for the collection of semivolatile
organic compounds. Environ. Sci. Technol. 26, 1551-1555.
Kurtz, D.A. and Atlas, E.L. (1990). Distribution of hexachlorocyclo-hexanes in the Pacific Ocean
basin, air and water. In: Long-Range Transport of Pesticides. Kurtz, D.A. (ed.), pp. 143-160.
Lewis Publishers, Chelsea, Michigan.
LaFlamme, R.E. and Hites, R.A. (1978). The global distribution of polycyclic aromatic
hydrocarbons in recent sediments. Geochim. Cosmochim. Acta 42, 298-303.
Lamb, B., Guenther, A., Gay, D. and Westberg, H. (1987). A national inventory of biogenic
hydrocarbon emissions. Atmos. Environ. 21, 1695-1705.
Lane, D.A., Johnson, N.D., Barton, S.C., Thomas, G.C., Schroeder, W.H. and Strachan, W.M.J.
(1988). Development and evaluation of a novel gas-and-particle sampler for semivolatile organic
compounds in ambient air. Environ. Sci. Technol. 22, 941-947.
Lane, D.A., Johnson, N.D., Hanley, M-J.J., Schroeder, W.H. and Ord, D.T. (1992a). Gas and
particle phase concentrations of alpha-hexachlorocyclohexane, gamma-hexachlorocyclohexane,
and hexachlorobenzene in Ontario air. Environ. Sci. Technol. 26, 126-133.
Lane, D.A., Schroeder, W.H. and Johnson, N.D. (1992b). On the spatial and temporal variations
in atmospheric concentrations of hexachlorobenzene and hexachlorocyclohexane isomers at
several locations in the province of Ontario, Canada. Atmos. Environ. 26A, 31-42.
Larsson, P and Okla, L. (1989). Atmospheric transport of chlorinated hydrocarbons to Sweden
in 1985 compared to 1973. Atmos. Environ. 23, 1699-1711.
135
-------
Larimer, J.S., Hoffman, E.J., Hoffman, G., Fasching, J.L. and Quinn, J.G. (1990) sources of
petroleum hydrocarbons in urban runoff. Water, Air, Soil Pollut. 52, 1-21.
Lewis, R.G., Martin, B.E., Sgontz, D.L. and Howes, Jr., I.E. (1985). Measurement of fugitive
atmospheric emissions of polychJorinated biphenyls from hazardous waste landfills. Environ. Sci.
Technol. 19,986-991.
Lewis, R.G., Coutant, R.W., Callahan, P.J. and Chuang, J.C. (1988) Phase distribution of
polynuclear aromatic hydrocarbons in ambient air. Presented at : 196th National Meeting of the
American Chemical Society, Los Angeles, CA.
Li, S.-M. and Winchester, J.W. (1990). Haze and other aerosol components in later winter
Arctic Alaska, 1986. J. Geophys. Res. 95, D2, 1797-810.
Ligocki, M.P and Pankow, J.F. (1985). Trace organic compounds in rain III. particle scavenging
of neutral organic compounds. Atmos. Environ. 19, 1619-1626.
Lindstrom, G., Rappe, C. and Sjostrom, M. (1989). Multivariate data analysis applied in studying
the distribution of PCDDs and PCDFs in human milk. Chemosphere 19, 745-750.
Lowenthal, D.H. and Rahn, K.A. (1985). Regional sources of pollution aerosol at Barrow,
Alaska during winter 1979 - 80 as deduced from elemental tracers. Atmos. Environ. 19, 2011-
24.
Mackay, D., Paterson, S. and Schroeder, W.H. (1986). Model describing the rates of transfer
processes or organic chemicals between atmosphere and water. Environ. Sci. Technol. 20, 810-
816.
Mackay, D. (1992) Source Investigation for Lake Superior, Report to the Virtual Elimination
Task Force, International Joint Commision, Windsor, Ontario.
Majewski, M.S., Glotfelty, D.E. and Seiber, J.N. (1989). A comparison of the aerodynamic and
the theoretical shape profile methods for measuring pesticide evaporation from soil. Atmos.
