Environ. Sci. Technol. 2001, 35, 863-873
Processes Influencing Rainfall
Deposition of Mercury in Florida
JANE L. GUENTZEL*
Departments of Marine Science and Chemistry,
Coastal Carolina University, P.O. Box 261954, Conway,
South Carolina 29528-6054
WILLIAM M. LANDING
Department of Oceanography, Florida State University,
Tallahassee, Florida 32306-4320
GARY A. GILL
Department of Marine Sciences, Texas A&M University,
5007 Avenue U, Galveston, Texas 77551
CURTIS D. POLLMAN
Tetra Tech, Inc., 408 West University Avenue, Suite 301,
Gainesville, Florida 32601
The primary goal of the Florida Atmospheric Mercury
Study (FAMS) was to quantify the atmospheric deposition
of Hg throughout Florida. Monthly integrated precipitation
and weekly integrated particulate samples were collected
at 10 sites in Florida for periods ranging from 2 to 5 yr.
The monthly rainfall across the state and the concentrations
of Hg in wet-only and bulk deposition increased by a
factor of 2-3 during the summertime “wet season” (MayOctober). These parallel increases in rainfall amount and
Hg concentration resulted in 5-8-fold increases in rainfall
Hg deposition during the wet season. The annual volumeweighted Hg concentrations ranged from 14 ( 2 to 16 ( 2
ng/L across southern Florida, and the annual rainfall Hg
fluxes ranged from 20 ( 3 to 23 ( 3 µg m-2 yr-1. The weekly
integrated particulate Hg concentrations in southern
Florida were low (4.9-9.3 pg/m3) and did not exhibit strong
seasonal variability. Considering the pronounced seasonal
pattern in rainfall Hg deposition, the relatively uniform
summertime rainfall Hg concentrations, and the low
concentrations of particulate Hg, we conclude that processes
other than particulate Hg transport and scavenging
govern rainfall Hg deposition in southern Florida. We
hypothesize that long-range transport of reactive gaseous
Hg (RGM) species coupled with strong convective
thunderstorm activity during the summertime represents
>50% of the Hg deposition in southern Florida. Model
calculations indicate that local anthropogenic particulate
Hg and RGM emissions account for 30-46% of the
summertime rainfall Hg deposition across the southern
Florida peninsula.
Introduction
The discovery of elevated levels of Hg (0.4-4.4 ppm) in large
freshwater game fish from southern Florida (1) and the death,
potentially attributed to Hg toxicosis, of one endangered
* Corresponding author e-mail: jguentze@coastal.edu; phone:
(843)349-2374; fax: (843)349-2545.
10.1021/es001523+ CCC: $20.00
Published on Web 01/26/2001
 2001 American Chemical Society
Florida panther (Felis concolor coryi) from a remote region
of southern Florida (2) prompted State and Federal agencies
to investigate the sources of Hg to Florida’s ecosystems.
Previous studies have demonstrated the importance of
atmospheric transport and deposition of particulate (aerosol)
Hg as a mechanism for the delivery of Hg to remote and
pristine ecosystems (3, 4). Mercury in precipitation results
from the scavenging of aerosol Hg and reactive gaseous forms
of Hg(II) (RGM). Elemental gaseous Hg has a very low
solubility in water and must first become oxidized before it
is efficiently scavenged by precipitation events. These various
forms of Hg in the atmosphere originate from natural
processes (25%) and anthropogenic activities (75%) (5). The
relatively long atmospheric residence time for elemental
gaseous Hg (1 yr) results in an elevated global atmospheric
background of anthropogenic Hg (6). In highly urbanized
and industrialized regions, the deposition of Hg can be
strongly influenced by contributions of aerosol Hg and
reactive gaseous Hg (RGM) from local sources. Hg deposition
in rural regions is usually significantly lower than in urbanized
areas because local anthropogenic emissions are attenuated,
by deposition and dilution, as the distance from urban
sources increases. This conceptual model has been invoked
to describe the deposition of Hg in many regions (6-8).
Determining the atmospheric loading of Hg to Florida
was therefore a necessary first step toward understanding
the processes that govern the cycling of Hg in Florida’s aquatic
ecosystems. As a result, the Florida Atmospheric Mercury
Study (FAMS) was initiated in the fall of 1992. The primary
objective of the FAMS project was to quantify geographical
and seasonal gradients in atmospheric Hg deposition across
Florida. Other objectives of the study included determining
whether subregional gradients in Hg deposition occurred
and whether Hg deposition could be correlated with regions
(“hotspots”) of alarmingly high Hg concentrations (1-4 ppm)
in fish from Everglades Water Conservation Area WCA-3.
The project included studying the partitioning between wet
and dry deposition; investigating the speciation of Hg in
precipitation, throughfall (9), and gaseous Hg (10); and
identifying the possible sources of Hg using multiple chemical
tracers (11, 12). This paper discusses the concentrations of
Hg in aerosols and precipitation and the rates of rainfall Hg
deposition at 10 locations in Florida and a site in Barbados.
In addition, we propose a model to account for the sources
of the elevated rainfall Hg deposition over the southern
Florida peninsula.
Two meteorological processes that are characteristic of
the southern Florida peninsula during the “wet season”
(May-October) are the strong synoptic southeast and easterly
winds associated with the tropical North Atlantic trade winds
and the formation of deep convective thunderstorm cells.
The tall (12-16 km) convective thunderstorms are generated
when moist air from the sea breeze is carried aloft by the hot
air masses rising off the southern peninsula of Florida. In
contrast to low altitude frontal storms, tall convective
thunderstorms can scavenge particulate Hg and watersoluble RGM from the middle and upper troposphere. We
propose that the southeasterly trade winds regularly resupply
reactive gaseous Hg species to the atmosphere over the
southern peninsula of Florida from May to October. As a
result, the tall convective thunderstorms are exposed to a
fresh atmospheric burden of “background” RGM on an almost
daily basis during the summer months. We present arguments
that anthropogenic sources of RGM in Dade and Broward
Counties, primarily municipal solid waste and medical waste
incinerators, supply less than half of the daily summertime
VOL. 35, NO. 5, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
863
FIGURE 1. Locations of the nine atmospheric sampling towers (LB,
FM, FS, EG, TT, CK, AT, EN, and CV) and two ground-based sites
(EGG and FL) used in the FAMS project.
RGM burden in the southern Florida airshed. We estimate
that a greater mass of RGM enters the airshed due to daily
ventilation with background air containing scavengable RGM
and aerosol Hg. Oxidation of gaseous elemental Hg within
the airshed over southern Florida could supply an additional
8% of the daily RGM burden in the boundary layer.
