Ecological Indicators 36 (2014) 392399

Contents lists available at ScienceDirect

Ecological Indicators
journal homepage: www.elsevier.com/locate/ecolind

Evaluating relationships between mercury concentrations in air and

in Spanish moss (Tillandsia usneoides L.)
Kathryn T. Sutton , Risa A. Cohen, Stephen P. Vives
Department of Biology, Georgia Southern University, P.O. Box 8042, Statesboro, GA 30460, USA

a r t i c l e

i n f o

Article history:
Received 13 March 2013
Received in revised form 15 July 2013
Accepted 5 August 2013
Keywords:
Bioindicator
Atmospheric pollutant
Epiphyte
Bromeliad
Southeastern United States

a b s t r a c t
Measurement of mercury vapor is essential given that it is transported globally, and once deposited
can be converted to methylmercury, a dangerous neurotoxin. A study was conducted in southeastern
Georgia and northern Florida, USA to determine whether Spanish moss, an epiphytic vascular plant common to the southeastern United States, has the ability to retain mercury in its tissues over time, and
to detect atmospheric mercury at relatively low concentrations from nonpoint sources. Spanish moss
plants exposed to 10 and 100 ambient concentrations of mercury vapor increased tissue concentrations by 13.7 11% and 74.1 17% respectively, and then retained the mercury over two weeks following
removal from the source. There was a strong trend of increasing Spanish moss mercury concentrations
with increasing air concentration in resident populations across urban, rural inland, coastal and industrial sites. Transplanted Spanish moss around an industrial site contaminated with mercury exhibited a
164.8 8.7% increase in mercury concentration after two weeks. Mercury concentration of Spanish moss
transplanted to rural inland sites also increased after two weeks, while the change in transplant concentration at coastal sites was more variable. This study shows that Spanish moss possesses characteristics
important for use as a bioindicator of atmospheric mercury, and can potentially be adapted as a tool for
obtaining time-integrated atmospheric mercury data to add to existing atmospheric mercury monitoring
programs.
2013 Elsevier Ltd. All rights reserved.

1. Introduction
Comprehensive measurement of atmospheric mercury is essential given that its deposition can lead to methylation, and therefore
conversion to a toxin that can bioconcentrate in food webs
(Wagemann et al., 1995), be stored in sediments and biota
(Canrio et al., 2007; Williams et al., 1994) or be re-released to
the air through transpiration by plants or volatilization of the
gaseous form, Hg0 (Lindberg et al., 2002; ODriscoll et al., 2007).
Atmospheric mercury concentrations and deposition patterns are
typically estimated using mechanical methods, such as gold trap
systems (Calasans and Malm, 1997), and the Differential Absorption
Lidar technique which measures atmospheric mercury concentrations along a laser beam (Grnlund et al., 2005). However, obtaining
gaseous atmospheric mercury measurements that are spatially
comprehensive and temporally integrated using the above methods requires many replicate samples, which can be laborious and
costly (Carballeira and Fernndez, 2002; Figueiredo et al., 2007).

Corresponding author. Tel.: +1 912 478 1228; fax: +1 912 478 0845.
E-mail addresses: ks02695@georgiasouthern.edu (K.T. Sutton),
rcohen@georgiasouthern.edu (R.A. Cohen), svives@georgiasouthern.edu
(S.P. Vives).
1470-160X/$ see front matter 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.ecolind.2013.08.011

A potential solution is the use of bioindicators, which offer an

integrative method to estimate atmospheric concentrations and
deposition patterns (De Temmerman et al., 2004).
Plants are often used as bioindicators of heavy metal air pollution (e.g. Brighigna et al., 1997; Davis et al., 2007; Figueiredo et al.,
2004; Steinnes et al., 2003) largely because they have a high capacity to accumulate heavy metals into their tissue over long time
intervals (Alves et al., 2008; Calasans and Malm, 1997). A plant
that is an effective bioindicator of atmospheric heavy metals such
as mercury must not only be able to take up metal pollutants but
also retain pollutants in tissue for a period of time necessary for
tissue concentrations to be representative of average atmospheric
concentrations (Baker, 1981; Falla et al., 2000). In addition, plants
with a wide geographical distribution are effective in ascertaining impacts of atmospheric pollution over large spatial scales. For
example, lichens have been used to assess mercury contamination
surrounding specic sources, such as a chlor alkali plant (Xanthoria parietina; Grangeon et al., 2012), as well as across the Praded
and Glacensis ecoregions in Poland and the Czech Republic (Hypogymnia physodes; Kos et al., 2012). Mosses (true bryophytes) may
also be used as bioindicators of atmospheric heavy metal pollution;
Harmens et al. (2010) sampled moss species throughout Europe
from 1990 to 2005, and found lower tissue metal concentrations
coincident with the decline in heavy metal emissions during that