Environ. 23, 929-928.
Majewski, M.S., Glotfelty, D.E., Tha Paw U.K. and Seiber, J.N. (1990). A field comparison of
several methods for measuring pesticide evaporation rates from soil. Environ. Sci. Technol. 24,
1490-1497.
Manchester-Neesvig, J.B. and Andren, A.W. (1989). Seasonal variation in the atmospheric
concentration of polychlorinated biphenyl congeners. Environ. Sci. Technol. 23, 1138-1148.
Masclet. P., Movier, G. and Nikolaou, K. (1986). Relative decay index and sources of polycyclic
aromatic hydrocarbons Atmos. Environ. 20, 439-446.
MBB (1986). Non-ferrous Metal Works of the World. Metal Bulletin Books Ltd., 4th Edition.
Park House, Surrey, U.K.
136
-------
McConnell, L.L. (1992) Air-Water Exchange of Chlorinated Organic Compounds. Ph.D
dissertation, Dept. of Chemistry and Biochemistry, University of South Carolina, Columbia, SC ,
190pp.
McConnell, L.L., Gotham, W.E. and Bidleman, T.F (1992). Gas exchange of
hexachlorocyclohexanes in the Great Lakes. Environ. Sci. Technol. (in press).
McDow, S.R. and Huntzicker, J.J, (1990). Vapor adsorption artifact in the sampling of organic
aerosol: face velocity effects. Atmos. Environ. 24A, 2563-2571.
McVeety, B.D. and Hites, R.A. (1988). Atmospheric deposition of polycyclic aromatic
hydrocarbons to water surfaces: a mass balance approach. Atmos. Environ. 22, 511-536.
Miller, D.F., Schorran, D.E., Hoffer, I.E., Rogers, DP., White, W.H. and Macias, E.S. (1990)
An analysis of regional haze using tracers of opportunity. J. Air Waste Mgmt. Assoc. 40, 757-
761.
Miller, M.S, Friedlander, S.K. and Hidy, G.M (1972) A chemical element balance for Pasadena
aerosol. J. Colloid Interface Sci. 39, 165-178.
Mischke, T., Brunetti, K., Acosta, V., Weaver, D. and Brown, M. (1985). Agricultural Sources of
DDT Residues in California's Environment, Environmental Hazards Program, California Dept. of
Food and Agriculture, Sacramento, California, 42 pp.
Mobley, J.D., (1992). Procedures for verification of emission inventories. U.S.
Environmental Protection Agency, Final Draft Report, Research Triangle Park, NC, 29
September, 1992.
Mueller, P.K., Hidy, G.M, Basket, R.L., Fung, K.K., Henry, R.C., Lavery, T.F., Warren,
K.K. and Watson, J.G. (1983). The Sulfate Regional Experiment: Report of Findings 3, Palo
Alto, CA., Electric Power Research Institute, EA-1901.
Muller, R., Lach, G. and Parlar, H. (1988). Vergleichende untersuchungen uber das vorkommen
von toxaphenruckstanden. Chemosphere 17, 2289-2298.
Muller, M.D., Schlabach, M. and Oehme, M. (1992). Fast and precise determination of alpha-
HCH enantiomers in environmental samples using chiral high resolution gas chromatography.
Environ. Sci. Technol. 26, 566-569.
Murphy, T.J. (1983). In: Physical Behavior of PCBs in the Great Lakes. Mackay, D. and
Paterson, S. (eds.). Ann Arbor Science, Ann Arbor, Michigan.
Murphy, T.J., Formanski, L.J., Brownawell, B. and Meyer, J.A. (1985). Polychlorinated biphenyl
emissions to the atmosphere in the Great Lakes region, municipal landfills and incinerators.
Environ. Sci. Technol. 19, 942-946.
NAPAP (1990). Acidic Deposition: State of Science and Technology, Report 8.
137
-------
Nash, R.G. and Hill, B.D. (1990) Modeling of pesticide volatilization and soil Celine under
controlled conditions. In: Long-Range Transport of Pesticides. Kurtz, D.A (ed ). pp. 17-28.