Experimental Methods
Site Locations. The FAMS monitoring network consisted of
10 sites in Florida (Figure 1). The Caryville site of the Florida
Acid Deposition Network (CV) and Lake Barco (LB) in the
Katherine Ordway Reserve represented north and north
central Florida, respectively. The sites in southern Florida
were located at the Terry Park recreational complex in Ft.
Myers (FM), the ranger’s compound of the Fakahatchee
Strand State Preserve (FS), the Tamiami Trail Ranger Station
(TT) of the Everglades National Park at “40-mile bend” on
U.S. Highway 41, the Beard Research Center in the Everglades
National Park (EG), the Andytown substation of Florida Power
& Light (AT), the South Florida Water Management District
Everglades Nutrient Removal Project (EN), and the state park
on Little Crawl Key (CK). Each of these sites was equipped
with a 15-m aluminum staircase tower with outboard
sampling platforms (UpRight Inc.). The samples were collected on top of the towers to minimize the influence of
birds and insects and locally generated dust and pollen (13,
14).
Two ground-based sites were added to the network in
1995 (Figure 1). The EGG site was established in May 1995
to sample Hg in throughfall and was located under a 4-6-m
stand of brazilian pepper trees (Schinus terebinthifolius)
approximately 50 m from the EG tower. These trees have
foliage year-round. The Ft. Lauderdale site (FL) rain sampler
was positioned in July 1995 on a 1.5-m wooden sampling
platform located in a grassy field at the University of Florida
Institute of Food and Agricultural Sciences regional facility
in the Ft. Lauderdale suburb of Davie, FL. This site is
approximately 22 km east of the AT site, which borders
the eastern edge of the Everglades Water Conservation
Area WCA-3, and is less than 2 km from a municipal solid
864
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 5, 2001
waste incinerator. Samples were also collected atop the
NSF-sponsored Atmosphere-Ocean Chemistry Experiment
(AEROCE) sampling tower on remote, windward “Ragged
Point” in Barbados (BS) (14).
Sample Collection. The success of this project was
dependent upon the use of rigorous sample collection
protocols coupled with processing and analysis under “clean
lab” conditions. In the clean lab at Florida State University,
extensive detergent, solvent, and acid washing procedures
were used to clean the sampling equipment (15-17).
Thorough discussions of the QA/QC procedures, cleaning
procedures, equipment design, and sampling methods have
been published by Landing et al. (18), Guentzel et al. (15),
Pollman et al. (19), Gill et al. (10), Landing et al. (16), and Gill
and Fitzgerald (17).
Monthly integrated bulk and wet-only deposition samples
were collected at the 9 tower sites for periods ranging from
2 to 5 yr. Monthly integrated wet-only deposition samples
were collected at the two ground-based sites (EGG and FL).
Monthly integrated bulk deposition samples were collected
on the Barbados (BS) tower. The FAMS bulk deposition
collectors were fabricated using Teflon components, FEP
Teflon receiving bottles, and polycarbonate funnels (6, 15,
16). Duplicate bulk deposition samplers were deployed at
each site, and the upward-facing polycarbonate funnels were
left uncovered for the entire month-long deployment period.
Immediately before the samples were recovered, the funnels
were each rinsed with two 70-mL aliquots of ultrapure water
acidified with triple distilled sub-boiling quartz-distilled HCl
(0.045 M 3×Q-HCl). Thus, the bulk deposition samples
represent wet deposition plus any dry deposition that may
have accumulated on the inner surface of the funnel (16).
The wet deposition samples were collected using a
modified Aerochem Metrics 301 wet/dry deposition sampler
equipped with a polycarbonate roof, splash guard, Tefloncovered seal, and Teflon-coated arms (16). One 1-L polyethylene bottle (pH and major ions) and two 1-L FEP Teflon
receiving bottles were nested inside the wet bucket and were
attached to polycarbonate funnels using PTFE Teflon couplings. Rainfall amounts exceeding 27 cm/month were not
collected due to the volume limitation of our receiving bottles.
This loss amounted to less than 5% of the annual rainfall at
10 of the 12 sampling locations (Table 1). Excluding the few
periods when the samples overfilled, the excellent agreement
between the sample volumes and the rain gauge volumes
indicates that the rain events were quantitatively collected
over the course of the monthly deployment (sample/gauge
) 1.02 ( 0.01 SD; n ) 822; ref 16).
Equipment blanks and field blanks for the 30-day deployment of bulk and wet deposition were negligible (<12%) when compared to the total mass of Hg in the samples
(15, 16). The sampling and analytical methods were extensively tested and are free from significant contamination or
loss of Hg created by the 30-day integration times and the
design of the sampling equipment (15, 16). Variability for
total Hg between co-located samples ranged from 1 to 5%.
Intercomparison exercises with other laboratories have
confirmed the accuracy of the sampling and analytical
procedures used for total Hg analysis at FSU (15, 16, 20).
Particulate (aerosol) Hg was collected at 40-60 L/min on
47 mm of 0.4-µm acid-cleaned polypropylene filters (MSI,
Inc.) and integrated for 6 days out of each week (ThursdayTuesday). The open-faced polypropylene filter holders were
mounted facing outward into free air under a Plexiglas rain
shroud mounted on top of the tower. Field blank filters were
placed in identical filter holders within the rain shrouds. No
significant differences ((10%) were observed between particulate Hg samples collected for integration periods of 6
days vs summing six consecutive 24-h integrated samples
(16).