K.T. Sutton et al. / Ecological Indicators 36 (2014) 392399

time period. Furthermore, Suchara and Sucharov (2008) reported

similar mercury concentrations in the moss, Hypnum cupressiforme,
as well as humus, oak and pine barks around a contaminated chloralkali plant, suggesting utility as bioindicators of mercury levels.
Spanish moss (Tillandsia usneoides L.), an epiphytic bromeliad,
is a potential bioindicator of atmospheric mercury concentration
(Figueiredo et al., 2007; Husk et al., 2004). Spanish moss accumulates heavy metals from the environment in its tissue due to high
surface area, the absence of a cuticle, and high cation exchange
capacity (Alves et al., 2008; Calasans and Malm, 1997). In addition,
roots of this bromeliad are virtually non-existent or are used only
as holdfasts, meaning nutrients, water, and atmospheric pollutants
are absorbed directly from the air through scales on the plant surface (Alves et al., 2008; Amado Filho et al., 2002; Wannaz et al.,
2006). Therefore, pollutants measured in tissue may be related to
atmospheric sources without complication due to uptake from soil
or other media (Carballeira and Fernndez, 2002). This plant also
has a low sensitivity to heavy metal exposure, likely because metals
are retained in its outer scales and not translocated to internal
mesophyll bers (Alves et al., 2008; Amado Filho et al., 2002; Bastos
et al., 2004; Figueiredo et al., 2007).
Another advantage is that Spanish moss can be easily
transplanted with little mortality, which allows for estimating
deposition patterns in areas where no suitable bioindicator plants
are present, and establishing baseline measurements of deposition
in areas where there is no historical data (Falla et al., 2000). Calasans
and Malm (1997) transplanted Spanish moss inside a chlor-alkali
plant with air mercury levels between 1 and 64 g m3 , and after
15 days found tissue mercury increased to 13,500 times greater
than control plants. In addition, transplanted Spanish moss inside
and around gold shops reached 26 ppm, which was 300 times
higher than control plants, and plant concentrations decreased
with increasing distance, up to 20 m from the shops (Malm et al.,
1998). Although Spanish moss accumulates mercury either indoors
or in very close proximity to emission point sources (Calasans and
Malm, 1997; Malm et al., 1998), it is unclear whether Spanish moss
takes up mercury in proportion to atmospheric concentrations
from nonpoint emissions sources over large geographic areas.
It is also unknown whether Spanish moss retains mercury in
tissue after exposure to variable atmospheric concentrations. The
studies by Calasans and Malm (1997) and Malm et al. (1998)
showed Spanish moss transplants take up mercury when exposed
to a highly elevated and constant source of emission. However,
atmospheric mercury concentrations in other areas are typically
variable due to sporadic emissions and changing wind and precipitation patterns (Conti and Cecchetti, 2001; Schroeder and Munthe,
1998). Therefore, if Spanish moss loses tissue mercury taken up
after intermittent changes in atmospheric concentration, it would
not be representative of average air concentrations. Mercury must
be retained in tissue to effectively use Spanish moss as a biomonitor
of atmospheric mercury concentrations.
The purpose of this study was to determine the utility of Spanish
moss as an indicator of atmospheric concentrations of mercury in a
range likely to be encountered in the eld. We hypothesized that:
(1) mercury taken up by Spanish moss is retained in tissue following uptake; and (2) mercury concentrations in Spanish moss tissue
are associated with atmospheric concentrations over a wide geographic range. To test our hypotheses, we conducted eld studies in
southeastern Georgia and northern Florida, USA. We predicted that
mercury taken up by Spanish moss from the atmosphere would
be retained in tissue regardless of exposure concentration, and
expected air mercury concentrations to not only differ geographically due to nonpoint emission sources, but also that concentrations
of mercury in Spanish moss tissue would reect these differences.
Results of this study are used to address the capabilities of Spanish
moss for the temporal integration of atmospheric mercury from

393

point and nonpoint sources, and its utility as a tool to augment

existing atmospheric mercury monitoring protocols.
2. Materials and methods
2.1. Retention experiment
To determine how long mercury is retained following uptake
into Spanish moss tissue, plants with high tissue concentrations
of mercury were allowed to depurate in a low mercury environment. First, plant segments were collected haphazardly at the Salt
Marsh Ecosystem Research Facility (SERF) at Skidaway Island, GA,
USA (31 55 39 N, 81 2 33 W) from an established population
with low background tissue mercury levels (0.050.16 g Hg g1
tissue). Segments were 1020 cm of new growth from distal ends
of each plant at least 1 m from the ground to ensure plants did not
have contact with soil. In the laboratory, plants were washed with
mercury-free tap water to remove any particulate mercury on the
plant surface. Segments were then mixed together so individual
replicates would be selected at random, and a subsample (n = 6)
was analyzed for background mercury concentration.
Next, Spanish moss segments (total biomass, 200 g) were placed
into each of two glass containers and onto latticed cardboard over
a pool of volatilizing liquid elemental mercury for one week. Two
treatment groups were obtained by exposing Spanish moss to vapor
concentrations resulting in tissue mercury concentrations of either
10 (1.3 0.09 g g1 ) or 100 (12.2 0.26 g g1 ) that of ambient Spanish moss concentrations (0.13 0.003 g g1 ). Individual
segments of Spanish moss from the two treatment groups were
haphazardly tied to individual trees at least 1 m apart along a 100 m
transect in the forest at the SERF site. At 2, 7, and 14 days, 5 segments from each treatment were collected and analyzed for tissue
mercury concentration.
Spanish moss tissue mercury concentration data (in this and all
subsequent experiments) were tested for equality of variance using
Levenes test and for normality using the ShapiroWilk W test. Data
not meeting parametric assumptions were either log transformed
or analyzed with nonparametric tests. To determine whether Spanish moss tissue retained mercury and whether responses were
concentration-dependent, ANCOVA was conducted following log
transformation on treatment group (starting tissue mercury concentration) with sampling time as the covariate.
2.2. Resident population study
A eld study was conducted to examine the relationship
between mercury concentration in resident Spanish moss populations and air mercury concentration. Spanish moss from
established populations was collected in southeastern Georgia and
northern Florida, USA between March and September 2011 from
sites categorized as urban, coastal, inland, and industrial (Fig. 1).
The urban sites of Jacksonville, FL and Savannah, GA were expected
to have high mercury levels in air and in Spanish moss because
these are urban centers with large human populations (>500 people km2 ; US Census, 2011). The LCP Chemicals EPA Superfund Site
in Brunswick, GA was also predicted to have high mercury levels
because the site has a history of mercury pollution in soils and
sediments from industrial activity (Gardner et al., 1978; US EPA,
2011; Windom et al., 1976; Winger et al., 1993). In contrast, Georgia
coastal barrier islands (Cumberland Island, Ossabaw Island, and
Sapelo Island) were predicted to have lower mercury concentrations than urban and industrial sites due to limited public access,
a relative lack of development, and location >30 km from urban
centers. However, because the barrier islands are located east of
urban centers, it is possible that prevailing winds could transport

