EPA-453/R-94-086
RELATIVE ATMOSPHERIC LOADINGS OF TOXIC
CONTAMINANTS AND NITROGEN TO THE GREAT WATERS
A report prepared for:
Melissa McCullough
Great Waters Program Coordinator
Pollution Assessment Branch, ESD (MD-13)
Office of Air Quality Planning and Standards
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
by:
Joel E. Baker1, ITiomas M. Church2, Steven J. Eisenreich3,
William F. Fitzgerald4, and Joseph R. Scudlark2
'Chesapeake Biological Laboratory, Center for Environmental and Estuarine Studies, The
University of Maryland System, Solomons, MD 20688
2College of Marine Studies, University of Delaware, Lewes, DE 19958
3Gray Freshwater Biological Institute and Department of Civil and Mineral Engineering,
University of Minnesota, Navarre, MN 55392
"Department of Marine Sciences, University of Connecticut, Groton, CT 06340
Revision Date: 15 March 1993
-------
RELATIVE ATMOSPHERIC LOADINGS OF TOXIC
CONTAMINANTS AND NITROGEN TO THE GREAT WATERS
A report prepared for:
Melissa McCullough
Great Waters Program Coordinator
Pollution Assessment Branch, ESD (MD-13)
Office of Air Quality Planning and Standards
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
by:
Joel E. Baker1, Thomas M. Church2, Steven J. Eisenreich3,
William F. Fitzgerald4, and Joseph R. Scudlark2
Chesapeake Biological Laboratory, Center for Environmental and Estuarine Studies, The
University of Maryland System, Solomons, MD 20688
2College of Marine Studies, University of Delaware, Lewes, DE 19958
3Gray Freshwater Biological Institute and Department of Civil and Mineral Engineering,
University of Minnesota, Navarre, MN 55392
"Department of Marine Sciences, University of Connecticut, Groton, CT 06340
Revision Date: 15 March 1993
-------
RELATIVE ATMOSPHERIC LOADINGS OF TOXIC
CONTAMINANTS AND NITROGEN TO THE GREAT WATERS
A report prepared for:
Melissa McCullough
Great Waters Program Coordinator
Pollution Assessment Branch, ESD (MD-13)
Office of Air Quality Planning and Standards
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
by:
Joel E. Baker1, Thomas M. Church2, Steven J. Eisenreich3,
William F. Fitzgerald4, and Joseph R. Scudlark2
'Chesapeake Biological Laboratory, Center for Environmental and Estuarine Studies, The
University of Maryland System, Solomons, MD 20688
2College of Marine Studies, University of Delaware, Lewes, DE 19958
3Gray Freshwater Biological Institute and Department of Civil and Mineral Engineering.
University of Minnesota, Navarre, MN 55392
"Department of Marine Sciences, University of Connecticut, Groton, CT 06340
Revision Date: 15 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.
-------
ACKNOWLEDGEMENTS
The authors wish to acknowledge the intellectual and logistical contributions made by
Dianne Leister of University of Maryland. Ms. Brenda Yates, also from the University of
Maryland, typed and edited sections of this report. Dr. Cliff Davidson (Carnegie Mellon
University) constructively reviewed of a draft of this document and shared his insightful reviews
of wet and dry depositional processes. We thank Dr. Anders Andren (University of Wisconsin)
for leading a stimulating review of this report.
We thank Melissa McCullough, Amy Vasu, and Joanne Foy (U.S. EPA/OAQPS) for
providing us the opportunity to collaborate in the preparation of this document.
-------
TABLE OF CONTENTS
A. EXECUTIVE SUMMARY
A.I OVERVIEW
A.2 CASE STUDY MASS BALANCES ......................... «i
A.2. 1 Trace Elements ................................ yi
A.2.2 Semivolatile Organic Contaminants .................... ix
A.2.3 Mercury .................................... xiii
A.2.4 Nitrogen .................................... xix
A. 3 CURRENT UNCERTAINTIES IN GREAT WATERS MASS
BALANCES ..................................... xxii
1.0 INTRODUCTION AND SCOPE OF THIS REPORT .................. 1
2.0 RELATIVE LOADINGS: THE MASS BALANCE PARADIGM .......... 2
3.0 EVIDENCE OF ATMOSPHERIC DEPOSITION TO THE GREAT WATERS . . 4
3.1 TRACE ELEMENTS - Evidence of Deposition ................. 4
3.2 SEMIVOLATILE ORGANIC CONTAMINANTS - Evidence of
Deposition ........................................ 12
3.3 MERCURY - Evidence of Deposition ....................... 14
3.3.1 The Global Mercury Cycle ......................... 14
3.3.2 Regional Mercury Cycling and Localized Deposition
in North America ............................... 21
3.3.3 Localized Atmospheric Hg Deposition: Sweden ............ 27
3.4 NITROGEN - Evidence of Deposition ....................... 32
4.0 CURRENT UNDERSTANDING OF THE SPECIATION OF CHEMICALS IN
THE ATMOSPHERE AND IN PRECIPITATION ................... 36
4. 1 TRACE ELEMENTS - Speciation ......................... 36
4.2 SEMIVOLATILE ORGANIC CONTAMINANTS - Speciation ....... 37
4.3 MERCURY ....................................... 43
4.3.1 Mercury Speciation in Precipitation ..................... 44
4.4 NITROGEN - Speciation ............................... 50
-------
5.0 CURRENT UNDERSTANDING OF WET DEPOSITIONAL PROCESSES .... 51
5.1 GAS SCAVENGING BY PRECIPITATION 51
5.2 AEROSOL SCAVENGING BY PRECIPITATION 52
5.3 FIELD VERIFICATION OF SCAVENGING MECHANISMS 53
5.4 EVIDENCE FROM FOG WATER STUDIES OF ALTERNATE WET
SCAVENGING MECHANISMS 53
6.0 CURRENT UNDERSTANDING OF DRY DEPOSITIONAL PROCESSES .... 55
6.1. DRY AEROSOL DEPOSITION 55
6.1.1 Concepts and Models 55
6.1.2 Field Measurements 57
6.2 GAS ABSORPTION AND VOLATILIZATION 59
6.2.1 Importance in the cycling of organic compounds 59
6.2.2 Concepts and Models 59
6.2.3 Air and water concentrations 65
6.2.4 Mass Transfer Coefficients 66
6.2.5 Field Measurements 68
7.0 EVALUATION OF CURRENT SAMPLING AND ANALYTICAL PROCEDURES75
7.1 TRACE ELEMENTS - Evaluation of current methodologies 75
7.1.1 Atmospheric Sampling - trace elements 75
7.1.2 Precipitation Sampling - trace elements 75
7.2 SEMIVOLATILE ORGANIC CONTAMINANTS - Evaluation of current
methodologies 77
7.2.1 Atmospheric Sampling - SOCs 77
7.2.2 Precipitation Sampling - SOCs 77
7.3 MERCURY 79
7.3.1 Atmospheric Hg 79
7.3.2 Mercury: Precipitation Sampling and Analysis 81
-------
LIST OF FIGURES
1 Mass Balance Paradigm in Lakes .............................. *v
2 Mass Balance Paradigm in Estuaries ............................ v
3 Cobalt Mass Balance for the Delaware Bay (Church, 1986) .............. vii
4 Cadmium Mass Balance for the Delaware Bay (Church, 1986) ........... viii
5 Mass Balance of PCBs in Lake Superior .......................... x
6 Atmospheric Depositional Fluxes to the Great Lakes and to the Chesapeake
Bay (Strachan and Eisenreich, 1988; Baker and Eisenreich, 1993; Church
et al., 1993) ........................................... xii
7 Mass Balance Model for Hg in the Treatment Basin of Little Rock Lake,
Wisconsin (Adapted From Fitzgerald et al., 1991 and Based on Work and
Preliminary Budgets Appearing in Wiener et al., 1990, Fitzgerald and
Watras, 1989, and to Appear in Watras et al., 1992) ................. xiv
8 Mass Balance Model for Monomethyl Hg in the Treatment Basin of Little
Rock Lake, Wisconsin (Adapted and Based on Work From Fitzgerald et al.,
1991; Wiener et al., 1990; Hurley et al., 1991; Bloom et al., 1991, and
Watras, et al., 1992) ..................................... xvi
9 Flows of Hg in g yr"1 for a Typical Southern Swedish Lake With an Area
of 1 km2 and a Drainage Area of 10 km2. Atmospheric Depositional Fluxes
of Hg are for Precipitation and do not Include Dry Deposition. Adapted
From Johansson, et al., 1991, and Lindqvist et al., 1984 ............... xvii
10 An Annual Nitrogen Loading Budget for the Delaware Bay .............. xx
11 Inter-Annual Trends in Precipitation Trace Element Concentrations, Lewes,
DE ................................................. 11
12 Trace Element Fluxes at Various Locations in North America (mg/m2-year) .... 15
13 Latitudinal Distribution of Total Gaseous Hg (TGM;ng nr3) over the Pacific
Ocean Between 1980 and 1986. Adapted From Fitzgerald (1989) .......... 17
14 The Major Species, Fluxes, and Reservoirs for the Physical and Biogeo-
chemical Cycling of Hg in the Atmosphere and Within Lakes (Adapted From
Hudson et al., 1992) ....... . .............................. 18
-------
7.4 NITROGEN 84
7.4.1 Atmospheric Sampling 84
7.4.2 Precipitation Sampling 85
8.0 CASE STUDIES 85
8.1 MASS BALANCE OF TRACE ELEMENTS IN ESTUARIES 85
8.2 MASS BALANCE OF SOCS: A PCB BUDGET FOR LAKE
SUPERIOR, 1986 86
8.3 MERCURY MASS BALANCES
- 8.3.1 Wisconsin Seepage Lakes 88
8.3.2 Drainage Lakes in Sweden 92
8.3.3 Atmospheric Mercury Speciation:
Biogeochemical Implications 92
8.3.4 Summary of Mercury Mass Balances 100
8.4 NITROGEN MASS BALANCES IN COASTAL WATERS 101
8.4.1 Total Nitrogen 101
8.4.2 Dissolved Organic Nitrogen (DON) 107
8.4.3 Estimated Response of Nitrogen Loadings to
1990 Clean Air Act Amendments 108
9.0 CURRENT UNCERTAINTY IN GREAT WATERS MASS BALANCES .... 110
REFERENCES 112
-------
LIST OF FIGURES (CONT'D)
15 Whole Basin Accumulation Rates for H, (ug nr2 yr •') are Plotted Against
the Terrestrial Catchment Area to Lake Area Ratio. Modern Rates Based on
the Past 10 Years are Indicated by the Filled Squares, While the Pre-
industrial Estimates (Before ca. 1850) are Indicated by the Filled Circles
(Adapted from Swain et al., 1992) 22
16 Estimates of the Net Increase in H, due to the Atmospheric Deposition
Compared to Predicated Geological Contributions ("Background") Along a
Track Between Northwestern Minnesota and Eastern Michigan (Adapted From
Nater and Grigal, 1992). The Stations Appear in Figure 17 25
17 Sampling Sites in the Five Zones Established Across the Great Lake State
(Adapted From Nater and Grigal, 1992) 26
18 Stations in the Nordic Network Study of Atmospheric Hg During 1985 to 1989
(From Iverfeldt et al., 1991) 30
19 Mercury in Precipitation Along the Nordic Sampling Network (Sites in Figure
18) 31
20 Total, Reactive and Methylmercury in Rain Collected in Wisconsin at Little
Rock Lake Reference Basin, in 1989 and Max Lake in 1990, Adapted From
Fitzgerald et al., 1992 and Mason et al., 1991 48
21 PAH Speciation in Chesapeake Bay Rainfall (Leister and Baker, 1992) 54
22 Air-Water Exchange 60
23 Stagnant Two Film Model 62
24 Water Phase Transfer Velocity Versus Wind Speed (Modified From Liss and
Merlivat, 1986) 67
25 Air-Water Exchange Fluxes of PCBs and PAHs, Green Bay 74
26 Operationally Defined Species of Hj Based on the Wet Digestion and
Reduction/Sparging Procedures (Adapted From Lindqvist et al., 1991) 83
27a Modelling the Potential Pathways for the Production and evasion of H,0 in
Epilimnion of the Treatment Basin of Little Rock Lake, Wisconsin. The
Amqunts of Hj° Produced by Demethylation and Direct Reduction of Hg(II)
are Estimated and Related to the Input of HgR by Atmospheric Deposition
for the August 1989 and August 1990 experiments 97
-------
LIST OF FIGURES (CONT'D)
27b Modelling the Potential Pathways for the Production and evasion of Hj° in
Epilimnion of the Pallette Lake, Wisconsin. The Amounts of Hg° Produced
by Demethylation and Direct Reduction of H,(II) are Estimated and Related
to the Input of HgR by Atmospheric Deposition for the August 1989 and
August 1990 experiments 98
28 Seasonal Differences in the Relative Atmospheric Loading of Inorganic
Nitrogen to the Delaware Bay 103
29 Seasonal Variability in Nitrogen Deposition at Lewes, DE 105
30 Episodic Atmospheric Wet Deposition of Nitrate and Ammonium at Lewes, DE,
1990 106
31 Projected Nitrogen Emissions and Nitrate Deposition Rates to the Chesapeake
Bay Under Three Control Scenarios 109
-------
LIST OF TABLES
1 A PCB Budget for Lake Superior - 1986 . . xi
2 Current Uncertainties in Estimating the Role of Atmospheric Deposition
in Trace Element Mass Balances to the Great Waters xxiv
3 Current Uncertainties in Estimating the Role of Atmospheric Deposition
in Semivolatile Organic Contaminant Mass Zalances to the Great Waters
(1993) xxv
4 Current Uncertainties in Estimating the Role of Atmospheric Deposition
in Mercury Mass Balances to the Great Waters (1993) xxvi
5 Current Uncertainties in Estimating the Role of Atmospheric Deposition
in Nitrogen Mass Balances to the Great Waters (1993) xxvii
6 Emissions of Trace Metals from Natural Sources to the Atmosphere
(xlO6 kg/yr, From Nriagu 1989) 5
7 Worldwide Emissions of Trace Metals from Industrial Sources. Units:
xlO6 kg/yr (From Nriagu and Pacyna, 1988) 6
8 Atmospheric Versus Riverine Inputs of Trace Metals in the Ocean (From
Nriagu 1992) 8
9 Atmospheric Inputs of Trace Metals to Mid-Atlantic Coastal Marine Systems
(From Church et ai, 1988) 10
10 Input-Output Calculations for PCBs and Benzo[a]pyrene to the Great Lakes
(Strachan and Eisenreich, 1988) 13
11 Global Atmospheric Mercury Budget 20
12 Time Integrated Estimates of Mercury Deposition, as Determined in Peat
From an Ombrotrophic Portion of Arlberg Bog, Minnesota (Benoit et al,
1992a; 1992b) 24
13 Annual Deposition and Volume-Weighted Concentration Averages for Mercury
in Precipitation at Three Locations in Minnesota During 1988
and 1989 28
14 Atmospheric Wet Depositional Fluxes of Total Mercury to Various Stations
in the Nordic Countries 29
-------
LIST OF TABLES (CONT'D)
15 Calculated Nitrogen Loadings to Chesapeake Bay Watershed, 1984* ...... 33
16 Atmospheric Input of Nitrogen to Coastal Waters ................. 34
17 Size Distribution of Measurements of Organic Compounds ........... 42
18 Influence of Paniculate Mercury on the Composition of IL. in
Precipitation (After Iverfeldt, 1991) .......................... 45
19 Summary of the Average Concentration of Hj Species Observed in Wet
Deposition From Northcentral Wisconsin ...................... 46
20 Calculated Concentration of Hj in Rain Based on a Scavenging Ratio,
W = 600 (Range 200-1200) and Using the Formula W = C^ (pg/kg)
X 1.2 kg/m3 C^ (pg/m3). The Average Values were Calculated using
an Average Scavenging Ratio (W) of 600 While the Variability was
Estimated using a Range for W of 200 or 1200 and the Actual Paniculate
Concentration Extremes Found at These Sites. The Values for W Were
Taken From the Data for Lead Reported by Maring et al. (1989). Table
Adapted From Mason et aL 1992 ........................... 47
21 Henry's Law Constants of Semi volatile Organic Contaminants ......... 64
22 Some Empirical Relationships Between k,, and kg and Windspeed ....... 69
23 Mass Balance of PAHs and PCBs in Siskiwit Lake, Isle Royale, Lake
Superior .......................................... 71
24 Estimated Air- Water Fluxes of PCBs (+ Flux = Volatilization) ........ 73
25 Annual Hb Depositional Fluxes in Northcentral Wisconsin Between
October 1988 and October 1990 (From Fitzgerald et al. 1992) ......... 90
26 Average Annual Hg Deposition to Little Rock Lake, Wisconsin During
October 1988 to 1990 (*Dry Deposition Not Included) .............. 91
27 Degree of Saturation for Elemental Mercury (Hg°) in Northcentral
Wisconsin Lakes ..................................... 94
28 Estimated Average Evasional Fluxes for August 1989 and August 1990
for ious Northcentral Wisconsin Lakes. Fluxes are in pmol m'2 day"1,
Calculated Using a Transfer Velocity of 1.5 cm hr1 (0.36 m day1) ...... 96
-------
A. EXECUTIVE SUMMARY
A.I OVERVIEW
The 1990 Amendments to the Clean Air Act (CAA) recognize a geochemical paradigm
which has been evolving during the past 30 years. Pollutants emitted into the atmosphere are
transported for various distances and may later deposit in aquatic systems far removed from their
original sources. This atmospheric deposition may seriously degrade water quality. Linkages
between the atmosphere and surface waters operate on several scales, ranging from trace metal
contamination of ponds adjacent to smelting operations to regional episodes of acidic deposition
to global dispersion of organochlorine insecticides and greenhouse gases. As increasingly
stringent controls are applied to conventional (i.e., point) sources of pollutants to surface waters,
the relative importance of diffuse, non-point sources of contamination, including loadings from
the atmosphere is increasing. This fact requires a fundamental revaluation of our approach to
controlling air and water quality, as it has become increasingly evident that one cannot achieve
water quality objectives if corresponding air quality goals are neglected.
Section 112(m) of the 1990 CAA Amendments requires the U.S. Environmental
Protection Agency and the National Oceanic and Atmospheric Administration to estimate the
importance of atmospheric deposition of hazardous air pollutants to the Great Lakes, Lake
Champlain, the Chesapeake Bay, and other coastal waters (collectively dubbed the Great
Waters). This Section requires not only that gross atmospheric contaminant loadings to each
water body be documented, but, more importantly, that the relative importance of those loadings
compared to those from all other possible sources be quantified. Further, the agencies are
required to determine whether atmospherically-derived contamination results in exceedences of
water quality standards, and to estimate the fraction of contaminants accumulating in biota which
are derived from the atmosphere. Simply stated, Section 112(m) requires the agencies to
construct quantitative chemical mass balances for relevant contaminants in each of the Great
Waters. This is clearly a tall order.
In this paper, we first summarize the current understanding of atmospheric depositional
processes for trace elements, mercury, nitrogen, and synthetic and combustion-derived organic
contaminants. After addressing the question of whether gross contaminant loadings from the
atmosphere can be estimated, we address whether both the conceptual understanding and the
necessary data to construct defensible mass balances of these chemicals in the Great Waters are
available. Specifically, we address the following questions:
1. What is the consensus view of our current understanding of the fundamental processes
resulting in atmospheric deposition?
During the past three decades, the scientific community has recognized the importance
of atmospheric deposition as a source of contaminants to surface waters and has
developed and refined relatively simplistic models describing the fundamental physical
and-chemical processes responsible for atmospheric deposition. At the present time, our
-------
knowledge of processes in the atmosphere and in surface waters are insufficient to
adequately predict the magnitude and impact of atmospheric deposition to the Great
Waters
2. Do we have the tools to determine atmospheric deposition rates with the accuracy and
precision required to make the regulatory decisions required by the Great Waters
Program?
Methodologies of suitable accuracy and precision exist to determine the rate of
atmospheric deposition of many chemical species to specific locations in the Great
Waters. Specifically, deposition of trace elements and some organic contaminants during
rainfall can be adequately measured. Conversely, deposition of aerosol-bound
contaminants and exchange of gaseous contaminants across the air-water interface can
only be estimated by indirect methods, and is, at present, poorly known.
3. Do we have the conceptual understanding required to estimate the relative atmospheric
loadings of contaminants to the Great Waters?
The mass balance paradigm, in which aJl contaminant loadings to and sinks from a water
body are identified and quantified, provides an appropriate conceptual framework in
which to estimate the importance of atmospheric deposition as a source of contaminants
relative to all other sources to the Great Waters. While some of these sources and sinks
are difficult to quantify and require new research initiatives, development of the mass
balance framework is straightforward.
4. Do we currently have data of sufficient accuracy and precision to estimate relative
atmospheric loadings of contaminants to the Great Waters?
In order to estimate the importance of atmospheric deposition relative to all other sources
of contaminants to the Great Waters, it is necessary to construct mass balances for each
contaminant. With very few exceptions, this has not been possible due to a lack of
consistent, coherent, and coincident measurements of all loadings to a water body.
While technically possible, such measurements require a significant, long-term
commitment in order to generate adequate information to construct scientifically-credible
mass balances.
5. What specific studies are required to improve our ability to address the relative loadings
questions of the Great Waters Program?
Our current understanding of many processes is inadequate to make the evaluations
required within Section 112(m). Specific studies of aerosol behavior and deposition, of
absorption of gases and volatilization of dissolved contaminants, and of the reactivity and
bioavailability of deposited contaminants are needed. Coordinated studies in the
atmosphere and within the Great Waters are required in order to fully understand the
• •
u
-------
linkages between these two media. Improvements and validations of predictive,
integrated atmospheric emission, transport, and deposition models are required in order
to provide the predictive ability required by the Agency.
To address these questions, we first review the current understanding of atmospheric deposition
processes, and then apply this understanding by constructing mass balances for the trace
elements cobalt and cadmium and the nutrient nitrogen in the Delaware Bay, for mercury in a
Wisconsin seepage lake, and for polychlorinated biphenyls (PCBs) in Lake Superior.
A.2 CASE STUDY MASS BALANCES
The framework for interpreting the relative inputs of chemicals into a body of water is
a mass balance or input-output budget. The boundaries are often defined as the water column
of the water body, and mass exchanges across the air-water, sediment-water, and land-water
interfaces are inputs and outputs. Considering the case of a fresh water body as a chemical
reactor (Figure 1), contaminants may enter by riverine flow (dissolved and paniculate matter),
groundwater flow (dissolved matter), atmospheric deposition (in the form of gas and particle
scavenging by rain and snow, dry particle deposition, and gas absorption at the air-water
interface), sediment and benthic layer exchange (dissolved and particulate), and in situ
production. Chemical outputs from the lake volume include volatilization, riverine and
connecting channel outflow (dissolved and particulate), chemical and biological degradation,
sedimentation and burial of particles, and groundwater output (dissolved). Chemical species
entering the water column may undergo turbulent and diffusive mixing and reactions which result
in a change of speciation (e.g., dissolved to particulate; oxidation or reduction; hydrolysis and
complexation).
The components of a mass balance model of estuarine ecosystems such as the Chesapeake
Bay and the Delaware Bay differ from those of lakes because of the importance of tidal exchange
of water, solutes, and particles, and because of the frequently higher productivity of coastal
marshes and waters (Figure 2). Over the period of several days to weeks, tidal flow flushes
material into and out of the estuary and exerts a significant control on estuarine water quality.
Therefore, the water residence times in estuaries are often significantly less than those in large
lakes. However, estuaries are also efficient traps of particulate matter, leading to longer
residence times of particle-reactive chemicals. For example, the hydraulic residence time of the
Chesapeake Bay is less than one year, while the water column residence time of particles is
significantly longer. For estuaries, the time rate of change in chemical concentration in the
water column is similar to that of a fresh water system, with the inclusion of a tidal component.
111
-------
atmospheric transport
natural and
anthropogenic
•
Freshwater Contaminant
and
Nutrient Flows
river fluvial transport
ground water exchange
dry deposition
air/water exchange
chemical
and biological
reactions
sediment
exchange
benthic
exchange
I
outflow
Figure 1. Mass balance paradigm in lakes
-------
atmospheric transport
natural and
anthropogenic
Estuarine Contaminant
and
Nutrient Flows
dry deposition
air/water exchange
\
river fluvial transport
ground water exchange
R
chemical tida| exchange /
and biological
salt marsh
exchange
sediment
exchange'
Figure 2. Mass balance paradigm in estuaries
-------
A.2.1 Trace Elements
Mass balances of several trace elements in the Delaware estuary were estimated by
Church et al. (1986). The primary fluvial fluxes were calculated as the seasonally averaged
concentrations at the zero salinity end member times the riverine discharge. In this sense, any
ground water fluxes downstream of this point were ignored. Some unknown quantity of the
fluvial flux is comprised of atmospheric fallout onto the watershed. The secondary source of
trace elements from the surrounding salt marshes were estimated by multiplying the non-
conservative maximum concentration in a representative tidal creek times the net exchange
volume of the tidal wedge integrated over the tidal cycle. The primary components of this flux
come from atmospheric fallout into the marsh watershed and from the diagenetic release of
sedimentary components. Trace elements can enter the salt marsh with both upstream terrestrial
and downstream marine paniculate matter. The oceanic tidal inputs to the bay are difficult to
quantify, but have been attempted using a two layer model and salt balance (Church et al. 1986).
Using such an approach, one can close the balance between the trace element sources (rivers and
salt marshes) and the sinks (sediment burial and oceanic export) from the Delaware estuary
within a factor two. It is thought that the major unknown source in this balance may be those
trace elements which enter from ungauged groundwater. These groundwater fluxes may be as
great as those in the gauged fluvial sources.
Two examples of trace element budgets for Delaware Bay are shown in Figures 3 and
4. These budgets are notable in that they are among the few that are rather complete and
realistic balances as they consider both fluvial and atmospheric fluxes, a multitude of exchanges
(including tidal, intertidal, benthic), and output as well as input terms to the estuarine loading.
As the benthic fluxes are determined from incubated box cores taken in softer sediments, the
magnitude of the flux may not be representative of the whole bay. The direction of the benthic
flux measured by this method is a likely indication of the bay-wide benthic flux, however. The
first example is cobalt, an element which although toxic has a behavior much like those of the
more abundant non-toxic crustal trace elements, such as iron and aluminum. Fluvial transport
is half the total input and is dominated by paniculate loading (Figure 3). Surprisingly, the salt
marsh input of cobalt is equivalent to the river loading, and is dominated by dissolved cobalt.
This is a consequence of the diagenetic remobilization of redox trace elements by acidic sulfide
reoxidation, which dissolves oxides in surface sediments. As these particles are originally
carried into the salt marsh by tidal waters from the bay, this input may be considered secondary
to the fluvial term. The atmospheric input of cobalt to the Delaware Bay is comparatively small,
and from limited studies probably occurs in equal amounts of wet and dry deposition. As a
particle-reactive element, the majority (87%) of the cobalt loading in Delaware is buried in the
bottom sediments of the estuary and not exported to coastal waters. Benthic flux measurements
suggest that cobalt diffuses into the sediments, which indicate the diagenetic capacity of the
sediments to assimilate dissolved cobalt is at least equivalent to the rate of paniculate cobalt
burial.
A cadmium mass balance for the Delaware Bay is shown in Figure 4. Cadmium is both
vi
-------
atmospheric transport
natural and
anthropogenic
Fluxes in grams/second
wet deposition
+ 0.02
i'
dry deposition
0.02
air/water exchange
R
river
+ 0.62
fluvial transport
groundwater exchange
chemical
and biological
reactions
-0.17 /
tidal exchange
salt marsh
exchange
+ 0.67
sediment
exchange
-1.14
benthic
exchange
-2.6
Figure 3. Cobalt mass balance for the Delaware Bay (Church, 1986).
-------
atmospheric transport
natural and
anthropogenic
Fluxes in grams/second
wet deposition
I 4-0.02
T ai
dry deposition
air/water exchange
rver
^ - s
^V^<°S \
J>J><~'*i«»r'^.«
+ 0.20
fluvial transport
R
I
groundwater exchange
chemical
and biological
reactions
-0.19
tidal exchange /
can
salt marsh
exchange
+ 0.06
sediment
exchange
-0.09
benthic
exchange
+ 0.6
Figure 4. Cadmium mass balance for the Delaware Bay (Church, 1986).
-------
very toxic and displays strong nutrient-like geochemistry. Again, the fluvial input dominates
the cadmium loading to the Delaware Bay, but, unlike cobalt, the salt marsh input is only 10%.
This is because the diagenetic processes in the sediments produce insoluble authigenic cadmium
sulfide precipitates which are more immune to the redox processes of recycling. The
atmospheric input of cadmium is at least 10% of the total input, and in other less urbanized
estuaries would probably be more. This is the consequence of high atmospheric loadings for
cadmium from upwind combustion sources. Most of the estuarine cadmium loading (68%) is
exported to coastal waters, most likely due to strong organic complexing which decreases
reactivity to removal reactions, and to nutrient-like regeneration during bioassimilation. The
measured benthic flux of cadmium out of the sediments is seven times larger than the cadmium
burial rate, and exceeds both the net input and export fluxes. This reflects the large extent of
cadmium benthic recycling and demonstrates the potential effects of activities such as dredging
which could disrupt this cyclic balance.
A.2.2 Semivolatile Organic Contaminants
Of the potential hydrophobic organic chemicals for which a mass balance can be
demonstrated, polychlorinated biphenyls (PCBs) have been studied the most because of their
bioaccumulation, persistence, ubiquitous distribution in the environment, and alleged toxicity.
Eisenreich and co-workers (e.g., Eisenreich, 1987; Baker and Eisenreich, 1990; Baker et al.,
1991; Jeremiason et al. 1993) have accumulated sufficient information on the inventories and
cycling of PCBs in Lake Superior that a PCB budget may be constructed (Figure 5). Lake
Superior is the second largest lake on earth after Lake Baikal, is the largest of the Great Lakes
possessing >50% of its water volume and approximately 20% of the surface freshwater on
earth, has a large lake area to watershed ratio, has a long water residence time of — 170 years,
is oligotrophic, and is driven primarily by atmospheric interactions. Inputs to the lake include
riverine flows (including municipal/industrial discharges) and atmospheric deposition. PCBs
may be lost from the lake by riverine flow through the St. Mary's River, by sedimentation, by
chemical or biological degradation, and by volatilization.
The inventories of PCBs in the Lake Superior ecosystem are:
atmosphere: -200 kg (- 1.2 ng/m3)
water column: ~ 7200 kg (- 0.6 ng/L)
sediment: ~ 5000 kg (- 6 ng/cm2)
Riverine inflow and outflow are estimated to contribute 20 to 50 kg/yr and 40 kg/yr,
respectively, to the mass balance (Table 1). Eisenreich and Strachan (1992) estimate that
atmospheric deposition of PCBs in the late 1980's was - 167 kg/yr, 125 kg/yr in wet deposition,
and 32 kg/yr in dry particle deposition. The burial of PCBs in bottom sediments is ~ 10 to 50
kg/yr based on detailed analysis of 210Pb-dated sediment cores over the whole lake (Eisenreich,
1987; Baker et al, 1991; Jeremiason et a/., 1993). The assumption is made that chemical and
biological degradation reactions are negligible in the mass balance.
IX
-------
i PCS BUDGET FOR LAKE SUPERIOR 1986 |
Rivers
-20-50 kg/yr
Atmospheric Deposition
Wet 125
Dry 32
Total! 67 kg/yr
Atmosphere
200kg
Volatilization
600-4200 kg/yr
670-750 kg/yr
720 kg/ yr
Particle
Settling
•3000 kg/yr
Water Column
-7200 kg
Recycling
-2950 kg/yr
Outflow
-40 kg/yr
Burial
10-50 kg/yr
Sediment
-5000 kg
'Water Column -1980 - 12000 kg
-1986 - 7200 kg
-1992 - 2200kg
'Linear Water Column Loss Rate
1980 to 1986: - 800 kg/yr
1986 to 1988: - 790 kg/yr
*1st Order Loss Rate
1986: - 500 kg/yr
1 Baker and Eisenreich (1990)
2 Mass Balance
3 Swackhamer et al. (1988)
Figure 5. Mass balance of PCBs in Lake Superior
-------
TABLE 1
A PCB BUDGET FOR LAKE SUPERIOR - 1986
Inputs (kg/yr):
Rivers:
Atmosphere:
Wet
Dry
Total
TOTAL
20 to 50
125
32
167
187-217
Outputs (kg/yr):
St. Mary's River
Sedimentation:
Reaction:
Volatilization
Linear Water Column Loss:
TOTAL
40
10-50
7
600-4200
800
840-890
PCBs are lost from the lake by volatilization at a rate of about 600 to 4200 kg/yr (Baker and
Eisenreich 1990) based on air-water gradients and estimated mass transfer coefficients.
Swackhamer et al (1988) estimated PCB volatilization from Siskiwit Lake on Isle Royale in
Lake Superior to be at a rate comparable to an annual loss from Lake Superior of about 720
kg/yr. Measurements of water column PCBs since 1978 suggest a linear loss rate of - 800
kg/yr (1.3 ng/L in 1978 to 0.18 ng/L in 1992). Using the decrease of PCB concentrations in
the water column in the mass balance suggests volatilization is about 670 to 750 kg/yr.
According to this mass balance calculation, atmospheric deposition contributed 77% to
89% of the total inputs of PCBs to Lake Superior in 1986, similar to the earlier calculations of
Strachan and Eisenreich (1988). Annual atmospheric loadings to the Great Lakes are similar to
those to the Chesapeake Bay, reflecting rapid atmospheric mixing and transport over North
America (Figure 6). PCB losses from the lake occur primarily by volatilization which represents
nearly 90% of total losses, while burial in bottom sediments represents only about 5% of total
PCB losses. This finding is consistent with the earlier calculations of Strachan and Eisenreich
(1988), is near the lower end of that estimated by Baker and Eisenreich (1990), and about equal
to the estimate of Swackhamer et al. (1988) based on their Siskiwit Lake studies. Given the
magnitude and uncertainty of the field measured volatilization rates, this process represents a
critical need in the relative loading paradigm.
The estimated first order residence time of PCBs in the water column of Lake Superior
based on the decrease in concentrations over the last 10 to 15 years is 5 to 6 years (Jeremiason
et al., 1993). The majority of the decrease in water column concentrations is attributed to
volatilization, the rate of which will decrease with decreasing water concentrations. Based on
an ecosystem loss rate of 850 kg/yr and an ecosystem inventory of 12,400 kg, a steady state
residence time is about 15 years. The system is, of course, not at steady state and the overall
system response can only be correctly calculated using dynamic models showing changes in
inputs, outputs, and inventories with time.
XI
-------
30 -
20 -
Great Lakes Region
Eisenreich and Strachan, 1992
Wet Flux
Dry Flux
X!
•—i
&H
O
o
30
20
10
Chesapeake Bay
Leister and Baker, 1993
Church et al., 1993
T-PCBS PHE FLA B[K]F B[B]F B[A]P PB CD
i
AS
r
<
- 1
- 1
- 0
1.
1
0
Figure 6. Atmospheric depositional fluxes to the Great Lake
and to the Chesapeake Bay (Strachan and Eisienreich. 1988:
Baker and Eisenreich. 1993; Church et ai.. 1992).
-------
The mass balance paradigm is a necessary framework to estimate relative loadings of
chemical constituents to lakes and estuaries. To correctly do so requires the measurement of
concentrations, inventories, and fluxes over time in a precise and accurate manner to statistically
demonstrate differences in absolute and relative loadings.
A.2.3 Mercury
The prominence of atmospheric mobilization and depositional processes in the global
biogeochemical cycling of mercury is well known. Atmospheric mercury emissions associated
with contemporary human endeavors are comparable to natural emissions, and atmospheric
deposition is, in general, the principal input of mercury to natural waters. Current international
human-health and environmental concerns associated with elevated levels of monomethylmercury
(MMHg) in freshwater and marine piscivorous fish have focused attention on mercury as a
pollutant, and on its atmospheric cycle. Most of the mercury species in the troposphere are in
the vapor phase, and consist almost entirely of elemental mercury (Hg°). Yet, recent studies
point to paniculate mercury cycle and scavenging of particle associated mercury as the principal
source of mercury in deposition. Much information is needed about sources, chemical
composition, physical state, and direct impact of mercury compounds to the Great Waters.
Additionally, we must investigate processes associated with the post-depositional in situ bacterial
conversion of mercury species to more toxic forms, especially MMHg, which is the principal
form of mercury in fish. Human exposure to methylmercury compounds comes almost
exclusively from the consumption of fish and fish products.
•
Environmental mercury research is improving. There is a heightened awareness of the
need for accurate and broader measurements of mercury in the environment, and for the
incorporation of ultra trace-metal clean sampling and analytical protocols into mercury research.
Recent analytical developments and trace-metal-free methodologies allow for the determination
of total mercury (HgT), reactive mercury (Hgn), inorganic mercury [Hg(II)], elemental mercury
(Hg°) and alkylated mercury species [monomethyl mercury (MMHg); dimethyl mercury
(DMHg)] at the picomolar to femtomolar level in air, water, and precipitation.
The importance of atmospheric mercury deposition in the aquatic biogeochemistry of
mercury in the Great Waters has been demonstrated in two major mercury investigations: 1) for
seepage lakes as part of The Mercury in Temperate Lakes (MTL) Program in Wisconsin, and
2) for drainage lakes as part of the broadly based investigation, Mercury in Swedish
Environment. Both studies indicate that small increases in atmospheric depositional fluxes of
mercury could yield enhanced mercury concentrations in fish. These two investigations provide
a framework for assessing the quality of the available information for the atmospheric cycling
of mercury, and identifying the parts of the cycle where information is needed, especially as it
relates to the impact of atmospheric mercury deposition to the Great Waters.
For a model temperate seepage system, Little Rock Lake (summarized in Figure 7) total
atmospheric mercury deposition (HgT) of ca. 10 Mg nr2 yl (ca. 66% wet and 33% dry
xiii
-------
Total Mercury: Little Rock Lake (Treatment Basin)
Sedimentation
Sediment
8.9 ±1.3 ng/g
1 gram/mm
Figure 7. Mass balance model for Hg in the treatment basin of Little Rock Lake,
Wisconsin (adapted from Fitzgerald et al., 1991 and based on work and
preliminary budgets appearing in Wiener et al., 1990, Fitzgerald and
Watras, 1989, and to appear in Watras et al., 1992).
-------
deposition) readily accounts for the total mass of mercury in fish, water and accumulating in the
sediments. The ecosystem appears delicately poised with respect to atmospheric inputs, since
a relatively small fraction of the input (<10%) can supply the estimated accumulation of
mercury in fish. This suggests that modest increases in atmospheric mercury loading could lead
directly to elevated levels in the fish stock. As summarized in Figure 8, atmospheric deposition
of MMHg is insufficient to account for the amounts of MMHg observed in biota, thereby
indicating the need for in-lake synthesis. Chemical and physical speciation measurements of
paniculate matter and precipitation point to scavenging of atmospheric paniculate mercury as
the source of mercury in rain. Gaseous mercury in lake water is principally Hg°, and the
evasional fluxes of Hg° are significant. Moreover, the in-situ production and efflux of Hg° could
provide a potential buffering and/or amelioration role in aqueous systems.
The depositional results for mercury as established by the Little Rock Lake budget are
placed in larger regional and geographic perspective. There is broad agreement among the
Swedish work, the Nordic countries precipitation network, complementary whole lake
depositional experiments in Wisconsin and Minnesota, and the results from the MTL program
in Wisconsin. For example, simulation of the mercury flows into and out of a typical Swedish
lake in the southern half of Sweden (Figure 9) clearly demonstrates that: 1) atmospheric mercury
deposition is the preeminent source of mercury to a drainage lake, and 2) evasional fluxes of Hg°
are significant, though the estimates require refinement. One striking difference between the
drainage and seepage lake modelling is the significant portion of the mercury input that is stored
in forest soils of the catchment. On average, present atmospheric deposition is greater than the
output of mercury in run-off waters by about a factor of 10. Thus, even if anthropogenic
mercury inputs were to cease, modern mercury deposition that has accumulated in the soil would
continue to be released to the lakes from the forest soils. Indeed, 70 to 80% of the mercury in
the catchment is anthropogenic, and as a consequence, the watershed transport of mercury to the
lakes will remain elevated for long periods of time, perhaps several centuries.
Atmospheric deposition dominates the flux of mercury to lacustrine systems and the open
ocean, and it appears that modest increases in atmospheric mercury loading could lead directly
to enhanced levels of mercury in biota. The U.S. and Scandinavian studies of current and
historical mercury deposition show broad agreement. Mid-latitudinal preindustrial depositional
fluxes of total mercury were ca. 4 /zg nr2 yr1, while present day annual fluxes may vary
between ca. 10 /zg nr2 yr1 in rural semi-remote regions to > ca. 25 fj.g nv2 yr1 in places where
the presence of local/regional mercury sources is pronounced. The influence of anthropogenic
activities on the total mercury cycling is evident, and site specific research must be conducted
to assess the impact of human-related interferences in particular localities. However, the more
important and subtle concerns are associated with the physical and chemical speciation of
mercury deposition. For example, the presence of a significant regional paniculate mercury
cycle is found in specific chemical analysis of mercury in atmospheric paniculate matter and
precipitation. A portion of the HgT observed in precipitation at Little Rock Lake and in
Scandinavian regions, is in a particulate form which is not derived from the oxidation of Hg°
in the atmosphere. Moreover, significant differences are evident in the deposition of HgR, and
differences in HgR inputs may have profound effects on the Hg° and MMHg cycle in natural
xv
-------
Methylmercury: Little Rock Lake (Treatment Basin)
Dissolved
0.06 ±0.03 ng/L
0.02 gram
Seston
5.013.0 ng/g
0.02 gram
Fish
156 ±4 ng/g
0.15 gram
Sedimentation
2
Figure 8. Mass balance model for monomethyl Hg In the treatment basin of Little Rock
Wisconsin (adapted and based on work from Fitzgerald et al., 1991; Wiener t
1990; Hurley et al., 1991; Bloom et al., 1991, and Watras, et al., 1992).
-------
Figure 9.
Flows of Hg in g yr for a typical southern Swedish lake with an area of 1 ka
and a drainage area of 10 km . Atmospheric depositional fluxes of Hg are for
precipitation and do not include dry deposition. Adapted from Johansson, et al
1991, and Lindqvist et al., 1984.
-------
waters.
The production and evasion of Hg° in natural waters is a major feature of the aquatic
biogeochemical cycling of mercury. Significant effluxes of Hg° have been observed in seepage
lakes in Wisconsin as well as in a diverse range of fresh and ocean waters. Thus, Great Waters
with aquatic conditions favoring Hg° production would be less likely to have elevated levels of
mercury in fish. Although, such conditions are poorly known, Kg* production appears to
correlate with the availability and supply of HgR (the Hg(II) substrate) whether it is
atmospherically derived as in seepage lakes or supplied principally through upwelling as in the
equatorial Pacific. The evidence suggests that HgR found in precipitation and atmospheric
paniculate matter is derived from the atmospheric oxidation products of Hg° in the atmosphere.
This form of Hg is labile and highly reactive in aqueous systems and readily available, for
example, to participate in competitive reactions associated with methylation, reduction to Hg°,
uptake by biota, and sequestering with humics. The other fraction of the HgT in deposition is
the operationally defined strongly bound mercury portion ("unreactive" mercury ) and its
environmentally activity is not known. This fraction is probably associated with soot and may
be strongly bound or sequestered in some type of sulfur-carbon association. However,
unreactive mercury species could be solubilized under anoxic and/or sulfitic conditions in natural
waters and sediments to yield a species such as Hg(HS)2° which can be bacterially methylated.
Much study is needed since very few details of the processes affecting production and
destruction of Hg° in the atmosphere and natural, waters are known. There are many questions
concerning short-time scale spatial and temporal variability as well as the importance of
photoreduction reactions and redox boundaries (i.e., oxic/anoxic transition zones) in the
production of Hg°. In addition, the relationships among phytoplankton productivity, microbial
populations (e.g., bacterial reduction) on the activity of Hg° should be evaluated. Broadly based
Hg° investigations are required, particularly those including atmospheric speciation research,
ancillary biological studies and concurrent methylation investigations. Seasonal and spatial data
for atmospheric Hg deposition and the evasion of Hg° are limited. This points toward a need to
refine input to and efflux estimates from lake and coastal waters and to assess, quantitatively,
their influences on the overall cycling of Hg in Great Waters.