Lewis Publishers, Chelsea, Michigan.
Neme, C. (1990). Electric utilities emissions of toxic air pollutants. Center for Clean Air
Policy, Washington, D.C.
Ngabe, B. (1992). Organic Contaminants in Air and Runoff Water. Ph.D. dissertation, Dept. of
Chemistry and Biochemistry, University of South Carolina, Columbia, SC, 157 pp.
Nishioka, M.G., Howard, C.C., Contos, D.A., Ball, L.M. and Lewtas, J. (1988). Detection of
hydroxylated nitro aromatic and hydroxylated nitro polycyclic aromatic compounds in an ambient
air paniculate extract using bioassay directed fractionation. Environ. Sci. Technol. 22, 908-915.
Nishioka, M.G. and Lewtas, J. (1992). Quantification of nitro- and hydroxylated nitro-aromatic
PAHs in selected ambient air daytime winter samples. Atmos. Environ. 26A, 2077-2087.
Nriagu, J.O. and Pacyna, J.M. (1988). Quantitative assessment of worldwide contamination of
air, water and soils by trace metals. Nature, 333, 134 - 139.
Pacyna, J.M. (1980). Coal-fired power plants as a source of environmental contamination by trace
metals and radionuclides. Habilitation thesis, Technical University of Wroclaw, Poland.
Pacyna, J.M., Ottar, B., Tomza, U. and Maenhaut. W. (1985). Long-range transport of trace
elements to Ny Alesund, Spitsbergen. Atmos.Environ. 19, 857 - 65.
Pacyna, J.M. and Oehme, M. (1988). Long-range transport of some organic compounds to the
Norwegian Arctic. Atmos. Environ. 22, 243 - 57.
Pacyna, J.M. and Munch, J. (1988). Atmospheric Emissions of Arsenic, Cadmium, Mercury,
and Zinc in Europe in 1982. Norwegian Institute for Air Research, NILU OR 17/88,
Lillestram, Norway.
Pacyna, J.M. and Ottar, B. (1989). Origin of natural constituents in the Arctic aerosol.
Atmos. Environ. 23, 809-815.
Pacyna, J.M., Bartonova, A., Cornille, P. and Maenhaut, W. (1989). Modelling of long-
range transport of trace elements. A case study. Atmos.Environ. 23, 107 - 114.
Pacyna, J.M. (1989). Technological parameters affecting atmospheric emissions of trace elements
from major anthropogenic sources. In: Control and Fate of Atmospheric Trace Metals. J.M.
Pacyna and B. Ottar (eds,). Kluwer Academic Publishers, Dordrecht, the Netherlands.
Pacyna, J.M. and Shaw, G.E. (1991). Arctic Haze and Air Pollution. Invited overview paper
at the Int.Conf. on the Role of the Polar Regions in Global Change, Fairbanks, AL, 11-15
June, 1990.
138
-------
Pacyna, J.M. (1992). The Baltic sea Environmental Programme. The Topical Area Study for
Atmospheric Deposition of Pollutants. Final Technical Report. Norwegian Institute for Air
Research, NILU OR 46/92, Lillestrem, Norway.
Pait, A.S., DeSouza, A.E. and Farrow, D.R.G. (1992). Agricultural Pesticides in Coastal Areas:
A National Summary, Office of Ocean Resources Conservation and Assessment, National
Oceanic and Atmospheric Administration, Rockville, Maryland, 112 pp., (pre-publication copy).
Pankow, J.F. (1987). Review and comparative analysis of the theories on partitioning between
the gas and aerosol paniculate phases in the atmosphere. Atmos. Environ. 21, 2275-2283.
Pankow, J.F. and Bidleman, T.F. (1992). Interdependence of the slopes and intercepts from log-
log correlations of measured gas-particle partitioning and vapor pressure. I. Theory and analysis
of available data. Atmos. Environ. 26 A, 1071-1080.
Parmele, L.H., Lemon, E.R. and Taylor, A.W. (1972). Micrometeorological measurement of
pesticide vapor flux from bare soil and corn under field conditions. Water, Air, Soil Pollut. 1,
433-451.