394

K.T. Sutton et al. / Ecological Indicators 36 (2014) 392399

Fig. 1. Field sites in southeastern Georgia (Magnolia Springs state park, MS; George
L. Smith state park, GLS; Private Land, PL; Savannah, SV; Ossabaw Island, OI; Sapelo
Island, SI; Cumberland Island, CI; Brunswick LCP site, BW) and northern Florida
(Jacksonville, JX) used in resident population and coastal vs. inland transplant studies.

mercury from urban and industrial sites to the coast. Therefore,

Spanish moss was also sampled from Magnolia Springs and George
L. Smith state parks, which are located west of (therefore upwind
from) urban and industrial sites. The state parks are also rural
sites that also have limited infrastructure and public access. At
each eld site, three air samples were collected (following protocol
described in Section 2.3.3). In addition, 20 strands of Spanish moss
new growth (2.5 g) were collected haphazardly from distal ends
of plants, each from an individual tree, and placed immediately in
nitric acid, xing mercury in solution. Strands were between 10 and
20 cm, which, based on slow maximum growth rates calculated by
Martin et al. (1981) (270 m day1 number of leaves1 ) represented new plant growth exposed to the atmosphere within the
growing season. Tissue mercury concentrations of Spanish moss at
sites with differing human inuence (urban, coastal, inland, industrial) were measured in the laboratory within 48 h of collection and
mean tissue concentrations were correlated to air concentrations
for each site using the Pearson correlation coefcient.
2.3. Transplant studies
Spatial patterns of atmospheric mercury uptake by Spanish
moss were assessed via transplant studies, conducted at a site
with mercury contamination and at sites without direct emission
sources. All Spanish moss plants were collected for transplanting
from the SERF site following the collection protocol described in
Section 2.1.
2.3.1. Transplants to a mercury-contaminated site
Spanish moss with low background levels of mercury was transplanted to the vicinity of the LCP Chemicals EPA Superfund site in
Brunswick, GA, a 200250 ha area consisting of mostly salt marsh
habitat, in which industrial activities have operated since 1919,
including an oil renery, a paint manufacturing company, a power
plant, and a chlor-alkali plant (US EPA, 2011). These operations
have contaminated the soils and groundwater with PCBs, volatile
organics, and mercury (US EPA, 2011). Mercury was discharged
to soils and sediments from the chlor-alkali plant at a rate of
1 kg Hg day1 for six years until 1972, resulting in high mercury
levels (up to 27 ppm) in the marsh sediments surrounding the site

(Gardner et al., 1978; Windom et al., 1976). Although 25,000 tons

of contaminated soils were removed in 1998, Frischer et al. (2000)
measured mercury in marsh sediments adjacent to the site to average 10 g g1 . Therefore, mercury may still be present surrounding
the LCP Chemicals site and emitting gaseous mercury over a broad
area.
In July 2011, three 170 m transects were established in the salt
marsh west of and parallel to the LCP Chemicals site at distances
of 0.4, 0.6, and 1.3 km. Along each transect, Spanish moss transplants were tied to 15 PVC poles and placed 12.5 m apart. Each
transplant consisted of four 5 g bundles of Spanish moss suspended
approximately 2.5 m above sediment to avoid inundation by tides.
At transplant placement, three air samples were collected (following the protocol described in Section 2.4.2) at the closest transect
(0.4 km from the LCP Chemicals site) and three were collected at the
farthest transect (1.3 km from the site) at the center of each transect. After two weeks, air sampling was repeated and transplanted
Spanish moss samples were collected and xed in nitric acid. Tissue mercury concentration of each sample was determined within
48 h. Air mercury concentration was compared across transects
and times of transplant placement and collection using two-way
ANOVA. Final tissue mercury concentration was compared to initial concentrations and to the resident population at the site using
one-way ANOVA.

2.3.2. Transplants to coastal and inland sites

Spanish moss transplants were also placed near wetlands at
both coastal and inland sites in southeastern Georgia, USA to test
whether tissue mercury reects air concentrations from nonpoint
sources over a wide geographic scale (approximately 14,000 km2 ;
Fig. 1). At coastal sites (Sapelo Island, Ossabaw Island, and Cumberland Island) transplants were placed within 100 m of salt marshes,
while transplants at inland sites (Magnolia Springs and George L.
Smith state parks, and a 45 ha, undeveloped tract of private forested
land) were placed near freshwater wetlands. Between June and
September 2011, 20 Spanish moss transplants (5 g) were tied to
individual trees 10 m apart along a 200 m transect at each site. After
two weeks, 2.5 g of each transplant was collected (excluding Magnolia Springs state park and Ossabaw Island, which were collected
after three and four weeks, respectively). In addition, Spanish moss
samples from the established, natural population were also collected for comparison to transplants at each site (with the exception
of the private land that did not have a Spanish moss population).
Spanish moss samples were placed immediately into nitric acid and
analyzed for mercury concentration within 24 h. Three air samples
were collected in the center of each transect at times of transplant placement and collection except during transplant placement
at Cumberland Island due to equipment malfunction. Final tissue
mercury concentrations of transplants to coastal and inland sites
were compared to initial tissue levels and to resident populations
present at the sites using either one-way ANOVA followed by Tukey
HSD multiple comparisons or nonparametric KruskalWallis tests
followed by MannWhitney U-tests with a Bonferroni correction
for multiple comparisons. Air mercury concentrations at transplant
placement and collection were compared using t-tests for each
coastal and inland site.

2.3.3. Spanish moss and air sample collection

For each study, 2.5 g Spanish moss samples were placed into
pre-weighed teon vials with 6 ml nitric acid (HNO3 ) in the eld to
x mercury into solution, and transported on ice and in the dark so
that mercury concentrations would not change in transit to the laboratory (Southworth et al., 1958). Air samples were collected using
a diaphragm air pump to force 10 L of ambient air over 10 min into

K.T. Sutton et al. / Ecological Indicators 36 (2014) 392399

Treatment
Time
Time treatment

df

1
3
3

1229.89
6.77
3.03

<0.0001
0.002
0.05

a glass tube lled with gold sand to amalgamate mercury (Sutton,

unpublished data).

