Determination of dissolved gaseous mercury in seawater of Minamata Bay and estimation for mercury exchange across air–sea interface

Marine Chemistry 168 (2015) 9–17

Contents lists available at ScienceDirect

Marine Chemistry journal homepage: www.elsevier.com/locate/marchem

Determination of dissolved gaseous mercury in seawater of Minamata Bay and estimation for mercury exchange across air–sea interface Kohji Marumoto ⁎, Shoko Imai Environmental Chemistry Section, National Institute for Minamata Disease, 4058-18 Hama, Minamata-shi, Kumamoto 867-0008, Japan

a r t i c l e

i n f o

Article history: Received 1 February 2014 Received in revised form 20 August 2014 Accepted 8 September 2014 Available online 13 October 2014 Keywords: Dissolved gaseous mercury Mono methyl mercury Total Hg Hg evasion Air–sea exchange

a b s t r a c t Dissolved gaseous Hg (DGM) in seawater and atmospheric gaseous Hg were measured at six sites in Minamata Bay to investigate mercury (Hg) evasion flux from the sea surface. Minamata Bay was severely polluted with mono-methyl Hg (MMHg). Total Hg and MMHg, seawater characteristics such as water temperature and salinity, and meteorological parameters were also observed to estimate the air–sea exchange of Hg. The mean concentration of DGM was 116 ± 76 pg L−1 (N = 75), ranging from 19 to 442 pg L−1, and the concentrations were higher in summer than in other seasons. DGM concentration showed a significantly positive correlation with solar radiation, and air and water temperatures. Inversely, DGM showed a significantly negative correlation with salinity and redox potential (ORP). Hg evasion fluxes from the sea surface of the bay were calculated using a two-layer gas exchange model and ranged between 0.11 and 33 ng m−2 h−1 (mean, 5.4 ± 6.3 ng m−2 h−1). The estimated flux was slightly higher in the spring and fall when wind speed increased because the gas exchange coefficient used for estimating Hg evasion flux strongly depends on wind speed. The annual evasion flux of Hg from the sea surface of Minamata Bay was estimated to be 47 ± 56 μg m−2, which was on the same order of magnitude as the direct atmospheric deposition flux of Hg (24 μg m−2) during the observation period. Therefore, Hg evasion from the sea surface likely plays an important role in the Hg cycle of Minamata Bay. © 2014 Elsevier B.V. All rights reserved.

1. Introduction In recent years, there has been growing interest in the emission sources, distribution and long-range transport of atmospheric mercury (Hg). Active discussions on the reduction of Hg emission from human activities have been undertaken by UNEP (United Nations for Environmental Programs) and the Minamata Convention most recently held in October, 2013 (UNEP, 2013). To assess the risks of long-term exposure to human health, much more knowledge of emission sources, transport and deposition of atmospheric Hg is necessary. Although atmospheric Hg is emitted as a result of human activities such as the combustion of fossil fuels, waste incineration and gold mining, Hg is also emitted from natural sources such as volcanic activity and soil weathering (Nriagu, 1989). In addition, re-emission of Hg (both natural and anthropogenic) that had been previously deposited into soil and the oceans also plays an important factor in the global Hg cycle (Amos et al., 2013; UNEP, 2013). Amos et al. (2013) suggested that re-emission of legacy anthropogenic Hg from surface reservoirs accounts for 60% of present-day atmospheric deposition, compared to 27% from current human activities, which has been estimated at about 2000 ton year−1 (Pacyna et al., 2006; Streets et al, 2011). In order to determine the cost-effectiveness of measures to reduce the environmental risk of Hg ⁎ Corresponding author. Tel.: +81 966 63 3111; fax: +81 966 63 7822. E-mail address: [email protected] (K. Marumoto).

http://dx.doi.org/10.1016/j.marchem.2014.09.007 0304-4203/© 2014 Elsevier B.V. All rights reserved.