TABLE 1. Annual and Overall (1992-1996) Volume-Weighted Mean Concentrations and Fluxes of Hg for the FAMS Sampling Locationsa
VOL. 35, NO. 5, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
1993 rain (cm)
[wet] (ng/L)
[bulk] (ng/L)
wet flux (µg m-2 yr-1)
bulk flux (µg m-2 yr-1)
1994 rain (cm)
[wet] (ng/L)
[bulk] (ng/L)
wet flux (µg m-2 yr-1)
bulk flux (µg m-2 yr-1)
1995 rain (cm)
[wet] (ng/L)
[bulk] (ng/L)
wet flux (µg m-2 yr-1)
bulk flux (µg m-2 yr-1)
1996 rain (cm)
[wet] (ng/L)
[bulk] (ng/L)
wet flux (µg m-2 yr-1)
bulk flux (µg m-2 yr-1)
averaged rain (cm)
[wet] (ng/L)
[bulk] (ng/L)
wet flux (µg m-2 yr-1)
bulk flux (µg m-2 yr-1)
uncollected rain (%)e
CV
LB
FM
FS
TTb
134
11.4
11.1
15.3
14.9
148
9.4
8.9
13.9
13.2
141 ( 10
10 ( 1
10 ( 3
14 ( 1
14 ( 1
0.6
106
12.4
13.4
13.1
14.2
137
13.8
13.9
18.9
19.0
123
12.1
10.9
14.9
13.4
113
11.1
10.3
12.5
11.6
123 ( 14
13 ( 2
12 ( 2
16 ( 3
15 ( 3
3.4
136
17.2
16.4
23.4
22.3
127
18.0
17.8
22.9
22.6
165
13.3
11.7
21.9
19.3
121
17.4
16.8
21.1
20.3
137 ( 10
16 ( 2
16 ( 2
22 ( 2
22 ( 2
1.4
135
15.4
14.5
20.8
19.6
144
18.8
16.9
27.1
24.3
213
12.9
11.8
27.5
25.1
131
14.9
14.4
19.5
18.9
156 ( 38
15 ( 3
14 ( 2
23 ( 3
22 ( 3
9.8
140
17.4
16.7
24.4
23.4
150
13.1
12.1
19.7
18.2
102
15.5
15.2
15.8
15.5
131 ( 21
15 ( 3
15 ( 3
20 ( 3
20 ( 3
4.9
ENc
131
15.6
15.2
20.4
19.9
149
14.7
14.4
21.9
21.5
140(43
15 ( 2
15 ( 2
21 ( 2
21 ( 2
0.0
AT
127
14.1
12.2
17.9
15.5
104
13.2
12.7
13.7
13.2
140 ( 43
14 ( 2
12 ( 2
20 ( 6
17 ( 5
4.6
FL
123
12.9
15.9
135(16
14 ( 1
19 ( 3
4.7
EG
177
13.8
12.9
24.4
22.8
167
14.1
12.9
23.5
21.5
114
13.3
12.3
15.2
14.0
148 ( 30
14 ( 2
13 ( 2
21 ( 3
20 ( 3
3.0
EGG
114
15.4
17.6
148 ( 30
15 ( 4
21 ( 3
1.3
CK
127
10.7
9.1
13.6
11.6
68
10.9
11.1
7.4
7.5
101 ( 30
11 ( 1
10 ( 3
11 ( 1
12 ( 3
5.7
BS
100
6.5
6.5
86
6.4
5.5
93 ( 10
6.4 ( 1
4.8 ( 1
3.8
a Annual values are reported when a complete calendar year was sampled. b The April 1995 data were substituted for April 1994 because tower damage precluded sampling. c The April 1996 data were substituted
for April 1995 because tower damage precluded sampling. d Average annual values were calculated by volume-weighted averaging for 12-month periods from the onset of the site establishment. Errors were assessed
by computing a series of 12-month averages starting in January, April, July, and October of each year and then propagated using standard techniques. e Uncollected rain signifies overfilling of the receiving bottles
beyond 27 cm/month (see Figures 2-4); however, Hg fluxes were calculated using the Hg concentrations and the rain depths from the tipping-bucket rain gauge record.
9
865
Analytical Methods. The rain samples were acidified with
7.5 mL of 6 M 3×Q HCl (added prior to deployment) and 3
mL of 7.5 M 1×Q HNO3 (added after sample recovery) per
liter of sample, hermetically sealed in the FEP Teflon bottles
and digested in a low-wattage UV flux (730 µW/cm2; 254 nm)
for 48 h (15). Subsamples (100-300 mL) were reduced with
25 mL of 1 M NaBH4 followed by gas-phase stripping and Hg
amalgamation onto gold-coated quartz grains. The Hg was
quantified using cold vapor atomic fluorescence (21, 22).
The analytical detection limit, based on three times the
standard deviation of the purge blank, was 0.05 ng/L for a
sample volume of 200 mL.
The aerosol samples were digested using polymer-jacketed
PTFE Teflon microwave digestion bombs (Parr, Inc.). The
total dissolution treatment uses a mixture of 6 M 3×Q HCl/
15.9 M 1×Q HNO3/28.9 M Ultrex HF (J. T. Baker) (23). The
digest was analyzed for total Hg using SnCl2 reduction and
CVAFS (15). Using a typical air sample volume of 300 m3, the
procedural detection limit was 1 pg/m3. Frequent digestion
and analysis of NIST-2704 standard reference material
(Buffalo River Sediment) yielded 97 ( 7% (n ) 34) recovery
for total Hg.
Results and Discussion
Geographical and Seasonal Patterns of Hg in Precipitation.
The monthly integrated rainfall volumes and wet-only Hg
concentrations and fluxes for the nine Florida tower sites,
the ground-based EGG and FL sites, and the Barbados
AEROCE tower are shown in Figures 2-4. The annual volumeweighted mean rainfall Hg concentrations and fluxes are
summarized in Table 1. The annual rainfall volumes across
the southern Florida peninsula ranged from 102 to 213 cm,
with an average of 70% of the rainfall occurring during the
wet season (May-October). This rainfall pattern is generated
by the almost daily occurrence of strong convective thunderstorms across southern Florida during the summer
months. The concentrations of Hg in wet-only and bulk
precipitation samples varied seasonally, with the concentrations being 2-3 times higher during the wet season months.
The volume-weighted mean concentrations for the urban
(FM, AT, FL, EN) and rural (FS, TT, EG, EGG) sites in southern
Florida were very similar, while the northern (CV, LB) and
marine sites (CK, BS) were significantly lower (Table 1). The
monthly integrated bulk and wet deposition fluxes of Hg at
all of the sites showed strong seasonal dependency, with
70-90% of the Hg deposition occurring during the summer.
Both the rainfall volumes and rainfall Hg concentrations
increased by factors of 2-3, resulting in a 5-8-fold increase
in rainfall Hg deposition during the wet season (Figures 2-4;
Table 1). The occasional high concentration samples collected
during the dry season were associated with very low rainfall
volumes (0.1-1.0 cm) and accounted for less than 10% of
the annual rainfall Hg deposition. The annual volumeweighted Hg fluxes at the urban (FM, AT, FL, EN) and rural
(FS, TT, EG) sites in southern Florida were not significantly
different from each other, demonstrating that there was no
significant east to west gradient in Hg deposition across the
southern Florida peninsula (Table 1). The Hg fluxes from
sites located in northern and central Florida are 20-30%
lower and are similar to measurements from mid-continental
(US) areas that receive low altitude frontal storms (3, 7). The
marine sites (CK, BS) had the lowest Hg fluxes; similar to
background marine fluxes reported by refs 24-26.