30
Tissue Hg Concentration (g g -1)

Table 1
ANCOVA results for Spanish moss tissue Hg concentration of samples in retention
experiment (Section 3.1).

25
20
10x ambient

15

100x ambient
10
5
0
0

2.4.1. Tissue
In the laboratory, 0.5 ml of 201 Hg stable isotope spike solution
was added to solutions of Spanish moss and HNO3 and microwaved
(CEM-MDS-2100) at 71.1 C and a pressure of 30.4 kg cm2 to break
down and liquefy tissue. An 0.5 ml aliquot of each sample was
then passed through an Inductively-Coupled Plasma Mass Spectrometer with a detection limit of 0.036 ng l1 , and the ratio of the
added 201 Hg to the more abundantly occurring 202 Hg stable isotope was compared to determine the total mercury concentration
(g Hg g1 ) in each tissue sample (Smith, 1993). NOAA CRM 2976
mussel tissue standards with a known mercury concentration of
61.0 0.0036 g kg1 were also used for comparison and protocol
standardization (Smith, 1993).

8
Time (Days)

10

12

14

16

0.3
0.25
0.2
0.15
0.1
0.05
0
0

2.4.2. Air
A nichrome wire coil was placed around each gold sand-lled
glass tube and heated, which re-volatilized the amalgamated mercury. The re-volatilized mercury was swept by argon gas into a Cold
Vapor Atomic Fluorescence Spectrophotometry detector (Tekran
2500), and an integrator (HP 3394) displayed peak uorescence
from each air sample, with a detection limit of 0.13 ng l1 . Peak uorescence was compared to a standard curve of water with known
mercury concentrations to calculate the air mercury concentration
of each sample (adapted from Bloom and Fitzgerald, 1988).

Fig. 2. Mean total mercury concentration in Spanish moss tissue over two weeks in
a low atmospheric mercury environment. Tissue started with either 100 or 10
concentrations of mercury found in resident plants (n = 5).

Tissue Hg Concentration (g g )

2.4. Sample analysis

395

0.005

0.01

0.015

0.02
0.025
0.03
0.035
Air Hg Concentration (g m )

0.04

0.045

0.05

Fig. 3. Relationship between mean air and tissue mercury concentrations in resident
Spanish moss plants at urban (Jacksonville, ; Savannah, ), industrial (Brunswick
LCP, ), coastal (Ossabaw Island, ; Cumberland Island, ; Sapelo Island, ), and
inland (George L. Smith state park, ; and Magnolia Springs state park, ) sites.

3. Results
3.1. Retention experiment
Following removal from a source of mercury vapor, Spanish moss plants retained mercury for up to 15 days. Plants in
both the 100 and 10 treatments exhibited an increase in
mercury concentration over the depuration period (100, mean
increase = 74.1 17%; 10, mean increase = 13.7 11%), and there
was a strong trend toward a higher rate of increase in the100
treatment compared to the 10 treatment (Table 1 and Fig. 2).
3.2. Resident population study
There was a strong trend of increasing Spanish moss mercury
concentrations with increasing air concentration across sites (Pearson correlation; rp = 0.70; n = 8; p = 0.05; Fig. 3). The population from
the industrial LCP Chemicals site had the highest mercury concentration both in air and in Spanish moss, while urban sites had the
lowest mercury levels. Air mercury concentrations within locations
were highly variable relative to tissue mercury concentrations;
coefcients of variation (CV) for air were approximately twice those
for Spanish moss tissue across sites (32.9105.3% vs.16.960.6%,
respectively).

Fig. 4. Mean mercury concentration of Spanish moss tissue prior to transplanting

(n = 10), after two weeks of exposure at three distances from the LCP Chemicals EPA
Superfund site in Brunswick, GA (n = 15), and of a resident Spanish moss population
present at the site (n = 7). Error bars are one standard error of the mean (SEM).

3.3. Transplant studies

3.3.1. Transplants to a mercury-contaminated site
Mercury concentrations in transplanted Spanish moss increased
by 164.8 8.7% from initial concentrations after two weeks of
exposure to eld conditions at the LCP site (one-way ANOVA;
F4,56 = 20.7; p < 0.001; Fig. 4), and was similar to the resident population with a trend toward higher mercury concentration in the
resident population (TukeyKramer HSD; 0.4 km distance, p = 0.14;
0.6 km distance, p = 0.08; 1.3 km distance, p = 0.08). There was no

396

K.T. Sutton et al. / Ecological Indicators 36 (2014) 392399

D. Private Land

A. Sapelo Island
0.3

0.3

0.25
0.2
0.15

0.25

0.15

0.1

0.1

0.05

0.05

Tissue Hg Concentration (g g-1)

E. Magnolia Springs

0.3

0.3

0.25

0.25

0.2

0.2

0.15

0.15

A
0.1

0.05

0.05

C. Ossabaw Island

F. George L. Smith

0.3

0.3

0.25
0.2

B. Cumberland Island

0.1

0.2

0.25
0.2

0.15

0.15

0.1

0.1

0.05

0.05

B
A

Initial

Final

Resident

Initial

Treatment

Final Resident
Treatment

Fig. 5. Initial (n = 10) and nal (n = 20) tissue mercury concentration (g g1 ) one SEM in Spanish moss tissue transplanted to coastal (Sapelo Island [A], Cumberland Island
[B], and Ossabaw Island [C]) and rural inland (Private Land [D], Magnolia Springs [E], and George L. Smith [F]) sites in comparison to the resident population concentration
(n = 20) at each location.

difference in air mercury concentration between placement and

collection times (two-way ANOVA; F1,8 = 0.53; p = 0.49) or transect location (two-way ANOVA; F1,8 = 1.16; p = 0.31). The average
air mercury concentration (combined initial and nal air samples)
was an order of magnitude higher than that for offshore ambient
air (1.6 103 g m3 ), the standard for low air mercury concentration.