emissions due to human activities, it is important that Hg amounts released by legacy impacts and natural sources should be estimated as precisely as possible. In Japan, the annual Hg emission attributable to human activities such as coal-fired power plants and waste incinerators was estimated at 19 to 35 ton based on Hg emission data from each source (Kida and Sakai, 2005), but only limited data is available on Hg emissions from natural sources and the re-emission of deposited Hg (Nakagawa, 1999; Marumoto and Sakata, 2005; Narukawa et al., 2006). Minamata Bay is located in the southwestern part of the Kyushu Islands and was the site of a significant discharge of mono-methyl Hg (MMHg) in contaminated wastewater from the Chisso Corporation's Minamata factory into the bay from 1932 to 1965 (Kumamoto Prefecture, 1998). People living in the coastal areas around the bay and Yatsushiro Sea ate contaminated seafood harvested in these areas and suffered nervous disorders caused by MMHg poisoning. From 1977 to 1990, environmental pollution controls, including a reclamation project in which sediment containing more than 25 ppm (μg g−1) of total Hg was dredged from the bay and deposited along the coastline (Kumamoto Prefecture, 1998). However, the average concentration of total Hg in the surface sediment of the bay in 2010 remained high at 3.1 μg g−1 (Tomiyasu et al., 2014). This value is higher than the average concentration in the surface sediment in Tokyo Bay (0.43 ± 0.09 μg g−1), which receives inflow from the largest metropolitan area (Sakata et al., 2006). In addition, Kindaichi and Matsuyama (2005)

10

K. Marumoto, S. Imai / Marine Chemistry 168 (2015) 9–17

reported that the Hg concentration in some fish harvested from Minamata Bay exceeded 1 ppm, which is more than two times higher than the 0.4 ppm set by Japanese provisional regulation for total Hg in fish muscle (Ministry of Health and Welfare, Japan, 1973). The main source of MMHg in seawater remains unclear and comprehensive research including estimation of the input and output of Hg in the bay is needed to understand the sources and chemical reactions of Hg species. To investigate Hg evasion from the sea surface of Minamata Bay, which one of the removal processes of Hg from the bay, we measured volatile dissolved gaseous Hg (DGM) in surface seawater twelve times over a period of 1 year. In addition, total Hg and MMHg in the surface seawater were also measured. Moreover, atmospheric Hg concentration at the sea surface, meteorological parameters and water conditions were also monitored and used to calculate the Hg evasion flux using a gas exchange model across the air–sea interface (Liss and Slater, 1974). 2. Methods 2.1. Sampling The sampling sites in Minamata Bay are shown in Fig. 1. Seawater sampling for the measurement of DGM and other parameters was

carried out three times in each season from July 2012 to May 2013: summer (June–August), fall (September–November), winter (December–February) and spring (March–May). Observations on 18 October 2012 were conducted only at the coastal sites from St. 4 to St. 6 due to an approaching typhoon. Seawater was collected in a sampler that prevents volatilization loss of DGM from the sample (Marumoto et al, 2012). A schematic of the DGM sampler is presented in Fig. 2. Briefly, the DGM sampler has a Teflon ball in the bottom that is pushed up when the sampler takes in water and falls down, closing the opening, when the sampler is pulled from water. The sampler can be used to collect surface seawater without disturbing the water column and any decanting processes. The mean volatilization loss of DGM due to decanting processes is − 15 ± 10% (Marumoto et al, 2012). Separate from DGM sampling, 500 ml of surface seawater was collected in a Teflon bottle for analysis of total Hg and MMHg. Solar radiation above and in the sea was recorded using a pyrheliometer (Prede Co., Ltd.; PCM-01W) and water temperature, water density, salinity, conductivity, pH and oxidation–reduction potential (ORP) in the surface seawater were observed using a multi-parameter water quality measurement system with six sensors (Horiba, Co., Ltd.; U-22XD). Wind speed was recorded continuously using an ultrasonic anemometer and a vane-type anemometer at the rooftop of the Minamata Disease Archives (at a height of 23 m above sea level),

Sea of Japan

Pacific Ocean Minamata Disease Archives (Weather Station)

Chemical factory

ST-3 Yatsushiro Sea

(17 m)

ST-2 ST-1

ST-6 (10 m)

(17 m)

ST-5

(22 m)

(7 m)

Minamata Bay ST-4

Fukuro Bay

(8 m)

Fig. 1. Location of Minamata Bay and sampling sites in the bay. The water depths at each sampling site at high tide are indicated in parentheses. The tidal range of the bay is about 2 m. Meteorological parameters were continuously measured on the roof top of the Minamata Disease Archives.