Paterson, S., Mackay, D., Tarn, D. and Shiu, W-Y. (1990). Uptake of organic chemicals by
plants: a review of processes, correlations, and models. Chemosphere 21, 297-331.
Peters, J.A. (1981). An environmental assessment of POM emissions from residential wood-fired
stoves and fireplaces. In: Proceedings of the Fifth International Symposium on Polynuclear
Aromatic Hydrocarbons, Cooke, M. and Dennis, A. (eds.), pp. 571-582. Battelle Press,
Columbus, Ohio.
Piccot, S.D., Watson, J.J. and Jones, J.W. (1992). A global inventory of volatile organic
compound emissions from anthropogenic sources. J. Geophys. Res. 97, 9897-9912.
Pistikopoulos, P., Masclet, P and Movier, G. (1990). A receptor model adapted to reactive
species: PAH evaluation of source contributions in an open urban site. I. paniculate
compounds. Atmos. Environ. 24A, 1189-1197.
Pitea, D., Todeschini, R., Lasagni, M., Moro, G., Bonati, L. and Chiesa, G. (1989). The
combustion of minicipal solid wastes: PCDD and PCDF in MSW and in emissions, a
chemometric approach. Chemosphere 19, 751-757.
Podoll, R.T., Jaber, H.M. and Mill, T. (1986). Tetrachlorodibenzo-dioxin: rates of volatilization
and photolysis in the environment. Environ. Sci. Technol. 20, 490-492.
Prospero, J.M. and Seba, D.B. (1972). Some additional measurements of pesticides in the lower
atmosphere of the northern equatorial Atlantic Ocean. Atmos. Environ. 6, 363-364.
Pryor, S.C. and Hoffer, T.E. (1992). A case study of pollutant transport from Los Angeles to the
desert southwest. Atmos. Environ. 26A, 243-250.
139
-------
Radian (1992). Software Technical Specification for the Emission Factor Information System.
Draft Report. Radian Corporation, Research Triangle Park, NC.
Rahn, K.A. and Lowenthal, D.H. (1984). Elemental tracers of distant regional pollution
aerosols. Science. 23, 132-139.
Rahn, K.A. and Lowenthal, D.H. (1985). Pollution aerosol in the Northeast: Northeastern-
Midwestern contributions. Science. 228, 275-284.
Ramdahl, T., Alfheim, I. and Bjorseth, A. (1983), In: Mobile Source Emissions Including
Polvcvclic Organic Species. Rodia, D., Cooke, M. and Haroz, R.K. (eds.), pp. 277-298. D.
Reidel Pub. Co., Dordrecht, The Netherlands.
Ramdahl, T., Zielinska, B., Arey, J., Atkinson, R., Winer, A.M. and Pitts, Jr., J.N. (1986).
Ubiquitous occurrence of 2-nitrofluor-anthene and 2-nitropyrene in air. Nature 321, 425-427.
Rapaport, R.A., Urban, N.R., Capel. P.D., Baker, I.E., Looney, B.B., Eisenreich, S.J. and
Gorham, E. (1985). "New" DDT inputs to North America - atmospheric deposition.
Chemosphere 14, 1167-1173.
Rappe, C., Marklund, S., Kjeller, L-0. and Lindskog, A. (1989). Long range transport of PCDDs
and PCDFs on airborne particles. Chemosphere 18, 1283-1290.
Rappe, C. (1992). Sources of PCDDs and PCDFs -- introduction, reactions, levels, patterns,
profiles, and trends. Chemosphere 25, 41-44.
Rice, C.P., Samson, P.J. and Noguchi, G.E. (1986). Atmospheric transport of toxaphene to Lake
Michigan. Environ. Sci. Technol. 20, 1109-1116.
Risebrough, R.W., Huggett, R.J., Griffin, J.J. and Goldberg, E.D. (1968). Pesticides: trans-
Atlantic movement in the northeast trades. Science 159, 1233-1236.