This uniformity of Hg deposition across southern Florida
is in stark contrast to what we have observed for other trace
elements measured on the same samples (11, 12). The
concentrations of V, Ni, Cu, Zn, and Pb at the urban sites
(FM, AT) were 20-80% higher than at the rural sites (EG, FS,
TT), with 4-fold higher concentrations found during the
winter months. In addition, the concentrations of these
866
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 5, 2001
elements in bulk deposition samples were significantly greater
(30-100%) than the concentrations in wet-only rain samples.
These observations highlight the importance of urban aerosol
emission sources for many of the trace elements, and
emphasize how different their rainfall deposition patterns
are from what we observed for Hg.
Particulate Hg. The weekly integrated concentrations of
particulate Hg were low and did not show obvious seasonality
(Figures 2-4; Table 2). This contrasts with reports of
particulate Hg measured in the northeastern and midwestern
United States, where wintertime values were significantly
higher than summertime values (27, 28). The average volumeweighted concentrations ranged from 2.0 to 9.3 pg/m3 (Table
2). The lowest values were at the CK site and were similar to
remote marine measurements reported by Fitzgerald (24)
and Lamborg (26). The highest average particulate Hg
concentrations were observed at the urban FM site in 1994
and 1996, while the EN site had the highest annual average
concentration in 1995 (Table 2). All of these values are lower
than particulate Hg levels reported for rural areas of the midcontinental United States (27, 28). Previous studies illustrating
the dominance of local pollution sources reported concentrations of particulate Hg ranging from 22 to 1230 pg/m3 (6,
29, 30). Indeed, samples of 24-h integrated particulate Hg at
our urban FL site in August 1995 (21 ( 29 pg/m3; range 2-120
pg/m3; n ) 27; ref 16) were significantly higher than
simultaneous measurements at our suburban AT site (6.5 (
5.6; range 2-10 pg/m3; n ) 28; ref 16) and agreed well with
measurements made two summers earlier in surrounding
Broward County (20-120 pg/m3) by Dvonch et al. (31). These
data demonstrate that while the levels of particulate Hg can
be higher proximal to the urban areas of southern Florida,
the combined effects of localized wet and dry deposition
yield significantly lower levels of particulate Hg across the
rural southern Everglades. The low concentrations and
absence of strong seasonal variability in our particulate Hg
measurements suggest that aerosols do not contribute
significantly to the Hg observed in precipitation over southern
Florida.
Using our average particulate Hg concentrations and a
fine aerosol deposition velocity of 0.1 cm/s (32), we estimate
dry-deposition fluxes of particulate Hg ranging from 0.04 to
0.26 µg m-2 yr-1. Even if the deposition velocity and the
particulate Hg concentrations combined to yield an order of
magnitude increase, these fluxes are small as compared to
the average observed rates of rainfall Hg deposition in Florida
(11-23 µg m-2 yr-1; Table 1). Thus, while particle transport,
dry deposition, and rainfall scavenging have been invoked
in other studies to explain Hg concentrations in precipitation
and geographical gradients in Hg deposition (3, 6, 7, 26, 29,
30, 33), these processes do not appear to significantly
influence rainfall Hg deposition in southern Florida.
In a brief report describing the first two years of the FAMS
data, Guentzel et al. (15) showed that the wintertime rainfall
Hg concentrations decreased significantly as the rainfall
volume increased, while the summertime concentrations
seemed to remain elevated regardless of the rainfall amount.
For any rainfall species to exhibit a “washout effect”, the rate
at which that species is supplied for rainfall scavenging must
be lower than the rate at which it is actually scavenged. Our
complete data set shows that the monthly integrated
wintertime Hg concentrations do indeed exhibit the washout
effect, as do the sodium and non-sea-salt sulfate concentrations (Figure 5). However, the summertime wet season Hg
concentrations clearly do not exhibit the same washout trend
and, in fact, had generally higher and more uniform Hg
concentrations for any given rainfall amount relative to the
wintertime samples. The data suggest that the nature of the
scavengable Hg reservoir in the airshed over southern Florida
is different between the winter dry season and the summer
FIGURE 2. Time series of rainfall, rainfall Hg concentrations, and rainfall Hg deposition at the (A) Caryville, (B) Lake Barco, (C) Little Crawl
Key, and (D) Barbados AEROCE atmospheric sampling tower sites. See Figure 1 for the site locations and Table 1 for annual summaries
of the data from each site. Data gaps were due to lightning strikes, power surges, or tower damage from motor vehicles or severe storms.
wet season. During the wet season, the reservoir of scavengable Hg is either much larger and/or is more rapidly
replenished. In a separate study (9), we found that Hg
concentrations in throughfall (the EGG site; Table 1) were
similar to those in ambient rain (the EG site, Table 1), further
supporting our conclusion that dry deposition of particulate
Hg does not appear to be very significant in the Everglades
region.
Reactive Gaseous Mercury: An Important Source of Hg
in Rainfall. The low concentrations of particulate Hg, the
VOL. 35, NO. 5, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
867
FIGURE 3. Time series of rainfall, rainfall Hg concentrations, and rainfall Hg deposition at the (A) Ft. Myers, (B) Fakahatchee Strand, (C)
Tamiami Trail, and (D) Beard Research Center atmospheric sampling tower sites. See Figure 1 for the site locations and Table 1 for annual
summaries of the data from each site. Data gaps were due to lightning strikes, power surges, or tower damage from motor vehicles or
severe storms.
absence of a significant geographical gradient in rainfall Hg,
and the strong seasonal cycle of rainfall Hg deposition
observed across the southern Florida peninsula suggest that
processes other than aerosol transport and scavenging
dominate wet deposition of Hg in this region. Moreover, as
868
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 5, 2001
discussed above, a large or rapidly replenished reservoir of
scavengable Hg is required to produce the observed relationship between high amounts of rainfall and high Hg
concentrations during the wet season. Gill et al. (10) and
Guentzel et al. (15) argued that there must be a large
FIGURE 4. Time series of rainfall, rainfall Hg concentrations, and rainfall Hg deposition at the (A) Everglades Nutrient Removal Project,
(B) Andytown, and (C) Ft. Lauderdale (Davie). See Figure 1 for the site locations and Table 1 for annual summaries of the data from each
site. Data gaps were due to lightning strikes, power surges, or tower damage from motor vehicles or severe storms.