K. Marumoto, S. Imai / Marine Chemistry 168 (2015) 9–17

Gold trap for DGM Soda lime column for removing water vapor

N2 purge gas

Gold trap for removing mercury from the purge gas

The purge gas including DGM

Teflon (FEP) tube

11

sample was purged with N2 gas at a flow rate of 0.75 L min−1 for 30 min and DGM is directly collected on a gold amalgamation trap without the need for a decanting process. The trap loaded with the DGM from the seawater sample was stored in a glass test tube and sealed with a butyl rubber stopper until analysis by thermal desorption-cold vapor fluorescence spectrometry (CVAFS, Nippon Instruments Corporation; RA-FG+). A blank was also prepared by conditioning a trap through the N2 purge of 800 ml of ultrapure water. The blank equivalent concentration was 4.8 ± 0.9 pg L−1 (N = 27) and the method detection limit (MDL), defined as three times the standard deviation of the blanks, was 2.7 pg L− 1 in this study. In addition, duplicate analysis was conducted five times, and the differences were within 10% in this study. Atmospheric total gaseous Hg (TGM) was collected onto a gold amalgamation trap at a flow rate of 0.5 L min−1 for about 20 min and assayed. The MDL for TGM was 0.03 ng m−3 (N = 3) based on an air sample volume of 10 L. This MDL is more than 50 times lower than the TGM concentration in ambient air. 2.3. Estimation of mercury evasion flux Hg evasion to the atmosphere occurs from surface seawater which is supersaturated with DGM. Although DGM consists of Hg(0) and dimethyl Hg, it is well known that almost all DGM in the surface seawater is elemental form (Mason et al., 1995; Mason, 2013; Horvat et al., 2003). Assuming that all of DGM is elemental Hg (Hg(0)), saturation, S (%) is an indicator for Hg emission and can be obtained from the following equation,

85 cm

S¼ Glass frit

Ball stopper Teflon (PTFE) ball

Glass valve cock

0

H  Cw  100; C air

ð1Þ

where Cw (pg L−1 = ng m− 3) is the DGM concentration and Cair (ng m−3) is the TGM concentration in air. H′ is a Henry's law coefficient (dimensionless number) calculated using the equations provided by Andersson et al. (2008a). In the case that Hg saturation (S) exceeds 100%, surface seawater is supersaturated with Hg(0) and Hg evasion will occur. Conversely, if it is less than 100%, Hg(0) is deposited from the atmosphere to the seawater (Andersson et al., 2008b). The Hg evasion (or deposition) flux is calculated from data on DGM and other factors using the gas exchange model proposed by Liss and Slater (1974), as shown in the following,
 C F ¼ K w C w − air0 ; H

4.0 cm i. d. This attachment is installed after seawater sampling Fig. 2. Schematic of a seawater sampler with features to prevent volatilization loss of Hg in water samples for DGM determination.

which is located along the coastal area of the bay (Fig. 1). Other meteorological parameters such as air temperature, relative humidity, air pressure and precipitation depth were also recorded continuously at the same site.

2.2. DGM and TGM measurements DGM was collected on the gold amalgamation trap upon sampling as follows. After sampling, the volume of the seawater sample was adjusted to 800 ml using the glass cock attachment installed in the bottom of the sampler (see Fig. 2). A gold amalgamation trap fitted with a soda lime column (inner diameter, 6.0 mm × length, 100 mm) for removing water vapor was attached to the lid of the sampler, and the seawater

ð2Þ

where Kw (cm h−1) is the gas exchange velocity, which primarily depends on wind speed on sea surface and Schmidt number representing material transfer in a fluid (Wanninkhof, 1992; Nightingale et al., 2000; Loux, 2004) though it is well known that Kw is affected by the stability of the boundary layer, bubble production and reactivity of gas (Merlivat and Memery, 1983; Jähne et al., 1987). Wind speed is a parameter that is easy to measure and it captures much of the variability in turbulence at the air–sea exchange interface. Eq. (3) developed by Nightingale et al. (2000) was adopted in this study. Kw (cm h−1), the gas exchange velocity, was determined as   2 K w ¼ 0:222  u10 þ 0:333  u10 