The strongly bound Hg components in the atmosphere are likely to have different
geographic origins than the HgR species. Soot associated Hg particles, for example, will
probably have an anthropogenic source and a local/regional origin and tropospheric residence
times should range from days to weeks. At present, little is known about this part of the
atmospheric Hg cycle. The HgR fraction, as suggested, is most probably derived from the
oxidation of Hg°. As a consequence, HgR will be coupled to the global mobilization of Hg° and
its anthropogenic and natural sources.
These observations illus'rue the value of the chemical speciation approach to our
developing understanding of the . cling of Hg in nature. Indeed, they force us ask and address
the following general question: How do such speciation changes in the depositional fluxes of Hg
affect the cycling of Hg in aquatic systems, and what causes the variation in the HgT and HgR
xviii
-------
composition found in deposition? At present, there are no unequivocal answers to questions
concerning the sources and variability of the atmospheric Hg species. The significant aspects
of physico-chemical speciation and partitioning in the atmospheric cycling of Hg, and their
influences on deposition, water-air exchange and the biogeochemical behavior and fate of
mercury in aqueous systems have been identified. The present level of knowledge has been
evaluated and summarized. The precision limits for measurements are given, their relative
importance estimated and a research priority delineated.
A.2.4 Nitrogen
It is increasingly apparent that atmospheric deposition provides an important external
source of readily-available nitrogen to coastal waters. A major shortcoming common to most
recent atmospheric nitrogen deposition studies is that deposition has been evaluated relative to
external sources only. The failure to evaluate atmospheric deposition in the context of a
balanced estuarine nitrogen budget has been largely dictated by the reality that our basic
knowledge of the nitrogen cycle in coastal waters is often less than quantitative. However, to
accurately assess the ecological response of atmospheric nitrogen deposition, the overall nutrient
dynamics may-be as important as gauging the relative atmospheric loading. A system-oriented
approach would consider nitrogen sinks, retention, export, internal cycling and rates of
transformation. For example, while the overall nitrogen loading in the Delaware Bay estuary
is estimated to be ten times that in the nearby Chesapeake Bay (Nixon et al,, 1986), it does not
experience the eutrophication problems encountered in the Chesapeake. The seasonally in
atmospheric deposition is also an important consideration in coastal nitrogen dynamics. For
example, while most first order studies to date have examined atmospheric nitrogen inputs on
an annual basis, eutrophication is not a serious problem in the Chesapeake Bay during the fall
and winter. Thus, an examination of atmospheric nitrogen loading during the spring and
summer is probably more pertinent, although the feasibility of seasonally-based nitrogen
emission control policies is debatable from a management perspective.
To illustrate the current state of our understanding, a annual budget for dissolved
inorganic nitrogen (NCy + NH/) in the Delaware Estuary can be examined (Figure 10). This
budget does not include organic nitrogen, or dissolved/paniculate partitioning. The Delaware
Bay was chosen for this case study for several reasons. First, marine systems tend to be
nitrogen limited while freshwater systems are phosphorus limited. Second, in contrast to the
Chesapeake Bay, Delaware Bay is a more simple system to examine in terms of hydrology and
nutrient inputs from large population centers. The freshwater boundary for the nitrogen mass
balance in Figure 10 is the fall line at Trenton, New Jersey, above the heavily urbanized and
industrialized Philadelphia-Camden-Wilmington corridor. The marine boundary for the mass
balance extends between Cape May, New Jersey and Cape Henlopen, Delaware. The specific
details of the derivation of each flux term can be found in Scudlark and Church (1993).
The primary nitrogen inputs to the Delaware Bay estuary are provided by point
discharges and fluvial transport, the assessment of which appears to be fairly well constrained.
Contrary to the traditional "outwelling" theory describing salt marsh nutrient dynamics,
xix
-------
atmospheric transport
natural and
anthropogenic
Ij47±5
.. Mori
Fluxes in moles/second
Isources =114 moles/second
Isinks = 193 moles/second
air/water exchange
Djrect:
Indirect:
R
3±2
5±5
dry deposition
+ 27±3
river fluvial transport
groundwater exchange
I
9
chemical tidal exchange / (?)
and biological
reactions
-177±18
salt marsh
exchange
+ 2 + 2
sediment
exchange1
22±11
benthic
exchange'
-16±5
ocean
\
\
i
i
\
t
Figure 10. An annual nitrogen loading budget for the Delaware Bay
-------
pervasive coastal wetlands do not appear to provide a significant source of inorganic nitrogen
to the Delaware Bay. The "sediment exchange" term in Figure 10 represents the benthic flux
of recycled nitrogen (primarily NJV) and has been estimated from several independent studies.
Nitrogen fixation is thought to be relatively minor in marine systems, while the relative
importance of the direct absorption of gaseous ammonia from the atmosphere is unknown. Shelf
exchange, a unique feature of coastal waters, is discussed in Section 8.4.1. Direct wet
deposition of nitrogen from the atmosphere has been estimated based on five monitoring sites
in the Delaware Bay watershed, and is assumed to comprise 50% of the total (wet4-dry)
atmospheric loading. Indirect atmospheric input via watershed export was estimated based upon
a transmission factor of 0.1 (i.e., 90% of the nitrogen deposited on the drainage basin is retained
within the landscape or is lost in feeder tributaries prior to entering the bay proper.
Phytoplankton uptake provides a dominant, though arguably temporary, sink of nitrogen
from the water column. The "benthic exchange" term in Figure 10 refers to sedimentary
denitrification, which results in a net loss of gaseous end products (Nj, N2O) from the system.
It is clear from FigurelO that the nitrogen sinks, primarily phytoplankton uptake (£=193
moles/s), are not nearly balanced by the sources (£=114 moles/s). In this context, atmospheric
inputs provide less than 10% of the primary production requirements. Potentially important non-
quantified source terms which would help close this deficit include water column
remineralization, ungauged groundwater input, and tidal exchange.
The primary uncertainties associated with atmospheric nitrogen deposition to coastal
waters are related to (1) estimation of dry depositional fluxes, and (2) gauging watershed
retention. Estimation of dry deposition by inferential means is currently limited by our basic
ability to accurately measure atmospheric NO3" concentrations. Other potentially important
research needs, for which less is known are (3) accurately assessing the concentration and
deposition of organic nitrogen in precipitation and aerosols, and (4) quantifying the gas/water
mass transfer of NH3 .and HNO3 to surface waters. In recognition of the increased focus on
deposition in the coastal zone, wet and dry deposition monitoring networks should make a
greater effort to include coastal sites. Process-oriented studies addressing the effects of urban
plumes are also warranted.
xxi
-------
A.3 CURRENT UNCERTAINTIES IN GREAT WATERS MASS BALANCES
There has been tremendous improvements in our understanding to the role of atmospheric
deposition in supplying trace elements, mercury, nitrogen, and organic contaminants to surface
waters. Many advances have been made in the basic sampling and analytical tools required to
make reliable field measurements of atmospheric inventories and wet depositional fluxes. Such
measurements require extreme care and are, therefore, not necessarily suitable for large, routine
monitoring networks. Nonetheless, during the past ten years it has become possible to
accurately measure atmospheric concentrations, speciation, and wet depositional fluxes of many
chemicals, and excellent records are evolving at several locations (e.g., Lewes, DE; Chesapeake
Bay, Great Lakes region). Due to the inherent variability in atmospheric processes, long-term
records, on the scale of decades, will be required to assess changes in atmospheric deposition
loadings to the Great Waters. It is left up to the responsible agencies to develop and maintain
these long-term programs.
To prioritize future research efforts, the authors of this report have estimated the current
uncertainties in the fundamental atmospheric depositional processes and assessed their relative
importance (Tables 2-5). It is important to note that the "achievable uncertainty" refers to the
optimal precision obtainable for a measured parameter at a single site. As such, these estimates
should be viewed as goals rather than an accurate, assessment of the quality of currently available
da: from the Great Waters. In general, more reliable measurements of wet deposition are
available compared to either dry aerosol deposition or gaseous exchange. While wet deposition
can be measured in the field, our ability to predict (e.g., model) contaminant scavenging from
the atmosphere by precipitation is highly uncertain, perhaps no better than to within one to two
orders of magnitude. Specific studies of wet depositional processes, especially those employing
novel geochemical tracers and airborne sampling, are required. While it is important to continue
and expand wet deposition measurements and research, much of the research effort must be
placed in improving our ability to measure and model dry deposition. In particular, our
estimates of dry aerosol deposition are hindered both by a lack of aerosol size distribution
information and by our generally poor understanding of the micrometeorological environment
above water surfaces. The potentially large contaminant fluxes resulting from the rapid settling
of supermicron particles near emission sources (Hoi sen et al., 1991) as well as the possible
elevated fluxes during short-term, intensive meteorological events deserve further study.
Mass balance calculations and measurements of concentration gradients in the field
strongly suggest that many organic contaminant are degassing from the Great Lakes, especially
during warm summer months. Chemicals such as PCBs, which are no longer produced, may
be leaving the Great Lakes back into the atmosphere to be transported and deposited to the
world's oceans and to the polar ice pack. The processes by which the Great Waters give these
chemicals back to the atmosphere clearly need to be understood, both to predict contaminant
levels in these water bodies and to characterize the global redistribution of these persistent
chemicals.
xxu
-------
In summary, the significant progress made during the past two decades has provided
many of the tools required to answered the questions posed by Section 112(m) of the 1990
Amendments to the Clean Air Act. While much remains to be done, the regulatory community
will be well served to continue to adopt the geochemical mass balance, emphasizing processes
and fluxes of materials, as a rational framework for future endeavors
XXII!
-------
TABLE 2
CURRENT UNCERTAINTIES IN ESTIMATING THE ROLE OF ATMOSPHERIC
DEPOSITION IN TRACE ELEMENT MASS BALANCES TO THE GREAT WATER!
(1993)
(+ = low, + + = medium, + + + = high)
Wet Deposition
Precipitation volume
Precipitation concentration
Diss./paft. distribution
Speciation
Dry Deposition
Aerosol concentration
Aerosol deposition velocity
Aerosol size distribution
Surface microlayer reflux
Episodic events/turbulence
Gaseous exchange
Overall Loadings Estimates"11"
Wet Loading
Dry Loading
Minimum Technical
Uncertainity"
+
+
+ +
+ + +
+
+ + +
+ +
+ + +
+ + +
+ + +
+ +
+ + +
Importance
to Loading
Calculations
+ + +
+ + +
+ +
+ +
+ +
+ -f +
+ +
•f
+ +
+
+ + +
+ + +
Research
Priority
+
+ +
+ +
+ + +
+
+ + +
+ +
+
+ +
+ +
In the authors' judgement, this is the highest precision which could be obtained when
suitably qualified personnel use the best current methods for an adequate period of
time. The actual uncertainity in many of these parameters for specific locations are
much greater due to a lack of quality information.
~8The parameters listed above are site- and time-specific. Additional uncertainity is
introduced in the extrapolation of these parameters to estimate spatially- and
temporally-integrated contaminant loading rates.
-------
TABLE 3
CURRENT UNCERTAINTIES IN ESTIMATING THE ROLE OF ATMOSPHERIC
DEPOSITION IN SEMIVOLATILE ORGANIC CONTAMINANT MASS
BALANCES TO THE GREAT WATERS (1993)
(+ = low, 4-4- = medium, 4-4- + = high)
-
Wet Deposition
Precipitation volume
Total atmospheric concentration
Gas/aerosol distribution
p" > 10's ton-
ic4 torr > p° > 10'3 torr
p° < 10-* ton-
Total precipitation concentration
Aerosol scavenging coefficient
Gas scavenging coefficient
Dry Deposition
SOC aerosol size distribution
Aerosol deposition velocity
Gas Exchange
mass transfer coefficient
total SOC concentration in water
SOC speciation in water
Henry's Law constant = f(T)
Overall Loadings Estimates"9'
Wet Loading
Dry Loading
Minimum
Technical
Uncertainity"
+
+
+ +
+ 4-4-
4- +
+ +
4-4- +
+ 4-
4-4-4-
+ 4-4-
+ +
4-
4- +
+ +
4- +
+
Importance
to Loading
Calculations
4-4- +
+ 4-
4-4-4-
4-4-4-
+ 4- +
+ + +
+ + +
+ + +
+ +
+ + •+•
4- + +
+ + +
+ + +
+ 4- +
+ + +
+ + 4-
Researcb
Priority
•
4-
4-4-
4-4-
4-4-4-
4- -t-
+ +
4-4-4-
4-4-
4- 4-
+ 4-4-
4-4-
4-4-
4-4-4-
4-4-
In the authors' judgement, this is the highest precision which couid be obtained when suitably qualified
personnel use the best current methods for an adequate period of time. The actual uncertainity in many of these
parameters for specific locations are much greater due to a lack of quality information.
"*The parameters listed above are site- and time-specific. Additional uncertainity is introduced in the
extrapolation of these parameters to estimate spatially- and temporally-integrated contaminant loading rates.
-------
TABLE 4
CURRENT UNCERTAINTIES IN ESTIMATING THE ROLE OF ATMOSPHERIC
DEPOSITION IN MERCURY MASS BALANCES TO THE GREAT WATERS (1993)
(+ = low, + + = medium, + -f + = high)
Wet Deposition
Precipitation volume
Gas scavenging (Hg°)
Aerosol scavenging
Atmospheric concentrations
Diss./part. distribution
Speciation
Dry Deposition
Aerosol concentration
Aerosol size distribution
Aerosol composition
Aerosol reactivity
Mercury Vapor Exchange
Chemical Speciation
Atmospheric oxidation
Gas exchange (flux)
Overall Loadings Estimates"'1"
Wet Loading
Dry Loading
Minimum Technical
Uncertainiry*
+
+ + +
+ + +
+ +
+ +
+ +
+ +
+ + +
+ + +
+ + +
+
+ + +
+ + +
+ +
+ + +
Importance to
Loading
Calculations
+ + +
+ +
+ +
+ +
+ +
+ + +
+ +
+ +
+ +
+ + +
+
+ + +
+ + +
+ + +
+ + +
Research
Priority
+
+ +
+ +
+ +
+ +
+ -r +
+ +
+ +
+ +
+ + +
+
+
+ + +
In the authors' judgement, this is the highest precision which could be obtained when suitably qualified
personnel use the best current methods for an adequate period of time. The actual uncertainly in many of th<
parameters for specific locations are much greater due to a lack of quality information.
*The parameters listed above are site- and time-specific. Additional uncertainity is introduced in the
extrapolation of these parameters to estimate spatially- and temporally-integrated contaminant loading rates.
-------
TABLE 5
CURRENT UNCERTAINITIES IN ESTIMATING THE ROLE OF ATMOSPHERIC
DEPOSITION IN NITROGEN MASS BALANCES TO THE GREAT WATERS (1993)
(4- = low, 44 = medium, 444- = high)
Parameter
Wet Deposition
Precipitation volume
Atmospheric concentration
and speciation
Dry Deposition
NO2 aerosol cone.
DON aerosol cone.
NHS aerosol cone.
NO3' aerosol cone.
Aerosol size distribution
Gas/water partitioning
Overall Loadings Estimates"11"
Wet Loading
Dry Loading
Watershed transmission
. Minimum Technical
Uncertainity"
4-
(NH4+, NO,") 4-
(DON) 444-
4-4-4-
444-
4-
4-4-
4- •
44-4
4-
4-4
4-4-4-
Importance
to Loading
Calculations
4-4-4
4-4-4-
+ 4
4-4
4-4
4-4-4-
4-4- +
+
44
4-4-4
4-4-4-
4-4-4-
Research
Priority
4-
4-
4-4 +
44
4-4 +
4-4-
4-4
4-
444-
*In the authors' judgement, this is the highest precision which could be obtained when
suitably qualified personnel use the best current methods for an adequate period of
time. The actual uncertainty in many of these parameters for specific locations are
much greater due to a lack of quality information.
""The parameters listed above are site- and time-specific. Additional uncertainly is
introduced in the extrapolation of these parameters to estimate spatially- and
temporally-integrated contaminant loading rates.
-------
Relative Atmospheric Loadings... Revision Date: 15 March 1993
1.0 INTRODUCTION AND SCOPE OF THIS REPORT
The 1990 Amendments to the Clean Air Act (CAA) recognize a geochemical paradigm
which has been evolving during the past 30 years. Pollutants emitted into the atmosphere are
transported at various distances and may later deposit in aquatic systems far removed from their
original sources. This atmospheric deposition may seriously degrade water quality. Linkages
between the atmosphere and surface waters operate on several scales, ranging from trace metal
contamination of ponds adjacent to smelting operations to regional episodes of acidic deposition
to global dispersion of organochlorine insecticides and greenhouse gases. As increasingly
stringent controls are applied to conventional (i.e., point) sources of pollutants to surface waters,
the relative importance of diffuse, non-point sources of contamination, including loadings from
the atmosphere is increasing. This fact requires a fundamental revaluation of our approach to
controlling air and water quality, as it has become increasingly evident that one cannot achieve
water quality objectives if corresponding air quality goals are neglected.
Section 112(m) of the 1990 CAA Amendments requires the U.S. Environmental
Protection Agency and the National Oceanic and Atmospheric Administration to estimate the
importance of atmospheric deposition of hazardous air pollutants to the Great Lakes, Lake
Champlain, the Chesapeake Bay, and other coastal waters (collectively dubbed the Great
Waters). This section requires not only that gross atmospheric contaminant loadings to each
water body be documented, but, more importantly, that the relative importance of those loadings
compared to all those from all other possible sources be quantified. Further, the agencies are
required to determine whether atmospherically-derived contamination results in exceedences of
water quality standards, and to estimate the fraction of contaminants accumulating in biota which
are derived from the atmosphere. Simply stated, Section 112(m) requires the agencies to
construct quantitative chemical mass balances for relevant contaminants in each of the Great
Waters. This is clearly a tall order.
In this paper, we first summarize the current understanding of atmospheric depositional
processes for trace elements, mercury, nitrogen, and synthetic and combustion-derived organic
contaminants. After addressing the question of whether gross contaminant loadings from the
atmosphere can be estimated, we address whether both the conceptual understanding and the
necessary data to construct defensible mass balances of these chemicals in the Great Waters are
available. Specifically, we address the following questions:
1. What is the consensus view of our current understanding of the fundamental processes
resulting in atmospheric deposition?
2. Do we have the tools to determine atmospheric deposition rates with the accuracy and
precision required to make the regulatory decisions required by the Great Waters
Program?
-------
Relative Atmospheric Loadings... Revision Date: 15 March 1993
3. Do we have the conceptual understanding required to estimate the relative atmospheric
loadings of contaminants to the Great Waters?
4. Do we currently have data of sufficient accuracy and precision to estimate relative
atmospheric loadings of contaminants to the Great Waters? and,
5. What specific studies are required to improve our ability to address the relative loadings
questions of the Great Waters Program?
2.0 RELATIVE LOADINGS: THE MASS BALANCE PARADIGM
The framework for interpreting the relative inputs of chemicals into a body of water is
a mass balance or input-output budget. The boundaries are often defined as the water column
of the water body, and mass exchanges across the air-water, sediment-water, and land- water
interfaces are inputs and outputs. Considering the case of a fresh water body as a chemical
reactor (Figure 1), contaminants may enter by riverine flow (dissolved and paniculate matter),
ground water flow (dissolved matter), atmospheric deposition (in the form of gas and particle
scavenging by rain and snow, dry particle deposition, and gas absorption at the air- water
interface), sediment and benthic layer exchange (dissolved and paniculate), and in situ
production. Chemical outputs from the lake volume include dissolved gas volatilization, riverine
and connecting channel outflow (dissolved and paniculate), chemical and biological degradation,
sedimentation and burial of particles, and ground water output (dissolved). Chemical species
entering the water column may undergo turbulent and diffusive mixing and reactions which result
in a change of speciation (e.g., dissolved to paniculate; oxidation or reduction; hydrolysis and
complexation). In a general sense, the time rate of change in chemical concentration in the
water column is:
V dc/dt = EQjC, - EQ0CTtir
Wet Deposition Air-Water Exchange Dry Particle Deposition
dc
Where V — represents the time rate of change in total chemical mass in the well-mixed water
at
column (mols/yr), V = lake volume (m3); A = lake surface area (m2); CTiH20, CPiH2o, Cd-H20, C«d
= total, paniculate, dissolved and surface sediment concentrations of chemical (mols/m3); CTrtin
= total rain concentrations (mol/m3); J = annual precipitation amount (m/yr); Q; and Q0 =
volumetric hydraulic inflow and outflow rates (m3/yr), respectively; p atmospheric partial
pressure of the gas (Pa); H = Henry's Law constant [(Pa.m3)/mol]; k,, k,, and k^, = rate
coefficients (yr1) describing sedimentation, resuspension, and other reactions acting on the
chemical (i.e., photolysis, hydrolysis, biodegradation); Vd = dry particle deposition velocity
-------
Relative Atmospheric Loadings... Revision Date: I5 March 1993
(m/yr); and Cp,^ = the concentration of chemical in the atmospheric particle phase. If the rate
of chemical input equals the rate of chemical output (i.e., no net change in water column
inventory with time), the residence time (r) is then [MT/E(fluxes in or out)], where MT is the
total mass of the chemical. The residence time is a measure of the average time it takes for a
molecule of a chemical entering the water column to leave via the sum of all loss processes.
For example, the water residence times in Lake Superior and Lake Michigan are about 170 years
and 100 years, respectively, while the residence time for non-conservative species such as PCBs
and lead are only 1 to 5 years.
The components of a mass balance model of estuarine ecosystems such as the Chesapeake
Bay and the Delaware Bay differ from those of lakes because of the importance of tidal exchange
of water, solutes, and particles, and because of the frequently higher productivity of coastal
marshes and waters (Figure 2). Over the period of several days to weeks, tidal flow flushes
material of the estuary and exerts a significant control on estuarine water quality. Therefore,
the water residence times in estuaries are often significantly less than large lakes. However,
estuaries are also efficient traps of paniculate matter, leading to longer residence times of
particle-reactive chemicals. For example, the hydraulic residence time of the Chesapeake Bay
is less than one year, while the water column residence time of particles are significantly longer.
For estuaries, the equation describing the time rate of change in chemical concentration in the
water column is similar to that of a fresh water system, with the inclusion of a tidal component.
The construction of a chemical mass balance of a lake or estuary demands that chemical
inputs, outputs and internal losses and gains be quantified. For the Great Lakes, mass budgets
have been attempted for semivolatile organic contaminants (SOCs) such as polychlorinated
biphenyls (PCBs) (Eisenreich 1987; Swackhamer and Armstrong 1986, Swackhamer ei al. 1988;
Strachan and Eisenreich 1988; DePinto et al. 1992); and PAHs (McVeety and Kites 1988,
Strachan and Eisenreich 1988). Inputs usually quantified are atmospheric deposition, riverine
inflow and outflow, air-water exchange, and sedimentation. Processes often ignored or poorly
quantified are chemical and biological degradation, groundwater flows, benthic exchange, and,
in the case of estuaries, tidal exchange. Quantification of riverine inputs and outputs must
recognize the inherent seasonal variability of flows, and, therefore, loadings. Likewise, air-
water exchange of SOCs is a strong function of water temperature and wind conditions,
necessitating seasonally dependent studies. Mass budget studies incorporating losses of chemical
to sedimentation must obtain a sufficient number of spatially-representative sediment
accumulation rates, a condition normally not encountered in mass balance studies. Contaminated
sedimentary deposits may release nutrients, metals, and organic contaminants by diffusive and
advective processes on a continuous or seasonal cycle, or in catastrophic episodes such as
hurricanes, all contributing to the mass budget of the water column.
Until recently, atmospheric deposition of toxic chemicals to the Great Waters was a
complete mystery. Now, reasonable estimates of wet deposition of trace elements and SOCs are
being generated, but dry particle deposition is still very uncertain. Air-water exchange is onlv
-------
Relative Atmospheric Loadings... Revision Date: 15 March 1993
now becoming recognized as an important process (e.g., Achman et al. 1993a,b; McConnell ei
al. 1993; Atlas et al. 1986; Strachan and Eisenreich 1988).
The mass balance paradigm is a logical framework to establish the relative loadings of
chemicals to the Great Waters. The first question is what is the magnitude of atmospheric
deposition of target nutrients, trace metals, and organic chemicals to the Great Waters. Nex..
one must ask what is the magnitude of input of these target species from all other sources. Only
then can the relative importance of each source be determined. This report details the processes
by which trace elements, mercury, nitrogen, and organic chemicals are removed from the
atmosphere and are loaded to the Great Waters, reports on the magnitude of chemical input from
the atmosphere, and present case studies where the mass balance paradigm has been used to
determine the relative importance of the atmosphere as a source of contamination to the Great
Waters.
3.0 EVIDENCE OF ATMOSPHERIC DEPOSITION TO THE GREAT
WATERS
3.1 TRACE ELEMENTS - Evidence of Deposition
Trace elements are those in rarer geochemical abundance compared to major crustal
elements. These elements generally pass more rapidly through global reservoirs due to their
reactivity in terms of abiotic hydrolysis reactions or biotic uptake as nutrients or toxics. In the
atmosphere, trace elements are generally the result of global biogeochemical processes that
include weathering, emissions, and transport. Aeolian transport can provide substantial trace
element inputs to water bodies adjacent or downwind to major atmospheric sources or those
characterized by a low ratios of watershed-to-surface area.
The atmosphere receives and processes trace elements from both natural and
anthropogenic sources (Nriagu 1989, 1992; Nriagu and Pacyna 1988). Natural sources include
both geochemical and biochemical processes and industrial sources include burning of fossil fuel
and metalliferous production (Table 6). Natural and anthropogenic processes suspend fugitive
dusts during dry weather and agricultural tilling (Table 7). Likewise, volatilization results from
either low temperature emissions (such as evapo-transpiration or surface evaporation), or high
temperature combustion (such as volcanoes, forest fires, and combustion of fossil fuels; Tables
6 and 7). Dry deposition is largely responsible for the removal of fugitive aerosols greater than
a micron or in mass median diameter from the atmosphere. However, sub-micron aerosols or
combustion condensates of trace metals are subjected to long atmospheric residence times and
subsequent transport of days to weeks over the continents and eventually over inland or coastal
water bodies. Wet deposition in the form of precipitation becomes an important means of
atmospheric trace metal scavenging from the troposphere for these sub-micron aerosols (Junge
-------
TABLE 6
EMISSIONS OF TRACE METALS FROM NATURAL SOURCES TO THE
ATMOSPHERE (xlO6 kg/yr, FROM NRIAGU 1989)
Element
As
Cd
Cr
Co
Cu
Hg
Mn
Mo
Ni
Pb
Sb
Se
V
Zn
Soil-Derived
Dust
2.6
0.21
27
4.1
8.0
0.05
221
1.3
11
3.9
0.78
0.18
16
19
Seasalt
Sprays
1.7
0.06
0.07
0.07
3.6
0.02
0.86
0.22
1.3
1.4
0.56
0.55
3.1
0.44
Volcanoes
3.8
0.82
15
0.96
9.4
1.0
42
0.40
14
3.3
0.71
0.95
5.6
9.6
Forest
Fires
0.19
0.11
0.09
0.31
3.8
0.02
23
0.57
2.3
1.9
0.22
0.26
1.8
7.6
Biogenic
Sources
3.9
0.24
1.1
0.66
3.3
1.4
30
0.54
0.73
1.7
0.29
8.4
1.2
8.1
Total
12
1.4
43
6.1
28
2.5
317
3.0
29
12
2.6
10
45
45
-------
TABLE 7
WORLDWIDE EMISSIONS OF TRACE METALS FROM INDUSTRIAL SOURCES.
UNITS: xlO6 kg/yr (FROM NRIAGU AND PACYNA, 1988)
Process
Energy Production
Mining
Smelting and Refining
Manufacturing
Processes
Commercial
Applications
Waste Incineration
Transportation
TOTAL
As
2.2
.06
12.
-
2.0
.31
-
19.
Cd
.79
-
5.4
.60
-
.75
-
7.6
Cr
13.
-
-
17.
-
.84
-
3.1
Cu
8.0
.42
23.
2.0
-
1.6
-
35.
Hg
2.3
-
.13
-
-
1.2
-
3.6
Mn
12.
.62
2.6
15.
-
8.3
-
38.
Ni
42.
.80
4.0
4.5
-
.35
-
52.
Pb
13.
2.6
46.
16.
4.5
2.4
248
332
Sb
1.3
.10
1.4
-
-
.67
-
3.5
Se
3.8
.16
2.2
-
-
.11
-
6.3
Sn
3.3.
-
1.1
-
-
.81
-
5.1
V
84.
-
.06
.74
-
1.2
-
86.
Zn
17.
.46
72.
33.
3.2
5.9
-
132
Ti
I.I
-
-
4.0
-
-
-
5.1
-------
Relative Atmospheric Loadings... Revision Date: 15 March 1993
1977). Non-crustal metals are typically enriched in sub-micron aerosols over the remote areas.
As such, the atmosphere can act as a powerful agent directly transporting and rapidly depositing
continental materials to adjacent water bodies and their watersheds. In doing so, atmospheric
deposition may largely determine the trace element input, composition, and biological exposure
in certain water bodies.
Due to the typically low concentrations encountered and high potential for extraneous
contamination, the accurate sampling and analysis of trace elements in the atmosphere and in wet
deposition requires special equipment, rigorous procedures, and sensitive analytical equipment.
This has largely limited the extent and quality of previous studies which attempted to evaluate
the importance of atmospheric trace element deposition to inland and coastal waters (Galloway
et al. 1982, Barrie et al. 1987). What follows is a critical review of these studies to date
recognizing that mercury is treated separately from trace elements in this report.
The first successful attempts to measure trace elements in the atmosphere were done over
the ocean. The Sea Air Exchange Program (SEAREX) of the NSF sampled trace elements in
aerosol and in precipitation over the Atlantic (Duce et al. 1976) and in aerosols over the Pacific
(Arimoto et al 1985, 1987, 1990). Initial SEAREX results from over the Atlantic drew early
attention to the hypothesis that atmospheric deposition is an important, if not dominant, source
of trace elements to oceanic surface waters. French scientists supported this hypothesis in the
sub-tropical Atlantic by showing the similarity of trace element enrichments in oceanic aerosols
compared with those in oceanic particulates (Buat-Menard and Chesselet 1979). Atmospheric
deposition of trace elements to the Sargasso Sea (Jickells et al. 1984, Church et al. 1984)
dominates the inputs to and flux from the ocean water column (Jickells et al. 1987). These
results are likely to extrapolate into the more northern Atlantic, where the continental sources
are well defined (Church et al. 1990), the concentrations comparable (Church et al. 1991), and
the transport in and scavenging from the upper troposphere is more efficient (Church et al.
1992). In the Pacific, seasonally (spring) large amounts of aeolian dust and associated trace
elements are transported from southeast Asia over distances of tens of thousands of kilometers
(Duce et al. 1980). The corrosive redox conditions of this dust during transport and wet
scavenging can lead to the extraordinary large concentrations of iron (Zhuang et al. 1992) and
other trace elements (Maring et al. 1989) in precipitation. Nevertheless, during similar Saharan
dust transport over the sub-tropical Atlantic, much of the crustal trace elements remain as
insoluble particles in the precipitation (Lim and Jickells 1990). Geochemical budgeting which
demonstrates the important role of atmospheric deposition to the ocean has subsequently been
tested and confirmed for a number of inland ocean water bodies such as the Baltic,
Mediterranean, and North Seas (Jickells et al 1989), and even on a global basis (Duce et al.
1991). Atmospheric deposition exceeds riverine inputs to the world oceans for many trace
elements (Table 8, Nriagu 1992).
Atmospheric trace element deposition is important to coastal and inland waters (Patterson
and Settle 1987, Nriagu 1992). The sources and transport of aerosol trace elements has been
-------
TABLE 8
ATMOSPHERIC VERSUS RIVERINE INPUTS OF TRACE METALS
IN THE OCEAN (FROM NRIAGU 1992)
Element
As
Cd
Cr
Cu
Hg
Mo
Ni
Pb
Sb
Se
Zn
Dissolved in Rivers
(ng/L)1
47
2.1
172
115
0.82
12
135
8.5
14
35
165
River Input
(xlO'g/yr)2
1.6
0.07
5.8
4.0
0.03
0.41
4.6
0.29
0.48
1.2
5.6
Atmospheric Input
(xl09g/yr)3
5.8
3.2
-
34
1.7
-
25
88
-
-
136
'Based on published data obtained using the ultra-clean laboratory procedures.
2Assuming total discharge of water by rivers to be 3.4 x 1016 L yr"1.
3From GESAMP 1989.
-------
Relative Atmospheric Loadings... Revision Date: 15 March 1993
evaluated in the Northeast United States (Rahn and Lowenthal 1984, 1985), and the record of
their recent input have been inferred from profiles recorded in salt marshes (McCaffrey and
Turekian 1980, Bricker-Orso and Nixon 1989), lake sediments (Norton and Kahl 1992, Flegal
et al 1989), and forest soils (Friedland 1992). The longest current trace elements deposition
record to coastal waters has been made at Lewes, Delaware by the University of Delaware at
an atmospheric research site near Cape Henlopen on the mid-Atlantic coast. Wet deposition has
been collected on an event basis for trace metals and analyzed since 1982 (Church et al. 1984,
Church and Scudlark 1992). The success of this record has come from the development of
suitably clean protocols and sufficiently accurate and precise analytical techniques developed
specifically for trace elements in precipitation (Tramontane et al. 1987, Scudlark et al. 1992).
The data from Lewes, Delaware (Church and Scudlark 1992), show that the concentrations of
trace elements in rural precipitation far exceed those of most surface waters and are in excess
of natural crustal or sea water sources. Trajectory analyses of air masses associated with
precipitation events reveal that the excess sources of trace elements are similar to those for acid
precipitation, namely emissions from the combustion of fossil fuels or metal refining. Wet
depositional fluxes of many trace elements at the Lewes, Delaware site are relatively constant
over the past eight years, with the notable exception of lead. Concentrations of lead in
precipitation and wet depositional fluxes decline from the de-leading of automobile gasoline
(Figure 11). This same decreasing trend of lead deposition has been confirmed as well in the
northern Great Lakes area of Minnesota (Eisenreich et al. 1986).
The magnitude of trace element deposition at the Lewes, Delaware site is compared to
other coastal inputs in Table 9. Although the atmospheric depositional flux of trace elements
into the watershed can dominate that which crosses the fall line (Church et al. 1988), the amount
falling directly into the open estuarine waters of Delaware Bay appear to be minor (Church.
1987). However, the amount of atmospheric input for metals such as Cd, Pb, and Zn appear
to dominate the riverine/estuarine inputs into coastal waters of the mid-Atlantic bight (Church
1987, 1992). However, there are several problems in deconvoluting the fraction of trace metals
entering the estuary directly from weathering versus indirectly from the run-off of trace elements
previously deposited from the atmosphere. Part of the problem involves complex trace element-
watershed interactions (Lindberg and Turner 1988).
Other studies which document the atmospheric deposition of trace elements to coastal
waters such Puget Sound (EPA 1991) are limited in length of record, location of sampling
points, and analytical procedures. The Great Lakes Atmospheric Deposition (GLAD) Network
has measured wet deposition of trace elements to the Great Lakes (Klappenbach 1992). Most
recently, the Chesapeake Bay Atmospheric Deposition Study (CBADS) also documented dry and
wet deposition of trace elements (Baker et al. 1992). The relatively uniform atmospheric fluxes
between the CBADS and Lewes sites (Figure 12) suggest the importance of rather distant sources
being transported to this important estuarine system (Wu et al. 1993). A comparison of the wet
depositional fluxes (Figure 12) between the various aquatic systems suggest regional uniformity
in the Eastern United States. Greater fluxes in the past are not supported by the
-------
TABLE 9
ATMOSPHERIC INPUTS OF TRACE METALS TO MID-ATLANTIC COASTAL
MARINE SYSTEMS (FROM CHURCH er al., 1988)
Element
Al
Cd
Cr
Cu
Fe
Mn
Ni
Pb
Zn
Atmospheric Component (%)
Del. Watershed
—
—
—
41
64
-58
43
—
32
Del. Estuary
—
~
—
4
2
—
2
-'
4
Mid-Atlantic Bight
—
—
~
18
~
4
30
96
64
-------
D5
3-
d
c
o
O
•D
-------
Relative Atmospheric Loadings... Revision Date: 15 March 1993
Lewes record, and it is possible that these earlier measurements are biased by poor analytical
protocols. The high trace element fluxes reported for Commencement Bay, Puget Should may
reflect the proximity to local emissions or bias resulting from bulk deposition sampling (EPA
1991).
3.2 SEMIVOLATILE ORGANIC CONTAMINANTS - Evidence of Deposition
Organic chemicals emitted into the atmosphere by anthropogenic activities may be
transported long distances, chemically converted, and deposited at sites distant from their
emission sources. The presence of persistent semivolatile organic compounds (SOC) in the
Arctic snowpack and food chain (Gregor and Gummer 1989, Hargrave et al. 1988, Welch et al
1991, Hinckley et al 1990, 1991, Patton et al., 1989, Gotham and Bidleman 1992) and in the
Antarctic (Tanabe et al. 1983) is ample evidence that long range atmospheric transport and
subsequent deposition is an important global pathway for these chemicals. The occurrence of
current-use agrichemicals on non-target crops adjacent to treated fields (i.e., Seiber and
McChesney 1991) is further evidence of the importance of atmospheric transport and deposition
over shorter spatial scales. The role which atmospheric transport and deposition plays in
delivering SOCs to urban areas has been seldom studied and is consequently poorly understood.
While it may seem intuitive that atmospheric fluxes pale in comparison to easily identified and
measured point sources in urban areas, persistently high SOC levels in urban areas and enhanced
depositional processes may result in significant loadings to near by water bodies. Recent studies
in Chicago (Holsen et al. 1991) measured extremely large depositional fluxes of polychlorinated
biphenyls (PCBs) from the atmosphere, which were attributed to settling of large, PCB-laden
aerosols.
In the upper Great Lakes region, atmospheric deposition is believed to be the major
source of SOCs (Eisenreich et al. 1981, Strachan and Eisenreich, 1988). For example,
atmospheric deposition supplies 90%, 63%, and 58% of the PCB loading to Lakes Superior,
Huron, and Michigan, respectively (Table 10). The atmosphere is also the dominant source of
lead to the Great Lakes, exceeding 95% of estimated external sources to lakes Superior.
Michigan, and Huron (Strachan and Eisenreich 1988). Although regulatory bans on the
production and use of PCBs and other persistent SOCs (i.e., dieldrin) have resulted in decreased
levels of these compounds in the Great Lakes fisheries, this decrease has been tempered by the
continuing deposition from the atmosphere and from internal SOC recycling. Interestingly.
atmospheric inventories of these SOCs over the upper Great Lakes have not changed appreciably
since the late 1970's, suggesting a dynamic exchange between the atmosphere and the much
larger terrestrial reservoir (Baker and Eisenreich 1990, Manchester-Neesvig and Andren 1989).
Such recycling will likely result in continued atmospheric deposition of SOCs to remote
environments for some time into the future.
12
-------
TABLE 10
INPUT-OUTPUT CALCULATIONS FOR PCBs and BENZO[a]PYRENE
TO THE GREAT LAKES
(STRACHAN AND EISENREICH, 1988)
Lake
Input(kg/yr)
% Atmospheric
Output (kg/yr)
% Volatilization
PCBs
Superior
Michigan
Huron
Erie
Ontario
606
685
636
2520
2540
90
58
63
20
13
2190
7550
2760
2390
1320
86
68
75
46
53
Benzo[a]pyrene
Superior
Michigan
Huron
Erir
Ontario
72
208
290
122
155
96
86
80
79
72
314 '
6250
1370
3720
1290
19
6
31
15
33
-------
Relative Atmospheric Loadings... Revision Date: 15 March 1993
Recently, atmospheric depositional fluxes of SOCs (Leister and Baker 1993) and trace
elements (Wu el al 1992,1993, Scudlark et al. 1993) to the Chesapeake Bay were measured.
In general, fluxes of SOCs to the Chesapeake Bay are larger than those to relatively remote
regions of the Great Lakes (Figure 6), but smaller than those measured in urban areas (e.g.,
Ligocki et al. 1985a and 1985b, Holsen et al. 1991), reflecting the various types of air masses
travelling over the Chesapeake Bay (i.e., marine versus urban). It is difficult to place these
loadings in perspective because estimates of contaminant loadings from other sources to the
Chesapeake Bay are highly uncertain. Nonetheless, Leister and Baker (1993) estimate that the
atmospheric loadings of PCBs and PAHs directly to the surface of the Chesapeake Bay may be
comparable to those discharged from the Susquehanna River, the dominant tributary supplying
60% of the bay's freshwater.
Atmospheric loadings to Commencement Bay and Puget Sound have been estimated using
bulk deposition collectors (EPA, 1991). Loadings of PAHs and trace elements are significantly
greater than those measured either to the Great Lakes or to the Chesapeake Bay, often by more
than two orders of magnitude (Figure 12). These elevated loadings may reflect proximity of the
sampling sites to the many emission sources around Commencement Bay, or may have resulted
from oversampling by the bulk deposition samplers. In either case, it is unclear how to
extrapolate these high, perhaps localized, loadings to larger geographic areas (e.g. the entire
Puget Sound).
Sources and loadings of contaminants to the Massachusetts Bay system was studied by
Werme and Menzie (1991). Based largely upon a compilation of monitoring data and literature
values, they conclude that the atmosphere can be an importance source of PAHs and PCBs to
these urban waters, although the Merrmack River, North Shore, and Boston Harbor drainage
areas are the dominant source of trace metals and synthetic organics. The authors emphasize
the considerable uncertainty inherent in developing contaminant loadings inventories using data
from disparate sources.
3.3 MERCURY - Evidence of Deposition
3.3.1 The Global Mercury Cycle
To paraphrase Einstein, "Nature is not malicious, but subtle", and environmental
pollutants cycle in elusive ways. Mercury for example, has proved to be one of the most
challenging and insidious contaminants measured in the environment. There is widespread
evidence in the United States, Canada and Europe of tissue concentrations of mercury in fish
(even in pristine regions) that exceed local, national and international public health guidelines.
This situation represents a serious human health concern as well as a significant economic threat
to commercial and sport fishing industries. Anthropogenic mercury is derived principally from
coal combustion, smelting and waste incineration. Most mercury is "invisibly" transferred
14
-------
X
JD
LL
JD
CL
z
c
2
*.
O
"O
O
LL
C
M
Lewes
CBADS
GLAD Puget Snd. N. Atlantic
Figure 12. Trace element fluxes at various locations in North America
(mg/m2-year)
-------
Relative Atmospheric Loadings... Revision Date: 15 March 1993
through the atmosphere as a gas, Hg°, which is eventually oxidized, scavenged, and deposited
with precipitation. Our understanding of the environmental cycling of Hg°, though improving,
is quite limited. There is also a correspondingly important but poorly understood atmospheric
mobilization of mercury associated with particles. The atmospheric paniculate mercury cycle
may be more significant than that of Hg° in its potential to adversely affect aquatic systems on
local and regional scales. In natural waters, and as shown in Fig. 14, atmospherically derived
Hg is transformed by bacteria to a very toxic organic form, monomethyl mercury, which is
biologically amplified and concentrated in fish muscle tissue. While game fish filets (e.g.,
swordfish, tuna, mako shark, and walleye pike) often show monomethyl mercury levels greater
than 1 part per million, the concentrations in water are commonly less than 1 part per trillion.
An amplification of more than a million times has occurred. The amounts of mercury in the air
and water are extraordinarily minute such that, regardless of the highly sophisticated nature of
the equipment, the study of mercury and other metals in the atmosphere and in aqueous systems
requires ultra-clean and rigorous trace metal analytical protocols. Environmental scientists and
engineers have been slow to incorporate clean laboratory expertise into their trace metal studies.