Rosen, H. and Hansen, A.D.A. (1984). Role of combustion-generated carbon particles in the
absorption of solar radiation in the Arctic haze. Geophys. Res. Lett. 11, 461 - 4.
Ross, L.J., Nicosia, S., McChesney, M.M., Hefner, K.L., Gonzalez, D.A. and Seiber, J.N. (1990).
Volatilization, off-site deposition, and dissipation of DCPA in the field. Science 19, 715-722.
Saeger, M., et al. (1989). The 1985 NAPAP Emissions Inventory (Version 2): Development of
the Annual Data and Modelers' Tapes. U.S. Environmental Protection Agency, Research Triangle
Park, NC, Report No. EPA-600/7-89-012a (NTIS PB88-225792).
Schantz, M.M., Benner, B.A., Chesler. S.N., Koster. B.J., Hehn, K.E., Stone, S.F., Kelly, W.R.,
Zeisler, R. and Wise, S.A. (1990). Preparation and analysis of a marine reference sediment
reference material for the determination of trace organic constituents. Fres. Z. Anal. Chem. 338,
501-514.
140
-------
Scheff, P.A. and Wadden, R.A. (1991) Receptor modeling for volatile organic compounds. In:
Receptor Modeling for Air Quality Management. Hopke, P.K. (ed.), pp. 213-253. Elsevier, New
York.
Scheff, P.A., Wadden, R.A., Bates, B.A. and Aronian, P.F (1989). Source fingerprints for
receptor modeling of volatile organics. J. Air Pollut. Com. Assoc. 39, 469-478.
Scholz, M.T. and Voldner, E.G. (1992). Estimation of pesticide emissions to the air resulting
from agricultural application. Presented at the Ninth World Clean Air Congress and Exhibition,
Montreal, Aug. 30 - Sept. 4.
Schonherr, J. and Reiderer, M. (1989). Foliar penetration and accumulation of organic chemicals
in plant cuticles. Rev. Environ. Toxicol. Chem. 108, 1 -70.
Schroeder, W.H. and Lane, D.A. (1988). The fate of toxic airborne pollutants. Environ. Sci.
Technol. 22, 240-246.
Schroeder, W.H., Lindquist, O., Munthe, J. and Xiao, Z. (1992) Volatilization of mercury from
lake surfaces. Sci. Total. Environ. 125,47-66.
Schulz, D.E., Petrick, G. and Duinker, J.C. (1989). Complete characterization of polychlorinated
biphenyl congeners in commercial Aroclor and Clophen mixtures by multidimensional gas
chromatography - electron capture detection. Environ. Sci. Technol. 23, 852-859.
Schwartz, T.R., Stalling, D.L. and Rice, C.L. (1987). Are polychlorinated biphenyl residues
adequately described by Aroclor mixture equivalents? Isomer-specific principal component
analysis of such residues in fish and turtles. Environ. Sci. Technol. 21, 72-76.
Seiber, J.N., Madden, S.C., McChesney, M.M. and Winterlin, W.L. (1979). Toxaphene
dissipation from treated cotton field environments: component residual behavior on leaves, and in
air, soil, and sediments determined by capillary gas chromatography. J. Agric. Food Chem. 27,
284-291.
Semb, A. and Pacyna, J.M. (1988). Toxic Trace Elements and Chlorinated Hydrocarbons:
Sources, Atmospheric Transport and Deposition. Nordic Council of Ministers, NORD 1988:
74, Copenhagen, Denmark.
Sheridan, P.J. (1989). Characterization of size segregated particles collected over Alaska and
the Canadian high Arctic, AGASP-II flights 204-206. Amos.Environ. 23, 237 - 86.
Simoneit, B.R.T. (1989). Organic matter in the troposphere — V application of molecular marker
analysis to biogenic emissions into the troposphere for source reconciliations. J. Atmos. Chem. 8,
251-275.
Slivon, L.E., Gebhart I.E., Hayes, T.L., Alford-Stevens, A.L. and Budde, W.L. (1985).
Automated procedures for mass spectrometric determination of polychlorinated biphenyls as
isomer groups. Anal. Chem. 57, 2464-2469.