TABLE 2. Annual Average Particulate Hg Concentrations at FAMS Tower Sitesa
particulate Hg (pg/m3)
1994
station
Caryville
Lake Barco
Ft. Myers
Fakahatchee Strand
Tamiami Trail
Everglades Nutrient Removal Project
Andytown
Beard Research Center
Little Crawl Key
a
1995
1996
overall
site ID mean ( SD median mean ( SD median mean ( SD median mean ( SD median
CV
LB
FM
FS
TT
EN
AT
EG
CK
b
4.8 ( 2.3
7.9 ( 7.0
5.0 ( 2.2
4.6 ( 2.0
b
4.5 ( 2.0
4.1 ( 2.2
1.8 ( 1.3
4.3
5.9
4.9
4.5
4.2
3.8
1.4
5.0 ( 3.1
7.5 ( 8.6
7.0 ( 3.3
4.6 ( 1.9
4.0 ( 2.2
8.2 ( 4.4
5.5 ( 4.0
4.7 ( 2.9
1.6 ( 1.3
4.2
4.3
6.1
4.7
3.5
8.0
4.3
3.8
1.1
6.6 ( 3.6
5.1 ( 3.1
11.8 ( 7.3
6.2 ( 2.7
6.1 ( 3.6
10.2 ( 5.3
6.5 ( 3.5
5.3 ( 3.7
2.4 ( 1.4
Samples are integrated for 6 full days per week (Thursday-Tuesday). All concentrations are in pg/m3.
contribution from highly reactive water-soluble gaseous
Hg(II) species (RGM) to the rainfall deposition of Hg in
Florida. They further pointed out that the relatively uniform rainfall Hg concentrations and fluxes across southern
b
6.0
4.6
9.5
5.6
5.4
9.0
5.8
3.9
2.1
6.0 ( 3.5
5.7 ( 5.4
9.0 ( 6.3
5.2 ( 2.3
5.0 ( 2.8
9.3 ( 5.0
5.7 ( 3.6
4.9 ( 3.3
2.0 ( 1.4
5.2
4.4
7.3
5.1
4.5
8.5
4.9
4.0
1.6
Site not yet established.
Florida required that the scavengable Hg coming from all
sources into the airshed had to be relatively uniformly
distributed on time scales of days to weeks during the wet
season.
VOL. 35, NO. 5, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
869
FIGURE 5. Rainfall Hg concentration vs rainfall amount for monthly integrated samples from (A) urban sites, (B) rural sites, (C) rural sites
(sodium), and (D) rural sites (non-sea-salt sulfate). The volume-weighted average Hg concentrations are provided for the wet and dry
seasons.
To a first approximation, the anthropogenic emission
sources in Dade and Broward Counties can be thought of as
RGM plumes originating at the base of an open window,
along the southeastern coast of Florida, through which the
trade winds blow during the wet season. These plumes are
driven generally westward and are diluted by ambient
background air. Tall convective thunderstorms, with a
frequency on the order of 15-20 per month at each FAMS
site, scavenge the RGM from the airshed and deposit it
relatively evenly across the area covered by the FAMS network
in southern Florida at a rate of 3-5 µg m-2 month-1 (g3
kg/day over the Florida peninsula south of the southern shore
of Lake Okeechobee; 34 × 109 m2). These tall convective
thunderstorms are reported to entrain approximately 60%
of their air from the boundary layer and 40% from the free
troposphere (34). Thus, we can estimate what fraction of the
Hg in wet deposition results from the various RGM sources
by evaluating the impact each source has on the daily
inventory of RGM in the boundary layer and in the free
troposphere over southern Florida. The means and standard
deviations in the following estimates result from a Monte
Carlo simulation (n ) 1000) where the various input
parameters in each calculation are allowed to vary randomly
over the quoted parameter range.
There are relatively few quantified sources of RGM to the
airshed over southern Florida. Dvonch et al. (35) estimated
that the main anthropogenic RGM emission sources, municipal and medical waste incinerators and oil combustion
sources, in the Miami/Ft. Lauderdale region emitted 2.5 (
0.5 kg of RGM/day. In the Florida Pilot Mercury TMDL Study
(36), all Hg emission sources in southern Florida cataloged
870
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 5, 2001
in the EPA emissions inventory were adjusted using the total
Hg emissions values and RGM concentrations reported by
Dvonch et al. (35), yielding a range of aerosol Hg plus RGM
emission from 2.6 to 4.0 kg/day.
During the wet season in southern Florida, the source for
background air is primarily long-range transport due to the
synoptic summertime trade winds. Accepting the global
atmospheric Hg residence time of 1 yr (5), which requires a
net oxidation rate of 0.3% per day, a boundary layer air mass
that does not experience frequent scrubbing by rain events
would be expected to gain on the order of 3-5 pg of RGM
m-3 day-1. Trade wind transport also carries large quantities
of African dust over southern Florida, demonstrating that a
significant fraction of the air mass escapes scrubbing during
the week-long transport time across the Atlantic (37). It is
therefore reasonable to expect these same air masses to arrive
over southern Florida with a burden of accumulated RGM
plus aerosol Hg. Time series measurements of RGM at our
remote CK site showed concentrations increasing from 5
pg/m3 in the winter to 30 pg/m3 in the summer (38).
Measurements of RGM in the marine troposphere east of
Miami (aircraft sampling between 800 and 2600 m) showed
RGM values ranging from 50 to 340 pg/m3 during the late
summer (39). These RGM measurements were all made using
an ion-exchange filter pack technique (39) that has been
shown to generally underestimate RGM concentrations
relative to a mist chamber technique (32) or relative to KClcoated annular denuders (40). It would be clearly incorrect
to assert that there should be no scavengable Hg in the
background air masses crossing over southern Florida, and
our data would suggest that the background boundary-layer
RGM concentrations are at least 25 pg/m3.
Whether local RGM sources in southern Florida or RGM
associated with background air masses dominates the wet
and dry deposition of RGM in southern Florida thus depends
on the “replacement time” for RGM in the airshed over the
southern peninsula:
long-range RGM transport rate (g/day) )
[RGM]bkg (g/m3) × A (m2) × H (m)
T (day)
where T is the replacement (or ventilation) time, [RGM]bkg
is the concentration in the background air, A is the area of
the southern Florida peninsula south of Lake Okeechobee
(34 × 109 m2), and H is the layer height (2600 m for the
boundary layer and 9400 m for the free troposphere (41, 42).
The observed [RGM]bkg of 25-35 pg/m3 and a replacement
time of 1 day yields long-range RGM transport rates of 2-3
kg/day, comparable to the local emission rates. The same
analysis can be applied to the free troposphere, where local
RGM emissions would be a minor term. Because of the much
deeper layer, the background [RGM] required to generate
2-3 kg/day is only 6-10 pg/m3 for a 1-day replacement time.