The prominence of atmospheric mobilization and depositional processes in the global
biogeochemical cycling of mercury is well recognized and described in a variety of mass
balance formulations of the global mercury cycle (e.g., Wollast et al. 1975, NAS 1978; Slemr
et al 1981, Lindqvist and Rodhe 1985, Fitzgerald 1986, Nriagu 1989; Fitzgerald and Clarkson
1991, Lindqvist et al. 1991). Although the significance of the atmosphere was evident in early
models, environmental assessments of source strengths for natural and anthropogenic processes
were often in error because they lacked accurate information about critical aspects of the
mercury cycle. Current international human-health and environmental concerns associated with
elevated levels of monomethyl mercury in freshwater and marine piscivorous fish have focused
attention on mercury as a pollutant. Consequently, there has been an expansion in Hg research,
a heightened awareness of the need for an accurate and broader environmental data base for Hg
in the environment, and the incorporation of ultra trace-metal clean sampling and analytical
protocols into Hg research (see Evaluation of Current Sampling and Analytical Procedures
Section for further details). In addition, new information is being communicated effectively.
Two international meetings dealing with mercury in the environment have taken place recently,
with a third meeting planned for 1994. The geochemical view of the global mercury cycle has
improved significantly, and present estimates for mercury fluxes to the earth's surface and for
the mercury content of active reservoirs are converging. The agreement among recently
published budgets for the atmospheric cycling of mercury is quite satisfactory, given the
uncertainties associated with global scale estimates (Fitzgerald 1986, Nriagu 1989, Fitzgerald
and Clarkson 1991, Lindqvist et al, 1991).
Elemental mercury concentrations in the marine boundary layer decrease between the
northern and southern hemisphere over the Atlantic and Pacific Oceans (see Figure 13, for
Pacific data). This interhemispheric distributional pattern characterizes a trace atmospheric gas
whose primary sources, on a unit area basis, are continental and likely anthropogenic. Trace
16
-------
g
Ofl
bfl
ffi
O
W
O
o
H
°GoO 0%^
40S 30 20 10 0 10 20 30
Latitude
40
60 60N
Figure 13. Latitudinal distribution of total gaseous Hg (TGMjng m~3) over the Pacific
Ocean between 1980 and 1986. Adapted from Fitzgerald (1989).
-------
74
;v,^,,,:.,.^^^
saotvcd
'|f< hq (Unr«activ«)particu
A-Atnonc
t- ;> Raft linvitd proctss
^=s C^uiibrum proccu
Figure 14. The major species, fluxes, and reservoirs for the physical and biogeochet
cycling of Kg in the atmosphere and within lakes (adapted from Hudson ei
1992).
-------
Relative Atmospheric Loadings... Revision Date: IS March 1993
gas modeling of mercury yields an average tropospheric residence time of total gaseous mercury,
assumed to be Hg°, of about 1-year (Fitzgerald et al, 1981). Confirmation of this relatively long
average residence time is provided from estimates of annual mercury deposition to the earth's
surface using a steady state model for the global mercury cycle (Fitzgerald 1986, Table 11).
Thus, Hg° from both natural and anthropogenic sources can be readily mixed
intrahemispherically. Interhemispheric mixing allows northern hemispheric emissions of Hg°
to be transported to the atmosphere of the southern hemisphere. While the broad dispersion of
mercury has reduced some localized impact from human related emissions of mercury, it may
have led to the geographically large problem of elevated mercury concentrations in fresh water
and marine fish that are far removed from local sources.
The major fluxes associated with the global atmospheric Hg cycle are summarized in a
mass balance format in Table 11. The major species, fluxes, and reservoirs for the physical and
biogeochemical cycling of Hg in the atmosphere and within lakes is shown in Figure 14, which
has been adapted from the MCM Lake mercury model developed by Hudson et al. (1992).
Estimates of the total annual emission of mercury to the atmosphere range from 5 to 7.5 x 109
g/year. Fitzgerald (1986) noted that this estimate is much smaller than even values for
"preindustrial fluxes" used in many models. Further, and as summarized in Table 11,
atmospheric Hg emissions associated with contemporary human endeavors are comparable to
those from natural sources. Estimates for annual anthropogenic Hg emissions are between 2 and
4.5 x 109 g/year, which represents about 30 to 90% of the total annual mercury input to the
atmosphere. Elemental mercury evasion from the oceans and other natural waters is a significant
source of atmospheric mercury, and may account for 25% to 40% of the annual fluxes of
mercury. Marine studies demonstrate that in situ synthesis of volatile Hg, which is principally
Hg° in the mixed layer (Kim and Fitzgerald 1986, Mason and Fitzgerald 1990, 1991, 1992) and
its subsequent evasion at the water-air interface are major features of the global Hg cycle
(Fitzgerald et al. 1984, Iverfeldt 1988). Most recently, fresh water investigations by Vandal et
al., (1991) in Wisconsin and Xiao et al. (1991) in Sweden have shown a similar and important
in-lake Hg° cycle which yields significant Hg° fluxes to the atmosphere. Other natural volatile
sources of Hg such as volcanic emanations, biological mobilization, and forest fires can
contribute about 30 to 60% to the yearly emissions. The comparatively small estimate for the
fluvial flux of Hg illustrates the preeminence of the atmosphere in the transfer of mercury to the
world's oceans (Table 11).
Anthropogenic interferences within the biogeochemical cycle of Hg present perplexing
and challenging problems. We must be concerned not only with sources, chemical composition,
physical state, and direct impact of Hg compounds to natural waters, but with the post-
depositional in situ bacterial conversion of Hg species to more toxic forms, especially
monomethyl mercury. Monomethyl mercury is the principal form of mercury in fish (Westoo
1966, NAS 1978), and it is considerably more toxic than either Hg° or other mercury species.
Human exposure to methyl mercury compounds comes almost exclusively from the consumption
offish and fish products.(WHO 1976), although, in certain populations, consumption of marine
19
-------
TABLE 11
GLOBAL ATMOSPHERIC MERCURY BUDGET
Source
Deposition
Mercury Flux
(109 g/year)
5-6
6
7.5
Reference
=====
Fitzgerald, 1986
Slemretal., 1981
Lindqvistetal., 1991
Emissions
Anthropogenic
Natural
Oceanic Sources
Equatorial Pacific
Volcanic
Other Continental Sources
- Crustal Degassing
- Forest Fires
- Biological Mobilization
Fluvial Hg Input
2
3.6
4.5
2.5
3
2
0.2
0.06
0.6
1-2
0.2
Watson, 1979
Nriagu and Pacyna, 1988
Lindqvistetal., 1991
Nriagu, 1989
Lindqvist et al., 1991
Kim and Fitzgerald, 1986
Fitzgerald, 1986
Varekamp and Buseck, 1986
Fitzgerald and Clarkson, 1991
Gill and Fitzgerald, 1987
-------
35
TU. 30 H
CO
CD
D
E
D
O
o
co
DD
25 H
20 H
15H
10H
0
Modern = 12.5 + 3.27 x
Mt
D
D
D
K
±3
M
D
O
Preindustrial = 3.7 + 0.83 x
12345
Catchment area/lake area
-2 -1
Figure 15. Whole basin accumulation rates for Hg (ug m yr ) are plotted against the
terrestrial catchment area to lake area ratio. Modern rates based on the
past 10 years are indicated by the filled squares, while the preindustrial
estimates (before ca. 1850) are indicated by the filled circles (adapted
from Swain et al., 1992).
-------
Relative Atmospheric Loadings... Revision Date: IS March 199:
mammals is a significant source (Fitzgerald and Clarkson 1991). The importance of atmospheric
mercury deposition in the aquatic biogeochemistry of Hg has been demonstrated for seepage
lakes as part of The Mercury in Temperate Lakes (MTL) Program in Wisconsin (Fitzgerald «
al. 1991), and for drainage lakes in Sweden (Lindqvist et al 1991). Both studies indicate that
small increases in atmospheric depositional fluxes of mercury could result in enhanced mercury
concentrations in fish, as suggested by Fitzgerald and Watras (1989). These two investigations
will serve as a benchmark for assessing the quality of the available information about the
atmospheric cycling of mercury and for identifying parts of the cycle where information is
needed, especially as it relates to the impact of atmospheric Hg deposition to the Great Waters.
3.3.2 Regional Mercury Cycling and Localized Deposition in North America
A recent and very convincing report documenting increasing rates of atmospheric Hg
deposition to lakes in Minnesota and Wisconsin was recently published (Swain et al. 1992). They
employed an innovatively simple but effective mass-balance approach to Hg flux information
obtained from the sediment record of seven "relatively undisturbed" lakes in Minnesota and
Wisconsin. For each lake, multiple cores (7 to 15) were taken, the strata were dated and
sedimentation rates established from the 210Pb chronology. Mercury was measured, whole lake
Hg fluxes were determined, and the preindustrial and modern atmospheric Hg inputs were
inferred.
The results are presented in Figure 15 which has been adapted from Swain et al (1992).
Whole basin accumulation rates for Hg (/zg nr2 yr1) are plotted against the terrestrial catchment
area to lake area ratio. Modern mercury accumulation rates based on the past 10 years are
indicated by the filled squares, while the preindustrial estimates (before ca. 1850) are indicated
by the filled circles. The intercept of the regression line (catchment area = zero) yields a value
that represents the rate of regional atmospheric deposition to the surface of the lake. In addition,
the ratio of the intercept to the slope provides the portion of the atmospheric Hg depositional
flux to the catchment that is transported to the lake. As indicated in the regression equations
(Figure 15), the estimate for present atmospheric deposition of mercury to Minnesota and
Wisconsin is 12.5 /xg nr2 yr1, and preindustrial value is 3.7 ^g nr2 yr-1. This represents an 3.4
fold increase in deposition in recent times. These estimates of substantial anthropogenic
enhancement to the mid-continental Hg cycle at approximately 2% yr1 for the past 140 years
are consistent with predictions on a global scale (Fitzgerald and Clarkson 1991). For example,
approximately two-thirds of the total world's production of Hg has taken place during this
century, and anthropogenic releases of Hg to the environment have increased about 3-fold since
1900 (Andren and Nraigu 1979).
The depositional results for Hg as established by the Little Rock Lake budget (Section
8.3), and the whole lake experiments by Swain et al. (1992) yield a consistent estimate of
21
-------
TABLE 12
TIME INTEGRATED ESTIMATES OF MERCURY DEPOSITION, AS DETERMINED IN
PEAT FROM AN OMBROTROPHIC PORTION OF ARLBERG BOG, MINNESOTA
(BENOIT £/
-------
Relative Atmospheric Loadings... Revision Date: 15 March 1993
present day mercury deposition rates to the mid-continental U.S. Further support is provided
by Mierle (1990), who reported a mean wet depositional flux of 10.2 /xg nr2 yr"1 at Dorset,
Ontario, which is about 1000 km east of the Minnesota and Wisconsin lakes. This agreement
may be misleading and bears careful scrutiny because it suggests that the atmospheric
contributions to the mid-continental North America are almost exclusively associated with the
global Hg cycle. As noted, Hg°, the principal form of Hg in the atmosphere, has a tropospheric
residence time of the order of a year, allowing it to be widely dispersed before oxidization and
removal by precipitation or dry deposition. Primary paniculate Hg emissions will contribute
principally to local and regional deposition. Therefore, one should expect to find the influence
of continental sources of paniculate Hg in mid-continental Hg deposition. This local-regional
paniculate Hg would be superimposed on an increasing background of Hg deposition associated
with the global cycling of Hg°. Nater and Grigal (1992) have found such a pattern of regional
Hg deposition in organic litter and surface mineral soil at forested sites along a 1000 km track
from northwestern Minnesota to eastern Michigan. A summary of their results is presented in
Figure 16, while the sample locations are indicated in Figure 17. According to these authors,
"the observed gradient parallels changes in wet sulphate deposition and human activity along the
transect, suggesting that the regional variation in mercury content is due to deposition of
anthropogenic mercury, mostly in the paniculate form."
An examination of the Hg accumulation profiles versus age of the sediments from
representative cores presented in Figure 1 of the Swain et al. (1992) work suggests that the
whole lake regression analysis will not yield consistent results for other time periods (e.g., 1950
to 1960). Moreover, the number of pre-1980 peaks in the sedimentary record indicates that Hg
deposition was variable and possibly larger in the 1950s compared to estimates for the last
decade. Temporal variations and localized contributions to atmospheric Hg inputs during the
past century are quite evident in the Hg accumulation record for an ombrotrophic portion of
Arlberg Bog in northeastern Minnesota (Benoit et al. 1992a, 1992b). Arlberg bog is located in
St Louis County, Minnesota near the town of Cloquet, about 50 km west of Duluth. The
average Hg accumulation (/xg nr2 yr1) rates versus time are summarized in Table 12. Average
recent Hg deposition is 24.5 ± 7.9 /xg nr2 yr1, which is about twice the estimate from the Little
Rock Lake Study (Section 8.3) and the Swain et al. (1992) results (Figure 15). Also, and in
contrast to the Nater and Grigal (1992) investigation (Figures 16 and 17), atmospheric Hg
deposition is greater at this northeastern Minnesota location than in Wisconsin. Local/regional
scale gradients in the Hg deposition are evident in temporal deposition where the mean Hg
accumulation was approximately 38 ± 12 /xg m'2 yr'1 between 1935 and 1980, and 10 ± 3 /ig
m~2 yr1 during 1750 to 1935. Finer resolution from two peat cores indicates that the pre-1900
atmospheric deposition was ca. 4 ± 1 /xg m'2 yr1. Thus, the estimates for preindustrial Hg
fluxes from the atmosphere as determined from the Hg distributions preserved in lake sediments
and an ombrotrophic peat bog are identical. However, the estimates for recent Hg accumulation
differ, and point towards atmospheric scavenging and deposition of Hg in particles near their
emission sources.
23
-------
Backaround
Net change
Figure 16. Estimates of the net increase in Hg due to atmospheric deposition compared
to predicted geological contributions ("background") along a track betwe-en
northwestern Minnesota and eastern Michigan (adapted from Nater and Grigal,
1992). The stations appear in Figure 17.
-------
o*ag *
^ 0°
• »
0»
:
:
\^
ZONE
1
o» 5
';°
o
"*i • _
\ \
i •
:
-J1 Q
»
ZONE
2
0 ° o/ y-
_L^<
^^T C
V**
• B
% '
•
\
\.
; ^>s
• Z^DNE
: 3
" I
O.j
••••
i
Minnesota
ZONE/
Wisconsin
Michigan
Figure 17. Sampling sites in the five zones established across the Great Lake State
(adapted from Nater and Grigal, 1992).
-------
TABLE 13
ANNUAL DEPOSITION AND VOLUME-WEIGHTED CONCENTRATION
AVERAGES FOR MERCURY IN PRECIPITATION AT THREE LOCATIONS
IN MINNESOTA DURING 1988 AND 1989
(Adapted from Glass et al, 1991)
Site
Duluth
Marcell
Ely
Duluth
Marcell
Ely
Year
1988
1988
1988
1989
1989
1989
Deposition
(Mg/m2)
19.9
15.7
16.7
6.5
13.0
41.9
Concn
(ng/L)
22.6
17.7
19.7
10.5
18.0
81.4
Precipitation
Depth (cm)
88.5
88.8
84.9
62.1
72.2
51.4
Sampling
Period (Weeks)
52
48
47
52
41
52
-------
Relative Atmospheric Loadings... Revision Date: 15 March 1993
The importance of local Hg emissions and deposition are indicated in another Minnesota
study. Glass et al (1991) found a two year average flux of 19 ± 12 ng nr2 yr-1 for rain and
snow at three sites (Duluth, Marcell, and Ely) in Northern Minnesota. If we assume that an
additional 5 ^g m'2 yr-1 would be contributed by dry deposition, the results are comparable to
the Benoit et al. (1992a,b) estimates for recent total Hg deposition. This is a further indication
of the importance of local sources and site variability in the geographic pattern of atmospheric
Hg deposition. As summarized in Table 13, the interannual variations are substantial for the
Duluth and Ely sites. The results for Ely, a relatively remote location bordering the Boundary
Waters Canoe Area in the Superior National Forest, are particularly puzzling because of the
anomalously high concentrations > 100 ng Hg/L of precipitation and the corresponding elevated
depositional fluxes reported for the spring of 1989. These unusually high levels are most
probably artifacts reflecting contamination incurred during the sample collection and processing
procedures. This is a problem common to the study of Hg and other trace metals in the
environment (see for example, Patterson and Settle 1976, Fitzgerald and Watras 1989, Fitzgerald
et al. 1991). We do note that "events" of this magnitude have been observed in the Swedish
mercury depositional studies described below.
3.3.3 Localized Atmospheric Hg Deposition: Sweden
The importance of local and regional atmospheric deposition is demonstrated by the
extensive study of the tropospheric cycling of Hg over the Nordic countries (Iverfeldt 1991a,b).
The results are presented in Table 14, for locations shown in Figure 18 that were chosen to
examine south - north gradients of Hg in air and precipitation over a distance of approximately
1500 km, which is comparable to the range covered between northwestern Minnesota and eastern
Michigan in the Nater and Grigal (1992) work. In general, average total Hg in precipitation
increases from ca. 8 ng/L at the most northerly stations to ca. 40 ng/L at the most southerly site
(Figure 19). The latitudinal depositional pattern is particularly pronounced in Norway where with
the smallest Hg fluxes observed at the northernmost stations of Overbygd and Jergul (5 and 3
fj.g m"2 yr1- respectively) and the largest at Birkenes (35 ng nr2 yr1), a port on the southern tip
of Norway. The variations in annual deposition of Hg are related to the locality and annual wet
deposition. For example, Karvatn, in southeast Norway, with a large annual precipitation of
1430 mm, has a flux of 13 /xg m'2/yr, while the southern-most site in the network (Keldsnor,
Denmark), with a comparable yearly flux of 17 /xg m'2 yr1, has a annual rainfall that is a factor
of 3 lower (430 mm). The depositional pattern for the Nordic study closely resembles the
average range for the global depositional pattern of Hg as estimated from a limited data base by
Fitzgerald (1986).
The importance of regional European sources to Hg deposition in the Nordic countries
is evident. The highest levels of Hg in precipitation are associated with air mass trajectories
from the south/southeast, mostly from eastern part of Europe. Moreover, episodic effects were
evident where concentrations of total Hg in precipitation were > 100 ng/L for southerly air
trajectories and these high levels correlated with "exceptionally high concentrations" of soot and
27
-------
TABLE 14
ATMOSPHERIC WET DEPOSITIONAL FLUXES OF TOTAL
MERCURY TO VARIOUS STATIONS IN THE NORDIC COUNTRIES
After Iverfeldt, et al. (1991)
Station
Keldsnor, DK
Aspvreten, S
Rorvik, S
Birkenes, N
Tikkakoski, SB
Vindeln, S
Karvatn, N
Overbygd, N
Jergul, N
No.
1
2
3
4
5
6
7
8
9
'Precipitation
(mm)
430
520
770
1730
720
650
1430
540
340
Deposition Rate
(/zg/m2-yr)
17
10
27
35
11
7
13
5
3
-------
Figure 18. Stations in the Nordic network study of atmospheric Hg during 1985
to 1989 (from Iverfeidt et al.t 1991).
-------
60 F
40
20
i
w
I 1
I i i
123456789
123456769
3234567B9
Station (no.)
r 20
0) 15
c
U> 5
0
—' 1 I 1 I
-
- j i • {
-
— , , , , 1
* M *
, 1 1
* '—
-
-
1 i
,—
^ 4
L
O)
I2
1
o»
z
0
^1 1 1 1
•
.
: »M
b
•
-,
' ' ' ' T ' '—
•
t 1 , :-
A 1 J.
4- 1 1 "
' T
,—
igure 19
Mercury in precipitation along the Nordic Sampling Network (Sites
in Figure 18).
-------
Relative Atmospheric Loadings... Revision Date: 15 March 1993
sulfate, as well as low pH. The lead and cadmium concentrations were also about a order of
magnitude greater (Iverfeldt, 1991b). In earlier studies in Sweden, Brosset (1987) and Brosset
and Iverfeldt (1989) reported a strong correlation between Hg, soot and air trajectories from the
eastern part of Europe. Recently, Xaio et al.t 1991 reported a good correlation (r= 0.74)
between the paniculate Hg fraction and soot for air samples collected in Goteborg, Sweden.
3.4 NITROGEN - Evidence of Deposition
Nitrogen and phosphorus are the two primary nutrients required to sustain aquatic
biological production. A general ecological axiom is that primary production in freshwater
systems is phosphorus limited, while marine systems are nitrogen limited (Hecky and Kilham,
1988). In remote regions, aquatic productivity is often limited by the availability of one or both
of these nutrients. However, as the result of anthropogenic inputs, these nutrients are often
present at concentrations in gross excess of basic requirements, resulting in a deleterious
condition known as eutrophication.
Although a number of earlier works (e.g. Correll and Ford 1982, Smullen et al 1982)
recognized the ecological significance of the atmospheric deposition of nitrogen in a chemical
mass balance for aquatic systems, until recently its import was largely overlooked or
underestimated. Much of the current interest has resulted from the recent studies of Fisher et
al. (1988), Fisher and Oppenheimer (1991), and others (Tyler 1988, Hinga et al. 1991) which
focus on Chesapeake Bay. These reports concur that 25-40% of the nitrogen loading to
Chesapeake Bay is derived from atmospheric deposition (Table 15). Although these studies bring
to light the inherent uncertainties in such estimates (particularly with respect to dry deposition
and watershed loading estimates), they have underscored the importance of atmospheric
deposition and have forced a serious re-examination of eutrophication mitigation strategies in
water quality management decisions.
A summary of studies which document the atmospheric input of nitrogen to coastal waters
is provided in Table 16. The initial results of Fisher and Oppenheimer (1991) for the
Chesapeake Bay have been corroborated by several researchers using somewhat differing
approaches. For the other major east coast estuaries, the lower relative atmospheric loading of
nitrogen primarily reflects the greater degree of anthropogenic influence. For the subestuaries
listed, the atmospheric contribution ranges from 7% for Laholm Bay (impacted by heavy
agricultural nitrogen inputs) to 100% for Ochlockonee Bay (in an isolated forested watershed
which receive minimal anthropogenic inputs). For the coastal seas, the atmospheric input
includes direct deposition to the water surface, as little is quantitatively known about the fluvial
transport of nitrogen from estuaries to the shelf.
32
-------
TABLE 15
CALCULATED NITROGEN LOADINGS TO CHESAPEAKE BAY WATERSHED, 1984'
Source
Precipitation:
Nitrate
Ammonium
Animal Waste
Fertilizer
NFS Subtotal
Point Sources
Total
10* kg N.yr1
151
54
195
158
588
41
628
kg N.ha-'.yr1
9.2
5.1
11.9
9.6
35.9
2.5
38.3
% of Total
24
13
31
25
-
7
100
% of NPS
26
14
33
27
100
-
-
•Includes the bay.
From: Fisher and Oppenheimer, 1991.
-------
TABLE 16
ATMOSPHERIC INPUT OF NITROGEN TO COASTAL WATERS
Major Estuaries
Chesapeake Bay
(1)
(2)
(3)
Delaware Bay (4)
Narragansett Bay (2)
New York Bay (2) .
Long Island Sound (11)
Sub-Estuaries
Potomac River (5)
Rhode River, MD (6)
Neuse River, NC (7)
Rehoboth/Indian River Inland Bays, DE (8)
ILaholm Bay, Sweden (2)
Ochlockonee Bay, FL (2)
Coastal Seas
North Sea (9)
New York Bight (10)
Baltic Sea (12)
Percent Input
39
35
25*
14
12
10
7*
28
40
23
8 '
7
100
27**
13**
25
**,
•nitrate only
deposition to water surface only
(1) Fisher and Oppenheimer 1991
(2) Hingae/. al, 1991
(3) Tyler 1988
(4) Scudlark and Church 1993
(5) Jaworski et. al, 1992
(? Correll and Ford 1982
(7) Fisher et. al, 1988
(8) Ritter 1986
(9) Lancelots, al, 1987
(10) Sinderman and Swanson, 1979
(11) NOAA 1986
(12) Larsson et. al, 1985
-------
Relative Atmospheric Loadings... Revision Date: 15 March 1993
The largest uncertainties associated with estimating the relative role of atmospheric
deposition of nitrogen to inland and estuarine waters are associated with (1) accurately
quantifying dry depositional flux, and (2) determining the indirect atmospheric loading via
watershed runoff. The inherent uncertainties in dry flux measurements are subsequently
discussed in section 6.0. In terms of wet/dry apportionment of atmospheric deposition, a
commonly-employed approximation is that dry and wet deposition provide equal contributions
to atmospheric flux. The regional deposition models of Levy and Moxim (1987), Logan (1983),
and Sirois and Barrie (1988) indicate that dry deposition accounts for 46-63% of the total
atmospheric deposition of [NO+NO2-fHNO3+paniculate NOJ. Employing the vegetative
throughfall approach, Lovett and Lindberg (1986) gauge the dry deposition of NO3 to be 60%
of the total. For NH3 and NIV, dry deposition has been reported to comprise 30-63% of the
total deposition (Walcek and Chang 1987, Lindberg et al 1986).
Based on the watershed mass balance approach, Fisher and Oppenheimer (1991) estimate
that dry deposition accounts for 40-62% of the total atmospheric nitrogen deposition to
Chesapeake Bay. Using the same approach, Hinga et al. (1991) estimate that dry deposition
comprises 42-61 % of the total atmospheric nitrogen flux. Recently, the National Dry Deposition
Network, a component of the National Acid Precipitation Assessment Program, have initiated
inferential dry deposition measurements of gas-phase HNO3 plus paniculate NO3~ at eight sites
in the eastern U.S. Although the relative contribution of dry deposition reported for 1984-87
(30-45%) are somewhat less than the other cited studies, these measurements exclude other
potentially significant nitrogen species (such as gas-phase NO2).
Consistent with our understanding of atmospheric nitrogen emissions and reactivity, for
more remote marine areas, far removed in space and time from continental sources, the relative
contribution of nitrogen dry deposition is believed to comprise no more than 25% of the total
deposition (Duce et al. 1991). There is relatively little known about the direct transfer of gas-
phase nitrogen to water surfaces. For HN03, Lewis (1983) reports that surface waters are a
more efficient collector than a dry surrogate surface, but that the opposite is true for NH4. The
air-water gas flux of NO and NO2 is assumed to be minor since these gases are relatively
insoluble and unreactive in water. In coastal marine areas, the deposition of vapor-phase HNQ,
will also be enhanced via the selective scavenging by alkaline marine aerosols (according to the
previously elucidated reaction).
In instances where the drainage basin is large relative to the open water (for Chesapeake
Bay this ratio is approximately 15:1), the indirect loading can actually exceed the direct
deposition to surface waters. The degree of watershed retention, transformation and export of
atmospherically-deposited nitrogen depends ultimately on the land usage and geomorphology
(Walcek and Chang 1987). Processes which would dictate the degree of watershed retention
include the amount, rate and physical nature (rain vs. snow) of precipitation, plant uptake.
adsorption and accumulation in soils, removal via crop harvesting, volatilization and
denitrification in soils. Furthermore, due to in-stream utilization, only 50-80% of the nitrogen
35
-------
Relative Atmospheric Loadings... Revision Date: 15 March 1993
entering a feeder tributary from the watershed actually enters the bay proper (Tyler 1988, Hinga
et al. 1990 and references therein).
4.0 CURRENT UNDERSTANDING OF THE SPECIATION OF
CHEMICALS IN THE ATMOSPHERE AND IN PRECIPITATION
4.1 TRACE ELEMENTS - Speciation
Recent investigations in marine and freshwater systems have documented the importance
of chemical speciation in the understanding and description of trace element/biosphere
interactions. Chemical speciation in aquatic systems reflects a dynamic interaction of biological,
chemical, physical, and geological processes. This paradigm has long been applied from a
biological perspective in examining the biotransformation of nutrients (i.e., the nitrogen cycle).
As analytical methodology has advanced, analogous cycles have become evident for a variety
of trace elements as well. Evidence indicates that the chemical form or "speciation" of a trace
element is perhaps more important than the total metal concentration in dictating its biological
availability, reactivity, and ultimate toxicity. For example, considerable work has been
performed by Morel and co-workers (e.g., Morel and Hudson, 1985) which examines the
complexation and bioavailability of copper in aquatic systems. Similarly, the Cr(VI) and the
As(III) chemical oxidation states exhibit the greatest toxicity in surface waters and sediments.
Trace elements are present in the atmosphere in a variety of aerosol sizes resulting from
a variety of natural and anthropogenic sources. The most visible form of aerosols in the
stratosphere is from condensation processes during atmospheric redox reactions involving
primarily sulfuric acid or neutral sulfate (Junge 1977). The most visible form in the troposphere
is that of water vapor condensation (clouds and fog) and dust or haze. Aerosol trace elements
exist in super-micron sizes in dust and submicron sizes in haze. The log normal size distribution
results from both settling and surface dependent scavenging processes. Both the concentration
and enrichment factors (defined later) for trace metals increase with decreasing size in a log
normal fashion (Duce et al. 1976). The form of trace metal in aerosols depends critically on
their origin. In the case of stratospheric or tropospheric haze, they are often the condensation
products of volatile emissions. Clouds contain the scavenged and redox solubilized trace
elements that result from cloud scavenging and formation. In the case of dust, trace elements
are a component of the inherent or adsorbed products of resuspension processes at the earth's
surface. Thus rainfall chemistry is the super-condensed product of a primary aerosol and its
scavenging chemistry. Most non-crustal atmospheric trace elements are vaporized during high
temperature combustion, which preferentially condense on (Linton et al. 1976) or form (Smith
et al. 1979, Natusch et al 1974, Ondov et al 1979, Shendrikar et al 1983, and others) sub-
micron aerosols. However, the low boiling points of some elements (specifically mercury.
arsenic, and selenium) allows for a significant vapor phase at standard atmospheric temperatures
36
-------
Relative Atmospheric Loadings... Revision Date: 15 March 1993
and pressures.
Both arsenic (As) and selenium (Se) are greatly enriched in coal. As a result, their
global emission inventories are dominated by coal combustion from electric power generation,
non-ferrous metal smelting and manufacturing (Mosher and Duce 1987, Walsh et al. 1979b).
The work of Andren et al. (1975), Ondov et al. (1979), Shendrikar et al. (1983) and others
revealed the presence of vapor-phase elemental selenium in coal-fired power plant emissions,
presumably as Se(IV) oxide or elemental Se(0). Mosher and Duce (1983, 1987) also present
indirect evidence for the existence of vapor-phase selenium in the atmosphere, which may
represent natural dimethyl selenide emissions (Jaing et al. 1983). Similarly, Walsh et al.
(1979a) report the existence of vapor-phase As in a variety of marine and continental regimes.
However, similar measurements at Lewes, DE did not reveal any vapor-phase As (M. Kitto,
Univ. of Maryland, personal comm.). Alkylated arsenic compounds (dimethyl arsine and
trimethylarsine) have also been detected at extremely low concentrations (0.1-0.9 ng/m3) over
soil (Johnson and Braman 1975). Such forms are thought to rapidly oxidize to stable oxoacids
such as dimethlyarsenic (cacodylic) acid. However, limited measurements of methylated forms
of Se and As in precipitation have failed to detect measurable quantities (Andreae, 1980; Cutter
and Church, 1986; Scudlark and Church, 1988).
Methylated forms of As (e.g., dimethylarsenic acid) have also been detected in aquatic
environments (Andreae 1979, and others), which has been shown to be the result of biologically-
mediated reactions. However, due to their high solubilities and low vapor pressures, it is
unlikely that such forms tend to appreciably partition into the atmosphere (Andreae 1980).
Thus, based on the evidence currently available, while the presence of vapor phases of As and
Se may influence their atmospheric reactivity, transport and scavenging, they do not appear to
contribute significantly to the net depositional fluxes of these elements.
There exists a paucity of trace element speciation measurements in precipitation.
Speciation data are limited to measurements of relative particle loadings and metalloid oxidation
states. Most trace elements are dissolved in precipitation, except crustal elements under more
dusty conditions (Lim et al, 1991). It has been suggested that the As+3/As+5 and the
Se(IV)/Se(VI) ratios reflect the redox poise of the attendant air mass, as dictated by the variable
concentrations of chemical oxidants (e.g., O3, H2O2) and reductants (e.g., SO2; Scudlark and
Church, 1988; Cutter and Church, 1986). Preliminary results of zinc speciation indicates that
zinc is almost exclusively in the uncomplexed state, as might be anticipated for an acidic"
rainwater matrix (Lewis and Church, unpublished data).
4.2 SEMTVOLATILE ORGANIC CONTAMINANTS - Speciation
The transport, fate, atmospheric residence time, and removal processes for organic
compounds a^e largely due to the distribution of the semi-volatile organic chemical (SOC)
between the gas and particle phases (Bidleman 1988. Junge 1977, Ligocki et al. 1985a.b,
37
-------
Relative Atmospheric Loadings... Revision Date: 15 March 1993
Eisenreich et al. 1981, Pankow 1987, Duinker and Bouchertall 1989, Ligocki and Pankow 1989,
Foreman and Bidleman 1990). In general, gas-particle distributions are a function of the vapor
pressure of the organic compound, the ambient temperature, and the concentration, size, and
composition of atmospheric paniculate matter. Junge (1977) described gas-particle distributions
in the atmosphere using a Langmuir isotherm as:
- [Cp/(cp + c,)] = c0/(p° + cff) [2]
where is the fraction of the compound in the particle phase, cp and c, are particle and gas
phase concentrations of the SOC in the atmosphere (ng/m3), respectively, p° is the vapor
pressure of the SOC (torr), c is a constant (equal to 0.13 in Junge 1977), and 6 is the
concentration of total suspended matter in the atmosphere expressed as surface area (cmVcm3).
Junge (1977) and' others have used this equation to describe the general distribution of SOCs
between the gas and particle phase in the atmosphere as a function of increasing p° by plotting
(Cp/cg) (= log (1-<£/<£)) vs log cp° where the intercept is log (1/cT). It may be expected that c
is a function of compound class and therefore constant within that class, and representative of
the difference of the enthalpy of desorption of the compound from paniculate matter and
vaporization of the pure compound (Pankow 1987). Foreman and Bidleman (1990) found that
the distribution of n-alkanes, PCBs, PAHs, and other organochlorine pesticides followed the
relationship below for Denver aerosol:
log [A(TSP)/F] = 0.830 log p° = 7.109 [3]
whereas in previous studies, PAHs and organochlorines partitioned to paniculate matter
differently (Foreman and Bidleman, 1987, Bidleman et al. 1986).
Yamasaki et al. (1982) collected PAHs in the urban atmosphere of Tokyo using a glass
fibre filter followed by a polyurethane foam (PUF) adsorbent. They examined the gas-panicle
distributions as a function of total suspended paniculate concentrations (TSP):
lx)g K = log (A(TSP)/F) = m/T + b [4]
where A and F are the concentrations of SOC in the operationally-defined adsorbent and filter,
respectively, m and b are constant dependent on compound, and T is ambient temperature in K.
Pankow (1987) has shown that, if sampling artifacts are absent or do not affect the gas-particle
distribution, then
log cg/cp = log C/TSP + log p° [5]
where C is a temperature-dependent constant, and C/TSP = 1/C0. Thus, G/P distributions in
the atmosphere are clearly a function of TSP, a surrogate of surface area concentration, a
constant related to the difference between the enthalpy of desorption of the SOC from paniculate
38
-------
Relative Atmospheric Loadings... Revision Date: 15 March 1993
matter and the enthalpy of vaporization of the pure compound, and T, Pankow (1987) further
developed the above relationship using Langmuir sorption theory given
K = A(TSP)/F = (990*2TMRT)l/2/(NgATSpN0S0t0RTeQ1RT) [6]
where ATSP = specific surface area of aerosol (cm2/ug), N, = moles of sorption sites per cm2
of aerosol, M = molecular weight of compound, N0 = Avagadro's number, s0 = surface area
per sorption site (cm2), t,, = characteristic vibration time (10~13 to 10'12 sec), Ql = enthalpy for
surface desorption, R = universal gas constant, and T = absolute temperature (K). For a series
of similar compounds,
K = 1.6 x 104p°/NiATSPexp([{Ql-Qv}/RT] [7]
where Ns is the number of moles of sorption sites per c2 of paniculate matter surface area.
When sorption to the surface is liquid-like, then {Ql-Qv} is 0, and the equation simplifies to
K = 1.6x 104p°/NsATSPT [8]
Thus K is determined by the vapor pressure of the compound, the characteristic surface area of
the TSP, and ambient temperature. Numerous examples of the application of the relationship
of log A(TSP)/F or vs p°L or 1/T for PAHs and organochlorines are in the literature (Junge
1977, Yamaskai et al 1982, Bidleman et al. 1986, Foreman and Bidleman 1990, Hermanson
and Kites 1989, Baker and Eisenreich 1990, Bidleman 1988, Duinker and Bouchertall 1989,
Ligocki and Pankow 1989, Manchester-Neesvig and Andren 1988, Eitzer and Kites 1989).
Pankow (1987, 1988) and co-workers (e.g., McDow 1986, Hart 1990, Hart et al 1992,
Ligocki and Pankow 1989) and Bidleman (1988) and co-workers (Gotham 1990, Foreman and
Bidleman 1990, Bidleman et al. 1986) remind us that the above equations refer to the
equilibrium non-specific physical binding of the SOC to atmospheric particles. Possible
explanations for the G/P distribution not reflecting equilibrium partitioning include slow kinetics
of partitioning (Rounds and Pankow 1990), presence of a non-exchangeable fraction of the
particulate-bound compound (Pankow 1988), variability in the sorption characteristics of the
atmospheric aerosol (Ligocki and Pankow 1989), and the presence of sampling artifacts (McDow
1986, Hart 1990, Ligocki and Pankow 1989, Bidleman et al. 1986, Gotham 1990).
Non-exchangeable material in the paniculate phase would result in a plot of log A(TSP)/F
vs v° that is non-linear. Log A(TSP)/F would be lower than expected, and 0 would be higher
than expected. Compounds that may exhibit non-exchangeable behavior include PAHs. A
portion of the total aerosol PAH may be occluded within a soot or flyash particles and may not
be available to partition to the gas phase. Organochlorines such as PCBs and HCHs are not
expected to reflect this behavior. Foreman and Bidleman (1987) and Ligocki and Pankow (1989)
have witnessed negative deviations for some PAHs. Pankow (1988) has developed the
39
-------
Relative Atmospheric Loadings... Revision Date: 15 March 1993
expression <£ = [(100-x)/l+Cp°/TSP) + x]/100 where x is the fraction of non-exchangeable
material. When x = 0, the equation reverts to the Junge-Pankow relationship (equation [2]).
The direction of the deviation from the equilibrium condition will always be negative. The
magnitude of the deviation increases with increasing p° and x. Pankow (1988) has shown that
the "magnitude of the effect can be significant even when the bound fraction is as low as a few
percent; the more volatile the compound, the larger the effect." It is difficult to quantify the
effect of a non-exchangeable fraction in field data since other factors such as variable sorptive
characteristics of atmospheric paniculate matter also play a role, although Bidleman et al.
(1986), Ligocki and Pankow (1989), and Foreman and Bidleman (1990) observed that the
magnitude of log K varies little with the source of the atmospheric aerosol.
The times to sorptive equilibrium of SOC gases on atmospheric particles are largely
unknown. Rounds and Pankow (1990) adapted a radial diffusion model to the time scales for
sorptive equilibrium to both atmospheric aerosols and filter-laden particles. In general, they
concluded that SOCs having K < 0.1 mVug, the time to sorptive equilibrium is on the order
of hours or less, and probably not a problem. For particles collected on filters, the potential for
deviations from SOC equilibrium is significant, and depends on the kinetics of the sorption
process. In general, the magnitude of the artifacts depend on the duration of sample collection,
TSP, the gas phase concentration and temperature. The effect is projected to be most significant
for compounds of higher p° that achieve rapid sorptive equilibrium. In this case, the particles
on the filter will be at equilibrium with the last parcel of air, which is likely not to reflect the
average over the sampling interval. This model is not yet calibrated or verified for atmospheric
particles.
Sampling artifacts include adsorption of SOC gases onto the filter, overestimating the
particulate fraction, and "blow-off of SOC gases sorbed to particles collected on the filter,
underestimating the particle fraction. Several researchers demonstrate that blow-off during a
collection period is significant (Cautreels and Van Cauwenberghe, 1978, Van Vaeck et al. 1984)
although more recent studies suggest sorption is a bigger problem. Although there is strong
evidence to suggest that SOCs may sorb to filters in laboratory and field studies (Bidleman 1988,
Bidleman et al. 1986, Coutant et al. 1988, Ligocki and Pankow 1989, Gotham 1990, McDow
and Hutzicker 1990, Hart 1989), the filter-adsorbent combination may not exhibit the problems
actively discussed in the literature (Lane et al. 1988, Leister and Baker 1993, Kaupp and Umlauf
1992). Zhang and McMurry (1991) calculated that evaporative losses (blow-off) of adsorbed
species during atmospheric aerosol sampling is not significant for most SOCs at the pressure
drops experienced in typical organic samplers. Hart (1989) has shown that sorption of PAHs
and some organochlorines to the quartz front filter does not appreciably alter determination of
real G/P distributions. Hart (1989) and McDow and Huntzicker (1990) suggest that
volatilization losses and adsorptive gains can be minimized by reducing sampling times as much
as possible to minimize fluctuations in temperature and atmospheric concentrations. In addition,
a quartz front filter and a backup Teflon filter can be used to estimate the contribution of gas
adsorption. These recommendations have been incorporated into a new SOC hi-volume sampler
40
-------
Relative Atmospheric Loadings... Revision Date: 15 March 1993
by Hart et al. (1992). Also, many researchers now incorporate two quartz filters in series on
a routine basis to estimate field sorption of SOC gases to the filter matrix (Leister and Baker,
1993). The first filter collects the particles and sorbed gases while the second collects only
sorbed gases. It is now suspected that this approach overestimates the magnitude of the needed
correction (Hart 1989, Hart et. al., 1992). Kaupp and Umlauf (1992) used a conventional glass
fibre filter, a low pressure cascade impactor, and a electrostatic precipitator each combined with
gas adsorbents to evaluate the importance of the sampling artifacts in the field. They concluded
that the G/P distributions were better predicted by the normal filter-adsorbent combination then
from the cascade impactor because of particle aggregation in the impactor reducing the
exchangeable gas phase concentrations. In any event, they conclude that sampling artifacts for
organochlorines such as PCBs, DDT, HCHs, etc. and some PAHs are minimal when proper
protocols and media are followed. N-alkanes and other waxy natural materials may still exhibit
deviations from equilibrium behavior because they may be transported as waxy aggregates for
which the correlative physical parameter is p°,, the solid vapor pressure.
Seasonal variations in atmospheric SOC concentrations and G/P distributions have been
actively investigated (Manchester-Neesvig and Andren 1989, Hermansen and Kites 1989, Hoff
et al. 1992a) and must inevitably reflect source emissions and the effect of temperature on vapor
pressure. Hoff et al. (1992a) determined atmospheric concentrations of PCB congeners and
many organochlorine pesticides in 143 samples collected in 1988-89 in Egbert, Ontario. This
high resolution data set provides the definitive evidence of the seasonal (i.e., temperature)
control on atmospheric SOC concentrations. Hoff et al. (1992b) developed atmospheric source
functions to describe the seasonal behavior of SOCs. Not surprisingly, SOC concentrations were
correlated nicely to temperature and to p°L. The seasonal source functions have been combined
with temperature-dependent gas/particle distributions calculated using the Junge-Pankow model
to estimate annual wet' and dry loadings of numerous chemicals in the Great Lakes region
(Eisenreich and Strachan, 1992). The importance of atmospheric speciation of SOCs and the
potential applications are clearly described in this report.