141
-------
Smith, R.M.. O'Keefe. P W., Aldous. KM., Valente, H , Connor, S.P and Donnelly, R.J. (1990).
Chlorinated dibenzofurans and dioxins in atmospheric samples from cities in New York. Environ.
Sci. Technol. 24, 1502-1506.
Spencer, W.F. and Cliath, M.M. (1990). Movement of pesticides from soil to atmosphere. In:
Lone-Ranee Transport of Pesticides. Kurtz, D.A. (ed.), pp. 1-16. Lewis Publishers, Chelsea,
Michigan.
Spies, R.B., Andresen, B.D. and Rice, Jr., D.W. (1987). Benzthiazoles in estuarine sediments as
indicators of street runoff. Nature 327, 697-699.
Standley, L.J. and Simoneit, B.R.T. (1987). Characterization of extractable plant wax, resin, and
thermally matured components in smoke particles from prescribed bums. Environ. Sci. Technol.
21, 163-169.
Stevens, R.K., Dzubay, T.G., Shaw, R.W., Jr., McClenny, W.A., Lewis, C.W. and Wilson,
W.E. (1980). Characterization of the aerosol in the Great Smoky Mountains. Environ. Sci.
Technol. 14, 1491-1498.
Stevens, R.K., Zweidinger, R.F., Lewis, C.W. and Dzubay, T.G. (1989). Volatile hydrocarbons as
mobile source tracer species for receptor modeling. In: Man and his Ecosystem, Proceedings of
the Eighth World Clean Air Congress, Brasser, L.J., Mulden, W.C. (eds.), pp. 201-205. The
Hague, The Netherlands, Elsevier Science Publishers, Amsterdam.
Strachan, W.M.J. and Eisenreich, S.J. (1988). Mass Balancing of Toxic Chemicals in the Great
Lakes: The Role of Atmospheric Deposition. Appendix 1 from the Proceedings of the Workshop
on the Estimation of Atmospheric Chemicals to the Great Lakes Basin, Scarborough, Ontario,
Joint Water Quality Board/Science Advisory Board/International Air Quality Advisory Board of
the International Joint Commission, Windsor, Ontario, 113 pp.
Sturges, W.T. and Barrie. L.A. (1989). Stable lead isotopes in Arctic aerosols: Evidence for
the origin of arctic air pollution. Atmos. Environ. 23, 2513 - 19.
Swackhamer, D.L., McVeety, B.D. and Kites, R.A. (1988). Deposition and evaporation of
polychlorobiphenyl congeners to and from Siskewit Lake, Isle Royale, Lake Superior. Environ.
Sci. Technol. 22, 664-672.
Swackhamer, D.L. and Hites, R.A. (1988). Occurrence and bioaccumulation of organochlorine
compounds in fishes from Siskewit Lake, Isle Royale. Environ. Sci. Technol. 22, 543-548.
Swackhamer, D.L. and Eisenreich, S.J. (1991). Processing organic contaminants in lakes. In:
Organic Contaminants in the Environment: Sources. Environmental Pathways, and Effects.
Jones, K.C. (ed.), pp. 33-86. Elsevier Science Publishers.
Lakes. Vol. Ill, Schmidtke. N.W (ed.), pp. 389-427. Lewis Publishers, Chelsea, Michigan.
142
-------
Swanson, S.E., Rappe, C., Malmstron, J. and Kringstad. K.P (1988). Emissions of PCDDs and
PCDFs from the pulp industry. Chemosphere 17, 681 -691.
Sweet, C.W. and Vermette, S.J. (1992). Toxic volatile organic compounds in urban air in Illinois.
Environ. Sci. Technol. 26, 165-173.
Takada, H., Onda, T. and Ngura, N. (1990). Determination of polycyclic aromatic hydrocarbons
in urban street dusts and their source materials by capillary gas chromatography. Environ. Sci.
Technol. 24, 1179-1186.
Takeoka, H., Ramesh, A., Iwata, H., Tanabe, S., Subramanian, A.N., Mohan, D., Magendran, A.
and Tatsukawa, R. (1991). Fate of the insecticide HCH in the tropical coast area of south India.