Another source for RGM that is not related to local RGM
emission is the homogeneous gas-phase oxidation of elemental Hg vapor by ozone. Using an average ozone
concentration of 40-60 ppb (1.4 × 1018 atoms/m3), an
elemental Hg vapor concentration of 1.5 ng/m3 (4.5 × 1012
atoms/m3), and a second-order rate constant of 2.6 × 10-21
m3 atom-1 day-1 (43) yields RGM production of 4-6 pg m-3
day-1. While ozone oxidation is not the only atmospheric
reaction producing RGM (44), it is perhaps the one that is
the best characterized. Oxidation of elemental gaseous Hg
would therefore produce 0.43 ( 0.10 kg/day of RGM in the
boundary layer over southern Florida given a production
rate of 4-6 pg m-3 day-1, a boundary layer depth of 18003200 m, and an area of 34 × 109 m2. This represents 17 ( 5%
of the local RGM emission. In the free troposphere, we assume
that the average RGM production rate from the bimolecular
ozone oxidation of elemental Hg would be reduced by 75%
(due to the reduced pressure) to 1-1.5 pg m-3 day-1. RGM
production by this mechanism in the free troposphere (260012 000 m over 34 × 109 m2) would then yield 0.38 ( 0.10
kg/day.
Thunderstorm scavenging from the boundary layer during
the summer wet season would therefore include roughly
equal proportions of RGM from local sources (2.5 ( 0.5 kg/
day) and from long-range transport (2-3 kg/day) plus insitu oxidation of elemental Hg (0.43 kg/day). RGM scavenged
from the free troposphere would come primarily from longrange transport (>2.5 kg/day) plus in-situ elemental Hg
oxidation (0.38 kg/day). Local RGM sources would therefore
account for approximately 30 ( 7% of the Hg in wet deposition
(2.5 ( 0.5 kg/day out of 8.3 ( 0.9 kg/day total, assuming
equivalent thunderstorm entrainment and RGM scavenging
efficiency for the boundary layer and the free troposphere).
Local sources would account for 46 ( 11% of the Hg in
summertime rainfall (2.5 ( 0.5 kg/day out of 5.4 ( 0.7 kg/day
total) if one made the extreme assumption that there is no
RGM scavenging from the free troposphere.
What fraction of the rainfall RGM deposition is expected
to come from local sources over the course of a full year?
From our data, we see that roughly 16% of the annual rainfall
Hg deposition in southern Florida occurs during the November-April dry season. If we make an extreme assumption
that long-range transport disappears completely while insitu elemental Hg oxidation matches the wet season value
(0.43 ( 0.1 kg/day), then local RGM sources (2.5 ( 0.5 kg/
day) would account for 85 ( 23% of the daily boundary layer
RGM supply during the dry season. Since winter (frontal)
storms are less likely to scrub much of the free troposphere,
we assign e14% of the annual rainfall Hg flux to local RGM
sources during the dry season (85% local source supply times
16% of the annual rainfall Hg deposition). During the wet
season, local RGM sources account for 25-39% of the annual
rainfall Hg deposition (30-46% local source supply times
84% of the annual rainfall Hg deposition).
These conclusions contrast sharply with those reached
by Dvonch et al. (35, 45). During the SoFAMMS experiment,
they collected over 300 24-h integrated rain samples in the
Miami/Ft. Lauderdale area during August 1995 and analyzed
them for Hg and a number of other trace elements. Using
multi-variate statistical processing and detailed analysis of
individual storm events, they concluded that local Hg
emission sources had a significant impact on the rainfall Hg
concentrations and deposition, accounting for 71-73% of
the rainfall Hg deposition in their study domain. According
to our calculations, local RGM sources could account for
more than 70% of the wet deposition across southern Florida
only by making the extreme assumptions that there is no
RGM scavenging from the free troposphere and that the
scavengable RGM plus aerosol Hg in the background
boundary layer air is less than 5 pg/m3.
Furthermore, the FAMS data from August 1995 show that
the rainfall Hg concentrations and deposition at many of the
other FAMS sites in southern Florida (Figures 2-4) were
significantly higher than at the SoFAMMS Davie and Andytown sites (45). The Fakahatchee Strand site, over 60 km
southwest of the Andytown site, had the highest rainfall Hg
deposition of the FAMS sites during August 1995 (18 ng/L;
9.1 µg/m2). The CK site (19 ng/L; 4.37 µg/m2) is even farther
away, over 150 km to the south-southwest of the Andytown
site (15 ng/L; 4.35 µg/m2). One would expect rainfall washout
of particulate Hg and RGM from local emission sources in
the Miami/Ft. Lauderdale area to yield progressively lower
(not higher) rainfall Hg concentrations and deposition as
one moved away from the urban area.
The MM5 meteorological model was used to generate
forward trajectories from the urban emission sources and
back trajectories from the CK site for every rain event during
August 1995 at the CK site (46). Combining trajectories with
daily rainfall amounts, Green et al. (46) concluded that less
than 16% of the rainfall Hg collected at the CK site during
August 1995 was associated with air masses that had passed
near the RGM emission sources in the Miami/Ft. Lauderdale
urban area.
Our conclusions also conflict with model results from
Bullock (47), who used the EPA Hg emissions inventory and
the RELMAP model to predict Hg deposition across the
continental United States. In southern Florida, the RELMAP
model predicted a very strong east-to-west geographical
gradient in rainfall Hg deposition, from >30 µg m-2 yr-1 near
urban Miami/Ft. Lauderdale to <10 µg m-2 yr-1 in the remote
Everglades (47). In contrast, our data shows nearly uniform
rainfall Hg deposition across this region, including suburban
Ft. Lauderdale to the east (the AT and FL sites; Table 1), and
urban Ft. Myers to the west (the FM site). Most of the rainfall
Hg deposition in the RELMAP model was due to locally
emitted RGM because the model simulations did not include
any RGM or particulate Hg in the background air masses (47,
48). As we concluded above and as Constantinou et al. (49)
demonstrate in their plume dispersion/deposition model,
the fraction of rainfall Hg that is due to RGM from any local
source is completely governed by how much scavengable
RGM and particulate Hg are brought into the model domain
with the background air.
There is further evidence that municipal and medical solid
waste (MSW) incinerators in southern Florida are not the
dominant RGM sources for rainfall Hg over southern Florida.