ScHg ScCO2

!−0:5 ;

ð3Þ

where ScHg and ScCO2 are the Schmidt number of Hg and carbon dioxide, respectively. The Schmidt number is determined by dividing the kinematic viscosity of seawater (cm2 s−1) by the diffusion coefficient (cm2 s− 1) of gas in water. ScHg and ScCO2 were calculated based on

K. Marumoto, S. Imai / Marine Chemistry 168 (2015) 9–17

u10 ¼

ð10:4uz Þ ; ð1nðzÞ þ 8:1Þ

ð4Þ

(a) (pg L-1)

gas exchange parameterization provided by Soerensen et al. (2010). The u10 in Eq. (3) is the wind speed at a standardized height of 10 m and Eq. (4) was used to transform the measured wind speed at 25 m to u10 (Schwarzenbach et al., 1993), as

DGM concentration

12

300 200 100 0

3. Results and discussion 3.1. Concentrations of DGM, MMHg and total Hg Fig. 3 shows the spatial distribution of DGM, MMHg, total Hg and the ratio of DGM for total Hg (%DGM) in the surface seawater of Minamata Bay. Concentrations at the coastal sites (from St. 4 to St. 6) were about two times higher than those at the more open sea sites (from St. 1 to St. 3). Statistical analysis revealed significant differences in concentrations between the coastal sites and the more open sea sites (t-test, P b 0.05 for DGM and P b 0.01 for MMHg and total Hg), indicating that these Hg species may be sourced from land. However, there was no difference in %DGM between the coastal sites and the more open sea sites. The mean concentrations of these Hg species in each season are shown in Table 1. Although the concentrations of MMHg and total Hg at the more open sea sites and the coastal sites had no significant seasonal variations, DGM showed seasonal variation with a higher concentration in summer than in other seasons (t-test, P b 0.01). The same seasonal

(pg L-1)

MMHg concentration

St.1

St.2

St.3

St.4

St.5

St.6

St.1

St.2

St.3

St.4

St.5

St.6

St.1

St.2

St.3

St.4

St.5

St.6

St.1

St.2

St.3

St.4

St.5

St.6

90 60 30 0

(pg L-1)

(c) Total Hg concentration

Total Hg in unfiltered seawater was measured using EPA Method 1631 (U.S. EPA, 2002). After sample collection, a 100 ml aliquot was preserved by adding 5 ml of concentrated HCl and 1 ml of BrCl (0.2 mol L−1), and the mixture was stored in a refrigerator until analysis. Prior to Hg analysis, 0.2 ml of 10% (w/v) NH2OH⋅HCl was added to the sample and the mixture was allowed to stand for 5 min to avoid the negative effects of free halogens on the analysis. Total Hg concentration in the sample was determined using a cold vapor atomic absorption spectrometer with gold amalgamation (Nippon Instruments Corporation; MA2000) following Hg(0) generation using 2 ml of 20% (w/v) SnCl2 as the reducing agent. The MDL for the total Hg was 210 pg L−1 when the blank solutions (N = 10) using ultrapure water were measured and the total Hg concentration in samples was generally more than 10 times higher than the MDL. The analytical procedure for MMHg determination on unfiltered seawater samples uses solvent extraction with dithizone-toluene (Dz) solution (Ministry of the Environment, Japan, 2004). MMHg and other Hg species in a 400 ml aliquot were concentrated into 3 ml of Dz solution after the removal of the matrix, and the pH was adjusted by sequentially adding and mixing 2 ml of 20 mol L−1 H2SO4, 1 ml of 0.5%(w/v) KMnO4, 4 ml of 10 mol L−1 NaOH, 0.8 ml of 10%(w/v) NH2OH·HCl and 0.8 ml of 10%(w/v) EDTA. Prior to analysis, MMHg in the Dz solution was extracted into 0.8 ml of 5 μg ml−1 Na2S solution through several steps. The MMHg concentration in the Na2S solution was determined using a cold vapor atomic fluorescence spectrometer coupled to a gas chromatograph with thermal desorption (GC-CVAFS; Tekran Inc., model 2500), following derivatization using sodium tetra ethylene borate, NaB(C2H5)4, and the adsorption of the generated CH3HgC2H5 by a Tenax trap (Logar et al., 2002). The MDL for MMHg was 4.5 pg L−1 when the blank solutions (N = 8) using ultrapure water were measured. In addition, the recovery of MMHg was 97.6 ± 4.7% (N = 4) based on the recovery of a spike of a known concentration of MMHg obtained from alkaline dissolution for Dorm-2, which is a standard material for MMHg in dogfish.