Although particle size plays a critical role in atmospheric removal processes, little is
known about particle size distributions of atmospheric SOCs. Previous studies of size
distributions of trace organic compounds are summarized in Table 17. While several studies of
PAH size distributions have been reported, there is only one for PCBs (Holsen et al. 1991) and
this is only for super coarse particles. Furthermore, the time (minimum sampling time is 1 day)
and size resolution of the available data is relatively coarse. Most data have been measured with
Sierra (minimum size cut of 0.5 /*m) or Anderson (minimum size 1.1 /zm) high volume
impactors. Mass median diameters of the various SOCs reported with these samplers typically
range from 0.5 to 2 nm. Kertesz-Saringer et al. (1971) used a Casella impactor with a
minimum size cut of 0.28 /xm and found that 30 to 50% of the benzo[0]pyrene was collected on
the glass fibre after filter (<0.28 urn). Miquel and Friedlander (1978) used a Hering low
pressure impactor (minimum size cut of 0.06 jum) and found that 75 to 85% of the
benzo[a]pyrene and coronene in Los Angeles aerosols is associated with particles smaller than
41
-------
TABLE 17
SIZE DISTRIBUTION OF MEASUREMENTS OF ORGANIC COMPOUNDS
t'heinicnl
PA Ms
Uenzo( fijpyrene
Henzo[|A|nuro»nlhene
PAHs
Denzo|«i|pyrene
Coronene
Ether Extractable
Oiganics
PAHs
Aliphatic and Carboxylic
Acids
PAHs
PCBs
Air Volume
(m1)
28000
200
1631
800-1600
4.3
2000
2500-40000
800-2000
12000-18000
Sampling Time
(Days)
15
3 to 30
2
1 to 2
3
I to 2
7 to 35
1
5.5 to 8.3
Sampler
Horizontal
Elutriator
Casella
Impactor
Anderson Hivol
Impactor
Anderson Hivol
Impactor
Hering Low
Pressure Impactor
Sub 2~pm; Hivol
Anderson and
Sierra
Hivol Impactors
Anderson Hivol
Impactor
Noll Rotary
Impactor
Substrates
OFF
Stages: Glass
AF:GFF
Stages: OFF
AF: OFF
Stages: OFF
AF: GFF
Stages: Quartz
GFF
Stages: GFF
AF: GFF
Stages: GFF
AF: GFF
Mylar Strips;
Apiezon L
Reference
Demaio and Comz (1966)
Kertesz-Saringer et al. (1971)
Albagli et al. (1974)
Pierce and Katz (1975)
Miquel and Friedlander (1978)
Ketserides and Eiclimann (1978)
Van Vaek et al. (1979)
Van Vaeck and Van Cauwenbergle
(1980)
Van Vaeck et al. (1979)
Van Vaeck and Van Cauwenbergle
(1985)
Katz and Cohen (1980)
Holsenetal. (1991)
-------
Relative Atmospheric Loadings... Revision Date: IS March 1993
0.26 urn. However, this impactor operates at pressures significantly below atmospheric, raising
the possibility that volatilization of sorbed PAHs may occur during sampling. Thus for SOC
data reported to date, mean particle sizes tend to decrease with decreasing impactor cut point.
Measurements of SOC size distributions should be made with impactors having size cuts as small
as possible. The state-of-the-art for lower size cuts is about 0.05 pm (e.g., MOUDI, Marple
etal. 1991).
The data of Holsen et al (1991) are interesting in that high PCB concentrations were
measured in particles of diameter 2 to 20 pm having estimated deposition velocities of about 4
to 6 cm/sec. If this phenomenon is a general one, then dry deposition of large particles emitted
in urban/industrial centers and containing high concentrations of SOCs could dominate
atmospheric fluxes to nearby coastal areas (e.g., Lake Michigan near Chicago, IL; Chesapeake
Bay near Baltimore, MD; Lake Ontario near Toronto-Hamilton; Long Island Sound near the
New York City and New Jersey metropolitan areas.
Most of the measurements in Table 17 were made with Sierra and Anderson high volume
impactors. Both of these instruments use glass fibre filters as sample collection substrates on
the impactor stages. This leads to two concerns. First, because the impactor jets can penetrate
into the OFF, particles that are too small to be collected inertially may be collected by filtration
on the filters (Willeke, 1975). Thus a portion of the SOCs that are reportedly associated with
coarse particles may, in fact, be from small particles. Secondly, sampling artifacts with glass
fibre or quartz filters have been raised. Grosjean (1983) reported that when PAHs were
collected on both glass and Teflon filters, the glass/Teflon ratio ranged from 0.25 to 0.8.
Furthermore, some organic gases may adsorb on to glass fibre filters (McDow 1986; Hart 1989).
Substrates such as Teflon foil are less likely to lead to such artifacts. More recent examinations
suggest that the gas adsorption artifact using glass fibre filters may not be as large as first
hypothesized for PAHs and organochlorines, somewhat greater for N-alkanes (McDow 1986;
Hart 1989; Ligocki and Pankow 1989; Foreman and Bidleman 1991).
4.3 MERCURY SPECIATION IN THE ATMOSPHERE
Most of the Hg species in the troposphere are in the vapor phase (Braman and Johnson
1974, Fitzgerald and Gill 1979, Fitzgerald et al. 1981, Slemr et al. 1981), and consist almost
entirely of elemental mercury (Hg°), as demonstrated by Kim and Fitzgerald (1986). Improved
trapping, separation, and detection procedures developed by Bloom and Fitzgerald (1988) have
refined our understanding of the partitioning of the vapor phase. These authors showed that Hg°
accounts for 95 to 100% of the total vapor phase concentration in a coastal/urban location on
Long Island Sound. The remainder species of the vapor phase concentration was monomethyl
mercury (MMHg). Greater than 99% of the total mercury present in the near surface marine
atmosphere is Hg° (Mason etal. 1992). Moreover, recent studies in mid-continental northcentral
Wisconsin are showing a similar partitioning with the Hg° fraction generally > 99%.
43
-------
Relative Atmospheric Loadings... Revision Date: 15 March 1993
Monomethyl mercury is much more soluble in water than Hg°, and small quantities are present
in precipitation. The source of MMHg in the atmosphere is not known.
4.3.1 Mercury Speciation in Precipitation.
An average scavenging ratio (i.e. concentration in rain/concentration in air) observed for
mercury in rain collected in Wisconsin was 437 (Fitzgerald et al. 1991), and this value is
comparable to scavenging ratios found for metals such as lead (Maring et al. 1989, Church et
al. 1991). Although there are limitations to the scavenging ratio approach, it appears that
atmospheric Hg wet deposition is analogous to other trace metals (i.e., Pb, Cu, Zn) which exist
as particles in the atmosphere (Maring et al. 1989, Buat-Menard 1985). Similar conclusions
regarding the prominence of the atmospheric paniculate Hg cycle in conveying Hg to natural
waters were reached in Swedish work (Brosset 1987, Lindqvist et al. 1991, Iverfeldt-1991a).
The influence of paniculate Hg on the composition of Hg in precipitation is shown in Table 18
from Iverfeldt (1991a). He found an on average, > 67% and > 79% of the total mercury in
precipitation was filterable at the 0.4 /xm size range for two sites, Overbygd, Norway and
Keldsnor, Denmark, respectively, in the Nordic Countries Network. The results also show that
Hg is associated with large particles (> lOtim) as well as particles between 0.4 /xm and 10 ^im
size classes.
Mercury speciation in precipitation from the MTL Wisconsin studies (Fitzgerald et al.
1991, 1993) is shown in Table 19, where measurements of total mercury (HgT), reactive
mercury (Hg,0 and monomethyl mercury in wet deposition (snow and rain) from these mid-
continental rural temperate lacustrine environs are summarized for a two year period (1989 and
1990). Details of the Hg speciation for rain are presented in Figure 20. Several general features
are evident in the broad geochemical view provided by the average speciation results in Table
19. Firstly, the average HgT was similar for both years (52.5 ± 24.0 and 49.3 ± 20.8 pM)
while the average HgR was higher during 1990 (41.0 ± 20.6 pM) than during 1989 (13.7 ±
10.6 pM). As noted by Fitzgerald, et al., (1991), the difference between HgT and HgRis not
due to MMHg, which is present in very small quantities (< ca. 1 pM; see Figure 20). Rather,
the difference is due principally to strong Hg associations with organic substances that are
destroyed by the powerful oxidant (BrCl) used in the determination of HgT (see Analytical
section for details). The authors also suggest that this strongly bound Hg fraction is associated
with atmospheric particulates containing organics which may have a significant sulfur content.
This interpretation is part of a more general atmospheric paniculate Hg scavenging hypothesis
emphasizing the influence of atmospheric paniculate Hg on the composition in rain. Support
for this postulate comes from the scavenging ratio estimates, and the Iverfeldt (199la) work
(Table 18) showing that the filterable or colloidal species dominate the HgT in rain.
Predicted HgT concentrations in rain based solely on the scavenging of atmospheric
paniculate Hg are summarized in Table 20. Recent data from the equatorial Pacific are also
44
-------
TABLE 18
INFLUENCE OF PARTICULATE MERCURY ON THE COMPOSITION
OF H, IN PRECIPITATION (AFTER IVERFELDT, 1991)
Station
Keldsnor, Denmark
Overbygd, Norway
Avg
64
42
> 10 fim
Range
8-93
8-64
SD
22
20
n
12
6
Avg
MMMMMMM
79
67
>0.4/im
Range
•
49-92
24-92
SD
a^—
12
27
n
=— —
10
5
-------
TABLE 19
SUMMARY OF THE AVERAGE CONCENTRATION OF Hg SPECIES OBSERVED
IN WET DEPOSITION FROM NORTHCENTRAL WISCONSIN
Sampling Period
Rain 1989
Rain 1990
Snow 89/89
Snow 89/90
N'
12
9
6
3
HgR (PM)
13.7 ± 10.6
41,0 ± 20.6
17.5 ± 12
8.0 ± 0.75
Total (pM)
52.5 ± 24.0
49.3 ± 20.8
30.0 ± 4.5
14.9 ± 3.9
MMHg (pM)
0.78 ± 0.34
0.37 ± 0.16
0.24 ±0.11
0.52 ± 0.20
*N = number of deposition events sampled
-------
TABLE 20
CALCULATED CONCENTRATION OF Hg IN RAIN BASED ON A SCAVENGING RATIO,
W = 600 (RANGE 200-1200) AND USING THE FORMULA W = Crain (pg/kg) X 1.2 kg/m3
* C* (pg/m3). THE AVERAGE VALUES WERE CALCULATED USING AN AVERAGE
SCAVENGING RATIO (W) OF 600 WHILE THE VARIABILITY WAS ESTIMATED USING
A RANGE FOR W OF 200 OR 1200 AND THE ACTUAL PARTICULATE
CONCENTRATION EXTREMES FOUND AT THESE SITES. THE VALUES FOR W WERE
TAKEN FROM THE DATA FOR LEAD REPORTED BY MARING et al (1989). TABLE
ADAPTED FROM MASON et al. 1992
Region
Pacific
Wisconsin 89*
Wisconsin 90*
Particle Concentration
pg/m3
Range
0.2-6
5-62
7-77
Ave.
2.6
22
37
Rain Concentration
Calculated Measured
Range
0.2-30
4-310
6-385
Avg.
6
55
93
Avg.
14
51
49
The average particulate concentration and range (all months) found in Wisconsin during 1989
and 1990.
-------
o
h-
o:
LJ
O
O
o
i o -
14-
12-
10-
8-
6-
4-
2-
n
•
•
i
_
m
n
r
] MMHg mm TOTAL ESS REACTIVE
m
I
\
:1
1n
I
n
i
n
0
*
3
»
i
fi
n
n
:
•
n
•
•
•
•
_
mt
1.6
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0
22 30 18 13 19 22 15 16 7 17 19 27 9 27 5
JU JY AG JU • JY AG SP
1989 1990
COLLECTION DATE
Figure 20. Total, reactive and methylmercury in rain collected in Wisconsin at
Little Rock Lake Reference Basin, in 1989 and Max Lake in 1990,
adapted from Fitzgerald et al., 1992 and Mason et al., 1991.
-------
Relative Atmospheric Loadings... Revision Date: 15 March 1993
included to illustrate and emphasize the integral role paniculate Hg plays in determining the
composition of Hg in rain. As discussed in Mason et al., (1992) the differences in the Hg
content between the equatorial Pacific and mid-continental rains correspond to the atmospheric
paniculate Hg distribution between these regions. For example, during the studies, the average
paniculate Hg concentration was 2.6 pg/m3 in the equatorial Pacific and 22 pg/m3 and 37 pg/m3
in Wisconsin in 1989 and 1990, respectively. As illustrated in Table 20, estimates of expected
rain concentrations due solely to paniculate Hg scavenging are comparable to the observations
at both locations. These predictions suggest that the variation in HgR and HgT between regions
and the temporal differences at Wisconsin are a function of the differences in paniculate Hg
composition and burden. Also, as suggested earlier (Fogg and Fitzgerald 1979; Fitzgerald et
al. 1983), this depositional behavior indicates that while most of the Hg in the atmosphere is Hg°
(> 95%), it is not oxidized and solubilized in processes leading directly to the formation of
precipitation. A more general gas-to-particle atmospheric oxidation processes is inferred.
Further, Fitzgerald et al. (1991, 1993) suggest that HgR found in precipitation and atmospheric
paniculate matter is derived from the atmospheric oxidation products of Hg° in the atmosphe ..
This form of Hg is labile and highly reactive in aqueous systems and readily available, for
example, to participate in competitive reactions associated with methylation, reduction to Hg°,
uptake by biota, and sequestering with humics. The other fraction of the HgT in deposition is
the operationally defined strongly bound Hg portion ("unreactive" Hg ), which raises complex
biogeochemical questions as well. For example, are these particles environmentally active?
This fraction is likely to be associated with soot and may be strongly bound or sequestered in
some type of sulfur-carbon association (Brosset 1987). Perhaps, this unreactive Hg can be
solubilized under anoxic and/or sulfitic conditions to yield a species such as Hg(HS)20 which can
be bacterially methylated. In lakes, coastal waters, rivers and estuaries, this process could take
place at the sediment-water interface as well as in the low oxygen waters of the water column.
This would be an insidious process where an apparently unreactive component under oxic
conditions would yield MMHg in low oxygen zones of natural waters.
There is a consensus that the principal source of HgR in precipitation is the oxidation and
dissolution of atmospheric Hg°, and two atmospheric reaction pathways have been postulated.
The first argues that HgR is derived from a generalized atmospheric oxidation of Hg°, using
oxidants such as 03 or OH radicals. This reaction may occur heterogeneously and involve Hg°
adsorbed to particles (Fitzgerald et al. 1991, 1992). Such a particle conversion hypothesis is
supported by Hg(II) washout calculations, and from the direct physical and chemical analysis of
rain showing most of the Hg is associated with particles (Iverfeldt et al. 1991a,b, Fitzgerald ei
al. 1992, Mason et al. 1992). Alternatively, the work of Iverfeldt and Lindqvist (1986) suggests
that O3 oxidation of Hg° in clouds could be an important mechanism contributing to HgR in rain.
The authors predicted an oxidation rate of 0.01 hr1 (88 yr1) for conversion of Hg° to Hg(II) in
clouds, assuming 1 g nr3 of liquid water and 23 ppb O3 (a value similar to observed background
concentrations). In the absence of O3, the reaction rate is three orders of magnitude slower.
However, a residence time of Hg in the atmosphere of approximately 1 yr (Fitzgerald, 1989)
based on the global cycle yields an overall conversion rate of approximately 1 yr1, assuming all
49
-------
Relative Atmospheric Loadings... Revision Date: 15 March 1993
the Hg in rain is derived from oxidation of Hg°. This calculation represents a maximum
conversion rate. Therefore, the reaction investigated by Iverfeldt and Lindqvist (1986), which
has a substantially larger rate constant, should not be a predominant mechanism for the oxidation
of Hg° in the atmosphere if the reaction occurs at rates comparable to those found in the
laboratory. Munthe and Lindqvist (1989) and Munthe (1991) modified this model by suggesting
that rapid sulfite (SO3=) complexation of Hg2* in cloud water would yield [Hg(SOj)2=] with
subsequent reduction of Kg2* to Hg°, thereby serving as a potential reverse reaction which limits
the net amount of Hg° solubilized. However, insufficient amounts of atmospheric sulfur species
are available in the atmosphere over most of the earth's surface (i.e., oceans). Thus, alternative
gas to particle conversion processes such as suggested in the other hypothesis must be providing
pathways for formation of the HgR compounds found in rain.
4.4 NITROGEN - Speciation
The primary component of the earth's atmosphere is unreactive N2 gas. During high
temperature combustion processes, such as the combustion of fossil fuels in transportation and
utility sources, atmospheric nitrogen reacts to form oxides of nitrogen (NO, = NO + N02).
Under varying conditions, the nitrogen oxides can further react via a complex series of
photochemical pathways to form a suite of reactive nitrogen species. Although it has been
demonstrated that these compounds can be produced from natural sources as well (e.g. , NO
production by lightning and down mixing of stratospheric NO), anthropogenic emissions are now
known to dominate the tropospheric nitrogen budget for eastern North America (Singh 1987,
Logan 1983, Galloway and Whelpdale 1980). In fact, the U.S. EPA (1982) estimates that more
than 80% of the North American NOX emissions are from industrial sources.
In contrast with nitrogen oxides, the dominant source of ammonia (NH3) emissions
appears to be biogenic, in particular those associated agricultural practices (e.g. , the
decomposition of animal excrement and fertilizer production/application, NRC 1979, Apsimon
et al. 1987). However, there is still some uncertainty about the specific sources of NH5
emissions. Not only is NH3 important as a major source of atmospherically deposited nitrogen,
but also its atmospheric reactivity greatly influences the formation and deposition of other
nitrogen species (for example, the formation of particulate NI^NOa aerosol).
Anthropogenic NO is rapidly oxidized to NO2, and ultimately to gaseous HNO3 by
photochemically-catalyzed gas-phase reactions. Thus, NO, and NO2 tend to be deposited
(primarily by dry removal processes) fairly close to their source, while vapor-phase HNOs is
more inclined to be removed (via wet and dry processes) at a greater distance from its source.
Both NH3 and HNO3 are readily adsorbed onto surfaces, particularly water. For example.
Lovett and Lindberg (1986) estimate that as much as 75% of the total nitrogen deposition to a
forested watershed was in the form of HNO3 vapor.
50
-------
Relative Atmospheric Loadings... Revision Date: 15 March 1993
Vapor-phase HNO3 can also react with NH3 to form paniculate NH4NO3 . In the marine
environment, vapor-phase HNO3 is also selectively scavenged by alkaline sea-salt aerosols via
the generalized reaction (Junge 1956, Brimblecombe and Clegg 1988, Savioe and Prospero 1982,
Memane and Mehler 1987, and many others):
HN03(g) + NaClw ---> NaN03(f) + HCl^ [9]
It has also been shown that gaseous N02 is similarly scavenged by sea-salt aerosols, particularly
in urban environments, according to the generalized reaction (Finlayson-Pitts 1983):
2NO2(l) -I- NaCl(i) — > ClNOtt) + NaNO3(I) [10]
From an ecological perspective, the primary species of atmospherically-derived nitrogen
are nitric acid vapor (HNO3), paniculate nitrate (NO3"), ammonia (NH3), paniculate ammonium
(N1V), nitric oxide (NO), and nitrogen dioxide (NOj) (Hanson and Lindberg 1991). Less
information exists on the deposition of other potentially-important reactive nitrogen species, such
as nitrous acid (HNOj), dinitrogen pentoxide (N2O5), or peroxyacetyl nitrate (PAN). Recent data
for the wet deposition of dissolved organic nitrogen (DON) suggest this class of compounds may
also contribute significantly to nitrogen loading in aquatic ecosystems.
5.0 CURRENT UNDERSTANDING OF WET DEPOSITIONAL
PROCESSES
5.1 GAS SCAVENGING BY PRECIPITATION
Semivolatile organic contaminants (SOCs) are incorporated into water droplets in the
atmosphere by a variety of processes. Gas phase SOCs partition across the droplet surface and
become dissolved in the bulk liquid. At equilibrium, the magnitude of this partitioning is
described by the SOC's Henry's Law constant, which may be approximated as the ratio of its
subcooled liquid vapor pressure to it aqueous solubility at the ambient temperature:
(W- [Q, / Hifn] x MW, x 106 [11]
where Ciippugll is the concentration of nonreactive species / in precipitation resulting from gas
dissolution (ng/L), Q., is the concentration of i in the gas phase (Pa), H^ is the Henry's Law
constant of / at the appropriate temperature (Pa-m3/mole), and MW; is the molecular weight of
/ (g/mole). Raindrops are well-mixed due to the turbulence induced by falling through the air
column, minimizing local within-drop concentration gradients. Because mass transfer
coefficients of nonreactive organic chemicals across the air-water interface are reasonably large
51
-------
Relative Atmospheric Loadings... Revision Date: 15 March 1993
(e.g., IQr11 mol/[cm2-s-Pa] for PCB congeners), it is generally believed that dissolved
concentrations of nonreactive volatile chemicals in falling raindrops are in equilibrium with the
corresponding gaseous concentrations in surrounding atmosphere (Scott 1981).
Gaseous species which react in the aqueous phase are much more efficiently scavenged
by precipitation, as the aqueous reactions deplete the dissolved species concentration and
maintain the diffusive gradients. For example, ammonia is efficiently scavenged by acidic
precipitation due to the rapid protonated of ammonia to ammonium. There is some evidence in
the literature (e.g., Schomburg et al., 1991) that surface-active organic matter may sequester
hydrophobic organic contaminants within rain droplets, effectively maintaining the gas-aqueous
concentration gradients across the droplet surface. The scavenging of reactive species is more
appropriately modeled as partitioning into an infinite sink.
5.2 AEROSOL SCAVENGING BY PRECIPITATION
Paniculate metals and aerosol-bound SOCs can be entrained in precipitation by both
below- and in-cloud scavenging mechanisms. Depending on the species, the ambient
precipitation pH, and the chemical properties of the scavenged parent aerosol, particulate phases
are subsequently solubilized to varying degrees before, during, or after deposition. From the
perspective of evaluating atmospheric deposition, the wet flux of trace elements and SOCs
includes both dissolved and particulate forms. However, from an ecological perspective, the
dissolved component is probably of greatest interest as it is the most readily available. Based
on a review of recent data reported at various world-wide locations (Nguyen et al. 1990,
Nurnberg et al. 1984, Gatz and Chu 1986, Lim and Jickells 1990, Scudlark and Church 1993),
the phase distribution of trace metals in precipitation appears to be highly variable. Generally.
mineral aerosols such as Al are less soluble at ambient pH levels than high temperature
combustion condensates such as lead. Often, the speciation of trace elements are altered during
sample collection and subsequent acidification.
There is supporting evidence in the literature that the removal of contaminants from the
atmosphere by precipitation is dependent on the size distribution of atmospheric aerosols. Slinn
et al. (1978) derived a semi-empirical relationship between the "collision efficiency" of raindrops
and aerosol size. The efficiency of rain to remove particles is predicted to be lowest between
0.1 and 1 /xm and increases with decreasing raindrop size. Doskey and Andren (1981)
calculated aerosol washout ratios for PCBs for a range of aerosol diameters using the
relationship described by Slinn et al. (1978) and indicate that PCBs are most likely enhanced in
submicron aerosol size fractions. Scavenging ratios for several trace elements have also been
shown to depend upon aerosol size (Tschiersch et al. 1989).
52
-------
Relative Atmospheric Loadings... Revision Date: 15 March 1993
5.3 FIELD VERIFICATION OF SCAVENGING MECHANISMS
The above description of scavenging by precipitation has been evaluated by coincident
collection of air and precipitation samples by several investigators at several locations. Ligocki
et ai (1985a) reported gas phase scavenging coefficients for a variety of SOCs measured in
Portland, OR. Values of the gas phase scavenging coefficients measured at 8°C were three to
six times higher than those predicted using Henry's Law constants at 25°C. By correcting these
constants to the ambient temperature, they demonstrated that gas and dissolved phases was near
equilibrium for several PAHs and other low molecular weight SOCs. Field determined overall
(gas + aerosol) scavenging coefficients for many SOCs with vapor pressures less than 10~5 ton-
are often substantially larger than those attributable to gas scavenging along, suggesting that
aerosol scavenging is an important process. Scott (1981) suggests that in-cloud scavenging my
result in aerosol scavenging coefficients on the order of 106. Comparable values for below-cloud
scavenging of trace elements range from 103 to 10s (Slinn et al 1978, Slinn 1983, Talbot and
Andren 1983).
More recently, Leister and Baker (1993) have investigated SOC scavenging in the
Chesapeake Bay region. Integrated wet-only precipitation samples and air samples were
collected at a station adjacent to the bay since July 1990, using a large volume precipitation
collector which isolates 'particulate' and 'dissolved' SOCs (as operationally defined by filtration;
Baker et al., 1986). In that study, SOC scavenging by precipitation was highly variable, with
both the concentration and physicochemical speciation of the SOCs varying substantially between
sampling periods (Figure 21). For example, polycyclic aromatic hydrocarbons concentrations
in precipitation collected between 26 June and 10 July 1990 are more than ten times greater than
those in samples collected in August. Interestingly, PAHs in precipitation collected between 9
and 15 August 1990 existed primarily in the dissolved phase (as determined by in situ filtration)
while those collected between 15 and 28 August 1990 were more evenly distributed between
dissolved and particulate phases. We interpret this as direct evidence that both the magnitudes
and the relative importance of various scavenging mechanisms are highly variable. Clearly, the
use of a single set of scavenging coefficients to describe the transport for each organic chemical
is in error.
5.4 EVIDENCE FROM FOG WATER STUDIES OF ALTERNATE WET
SCAVENGING MECHANISMS
Recent studies of fog water have provided intriguing evidence that SOC scavenging by
water droplets in the atmosphere is controlled by processes other than gas dissolution and aerosol
scavenging. Glotfelty and co-workers (Glotfelty et al. 1987, Glotfelty et al. 1990, Schomburg
et al. 1991) measured concentrations of several agrichemicals in fog water and in the interstitial
air. The observed gas/dissolved concentration ratios were often much greater than their
53
-------
100
60
20
to
e
o
(—*
WM
O
u
10
•2 6
10
July 26 - July 10, 199(
August 9 - August 15, 1990
August 15 - August 28, 1990
dissolved
particulate
FLU ANT PYR CHR B[k]F I[123]P B[ghi]P
PHE FLA B[a]A B[b]F B[a]P D[ah]A
Figure 21. PAH speciation in Chesapeake Bay rainfall
(Leister and Baker, 1992)
-------
Relative Atmospheric Loadings... Revision Date: 15 March 1993
corresponding Henry's Law constants, suggesting that the fog water was significantly
supersaturated with these chemicals with respect to the gas phase concentrations. The extent of
apparent supersaturation was greater for those compounds with larger Henry's Law constants.
Glotfelty et al (1987) and, later, Schomberg et al (1991) hypothesized that this enrichment of
SOCs in fog water resulted from its complexation by surface-active natural organic matter which
was either adsorbed onto the droplet surfaces or was present as colloids within the droplets.
Further evidence of the role of natural and anthropogenic organic matter in the enrichment of
SOCs is presented in the collaborative studies of fog events in Dubendorf, Switzerland (Czuczwa
et al 1988, Leuenberger et al. 1988, Capel et al. 1990, Capel et al 1991). As with the earlier
studies by Glotfelty and coworkers, these investigators measured elevated levels of SOCs in
Dubendorf fogwater in far excess of calculated equilibrium values with the interstitial air. Capel
et al (1990) reported elevated concentrations of dissolved organic carbon (DOC) ranging from
31 to 260 mg C/L, resulting in significantly lower surface tensions in these urban fog waters.
High levels of DOC in surface waters are thought to complex hydrophobic SOCs, lowering their
aqueous phase fugacities. Assuming that these fogs scavenge SOCs by the same mechanisms
as those operating in clouds, these data suggest that natural organic matter in precipitation may
play a substantial, if not controlling, role in scavenging SOCs from the atmosphere.
The field measurements presented above strongly indicate that SOC scavenging by
precipitation involves more than the simple Henry's Law-type dissolution of SOC gases and
aerosol scavenging. Clearly, both precipitation and fogwater appear to be supersaturated with
SOCs relative to the surrounding air, suggesting either that mass transfer out of the droplets is
hindered or that alternate mechanisms result in complexed SOCS within the droplets.
6.0 CURRENT UNDERSTANDING OF DRY DEPOSITIONAL
PROCESSES
6.1. DRY AEROSOL DEPOSITION
6.1.1. Concepts and Models
Dry aerosol deposition results from the transport and accumulation of aerosol-associated
contaminants during periods without precipitation. The theoretical basis of dry aerosol
deposition has been discussed extensively in the literature, and the reader should consult the
reviews by Davidson and Wu (1988; 1989), Hicks et al (1980; 1986), McMahon and Denison
(1979), Sehmel (1980), and Hosker and Lindberg (1982). In general, the magnitude of the dry
aerosol contaminant flux is related to the concentration of aerosol-associated contaminant in the
air mass. This relationship between concentration and flux is quite complex and non-linear.
however, and depends upon characteristics of the atmosphere (e.g., physical stability), upon the
nature of the receptor surface (e.g., tree canopy versus water), and upon the properties of the
55
-------
Relative Atmospheric Loadings... Revision Date: 15 March 1993
depositing contaminant (e.g., reactivity, aerosol size).
The stability of the atmosphere determines the amount of turbulence and, therefore, the
magnitude of aerosol transport in the vertical direction. Near the Earth's surface, a momentum
boundary layer exist where the air velocity increases from zero within millimeters of the surface
to some constant value with height. A quasi-laminar sublayer of air comprises this millimeter
thick layer adjacent to receptor surfaces, where turbulent transport is greatly reduced. A surface
layer extends from several meters to tens of meters above the receptor surface. Within the
surface layer, vertical turbulent fluxes of momentum and heat are constant with height. Above
the surface layer, the planetary boundary layer include the entire region of the atmosphere in
which transport is influenced by interactions with the Earth's surface.
Dry aerosol deposition is modeled as a three step transport processes, in which aerosol
are carried from the planetary boundary and surface layers through the quasi-laminar layer and
are allowed to interact with the receptor surface. Aerosol transport from the planetary boundary
and surface layers results from eddy diffusion and sedimentation. The magnitude of eddy
diffusion depends upon the amount of turbulence which depends, in turn, on the atmospheric
stability. Heat exchange vertically in the atmosphere either contributes to or suppresses
turbulent energy. In the case where the air at ground level is colder that than aloft, the air
column is stable, turbulence is low, and aerosol deposition by eddy diffusion is limited.
Conversely, when the air at ground level is heated, the air column becomes unstable and the
increase in turbulence enhances aerosol transport rates. Heating of air at the Earth's surface
may result either from solar heating or from the seasonal cooling of large bodies of water.
Aerosol particles larger than one micron may be transported from the planetary boundary
and surface layers by sedimentation. Gravitational attraction accelerates aerosol particles
towards the Earth's surface and particles increase in velocity until drag forces offset the
gravitational forces, as described by Stoke's Law. The magnitude of this terminal settling
velocity (i.e., the steady-state velocity when drag and gravitational forces are equal) depends
upon the size, shape, and density of the aerosols. Because the terminal settling velocity
increases with the square of the aerosol diameter, gravitational settling is an important
component of dry aerosol depositional fluxes for those contaminants which are associated with
large aerosols.
Once transported to the quasi-laminar sublayer, aerosols move through this sublayer to
the receptor surface by a variety of processes, including eddy diffusion, interception, inertial
motion, and sedimentation (Davidson and Wu, 1988). Electrostatic forces, thermophoresis, and
diffusiophoresis may also assist transport in the sublayer. Aerosols reaching the receptor surface
may adhere to or bounce off the surface, depending on the characteristics of both the surface and
the aerosol. In addition, specific aerosols or their associated contaminants may react chemically
with the receptor surface, maintaining a large concentration gradient within the quasi-laminar
sublayer.
56
-------
Relative Atmospheric Loadings... Revision Date: 15 March 1993
Dry aerosol depositional fluxes have been modeled using a temporally- and spatially-
variable dry deposition velocity (vdAM length/time):
F = CJxv(U,, [12J
where F is the dry aerosol depositional flux (mass/area-time) and Q, is the concentration of
aerosol (or aerosol-associated contaminant) at height z and time / (mass/volume). The overall
resistance to transport of aerosols is equal to the reciprocal of the dry deposition velocity. This
resistance may be conceptually divided into component resistances resulting from transport in
the planetary boundary, surface, and quasi-laminar sublayer, and from interactions at the
receptor surface. For most contaminants, the resistance to transport in the planetary boundary
and surface layers is quite small relative to those in the sub-layer and at the surface, resulting
in large concentration gradients near the Earth's surface. Equations for aerosol deposition
velocities derived from the flux of air momentum, boundary layer transport, and interactions
with the surface are reviewed by Davidson and Wu (1988).
Aerosol deposition to water surfaces are influenced by the specific properties of the air-
water interface. Of particular importance is the exchange of momentum, heat, and water vapor,
the potential increase in aerosol size as particles incorporate water, and the role of aerosols
produced via sea salt ejection and bubble breaking. Slinn and Slinn (1980, 1981) presented a
two layer model of particle deposition to water which included transport through a constant flux
layer above the surface as well as transport through the surficial boundary layer. Their model
included the effect of slip at the water surface and of the growth of hygroscopic aerosols in the
high humidity environment at the air-water interface. The Slinn and Slinn model assumes that
the surface is a perfect sink (i.e., that there is no surface resistance). Fairall and Larsen (1984)
expanded their earlier model to include the influence of sea spray production and atmosphere
surface interactions. Williams (1982) accounted for the breaking of waves and the resulting
increase in sea surface roughness.
6.1.2. Field Measurements
Methodologies to estimate dry aerosol depositional fluxes have been reviewed previously,
including at the National Acid Precipitation Assessment Program (NAPAP) Dry Deposition
Workshop in March, 1986, as summarized by Hicks et al. (1986). Dry aerosol depositional
fluxes are measured directly using a variety of surface analysis methods, in which rates of
contaminant accumulation on particular surfaces are measured and depositional fluxes inferred.
Alternatively, atmospheric flux methods measure contaminant inventories and speciation in the
atmosphere and the corresponding micrometeorology necessary to model aerosol transport.
Although progress has been made, there are no methods to directly measure dry deposition.
Furthermore, there are no unambiguous, widely-accepted methods recognized for the indirec:
estimation of dry deposition. The differing approaches, as summarized and evaluated by Hicks
57
-------
Relative Atmospheric Loadings... Revision Date: 15 March 1993
et al. (1986), Nicholson (1988), Davidson and Wu (1989) and Lindberg (1989) are:
(a) Inferential micrometeorological techniques based on vertical airborne
concentration measurements at one or more levels (e.g., tower-based eddy
correlation). Transfer rates from the atmosphere to the surface are obtained via
modelling, or where possible, direct measurement of surrogate species (e.g.
ozone and NO3") and inferring a deposition rate. Due to the complexity and
variability in the various processes which control dry deposition, methods which
assume a constant deposition velocity from literature values have an inherent large
degree of uncertainty. Inferential dry deposition measurements are generally
limited to fine particles, under limited terrain conditions (e.g. uniform vegetation
and adequate fetch) and over relatively short (hrs) time scales.
(b) Surface analysis techniques, such as the use of surrogate surfaces(e.g. Teflon
plates, glass microslides), vegetative throughfall/stemflow, and foliar extraction,
(Hanson and Lindberg 1991). Surrogate surface techniques are generally
applicable only to fine particles over a small spatial scale (cm2 to irij), but
integrate deposition over relatively long time scales (10s to 100s of hours). A
major advantage of surface techniques is that vegetative surfaces more accurately
represent deposition to natural surfaces that other approaches cannot physically
or mathematically duplicate.
(c) Watershed mass balance approaches, which have the advantage of integrating
measurements over relatively large spatial (ca. 5-10 hectares) and temporal
(months-years) scales. The major disadvantage of this technique is that the
watershed studied must be well-characterized hydrologically and in terms of
internal sources and sinks.
(d) Regional-scale langrangian and eulerian models (e.g., Levy and Moxim 1989,
Eliassen et al. 1988), which are typically based on emission inventories rather
than concentration measurements. As such, the transfer rates are at the mercy of
the input data to the model.
(e) Isotopic tracers, which can be utilized to infer the total (wet + dry) deposition
as well as to apportion the wet and dry fractions (the I37Cs/210Pb ratio in
precipitation is more than 2X the corresponding ratio in submicron aerosol).
Isotopic tracer techniques have the advantage of yielding absolute identification
of atmospherically-derived material, and the ability to infer deposition over
relatively large spatial scales. Depending on the isotopes utilized, deposition can
be estimated on temporal scales ranging from individual events (87Sr/86Sr) to
decades (137Cs/210Pb). Other tracers (e.g. 35S, 214Pb, and 7Be) have been used to
specifically measure dry deposition.
58
-------
Relative Atmospheric Loadings... Revision Date: 15 March 1993
6.2 GAS ABSORPTION AND VOLATILIZATION
6.2.1 Importance in the cycling of organic compounds
Little attention has been paid to the role of air-water exchange of semi-volatile
organic chemicals (SOCs) in lakes, estuaries, and oceans. Whereas the transfer of low
molecular weight, volatile and biologically-mediated gases such as 02, C02, CH4, H2S,
and 1 and 2 carbon halocarbons have received considerable attention for water quality,
carbon cycling, global warming, and other reasons (e.g. Thompson and Zafiriou, 1983),
SOCs such as PCBs, DDT, HCHs, PAHs, and toxaphene have been largely ignored (see
Eisenreich et al. 1981, Doskey and Andren 1981, Atlas and Giam 1986, GESAMP
1989). These compounds are derived from major anthropogenic emissions in
urban/industrial centers and world-wide agriculture, are transported globally, and are
often concentrated in the northern hemisphere atmosphere from 20°N to 45°N (Stanley
and Kites, 1991; Ballschmiter et al. 1981, Tanabe and Tatsukawa, 1986 and references
therein). Recent studies suggest that the air-water exchange of SOCs plays an important
role in the mass balancing of inputs to large aquatic systems such as the Great Lakes, the
Mediterranean Sea, and the world's oceans (Strachan and Eisenreich 1988, Burns and
Villenuve 1988, GESAMP 1989, Atlas et al. 1986). For example, the calculated
volatilization flux of PCBs out of the North American Great Lakes (Mackay 1989,
Swackhamer and Armstrong 1986, Strachan and Eisenreich 1988, Eisenreich 1987) is
estimated to be comparable to sedimentation losses. Since planet Earth is approximately
70% water by area, it is not surprising that the problem is one of interfaces. Large
aquatic systems such as the Great Lakes, Chesapeake Bay, and the coastal ocean have
large surface areas for transfer of chemicals that are unfortunately often close to regions
of contaminant input. Air-water transfer of SOCs occurs in both directions; whether
absorption or volatilization dominates is discussed below. This section discusses the
concepts and models applicable to air-water exchange of SOCs and how these models
have been applied in the field.
6.2.2 Concepts and Models
The theory and concepts of air-water exchange and mass transfer of chemicals
across water surfaces have been presented (e.g., Liss and Slinn 1983, Brutsaert and
Jirkha 1984, Buat-Menard 1986, Wilhelms and Gulliver 1991, Schwarzenbach et al.
1992). Air-water exchange refers to the transfer of chemicals across an air-water
interface, or air-side and water-side boundary layers (Figure 22). The gas concentration
in the atmosphere (C..g) attempts to reach equilibrium with the concentration of gas
dissolved in water (C^.^J. When equilibrium is achieved, the ratio of the gas activities
59
-------
Air-Water Exchange
a,.g
a,p
Henry's Law
W,dlSS
w.p
Henry's Law: H
C /C
a,g w.diss
Flux w.a = K(°C)
Air
Interface
Water
Figure 22. Air-water exchange
-------
Relative Atmospheric Loadings... Revision Date: 15 March 1993
in air and water at constant temperature are represented by Henry's Law constant (H =
Ct.g/CWidUi). The direction of chemical transfer is from the water to the air (e.g.
volatilization) when the activity in the water exceeds the activity (gas phase
concentration) in air. Chemical transfer from the air to the water (e.g., absorption)
occurs when the activity in the air exceeds the chemical activity in water. The processes
of gas absorption and volatilization occur simultaneously, and together contribute to the
net flux. The magnitude of mass transfer is controlled by a mass transfer coefficient or
piston velocity and the concentration or activity gradient across the interface (Liss and
Slater, 1974). Thus, the direction and magnitude of gas transfer is a function of the free
concentrations in air and water (activity gradient), wind speed (water side turbulence),
temperature, characteristics of the water (water chemistry; surface films), and the
physicochemical properties of the chemical compound (Henry's Law constant; octanol-
water partition coefficient, vapor pressure, and Schmidt number, Sc).
Air-water exchange may be visualized as diffusive transfer of a chemical across
a stagnant film of 0.1 to 1.0 mm thickness. At low wind speeds, insufficient energy
exists to stir the air and water films or boundary layers, and a completely stagnant
boundary layer is established (Stagnant Two-Film Model). Higher wind speeds generate
more turbulence in the boundary layers, and parcels of air and water are forced rapidly
to the surface. Exchange is dependent on the renewal rate of air and water parcels
(Surface Renewal Model). In highly turbulent seas, gas exchange is enhanced by
continual breakup of the surface and generation of a large number of bubbles of great
surface area (Bubble Ejection Model). Under turbulence and wind conditions occurring
in the Great Lakes and coastal seas, the first two models are most applicable.
The Stagnant Film Model (Whitman 1923, Liss and Slater 1974) describes the
transfer of a chemical by diffusion across stagnant air and water films on either side of
the air-water interface (Figure 23). The bulk air and water compartments are assumed
to be well mixed and offer no resistance to gas transfer. The transfer of chemical at the
interface is assumed to be instantaneous, offering no resistance to transfer. The rate of
gas exchange or mass flux is equal to FglJ = KOL (Cw-diJS-C*), where KOL is the overall
mass transfer coefficient (m/d), and Cw-di51 and C* (mol/m3) are the free water
concentration and the water concentration in equilibrium with the partial pressure of the
gas in the atmosphere (P,atm; C* = P/H), respectively. If the concentrations in air and
water are expressed as mol/m3, KOL as m/d, and H as atm mVmol, then the chemical
flux has units of mol/m2 day and a positive flux reflects volatilization. The overall
resistance to mass transfer is the sum of the resistance across the air and water films:
I/KQL = 1/k, + RT/Hkg where k,, and kg are the water film and gas film transfer
coefficients (m/day), respectively, R is the universal gas constant (8.2 x 10"5
atm/m3/mol-K), and T is absolute temperature (K). The Henry's Law constant
influences' both the magnitude of KQL and the concentration gradient. Since H is a
function of temperature (-2.5 fold increase with a 10°C increase), the direction and
61
-------
zw
interface
-za
Diffusion I
Well Mixed Air
Stagnant Air Film
Henry's Law Equilibrium
; Stagnant Water Film
Well Mixed Water
Ca,g Cw/a Ca/w Cw,diss
Figure 23. Stagnant Two Film Model
-------
Relative Atmospheric Loadings... Revision Date: 15 March 1993
magnitude of gas flux is influenced by the water temperature. The resistance to transfer
may arise either in the water or gas film, or a linear combination of the two. Assuming
average values for k., (20 cm/hr) and k^ (2000 cm/hr) resistance to transfer at 25°C is
dominated by the water film when H> -10"3 atm mVmol, and by the gas film when
H< -105 atm mVmol. Resistance in both phases occurs when H is between these
extremes. Compounds such as PCBs, PAHs, and DDT isomers generally incur
resistances in both phases. SOCs with H< 10"5 atm mVmol volatilize slowly at a rate
dependent on H; k» dominates and the rate is controlled by diffusion through the air.
Intermediate and high molecular weight PAHs have H's in this range (Lyman et al.,
1990). Examples of Henry Law constants at 25°C are shown in Table 21.
The Surface Renewal Model (Higbie 1935, Danckwerts 1951) involves the
periodic renewal of parcels of air and water on either side of the interface with new
parcels turbulently mixed to the surface. In mathematical terms, FgtJ =
(HN/'(r1DJ+>/(rwDw)](Cw diii- P/H) where air parcel renewal rates r, and rw are the air and
water parcel renewal rates and D. and Dw the chemical diffusivities in air and water
(cmVsec), respectively (Schwarzenbach et al. 1992). This equation has the same form
as that describing gas flux by the stagnant film model with the exception of how the mass
transfer coefficients depend on diffusivities. K^ depends on D1 in the stagnant two film
model while k depends on D1/2 in the surface renewal model. Enhancements in the
surface renewal model have been reviewed by Bennett and Rathun (1972) and
Theofannous (1984). New developments in the surface renewal model yield a Boundary
Layer Model as described in Deacon (1977) and Hanratty (1991).
The stagnant film and surface renewal models incorporate unmeasurable
parameters that must be estimated in the field (e.g. the depth of the interfacial zone in
the stagnant film model and the renewal rates in the surface renewal model). The
stagnant two film model is used oftentimes for its simplicity even though it lacks
mechanistic accuracy. Either treatment yields comparable gas fluxes. For convenience,
the two film model may be expressed in terms of fugacity (e.g., see Mackay et al. 1986).