Marine Pollut. Bull. 22, 290-297.
Tan, Y.L. and Heit, M. (1981). Biogenic and abiogenic polynuclear aromatic hydrocarbons in
sediments from two Adirondack lakes. Geochim. Cosmochim. Ada 45, 2267-2279.
Tan, Y.L., Quanci, J.F., Borys, R.D. and Quanci, M.J. (1992). Polycyclic aromatic hydrocarbons
in smoke particles from wood and burning duff. Atmos. Environ. 26A, 1177-1181.
Tiernan, T.O., Wagel, D.J., Vanness, G.F., Garrett, J.H., Soldi, J.G. and Harden, L.A. (1989).
PCDD/PCDF in the ambient air of a metropolitan area in the U.S. Chemosphere 19, 541-546.
Tsai, W., Cohen, Y., Sakugawa, H. and Kaplan, I.R. (1991). Dynamic partitioning of semivolatile
organics in gas/particle/rain phases during rain scavenging. Environ. Sci. Technol. 25, 2012-
2023.
Tuncel, S.G., Olmez, I., Parrington, J.R., Gordon, G.E. and Stevens, R.K. (1985).
Composition of the fine particle regional sulfate component in the Shenandoah Valley.
Environ. Sci. Technol. 19, 529-537.
Tuominen, J., Salomaa, S., Pyysalo, H., Skytta, E., Tikkanen, L., Nurmela, T., Sorsa, J., Pohjola,
V., Sauri, M. and Himberg, K. (1988). Polynuclear aromatic compounds and genotoxicity in
paniculate and vapor phases of ambient air: effect of traffic, season, and meteorological
conditions. Environ. Sci. Technol. 22, 1228-1234.
Tysklind, M. and Rappe, C. (1991). Photolytic transformation of polychlorinated dioxins and
dibenzofarans in fly ash. Chemosphere 23, 1365-1375.
Uliasz, M. and Pielke, R.A. (1990). Application of the receptor oriented approach in mesoscale
dispersion modeling. In: Proc. of the 18th International Technical Meeting on Air Pollution
Modelling (NATO/CCMS), Vancouver, Canada, May 1990.
Van Vaeck, L. and Van Cauenberghe, K. (1984). Conversion of polycyclic aromatic
hydrocarbons on diesel paniculate matter upon exposure to ppm levels of ozone. Atmos.
Environ. 18, 323-328.
143
-------
Vial, J. Bruckman. L. and Buckert. J. (1992). Development of an emissions inventory of toxic
air contaminants for the Great Lake states. Presented at the Inter. Conf. .ind Course on
Emission Inventory Issues, Durham, NC, October 19 - 22, 1992.
Villeneuve, J-P., Holm, E. and Cattini, C. (1985). Transfer of chlorinated hydrocarbons in the
food chain lichen-to-reindeer-to-man. Chemosphere 14, 1651-1658.
Villeneuve, J-P., Fogelqvist, E. and Cattini, C. (1988). Lichens as bioindicators for atmospheric
pollution by chlorinated hydrocarbons. Chemosphere 17, 399-403.
Voldner, E.G. and Smith L. (1989). Production, Usage, and Emissions of Fourteen Priority Toxic
Chemicals. Appendix 2 to the Proceedings of the Workshop on the Estimation of Atmospheric
Chemicals to the Great Lakes Basin, Scarborough, Ontario, Joint Water Quality Board/Science
Advisory Board/International Air Quality Advisory Board of the International Joint Commission,
Windsor, Ontario, ISBN 1-895085-20-9, 94 pp.
Voldner, E.C. and Schroeder, W.H. (1989). Modelling of atmospheric transport and deposition
of toxaphene into the Great Lakes ecosystem. Atmos. Environ. 23, 1949-1961.
Welch, H.E., Muir, D.C.G., Billeck, B.N., Lockhart, W.L., Brunskill, G.J., Kling, H.J., Olson,
M.P. and Lemoine, R.M. (1991). Brown snow: a long range transport event in the Canadian
Arctic. Environ. Sci. Technol. 25, 280-286.