(b)

6000 4000 2000 0

(d) total Hg (%)

2.4. Analysis of total Hg and MMHg in seawater

Percentage of DGM to

where uz and u10 are wind speed at the heights of z and 10 m, respectively, and z is the height at which wind speed was measured.

30.0 20.0 10.0 0.0

Fig. 3. Spatial distribution of the concentrations (mean ± SD) of (a) DGM, (b) MMHg, (c) Total Hg and (d) the percentages of DGM to total Hg (%DGM) in Minamata Bay.

trends of DGM concentrations were observed in the South China Sea (Tseng et al., 2013) although the maximum in summer (200 fM = 40 pg L− 1) was one-tenth lower compared with the value (442 pg L−1) observed at St. 4 in Minamata Bay on 22 July 2012. In addition, %DGM at the more open sea sites was also higher in summer than in other seasons (P b 0.01, t-test). On the other hand, %DGM in summer at the coastal sites was also higher than in fall and was comparable to those in winter and spring. As shown in Fig. 4, DGM, MMHg and total Hg all showed a significantly positive correlation with each other (r = 0.49–0.54, P b 0.01) at the coastal sites with the highest MMHg observation (143 pg L−1 at St. 5 on 23 April 2013) excluded as an outlier. These correlations indicate that these Hg species have similar sources and/or that total Hg discharged into the bay induces the production of DGM and MMHg. On the other hand, at the more open sea sites, there were no correlations. 3.2. Relationships between Hg species and water and meteorological parameters Table 2 shows the correlation coefficients obtained for the relationships between each Hg species and water and meteorological parameters. DGM concentration and %DGM were positively correlated with air and water temperatures and were negatively correlated with salinity

K. Marumoto, S. Imai / Marine Chemistry 168 (2015) 9–17

13

Table 1 Dissolved gaseous Hg (DGM), mono-methyl Hg (MMHg), total Hg and the percentage of DGM for total Hg (%DGM) in the seawater of Minamata Bay. Seasons

More open sea sites (St.1–St.3) N

Summer Fall Winter Spring All data

(July–August) (September–November) (December–February) (March–May)

9 9 9 9 36

Coastal sites (St.4–St.6)

DGM

MMHg

Total Hg

(pg L−1)

(pg L−1)

(pg L−1)

172 82 54 68 94

9 11 8 12 10

1100 1670 800 1100 1130

± ± ± ± ±

36 30 9 29 54

± ± ± ± ±

4 3 3 6 4

± ± ± ± ±

%DGM to total

240 850 290 240 590

and ORP (P b 0.001) at the more open sea sites. Higher seawater temperatures and lower ORP are expected to increase Hg(0) production through the reduction of Hg(II). Salinity can also affect Hg complexation in seawater in which Hg(II) is likely to exist as a mercuric chloride in the 2− (Berlin et al., 2007). However, forms of HgCl2, HgCl− 3 and HgCl4

DGM vs MMHg

(a) 500

y = 4.78x + 34.3 r = 0.54 P< 0.001

DGM (pg L-1)

400 300 200 100 0

0

10

20

30

40

50

MMHg (pg/L)

(b)

DGM vs Total Hg 500

y = 0.022x + 76.9 r= 0.50, P< 0.01

DGM (pg L-1)

400 300 200 100 0

0

5000

10000

Total Hg (pg L-1)

(c)

MMHg vs Total Hg 50.0

MMHg (pg L-1)

40.0 30.0 20.0 10.0 0.0

y = 0.0025x + 14.6 r= 0.49, P< 0.01 0

5000

Total Hg (pg

10000

L-1)

Fig. 4. Correlation between Hg species concentrations at coastal sites from St. 4 to St. 6: (a) DGM versus MMHg, (b) DGM versus Total Hg, and (c) MMHg versus Total Hg.