Atkeson (50) reports that MSW incinerators in southern
VOL. 35, NO. 5, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
871
Florida have reduced their Hg emissions by almost a factor
of 10 since 1994. Despite this dramatic reduction in Hg
emission, the wet deposition of Hg at the EG site has not
decreased significantly. Our data in Figure 4 shows Hg wet
deposition of >17 µg m-2 yr-1 for an incomplete sampling
year in 1993 at the EG site and then Hg wet deposition of 24,
24, and 15 µg m-2 yr-1 in 1994, 1995, and 1996, respectively.
Data from the MDN network show Hg wet deposition at the
EG site (FL-11) of 17 µg m-2 yr-1 in 1996, 27 µg m-2 yr-1 in
1997, 20 µg m-2 yr-1 in 1998, 17 µg m-2 yr-1 in 1999, and 17
µg m-2 yr-1 in 2000 (January-September) (51). It appears
that the large reduction in local MSW incinerator Hg
emissions has not resulted in a significant decrease in rainfall
Hg concentrations or wet deposition of Hg at the EG site. In
contrast, it was recently demonstrated that significant
decreases in bulk Hg precipitation and deposition coincided
with reductions in anthropogenic Hg emissions from a remote
site in northern Wisconsin (52).
Dry Deposition of RGM. What impact do these various
RGM sources have on RGM dry deposition? Previous RGM
modeling studies have noted the significance of rapid
localized dry deposition, which depletes the RGM reservoir
before wet deposition can occur farther downwind (47). If
local emission sources of RGM (plus particulate Hg) are in
fact responsible for less than half of the daily RGM supply
in the summertime boundary layer, then those sources would
account for less than half of the RGM dry deposition. During
the winter, when long-range RGM transport is perhaps less
important, local emission sources could dominate the supply
of RGM to the boundary layer and may account for g85%
of the wintertime dry deposition.
Estimates of RGM dry deposition to Everglades vegetation
(16 µg m-2 yr-1; ref 9) are slightly lower than our measured
wet deposition rates (15-24 µg m-2 yr-1, Table 1). Extrapolating over the entire 34 × 109 m2 of southern Florida, dry
deposition of RGM would come to 1.5 kg/day. Added to the
2.5-3 kg/day summertime wet deposition of Hg (total wet
+ dry g4 kg/day), it appears that the identified local RGM
emission sources (2.5 ( 0.5 kg/day) are simply not adequate
to supply all of the necessary RGM, even if one could
somehow capture 100% of the emissions each day over the
southern Florida peninsula.
Implications for the Everglades Region. Our RGM source/
sink inventory described above is a first-order explanation
of Hg deposition and the complicated summertime meteorology in southern Florida. Continued efforts to quantify
background RGM in the boundary layer and the free
troposphere; to quantify local emission sources; and to model
Hg emission, transport, and scavenging will further explain
and reconcile the FAMS and SoFAMMS data. Lagrangian
deposition models (47, 48) and statistical source apportionment models (35) that aspire to explain the deposition of Hg
in southern Florida should be able to simulate the strong
seasonality we demonstrate with the FAMS rainfall Hg
deposition data as well as the uniformity in Hg deposition
across the entire southern Florida peninsula. Such modeling
efforts must include the impacts of RGM and particulate Hg
in background air.
The contrast between uniform Hg deposition and geographical “hot spots” in fish Hg concentrations from the
Everglades region further suggests that aquatic/terrestrial
Hg cycling processes, rather than atmospheric source
strength, are responsible for the hot spots. Ecosystem Hg
cycling models for the Everglades should be structured to
simulate the effects of the summertime pulses of reactive Hg
in rainfall.
Finally, it is important to recall that 60-70% of the Hg in
the modern global atmosphere results from anthropogenic
industrial activities (5). From our analysis of the data, we
conclude that significant reduction in rainfall Hg deposition
872
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 5, 2001
over the Florida Everglades will likely require reductions in
local and global Hg emissions.
Acknowledgments
This project was supported financially by grants from the
Florida Department of Environmental Protection, Florida
Power and Light, the Electric Power Research Institute, the
South Florida Water Management District, the U.S. Environmental Protection Agency, and the Florida Electric Power
Coordinating Group. The authors’ conclusions do not imply
endorsement by these funding agencies. We thank the Florida
Park Service, the National Park Service, the Keys Marine
Laboratory, the Ordway Preserve, and Ken Larson of the
Broward County Air Monitoring Program for providing
logistical support. We gratefully recognize the invaluable
assistance provided by the graduate students and support
personnel employed by FSU, Texas A&M, and KBN Engineering. We specifically acknowledge the following: John
Cooksey, Laurel Buttermore, Geoffrey Schaefer, Johan Schijf,
Jerome J. Perry Jr., Scott Sigler, Stephanie Smith-Moore, Jill
Brandenburger, Dave Oliff, Jim Winnie, Paul Ruscher, Melody
Owens, Matt Green, Ron Lehman, Dave Bare, Mike Arrants,
John and Gayle Swanson, Cornelius Shea, and the NSF/
AEROCE Program.
Literature Cited
(1) Lange, T. Florida Game and Freshwater Fish Commission,
personal communication.
(2) Jordan, D. Endangered Species Tech. Bull. 1990, XV (2).
(3) Sorenson, J. A.; Glass, G. E.; Schmidt, K. W.; Huber, J. K.; Rapp,
G. R., Jr. Environ. Sci. Technol. 1990, 24, 1716-1727.
(4) Fitzgerald, W. F.; Mason, R. P.; Vandal, G. M. Water Air Soil
Pollut. 1990, 56, 745-768.
(5) Mason, R. P.; Fitzgerald, W. F.; Morel, F. M. M. Geochim.
Cosmochim. Acta 1994, 58, 3191.
(6) Iverfeldt, A. Water Air Soil Pollut. 1991, 56, 251-263.
(7) Sorenson, J. A.; Glass, G. A.; Schmidt; K. W. Environ. Sci. Technol.
1994, 28, 2025-2032.
(8) Jensen A.; Iverfeldt, A. In Mercury Pollution: Integration and
Synthesis; Huckabee, J. W., Watras, C. J., Eds.; Lewis Publishers:
Boca Raton, FL, 1994; pp 221-229.
(9) Guentzel, J. L.; Landing, W. M.; Gill, G. A.; Pollman, C. D. Sci
Total Environ. 1998, 213, 43-51.
(10) Gill, G. A.; Guentzel, J. L.; Landing, W. M.; Pollman, C. D. Water
Air Soil Pollut. 1995, 80, 235-244.