3.3.2. Transplants to coastal and inland sites

The mercury concentration of transplants increased from initial
concentrations at inland sites Magnolia Springs (one-way ANOVA;
F2,46 = 44.22; p < 0.01) and George L. Smith (KruskalWallis;
H2 = 17.44; p < 0.01), and reached nal tissue mercury concentrations similar to the resident population (TukeyKramer HSD;
MS, p = 0.15; MannWhitney; GLS, p > 0.05; Fig. 5). At the privately owned inland site without a resident population, transplants
increased from initial levels (t-test; t27 = 6.97, p < 0.01). Transplant responses at coastal sites differed by location (Fig. 5). Tissue
mercury concentrations increased in transplants to Sapelo Island
relative to initial concentrations (one-way ANOVA; F2,46 = 24.57;
p < 0.01; TukeyKramer HSD, p = 0.003) but did not reach tissue
mercury concentration of the resident population (TukeyKramer
HSD; p < 0.001). Cumberland Island transplants remained similar to initial concentrations (TukeyKramer HSD, p = 0.92) but
transplant mercury concentrations were only one-third that in
tissue from the resident population (KruskalWallis; H2 = 19.17;
p < 0.01; MannWhitney; p = 0.0006). Transplants to Ossabaw
Island exhibited no difference among initial and nal tissue

Table 2
Total air mercury concentration (g m3 ) one standard error of the mean (SEM)
measured during Spanish moss transplant placement and collection at each eld
site (n = 4).
Site

Placement

CV%

Collection

Sapelo Island
Cumberland Island
Ossabaw Island
Private Land
Magnolia Springs
George L. Smith

0.007

0.004
0.075
0.003
0.000

56.44

102.08
48.63
32.86
0.00

0.011
0.012
0.006
0.000
0.041
0.006

0.00
0.00
0.02
0.00
0.00

0.00
0.00
0.00
0.00
0.02
0.00

CV%
44.55
52.69
89.83
2.00
94.71
44.42

mercury concentrations of transplants and tissue concentration of

the resident population (KruskalWallis; H2 = 3.67; p = 0.16).
Air concentrations at inland sites were more variable and
reached higher maximum values than at coastal sites (Table 2).
Inland CVs were 4-6X higher than coastal sites, and differed more
between times of placement and collection than coastal sites. For
example, the air concentration at Magnolia Springs was more variable during transplant collection than during placement (CV was
three times higher), yet the private land site air concentration was
more variable during placement (Table 2). The absolute air concentrations at inland sites were higher during transplant collection
than deployment at Magnolia Springs (t-test; t5 = 4.16; p = 0.009)
and George L. Smith (t-test; t4 = 3.43; p = 0.03), but were lower at
the private inland site during transplant collection (t-test; t6 = 4.49;
p = 0.004). In contrast, at Sapelo and Ossabaw Islands, the air concentrations were similar between times of transplant placement
and collection (t-test; t6 = 1.49; p = 0.19 and t-test; t6 = 0.63;

K.T. Sutton et al. / Ecological Indicators 36 (2014) 392399

p = 0.55 respectively) and the direction of air concentration magnitude increased from the time of sample placement and collection
(Table 2).

4. Discussion
We found that Spanish moss retains mercury in its tissue, an
essential characteristic for a time-integrative bioindicator of atmospheric mercury concentrations. Other epiphytic plants, such as
Sphagnum girgensohnii, also exhibit strong retention of mercury
in tissue for up to 4 weeks after removal from a source of mercury vapor (Lodenius et al., 2003). However, Spanish moss not only
retained mercury regardless of air mercury concentration, but also
continued taking up mercury in the eld over two weeks despite
having reached high tissue concentrations. Therefore, saturation of
mercury in tissue is unlikely to occur when exposed to ecologically relevant concentrations of atmospheric mercury. The similar
retention and continued uptake of mercury over time by Spanish
moss exposed to different concentrations should allow for timeintegrated measurements of atmospheric mercury, making this
plant useful for development as a bioindicator.
As expected, areas with different land use and emission sources
had differing atmospheric mercury concentrations, which were
reected in resident Spanish moss tissue concentrations. The strong
positive trend between air and tissue mercury concentrations is
likely attributed to the long exposure time (and therefore period
of integration) of resident populations to atmospheric mercury
concentrations. High variability in air concentration measured at
sites was expected due to the discrete sampling method used, in
which replicate one-time snapshot measurements of air mercury
concentration were collected. Wind patterns, circulation, and precipitation can impact the presence of gaseous mercury in an area
(Morel et al., 1998; Rothenberg et al., 2010; Schroeder and Munthe,
1998). Therefore, air measurements taken at one point in time at
sites without a constant source of mercury are not likely to represent average atmospheric mercury concentrations which may be
measurable using Spanish moss tissue concentrations.
The positive trend between air and Spanish moss tissue mercury concentration revealed the highest mercury concentrations in
resident Spanish moss and in air at the Brunswick LCP site and the
lowest concentrations at the urban sites sampled. The Brunswick
LCP site was predicted to have elevated concentrations given its
history of mercury contamination; however, our nding that urban
locations had the least mercury in air and in resident Spanish moss
of all sites sampled was unexpected. A possible explanation is that
prevailing wind patterns transported gaseous mercury to or from
target sampling areas (Kellerhals et al., 2003; Rothenberg et al.,
2010). Prevailing winds in the area where samples were collected
are east- and northeastward (Weber and Blanton, 1980). Therefore,
mercury emissions emanating from urban areas and sources west
or southwest of the urban centers may have been transported away
by winds at the time of sampling. For example, several coal-red
power plants are located within 20 km of the urban sites in our
study (US EPA, 2011). Plants such as these are estimated to contribute 2233% of mercury in rainfall in the United States (Landing
et al., 2010). Therefore, it is possible that prevailing winds transported emissions from both the cities themselves, and from nearby
coal-red power plants, offshore and away from urban areas.
Wind patterns alone cannot explain why rural inland and coastal
sites had higher mercury concentrations in air and in Spanish moss
populations than urban sites. Rural inland sites were not located
downwind from urban sites. Soerensen et al. (2010) determined
there is a 210 reduction in gaseous elemental mercury concentration in the atmosphere at distances of 40120 km from emission
sources, while Carballeira and Fernndez (2002) found the lowest