16.1 5.6 7.6 7.5 9.2

± ± ± ± ±

3.4 2.8 3.0 2.6 5.0

N

9 12 9 9 39

DGM

MMHg

Total Hg

(pg L−1)

(pg L−1)

(pg L−1)

263 92 101 105 137

26 21 19 32 24

3660 3350 1460 2030 2680

± ± ± ± ±

95 17 40 46 88

± ± ± ± ±

14 7 10 42 22

± ± ± ± ±

%DGM to total

2880 1250 700 1880 1970

10.1 3.5 8.4 8.2 7.2

± ± ± ± ±

4.9 2.4 4.8 4.5 4.8

Rolfhus and Fitzgerald (2001) suggested that higher salinity and lower dissolved organic carbon (DOC) might lead to reduce abundance of Hg organic complexes and to accelerate the increase of labile Hg(II), resulting in enhanced Hg(0) production. Soerensen et al. (2013) also reported that DGM concentrations were both negatively and positively correlated with salinity in the West Atlantic Ocean. A potential explained by the observed correlation between salinity and DGM may be freshwater transport because sampling at the more open sea sites was basically carried out at falling tide (Table S1). Fig. 5 shows the relationship between DGM concentration and salinity between the more open sea sites and the coastal sites. On all sampling days, DGM concentration and salinity at the more open sea sites increased with those at the coastal sites. This indicates that freshwater transport from the coastal area greatly contributes to the variability observed in DGM concentration among the more open sea sites. Correlations between total Hg and MMHg for the more open sea sites and the coastal sites were not observed. Further, they found no correlations among the seawater parameters at the more open sea sites, indicating that other factors such as seawater exchange between the bay and the Yatsushiro Sea (Rajar, et al., 2004), elution from sediments (Tomiyasu et al., 2008) and primary production (Horvat et al., 2003; Heimbürger et al., 2010) contributed to the observed variation. On the other hand, the concentrations of total Hg and MMHg at the coastal sites showed significant negative correlations with salinity (P b 0.01), which is reduced by the inflow of freshwater through precipitation, river water, runoff water and groundwater. Treated wastewater from the chemical factory located in the center of Minamata City flows into Minamata Bay near St. 6 (see Fig. 1). Freshwater discharge into the bay increases with increased precipitation because only a small river feeds into Fukuro Bay. Fig. 6 shows the relationship between the concentrations of total Hg and MMHg in the surface seawater at the coastal sites and integrated precipitation depths during the 24, 72 and 120 h before the sampling day. These Hg parameters were significantly correlated with the 72 h integrated precipitation depth (r = 0.72, P b 0.01 for total Hg; r = 0.55, P b 0.05 for MMHg), suggesting that total Hg and MMHg became higher in Minamata Bay due to increased Hg supply from the land through runoff water and groundwater because there were no correlations with these Hg parameters at the more open sea sites and the influence of freshwater discharge on total Hg and MMHg concentrations in surface seawater showed a time lag. Notably, mean total Hg was highest at St. 4, which is closest to the submarine groundwater discharge in the Yudo District (Tomiyasu et al., 2006). As total Hg in this study was measured on unfiltered water, which includes suspended solids derived from sediments, it is inferred that Hg in the groundwater and/or Hg in the sediments churned up by groundwater discharges are the main sources of total Hg at St. 4. The submarine groundwater discharge is a significant source of Hg in seawater in other areas such as Elkhorn Slough on the central California coast, USA (Black et al., 2009), Bangdu Bay and Hwasun Bay in Jeju Island, Korea (Lee et al., 2011), but it is not significant in Puck Bay, the Southern Baltic Sea (Szymczycha et al., 2013). As shown in Table 2, %DGM at the coastal sites showed no correlation with air and water temperatures, salinity and ORP, although DGM concentration had significant correlations with them. This may be