The presence of a surface film may add resistance to air-water transfer (Mackay 1982,
Liss 1983, Asher and Pankow 1991) by adding a third resistance film. However, surface
films in lakes and seas are sufficiently broken up by turbulence, except under the most
calm conditions, that they represent little or no resistance to SOC transfer, at least on
longer time scales. However, Mackay et al. (1991) argue that SOCs accumulate at the
air-water interface by thermodynamic association with structured water at the interface.
The net effect is to reduce transfer rates (as observed by experiments conducted by Asher
and Pankow (1989) by adding a layer of higher capacity which hinders the diffusion
process.
63
-------
TABLE 21
HENRY'S LAW CONSTANTS OF SEMIVOLATILE ORGANIC CONTAMINANTS
Compound
PCBs (1 to 10 Cl)
PAHs (1 to 6 rings)
Chlorobenzene
Tetrachlorobenzene
Hexachlorobenzene
Lindane fr-HCH)
a-HCH
2,3,7,8-TCDD
Log Henry's Law
Constant
(atm-mVmol)
-3.6 to -4.0
-7.4 to -4.0
-4.6
-2.4
-2.9
-5.6
-5.3
-4.5
-------
Relative Atmospheric Loadings... Revision Date: 15 March 1993
6.2.3 Air and water concentrations
Chemical activities in the gas and water phases are required to assess the direction
of gas transfer and activity gradient. On the air side, this requires explicit knowledge
of the gas phase concentration. Compounds with vapor pressures > 10"5 atm exist
predominantly in the gas phase (Junge 1977, Bidleman 1988) and their atmospheric
concentration is easily determined. Compounds with vapor pressures between 10"5 and
10~8 atm are distributed to varying extents between the gas and aerosol phases (Bidleman,
1988). Only the gas phase participates in air-water exchange. Sampling procedures
and/or modeling of gas-particle distributions are required to properly assess the gas phase
species of interest (see 4.2).
On the water side, the chemical activity is equivalent to the concentration of
freely dissolved, unassociated species. As portrayed in Figure 22, SOCs in marine and
fresh waters are distributed between the dissolved and paniculate phases according to the
hydrophobicity of the chemical, and the composition, concentration, and size of the
particle population (Karickhoff, 1984, Elzerman and Coates 1987, Schwarzenbach et al.
1992, Chiou 1990). SOCs partition into organic carbon-rich particles, the magnitude of
which may be correlated with the fractional organic carbon content of the particles (foe)
and K^,. Correlations take the form of Kp = aK^1" where Kp is the equilibrium partition
coefficient (ml/g) and a, b are constants (b — 0.7 to 1.0; a depends on compound). For
example, Schwarzenbach and Westall (1981) suggest log Kp = 0.72 log K^, + log foe
+ 0.49. Using this and similar correlations (Karickhof 1984, Lyman et al. 1990), the
fraction of chemical (fj) in the dissolved phase may be calculated as fd = l/[(TSM)kp+l]
where TSM is the total suspended matter concentration (g/L). At typical Great Lakes
TSM values of 0.5 to 2 mg/L, PCBs with log Kow values -4.5 to 7.5 are mostly in the
dissolved phase (Eadie and Robbins, 1987).
Methodologies to separate dissolved from paniculate phases includes the use of
high volume filtration through glass fiber quartz filters (nominal pore size of 0.5 to 1
um) and continuous flow centrifugation. This protocol yields "operational" separation
of dissolved and paniculate species because colloidal size particles pass into the
"dissolved phase". Field comparisons suggest that filtration provides a better estimate
of the aqueous activity of PCB congeners than does centrifugation (Swackhamer et al.
1993). Gschwend and Wu (1985) and Baker et al (1986) modeled the distribution of
SOCs between the dissolved, suspended, and colloidal phases. Dissolved and colloidal
natural organic matter also bind SOCs in dilute solution (e.g., McCarthy and Jimenez
1985, Landrum et al. 1987). Improved methods for measuring dissolved SOC
concentrations in natural waters and further characterization of the rates of chemical
uptake by aquatic particles are necessary to accurately predict chemical distributions
(Baker et al. 1991). Sproule et al (1991) describe such an in situ sensing device. A
further complication is that SOCs may be taken up by aquatic phytoplankton at
65
-------
Relative Atmospheric Loadings... Revision Date: 15 March 1993
sufficiently slow rates compared to growth kinetics that equilibrium calculations are
accurate only in oligotrophic ecosystem (low TSM) and/or at cold temperatures
(Swackhamer and Skogland 1991).
6.2.4 Mass Transfer Coefficients
Mass transfer of gases across the air-water interface is a function of wind speed,
waves (height, frequency), bubbles (breaking waves), and heat transfer. Laboratory wind
tunnel experiments over the last 20 years have elucidated the controls of transfer,
especially for gases experiencing liquid (water) control (H > 10-3 atm m3/mol) such as
02 and COz (for reviews, see Liss 1983, Liss and Merlivat 1986, Brutsaert and Jirka
1984; Wilhelms and Gulliver 1991). Figure 24 (modified from Liss and Merlivat, 1986)
shows that liquid phase controlled gases exhibit low transfer rates at low wind speeds
measured at 10 to 60 cm above the water surface, corresponding to a smooth surface
regime. Transfer rates increase at a transition at 5±3 m/s corresponding to a rough
surface regime and the presence of capillary waves. The next transition to higher k»
values occurs in the wave breaking regime at about 10 to 13 m/sec wind speed. K,, is
generally found to be proportional to Sc'2/3 at low wind speeds (i.e., k« a D273) under
calm conditions, and k, a Sc'm (i.e., k» a D1/2) in the rough surface regime. Laboratory
experiments provide reasonable agreement for k,, with measures of turbulence but may
not accurately mimic the complexity of the real environment with respect to bubble
formation, spray, boundary effects, etc.
Experimental approaches to determine gas transfer rates in the field include the
direct flux method (box), oxygen balance method, profile and eddy correlation
techniques, use of natural and bomb-produced I4C, the radon deficiency method, and the
use of tracer gases indigenous (e.g., methane) or added (SF6) to the water. Most of these
techniques are most applicable to determining gas transfer coefficients for compounds for
which the resistance to transfer is dominated by the water phase (02, C02, 03, most PCB
congeners, most PAHs, low MW chlorinated solvents). However, the eddy correlation
technique may be appropriate for determining gas transfer coefficients controlled by air
film resistance (e.g., water vapor). In such studies, kg has been found to vary linearly
with wind speed and therefore expressed as kg = CD x U,0 where CD is the drag
coefficient and U10 is the wind speed at 10 m height.
The tracer gas SF6 has now been used successfully in long term field experiments
of gas transfer and mixing in lakes, rivers, and oceans since the first report by
Wanninkhof and co-workers (Wanninkhof et al. 1985; 1987; 1991; Wanninkhof, 1986;
Ledwell ei al. 1986; Watson et al. 1991; Watson and Ledwell 1988; Watson and
Liddicoat 1985; and Upstill-Goddard et al. 1990). SF6 is an excellent tracer of water
controlled gases because is has steady, low, and well-defined concentrations in the
66
-------
150
120
u
_o ^
Q) ^
> E
i- U
(0
80
40
Smooth
Surface
Regime
kw"Sc-2/3
Rough
Surface
Regime
kw'Sc-1/2
X
X
Breaking Wave .
(Bubble) /
Regime
4 8 12 16 20
Wind Speed, Uj0 m/sec
Figure 24. Water phase transfer velocity versus wind speed
(modified from Liss and Merlivat, 1986)
-------
Relative Atmospheric Loadings... Revision Date: 15 March 1993
ambient atmosphere, is chemically and biologically inert, does not sorb to aquatic
particles, is analyzed precisely and accurately at trace concentrations, and its loss rate
to the atmosphere is sufficiently rapid to perform experiments to yield kw values on a
daily basis. Liss and Merlivat (1986) took the data of Wanninkhof et al. (1985) for gas
transfer on a small lake, successfully compared them to previous k, estimations, and
proposed that the following three relationships might be used to determine k» in various
wind regimes up to wind speeds of at least 13 m/sec (u is wind speed at 10m; Ic,, is
normalized to COz having a Sc number of 600; SF6 has a Sc number of - 1300):
kw = 0.17 u for u < 3.6 m/sec
kw = 2.85 u - 9.65 for 3.6 < u < 13 m/sec
k» = 5.9 u - 49.3 for u > 13 m/sec
These relationships have been applied by GESAMP (1989) to estimate gas transfer in the
world's oceans, and others have applied them to PCB exchange in the Great Lakes
(Achman et al. 1993). Wanninkhof and others have now conducted several SF6 tracer
studies in lakes of different surface areas and in the ocean and have observed a
dependence of kw on the size (fetch and turbulence) of the lake. Wanninkhof et al.
(1991) have proposed a power law relationship between k^ and wind speed, based upon
all data on SF6 transfer in lake experiments: k,, = 0.45 u1-64 (Scchem/oOO)1/2. The
correlation coefficient was 0.66 indicating there is still considerable scatter and
uncertainty in the dependence of kw on wind speed. Livingstone and Imboden (1992)
support their approach since k» is probably a function of both the mean wind speed and
the probability distribution of the wind speed. They suggest applying a Weibull wind
speed distribution with power law expressions to obtain the effect of wind on k,,. Table
22 is a compilation of literature correlations relating k,, and kg to wind speed from both
laboratory and field experiments.
6.2.5 Field Measurements
Air-water exchange of SOCs in large ecosystems has been difficult to quantify but
there is a growing appreciation that it is an important component in whole lake or
ecosystem mass balances, and plays a large role in the chemical entry into the food
chain. Atlas and Giam (1986), Atlas et al. (1986), and GESAMP (1989) have reported
that the major mode of PCBs and other organochlorine input to the world's oceans is
water absorption of atmospheric gases. Atlas et al. (1986) have calculated that of 16.9
x 10* g/yr PCBs entering the world's oceans, 70% resulted from air-water exchange.
GESAMP (1989) reported similar loadings. In the Great Lakes region, Strachan and
Eisenreich (1988) used mass balance calculations for the mid-1980's to suggest that the
SOC volatilization dominated all inputs and outputs (Table 10).
Approaches to estimating the direction and magnitude of gas exchange in large
68
-------
TABLE 22
SOME EMPIRICAL RELATIONSHIPS BETWEEN kw AND k^ AND WTNDSPEED
kg (cm/hr)
Mackay and Yuen (1983)
Schwarzenback et al. (1992)
kw (cm/hr)
Mackay and Yuen (1983)
Wanninkofh et al (1985)
Liss and Merlivat (1986)
(Sc = 600)
Wanninkhof et al (1991)
kg = 0.0065(6.1 + 0.63U10)05U10
k^, = 0.2U10 + 0.3
Km) = 1-75 x ltf(6.l + 0.63U10)°-5
kv,(SF« = W(-8.9 + UIO)
kw = 0.17U10 for Ujo < 3.6 m/sec
k., = 2.85U10 - 9.65 for 3.6 < UIO < 13 m/sec
kw = 5.9U10 - 49.3 for U10 > 13 m/sec
K, = 0.45U101-64(Scchem/600)°-3
-------
Relative Atmospheric Loadings... Revision Date: 15 March 1993
aquatic systems include mass balance calculations (based on literature data and
measurements in the field), and measurement of activity gradients in the field with
calculation of mass transfer coefficients. Atlas and Giam (1986), GESAMP (1989), and
Atlas et al (1986) have estimated the atmospheric deposition of organochlorine
compounds including PCBs and DDT to the world's oceans including gas exchange based
on published data in the literature. In general, the two film stagnant transfer model was
applied to the calculation of the global mean gradient across the air-water interface, and
mean global wind speeds to estimate mass transfer coefficients. GESAMP (1989) report
on the total deposition of organochlorines to the oceans. For example, the mean
atmospheric flux (/ig/m2-yr) for some compounds was estimated as: EHCH, 14: HCB,
77: E-DDT, 165; EPCBs, 239. With all the caveats presented, they emphasize that the
direction and magnitude of transfer are both uncertain but from 25 to 85% of the total
compound deposition is due to gas exchange. Atlas and Giam (1986) provide similar
estimates for the ocean. Atlas et al (1986) suggest that PCB transfer to the ocean
surface is —4.5 /ig/m2-yr, significantly lower than the above estimates, and is dominated
by air-to-water transfer of gas phase PCBs. Using similar approaches, Doskey and
Andren (1981), Murphy et al. (1983), Eisenreich et aL (1981), Strachan and Eisenreich
(1988), and Mackay et al. (1986) estimate that the direction of gas transfer for PCBs is
from the water to the air (i.e., volatilization) for the Great Lakes, and volatilization
represented a dominant component in both the atmospheric deposition and whole lake
cycling. The difference in PCB behavior between oceans and the Great Lakes (PCB gas
absorption versus volatilization) is likely due to the close proximity of atmospheric
sources and the presence of significant inputs to surface waters in the Great Lakes.
Strachan and Eisenreich (1988) estimated that volatilization represents about 45 % to 87%
of total PCB outputs from the Great Lakes. Swackhamer and Armstrong (1986)
estimated that PCB volatilization from Lake Michigan was about 5.6 j*g/m2-yr based on
a comparison of Lake Michigan and remote lake sediment cores. Swackhamer et al.
(1988) reported on the gas transfer of PCBs from a remote lake located on an island in
Lake Superior. They measured inputs and outputs from all sources except air-water
exchange and determined by mass balance that the volatilization flux was ~ 8.5 /ig/m2-yr
(Table 23). Larsson and co-workers (Larsson, 1983, 1985; Larsson et al., 1990) using
mass balance techniques in mesocosms and in the field, have shown that PCB
contaminated sediments are a source of PCBs available for volatilization to the
atmosphere.
Much less information is available on PAHs. McVeety and Hites (1988), using
the same approach on Siskiwit Lake, determined that low and medium molecular weight
PAHs were lost by volatilization in general proportion to their vapor pressure, and
ranged from 0 to 80% of total outputs (Table 23). Strachan and Eisenreich (1988)
estimated that benzo[a]pyrene volatilization represented approximately 2 to 19% of total
outputs from the Great Lakes in comparison to ~50% from Siskiwit Lake (McVeety and
Hites, 1988).
70
-------
TABLE 23
MASS BALANCE OF PAHs and PCBs IN SISKIWIT LAKE, ISLE RCYALE,
LAKE SUPERIOR
(McVeety and Kites 1988, Swackhamer et at., 1988)
Inputs Otg/m2-yr)
Atmosphere
Rain
Snow
Dry Particle
Total
Phenanthrene
0.35
0.2
2.85
3.4
BaP
0.15
0.05
0.5
0.7
PCBs
7.3
3.2
3.6
14.1
Outputs (/xg/m2-yr)
Sedimentation
Volatilization
Total
0.7
2.7
3.4
0.4
0.3
1.4
5.6
8.5
14.1
-------
Relative Atmospheric Loadings... Revision Date 15 March 1993
Achman et al. (1993a, b) have investigated the air-water transfer of PCBs and
PAHs across the air-water interface of Green Bay, Lake Michigan. The strategy was to
simultaneously collect air and water samples seasonally from aboard ship at several
locations in Green Bay, Lake Michigan in 1989. The fugacity gradients were calculated,
and wind speeds measured over the water were used to calculate daily mass transfer
coefficients. This is the same strategy that Baker and Eisenreich (1990) applied to
determining the flux of PCBs from Lake Superior. Table 24 reports the resulting PCB
data and compares them to air-water fluxes reported elsewhere. These data suggest that
volatilization fluxes of PCBs of about +50 to +200 ng/m2-day may be a common
phenomenon in large, uncontaminated freshwater systems such as the Great Lakes.
However, the direction and magnitude of gas transfer depends on season and location,
the relative concentrations of compound in the atmosphere and water, and the properties
of the chemical (especially H). Achman et al. (1993a,b) have compared the air-water
exchange of PCBs and PAHs in Green Bay, Lake Michigan in 1989. Figure 25 shows
the daily flux of individual PCBs and PAHs at the same location and time in Green Bay.
Remembering that H influences both the fugacity gradient and the magnitude of the mass
transfer coefficient, PCB behavior is dominated by volatilization while PAH behavior is
dominated by absorption. That is, PCBs are being lost from the bay and dramatically
influence the over-lake concentrations (Hornbuckle et al. 1993) while atmospheric PAHs
are a major source to the water. PAHs have lower H values and are dominated by gas
phase resistance to transfer. Baker and Eisenreich (1990) drew the same conclusion for
Lake Superior.
McConnell et al. (1993) have estimated hexachlorocyclohexane fluxes to the Great
Lakes based on air and water concentrations determined in 1989 and 1990. They
concluded that both a-HCH and g-HCH have net inputs from water to air by gas
exchange as follows (ng/m2/yr): a-HCH - Michigan, -2006: huron, -4559; Erie, -121;
Ontario, 1196; g-HCH-1691, Michigan; -1748, Huron; -1246, Erie; -1424, Ontario.
Although net flux was estimated to be from air to water, calculations showed that
volatilization dominated only under warm water conditions in August. The direction, and
certainly the magnitude, is a function of season as driven by water temperature and
changing air and water concentrations.
To demonstrate the importance of air-water exchange on ecosystem mass balances,
Figure 5 depicts the situation for PCBs in Lake Superior today (Jeremiason et al 1993).
Important points are: 1) Reservoir of PCBs is -9000 kg, most in water and sediments;
2) Inputs are ~220 kg/yr while outputs are -925 kg/yr, -85% of which is allocated
to volatilization; 3)-700 kg/yr PCBs have been lost from Lake Superior on average
from 1980 to 1992 very closely matching the volatilization flux; 4) The change in water
column concentrations of EPCBs of 1.3 ng/L in 1980 and 0.3 ng/L in 1992 results from
volatilization; and 5) The residence time of PCBs in Lake Superior is -4 to 5 years.
The magnitude of volatilization of PCBs suggests that PCBs are a major source to the
72
-------
TABLE 24
ESTIMATED AIR-WATER FLUXES OF PCBS
(+ FLUX = VOLATILIZATION)
LOCATION
Green Bay
Green Bay
Lake Superior
Lake Superior
Lake Superior
Siskiwit Lake
Lake Michigan
Lake Michigan
Lake Michigan
Lake Ontario
River Elm, Sweden
Oceans
Oceans
Oceans
Oceans
PCB FLUX (ng/m2-day)
+ 15 to + 1300 (1-3 m/sec)
+50 to + 1300 (4-6 m/sec)
+ 19 (still air)
+ 141 (5 m/sec)
+63
+23
Oto +13,000
+240
+ 15
+81
+50
-160 to -450
-4.5
-0.6 to -10
+ 12 to -35
REFERENCE
Achman et al (1993)
Achman et al. (1993)
Baker and Eisenreich (1990)
Baker and Eisenreich (1990)
Strachan and Eisenreich (1988)
Swackhamer et al (1986)
Doskey and Andren (1981)
Strachan and Eisenreich (1988)
Swackhamer and Armstrong (1986)
Mackay (1989)
Larsson et al. (1990)
GESAMP (1989)
Atlas and Giam (1986)
Atlas et al. (1986)
Iwata et al. (1993)
-------
GREEN BAY, LAKE MICHIGAN
o
cs
X
a
0 •
-25 -
i -50-
-75 -
-100
200
>s
O
is
I
•—»
K
C -200 -
-100 -
-300 /
-1000 '
-2000
Site 10, July 31, 1989
Wind Speed - 1 m/s Woter Temp. » 15.6 »C
PCBs
PAHs
Il,
4.5 5.0 5.5 6.0 4.0 5.0 6.0 7.0
-Log HLC (otm rr^/mol) -Leg HLC (otm m3/mol)
Site 10, October 22, 1989
{Wind Speed ° 6.5 m/s Woter Temp. • 7.6 *c)
PCBs PAHs
• r
1 <' I r
i !
5.5 4.0 5.0
6.0
7.3
-Log HLC (otm m^/moi) -Log HLC (otm m^/moi)
• Tiguxe 25. Air-water exchange fluxes of PCBs and PAHs, Green Bay
-------
Relative Atmospheric Loadings... Revision Date: 15 March 1993
atmosphere and contribute to global distributions.
7.0 EVALUATION OF CURRENT SAMPLING AND
ANALYTICAL PROCEDURES
7.1 TRACE ELEMENTS - Evaluation of current methodologies
7.1.1 Atmospheric Sampling - trace elements
The major techniques used to measure aerosol trace element concentrations include
filtration collection on polycarbonate/Teflon filters. Often this is proceeded by
differential size separation using dynamic or kinematic techniques such as successive
nests of cascade or orifice impactors. The limitation of such filtration techniques is that
they either exclude the finest fraction below the filter cut off where higher concentrations
of trace elements exist, or they fail to collect the largest fraction where most of the
depositional mass exists. Analysis of the filters include direct neutron activation, x-ray
fluorescence, or digestion followed by graphite furnace atomic absorption spectroscopy.
inductively coupled plasma emission spectroscopy, etc. Dry deposition of aerosols can
be measured by exposure of surrogate surfaces of various types, analyzed similar to
filters, as mentioned in the previous section of dry deposition of nitrogen.
7.1.2 Precipitation Sampling - trace elements
Due to their typically low concentrations and high potential for contamination, the
accurate determination of trace elements in the atmospheric environment requires special
equipment, rigorous procedures, and sensitive analytical equipment. Reviews of the
historical data on the concentration of trace elements in precipitation (Galloway et al.
1982; Barrie et al 1987) cast considerable doubt on the accuracy of reported values and
efficacy the techniques utilized. In many of these studies, it is readily apparent that the
authors failed to observe the requisite ultra-clean sampling and handling precautions
(e.g., Bately and Gardner 1977; Ross 1986), resulting in either gross contamination, or
conversely, irreversible losses of certain metals (e.g., lead) to container walls (Chan et
al. 1983). Furthermore, many laboratories simply do not possess the capabilities to
conduct routine analyses in the ng/L range. The reliability of most available databases
on the trace metal concentration in precipitation are thus compromised, as are the
resultant estimates of wet depositional fluxes.
Most of the first successful attempts to sample atmospheric trace metals were in
75
-------
Relative Atmospheric Loadings... Revision Date: 15 March 1993
remote, open-ocean locations, and involved manual collections and meticulous handling
precautions (e.g., Duce et al 1991). More recent studies over continental regions have
been conducted using a wide variety of collection devices, handling procedures and
analytical techniques. Several recent studies have attempted to specifically address the
complexities involved in accurate determination of the trace metal concentrations in
precipitation to derive standardized procedures (Ross 1986, Tramontane et al. 1987,
Keller et al. 1988, Vermette et al. 1992, Scudlark et al. 1992). From these studies,
several common considerations emerge:
1. Sampling must be conducted on a wet-only basis. Passive capture of dry fallout
is a serious source of contamination. Likewise, bulk (wet + dry) sampling
techniques are too easily corrupted to be considered reliable.
2. Polyethylene (LDPE or HDPE) and TeflonR are the only materials that are
compatible with sample contact.
3. To remove metal impurities present from manufacture or prior use, all plasticware
must be scrupulously cleaned, which usually involves successive leaching in a
series of acid solutions (HNO3 and HC1).
4. To prevent (or reverse) the adsorption of certain metals on the plastic container
walls, the sample must be acidified to below pH 2 with ultra-high purity acid.
5. Depending on the element and its concentration, analysis can be accomplished by
a variety of techniques, including graphite furnace atomic absorption
spectrophotometry, inductively-coupled plasma atomic emission
spectrophotometry, mass spectrometry, and polarography.
Due to the typically low concentrations in precipitation and high potential for
contamination, the accurate assessment of trace metals in atmospheric samples requires the strict
adherence to a rigorous quality assurance program. Components of such a program should
include the routine evaluation of procedural blanks, the use of externally-certified analytical
reference samples (e.g., EPA or NIST), and conducting inter-laboratory analytical comparisons.
including the use of redundant techniques where possible. Operational blanks provide a
comprehensive assessment of the background contamination during sampling and handling. A
"field blank", which mimics actual sampling procedures as closely as possible, should be
evaluated on a regular (e.g., monthly) basis, or if there is a significant change in the personnel,
methods, materials, site activities, or reagents associated with sampling. Such blanks can be
utilized to identify and remedy any source of contamination, and to correct apparent precipitation
concentrations for background trace metal levels. As a general rule, this correction is required
only when the blank contribution exceeds 10% of the average metal concentration (an
approximate limit of analytical confidence).
76
-------
Relative Atmospheric Loadings... Revision Date: 15 March 1993
7.2 SEMIVOLATILE ORGANIC CONTAMINANTS - Evaluation of current
methodologies
7.2.1 Atmospheric Sampling - SOCs
The separation of gases and particles in field studies is normally accomplished by passing
air through a glass fibre, quartz, or membrane filter followed by a solid adsorbent such as
polyurethane foam (PUF), XAD-2 resin, and Tenax GC. The filters are extracted and analyzed
separately by gas chromatography with electron capture or mass spectrometric detection. The
total concentration is the sum of the operationally-defined gas and particle SOC fractions.
Yamasaki et al. (1982), Bidleman et al (1986), Pankow (1987), and Bidleman (1988) suggest
that the G/P distribution may be determined by Cp/C, or A(TSP)/F which is a function of p° and
T. Sampling artifacts which alter the equilibrium G/P distribution such as "blow-off from
collected particles, sorption of gases onto the filter, and degradation and/or transformation of
SOCs on the filter may be serious practical problems. The diffusion denuder (Lane et al. 1988,
Coutant et al. 1989) is an alternative to the collection and speciation of SOCs. In a denuder,
the gases and particles pass through a denuder section consisting of parallel tubes or concentric
cylinders coated with an adsorbent efficient in trapping the SOC gas. The particles pass through
the denuder and are collected on a filter. A diffusion separator is similar in theory but
minimizes collection and analysis difficulties presented by the denuder and is under active
research (Turpin et al., 1992). In the diffusion separator, gases and particles are separated in
a short tube based on diffusion differences, and the G/P distribution is determined from the
fraction of SOC collected in the annular and core flow in comparison to theory. The denuders
should offer definitive data on the G/P distributions in the atmosphere.
7.2.2 Precipitation Sampling - SOCs
Previous studies have used bulk (Glotfelty et al. 1991, Webber 1983, EPA 1991) and wet-
only (McVeety and Kites 1988, Ligocki and Pankow 1985a, 1985b, Franz et al. 1991, Duinker
and Bouchertall 1989, Chan and Perkins 1989, Brun et al. 1991, Murray and Andren 1992)
deposition samplers to collect SOCs in precipitation. Total depositional fluxes estimated from
bulk deposition samples are generally quite high, strongly influenced by local contamination, and
thought to be unrepresentative of actual depositional rates.
Franz et al. (1991) systematically evaluated the performance of several SOC precipitation
samplers. Samplers with Teflon-coated and stainless steel collection surfaces, adsorbent and
solvent-based isolation systems, and with or without filters were codeployed in central Minnesota
for one year. Aside from operational problems (e.g., blown fuses, faulty motors), all samplers
efficiently collected SOCs in precipitation. Precipitation frequently backed-up in the sampler
77
-------
Relative Atmospheric Loadings... Revision Date: 15 March 1993
which relied upon gravity to pull the water through an in-line filter, resulting in possible
revolatilization of the SOCs. While the solvent-based isolation system (Chan and Perkins 19xx)
was efficient at collecting SOCs from precipitation, higher blank levels and problems with
transporting solvents may limit the use of this system. The study of Franz et al (1991) and
others (Murray and Andren 1992) suggest that rough and hydrophobic Teflon-coated collection
surfaces retain particles from precipitation and may adsorb gaseous phase HOCs from the
atmosphere. A major problem identified by Franz et al. was the relatively small sample sizes
collected by their samplers which had collection surfaces ranging from 0.081 to 0.2 m2,
complicating measurement in the sub-ng/L range.
A wet-only precipitation sampler which automatically collects rainfall and isolates and
preserves target HOCs, including polycyclic aromatic hydrocarbons (PAHs), polychlorinated
biphenyls (PCBs), and current use agrichemicals is currently in use in the Chesapeake Bay
Atmospheric Deposition Study (Leister and Baker, 1993) is designed to operate automatically
without intervention for two weeks and meets the following criteria:
1. Wet-only collection. The sampling train is opened to collect precipitation during events,
but remains sealed during dry periods to prevent inadvertent contamination.
2. Large sample size. Anticipated HOC concentrations in precipitation range from pg/L to
ng/L. To insure that the samples which are collected contain HOCs in excess of the
analytical detection limits, this sampler generally collects more than 10 L of precipitation
during each deployment period.
3. Inert sampling train. Due to the low analyte levels in the samples, the sampling train is
constructed of inert materials (stainless steel and Teflon) to avoid contamination of the
samples during the deployment. The sampling train must be cleaned with solvents in the
field between samples.
4. Immediate, in situ isolation of HOCs from the precipitation. Once collected, HOCs in
precipitation may revolatilize, especially if the sample warms during storage. In addition,
redistribution of HOCs between dissolved and paniculate phases may occur. To limit
these artifacts, the paniculate and dissolved HOCs are isolated and preserved during each
precipitation event.
5. Automated operation. This sampler is to be deployed unattended in the field for at least
two weeks.
The sampling train of the CBADS sampler consists of a 1 m2 funnel constructed from polished
316 grade stainless steel. The neck of the funnel is attached to a stainless steel vertical tube
containing two liquid level sensors followed by a filter holder assembly, which in turn is
connected to a resin column. Water is pulled through the sampling train by a peristaltic pump
78
-------
Relative Atmospheric Loadings... Revision Date: 15 March 1993
downstream from the resin column, and stored in a reservoir. The sampling train is covered by
a aluminum lid during dry periods which is swung away from the funnel at the beginning of each
precipitation event. This sampler has been in continual operation for 2-1/2 years on the
shoreline of the Chesapeake Bay, allowing wet depositional fluxes and SOC speciation to be
measured.
7.3 MERCURY
The wide spread human-health and environmental worries with mercury as a global and
local pollutant has led to growing improvements in the quality of mercury determinations in the
environment. It was painful to recognize the extent of inaccurate environmental mercury
measurements, particularly in atmospheric and aquatic investigations. Such flaws have been an
impediment to understanding the biogeochemical cycling of mercury and assessing the influence
from anthropogenic inputs of mercury. The need for accuracy, and high quality, critical
experimental designs which incorporate ultra trace-metal clean sampling and analytical protocols
must be recognized by scientists involved in mercury research. Here we will present a summary
of the state-of-the art with respect to the determination of mercury in the atmosphere and natural
waters.
7.3.1 Atmospheric Hg
In general, sampling techniques, apparatus, and protocols for the measurement of Hg in
the gaseous and paniculate phases follow the well-tested two-stage gold amalgamation
methodology developed by Fitzgerald and Gill (1979) for oceanic studies. Briefly, air collections
of particulate Hg are made using air filters having a nominal pore sizes approaching ca. 0.3 to
0.45 j*m, which is in accord with traditional atmospheric analytical chemical practices. A
variety of filter materials are used, including quartz wool plugs (e.g., Fitzgerald et al. 1992,
Iverfeldt 1991a,b), quartz and glass fiber filters (e.g., Fitzgerald et al. 1983, Mason ei al,
1992, Dumarey et al, 1979), polycarbonate, and teflon substrates. The airborne total gaseous
Hg phase (TGM) is operationally defined as the quantity of Hg which passes through the filter
and is collected on gold or gold-coated substrates, such as quartz sand or wool (Fitzgerald and
Gill 1979; Slemr et al, 1979). The TGM phase has been successfully partitioned into its
components using in-line arrangements of stacked columns containing selective adsorbents (e.g.,
gold; CarbosieveR: activated carbon, CarbotrapR). Initial partitioning studies indicated that TGM
was composed principally of Hg° (Fitzgerald et al 1981, 1983; Slemr et al. 1981). Most
recently, chromatographic separation procedures have been applied to identify specific organo-
mercury species, such dimethyl mercury in addition to Hg° (e.g., Bloom and Fitzgerald 1988,
Ballantine and Zoller 1984, Schroeder and Jackson 1984). Applications using the analytical
technique developed by Bloom and Fitzgerald (1988) have provided most of the information on
the chemical, speciation of the gas phase. The apparatus consists of three atmospheric Hg
sampling trains stacks containing the following trapping materials, arranged in sequence to form
79
-------
Relative Atmospheric Loadings... Revision Date: 15 March 1993
a sampling stack.
1. Paniculate filter (e.g., quartz wool) / Au (coated quartz sand): yields a total gaseous Hg
determination (TGM).
2. Paniculate filter (e.g., quartz wool) / CarbotrapR (graphitized carbon) / Au (coated quartz
sand): provides a collection that can be used for direct Hg speciation determinations of
alkylated Hg (i.e., MMHg; DMHg) by GC separation and atomic fluorescence detection
(CTS).
3. Paniculate filter (quartz wool plug): for paniculate Hg determinations (PM).
The quartz wool/Au system provides a measure of the TGM concentration (Hg° + alkylated
species), while the quartz wool/CarbotrapR/Au yields a specific determination of the
concentrations of DMHg and MMHg. Since Hg° is not trapped very efficiently on CarbotrapR,
most of the Hg° is found on the gold trap, and the total Hg° is obtained from the sum of both
traps (Au + CarbotrapR). Moreover, the sum of the organo mercury species and Hg° found in
the CarbotrapR / Au sampling train should equal the TGM from the separate gold collection.
This provides a convenient mass balance constraint. Additional details associated with these
various trapping materials can be found in Fitzgerald and Gill (1979), Kim and Fitzgerald
(1986), and Bloom and Fitzgerald (1988).
The total gaseous mercury (TGM) analyses are generally conducted by the two-stage gold
amalgamation technique (Fitzgerald and Gill, 1979) with detection by atomic fluorescence
spectroscopy [(AFS); Bloom and Fitzgerald, 1988]. Currently, the overall atmospheric TGM
methodology yields a detection limit of ca. 0.15 ng nr3 and a precision of between 10 and 15 %
at 1.5 ng m'3, based on sample volumes ranging from 0.5 to 2 m3. Paniculate Hg is determined
following pyrolysis of the quartz wool plug or filters and trapping on gold. Chemical speciation
of the gaseous phase is achieved through analysis of the graphitized substrate (CarbotrapR)
collections using gas chromatographic separation and AFS detection (Bloom and Fitzgerald
1988). The detection limit for the determination of monomethyl mercury and dimethyl Hg is
ca. 5 pg m"3 The particulate mercury can be analyzed using wet digestion procedures and
derivatization with tetraethylborate to determine HgT, HgR, and MMHg following procedures to
determine Hg species in precipitation (see below).
The precision of measurement can be readily improved by increasing the volumes of air
sampled. A precision of < 5% can be achieved, and would be required, for example, to study
a current question as to whether Hg° is presently increasing in the global atmosphere as a result
of anthropogenic emissions. Indeed, in a recent controversial paper based on non-synoptic data
from 7 oceanographic cruises of short duration, Slemr and Langer (1992) concluded that annual
atmospheric Hg increases of ca. 1.5% for the Northern Hemisphere and ca. 1.2% for the
Southern Hemisphere had occurred for the period between 1977 and 1990. While the inferred
80
-------
Relative Atmospheric Loadings... Revision Date: 15 March 1993
increases do agree with expectations, the precision of measurement appears inadequate, and the
experimental design does not account for short time scale variations of both a natural and
anthropogenic origin. For example, in a two-month study, Fitzgerald and co-workers (1989)
found variations in atmospheric Hg concentration over the northeast Pacific ocean that were
comparable to the changes reported for the 13 year period in the Slemr and Langer work. Thus,
the very important question of whether Hg is increasing in the atmosphere has not been properly
addressed. The desirable approach would use a sampling and analytical strategy similar to the
successful Atmospheric Lifetime Experiment Program (ALE). The ALE studies of
contemporary temporal changes in the atmospheric concentrations of the freons, methyl
chloroform and carbon tetrachloride (Golambek and Prinn 1986) show that two to three years
of on-site continuous measurements are necessary to deal satisfactorily with questions of natural
variability and to resolve the influence of pollution on constituents such as Hg° in the
atmosphere. In addition, measurements must be carried out in a network context. Stations should
selected in both the northern and southern hemispheres. These locations must be remote from
significant local and regional sources of Hg°. For example, the ALE network used sites on the
west coast of Ireland, Barbados, Hawaii (Mauna Loa), American Samoa (Cape Matatula), and
Tasmania (Cape Grim).
7.3.2 Mercury: Precipitation Sampling and Analysis
At present, the preferred method for collecting rain for mercury studies is on an event
basis by trained personnel employing ultra-clean techniques (Fitzgerald er al. 1991, 1992,
Iverfeldt 1991a,b). Mercury has been examined in rainfall obtained on a autosampling basis
(Glass et al. 1991) and in a total deposition mode (Lindqvist et al. 1991). The unattended
collection approach is risky, requires preservatives, and often produces artifacts. Pyrex glass
and Teflon collectors have been used quite successfully (e.g., Fitzgerald et al. 1991, Iverfeldt
1991a,b, Mason et al. 1992). A light funnel constructed from a molded TeflonR sheet, and
contained in an acrylic housing, with a removable acrylic lid is described in Fitzgerald et al.
(1992), and in Mason et al. (1992). It is designed so that rain entering the funnel contacts only
TeflonR parts which were rigorously acid-cleaned prior to use. The sampling protocol should
include a continuous program of funnel washings and blanks to insure the integrity of each
collection.
The following protocols are used by Fitzgerald and co-workers for rain and snow in the
Mercury in Temperate Lakes (MTL) study. The ultra-clean and laboratory prepared rain
apparatus (acrylic housing containing the Teflon" funnel) is placed in an appropriately selected
field site. The sampler had been sealed by large clean poly-bags as part of the laboratory
preparation. Before sampling, and from a downwind position, the rain collector, wearing clean
rainsuits and long poly gloves, removes and carefully stores the bags and acrylic lid. The funnel
is rinsed with about 1 L Q water (subboiled distilled low Hg water), from a 2 L TeflonR bottle
that will be used as the collection bottle. The second 1-L aliquot of water is used to take a funnel
81
-------
Relative Atmospheric Loadings... Revision Date: IS March 1993
blank. Half of the aliquot is rinsed through the funnel and collected and the other half is retained
for analysis without passing through the funnel. As noted, these samples are used to test and
monitor the funnels for artifacts and potential contamination. In general, there is no discernable
difference between the concentration of Hg in the rinse water and the funnel blank. The funnel
is exposed to the atmosphere for the duration of the event. Following the sampling, the rain
collectors are carefully covered. An aliquot of rain for ethylation/speciation determinations is
decanted into a 100 mL bottle prior to sample acidification. The 100 mL aliquot is frozen and
stored for analysis.
Snow samples are taken by scooping snow directly into 1 L TeflonR jars as soon as
possible after a significant snowfall (at least three inches of fresh snow). Great care is taken to
maintain sample integrity during collection and shipment. During snow collection operations,
the sampling personnel wear particle-free nylon suits (reserved for snow collection) and use arm-
length plastic gloves. Snow is collected moving into the wind and away from possible
contamination. The jars are wrenched tight, stored, and transported frozen to await analysis. The
snow samples for HgR and HgT determinations are acidified with analyzed reagent grade HC1 in
a Class 100 clean laboratory, resealed and allowed to thaw at room temperature. The sample for
chemical speciation is allowed to thaw without acidification. The samples are analyzed
immediately after thawing.
Three different procedures are used in the analysis of rain and snow samples and these
measurements provide information on the forms of mercury in precipitation. All procedures rely
on the production of volatile Hg species in solution, which are then purged from solution with
inert, Hg-free gas and trapped on adsorbing substrate. The determination of "reactive" or "acid-
labile" mercury (= Hgg) involves sample acidification to pH = 1, reduction of ionic Hg and
labile Hg to Hg° with SnCl2, aeration of the solution and collection on Au (Gill and Fitzgerald
1987b). The determination of "total" Hg (HgT) is similar to the reactive Hg procedure after
sample pretreatment with a strong oxidant, BrCl, followed by reduction of the BrCl with
NH2OH-HCL, before SnCl2 reduction, sparging and collection on Au (Bloom and Crecelius
1983). Oxidation of the solution with BrCl destroys many strong organo-metal associations and
decomposes monomethyl mercury, rendering bound Hg available for SnCl2 reduction. The
procedural blank for the reactive Hg determination is 0.025 ± 0.010 ng and the sample size is
generally 250 mL, resulting in a detection limit (defined as 3 x the standard deviation of the
blank) of 0.5 pM. The blank associated with the oxidation technique is 0.1 ng ± 0.03 ng for
a 250 mL sample with a corresponding detection limit of 2 pM. The operationally defined
species of Hg based on the wet digestion and reduction/sparging procedures are summarized in
Figure 26 (adapted from Lindqvist et al 1991). The operational definitions are given in the
figure; we note that Hg-IIa is identical to HgR as defined in the MTL studies.
Quantification and identification of mercury species in precipitation involves the ethylation
of dissolved Hg in solution using sodium tetraethylborate. The volatile ethyl-Hg derivatives, as
well as other volatile Hg species in solution (i.e. Hg° and dimethylmercury) are purged from
82
-------
Decreasing reducibiiity
Hg-llb
Hg-llc
Species
Hg-tol
Hg-n
Hg-IIa
Hg-IIb
Hg-IIc
Analyzed
Method
BrCl->SnCl2
Calculated
HCl->SnCl2
= Hg-II-Hg-IIa
= Hg-tot - Hg-fl
Other labels
total
reactive+non-reacu ve
reactive, acid labile, inorganic
non-reactive
inert
Figure 26. Operationally defined species of Hg based on the wet
'digestion and reduction/sparging procedures (adapted
from Lindqvist et al., 1991).
-------
Relative Atmospheric Loadings... Revision Date 15 March 1993
solution, concentrated on CarbotrapR, separated by cryogenic GC, and detected by atomic
fluorescence. The details of the procedure are outlined in Bloom (1989). This method allows
for the identification and measurement of methylmercury, as methylethylmercury, labile Hg(II),
as diethylmercury, and dimethylmercury and elemental mercury without ethylation. The
detection limits are 0.05 pM for methylmercury, and 0.1 pM for Hg(II).
7.3.3 Dissolved Gaseous Mercury in Natural Waters
Geochemical and analytical details associated with investigations the cycling of volatile
Hg can be found in Vandal et al (1991), Fitzgerald et al (1992), Mason et al (1992), Xaio et
al 1991). Briefly, lake water and seawater are collected with TeflonR coated Go-FloR sampling
bottles suspended from Kevlar11 hydrographic line, using plastic or TeflonR weights and
messengers. The bottle integrates ca. 1 m depth range. The water is analyzed immediately (ca.
1 to 3 hr) in the clean laboratory, and the samples are maintained near their in-situ temperature
until analysis. Dissolved gaseous mercury (DGM) measurements are made by sparging the
volatile species from solution using Hg-free argon and trapping on either Au or CarbotrapR
without pretreatment of the sample. The collection of Hg on Au allows for the measurement
of total DGM, and CTS collection allows for separation and identification of the volatile
dissolved species by cryogenic GC with AFS detection. The samples are purged in a 2-L pyrex
bubbler, and a total sample volume of 4-L is used for each determination. The detection limit
for total DGM is 5fM (femtomolar) and 3fM for dimethylmercury.
7.4 NITROGEN
7.4.1 Atmospheric Sampling
Measurement of ambient concentrations of nitrogen species can be divided into two groups
based on the reactivity of the nitrogen species. Highly reactive species, such as NO, NO^ and
PAN, which have lifetimes on the order of minutes to hours, must be measured in situ. The
more stable constituents, such as HNO3 and paniculate N03 are typically collected on filters and
subsequently analyzed.
A recent review of five major techniques for measuring gas-phase atmospheric NH:,
(Williams et al. 1992) suggests that such measurements can be reliably and accurately
performed. However, a similar review of 18 techniques for determining atmospheric HNO3
concentrations (Hering et al. 1988) indicated a general lack of agreement between the methods
(by as much as a factor of four). Very little is known about the deposition of other potentially
important airborne nitrogen species, especially NO2 and PAN (peroxyacetyl nitrate). Thus, for
the major nitrogen dry deposition, there would appear to be considerable uncertainty in
fundamental measurement.