397

tissue mercury concentrations in the moss species (Scleropodium

purum) 2030 km from a coal-red power plant. In the current
study, coal-red power plants were 76 km from inland sites and
therefore unlikely to contribute to gaseous mercury concentrations measured at these sites. However, wetlands were present
throughout the rural inland sites, and abundant bacteria in humic
environments increase rates of mercury methylation (Benoit et al.,
1999; Canrio et al., 2007; King et al., 2001). Methylated species
can then be taken up by organisms, or de-methylated to Hg (II) and
Hg0 and re-volatilized (Langer et al., 2001; Lindberg et al., 2002).
Therefore, local atmospheric mercury concentration is likely to be
elevated around wetlands (Williams et al., 1994; Zillioux et al.,
1993).
Tissue mercury content of Spanish moss transplanted to sites
with and without known mercury emission sources reected air
mercury concentrations. At the Brunswick LCP site, mean air
mercury concentration in the marsh surrounding the site was consistently an order of magnitude greater than background levels.
In addition, tissue concentrations in both transplants and the resident population of Spanish moss at the LCP Chemicals site were
elevated 34X above background levels reported in the literature
for Spanish moss (0.1 g g1 ; Calasans and Malm, 1997) and for
the moss species Pleurozium schreberi (0.060.09 g g1 ; Rinne and
Barclay-Estrup, 1980). Further, there was no difference in mercury
concentration in Spanish moss transplant or air measurements up
to 1.3 km from the site. These results suggest mercury emission
from the contaminated sediments throughout the site, particularly
since Malm et al. (1998) found mercury concentrations in Spanish
moss plants decreased 33-fold within 20 m from gold shops (point
sources of mercury emissions). It is possible that an atmospheric
event brought elevated concentrations of atmospheric mercury to
the site during the experiment. However, the trend toward higher
tissue mercury concentration in the resident population compared
to transplanted plants provides evidence that the area likely serves
as a constant source of gaseous mercury emissions, and indicated
that this plant is capable of rapidly responding to a constant, but
diffuse source of mercury in the eld.
At coastal and inland sites Spanish moss transplants were sensitive to even low levels of mercury from environments without a
constant source of mercury emission. With the exception of Cumberland and Ossabaw Islands, transplanted tissue had increased
mercury concentrations relative to starting concentrations. The
coastal site most likely inuenced by emission sources transported
by August northeastward winds was Sapelo Island. This site is
located 30.5 km northeast of a coal red power plant, and 35.5 km
northeast of the Brunswick LCP Chemicals site (US EPA, 2011).
However, transplants to inland sites with less potential to receive
wind-borne mercury also showed increased tissue mercury concentrations. Coastal sites did not have higher concentrations of
mercury in air and in Spanish moss than inland sites despite the
presence of greater amounts of wetland habitat at all sites. More
variable wind patterns on the coast (Weber and Blanton, 1980) may
have resulted in lower short-term coastal atmospheric mercury
concentrations, reected in lower tissue concentrations of transplants.

5. Conclusions
In this study we found Spanish moss has characteristics useful
for development as a bioindicator of atmospheric mercury concentrations. Spanish moss retains mercury in its tissue regardless of
air concentration, and the positive trend between mercury concentration in resident Spanish moss tissue and in air indicates that
this plant can be used to detect patterns in atmospheric concentration over both long periods of time and on a wide geographic

398

K.T. Sutton et al. / Ecological Indicators 36 (2014) 392399

scale. Transplants integrate atmospheric concentrations into tissue, and could potentially be used to assess average atmospheric
mercury concentrations in locations where there is no historical
data. Further, our transplant studies suggest Spanish moss may be
effective in monitoring changes in atmospheric mercury concentrations from diffuse sources on the order of weeks, to monitor
shorter-term changes in atmospheric concentrations (e.g. during
wildre events, or seasonal changes). The increased mercury concentration in Spanish moss at the polluted Brunswick LCP site
indicated that Spanish moss responds rapidly to changes in environmental mercury concentrations due to constant emissions from
a nonpoint source while transplants at inland sites and at Sapelo
Island revealed that Spanish moss takes up mercury present in
low, ambient concentrations at areas not adjacent to a known
source. Therefore, our ndings suggest that Spanish moss can
be used to compare ecologically relevant atmospheric mercury
concentrations across sites. In order to use Spanish moss tissue
mercury concentration as a quantitative measure of atmospheric
concentration, the direct relationship between air concentration
and mercury uptake by Spanish moss needs to be established.
However, a qualitative Spanish moss bioindicator that integrates
atmospheric mercury concentrations from nonpoint sources both
spatially and temporally can provide a valuable and cost-effective
tool to supplement existing monitoring programs.
Acknowledgments
We would like to thank Georgia Power (GPC Agreement #16658)
and the Georgia Southern University College of Graduate Studies
for providing funding for this study. We would also like to thank H.
Windom, C.R. Chandler, and C. Alexander for their valuable input
and assistance. Special thanks to D. Wells, C. Robertson, J. Johnson, C. Little and J. Cain for technical expertise. Finally, we thank R.
Achtziger and two anonymous reviewers for comments that greatly
improved the manuscript.
Appendix A. Supplementary data
Supplementary data associated with this article can be
found, in the online version, at http://dx.doi.org/10.1016/
j.ecolind.2013.08.011. These data include Google maps of the
most important areas described in this article.
References
Alves, E.S., Moura, B.B., Domingos, M., 2008. Structural analysis of Tillandsia usneoides
L. exposed to air pollutants in So Paulo City-Brazil. Water Air Soil Pollut. 189,
6168.
Amado Filho, G.M., Andrade, L.R., Farina, M., Malm, O., 2002. Hg localization in
Tillandsia usneoides L. (Bromeliaceae) an atmospheric biomonitor. Atmos. Environ. 36, 881887.
Baker, A.J.M., 1981. Accumulators and excluders strategies in the response of plants
to heavy metals. J. Plant Nutr. 3, 643654.
Bastos, W.R., Fonseca, M.F., Pinto, F.N., Rebelo, M.F., dos Santos, S.S., da Silveira,
E.G., Torres, J.P., Malm, O., Pfeiffer, W.C., 2004. Mercury persistence in indoor
environments in the Amazon Region, Brazil. Environ. Res. 96, 235238.
Benoit, J.M., Gilmour, C.C., Mason, R.P., Heyes, A., 1999. Sulde controls on mercury
speciation and bioavailability to methylating bacteria in sediment pore waters.
Environ. Sci. Technol. 33, 951957.
Bloom, N., Fitzgerald, W.F., 1988. Determination of volatile mercury species at
the picogram level by low-temperature gas chromatography with cold-vapor
atomic uorescence detection. Anal. Chim. Acta 208, 151161.
Brighigna, L., Ravanelli, M., Minelli, A., Ercoli, L., 1997. The use of an epiphyte (Tillandsia caput-medusae morren) as bioindicator of air pollution in Costa Rica. Sci. Total
Environ. 198, 175180.
Calasans, C.F., Malm, O., 1997. Elemental mercury contamination survey in a chloralkali plant by the use of transplanted Spanish moss Tillandsia usneoides (L.). Sci.
Total Environ. 208, 165177.
Canrio, J., Vale, M.C.C., Cesrio, R., 2007. Evidence for elevated production of
methylmercury in salt marshes. Environ. Sci. Technol. 41, 73767382.