Westerholm, R.N., Aimen, J., Ki, H., Rannug, J.U., Egeback, C-E., Gragg, K. (1991). Chemical
and biological characterization of paniculate, semivolatile, and gas-phase associated compounds
in diluted heavy-duty diesel exhausts: a comparison of three different semivolatile phase samplers.
Environ. Sci. Technol. 25, 332-338.
White, W.H., Macias, E.S., Miller, D.F., Schorran, D.E., Hoffer, T.E. and Rogers, D.P. (1990).
Regional transport of the urban workweek: methylchloroform cycles in the Nevada - Arizona
desert. Geophys. Res. Lett. 17, 1081-1084.
Wilber, G.G., Smith, L. and Malanchuk, J.L. (1992). Emissions inventory of heavy metals and
hydrophobic orbanics in the Great Lakes basin. In: Fate of Pesticides and Chemicals in the
Environment. Schnoor, J.L. (ed.), pp. 27-50. J. Wiley and Sons, New York.
Willis, G.H., McDowell, L.L., Southwick, L.M. and Smith, S. (1985). Toxaphene, methyl
parathion, and fenvalerate disappearance from cotton foliage in the mid-south. J. Environ. Qual.
14,116-120.
Willis, G.H. and McDowell, L.L. (1987). Pesticide persistence on foliage. Rev Environ.
Toxicol. Chem. 100
Wind, J.J. and Burke, L. (1990). Estimation of Atmospheric Deposition of Toxics to the
Chesapeake Bay Watershed, Prototype Emissions Inventory and Ptorotype
Concentration/Deposition Model., Prepared for the U.S. Environmental Protection Agency.
Chesapeake Bay Liaison Office, by American Management Systems, Inc
144
-------
Wise, S.A., Banner, B.A., Chesler, S.N., Hilpert, L.R., Vogt, C.R. and May, W.E. (1986)
Characterization of the polycyclic aromatic hydrocarbons from two standard reference material air
paniculate samples. Anal. Chem. 58, 3067-3077.
Wise, S.A., Hilpert, L.R., Rebbert, R.E., Sander, L.C., Schantz, M.M., Chesler, S.C. and May,
W.E. (1988). Standard reference materials for the determination of polycyclic aromatic
hydrocarbons. Fres. Z. Anal. Chem. 332, 573-582.
Wittlinger, R. and Ballschmiter, K. (1990). C6-C14 organohalogens (alpha- and gamma-HCH,
HCB, PCB, 4.4'-DDT, 4,4'-DDE, cis- and trans-chlordane, trans-nonachlor, anisols) in the lower
troposphere of the southern Indian Ocean. Fres. Z. Anal. Chem. 336, 193-200.
Woodrow, I.E., McChesney, M.M. and Seiber, J.N. (1990). Modeling the volatilization of
pesticides and their distribution in the atmosphere. In: Long-Range Transport of Pesticides.
Kurtz, D.A. (ed.), pp. 61-81. Lewis Publishers, Chelsea, Michigan.
World Meteorological Organization (1989). Atmospheric Inputs of Trace Species to the World
Oceans, Report of the Group of Experts on Scientific Aspects of Marine Pollution (GESAMP),
Report No. 38, IMO/FAO/UNESCOAVMOAVHO/IAEA/UN/UNEP.
Yokley, R.A., Garrison, A.A., Wehry, E.L. and Mamantov, G. (1986). Photochemical
transformation of pyrene and benzo(a)pyrene vapor-deposited on eight coal stack ashes. Environ.
Sci. Technol. 20, 86-90.
Zhang, X. and McMurray, P.H (1991). Theoretical aspects of evaporative losses of adsorbed or
absorbed species during atmospheric aerosol sampling. Environ. Sci. Technol. 25, 456-459.
Zweidinger, R.B., Stevens, R.K. and Lewis, C.W. (1990). Identification of volatile
hydrocarbons as mobile source tracers for fine-particulate organics. Environ. Sci. Technol.
24, 538-542.
145
------- |