84
-------
Relative Atmospheric Loadings... Revision Date: 15 March 1993
7.4.2 Precipitation Sampling
The procedures for determining wet deposition are more established, technically less
difficult and inherently less uncertain than those for dry deposition. Consequently, in
constructing mass balances, estimates of wet loading are fairly well constrained. Largely the
result of intense investigations of acid rain, reliable, automated wet-only samplers have been
developed for the collection of precipitation for chemical analysis (Galloway and Likens 1978).
Similarly, techniques for quantifying the major inorganic nitrogen components in precipitation
are common (Technicon 1973, DIONEX 1981).
One uncertainty in estimating nitrogen wet deposition is in integrating the spatial and
temporal variability. For example, data compiled by Jaworski et al. (1992) for atmospheric
nitrogen deposition at ten sites in the Chesapeake Bay watershed varied from 5.8 to 18.1
kg/ha/yr. The data of Fisher et al (1988) varies from 5.68 to 8.49 kg/ha/yr. A similar
assessment for six sites in the Delaware Estuary watershed (Scudlark and Church 1993) revealed
a 50% variability in the MM/ wet flux and 30% variability in the NCy wet flux.
8.0 CASE STUDIES
8.1 MASS BALANCE OF TRACE ELEMENTS IN ESTUARIES
Estimating the relative atmospheric input and throughput of toxics in an estuary requires
some assumptions which has severely limited the accuracy of such estimates in the past. For
trace elements, wet and dry deposition to entering the watershed can be largely assimilated by
vegetation and soils. Dry fall out accumulating in forest canopies must await washout. The
canopy itself can be a source or sink of trace elements, and vegetative ligands may complex and
solubilize trace elements (Lindberg and Turner 1988). Likewise, watershed deposition either
runs off with secondary weathering components or largely enters the local aquifer to become
a ground water component of base flow in streams. Thus, the proportion of atmospherically-
derived trace elements which actually cross the fall line into estuarine waters cannot be easily
deconvoluted without special studies and the use of tracers. In any case, accurate long term
fluvial trace element data are essential. This appears to be beyond the current capability of
large-scale monitoring programs (Windom et al. 1991).
Much of the flux of trace elements into estuaries occurs during episodic or seasonally
short periods of time that correspond to large storm events or spring freshets. Conditions of
measurement during these periods of time are often difficult, if not impossible. As such,
estuarine mass balances assume steady state between input and output fluxes which are not valid
during transient periods of episodic input or seasonal variations. Under such conditions, non-
steady state models based on continuity are required.
85
-------
Relative Atmospheric Loadings... Revision Date: IS March 1993
A mass balance of trace elements in the Delaware estuary has been performed using a
number of estimates and assumptions (Church 1988). The trace element sources were taken
from the following data sets. The primary fluvial flux was calculated as the seasonally averaged
concentration at the zero salinity end member times the riverine discharge. In this sense, any
ground water fluxes downstream of this point were ignored. Some unknown quantity of the
fluvial flux actually includes atmospheric fallout into the watershed. The secondary source from
surrounding salt marshes was based on the non-conservative maximum concentration in a
representative tidal creek times the net exchange volume of the tidal wedge integrated over the
tidal cycle. The primary components of this flux come from both atmospheric fallout into the
marsh watershed and from the diagenetic release of sedimentary components. Such components
can enter the salt marsh with both upstream terrestrial and downstream marine paniculate matter.
The oceanic tidal inputs to the bay are hard to quantify but have been attempted using a two
layer model and salt balance (Church et al 1986). Using such an approach, one can close the
balance between the trace element sources (rivers and salt marshes) and the sinks (sediment
burial and oceanic export) from the Delaware estuary within a factor two. It is thought that the
major unknown source in this balance may be that which enters from ungauged ground water
whose concentration and flux may be as great as that in the gauged fluvial sources.
The calculations of the percentage atmospheric sources for the Delaware estuary (Table
9) show that there is an net excess (atmospheric versus fluvial sources) of trace elements falling
into the watershed versus that which crosses the fall-line. The exception is manganese which
shows "negative excess" indicating other sources, which may include weathering, vegetative
sources, and benthic flux from the sediments. The amount of trace elements falling directly onto
the surface water of the Delaware Bay is small relative to the amount entering by fluvial means.
However, as stated above, the fluvial and salt marsh component may include both atmospheric
and groundwater components. Again, estimating tidal oceanic inputs, which requires accurate
estimates of residual circulation are the largest source of uncertainty in estuarine trace element
budgets.
8.2 MASS BALANCE OF SOCS: A PCB BUDGET FOR LAKE SUPERIOR, 1986
The components of a mass balance in aquatic systems such as the Great Lakes are shown
in the Figure 1. A mass balance constructed about the water column provides an understanding
of the transport and distribution of the chemicals in the lake, an estimate of the residence time
that a contaminant is in the ecosystem or any of its parts, and the essential framework for
determining the relative importance of various input or output sources. Of the potential
hydrophobic organic chemicals for which the mass balance can be demonstrated, PCBs have
been studied the most because of their bioaccumulation, persistence, ubiquitous distribution in
the environment, and alleged toxicity. Eisenreich and co-workers (e.g., Eisenreich, 1987; Baker
and Eisenreich, 1990; Baker et al, 1991; Jeremiason et al. 1993) have accumulated sufficient
•
86
-------
Relative Atmospheric Loadings... Revision Date: 15 March 1993
information on Lake Superior that a PCB budget may be constructed and is, therefore, a good
example of the paradigm. Lake Superior is the second largest lake on earth after Lake Baikal,
is the largest of the Great Lakes possessing >50% of its water volume and approximately 20%
of the surface freshwater on earth, has a large lake area to watershed ratio, has a long water
residence time of - 170 years, is oligotrophic, and is driven primarily by atmospheric
interactions (for PCBs). Inputs to the lake include riverine flows (includes municipal/industrial
discharges) and atmospheric deposition. PCBs may be lost from the lake by riverine flow (St.
Mary's River), sedimentation, degradation, and volatilization.
The inventories of PCBs in the Lake Superior ecosystem are estimated to be:
atmosphere: ~ 200 kg (- 1.2 ng/m3)
water column: ~ 7200 kg (~ 0.6 ng/L)
sediment: -5000 kg (~6 ng/cm2)
Riverine inflow and outflow are estimated to contribute 20 to 50 kg/yr and 40 kg/yr,
respectively, to the mass balance. Eisenreich and Strachan (1992) estimate that atmospheric
deposition of PCBs in the late 1980's was -167 kg/yr, 125 kg/yr in wet deposition, and 32
kg/yr in dry particle deposition. The burial of PCBs in bottom sediments is - 10 to 50 kg/yr
based on detailed analysis of Pb-210 dated sediment cores over the whole lake (Eisenreich, 1987;
Baker et al., 1991; Jeremiason et al., 1993). The assumption is made that chemical and
biological degradation reactions are negligible in the mass balance. PCBs are lost from the lake
by volatilization at a rate of about 600 to 4200 kg/yr (Baker et al. 1990) based on air-water
gradients and estimated mass transfer coefficients. Swackhamer et al. (1988) estimated PCB
volatilization from Siskiwit Lake on Isle Royale in Lake Superior to be about 720 kg/yr.
Measurements of water column PCBs since 1978 suggest a linear loss rate of — 800 kg/yr (1.3
ng/L in 1978 to 0.18 ng/L in 1992). Using the decrease of PCB concentrations in the water
column in the mass balance (below) suggests volatilization is about 670 to 750 kg/yr. Assuming
the mass budget is balanced by volatilization, then:
INPUTS-OUTPUTS-WATER COLUMN LOSS = VOLATILIZATION = 670 - 750 kg/yr.
According to this mass balance calculation, atmospheric deposition contributes 77% to 89% of
1986 inputs, similar to the earlier calculations of Strachan and Eisenreich (1988). PCB Losses
from the lake occur primarily by volatilization which represents nearly 90% of total losses;
sedimentation represents only about 5 %. This funding is consistent with the earlier calculations
of Strachan and Eisenreich (1988), is near the lower end of that estimated by Baker and
Eisenreich (1988), and about equal to the estimate of Swackhamer et al. (1988) based on their
Siskiwit Lake studies. Given the magnitude and uncertainty of the field measured volatilization
rates, this process represents a critical need in the relative loading paradigm.
The estimated residence time (1st order) of PCBs in the water column of Lake Superior
87
-------
Relative Atmospheric Loadings... Revision Date: 15 March 1993
based on the decrease in concentrations over the last 10 to 15 years is 5 to 6 years (Jeremiason
et al., 1993). The majority of the decrease in water column concentrations is attributed to
volatilization, the rate of which will decrease with decreasing water concentrations. Based on
an ecosystem loss rate of 850 kg/yr and an ecosystem inventory of 12,400 kg, a steady state
residence time is about 15 years. The system is, of course, not at steady state and the overall
system response can only be correctly calculated using dynamic models showing changes in
inputs, outputs, and inventories with time.
The mass balance paradigm is a necessary framework to estimate relative loadings of
chemical constituents to lakes and estuaries. To correctly do so requires the measurement of
concentrations, inventories, and fluxes over time in a precise and accurate manner to statistically
demonstrate differences in absolute and relative loadings.
8.3 MERCURY MASS BALANCES
8.3.1 Wisconsin Seepage Lakes
The tropospheric cycling, deposition and air-water exchange of mercury are being
investigated in the mid-continental lake region of Vilas County, northcentral Wisconsin
(Fitzgerald et al, 1991, 1992, Vandal et al. 1991). The work is part of a multidisciplinary study
of processes regulating the aquatic biogeochemistry of Hg in temperate lakes. The atmospheric
Hg flux data were evaluated in a well constrained mass balance Hg budget that was developed
for a representative seepage lake, Little Rock Lake (LRL). Little Rock Lake is an extensively
studied clear water system that has been divided with a sea curtain into two basins, one of which
was untreated (reference pH: 6.1) while the other (treatment) was experimentally acidified. The
first year's results for the treatment basin are summarized in Figure 7, from Fitzgerald et al.
(1991), and as adapted from MTL work to appear in Watras et al (1993), and from preliminary-
budgets by Fitzgerald and Watras (1989), and Weiner et al (1990).
The estimates of annual depositional fluxes noted for rain (4.5±2.0 ^g nr2 yr1), snow
(2.3+0.3 Mg nr2 y"1) and dry deposition (3.5±3.0 /*g nr2 yr1) yield a total Hg deposition of
10.3±3.6 fig nv2 yr1 for the temperate lake environs. This budget shows that the measured total
atmospheric Hg deposition accounts readily for the total mass of Hg in fish, water and
accumulating in the sediments of Little Rock Lake. In this mass balance, the data for the fish
and sediment were taken from Weiner et al (1990) and the data for the water and seston were
obtained from Watras et al (1993). The budget is balanced and the atmospheric Hg exchange
well constrained because the net deposition is balanced by the estimated accumulation of Hg in
the sediments. In addition, the approximate net accumulation of Hg in the biota of LRL (ca. 0.06
g, assuming a 40% turnover) can be supplied by < 10% of the annual Hg deposition.
Fitzgerald et al, (1991) note that "the ecosystem appears delicately poised with respect to
atmospheric inputs, since a relatively small fraction of the input readily accounts for the
-------
Relative Atmospheric Loadings... Revision Date: 15 March 1993
estimated accumulation of Hg in fish." It is noteworthy that atmospheric deposition of total Hg
to the LRL environs is comparable to the estimated depositional flux of Hg at 10 /xg nr2 yr1 for
the northeast Pacific Ocean (Fitzgerald, 1989).
Gaseous Hg the atmosphere and water at LRL was found to be principally Hg° (Vandal
et al., 1991). As shown in Figure 7, the evasional losses of Hg° are geochemically significant
accounting for about 7% of the input to the treatment basin of LRL. Vandal et al (1991)
demonstrated the biogeochemical importance of Hg° evasion on a broader basis using
experimental results from six temperate lakes including the reference and treatment basins of
LRL. Seasonal variations were observed with the highest levels of supersaturation for Hg°
occurring during the peak stratification period in August. A direct relationship between pH and
the degree of saturation for Hg° was also indicated. For example, the evasionr! flux of Hg°
from the reference basin of LRL (1.5+0.9 /*g nr2 yr-1) is about twice the value estimated for
the treatment basin. Further, Fitzgerald, et al. (1991) proposed a reactive Hg(II) substrate
hypothesis, in which they hypothesized that in-lake biological and chemical production processes
for Hg° and monomethyl mercury were in competition for the reactive Hg substrate, which was
suggested to be labile Hg (II) species (Hgn). They postulated that once Hg° is produced in the
aqueous phase, it is unreactive and eventually lost from the system. Thus, lakes with
limnological conditions favoring Hg° production would be less likely to have elevated levels of
Hg in the fish stock. In-lake production and water-air losses of Hg° might function as a potential
amelioration mechanism that reduces the HgR available for methylation. Conversely, higher
biological levels of Hg may occur in lakes where Hg° production processes are inhibited by
increased acidity.
Atmospheric depositional fluxes of monomethyl mercury (MMHg) were assessed in a less
refined budget for the treatment basin of LRL (Fitzgerald et al. 1991). The atmospheric inputs
were estimated using mean concentrations of MMHg observed in snow and rain, yielding 66±28
and 22±8 ng m'2 yr'1, respectively, for an annual flux of 88±29 ng/m2 -yr. While no dry
deposition information for MMHg was available, the preliminary data for wet deposition
suggested that atmospheric depositional fluxes were insufficient to account for the amounts of
MMHg observed in biota. For example, assuming that the MMHg inputs from the atmospheric
show a wet/dry partitioning similar to total Hg (i.e., 66% wet and 34% dry), then approximately
0.013 g of MMHg would be delivered to the treatment basin annually. This flux is about 22%
of the estimated yearly MMHg accumulation in the fish stock. An in-lake synthesis of MMHg
is implicated. This mass balance budget is illustrated in Figure 8, which has been developed
using information from Fitzgerald et al. (1991), Weiner et al. (1990), Hurley et al. (1991a,b),
Bloom et al (1991), and Watras et al (1993).
Additional data for atmospheric deposition of Hg in the Wisconsin MTL study have been
reported by Fitzgerald and his co-workers (Fitzgerald et al, 1992). These data are summarized
in Tables 25 and 26. The two year average broadens the basis for estimating Hg deposition in
these environs. The inputs of HgT and MMHg determined for the second year of the study are
89
-------
TABLE 25
ANNUAL Ht DEPOSITIONAL FLUXES IN NORTHCENTRAL WISCONSIN
BETWEEN OCTOBER 1988 AND OCTOBER 1990 (FROM FITZGERALD et al. 1992)
He Species
Total
Reactive
Methyl
Wet Deposition we m"2 v'1
1988/89
6.8 ± 2.0
2.5 ± 1.3
0.09 ± 0.03
1989/90
8.7 ± 3.7
7.1 ±3.5
0.07 ± 0.03
Dry Deposition we m'2 v'1
1988/89
3.5 ±3.0
No data
No data
1989/90
3.9 ±3.8'
No data
No data
"tstimated from a yearly average or 25± 23 pg m'-* obtained from the measurements made
in the winter and spring of 1989 and the summer data for 1990. A depositional velocity of
0.5 cm sec'1 is used (Fitzgerald et al. 1991).
-------
TABLE 26
AVERAGE ANNUAL He DEPOSITION TO LITTLE ROCK LAKE, WISCONSIN
DURING OCTOBER 1988 TO 1990 (*DRY DEPOSITION NOT INCLUDED)
Hg Species
Total
Reactive
Methyl
Annual
Oct
Deposition -u% m"2 v"1
1988 to Oct 1990
11.5 ±3.2
4
0
.8 ±1.6'
.08 ± 0.02'
-------
Relative Atmospheric Loadings... Revision Date: 15 March 1993
virtually the same as observed during the initial work. There is a significant difference in the
deposition of HgR to the lake region between years and the authors demonstrate that differences
in HgR inputs can have profound effects on the Hg° cycle. The role of atmospheric speciation
and its relationship to the aquatic biogeochemical cycling of Hg° and MMHg is considered in
Section 8.3.3 below.
8.3.2 Drainage Lakes in Sweden
Mercury fluxes to and from a model lake in southern half of Sweden are illustrated in
Figure 9, which has been adapted from Johansson et al. (1991). The lake area is 1 km2 and the
drainage area is 10 km2, and this proportion is representative of the average ratio for Swedish
lakes. Wet deposition of Hg is estimated at 20 /ig nr2 yr1, about 2.5 times the LRL value of 8
^g nr2 yr1 (Table 25). The main source of Hg to the lake is via atmospheric deposition. As
summarized in Figure 9, about 5 to 30% of the Hg input from the atmosphere (200 g yr1) to
the catchment will enter the lake through runoff and this waterborne flux (10 - 60 g yr1) will
represent between 50 and 100% of the input from wet deposition. The sediment accumulation
of mercury will depend on size of the lake, its biological and chemical character, and on
evasional losses (2 to 20 g yr1) of Hg at the water surface, which are estimated to range from
10 to 100% of the direct wet deposition of Hg (20 g yr1) to the lake.
There is broad agreement between the Swedish work and the results from the MTL
program in Wisconsin. For example, simulation of the Hg flows into and out of a typical
Swedish lake clearly demonstrates that atmospheric Hg deposition is the preeminent source of
Hg to a drainage lake, and that evasional fluxes of Hg° are significant, although the estimates
require refinement. One striking difference between the drainage and seepage lake modelling is
the significant portion of the Hg input that is stored in forest soils of the catchment. On
average, present atmospheric deposition is greater than the output of Hg in run-off waters by
about a factor of 10. Thus, even if anthropogenic Hg inputs were to cease, modern Hg
deposition that has accumulated in the soil would continue to be released to the lakes from the
forest soils. Indeed, Johansson and co-workers indicate that 70 to 80% of the Hg in the
catchment is anthropogenic (Lindqvist et al. 1991), and as a consequence, the watershed
transport of Hg to the lakes will remain elevated for long periods of time, perhaps several
centuries.
8.3.3 Atmospheric Mercury Speciation: Biogeochemical Implications
A biogeochemical coupling between HgR in atmospheric deposition and the Hg° cycle in
lakes has been found in the MTL Wisconsin investigation. As discussed (Table 19), there was
a substantial difference in the HgR composition in wet deposition between 1988-1989 and 1989-
1990 and the associated yearly supply of HgR to the lakes. A comparison of the estimated inputs
92
-------
Relative Atmospheric Loadings... Revision Date: 15 March 1993
was shown in Table 25. Approximately 3 times more active substrate was introduced in 1989-
1990 relative to 1988-1989, yet the HgT inputs between years were similar. Indeed, a direct
linkage between deposition of HgR and in-lake values of Hg° can be demonstrated from the lake-
air experiments conducted in the summers of 1989 and 1990. The biogeochemical significance
of the in-lake cycling and evasional losses of Hg° was shown by Vandal et al., (1991). In
general, the dissolved gaseous Hg (DGM) fraction in the Wisconsin seepage lakes consists
principally of Hg° with no significant contribution from volatile organic Hg species, (i.e.,
dimethylHg [detection limit of 3 femtomolar (fM)]). Experimentally, the air-water partitioning
of Hg° is determined from simultaneous measurements of Hg° in the atmosphere and in the
lakes. The degree of saturation (%S) for Hg° in lake water relative to the appropriate
temperature-corrected equilibrium with the atmosphere (Sanemasa, 1975) is determined using
the following relationship:
%S = [ (Cwller x H)/ CJ x 100 [13]
%S > 100 = supersaturation in water [14]
where H = Henry's Law Constant for Hg° and Cwaler and C^ = the concentration of Hg° in
water and air, respectively. A summary of the 1989 and 1990 Hg° data from August studies in
the Wisconsin seepage lakes is given in Table 27 (Fitzgerald et al. 1992). Notice that in 1990,
the lakes were highly supersaturated with values near the surface ranging from 10.4 (S =
1040%) in Max Lake to 44.6 (S = 4460%) times the equilibrium level in Crystal Lake. In
August 1989, supersaturation ranges were from ca. 1.4 to 12 times the saturation concentration.
Thus, these large %S values will translate into higher lake to atmosphere fluxes of Hg°than
reported previously (Vandal et al. 1991). The water-air transfer of Hg° is estimated from the
thin film gas - exchange model using the following relationship.
F = K (Clir H'1 - CW1(J [14]
where, F = gaseous Hg flux into (+) or out of (-) the lake, Ciir = air concentration of Hg°,
Cw.ter = water concentration of Hg°, H = Henry's Law Constant, and K = transfer velocity,
1.5 cm hr1 (0.36 m day1) for August, 1989 and 1990 (21°C to 24°C). Additional details for
lake-air exchange of Hg° calculations are given in Vandal et al. (1991).
A demonstration of the biogeochemical importance and dynamic nature of in-lake Hg°
production appears in Table 28, where the evasional fluxes from the northcentral Wisconsin
study lakes in August 1989 are contrasted with the August 1990 results. In general, with the
exception of Max Lake, effluxes of Hg° are ca. 2 to 6 larger in 1990 and interlake differences
are significant. Further, the atmospheric deposition of HgR was approximately three times
93
-------
TABLE 27
DEGREE OF SATURATION FOR ELEMENTAL MERCURY
IN NORTHCENTRAL WISCONSIN LAKES
Lake
Little Rock-Treat.
Little Rock-Ref.
Pallette Lake
Vandercook Lake
Max Lake
Little Rock-Treat.
Little Rock-Ref.
Pallette Lake
Vandercook Lake
Max Lake
Crystal Lake
Russett Lake
Date
August, 1989
August, 1990
Hg°(fM)
Range
83 - 107
135 - 200
60 - 355
85 - 163
283 - 297
181 - 490
214 - 358
92 - 640
281 - 570
182 - 546
90 - 785
179 - 1035
S(%)
Range
305 - 345
500 - 740
140 - 1180
315 - 600
990- 1100
920 - 2790
1220 - 2040
340 - 3650
1630 - 3250
1040-3110
350 - 4460
660 - 4060
S (%) Mean
and Std. Dev.
325 ± 28
620 ± 170
605 ± 530
458 ± 202
1045 ± 78
1703 ± 971
1536 ± 441
1476 ± 1370
2440 ± 810
2075 ± 1035
1583 ± 1953
2033 ± 1484
-------
TABLE 28
ESTIMATED AVERAGE EVASIONAL FLUXES FOR
AUGUST 1989 AND AUGUST 1990 FOR VARIOUS NORTHCENTRAL
WISCONSIN LAKES. FLUXES ARE IN pmol nr2 day1, CALCULATED
USING A TRANSFER VELOCITY OF 1.5 cm hr1 (0.36 m day"1)
Lake
Little Rock-Treat.
Little Rock-Ref.
Pallette
Vandercook
Max
Russett
Crystal
Average
Depositional Flux (Hgn)
Depositional Flux (HgT)
Evasional Flux
in August 1989
(pmol/m2 day)
25
50
85
50
98
-
-
62 ± 30
85
304
Evasional Flux
in August 1990
(pmol/m2 day)
167
120
221
92
57
143
274
153 ± 75
260
301
-------
Relative Atmospheric Loadings... Revision Date: 15 March 1993
greater in August, 1990 relative to August, 1989 (Table 28). The HgR input during August
1990 (260 pmol nr'-day1) was much higher than that of August 1989 (85 pmol nr2 day1). Yet
the total Hg (HgT) input was similar for both periods: 301 pmol nv2 day1 versus 305 pmol nr2
day1 (Table 28). A relationship between Hg° evasion and the input of HgR is apparent, as the
increases in Hg° supersaturation appear to respond to the supply of HgR and not to HgT.
The principal source(s) of the high levels of supersaturation in the surface waters is not
yet known. However, photocatalytic reduction of ionic Hg, bacterial demethylation reactions,
organic matter and/or photosynthetic processes may all play a role. Fitzgerald et al. (1992)
considered the two most probable sources of Hg° (demethylation and direct reduction) in detail,
and the results are illustrated in Figure 27, for (1) the estimated depositional HgR input and
evasional fluxes of Hg° for Little Rock Lake, (Treatment Basin) and Pallette Lake, respectively,
in August 1989 and 1990, and (2) the probable sources of Hg° - demethylation of Hg° and direct
reduction of Hg(II). The amounts of MMHg formed in the mixed layer was estimated using the
measured epilimnetic HgR concentrations in August, 1989 (Bloom, pers. comm.) and
measurements in 1990, along with water column methylation rates of 0.01'- 0.3% day1,
determined from laboratory spike experiments (Xun et al., 1987; Korthals and Winfrey, 1987;
Gilmour and Henry, 1991). The rate of Hg° formation by demethylation of MMHg must be less
than the MMHg formation rate. Thus, the calculations show that demethylation is a minor
source of Hg° in the epilimnion of the study lakes. Direct reduction of Hg(II) must be the
primary source of Hg° in the epilimnion of these lakes; a similar situation was found for the
mixed layer of the equatorial Pacific (Mason and Fitzgerald 1992).
The observed evasion in Little Rock Lake Treatment Basin could be maintained by Hg°
formation rates of 8 x 10"7 sec'1 (7% day1) in 1989 and 3.3 x 10"7 sec'1 (2.8% day1) for 1990.
At Pallette Lake, the observed evasion would require Hg° formation rates of 3.3 x 10"6 sec"1
(28% day1) in 1989 and 1.4 x 10"7 sec"1 (1.2% day1) for 1990. Abiotic production of Hg° (25
x 10"7 sec"1; 22% day1) in the presence of humic acids has been demonstrated in laboratory
studies (Alberts et al., 1974). Moreover, the reaction rates indicate that the system response to
HgR input is rapid with a pulse input of Hg(II) converted to Hg° in the absence of other reactions
in about 70 days at a conversion rate of 5% day1 (t'/2 = 14 days).
Although the Hg° data is limited to one set of measurements at each lake in each season,
these results support the postulate that the production, and subsequent evasion of Hg°, are
directly linked to the rate of supply of HgR. The reactive Hg concentration, therefore, is a
measure of the readily available substrate. These rates of conversion are similar to those
estimated for the equatorial Pacific (Mason and Fitzgerald 1992), suggesting that analogous
processes are involved in these two systems. The similarity between the HgT depositional inputs
for both seasons suggests further that the unreactive, strongly bound Hg fraction is not directly
available for conversion into Hg°. It is likely that the strongly bound fraction is transported into
the anoxic regions of the lakes (the hypolimnion) or the sediment before any remobilization into
a reactive form that can be converted into other Hg species.
95
-------
AIR PALLETTE LAKE, 8/89
Hg(ll) DEPOSITION
WET: 70
DRY: 15
Hg° EVASION
85
MMHg
PARTICULATE FLUX * ca. 0 (by diff.)
THERMOCLINE
ALL FLUXES IN pmol/m day
AIR PALLETTE LAKE, 8/90
Hg(ll)'DEPOSITION
WET: 215
DRY: 45
Hg° EVASION
221
MMHg
PARTICULATE FLUX
39
-------
LITTLE ROCK TREATMENT. 8/89
Hg(ll) DEPOSITION
WET: 70
DRY: 15
I
Hg° EVASION
25
<1
MMHg
<1
Hg(l|) > Hg'
>24
PARTICULATE FLUX
60 (by diff.)
THERMOCL1NE
ALL FLUXES IN pmol/m day
LITTLE ROCK TREATMENT, 8/90
Hg(ll) DEPOSITION
WET: 21 5
DRY: 45
Hg° EVASION
167
MMHg
PARTICULATE FLUX +
(by diff.)
THERMOCLINE
ALL FLUXES IN pmol/m2 day
Figure 27a. Modelling the potential pathways for the production and evasion of Hg In
epilimnion of the treatment basin of Little Rock Lake, Wisconsin. The
amounts of Hg° produced by demethylation and direct reduction of Hg(II) are
estimated and related to the input of HgR by atmospheric deposition for the
August 1989 and August 1990 experiments.
-------
Relative Atmospheric Loadings... Revision Date: 15 March 1993
In the Wisconsin lakes, there was no significant increase in the Hg° concentration during
winter ice cover (Vandal, pers. comm.). Iverfeldt (1988) attributed differences in the
concentrations of dissolved gaseous Hg, reactive, and total Hg in Framvaren fjord between
September 1983 and February 1985 to the lack of atmospheric input during ice cover. While
the lack of Hg° formation under the ice could be interpreted as evidence of Hg° production being
associated with a process requiring light (i.e. primary production or photo-reduction), it is also
suggestive of substrate limitation in winter through the lack of atmospheric input. The results
from these studies, Iverfeldt's work (1988) and Mason and Fitzgerald (1992) indicate that
formation of Hg° in the mixed layer of natural systems is a direct function of the rate of supply
of available Hg substrate.
In Table 28, the mass balances for the Wisconsin lakes show that with the exception of
Max Lake, atmospheric inputs of HgR and HgT were larger than the evasional losses of Hg°
during the August studies. Other processes removing (or sequestering) HgR from the mixed
layer limit the available substrate. Paniculate uptake and removal from the epilimnion is an
important removal mechanism that can deliver substrate for methylation to the hypolimnion or
the sediment interface. Net paniculate fluxes required to balance the increase in Hg in the
hypolimnion of Little Rock Lake Treatment Basin during summer stratification were estimated
to be 55 pmol nr2 day1 for the summer of 1989 (Hurley et al. 1991). This flux most likely
accounts for the difference between atmospheric input and Hg° evasion and suggests that this
process is a primary competing removal mechanism for epilimnetic Hg(II) substrate.
It is evident that the production and evasion of Hg° in natural waters is a major feature
of the aquatic biogeochemical cycling of Hg. Significant effluxes of Hg° have been observed in
seepage lakes in Wisconsin as well as in a diverse range of systems such as the open ocean
equatorial Pacific (Kim and Fitzgerald 1986), Davis Creek Reservoir, California (Gill and
Bruland, 1992), and drainage lakes in Sweden (Xaio et al. 1991). We note that Xaio et al,
made direct flux measurements over soils and lake waters in Sweden, and annual Hg° fluxes
from 2 to 20 ^g nv2 yr'1 were estimated for lake regions and < 1 ng nr2 yr1 for the coniferous
soils investigated. These lake-air fluxes of Hg° are 10 to 100 times the MTL estimates, while
Gill and Bruland estimate Hg° emissions at 21/xg nr2 yr'1. These higher fluxes coincide with the
larger amounts of available Hg in the Swedish lakes and in Davis Creek. Moreover, Mason et
al, (1992) have shown that Hg° emissions to the atmosphere are proportional to the availability
and supply of HgR (the Hg(II) substrate) whether it is atmospherically derived (as in seepage
lakes) or supplied principally through upwelling (as in the equatorial Pacific). It is particularly
striking that a large fraction of the HgR input to the northcentral Wisconsin lakes is returned to
the atmosphere. Indeed, and suggested previously (Fitzgerald et al, 1991), lakes with
limnological conditions favoring Hg° production would be less likely to have elevated levels of
Hg in fish. Moreover, there should be an inverse relationship between Hg° evasion and the
accumulation of Hg in sediments for a particular lake (Rada et al 1993).
Interlake variations in Hg° production and evasion are both expected and observed (Table
99
-------
Relative Atmospheric Loadings... Revision Date: 15 March 1993
28). This is consistent with a general physicochemical view of the Hg(II) substrate hypothesis.
For example, methylating and reducing processes compete to utilize Hg(II) species, and strong
sequestering with organic ligands, or inorganic interactions (e.g., with sulfitic ligands) could
reduce the activity of the substrate. Thus, inorganic and organic components, suspended matter,
pH, and biological productivity within a lake can alter the availability of the substrate. Vandal
et a/., (1991) presented evidence suggesting that increasing acidity may reduce the in-lake
production of Hg°, and that photosynthetic activity may enhance Hg° production. Very few
details of the processes affecting production and destruction of Hg° are known. There are many
questions concerning short-time scale spatial and temporal variability as well as the importance
of photoreduction reactions and redox boundaries (i.e., oxic/anoxic transition zones) in the
production of Hg°. In addition, the relationships among phytoplankton productivity, microbial
populations (e.g., bacterial reduction) and the activity of Hg° should be evaluated. Broadly based
Hg° investigations are required, particularly those including atmospheric speciation research,
ancillary biological studies and concurrent methylation investigations. Seasonal and spatial data
for atmospheric Hg deposition and the evasion of Hg° are limited. This points toward a need to
refine input to and efflux estimates from lake waters and to assess, quantitatively, their
influences on the overall cycling of mercury in lake systems.
These observations illustrate the value of the chemical speciation approach to our
developing understanding of the cycling of Hg in nature. Indeed, they force us to ask and
address the following general question: How do such speciation changes in the depositional
fluxes of Hg affect the cycling of Hg in aquatic systems, and what causes the variation in the
HgT and HgR composition found in deposition? At present, there are no unequivocal answers
to questions concerning the sources and variability of the atmospheric Hg species.
8,3.4 Summary of Mercury Mass Balances
Atmospheric deposition dominates the flux of Hg to lacustrine systems and the open
ocean, and it appears that modest increases in atmospheric Hg loading could lead directly to
enhanced levels of Hg in biota. The U.S. and Nordic studies of current and historical Hg
deposition show broad agreement. Mid-latitudinal preindustrial depositional fluxes of total Hg
were ca. 4 jig m'2 yr1, while present day annual fluxes may vary between ca. 10 ng m'2 yr'1 in
rural semi-remote regions to > ca. 25 jtg nv2 yr1 in places where the presence of local/regional
Hg sources is pronounced. The influence of anthropogenic activities on the total Hg cycling is
evident, and site specific research must be conducted to assess the impact of human-related
interferences in particular localities. However, the more important and subtle concerns are
associated with the physical and chemical speciation of Hg deposition. For example, the presence
of a significant regional particulate Hg cycle is found in specific chemical analysis of Hg in
atmospheric particulate matter and precipitation. Fitzgerald et al., (1991 and 1992) and Iverfeldt
(1991a,b) have shown that significant portion of the total Hg in precipitation and in particulate
matter is non-reactive to reduction with stannous chloride. Thus, a portion of the HgT observed
in precipitation at Little Rock Lake and in comparable Swedish regions, is in a particulate form
100
-------
Relative Atmospheric Loadings... Revision Date: 15 March 1993
which is not derived from the oxidation of Hg° in the atmosphere. Moreover, significant
differences are evident in the deposition of HgR and differences in HgR inputs may have profound
effects on the Hg° and MMHg cycle in natural waters.
8.4 NITROGEN MASS BALANCES IN COASTAL WATERS
8.4.1 Total Nitrogen
Fisher and Oppenheimer (1991) conducted a comprehensive assessment of the contribution
of atmospheric nitrogen deposition to the Chesapeake Bay using two approaches which differed
in their assumptions about watershed retention. In their first approach, non-point loadings to the
bay (atmospheric, animal waste, sewage and fertilizer) were calculated assuming equal retention
in the watershed. Their second approach takes into account the differential retention of nitrogen
based on the different land uses and the differing mobility of nitrogen from each source. These
authors conclude that 25% of the anthropogenic nitrogen loading to Chesapeake Bay is derived
from atmospheric NO3", with another 14% contributed from atmospheric NH,"1". To address the
uncertainties in watershed loading, Tyler (1988) utilized a "transmission factor" to account for
two processes. The first, being the most variable, accounted for retention within the watershed,
which was land-use specific. The second variable accounted for subsequent in-stream removal
due to denitrification. Utilizing this approach, Tyler estimated that atmospheric deposition
contributed 19-25% (with about a factor of two uncertainty) of the nitrogen loading to the
Chesapeake Bay.
Using an approach similar to Tyler's, but with more refined estimates of watershed
retention and in-stream removal, Hinga et al. (1991) contrasted the relative atmospheric loading
to four coastal ecosystems, including the Chesapeake Bay. This study also reported high and
low estimates based on varying assumptions. This sensitivity analysis revealed widely divergent
results (as much as a factor of 50). Overall, their "best estimate" prediction for the Chesapeake
Bay supports the earlier results of Fisher et al. (1988), concluding that atmospheric deposition
supplies 35% of the nitrogen loading. Estimates for other estuaries range from 12%
(Narragansett Bay) to 100% (Ochlockonee Bay), with this difference primarily reflecting
differences in the degree of urbanization and land use.
A major shortcoming common to the above studies is that atmospheric deposition was
evaluated relative to sources only. However, in terms of ecological impact and overall nutrient
dynamics, the throughput of nitrogen, including a consideration of estuarine sinks, is as
important as the gross atmospheric loading. For example, while the nitrogen loading to the
Delaware Bay is estimated to be 10 times greater than the Chesapeake (Nixon et al., 1986), the
assimilatory capacity in the Delaware is considerably greater.
101
-------
Relative Atmospheric Loadings... Revision Date: 15 March 1993
Jaworski et al (1992) have recently attempted a process-oriented mass balance approach
by evaluating nitrogen inputs and outputs to the Potomac River Basin, a sub-estuary of the
Chesapeake Bay. Of the five major source terms estimated based on direct measurements, they
concluded that atmospheric deposition provided 28% of the total nitrogen loading. Scudlark and
Church (1993) have examined atmospheric deposition to the Delaware Bay, a more heavily
urbanized coastal plain estuary adjacent to the Chesapeake Bay. They conclude that atmospheric
deposition provides about 16% of the nitrogen input on an annual basis. However, during late
spring and early summer, when the estuary is most nitrogen limited due to maximum rates of
primary productivity, atmospheric deposition is estimated to provide 25% of the total nitrogen
loading (Figure 28). The greater atmospheric contribution during summer is attributed to
increased atmospheric loading (Figure 28) coupled with minimum fluvial inputs because of low
river flow.
The ecological consequences of atmospheric nitrogen deposition is not limited to estuarine
waters. In fact, it is estimated that as much as 25 % of the nitrogen oxide emissions from North
America are advected eastward over the western Atlantic Ocean (Galloway and Whelpdale
1987), where they may be efficiently scavenged and deposited (Luke and Dickerson 1987). In
coastal (Paerl 1985; Paerl et al. 1990) and Gulf Stream (Willey and Cahoon, 1991) waters,
atmospherically-derived nitrogen inputs have been shown to enhance primary production. It has
also been suggested (Paerl 1988; Fanning 1989) that atmospheric inputs of inorganic nitrogen
is inducing in an ecological shift by oceanic phytoplankton from nitrogen to phosphorus
limitation (presumably with an accompanying shift in species composition as well). This
suggestion is challenged by Jickells et al. (1990).
Assessments of atmospheric deposition of nitrogen to oligotrophic waters in the open
ocean (Paerl 1985, Knap et al. 1986) over an annual time frame suggest that aeolian input
provides only a minor influence compared with upwelling (the primary source of available
nitrogen far from fluvial influence). However, evaluation of this impact on shorter time scales
(Owens et al 1992, Micheals et al. 1992) suggest that during episodic events, atmospheric
deposition can result in a significant fraction of "new production."
From an oceanographic perspective, nutrient input from shelf exchange processes
represents potentially significant source of inorganic nitrogen to estuarine waters which recent
published reports (e.g., Tyler 1988, Fisher et al 1991, Hinga et al. 1991) have neglected to
consider. Hydrologically, classic two-dimensional estuarine exchange involves the net seaward
transport of lower salinity water near the surface, with a compensating inflow of higher salinity
shelf water along the bottom. For many estuaries, such as the Delaware (Galperin and Mellor
1990) and the Hudson (Oey et al 1985), this two-dimensional exchange results in a net flux
(excluding tidal fluxes) of bottom water and associated nutrients into the estuary. Although the
bottom water nitrogen concentration is typically small when compared with that in surface water,
the large volume flux of water into the bay can result in a large nitrogen input. Consequently,
as discussed by Scudlark and Church (1993), such residual circulation may provide a major
102
-------
Annual
Municipal/Industrial 41.0%
Salt Marsh2.0%
Benthic19.0%
Combined LB Rivers 7.0%
Direct Atmospheres.0% ^~**«i Delaware Riven 7.0%
Indirect Atmosphere9.0%
Summer
Municipal/Industrial 49.0%
Combined LB River 4.0%
Benthic14.0% Delaware River7.0%
Direct Atmospheric 10.0% Indirect Atmospheric 16.0%
Figure 28. Seasonal differences in the relative atmospheric loading
of inorganic nitrogen to the Delaware Bay
-------
Relative Atmospheric Loadings... Revision Date: 15 March 1993
source of inorganic nitrogen to estuarine systems not previously considered. For the previously-
cited studies which focus on the role of atmospheric deposition in an estuarine nitrogen budget,
inclusion of a significant shelf exchange term would have serious mass-balance implications.
Thus, until the role of residual circulation is evaluated, the contribution of atmospheric
deposition estimated in these studies should be regarded as relative to landward or continental
sources only, and not to total dissolved inorganic nitrogen input.
The temporal variability in atmospheric deposition, fluvial input and primary production
are largely overlooked but important characteristics of nitrogen loading to coastal and estuarine
waters. For example, the atmospheric input reflected in Figure 28 is based on the long-term
(1978-1989) average precipitation composition and rainfall amount. However, the interannual
variability in wet deposition of nitrate and ammonium for the same period is as much as 51 %.
Therefore, assessments based on deposition from a single year may be misleading. On a
seasonal basis (Figure 29), maximum rates of nitrogen wet deposition are observed during
summer, when primary production in the estuary is most severely nitrogen limited (e.g., D'Elia
et al. 1986, Pennock 1987). Based on recent results from the National Dry Deposition Network,
the dry deposition of nitrogen appears to exhibit a similar, though less pronounced, seasonal
trend (Edgerton et al. 1991).
The episodic nature of wet deposition is revealed by the frequency distribution in Figure
30. While a majority of the events are typically associated with low flux, it is the small number
of exceptionally large events which drive the annual flux. For example, the ten largest episodes
in Figure 30 (about 10% of the total number) account for 38% of the total annual flux.
Furthermore, all of these dominant deposition events occurred during the summer, further
supporting the notion of seasonally varying nitrogen deposition. Despite the distinct temporal
variability in atmospheric nitrogen fluxes, the ecological response may occur on differing time
scales. Fisher et al. (1988) argue that nitrogen inputs during the winter and spring are largely
retained and recycled within the estuary. Similarly, atmospheric nitrogen deposited in the
watershed will accumulate during the summer and be transported in response to the hydrological
cycle (Boring et al. 1988). Thus, residual nitrogen from high input periods may persist in the
estuary and watershed sufficiently to influence annual productivity.
While such seasonally in the atmospheric flux of N03'is noted in both the Chesapeake
Bay (Lynch et al. 1989, Maxwell and Mahn 1987) and the Delaware Bay watersheds (Figure
29), it is not observed at all sites in the eastern U.S. (Calvert et al. 1985). This can be attributed
to either intra-annual variability in regional emissions, or related to the reactivity of NOX,
atmospheric transport and subsequent scavenging of NO3" (Wolff 1984, Lindberg 1982). For
NH/, summer deposition is generally maximum at all eastern U.S. locations, probably
reflecting increased emissions from biogenic sources during warmer weather.
A major limitation in assessing atmospheric nitrogen deposition is that the national wet
and dry deposition monitoring networks (e.g., NADP, NDDN) have been designed to
104
-------
.E
CO
"o
E
£
3-
0
ammonium
nitrate
A M
O N
D
Figure 29. Seasonal variability in nitrogen deposition
at Lewes, DE
-------
0.6 1.2 1.8 2.4 3.0
2
N Wet Deposition (mxnoles/m )
Figure 30. Episodic atmospheric wet deposition of
nitrate and ammonium at Lewes, DE, 1990.
-------
Relative Atmospheric Loadings... Revision Date. 15 March 1993
intentionally exclude bonafide coastal (i.e., shore based) sites in order to exclude complicating
influences of sea-salt aerosols on the species of interest (especially SO4'2). However, deposition
of nitrogen to coastal waters can significantly differ from estimated rates extrapolated from
inland sites by three mechanisms:
1. chemical, due to the scavenging of vapor-phase HNO3 by alkaline marine aerosols;
2. physical, due to the increased gravitational settling of hygroscopic aerosols under higher
relative humidities, and;
3. meteorological, whereby the atmospheric stability regime over water differs greatly from
that over land, due primarily to turbulence resulting from differences in water and air
temperatures.
Recently, the NOAA Atmospheric Nitrogen Input to Coastal Areas (ANICA) program has begun
to address this shortfall in the Chesapeake Bay, including characterization of the atmospheric
stability field over water using buoy-mounted meteorological sensors.
Similarly, atmospheric deposition monitoring networks have been designed so that the
sites are regionally representative of their airshed, typically located in rural or semi-remote
settings. However, this design inherently excludes the influence of large urban centers, many
of which are located along coastal waters. While the magnitude of deposition from the urban
emission plume is largely unknown, it would seem that studies to date which rely on national
and state monitoring data have underestimated atmospheric deposition.