Carballeira, A., Fernndez, J.A., 2002. Bioconcentration of metals in the moss Scleropodium purum in the area surrounding a power plant: a geotopographical
predictive model for mercury. Chemosphere 47, 10411048.
Conti, M.E., Cecchetti, G., 2001. Biological monitoring: lichens as bioindicators of air
pollution assessmenta review. Environ. Pollut. 114, 471492.
Davis, D.D., McClenahen, J.R., Hutnik, R.J., 2007. Use of the moss Dicranum montanum
to evaluate recent temporal trends of mercury accumulation in oak forests of
Pennsylvania. Northeast. Nat. 14, 2734.
De Temmerman, L., Bell, J.N.B., Garrec, J.P., Klumpp, A., Krause, G.H.M., Tonneijck,
A.E.G., 2004. Biomonitoring of air pollution with plants considerations for
the future. In: Klumpp, A., Ansel, W., Klumpp, G. (Eds.), Urban Air Pollution,
Bioindication, and Environmental Awareness. Cuvillier Verlag, Gttingen, pp.
337373.
Falla, J., Laval-Gilly, P., Henryon, M., Morlot, D., Ferard, J., 2000. Biological air quality
monitoring: a review. Environ. Monit. Assess. 64, 627644.
Figueiredo, A.M.G., Alcal, A.L., Ticianelli, R.B., Domingos, M., Saiki, M., 2004. The use
of Tillandsia usneoides L. as bioindicator of air pollution in So Paulo, Brazil. J.
Radioanal. Nucl. Chem. 259, 5963.
Figueiredo, A.M.G., Nogueira, C.A., Saiki, M., Milian, F.M., Domingos, M., 2007. Assessment of atmospheric metallic pollution in the metropolitan region of So Paulo,
Brazil, employing Tillandsia usneoides L. as biomonitor. Environ. Pollut. 145,
279292.
Frischer, M.E., Danforth, J.M., Healy, M.A.N., Saunders, F.M., 2000. Whole-cell versus
total RNA extraction for analysis of microbial community structure with 16S
rRNA-targeted oligonucleotide probes in salt marsh sediments. Appl. Environ.
Microbiol. 66, 30373043.
Gardner, W.S., Kendall, D.R., Odom, R.R., Windom, H.L., Stephens, J.A., 1978. The distribution of methyl mercury in a contaminated salt marsh ecosystem. Environ.
Pollut. 15, 243251.
Grangeon, S., Gudron, S., Asta, J., Sarret, G., Charlet, L., 2012. Lichen and soil as
indicators of an atmospheric mercury contamination in the vicinity of a chloralkali plant (Grenoble, France). Ecol. Indic. 20, 163169.
Grnlund, R., Sjholm, M., Weibring, P., Edner, H., Svanberg, S., 2005. Elemental mercury emissions from chlor-alkali plants measured by lidar techniques. Atmos.
Environ. 39, 74747480.
Harmens, H., Norris, D.A., Steinnes, E., Kubin, E., Piispanen, J., Alber, R., Aleksiayenak,
Y., Blum, O., Coskun, M., Dam, M., De Temmerman, L., Fernndez, J.A., Frolova,

K., Jeran, Z., Korzekwa,

M., Frontasyeva, M., Gonzlez-Miqueo, L., Grodzinska,

B.,
S., Krmar, M., Kvietkus, K., Leblond, S., Liiv, S., Magnsson, S.H., Mankovask,
Pesch, R., Rhling, ., Santamaria, J.M., Schrder, W., Spiric, Z., Suchara, I., Thni,
L., Urumov, V., Yurukova, L., Zechmeister, H.G., 2010. Mosses as biomonitors of
atmospheric heavy metal deposition: spatial patterns and temporal trends in
Europe. Environ. Pollut. 158, 31443156.
Husk, G.J., Weishampel, J.F., Schlesinger, W.H., 2004. Mineral dynamics in Spanish
moss, Tillandsia usneoides L. (Bromeliaceae), from Central Florida, USA. Sci. Total
Environ. 321, 165172.
Kellerhals, M., Beauchamp, S., Belzer, W., Blanchard, P., Froude, F., Harvey, B., McDonald, K., Pilote, M., Poissant, L., Puckett, K., Schroeder, B., Steffen, A., Tordon,
R., 2003. Temporal and spatial variability of total gaseous mercury in Canada:
results from the Canadian Atmospheric Mercury Measurement Network (CAMNet). Atmos. Environ. 37, 10031011.
King, J.K., Kostka, J.E., Frischer, M.E., Saunders, F.M., Jahnke, R.A., 2001. A quantitative
relationship that demonstrates mercury methylation rates in marine sediments
are based on the community composition and activity of sulfate-reducing bacteria. Environ. Sci. Technol. 35, 24912496.