(11) Landing, W. M.; Perry, J. J., Jr.; Guentzel, J. L.; Gill, G. A.; Pollman,
C. D. Water Air Soil Pollut. 1995, 80, 343-352.
(12) Perry, J. J., Jr.; Landing, W. M.; Guentzel, J. L.; Gill, G. A.; Pollman,
C. D. Abstract submitted to the Fourth International Conference
on Mercury as a Global Pollutant, Hamburg, Germany, August
4-8, 1996.
(13) Duce, R. A. Chemical Oceanography 10; Academic Press:
London, 1989; pp 1-14.
(14) Savoie, D. L.; Prospero, J. M.; Saltzmann, E. S. J. Geophys. Res.
1989, 94, 5069-5080.
(15) Guentzel, J. L.; Landing, W. M.; Gill, G. A.; Pollman, C. D. Water
Air Soil Pollut. 1995, 80, 393-402.
(16) Landing, W. M.; Guentzel, J. L.; Gill, G. A.; Pollman, C. D. Atmos.
Environ. 1998, 32, 909-918.
(17) Gill, G. A.; Fitzgerald, W. F. Mar. Chem. 1987, 20, 227-243.
(18) Landing, W. M.; Guentzel, J. L.; Gill, G. A.; Pollman, C. D.; Bare,
D. A. Report to the Florida Department of Environmental
Protection, Tallahassee, FL, 1993
(19) Pollman, C.; Gill, G.; Landing, W.; Guentzel J.; Bare, D.; Porcella,
D.; Zillioux, E.; Atkeson, T. Water Air Soil Pollut. 1995, 80, 285290.
(20) Bloom, N. S.; Horvat, M. A.; Watras C. J. Water Air Soil Pollut.
1995, 80, 1257-1268.
(21) Fitzgerald, W. F.; Gill, G. A. Anal. Chem. 1979, 51, 1714-1720.
(22) Gill, G. A.; Bruland, K. W. Environ. Sci. Technol. 1990, 24, 13921400.
(23) Eggimann, D. W.; Betzer, P. R. Anal. Chem. 1974, 48, 886-890.
(24) Fitzgerald, W. F. Chemical Oceanography 10; Academic Press:
London, 1989; pp 152-180.
(25) Gill, G. A.; Fitzgerald, W. F. Global Biogeochem. Cycles 1987, 1,
199-213.
(26) Lamborg, C. H.; Rolfhus, K. R.; Fitzgerald, W. F.; Kim, G. DeepSea Res. II 1999, 46, 957-977.
(27) Burke, J.; Hoyer, M.; Keeler, G.; Scherbatskoy, T. Water Air Soil
Pollut. 1995, 80, 353-362.
(28) Keeler, G. J.; Glinsorn, G.; Pirrone, N. Water Air Soil Pollut. 1995,
80, 159-168.
(29) Keeler, G. J.; Hoyer, M. E.; Lamborg, C. H. In Mercury Pollution:
Integration and Synthesis; Huckabee, J. W., Watras, C. J., Eds.;
Lewis Publishers: Boca Raton, FL, 1994; pp 231-241.
(30) Lamborg, C. H.; Fitzgerald, W. F.; Vandal, G. M.; Rolfhus, K. R.
Water Air Soil Pollut. 1995, 80, 189-198.
(31) Dvonch, J. T.; Vette, A. F.; Keeler, G. J.; Evans, G.; Stevens, R.
Water Air Soil Pollut. 1995, 80, 169-178.
(32) Lindberg, S. E.; Stratton, W. J. Environ. Sci. Technol. 1998, 32,
49-57.
(33) Hoyer, M.; Burke, J.; Keeler, G. J. Water Air Soil Pollut. 1995, 80,
199-208.
(34) Petersen, G.; Munthe, J.; Pleijel, K.; Bloxam, R.; Kumar, A. V.
Atmos. Environ. 1998, 32, 829-843.
(35) Dvonch, J. T.; Graney, J. R.; Keeler, G. J.; Stevens, R. K. Environ.
Sci. Technol. 1999, 33, 4522-4527.
(36) Florida Department of Environmental Protection, Florida Pilot
Mercury TMDL Study (Draft), 2000, http://www.dep.state.fl.us//
labs/hg/docs/tmdldraft/.
(37) Carlson, T. N.; Prospero, J. M. J. Appl. Meteorol. 1972, 11, 283297.
(38) Gill, G., Unpublished data.
(39) Prestbo, E. Mercury Speciation in the Boundary Layer and Free
Troposphere Advected to South Florida: Phase I-Reconnaissance;
Report Submitted to the Florida Department of Environmental
Protection, Tallahassee, FL, 1996.
(40) Sheu, G.-R.; Mason, R. P. Environ. Sci. Technol. (submitted for
publication).
(41) Kara, A. B.; Ruscher, P. H.; Elsner, J. B. Weather Forcasting 1998,
13, 921-933.
(42) Ruscher, P. Department of Meteorology, Florida State University,
personal communication.
(43) Hall, B. Water Air Soil Pollut. 1995, 80, 301-315.
(44) Seigneur, C.; Wrobel, J.; Constantinou, E. Environ. Sci. Technol.
1994, 28, 1589-1597.
(45) Dvonch, J. T.; Graney, J. R.; Marsik, F. J.; Keeler, G. J.; Stevens,
R. K. Sci. Total Environ. 1998, 213, 95-108.
(46) Green, M. Masters Thesis, Department of Meteorology, Florida
State University, 1998; 61 pp.
(47) Bullock, O. R.; Brehme, K. A.; Mapp, G. R. Sci. Total Environ.
1998, 213, pp 1-12.
(48) Bullock, O. R. In Mercury Study Report to Congress; U.S. EPA:
Washington, DC, 1997; 5-1-5-31.
(49) Constantinou, E.; Wu, X. A.; Seignur, C. Water, Air, Soil Pollut.
1995, 80, 325-335.
(50) Atkeson, T. Florida Department of Environmental Protection,
personal communication.
(51) National Atmospheric Deposition Program (NRSP-3)/Mercury
Deposition Network. NADP Program Office, Illinois State Water
Survey, 2204 Griffith Drive, Champaign, IL 61820, 2000.
(52) Watras, C. J.; Morrison, K. A.; Hudson, R. J. M.; Frost, T. M.;
Kratz, T. K. Environ. Sci. Technol. 2000, 34, 4051-4057.
Received for review July 26, 2000. Revised manuscript received December 4, 2000. Accepted December 5, 2000.
ES001523+
VOL. 35, NO. 5, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
873