8.4.2 Dissolved Organic Nitrogen (DON)
Along with NH3, organic nitrogen forms such as amines and amino acids can be released
to the atmosphere from the decomposition and volatilization of organic matter. However, due
to the paucity of reliable measurements of organic nitrogen in atmospheric deposition,
investigations of the role of atmospheric deposition of nitrogen in mass balances for aquatic
systems have focussed completely on inorganic forms of nitrogen. However, the few data
available would suggest that atmospherically-derived dissolved organic nitrogen (DON) may
provide a small, albeit significant source of external nitrogen to aquatic systems. For example,
Mopper and Zika (1987) reported an average concentration of DON (dissolved free amino acids
+ aliphatic amines) in the western Atlantic Ocean and Gulf of Mexico of approximately 7/xM.
Likens et al. (1983) detected primary amines in precipitation at Hubbard Brook, New Hampshire
and Ithaca, New York at concentrations averaging 6 /xg C/L. These data agree with observations
of Knap et al. (1986), Jickells et al. (1990) and Jickells et al. (unpublished), who measured a
total (persulfate oxidation) DON concentration in precipitation from the North Sea, Northeast
Atlantic and Bermuda of 6.3-8.7 pM. Similarly, Timperly et al. (1985) found that DON
(primarily urea) contributed substantially to the overall Nitrogen loading in a New Zealand lake.
107
-------
Relative Atmospheric Loadings... Revision Date: 15 March 1993
Furthermore, and perhaps as importantly, they established that such organic species are readily
assimilated by phytoplankton.
These organic nitrogen concentrations can be put in perspective by comparing them with
simultaneously-measured concentrations of the major inorganic nitrogen species (NO3" + NH/)
in precipitation. For the relatively polluted North Sea region, characterized by precipitation
concentrations typical of the eastern U.S., DON contributed 16% of the total nitrogen
deposition. For Bermuda, more representative of coastal waters of the eastern U.S., DON was
found to provide 25% of the total nitrogen wet deposition. Data summarized by Smullen et al.
(1982) for Chesapeake Bay indicate that atmospheric deposition of organic nitrogen (< 6.8
kg/ha/yr) would increase the total nitrogen deposition by 43%. In contrast, other recent studies
of DON in aerosols (Dodd et al. 1984), Rhode Island coastal precipitation (VanNeste et al.
1987), and North Atlantic precipitation (Gorzelska and Galloway 1990), these authors suggest
that the atmospheric deposition of organic-N in the eastern U.S. would be less than 10% of the
total nitrogen deposition.
Based on these limited measurements, it is difficult to resolve the apparent ambiguity. One
simple explanation is that the seemingly divergent results are due to pronounced seasonal and/or
spatial variability. If we conditionally accept a value for the average DON concentration in
precipitation of *7 /xmoles/1 (an approximate value on which a number of the cited studies
appear to converge), and assume that DON behaves similarly to DIN with respect to wet/dry
flux apportionment and watershed retention, the overall atmospheric loading to Delaware Bay
is estimated to increase by about 10-15% (Scudlark and Church 1993). However, since we do
not have reliable estimates of DON input from other estuarine sources, it is not possible to gauge
the impact of atmospheric DON input on an overall nitrogen mass balance. Clearly, a more
accurate assessment of the role of atmospherically-derived organic nitrogen in nutrient budgets
for the "Great Waters" will require more extensive measurements (both spatially and temporally)
of the total concentration and speciation of organic nitrogen compounds in precipitation and
surface waters, as well as identification of their sources.
8.4.3 Estimated Response of Nitrogen Loadings to 1990 Clean Air Act Amendments
Provisions of the 1990 Clean Air Act Amendments will "freeze" NOX emissions at 1990
levels, and presumably not greatly alter the rate of atmospheric nitrogen deposition to coastal
waters. However, in order to meet objectives for the reduction in urban ozone levels, further
emission reductions may be required. In a preliminary analysis, Buckley and Corio (1992)
examined the projected future NO% emissions and resultant atmospheric nitrogen deposition to
the Chesapeake Bay under various emission control scenarios (Figure 31). They predict that only
by assertively controlling emissions (44% reduction) over the minimum CAAA requirements
would a noticeable reduction (11 %) in atmospheric N03" loading to Chesapeake Bay be achieved.
It should be noted however that the CAAA do not specifically address sources of atmospheric
NH3, which comprises 30-40% of the total atmospheric nitrogen deposition.
108
-------
zu
^^
S 15
o
c
• MM
* MM
£
- 10
w
c
o
'to
5
x °
O
z
n
—
Scenario 0
©
^^
^
^
^*s^
M^ ~
+*~"*^^
0 ^•"~~ Scenario 1
"-^"^
_ l£"'r ^*^~-^^^^ -^ — "^
^V, ' ' "MlH^I ^'
«
* ^.
®- v^ Scenario 2 _
* «*
*^^_ 'vj*
"*" " " "**
I ! I I I
360
c
0
"O
c
(U
240 ^
o
^
c
o
^
"t/3
120 a
-------
Relative Atmospheric Loadings... Revision Date: 15 March 1993
9.0 CURRENT UNCERTAINTY IN GREAT WATERS MASS
BALANCES
There has been tremendous improvements in our understanding to the role of atmospheric
deposition in supplying trace elements, mercury, nitrogen, and organic contaminants to surface
waters. Many advances have been made in the basic sampling and analytical tools required to
make reliable field measurements of atmospheric inventories and wet depositional fluxes. Such
measurements require extreme care and are, therefore, not necessarily suitable for large, routine
monitoring networks. Nonetheless, during the past ten years it has become possible to
accurately measure atmospheric concentrations, speciation, and wet depositional fluxes of many
chemicals, and excellent records are evolving at several locations (e.g., Lewes, DE; Chesapeake
Bay, Great Lakes region). Due to the inherent variability in atmospheric processes, long-term
records, on the scale of decades, will be required to assess changes in atmospheric deposition
loadings to the Great Waters. It is left up to the responsible agencies to develop and maintain
these long-term programs.
To prioritize future research efforts, the authors of this report have estimated the current
uncertainties in the fundamental atmospheric depositional process and assessed their relative
importance (Tables 2-5). In general, more reliable measurements of wet deposition are available
compared to either dry aerosol deposition or gaseous exchange. While wet deposition can be
measured in the field, our ability to predict (e.g., model) contaminant scavenging from the
atmosphere by precipitation is highly uncertain, perhaps no better than to within one to two
orders of magnitude. Specific studies of wet depositional processes, especially those employing
novel geochemical tracers and airborne sampling, are required. While it is important to continue
and expand wet deposition measurements and research, much of the research effort must be
placed in improving our ability to measure and model dry deposition. In particular, our
estimates of dry aerosol deposition are hindered both by a lack of aerosol size distribution
information and by our generally poor understanding of the micrometeorological environment
above water surfaces. The potentially large contaminant fluxes resulting from the rapid settling
of supermicron particles near emission sources (Holsen et al., 1991) as well as the possible
elevated fluxes during short-term, intensive meteorological events deserve further study.
Mass balance calculations and measurements of concentration gradients in the field
strongly suggest that many organic contaminant are degassing from the Great Lakes, especially
during warm summer months. Chemicals such as PCBs, which are no longer produced, may
be leaving the Great Lakes back into the atmosphere to be transported and deposited to the
world's oceans and to the polar ice pack. The processes by which the Great Waters give these
chemicals back to the atmosphere clearly need to be understood, both to predict contaminant
levels in these water bodies and to characterize the global redistribution of these persistent
chemicals.
110
-------
Relative Atmospheric Loadings... Revision Date: 15 March 1993
In summary, the significant progress made during the past two decades has provided many
of the tools required to answered the questions posed by Section 112(m) of the 1991
Amendments to the Clean Air Act. While much remains to be done, the regulatory community
will be well served to continue to adopt the geochemical mass balance, emphasizing processes
and fluxes of materials, as a rational framework for future endeavors.
Ill
-------
Relative Atmospheric Loadings... Revision Date: 16 March 1993
REFERENCES
Achman, D.R., Hornbuckle, K.C., and Eisenreich, SJ. 1993. Volatilization of PCBs from
Green Bay, Lake Michigan. Environ. Sci. Tech., 27(1), 75-86.
Achman, D.R., Eisenreich, S.J., and Bidleman, T.F. 1993b. Air-water exchange of PAHs in
Green Bay, Lake Michigan, (in prep.).
Albagli, A., Oja, H., and Dubois, L. 1974. Size-distribution pattern of PAHs in airborne
particulates. Environ. Lett., 6, 241-251.
Alberts, J.J., Schildler, J.E., Miller, R.W. and Nutter, D.E. 1974. Elemental mercury evolution
mediated by hurftic acid. Science 184, 895-897.
Andreae, M.O. 1979. Arsenic speciation in seawater and interstitial water: the influence of
biological-chemical interactions on the chemistry of a trace element. Limnol. Oceanogr. 24:440-
452.
Andreae, M.O. 1980. Arsenic in rain and the atmospheric mass balance of arsenic. J.
Geophys. Res., 85(C8), 4512-4518.
Andren, A. and Nriagu, J.O. 1979. The global cycle of mercury. In: The Biogeochemistry of
Mercury in the Environment. Nriagu, J.O., Ed., Elsevier-North Holland, Amsterdam, pp. 1-21.
Andren, A.W., Klein, D.H., and Talmi, Y. 1975. Selenium in coal-fired steam plant emissions.
Environ. Sci. TechnoL, 9, 856-858.
Apsimon, H.M., Kruse, M., and Bell, J.N.B. 1987. Ammonia emissions and their role in acid
deposition. Atmos. Environ., 15, 1939-1946.
Arimoto, R., Duce, R.A., Ray, B.J., and Unni, C.K. 1985. Atmospheric trace elements at
Enewetak Atoll, 2, Transport to the oceans by wet and dry deposition, J. Geophys. Res., 90,
2391-2408.
Arimoto, R., R. A. Duce, B. J. Ray, A. D. Hewitt and J. Williams. 1987. Trace elements in
the atmosphere of America Samoa: Concentrations and deposition to the tropical south Pacific.
J. Geophys. Res., 92, 8465-8479.
Arimoto, R., Ray, B.J., Duce, R.H., Hewitt, A.D., Boldi, R., and Hudson, A. 1990.
Concentrations, sources and fluxes of trace elements in the remote marine atmosphere of New
Zealand. J. Geophys. Res.
112
-------
Relative Atmospheric Loadings... Revision Date: 16 March 1993
Asher, W. and Pankow, J.F. 1991. Predictions of gas-water mass transport coefficients by a
surface renewal model. Environ. Sci. Technol., 25(7), 1294.
Atlas, E., Bidleman, T.F., and Giam, C.S. 1986. Atmospheric transport of PCB to the oceans.
In PCBs and the Environment, Waid, J.S. (Ed.). Vol. I, Chapter 4, Boca Raton, FL, CRC
Press, p. 79-200. .
Atlas, E. and Giam C.S. 1986. Sea-air exchange of high molecular weight synthetic organic
compounds. In The Role of Air-Sea Exchange in Geochemical Cycling, P. Buet-Menard (Ed),
D. Reidel Publ.: Dordrecht, 295-323.
Baker, J.E., Capel, P.O., and Eisenreich, S.J. 1986. Influence of colloids on sediment-water
partition coefficients of PCB congeners in natural waters. Environ. Sci. Tech., 20, 1136-1143.
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. Tech., 24, 342-352.
Baker, J.E., Eisenreich, S.J., Eadie, B.J. 1991. Sediment trap fluxes and benthic recycling of
organic carbon, PAHs, and PCB congeners in Lake Superior. Environ. Sci. Tech., 25, 500-509.
Baker, I.E., Church, T.M., Ondov, J.M., Scudlark, J.R., Conko, K.M., Leister, D.L., and
Wu, Z.Y. 1992. Chesapeake Bay Atmospheric Deposition Study, Phase I: July 1990 - June
1991. Report to the Department of Natural Resources, Annapolis, MD.
Baker, I.E., Eisenreich, S.J., and Swackhamer, D.L. Field-measured associations between
PCBs and suspended solids in natural waters: an evaluation of the partitioning paradigm. In
Organics and Solids in Water, R.A. Baker (Ed.), Chapter 4, Lewis Publishers: Chelsea, MI,
79-89.
Ballantine, D. and Zoller, W.H. (1984). Collection and determination of volatile organic
mercury compounds in the atmosphere by gas chromatography with microwave plasma detection.
Anal. Chem. ,56, 1288-1293.
Ballschmiter, K., Buchert, H., Bihler, S., and Zell, M. 1981. Baseline studies of global
pollution: IV. The pattern of pollution from organochlorine compounds in the North Atlantic
as accumulated in fish. Fres. Z. Anal. Chem., 306, 323-339.
Barrie, L.A., Lindberg, S.E., Chan, W.H., Ross, H.B., Arimoto, R. and Church, T.M. 1987.
On the concentration of trace metals in precipitation. Armos. Environ., 21(5), 1133-1135.
113
-------
Relative Atmospheric Loadings... Revision Date: 16 March 1993
Bennett, J.P. and Rathun, R.E. 1972. Reaeration in open channel flow. U.S. Geological
Survey Professional Paper #737.
Benoit, J.M., Fitzgerald, W.F., and Damman, A.W.H. 1992a. Historical atmospheric mercury
deposition in the mid-continental United States as recorded in an ombrotrophic peat bog. In
Mercury as a Global Pollutant, Conference Volume Monterey, CA. (in review).
Benoit, J.M., Fitzgerald, W.F., and Damman, A.W.H. 1992b. The biogeochemistry of an
ombrotrophic peat bog: evaluation of use as an archive of atmospheric mercury deposition, (in
prep.)
Bidleman, T.F., Billings, W.N., and Foreman, W.T. 1986. Vapor-particle partitioning of
semivolatile organic compounds: estimation from field collections. Environ. Sci. Tech., 20,138-
1043.
Bidleman, T.F. and Foreman, W.T. 1987. Vapor-particle partitioning of semi-volatile organic
compounds. In Sources and Fates of Aquatic Pollutants. Kites, R.A., and Eisenreich, S.J.,
Eds., Advances in Chemistry Series #216, Am. Chem. Soc., 27-56.
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.
Bidleman, T.F., Atkinson, R., Adas, E.L., Bonsang, B., Burns, K., Keene, W., Miller, J.,
Rudolf, J., and Tanabe, S. 1988. Transport of organic compounds in the atmosphere: review
and recommendations. In Long Range Transport of Natural and Contaminant Substances. T.
Knap, Ed., NATO ASI Series.
Bidleman, T.F., McConnell, L.L., Kucklick, J.R., Gotham, W.E., and Hinckley, D.A. 1991.
Air-water exchange ofHCHs in the Great Lakes. US EPA Rept., Great Lakes National Program
Office, Chicago, IL.
Bloom, N.S. and Crecelius, E.A. 1983. Determination of mercury in seawater at
subnanogram per liter levels. Marine Chem. ,14, 49-59.
Bloom, N.S. and Fitzgerald, W.F. 1988. Determination of volatile mercury species at the
nanogram level by low temperature gas chromatography with an ultrasensitive cold- vapor
atomic fluorescence detector. Anal. Chem. Acta, 51, 1714-1720.
Bloom, N.S., Watras, C.J. and Hurley, J.P. 1991. Impact of acidification on the methylmercury
cycle of remote seepage lakes. Water, Air and Soil Pollution , 56, 477-491.
114
-------
Relative Atmospheric Loadings... Revision Date: 16 March 1993
Bloom, N.S. 1989. Determination of picogram levels of methylmercury by aqueous phase
ethylation, followed by cryogenic gas chromatography with atomic fluorescence detection. Can.
J. Fish. Aquat. ScL, 46, 1131-1140.
Boring, L.S., Swank, W.T., Waide, J.B., and Henderson, G.S. 1988. Sources, fates, and
impacts of nitrogen inputs to terrestrial ecosystems: review and synthesis. Biogeochemistry,
6, 119-159.
Bricker-Urso, S.A., Nixon, S., Cochran, J.K. 1989. Accretion rates and sediment accumulation
in Rhode Island salt marshes. Estuaries, 12(4), 300.
Brimblecomb, P. and Clegg, S.L. 1988. The solubility and behavior of acid gases in the
marine aerosol. J. Atmos. Chem., 7, 1-18.
Brosset, C. 1987. The behavior of mercury in the physical environment. Water, Air and Soil
Pollution, 34, 145-166.
Brun, G.L., Howell, G.D., O'Neill, HJ. 1991. Spatial and temporal patterns of organic
contaminants in wet precipitation in Atlantic Canada. Environ. Sci. Technol. 25, 1249.
Brutsaert, W. and Jirkha G.H. (Eds). 1984. Gas Transfer at Water Surfaces, D. Reidel:
Hingham, MA.
Buat-Menard, P. and Chesselet, R. 1979. Variable influences of the atmospheric flux on the
trace metal chemistry of oceanic suspended matter. Earth Planet. Sci. Lett., 42, 399-411.
Buat-Menard, P. 1985. Air to Sea Transfer of Anthropogenic Trace Metals. In The Role of
Air-Sea Exchange in Geochemical Cycling, Buat-Menard, P., Ed., D. Reidel Publ., Dordrecht
pp 477-496.
Buat-Menard, P. (Ed.) 1986. The Role of Air-Sea Exchange in Geochemical Cycling. NATO
ASI Series, D. Reidel: Doredreht, 549 p.
Buckley, I. and L. Corio. Clean Air Act Amendments may stabilize nitrate deposition to
Chesapeake Bay. Maryland Power Plant Research Program Research Update, 4(2), 2-4.
Calvert, J.G., A. Lazrus, G.L. Kok, B.G. Heikes, J.G. Walega, J. Lind, and C.A. Cantrell.
1985. Chemical mechanisms of acid generation in the troposphere. Nature 317, 27-35.
Capel, P.D., Gunde, R., Ziircher, F., and Giger, W. 1990. Carbon speciation and surface
tension of fog. Environ. Sci. Technol, 24., 722-727.
115
-------
Relative Atmospheric Loadings... Revision Date: 16 March 1993
Capel, P.O., Leuenberger, C., and Giger, W. 1991. Hydrophobic organic chemicals in urban
fog. Atmos. Environ., 25 A, 1335-1346.
Cautreels, W. and VanCauwenberghe, K. 1978. Experiments on the distribution of organic
pollutants between airborne paniculate matter and the corresponding gas phase. Atmos.
Environ., 12,1133.
Chan, C.H. and Perkins, L.H. 1989. Monitoring of trace organic contaminants in atmospheric
precipitation. 7. Great Lakes Res., 15,465-475.
Chan, W.H., F. Tomassini and B. Loescher. 1983. An evaluation of sorption properties of
precipitation constituents on polyethylene surfaces. Atmos. Environ. 77:1779-1786.
Chiou, C. 1990. Roles of organic matter, minerals, and moisture in sorption of non-ionic
compounds and pesticides by soil. In Hwnic Substances in Soil and Crop Sciences, MacCarthy,
P., Clapp, C.E., Malcolm, R.L., and Bloom, P.R., Eds., Soil Sci. Soc. Amer., Madison, WI,
111-160.
Church, T.M., Galloway, J.N., Jickells, T.D., and Knap, A.H. 1982. The chemistry of
western Atlantic precipitation at the mid-Atlantic coast and on Bermuda. J. Geophys. Res.,
87(C13), 11,013-11,018.
Church, T.M., Tramontane, J.M., Scudlark, J.R., Jickells, T.D., Tokos, J.J. Jr., Knap, A.H.,
and Galloway, J.N. 1984. The wet deposition of trace metals to the western Atlantic Ocean at
the mid-Atlantic coast and on Bermuda. Atmos. Environ., 18(12), 657-2664.
Church, T.M., Tramontane, J.M., and Murray, S. 1986. Trace metal fluxes through the
Delaware estuary. In Proc. of Int'l Council for Exploration of the Sea (ICES) Symposium
"Contaminant Fluxes through the Coastal Zone," Nantes, France, May 1984. Rapp. P.-v Reun.
Cons. int. Explor. Merc, 186, 271-276.
Church, T.M. 1987. Estuarine, Atmospheric and Groundwater Inputs to Continental Margins,
In U.S. Global Ocean Flux Study - Ocean Margins in GOFS, Report #6, University of Southern
Mississippi, March 10-12, 1987.
Church, T.M., Tramontane, J.M., Scudlark, J.R., and Murray, S.L. 1988. Trace metals in the
waters of the Delaware estuary, In Ecology and Restoration of the Delaware River Basin,. S.K.
Majumdar, E.W. Miller and L.E. Sage, Eds., The Philadelphia Academy of Sciences.
Church, T.M. 1988. Biogeochemical factors influencing the residence time of microconstituents
in a large tidal estuary, Delaware Bay. Marine Chemistry, 18, 393-406.
116
-------
Relative Atmospheric Loadings... Revision Date: 16 March 1993
Church, T.M., Arimoto, R., Barrier, L.A., Dehairs, F., Dulac, F., Jickells, T.D., Mart, L.,
Sturges W.T., and Toiler, W.H. 1989. The long-range atmospheric transport of trace elements;
a critical evaluation. Chapter 3, In Long Range Transport of Atmospheric Material Between
Oceans and Continents, Knap, A., Ed.,, Reidel Pubh, NATO Workshop, Bermuda, 1987.
Church, T.M., Veron, A., Patterson, C.C., Settle, D., Erel, Y., Maring, H.R., and Flegal
A.R. 1990. Trace elements in the North Atlantic troposphere: Shipboard results of precipitation
and aerosols. Global Biogeochem. Cycles, 4, 431-443.
Church, T.M., J.M. Tramontane, D.M. Whelpdale, M.O. Andreae, J.N. Galloway, W.C.
Keene, A.H. Knap, J. Tokos. 1991. Atmospheric and precipitation chemistry over the North
Atlantic Ocean: Shipboard results from April-May 1984. J. Geophys. Res., 96(D10), 18705-
18725.
Church, T.M. 1992. Sources, transport, and fate of toxic elements in the coastal marine
environment, In International Symposium on Marine Pollution, In Celebration of the 60th
Anniversary of National Cheng Kung University, 1931-1991, (Proceedings of Marine Pollution
Symposium - National Cheng Kung University, Taiwan, November, 1991), pp. 114-120.
Church, T.M. and Scudlark, J.R. 1992. Trace elements in precipitation at the mid-Atlantic
coast; a successful record since 1982. USDA Forest Service Report NC-150, Symposium
Proceedings, Philadelphia, PA, October 8* 1991, pp. 45-46.
Church, T., Veron, A., Patterson, C., and Settle, D. 1992. Trace metal scavenging from the
north Atlantic troposphere. In Precipitation Scavenging and Atmosphere-Surface Exchange, Vol.
1, The Gerogii Volume, Precipitation Scavenging Processes, (Proceedings of the Fifth
International Conference on Precipitation Scavenging and Atmospheric Surface Exchange
Processes, Richland Washington, 15-19 July, 1991), pp. 433-446.
Correll, D.L. and Ford, D. 1982. Comparison of precipitation and land runoff as sources of
estuarine nitrogen. Estuarine, Coastal and Shelf Science, 15:45-56.
Gotham, W.E. Jr. 1990. Chemical and Physical Processes Affecting the Transport and Fate
of Semivolatile Organic Contaminants in the Environment. Ph.D. Thesis, University of South
Carolina, 195 p.
Gotham, W.E. Jr., and Bidleman, T.F. 1992. Estimating the atmospheric deposition of
organochlorine contaminants to the arctic. Chemosphere, 22, 165-188.
Coutant, R.W., Callahan, P.J., Kuhlman, R., and Lewis, R.G. 1989. Design and performance
of a high volume compound annular denuder. Atmos. Environ., 23, 2205-2211.
117
-------
Relative Atmospheric Loadings... Revision Date: 16 March 1993
Cutter, G.A. and Church, T.M. 1986. Selenium in western Atlantic precipitation. Nature,
322(6081), 720-722.
Czuczwa, J., Leuenberger, C., and Giger, W. 1988. Seasonal and temporal changes of organic
compounds in rain and snow. Atmos. Environ., 22, 907-916.
Danckwerts, P.V. 1951. Significance of liquid-film coefficients in gas absorption. Ind. Engr.
Chem.,43, 1460-1467.
Davidson, C.I. and Wu, Y.L. 1988. Dry deposition of trace elements. In Acid Precipitation,
Adriano, D.C., Ed., Advances in Environmental Sciences, Springer-Verlag, New York.
Davidson, C.I. arid Wu, Y.L. 1989. Dry deposition of trace elements. In Control and Fate of
Atmospheric Trace Metals, Pacyna, J.M, and Ottar, B., Eds., Kluwer Academic Publishers,
Dordrecht, Holland.
Deacon, E.L. 1977. Gas transfer to and across an air-water interface. Tellus, 29, 363-374.
D'Elia, C.F., Sanders, J.G., and Boynton, W.R. 1986. Nutrient enrichment studies in a
coastal plain estuary: phytoplankton growth in large-scale, continuous cultures. Can. J. Fish.
and Aquatic Sci., 43,397-406.
Demaio, L. and Corn, M. 1966. PAHs associated with particulates in Pittsburgh air. J. Air.
Poll. Control. Assoc., 16, 67-71.
DePinto, J. 1992. Personal communication.
DIONEX Corporation. 1981. Determination of anions in acid rain. Application note 31,
Sunnyvale, Cal.
Dodd, R.L., Gundel, L.A., Benner, W.H., and Novakov, T. 1984. Non-ammonium reduced
nitrogen species in atmospheric aerosol particles. Sci. Total Environ., 36, 277-282.
Doskey, P.V. and Andren, A.W. 1981. Modeling the flux of atmospheric polychlorinated
biphenyls across the air/water interface. Environ. Sci. Technol., 15, 705-709.
Doskey, M.G. and Adriano, D.C. 1992. Trace metal impact on plants: Mediation by soil and
mycorrhizae. In The Deposition and Fate of Trace Metals in Our Environment , Verry, E.S.
and Vermett, S.J., Eds., USDA Forest Service General Technical Report NC-150, pp. 105-115,
from the Proceedings of the National Atmospheric Deposition Program National Trends
Network, October 8, 1991, Philadelphia, PA.
118
-------
Relative Atmospheric Loadings... Revision Date: 16 March 1993
Duce, R.A., Quinn, J.G., Olney, C.E. et al., Enrichment of heavy metals and organic
compounds in the surface microlayer of Narragansett Bay, Rhode Island. Science, 176, 161-
163.
Duce, R.A., Hoffman, G.L., Ray, B.J., Fletcher, I.S., Wallace, G.T., Fasching, J.L.,
Piotrowicz, S.R., Walsh, P.R., Hoffman, E.J., Miller, J.M., and Heffter, J.L. 1976. Trace
metals in the marine atmosphere: Sources and fluxes, In Marine Pollutant Transfer, Windom,
H.L. and Duce, R.A., Eds., Health and Company, Lexington, MA, pp. 77-119.
Duce, R.A., Unni, C.K., Ray, B.J., Prospero, J.M., and Merrill, J.T. 1980. Long-range
atmospheric transport of soil dust from Asia to the tropical north Pacific: Temporal variability.
Science, 209, 1522-1524.
Duce, R.A., Liss, P.S., Merrill, J.T., Atlas, E.L., Buat-Menard, P., Hicks, B.B., Miller,
J.M.,Prospero, J.M., Arimoto, R., Church, T.M., Ellis, W., Galloway, J.N., Hansen, L.,
Jickells, T.D., Knap, A.H., Reinhardt, K.H., Schneider, B., Soudine, A., Tokos, J.J.,
Tsunogai, S., Wollast, R., and Zhou, M. 1991. The atmospheric input of trace species to the
world ocean. Global Biogeochem. Cycles, 5,193-259.
Duinker, J.C. and Bouchertall, F. 1989. On the distribution of atmospheric PCB congeners
between vapor phase, aerosols, and rain. Environ. Sci. Technol., 23, 57.
Dumarey, R., Heindryckx, R., Dams, R., and Hoste, J. 1979. Determination of volatile
mercury compounds in air with the Coleman Mercury Analyzer system. Anal. Chim. Acta 107,
159.
Eadie, B.J. and Robbins, J.A. 1987. The role of paniculate matter in the movement of
contaminants in the Great Lakes. In Sources and Fates of Aquatic Pollutants, Hites, R. A. and
Eisenreich, S.J., Eds., ACS Advances in Chemistry Series #216, 319-364.
Edgerton, E.S., Lavery, T.F., and Prentice, H.S. 1991. National Dry Deposition Network Third
Annual Progress Report (1989) , EPA/600/3-91/018.
Eisenreich, S.J., Looney, B.B. and Thornton, J.D. 1980. Assessment of Airborne Organic
Contaminants in the Great Lakes Ecosystem. Science Advisory Board (Appendix B),
International Joint Commission, Windsor, Ontario, 150 pp.
Eisenreich, S J., Looney, B.B., and Thornton, J.D. 1981. Airborne organic contaminants in the
Great Lakes ecosystem. Environ. Sci. Tech., 15, 30-38.
Eisenreich, S.J., N.A. Metzer, N.R. Urban and J.A. Robbins. 1986. Response of atmospheric
lead to decreased use of lead in gasoline. Environ. Sci. Tech., 20, 171-174.
119
-------
Relative Atmospheric Loadings... Revision Date: 16 March 1993
Eisenreich, SJ. 1987. The chemical limnology of nonpolar organic contaminants:
polychlorinated biphenyls in Lake Superior. In Sources and Fates of Aquatic Pollutants, Kites,
R.A. and Eisenreich, S.J., Eds., ACS Advances in Chemistry Series #216, 393-470.
Eisenreich, SJ., Willford, W., and Strachan, W.MJ. 1989. In Intermedia Pollutant
Transport: Modeling and Field Measurements, D. Allen, Y. Cohen and I. Kaplan (Eds.),
Plenum Publ., N.Y. pp. 19-40.
Eisenreich, S.J., Achman, D.A., Hornbuckle, K.C., and Baker, J.E. 1991. Volatilization of
PCBs from the Great Lakes. In Air-Water Mass Transfer Second International Symposium,
Gulliver, J.S. and Wilhelms, S.C., Eds., U.S. Army Waterways Experiment Station/American
Society of Civil Engineering., ASCE Special Publication.
Eisenreich, S.J., Swackhamer, D.L., Lodge, K.L., Kawka, O.E., King, P., Achman, D.R., and
Johnson, G.A. 1992. Evaluation of Bioaccumulation Models for HOCs in Fish for Exposure
Assessment (PCDDs, PCDFs, PCBs, PAHs) University of Minnesota Report to the MN
Pollution Control Agency and the Legislative Commission on Minnesota Resources,
Minneapolis, MN.
Eisenreich, S.J., Baker, J.E., Franz, T., Swanson, M., Rapaport, R.A., Strachan, W.MJ., and
Kites, R.A. Atmospheric deposition of hydrophobic organic contaminants to the Laurentian
Great Lakes. In Fate of Pesticides and Chemicals in the Environment. J.L. Schnoor (Ed.), John
Wiley and Sons: New York, Chapter 3, 51-78.
Eisenreich, SJ. and Strachan, W.MJ. 1992. Estimating atmospheric deposition of toxic
substances to the Great Lakes. A Workshop held at the Canada Centre for Inland Waters.
Burlington, Ontario, January 31-February 2 1992.
Eitzer, B.D. and Kites, R.A. 1989. PCDDs and PCDFs in the ambient air of Bloomington,
In. Environ. Sci. Tech., 23, 1396-1401.
Eliassen, A., O. Hov, T. Iverson, J. Saltbones and D. Simpson. 1988. Estimates of airborne
transboundary transport of sulfur and nitrogen over Europe. EMEP/MSC-W 1/88, 79 pp.
Elzerman, A.W. and Coates, J.T. 1987. Hydrophobic organic compounds on sediments:
equilibria and kinetics of sorption. In Sources and Fates of Aquatic Pollutants, Kites, R.A. and
Eisenreich, S J., Eds., ACS Advances in Chemistry Series #216, 263-318.
EPA. 1991. Evaluation of the Atmospheric Deposition of Toxic Contaminants to Puget Sound.
EPA Report 910/9-91-027.
120
-------
Relative Atmospheric Loadings... Revision Date: 16 March 1993
Fanning, K. 1989. Influence of atmospheric pollution on nutrient limitation in the ocean.
Nature, 339, 460-463.
Finlayson-Pitts, BJ. 1983. The reaction of NO2 with NaCl and atmospheric implications of
NaOCl formation. Nature 306, 676-677.
Fisher, D.C., Ceraso, J., Mathew, T., and Oppenheimer, M. 1988. Polluted Coastal Waters:
The Role of Acid Rain, Environmental Defense Fund, New York.
Fisher, D.C. and Oppenheimer, M. 1991. Atmospheric nitrogen Deposition and the Chesapeake
Bay estuary. Ambio, 23(3), 102-108.
Fitzgerald, W.F. 1982. Evidence for anthropogenic mercury inputs to the oceans. EOS Trans.
Amer. Geophys. U., 63, 77.
Fitzgerald, W.F. and Gill, G.A. 1979. Subnanogram determination of mercury by two-stage
gold amalgamation and gas phase detection applied to atmospheric analysis. Anal. Chem., 51,
1714-1720.
Fitzgerald, W.F. and Gill, G.A., Hewitt, A.D. 1981. Mercury, a trace atmospheric gas. In:
Abstracts, Symposium on the Role of the Ocean on Atmospheric Chemistry. I AMP A Third
Scientific Assembly, Hamburg, Federal Republic of Germany, August 17-18.
Fitzgerald, W.F., Gill, G.A. and Hewitt, A. 1983. Air-Sea exchange of mercury. In Trace
Metals in Seawater Wong, C.S. et al (Eds.), NATO Conference Series, IV, Marine Science,
V.9, Plenum Press, New York pp 297-316.
Fitzgerald, W.F., Gill, G.A., and Kim, J.P. 1984. An equatorial Pacific Ocean source of
atmospheric mercury. Science, 224, 597-599.
Fitzgerald, W.F. and Gill, G.A. 1985. Depositional fluxes of mercury to the oceans. In
International Conference on Heavy Metals in the Environment, Vol. 1, Lekkas, T.O., Ed., CEP
Consultants, Edinburgh, p. 79-81.
Fitzgerald, W.F. 1986. Cycling of mercury between the atmosphere and oceans. In The Role
of Air-Sea Exchange in Geochemical Cycling, Buat-Menard, P., Ed., NATO Advanced Science
Institutes Series, Reidel Publishing Co., Dordrecht, The Netherlands, p. 363-408.
Fitzgerald, W.F. 1989. Atmospheric and oceanic cycling of mercury, In SEAREX Volume of
the Chemical Oceanography Series, Riley, J.P. and Chester, R., Eds. Academic Press, London.
p. 151-186
121
-------
Relative Atmospheric Loadings... Revision Date: 16 March 1993
Fitzgerald, W.F. and Watras, C.J. 1989. Mercury in the surficial waters of rural Wisconsin
lakes. Sci. Total. Environ., 87/88, 223-232.
Fitzgerald, W.F. and Clarkson, T.W. 1991. Mercury and mono methyl mercury: present and
future concerns. Environ. Health Perspectives, 96:159-166.
Fitzgerald, W.F., Mason, R.P., and Vandal, G.M. 1991. Atmospheric cycling and air-water
exchange of mercury over mid-continental lacustrine regions. Water, Air, Soil, and Pollution,
56, 745-767.
Fitzgerald, W.F., Mason, R.P. and Vandal, G.V. 1992. Air-Water cycling of mercury in lakes.
In Conference Volume, Mercury as a Global Pollutant, Monterey, CA. (in review).
Fitzgerald, W.F., Mason, R.P., and Vandal, G.M. (in press). Atmospheric cycling and air-water
exchange of mercury over mid-continental lacustrine regions. International Conference on
Mercury as an Environmental Pollutant, Gavle, Sweden, June 11-13, 1990.
Flegal, A.R., J.O. Nriagu, K.H. Coale, and S. Niemeyer. 1989. Isotopic tracers of lead
contamination in the Great Lakes. Nature, 339,.455-458.
Fogg, T.R. and Fitzgerald, W.F. 1979. Mercury in southern New England coastal rains. J.
Geophys. Res. 84, 6987-6988.
Foreman, W.T. and Bidleman, T.F. 1987. An experimental system for investigating vapor-
particle partitioning of trace organic pollutants. Environ. Sci. Tech., 21, 869-875.
Foreman, W.T. and Bidleman, T.F. 1990. Semivolatile organic compounds in the ambient air
of Denver, Colorado. Atmos. Environ., 24A, 2405-2416.
Franz, T.P., Eisenreich, S.J., and Swanson, M.B. 1991. Evaluation of precipitation samplers
for assessing atmospheric fluxes of trace organic contaminants. Chemosphere, 23, 343.
Friedland, AJ. 1992. The use of organic forest soils as indicators of atmospheric deposition
of trace metals. In The Deposition and Fate of Trace Metals in Our Environment, Verry, E.S.
and Vermett, S.J., Eds., USDA Forest Service General Technical Report NC-150, pp. 97-104,
from the Proceedings of the National Atmospheric Deposition Program National Trends
Network, October 8, 1991, Philadelphia, PA.
Galloway, J.N. and G.E. Likens. 1978. The collection of precipitation for chemical analysis.
Tellus 30, 71.
122
-------
Relative Atmospheric Loadings... Revision Date: 16 March 1993
Galloway, J.N. and Whelpdale, D.M.. 1980. An atmospheric sulfur budget for eastern North
America. Atmos. Environ., 14, 409-417.
Galloway, J.N., Thornton, J.D., Norton, S.A., Volchok, H.L., and McLean, R.A. 1982. Trace
metals in atmospheric deposition: A review and assessment. Atmos. Environ. 16, 1677-1700.
Galperin, B. and G.L. Mellor. 1990. A time-dependent, three-dimensional model of the
Delaware Bay and River system. Part 2: Three dimensional flow fields and residual circulation.
Estuarine Coastal and Shelf Sci., 31, 255-281.
Gatz, D.F. and Chu, L.C. 1986. Metal solubility in atmospheric deposition. In Toxic Metals
in the Atmosphere, John Wiley & Sons, New York, pp. 355-391.
GESAMP (Group of Experts on the Scientific Aspects of Marine Pollution) 1989. The
atmospheric input of trace species to the world ocean. Rep. Stud. 38, 111 pp., World Meteorol.
Organ., Geneva.
Gill, G.A. and Fitzgerald, W.F. 1987a. Mercury in surface waters of the open ocean. Global
Biogeochem. Cycles 1, 199-212.
Gill, G.A. and Fitzgerald, W.F. 1987b. Picomolar mercury measurements in seawater and other
materials using stannous chloride reduction and two-stage gold amalgamation with gas phase
detection. Marine Chem., 20, 227-243.
Gilmour, C.G. and Henry, E.A. 1991. Mercury methylation in aquatic systems affected by acid
deposition. Environ. Poll, 71, 131-169.
Glass, G.E., Sorensen, J.A., Schmidt, K.W., Rapp, G.R., Yap, D., and Fraser, D. 1991.
Mercury deposition and sources for the upper Great Lakes region. Water, Air, Soil, and
Pollution, 56, 235-249.
Glotfelty, D.E., Seiber, J.N., Liljedahl, L.A. 1987. Pesticides in fog. Nature, 325, 602-605,
Glotfelty, D.E., Majewski, M.S., and Seiber, J.N. 1990. Distribution of several
organophosphorus insecticides and their oxygen analogs in a foggy atmosphere. Environ. Sci.
Technol., 24, 353.
Glotfelty, D.E., Williams, G.H., Freeman, H.P., and Leech, M.M. 1991. Regional
atmospheric transport and deposition of pesticides in Maryland. In Long Range Transport of
Pesticides and Other Toxics. Kurtz, D., Ed., Lewis Publishers.
123
-------
Relative Atmospheric Loadings... Revision Date: 16 March 1993
Golambek, A. and Prinn.R. 1986. A global three-dimensional model of the circulation and
chemistry of CFC13, CF2C12, CHF3CC13, CC14, and N2O. J. Geophys. Res., 91, 3986-4002.
Gorzelska, K. and Galloway, J.N. 1990. Amine nitrogen in the atmospheric environment over
the North Atlantic Ocean. Global Biogeochem. Cycles, 4(3), 309-333.
Gregor, DJ. and Gummer, W.D. 1989. Evidence of atmospheric transport and deposition of
organochlorine pesticides and PCBs in Canadian Arctic snow. Environ. ScL Tech., 22,561-565.
Grosjean, D. 1983. PAHs in Los Angeles air from samples collected on Teflon, glass, and
quartz filters. Atmos. Environ., 17, 2565-2573.
Gschwend, P.M. and Wu, S.C. 1985. On the constancy of sediment-water partition coefficients
of hydrophobic organic pollutants. Environ. Sci. Technol, 19,90-96.
Hanratty, TJ. 1991. Effect of gas flow on physical absorption. In Air-Water Mass Transfer
Second International Symposium, Gulliver, J.S. and Wilhelms, S.C., Eds., U.S. Army
Waterways Experiment Station/American Society of Civil Engineering., ASCE Special
Publication, 10-33.
Hanson, PJ. and Lindberg, S,E. 1991. Dry deposition of reactive nitrogen compounds: a
review of leaf, canopy and non-foliar measurements. Atmos. Environ., 25A(8): 1615-1634.
Hargrave, B.T., Vass, W.P., Erickson, P.E., and Fowler, B.T. 1988. Atmospheric transport
of organochlorines to the Arctic Ocean. Tellus, 40B, 480-493.
Hart, K.M. 1989. A study of atmospheric n-alkanes and PAHs and their distributions between
the gaseous and paniculate phases. Ph.D. Thesis, Oregon Graduate Institute of Science and
Technology, 307 p.
Hart, K.M., Isabelle, L.M., and Pankow, J.F. 1992. High volume air-sampler for particle and
gas sampling: 1. Design and gas sampling protocol. Environ. ScL Tech., 26, 1048-1052.
Hecky, R.E. and Killham, P. 1988. Nutrient limitation of phytoplankton in freshwater and
marine environments: A review of recent evidence on the effects of enrichment. Limnol
Oceanogr., 33, 796-822.
Hering, S.V., D. R. Lawson, I. Allegrini, A. Febo, C. Perrino, M. Possanzini, I.E. Sickles,
K.G. Anlauf, A. Wiebe, B.R. Appel, W. John, J. Ondo, S. Wall, R.S. Braman, R. Sutton,
G.R. Cass, P.A. Solomon, DJ. Eatough, N.L. Eatough, E.C. Ellis, D. Grosgean, B.B. Hicks,
J.D. Womack, J. Horrocks, K.T. Knapp, T.G. Ellestad, R.J. Paur, WJ. Mitchell, M. Pleasant,
E. Peake, A. MacLean, W.R. Pierson, W. Brachaczek, H.I. Schiff, G.I. Mackay, C.W. Spicer,
D.H. Stedman, A.M. Winer, H.W. Biermann, and E.C. Tuazon. 1988. The nitric acid Shootout:
124
-------
Relative Atmospheric Loadings... Revision Date: 16 March 1993
Field comparison of measurement methods. Atmos. Environ. 22(8), 1519-1539.
Hermanson, M.H. and Kites, R.A. 1989. Long-term measurements of atmospheric PCBs in
the vicinity of Superfund dumps. Environ. Sci. Tech., 23, 1253-1258.
Hicks,B.B., M.L. Wesley.S.E. Lindberg, and S.M. Bromberg (Eds). 1986. Proceedings of the
Dry Deposition Workshop of the National Acid Precipitation Assessment Program.
NOAA/ATDD, Oak Ridge.TN.
Hicks, B.B., Wesley, M.L., and Durham, J.L. 1980. Critique of methods to measure dry
deposition: workshop summary. EPA-600/9-80-050. U.S. Environmental Protection Agency,
69 pp. (NTIS PB 81-126443).
Higbie, R. 1935. The rate of adsorption of a pure gas into a still liquid during short periods of
exposure. Trans. Am. Inst. Chem. Eng., 35, 365-389.
Hinckley, D.A., Bidleman, T.F., and Rice, C.P. 1991. Atmospheric organochlorine pollutants
and air-sea exchange of HCHs in the Bering and Chuckchi Seas. J. Geophys. Res., 96, 7201-
7213.