Kos, A., Rrajfur, M., Srmek,

I., Wacawek, M., 2012. Mercury concentration in
lichen, moss and soil samples collected from the forest areas of Praded and
Glacensis Euroregions (Poland and Czech Republic). Environ. Monit. Assess. 184,
67656774.
Landing, W.M., Caffrey, J.M., Nolek, S.D., Gosnell, K.J., Parker, W.C., 2010. Atmospheric
wet deposition of mercury and other trace elements in Pensacola, Florida. Atmos.
Chem. Phys. 10, 48674877.
Langer, C.S., Fitzgerald, W.F., Visscher, P.T., Vandal, G.M., 2001. Biogeochemical
cycling of methylmercury at Barn Island Salt Marsh, Stonington, CT, USA. Wetl.
Ecol. Manage. 9, 295310.
Lindberg, S.E., Dong, W., Meyers, T., 2002. Transpiration of gaseous elemental mercury through vegetation in a subtropical wetland in Florida. Atmos. Environ. 36,
52075219.
Lodenius, M., Tulisalo, E., Soltanpour-Gargari, A., 2003. Exchange of mercury
between atmosphere and vegetation under contaminated conditions. Sci. Total
Environ. 304, 169174.
Malm, O., Fonseca, M.F., Miguel, P.H., Bastos, W.R., Pinto, F.N., 1998. Use of epiphyte
plants as biomonitors to map atmospheric mercury in a gold trade center city,
Amazon, Brazil. Sci. Total Environ. 213, 5764.
Martin, C.E., Christensen, N.L., Strain, B.R., 1981. Seasonal patterns of growth,
tissue acid uctuations, and 14 CO2 uptake in the crassulacean acid
metabolism epiphyte Tillandsia usneoides L. (Spanish moss). Oecologia 49,
322328.
Morel, F.M.M., Kraepiel, A.M.L., Amyot, M., 1998. The chemical cycle and bioaccumulation of mercury. Annu. Rev. Ecol. Syst. 29, 543566.
ODriscoll, N.J., Poissant, L., Canrio, J., Ridal, J., Lean, D.R.S., 2007. Continuous analysis of dissolved gaseous mercury and mercury volatilization in the upper St.
Lawrence River: exploring temporal relationships and UV attenuation. Environ.
Sci. Technol. 41, 53425348.
Rinne, R.J.K., Barclay-Estrup, P., 1980. Heavy metals in feather moss, Pleurozium
schreberi, and in soils in NW Ontario, Canada. OIKOS 34, 5967.

K.T. Sutton et al. / Ecological Indicators 36 (2014) 392399

Rothenberg, S.E., McKee, L., Gilbreath, A., Yee, D., Connor, M., Fu, X., 2010. Evidence
for short-range transport of atmospheric mercury to a rural, inland site. Atmos.
Environ. 44, 12631273.
Schroeder, W.H., Munthe, J., 1998. Atmospheric mercuryan overview. Atmos. Environ. 32, 809822.
Smith, R.G., 1993. Determination of mercury in environmental samples by isotope
dilution/ICPMS. Anal. Chem. 65, 24852488.
Soerensen, A.L., Skov, H., Johnson, M.S., Glasius, M., 2010. Gaseous mercury in coastal
urban areas. Environ. Chem. 7, 537547.
Southworth, B.C., Hodecker, J.H., Fleisher, K.D., 1958. Determination of mercury in organic compounds: a micro and semimicromethod. Anal. Chem. 30,
11521153.
Steinnes, E., Berg, T., Sjbakk, T.E., 2003. Temporal and spatial trends in
Hg deposition monitored by moss analysis. Sci. Total Environ. 304,
215219.
Suchara, I., Sucharov, J., 2008. Mercury distribution around the Spolana chlor-alkali
plant (central Bohemia, Czech Republic) after a catastrophic ood, as revealed
by bioindicators. Environ. Pollut. 151, 352361.
United States Census Bureau (US Census), 2011. http://quickfacts.census.gov/qfd/

399

United
States
Environmental
Protection
Agency
(US
EPA),
2011.
http://www.epa.gov/superfund/sites/npl/nar1458.htm
Wagemann, R., Lockhart, W.L., Welch, H., Innes, S., 1995. Arctic marine mammals
as integrators and indicators of mercury in the arctic. Water Air Soil Pollut. 80,
683693.
Wannaz, E.D., Carreras, H.A., Prez, C.A., Pignata, M.L., 2006. Assessment of heavy
metal accumulation in two species of Tillandsia in relation to atmospheric emission sources in Argentina. Sci. Total Environ. 361, 267278.
Weber, A.H., Blanton, J.O., 1980. Monthly mean wind elds for the South Atlantic
Bight. J. Phys. Oceanogr. 10, 12561263.
Williams, T.P., Bubb, J.M., Lester, J.N., 1994. Metal accumulation within salt marsh
environments: a review. Mar. Pollut. Bull. 28, 277290.
Windom, H., Gardner, W., Stephens, J., Taylor, F., 1976. The role of methylmercury
production in the transfer of mercury in a salt marsh ecosystem. Estuar. Coast.
Mar. Sci. 4, 579583.
Winger, P.V., Lasier, P.J., Geitner, H., 1993. Toxicity of sediments and pore water from
Brunswick Estuary, Georgia. Arch. Environ. Contam. Toxicol. 25, 371376.
Zillioux, E.J., Porcella, D.B., Benoit, J.M., 1993. Mercury cycling and effects in freshwater wetland ecosystems. Environ. Toxicol. Chem. 12, 22452264.