Hinckley, D.A., Bidleman, T.F., Foreman, W.T., and Tuschall, J.R. 1990. J. Chem. & Eng.
Data, 35, 232-237.
Hinga, K.R., Keller, A.A., and Oviatt, C.A. 1991. Atmospheric deposition and nitrogen inputs
to coastal waters. Ambio, 20, 256-260.
Hoff, R.M., Muir, D.C,G., and Grift, N.P. 1992a. Annual cycle of polychlorinated biphenyls
and organohalogen pesticides in air in southern Ontario, air concentration data. Environ. Sci.
Tech., 26, 266-275.
Hoff, R.M., Muir, D.C.G., and Grift, N.P. 1992b. Annual cycle of polychlorinated biphenyls
and organohalogen pesticides in air in southern Ontario, atmospheric transport and sources.
Environ. Sci. Tech., 26, 276-283.
Holsen, T.M., Noll, D.E., Liu, S., and Lee, W. 1991. Dry deposition of polychlorinated
biphenyls in urban areas. Environ. Sci. Technol., 25, 1075-1078.
Hornbuckle, K.C., Achman, D.A., and Eisenreich, S.J. 1993. Over-water and over-land
concentrations of PCBs in Green Bay, Lake Michigan. Environ. Sci. Tech., 27, 87-98.
Hudson, R.J.M., Gherini, S.A., and Porcella, D.B. 1992. The MCM Lake Mercury Model.
Prepared for the Electric Power Research Institute by Tetra Tech, Inc., Lafayette, CA.
125
-------
Relative Atmospheric Loadings... Revision Date: 16 March 1993
Hurley, J.P., Watras, CJ. and Bloom, N.S. 199la. Mercury cycling in northern Wisconsin
seepage lakes: the role of paniculate matter in vertical transport. Water, Air and Soil Pollution
56, 543-551.
Iverfeldt, A. 1988. Mercury in the Norwegian fjord Framvaren. Marine Chem. 23, 441-456.
Iverfeldt, A. 199 la. Mercury in forest canopy throughfall water and its relation to atmospheric
deposition. Water, Air, Soil, and Pollution, 56, 553-563.
Iverfeldt, A. 1991b. Occurrence and turnover of atmospheric mercury over the Nordic
Countries. Water, Air, Soil, and Pollution, 56, 251-265.
Iverfeldt, A. and Lindqvist, O. 1986. Atmospheric oxidation of elemental mercury by ozone in
the aqueous phase. , Atmos. Environ., 20, 1567-1573.
Jaing, S., H. Robberecht and F. Adams. 1983. Atmos. Environ. ,17, 111-114.
Jaworski,N.A., P.M. Groffman, A. A. Keller and J.C. Prager. 1992. A watershed nitrogen mass
balance: The upper Potomac River basin. Estuaries 15(1), 83-95.
Jeremiason, J., Engstrom, D., Baker, J.E., and Eisenreich, S.J. 1993. in preparation.
Jickells, T.D., Knap, A.M., and Church, T.M. 1984. Trace metals in Bermuda rainwater. J.
Geophys. Res. 89(D1), 1423-1428.
Jickells, T., Church, T., and W. Deuser. 1987. Comparison of atmospheric inputs and deep-
ocean particle fluxes for the Sargasso Sea. Global Biogeochem. Cycles, 1, 117-130.
Jickells, T.D., R. Arimoto, L.A. Barrie, T.M. Church, F. Dehairs, F. Dulac, L. Mart, W.
Sturges, W. Zoller. 1989. The Long-range transport of trace elements: four case studies.
Chapter 10, In Long Range Transport of Atmospheric Material Between Oceans and Continents
Knap, A., Ed., Reidel Publ., NATO Workshop, Bermuda, 1987.
Jickells, T.D., A.M. Knap, R. Sherriff-Dow and J. Galloway. 1990. No ecosystem shift.
Nature 347, 25-26.
Johansson, K., Aastrup, M., Anderson, A., Bringmark, L., and Iverfeldt, A. 1991. Mercury
in Swedish forest soils and waters - Assessment of critical load. Water, Air and Soil Pollution
56, 267-281.
Johnson, D.L. and Braman, R.S. 1974. Distribution of atmospheric mercury species near
ground. Environ. Sci. Technol., 8, 1003-1009.
126
-------
Relative Atmospheric Loadings... Revision Date: 16 March 1993
Johnson, D.L. and R. S. Braman. 1975. Alkyl- and inorganic arsenic in air samples.
Chemosphere 6, 333-338.
Junge, C.E. 1956. Recent investigation in air chemistry. Tellus 8, 127-139.
Junge, C.E. 1977. Basic considerations about trace constituents in the atmosphere as related
to the fate of global pollutants, In Adv. Env. Sci. Techn., Vol. 8, edited by I.H. Suffett.
Karickhoff, S.W. 1984. Pollutant sorption in aquatic systems. /. Hydra. Div., Am. Soc. Civ.
Eng., 110, 707-735.
Keller, B.J., M.E. Peden and L.M. Skowron. 1988. Development of standard methods for the
collection and analysis of precipitation: Trace metals. Illinois State Water Survey Contract
Report 43B, Champaign, 111.
Katz, M. and Chan, C. 1980. Comparative distribution of eight PAHs in airborne particulates
collected by conventional high-volume sampling and by size fractionation. Environ. Sci. Tech.,
14, 838-843.
Kaupp, H. and Umlauf, G. 1990. Distribution of BaP, B[ghi]Per, and some organochlorine
compounds in atmospheric particulate matter with respect to particle size. In Organohalogen
Compounds, Vol. 1: Dioxin '90. Hutzinger, O. and Fiedler, H., Eds., EPRI-Seminar
Toxicology, Environment, Food, Exposure, Risk.
Kaupp, H. and Umlauf, G. 1992. Atmospheric gas-particle partitioning of organic compounds:
comparisons of sampling methods. Atmos. Environ., 26A, 2259-2268.
Kertesz-Saringer, M.E., Meszaros, M.E., and Varlonyi, T. 1971. Note on the size distribution
of B[a]P containing particles in urban air. Atmos. Environ., 5, 429-431.
Ketserides, G., and Eichman, R. 1978. Organic compounds in aerosol samples. Pageoph,
116, 274-282.
Kim, J.P. and Fitzgerald, W.F. 1986. Sea-air partitioning of mercury in the equatorial Pacific
Ocean. Science, 231, 1131-1133.
Kim, J.P. and Fitzgerald, W.F. 1988. Gaseous mercury profiles in the tropical Pacific Ocean.
Geophys. Res. Lett., 15, 40-43.
127
-------
Relative Atmospheric Loadings... Revision Date: 16 March 1993
Klappenbach, E.W. 1992. Analysis and deposition of trace metals around the Great Lakes. In
The Deposition and Fate of Trace Metals in Our Environment, Verry, E.S. and Vermett, S.J.,
Eds., USDA Forest Service General Technical Report NC-150, pp. 37-43, from the Proceedings
of the National Atmospheric Deposition Program National Trends Network, October 8, 1991,
Philadelphia, PA.
Knap, A., Jickells, T., Pszenny, A., and Galloway, J. 1986. Significance of atmospheric-
derived nitrogen on productivity of the Sargasso Sea. Nature, 320, 158-160.
Korthals and Winfrey 1982. Rates of mercury methylation.
Landmm, P.P., Nihart, S.R., Eadie, B.J., and Herche, L.R. 1987. Reduction in bioavailability
of organic contaminants to the amphipod Pontopreia hoyi by dissolved organic matter of
sediment interstitial waters. Environ. Toxicol. Chem., 6, 11-20.
Lane, D.A., Johnson, N.D., Barton, S.C., Thomas, G.H.S., and Schroeder, W.H. 1988.
Development and evaluation of a novel gas particle sampler for chlorinated organic compounds
in ambient air. Environ. Sci. Tech., 22, 941-947.
Larsson, P. 1983. Transport of 14C-labelled PCB compounds from sediment to water and from
water to air in laboratory model systems. Water Res., 10, 1317-1326.
Larsson, P. 1985. Contaminated sediments of lakes and oceans as sources of chlorinated
hydrocarbons for release to water and atmosphere. Nature, 317, 347-349.
Larsson, P., Okla, L., Ryding, S-O., and Westoo, B. 1990. Contaminated sediment as a
source of PCBs in a river system. Can. J. Fish. Aquat. Sci., 47, 446-754.
Ledwell, J.R., Watson, AJ. and Broecker, W.S. 1986. A deliberate tracer experiment in the
Santa Monica Basin. Nature, 323, 322-324.
Leister, D.L. and Baker, J.E. 1993. Atmospheric deposition of organic contaminants to the
Chesapeake Bay, Atmos. Environ., in press.
Levy, H. and W.J. Moxim. 1987. Fate of U.S. and Canadian combustion nitrogen emissions.
Nature 328, 414-416.
Levy, H. and W.J. Moxim. 1989. Simulated global distribution and deposition of reactive
nitrogen emitted by fossil fuel combustion. Tellus Ser.B. 41, 256-271.
Lewis, W.M. 1983. Collection of airborne materials by a water surface. Limnol. Oceanogr.,
28(6), 1242-1246.
128
-------
Relative Atmospheric Loadings... Revision Date. 16 March 1993
Ligocki, M.P., Leuenberger, C., and Pankow, J.F. 1985a. Trace organic compounds in rain -
-II. Gas scavenging of neutral organic compounds. Atmos. Environ., 19, 1609-1617.
Ligocki, M.P., Leuenberger, C., and Pankow, J.F. 1985b. Trace organic compounds in rain -
-III. Particle scavenging of neutral organic compounds. Atmos. Environ., 19, 1619-1626.
Ligocki, M.P. and Pankow, J.F. 1989. Measurements of the gas-particle distributions of
atmospheric organic compounds. Environ. Sci. Tech., 23, 75-83.
Likens, G.F., E.S. Edgerton and J.N. Galloway. 1983. The composition and deposition of
organic carbon in precipitation. Tettus 35B, 16-24.
Lim, B. and Jickells, T.D. 1990. Dissolved, paniculate, and acid leachable trace metal
concentrations in North Atlantic precipitation collected on the Global Change Expedition.
Global Biogeochem. Cycles, 4(4), 445.
Lim, B., Jickells, T.D., and Davies, J.D. 1991. Sequential sampling of particles, major ions
and total trace metals in wet deposition. Atmos. Environ., 25A, 745-762.
Lindberg, S.E. 1982. Factors influencing trace metal, sulfate, and hydrogen ion concentrations
in rain. Atmos. Environ., 16, 1701-1709.
Lindberg, S.E., G.M. Lovett D.D. Richter and D.W. Johnson. 1986. Atmospheric deposition
and canopy interactions of major ions in a forest. Science 231: 141-145.
Lindberg, S.E. and R.R. Turner. 1988. Factors influencing atmospheric deposition, stream
export, and landscape accumulation of trace metals in forested watersheds. Water, Air, and Soil
Pollution, 39:123-156.
Lindberg, S.A. 1989. Application of surface analysis methods to studies of atmospheric
deposition in forests. Proceedings of International Congress on Forest Decline Research: State
of Knowledge and Perspectives, pp 269-283.
Lindqvist, O., Jernelov, A., Hobansson, K., and Rodhe, H. 1984. Mercury in the Swedish
Environment: Global and Local Sources, National Swedish Environment: Global and Local
Sources. National Swedish Environment Protection Board, Solna, Sweden, p!05.
Lindqvist, O. and Rodhe, H. 1985. Atmospheric mercury - a review, Tellus 3376:136-159.
Lindqvist, O., Johansson, K., Aastrup, M., Andersson, A., Bringmark, L., Hovsenius, G.,
Hakanson, L., Iverfeldt, A., Meili, M. and Timm, B. 1991. Mercury in the Swedish
Environment-Recent research on causes, consequences and corrective methods. Special Report
129
-------
Relative Atmospheric Loadings... Revision Date: 16 March 1993
published in Water, Soil and Air Pollution, 55.
Linton, R.W.,, A. Loh, D.F.S. Natusch, C.H. Evans and P. Williams. 1976. Surface
predominance of trace elements in airborne particles. Science 191:852-854.
Liss, P.S. and Merlivat, L. 1986. Air-Sea Exchange Rates: Introduction and Synthesis. In
The Role of Air-Sea Exchange in Geochemical Cycling. P. Buat-Menard (Ed.), NATO ASI
Series #185, D. Rediel: Dordrecht, 113-128.
Liss, P.S. and Slater, P.O. 1974. Flux of gases across the air-sea interface. Nature, 247, 181-
184.
Liss, P.S. and Slinn, W.G.N. 1983. Air-Sea Exchange of Gases and Panicles, NATO ASI
Series #108, D. Reidel: Dordrecht: New York, 561 p.
Liss, P.S. 1983. Gas transfer: experiments and geochemical implications. In: Air-Sea Exchange
of Gases and Particles, NATO Advanced Science Institutes Series, P.S. Liss and W.G. Slinn
(eds), Reidel Press, Dordrecht, pp. 241-298.
Livingstone, D.M. and Imboden, D.M. 1992. The influence of wind speed on oxygen transfer
in natural water bodies. Tellus, in press.
Logan, J.A. 1983. Nitrogen oxides in the troposphere: Global and regional budgets. J.
Geophys. Res., 88, 10,785-10,807.
Lovett, G.M. and Lindberg, S.E. 1986. Dry deposition of nitrate to a deciduous forest.
Biogeochemistry, 2, 137-148.
Luke, W.T. and Dickerson, R.R. 1987. The flux of reactive nitrogen compounds from eastern
North America to the western Atlantic Ocean. Global Biogeochem. Cycles 1, 329-343.
Lyman, W.J., Rechl, W.F., and Rosenblatt, D.H. 1991. Handbook of Chemical Property
Estimation Methods.
Lynch, J.A., Grimm, J.W., and Corbett, E.S. 1989. Atmospheric deposition: spatial and
temporal variations in Pennsylvania. Penn State Environmental Resources Research Institute,
Report ER909.
McCaffrey, RJ. and Turekian, K.A. 1980. A record of the accumulation of sediment and trace
metals in a Connecticut salt marsh. In Estuarine Physics and Chemistry: Studies in Long Island
Sound, Saltzman, B., Ed., Academic Press, London.
130
-------
Relative Atmospheric Loadings... Revision Date: 16 March 1993
Mackay, D. 1982. Effects of surface films on air-water exchange rates. J. Great Lakes Res.,
8, 299-306.
Mackay, D. and Yuen, A.T.K. 1983. Mass transfer coefficients correlations for volatilization
of organic solutes from water. Environ. Sci. Tech., 17, 211-217.
Mackay, D., Patterson, S., and Schroeder, W. 1986. Model describing the rates of transfer
processes of organic chemicals between atmosphere and water. Environ. Sci. Tech., 20, 810-
816.
Mackay, D. 1989. Modeling the long term behavior of an organic contaminant in a large lake:
application of PCBs in Lake Ontario. J. Great Lakes Res., 15, 283-297.
Mackay, D., Shiu, W.Y., Valsaraj, K.T., Thibodeaux, L.J. 1991. Air-water transfer: the role
of partitioning. In Air-Water Mass Transfer Second International Symposium, Gulliver, J.S. and
Wilhelms, S.C., Eds., U.S. Army Waterways Experiment Station/American Society of Civil
Engineering., ASCE Special Publication.
Manchester-Neesvig, J. and Andren, A.W. 1989. Seasonal variation in the atmospheric
concentration of PCB congeners. Environ. Sci. Tech., 23, 1138.
Maring, H., C. C. Patterson, and D. M. Settle. 1989. Atmospheric input fluxes of industrial
and natural Pb from the westerlies to the mid-north Pacific (in) J. Riley and R. Chester (eds.)
Chemical Oceanography, Vol. 10, Acad. Press, pp. 83-106.
Marple, V.A., Rubow, K.L., and Behm, S.M. 1991. A microorifice uniform deposit impactor
(MOUDI): description, calibration, and use. Aerosol Sci. Tech., 14, 434-446.
Mason, R.P. and Fitzgerald, W.F. 1990. Alkylmercury species in the equatorial Pacific. Nature,
347, 457-459.
Mason, R.P. and Fitzgerald, W.F. 1991. Mercury speciation of open ocean waters. Water, Air
and Soil Pollution, 56, 779-789.
Mason, R.P. and Fitzgerald, W.F. 1992. The distribution and biogeochemical cycling of
mercury in the equatorial Pacific Ocean. Deep Sea Research (in press).
Mason, R.P., Fitzgerald, W.F. and Vandal, G.V. 1992. The sources and composition of
mercury in oceanic precipitation. J. Atmos. Chem., 14, 489-500.
Maxwell, C. and Mahn, S. 1987. The spatial and temporal distribution of precipitation
chemistry across Maryland in 1984. Volumes 1 and 2. Prepared for Maryland Power Plant
131
-------
Relative Atmospheric Loadings... Revision Date: 16 March 1993
Research Program, Annapolis, MD.
McCarthy, J.F. and Jimenez, B.D. 1985. Reduction in bioavailability to bluegills of polycyclic
aromatic hydrocarbons bound to dissolved humic material. Environ. ToxicoL Chem., 4, 511-
521.
McConnell, L.L., Gotham, W.E., and Bidleman, T.F. 1993. Gas Exchange of HCH in the
Great Lakes, submitted to Environ. Sci. Tech.
McDow, S.R. 1986. The effect of sampling procedures on organic aerosol measurement.
Ph.D. Thesis, Oregon Graduate Institute.
McDow, S.R. and Huntzicker, J.J. 1990. Vapor adsorption artifact in sampling of organic
aerosol: face velocity effects. Atmos. Environ., 24A, 2563-2571.
McVeety, B.D. and Kites, R.A. 1988. Atmospheric deposition of polycyclic aromatic
hydrocarbons to water surfaces; a mass balance approach. Atmos. Environ., 22, 511-536.
Memane, Y. and Mehler, M. 1987. On the nature of nitrate particles in a coastal urban area.
Atmos. Environ., 21(9), 1989-1994.
Mierle, G. 1990. Aqueous inputs of mercury to precambrian shield lakes in Ontario. Environ.
Toxicol. Chem., 9, 843-851.
Miguel, A. and Friedlander, S.K. 1978. Distribution of B[a]P and coronene with respect to
particle size distribution in Pasadena aerosols in the submicron range. Atmos. Environ., 12,
2407-2413.
Mopper, K. and Zikal, R.G. 1987. Free amino acids in marine rain: Evidence for oxidation and
potential role in nitrogen cycling. Nature, 325, 246-249.
Morel, F.M.M. and Hudson, RJ.M. 1985. The geobiological cycle of trace elements in aquatic
systems: Redfield revisited. In Chemical Processes in Lakes, Stumm, W., Ed., John Wiley &
Sons, 251-281.
Mosher, B.W., and Duce, R.A. 1987. A global atmospheric selenium budget. /. Geophys.
Res., 92, 289-298.
Mosher, B.W. and R.A. Duce. 1983. Vapor phase and paniculate selenium in the marine
atmosphere. J. Geophys. Res., 88, 6761-6768.
132
-------
Relative Atmospheric Loadings... Revision Date: 16 March 1993
Munthe, J. 1991. The redox cycling of mercury in the atmosphere. Ph.D. Thesis, U. of
Goteborg.
Munthe, J. and Lindqvist, O. 1989. The aqueous atmospheric chemistry of mercury.
Proceedings, Nordic Symposium on Atmospheric Chemistry, Stockholm/Helsinki, December,
1989.
Murphy, T.J., Pokojowczyk, J.C., and Mullin, M.D. 1983. Vapor exchange of PCBs with
Lake Michigan: the atmosphere as a sink for PCBs. In Physical Behavior of PCBs in the Great
Lakes. D. Mackay, S. Patterson, S.J. Eisenreich, M. Simmons (Eds.), Ann Arbor Science
Publishers: Ann Arbor, MI, 49-58.
Murray, M.W. and Andren, A.W. 1992. Precipitation scavenging of polychlorinated biphenyl
congeners in the Great Lakes region. Atmos. Environ., 26A, 883-897.
Nater, E.A. and Grigal, D.F. 1992. Regional trends in mercury distribution across the Great
Lake state, north central USA. Nature 358, 139-141.
NAS 1978. An Assessment of Mercury in the Environment. National Academy of Sciences
(NAS), Washington, 185 p.
National Research Council (NRC). 1979. Subcommittee on ammonia, Ammonia, Washington,
D.C.
Natusch, D.F.S., J.R. Wallace and C.A. Evans, Jr. 1974. Toxic trace elements: preferential
concentration in respirable particles. Science 183, 202-204.
Nguyen, V.D., Merks, A.G.A., Valenta, P. 1979. Atmospheric deposition of acid, heavy
metals, dissolved organic carbon, and nutrients in the Dutch delta area in 1980-1986. Sci. Total
Environ., 99, 77.
Nicholson, K.W. 1988. The dry deposition of small particles: a review of experimental
measurements. Atmos. Environ. 22, 2653-2666.
Nixon, S.W., C.A. Oviatt, J. Frithsen and B. Sullivan. 1986. Nutrients and the productivity of
estuarine and coastal marine ecosystems. J. Limnol Soc. South Africa 12, 43-71.
Norton, S.A. and Kahl, J.S. 1992. Paleolimnological evidence of metal pollution from
atmospheric deposition. In The Deposition and Fate of Trace Metals in Our Environment,
Verry, E.S and Vermett, S.J., Eds., USDA Forest Service General Technical Report NC-150,
pp. 85-95, from the Proceedings of the National Atmospheric Deposition Program National
Trends Network, October 8, 1991, Philadelphia, PA.
133
-------
Relative Atmospheric Loadings... Revision Date: 16 March 1993
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.
Nriagu, J.O. 1989. A global assessment of natural sources of atmospheric trace metals. Nature,
338, 47-49.
Nriagu, J.O. 1992. Worldwide contamination of the atmosphere with toxic metals. In The
Deposition and Fate of Trace Metals in Our Environment, Verry, E.S. and Vermett, S.J., Eds.,
USDA Forest Service General Technical Report NC-150, pp. 9-21, from the Proceedings of the
National Atmospheric Deposition Program National Trends Network, October 8, 1991,
Philadelphia, PA.
Nurnberg, H.W., Valenta, P., Nguyen, V.D. et al. 1984. Studies on the deposition of acid
and ecotoxic heavy metals with precipitates from the atmosphere. Fresnius Z. Anal. Chem.,
317, 314-323.
Oey,L.-Y., G.L. Mellor and R.I. Hires. 1985. A three-dimensional simulation of the Hudson-
Raritan estuary. Part II. Comparison with observations. J. Phys. Oceanogr. 15, 1693-1709.
Ondov, J.M., R.C. Ragaini, and A.H. Bierman. 1979. Emissions and particle-size distributions
of minor and trace elements at two western coal-fired power plants equipped with cold-side
electrostatic precipitators. Environ. Sci. and Technol. 13, 946-953.
Owens, N.J.P., Galloway, J.N., and Duce, R.A. 1992. Episodic atmospheric nitrogen
deposition to oligotrophic oceans. Nature, (in press).
Paerl, H.W. 1985. Enhancement of marine primary production by nitrogen-enriched acid rain.
Nature, 315-747.
Paerl, H. 1988. Nuisance blooms in coastal estuaries and inland waters. Limnol. Oceanogr., 33,
823-847.
Paerl, H.W., Rudek, J., and Mallin, M.A. 1990. Stimulation of phytoplankton production in
coastal waters by natural rainfall inputs: Nutritional and trophic implications. Marine Biology,
107, 247-254.
Pankow, J.F. 1987. Review and comparative analysis of the theories on partitioning between
the gas and aerosol particulate phases in the atmosphere. Amos. Environ., 21, 2277.
Pankow, J.F. 1988. The calculated effects of non-exchangeable material on the gas-particle
distributions of organic compounds. Atmos. Environ., 22, 1405-1409.
134
-------
Relative Atmospheric Loadings... Revision Date: 16 March 1993
Patterson, C.C. and D. Settle 1976. The reduction in orders of magnitude errors in lead
analysis of biological materials and natural waters by evaluating and controlling the extent and
sources of industrial lead contamination introduced during sample collection and analysis. In
Accuracy in Trace Analysis: Sampling, Sample Handling, and Analysis, LaFleur, P.O., Ed.,
U.S. National Bureau of Standards Special Publication 422, pp. 321-351.
Patterson, C.C. and Settle, D.M. 1987. Review of data on eolian fluxes of industrial and
natural lead to the lands and seas in remote regions on a global scale. Mar. Chem., 22, 137-
162.
Patton, G.W., Hinckley, D.A., Walla, M.D., and Bidleman, T.F. 1989. Airborne
organochlorines in the Canadian High Arctic. Tellus 41B, 243-255.
Pennock, J.R. 1987. Temporal and spatial variability in phytoplankton ammonium and nitrate
uptake in the Delaware Estuary. Estuar. Coastal and Shelf Sci. 24, 841-857.
Prospero, J.M., Nees, R.T., and Uematsu, M. 1987. Deposition rate of paniculate and
dissolved aluminum derived from Saharan dust in precipitation at Miami, Florida. J. Geophys,
Res., 92,723-731.
Rada, R.G., Winfrey, M.R., Wiener, J.G. and Powell, D.E. 1987. A comparison of mercury
distribution in sediment cores and mercury volatilization from surface waters of selected northern
Wisconsin lakes. Final report, Wisconsin DNR, Bureau of Water Resources Management,
Madison, Wisconsin.
Rada, R.G., Powell, D.E. and Wiener, J.G. 1993. Whole-lake burdens and spatial distribution
of mercury in surficial sediments in Wisconsin seepage lakes. Can. Journ. Fish Aquat. Sci. (in
press).
Rahn, K.A. and Lowenthal, D.H. 1984. Elemental tracers of distant regional pollution aerosols.
Science, 223, 132-139.
Rahn, K.A. and Lowenthal, D.H. 1985. Pollution aerosol in the Northeast: Northeastern-
Midwestern contributions. Science, 228, 275-288.
Ross, H.B. 1986. The importance of reducing sample contamination in routine monitoring of
trace metals in atmospheric precipitation. Atmos. Environ. 22(5), 937-943.
Rounds, S.A. and Pankow, J.F. 1990. Application of a radial diffusion model to describe
gas/particle sorption kinetics. Environ. Sci. Tech., 24, 1378-1386.
135
-------
Relative Atmospheric Loadings... Revision Date: 16 March 1993
Sanemasa, I. 1975. The solubility of elemental mercury vapor in water. Bull. Chem. Soc. Japan
48, 1795-1798.
Savioe, D.L., and J.M. Prospero. 1982. Particle size distribution of nitrate and sulfate in the
marine atmosphere. Geophys. Res. Lett. 9, 1207-1210.
Schemel, G.A. 1980. Particle and gas dry deposition; A review. Atmos. Environ., 14, 1155-
1163.
Schomburg, C.J., Glotfelty, D.W., and Seiber, J.N. 1991. Pesticide occurrence and distribution
in fog collected near Monterey, California. Environ. Sci. TechnoL, 25, 155-160.
Schroeder, W.H. and Jackson, R.A. 1984. An instrumental analytical technique for vapour-
phase mercury species in air. Chemosphere 13, 1041-1051.
Schwarzenbach, R.P. and Westall, J. 1981. Transport of nonpolar organic compounds from
surface water to groundwater. Environ. Sci. TechnoL, 15, 1360-1367.
Schwarzenbach, R.P., Gschwend, P.M., and Imboden, D.M. 1992. Dynamic Behavior of
Xenobiotic Organic Compounds in Aquatic Systems: Basic Principles and Modeling Concepts,
Wiley Interscience, New York.
Scott, B.C. 1981. Modeling wet deposition. In Atmospheric Pollutants in Natural Waters.
Eisenreich, S.J. Ed., Ann Arbor Science, Ann Arbor, Michigan, pp. 3-21.
Scudlark, J.R. and Church, T.M. 1988. The atmospheric deposition of arsenic at the mid-
Atlantic coast of North America. Atmos. Environ., 22(5), 937-943.
Scudlark, J.R., Church, T.M., Conko, K.M., and Moore, S.M. 1992. A method for the
automated collection, proper handling and accurate analysis of trace metals in precipitation.
USD A Forest Service Report NC-150, Symposium Proceedings, Philadelphia, PA, October 8,
1991, pp. 57-71.
Scudlark, J.R., Conko, K.M., and Church, T.M. 1993. Atmospheric wet deposition of trace
metals to Chesapeake Bay: CBAD Study Year 1 results., Atmos. Environ., in press
Scudlark, J.R. and Church, T.M. 1993. The atmospheric deposition of inorganic nitrogen to
Delaware Bay. Estuaries 16, in press.
Seiber, J.N., and McChesney, M.M. 1991. Fogwater deposition as a source of inadvertent
residues on nontarget crops. Presented at the 201" American Chemical Society National
Meeting, Atlanta.
136
-------
Relative Atmospheric Loadings... Revision Date: 16 March 1993
Shendrikar, A.D., D.S. Ensor, SJ. Cowen, GJ. Woffinden and M.W. McElroy. 1983. Size-
dependent penetration of trace elements through a utility baghouse. Atmos. Environ. 17, 1411-
1421.
Singh, H.B. 1987. Reactive nitrogen in the troposphere. Environ. Sci. Technol 21, 320-327.
Sirois, A. and L. A. Barrie. 1988. An estimate of the importance of dry deposition as a pathway
of acidic substances from the atmosphere to the biosphere in eastern Canada. Tellus 40B, 59-80.
Slemr, F. Seiler, W., and Schuster, G. 1979. The determination of total gaseous mercury levels
in background air. Anal Chim. Acta 110, 35-47.
Slemr, F., Schuster, G., and Seiler, W. 1985. Distribution, speciation and budget of
atmospheric mercury. J. Atmos. Chem., 407-434.
Slemr, F. and Langer, E. 1992. Increase in global atmospheric concentrations of mercury
inferred from measurements over the Atlantic Ocean. Nature 355, 434-437.
Slinn, W.G.N., Hasse L., Hicks, B.B., Hogan, A.W., Lai, D., Liss, PS., Munnich, K.O.,
Sehmel, G.A., Vittori, O. 1978. Some aspects of the transfer of atmospheric trace constituents
past the air-sea interface. Amos. Environ., 12, 2055-2087.
Slinn, S.A. and Slinn, W.G.N. 1980. Predictions for particle deposition on natural waters.
Atmos. Environ., 14, 1013-1016.
Slinn, W.G.N. 1983. Air-to-sea transfer of particles. In Air-Sea Exchange of Gases and
Particles., Liss, P.S. and Slinn, W.G.N., Eds., D. Reidel Publisher, Dordrecht, Holland.
Smith, R., Campbell, J.A., and Nelson, K.K. 1979. Concentration dependence upon particle
size of volatilized elements in fly ash. Environ. Sci. Tech., 13, 553-558.
Smullen, J.I., J.L. Taft, and J. Macknis. 1982. Nutrient and sediment loads to the tidal
Chesapeake Bay system, p. 147-162. In Chesapeake Bay Program Technical Studies: A
Synthesis. E.G. Macalaster, D.A. Barker and M. Kasper, Eds., ,U.S. Environmental Protection
Agency, Washington, D.C.
Sproule, J.W., Shiu, W.Y., Mackay, D. 1991. Direct in situ sensing of the fugacity of
hydrophobic chemicals in natural waters. Environ. Toxicol Chem., 10, 9.
Stanley, LJ. and Kites, R.A. 1991. Chlorinated organic contaminants in the atmosphere. In
Organic Contaminants in the Environment, Jones, K.C., Ed., Elsevier Applied Sciences:
London, 1-32.
137
-------
Relative Atmospheric Loadings... Revision Date: 16 March 1993
Strachan, W.MJ. and Eisenreich, S.J. 1988. Mass balancing of toxic chemicals in the Great
Lakes: the role of atmospheric deposition. International Joint Commission Report: Windsor,
Ontario, 160 p.
Swackhamer, D.L. and Armstrong, D.E. 1986. Estimation of atmospheric and non-atmospheric
contributions and losses of PCBs for Lake Michigan on the basis sediment records of remote
lakes. Environ. Sci. Technol., 20, 879-883.
Swackhamer, D.L., McVeety, B.D., and Kites, R.A. 1988. Deposition and evaporation of PCBs
to and from Siskiwit Lake, Isle Royale, Lake Michigan, Environ. Sci. Tech., 22, 664-672.
Swackhamer, D.L. and Skoglund, R.S. 1991. The role of phytoplankton in the partitioning of
hydrophobic organic contaminants in water. In Organic Substances and Sediments in Water,
Baker, R.A., Ed., Volume 2, 91-105.
Swain, E.B., Engstrom, D.R., Brigham, M.E., Henning, T.A., and Brezonik, P.L. 1992.
Increasing rates of atmospheric mercury deposition in midcontinental North America. Science
257, 784-787.
Talbot, R.W. and Andren, A.W. 1983. Relationship between Pb and 210Pb in aerosol
precipitation at a remote site in northern Wisconsin. J. Geophys. Res., 9, 474-496.
Tanabe, S. and Tatsukawa, R. 1986. In PCBs and the Environment, Waid, J.S., Ed., Vol. I,
Chapter 8, Boca Raton, FL, CRC Press, p. 143-162..
Technicon Industrial Systems. 1973. Ammonia in water and seawater. Industrial method No. 154-
71W, Terrytown,NY.
Theotannous, T.G. 1984. Conceptual models of gas exchange. In Gas Transfer at Water
Surfaces, Jirka, G.H., Ed., D. Reidel Publishing, Higham, MA, 271-282.
Tramontane, J.M., Scudlark, J.R., and Church, T.M. 1987. A method for the collection,
handling and analysis of trace metals in precipitation. Environ. Sci. Tech., 21, 749-753.
Tschiersch, J., Hietel, B., Schramel, P. 1989. Wet deposition of aerosol: test of the method.
J. Aerosol Sci., 20(8), 1181.
Turpin, B.J., Liu, S.P., Podolske, K., Gomes, M.S.P., McMurry, P.H., Eisenreich, S.J. 1992.
Design and evaluation of a novel diffusion separator for measuring gas-particle distributions of
SOCs. submitted to Environ. Sci. Technol.
138
-------
Relative Atmospheric Loadings... Revision Date: 16 March 1993
Tyler, M. 1988. Contributions of atmospheric nitrate deposition to nitrate loading in the
Chesapeake Bay. Maryland Department of Natural Resources, Report RP 1052, VERSAR, Inc.
Upstill-Goddard, R.C., Watson, A.J., Liss, P.S., and Liddicoat, M.I. 1990. Gas transfer
velocities in lakes measured with SF6. Tellus, 42B, 364-377.
U.S. Environmental Protection Agency. 1982. Air Quality Criteria for Oxides of Nitrogen,
EPA/600/8-82-026.
Vandal, G.M., Mason, R.P., and Fitzgerald, W.F. 1991. Cycling of volatile mercury in
temperature lakes. Water, Air, Soil, and Pollution, 56, 791-863.
Van Neste, A., R.A. Duce, and C. Lee. 1987. Methylamines in the marine atmosphere.
Geophys. Res. Lett. 14, 711-714.
Van Vaeck, L., Broddin, G., and Van Cauwenberghe K. 1979. Differences in particle size
distributions of major organic pollutants in ambient aerosols in urban, rural, and seashore areas.
Environ. Sci. Tech., 13, 1494-1502.
Van Vaeck, L., Broddin, G., Cautreels, W., and Van Cauwenberghe, K. 1979. Aerosol
collection by cascade impaction and filtration: an influence of different sampling systems on the
measured organic pollutant levels. Sci. Total Environ., 11,41-52.
Van Vaeck, L. and Van Cauwenberghe, K.A. 1980. Measurement of the particle size
distribution and gas phase concentration of organic pollutants in ambient air by programmed
mass fragmentography. In Advances in Mass Spectrometry, Heyden and Son, Ltd.: Vol. VIII:
London, 1436-1450.
Van Vaeck, L., Van Cauwenberghe, K.A., Janssens, J. 1984. The gas/particle distribution of
organic aerosol constituents: measurement of the volatilization artifact in hi-vol cascade
impactor sampling. Atmos. Environ., 18, 417-430.
Van Vaeck, L. and Van Cauwenberghe, K.A. 1985. Characterization parameters of particle
size distribution of primary organic constituents of ambient aerosols. Environ. ScL Tech., 19,
707-716.
Varekamp, J.C. and Buseck, P.R. 1986. Global mercury flux from volcanic and geothermal
sources. Applied. Geochem. 1, 65-73.
Vermette, S.J., Peden, M.E., Hamdy, S., Willoughby, T.C., Schroder, L.R., Lindberg, S.E.,
Owens, J.G.,(and Weiss, A.D. 1992. A pilot network for the collection and analysis of metals
in wet deposition. In The Deposition and Fate of Trace Metals in Our Environment, Verry, E.S.
139
-------
Relative Atmospheric Loadings... Revision Date: 16 March 1993
and Vermett, S.J., Eds.), USDA Forest Service General Technical Report NC-150, pp. 73-84,
from the Proceedings of the National Atmospheric Deposition Program National Trends
Network, October 8, 1991, Philadelphia, PA.
Walcek, CJ. and Chang, J.S. 1987. A theoretical estimate of pollutant deposition to individual
land types during a regional scale acid deposition episode. Atmos. Environ., 21, 1107-1113.
Walsh, P.R., R.A. Duce and J.L. Fasching. 1979a. Tropospheric arsenic over marine and
continental regions. J. Geophys. Res. 84, 1710-1718.
Walsh, P.R., R.A. Duce and J.L. Fasching. 1979b. Considerations of the enrichment, sources,
and flux of arsenic in the troposphere. J. Geophys. Res. 84, 1719-1726.
Wanninkhof, R. 1986. Gas exchange over the air-water interface determined with man made
and natural tracers. Ph.D. thesis, Columbia University, New York.
Wanninkhof, R., Ledwell, J.R., and Broecker, W.B. 1987. Gas exchange on Mono Lake and
Crowley Lake, California. J. Geophys. Res. 92(C13), 14567-14580.
Wanninkhof, R., Ledwell, J., and Crusius, J. 1991. Gas transfer velocities on lakes measured
with SF6. In Air-Water Mass Transfer, Wilhelms, S. and Gulliver, J., Eds. ASCE, New York,
441-458.
Wanninkhof, R., Ledwell, J.R., and Broecker, W.S. 1985. Gas exchange-wind speed relation
measured with SF6 on a lake. Science, 227, 1224-1226.
Wanninkhof, R.H. and Bliven, L.F. 1991. Relationship between gas exchange, wind speed,
and radar backscatter in a large wind-wave tank. J. Geophys. Res., 96C, 2785-2796.
Watras, CJ., Bloom, N.S., Fitzgerald, W.F. Wiener, J. G., Rada, R.. Hudson, R.J.M.,
Porcella, D.G. 1993. Sources and fates of mercury and methylmercury in remote temperate
lakes, (in prep.)
Watras, CJ. and Frost, T.M. 1989. Little Rock Lake: Perspectives on an experimental
ecosystem approach to seepage lake acidification. Arch. Environ. Contain. ToxicoL, 18, 157-
165.
Watson, AJ. and Liddicoat, M.I. 1985. Recent history of atmospheric trace gas concentrations
deduced from measurements in the deep sea: application to sulphur hexafluoride and carbon
tetrachloride. Atmos. Environ., 19, 1477-1484.
140
-------
Relative Atmospheric Loadings... Revision Date: 16 March 1993
Watson, A.J.. and Ledwell, J.R. 1988. Purposefully released tracers. Phil. Trans. R. Soc.
Lond.,A325, 189-200.
Watson, A.J., Upstill-Goddard, R.C. and Liss, P.S. 1991. Air-sea gas exchange in rough and
stormy seas measured by a dual-tracer technique. Nature, 349, 145-147.
Webber, D.B. 1983. Aerial flux of paniculate hydrocarbons to the Chesapeake Bay estuary.
Marine Poll. Bull., 14,416-421.
Weiner, J. G. 1987. Metal contamination of fish in low-pH lakes and potential implications for
piscivorous wildlife. Trans. N. Am. Wildl. Nat. Resour. Conf. 52, 645-657.
Weiner, LG., Fitzgerald, W.F., Watras, C.J., and Rada, R.G. 1990. Partitioning and
bioavailability of mercury in an experimentally acidified Wisconsin lake. Environ. Toxicol
Chem.,9, 909-918.
Welch, H.E., Muir, D.C.G., Billeck, B.N., Lockhart, W.L., Brunskill, G.J., Kling, HJ.t
Olson, M.P., Lenoine, R.M. 1991. Brown snow: a long-range transport event in the Canadian
Arctic. Environ. Sci. Technol., 25, 280.
Werme, C. and Menzie, C.A. 1991. Sources and Loadings of Pollutants to the Massachusetts
Bays, Report to the Massachusetts Bays Program, MBP-91-01, 247 pp.
Westoo, G. 1966. Determination of methylmercury compounds in foodstuffs I. Methylmercury
compounds in fish, identification and determination. Acta Chem. Scand., 20, 2131-2137.
Whitman, W.G. 1923. The two film theory of gas absorption. Chem. Metall Eng., 29, 146-
148.
WHO. 1976. WHO Environmental Health Criteria 1: Mercury. World Health Organization.
Geneva.
Wilhelms, S.C. and Gulliver (Eds.) 1991. Air-Water Mass Transfer, ASCE: New York, 797
P-
Willeke, K. 1975. Performance of the slotted impactor. AIHA Journal, 36, 683-691.
Willey, J.D. and Cahoon, L.B. 1991. Enhancement of chlorophyll a production in Gulf Stream
surface seawater by rainwater nitrate. Mar. Chem., 34, 63-75.
141
-------
Relative Atmospheric Loadings... Revision Date: 16 March 1993
Williams, R.M. 1982. A model for the dry deposition of particles to natural water surfaces.
Armos. Environ. 16, 1933-1938.
Williams, E.J., S.T. Sandholm, J.D. Bradshaw, J.S. Schendel, A.O. Langford, P.Q. Quinn,
P.J. LeBel, S.A. Vay, P.O. Roberts, R.B. Norton, B.A. Watkins, M.P. Buhr, D.D. Parish,
J.G. Calvert and F.C. Fehsenfeld. 1992. An intercomparison of five ammonia measurement
techniques./. Geophys. Res. 97(D11), 11,591-11,611.
Windom, H.L., Byrd, J.T., Smith, R.G., Huan, F. 1991. Inadequacy of NASQUAN data for
assessing metal trends in the nation's rivers. Environ. Sci. TechnoL, 25, 1137-1142.
Wolff, G.T. 1984. On the nature of nitrate in coarse continental aerosols. Atmos. Environ., 18,
977-981.
Wollast, R., Billen, F. and MacKenzie, F.T. 1975. Behavior of mercury in natural systems and
its global cycle. In: Ecological Toxicology Research. NATO Science Committee Conference On
Eco-toxicology, A.D. Mclntyre and C.F. Mills (ed.), Plenum Press, New York, pp. 145-166.
Wu, Z.-Y., J.M.Ondov, J.Z. Holland, and Z.C. Lin 1992. Dry deposition fluxes of elements
in Chesapeake Bay aerosol. Int'l. Aerosol Assoc., in press.
Wu, Z.-Y., J.M. Ondov, Z.C. Lin. 1993. Chesapeake Bay atmospheric deposition study, Year
1: Spatial and temporal trends in the concentrations of selected elements in aerosol particles.
Atmos. Environ. (Submitted).
Xiao, Z.F., Munthe, J., Schroeder, W.H., and Lindqvist, O. 1991. Vertical fluxes of volatile
mercury over forest soil and lake surfaces in Sweden. Tellus 4IB, 267-279.
Xun, L., Campbell, N.E.R. and Rudd, J.W.M. 1987. Measurement of specific rates of net
methylmercury production in the water column and surface sediments of acidified and
circumneutral lakes. Can. J. Fish Aquat. Sci. 44, 750-757.
Yamasaki, H., Kuwata, K., and Miyamoto, H. 1982. Effects of ambient temperature on aspects
of airborne PAH. Environ. Sci. Technol., 16,189.
Zhang, X. and McMurry, P.H. 1991. Theoretical analysis of evaporative losses of adsorbed
or absorbed species during atmospheric aerosol sampling. Environ. Sci. Technol., 25,456-459.
Zhuang, G., Z. Yi, R.A. Duce and P.R. Brown. 1992. Link between iron and sulphur cycles
suggested by detection of Fe(II) in remote marine aerosols. Nature, 355, 537-539.
142
------- |