14

K. Marumoto, S. Imai / Marine Chemistry 168 (2015) 9–17

Table 2 Correlation coefficients between Hg species and water and meteorological parameters. More open sea sites (St.1–St.3)

Coastal sites (St.4–St.6)

DGM

MMHg

Total Hg

%DGM

DGM

MMHg

Total Hg

%DGM

N

36

36

36

36

39

38⁎⁎⁎

39

39

Solar radiation Solar radiation in surface water Air temperature Wind speed pH DO Water temperature Salinity ORP

0.39 0.07 0.82⁎⁎ −0.34 −0.23 −0.23 0.84⁎⁎ −0.61⁎⁎ −0.82⁎⁎

0.21 0.09 0.10 0.16 0.07 −0.08 0.05 0.06 −0.08

0.11 0.19 0.18 0.11 0.09 −0.39 0.37 0.07 −0.21

0.42⁎ 0.38 0.33 −0.06 −0.22 −0.09 0.26 −0.54⁎⁎ −0.21

0.20 0.16 0.32 0.19 −0.30 −0.23 0.39 −0.47⁎ −0.14

0.13 0.10 0.13 −0.35 0.04 0.25 0.04 −0.10 −0.36

0.26 0.04 0.59⁎⁎

0.57⁎⁎ 0.39 0.67⁎⁎

0.34 −0.32 0.07 0.53⁎⁎ −0.58⁎⁎ −0.56⁎⁎

−0.25 −0.49⁎ −0.06 0.51⁎⁎ −0.81⁎⁎ −0.64⁎⁎

⁎ P b 0.01. ⁎⁎ P b 0.001. ⁎⁎⁎ The maximum of MMHg (143 pg L−1) was excluded as an outlier.

Salinity at more open sea sites (psu)

(b)

500

y = 0.51x + 21 r= 0.81, P< 0.01

400 300 200

(a) 24 h before

6000

24h before r= 0.33, P= 0.25

120 h before

72 h before r= 0.72, P< 0.01

5000 4000

120 h before r= 0.53, P= 0.055

3000 2000 1000 0

100 0

72 h before

0.0 50.0 100.0 150.0 Integrated precipitation depth (mm)

(b) 0

100

200

300

400

DGM concentration at the coastal sites (pg L-1) 40 35 30 25 20

y = 0.74x + 8.8 r= 0.965, P< 0.001

15 15

20

25

30

35

24 h before

500

40

MMHg concentrations at the coastal sites (pg L-1)

DGM concentration at more open sea sites (pg L-1)

(a)

in offshore seawater, which had lower DOC, compared to areas that are more heavily influenced by freshwater. This is consistent with our finding that %DGM was significantly higher at the more open sea sites than at the coastal sites in summer. Therefore, further research is needed to elucidate the interaction between DGM production and DOC.

Total Hg concentration at the coastal sites (pg L-1)

because the Hg supply from land greatly influences the variations in total Hg concentrations. The concentrations of DGM and MMHg were positively correlated with solar radiation above the sea surface at the coastal site. The correlation between DGM and solar radiation is consistent with the finding that Hg(0) production from divalent Hg is mediated by photochemical reactions (Amyot et al, 1997; Rolfhus and Fitzgerald, 2004; Qureshi et al., 2010). In addition, DOC facilitates the photochemical production of Hg(0) (Costa and Liss, 1999, 2000) and, in general, is more abundant in freshwater and coastal water than in open sea (Bauer and Bianchi, 2011; Ogawa and Tanoue, 2003). Although we did not measure DOC in this study, it is expected that the DOC concentration would be higher at the coastal sites, which are more influenced by freshwater than are the more open sea sites. Soerensen et al. (2013) reported that the proportion of DGM for filtered Hg was higher

72 h before

120 h before

60 50

40 30 20

24 hours r= 0.17, P= 0.56

72 hours r= 0.55, P< 0.05

120 hours r= 0.39, P= 0.17

10 0 0.0 50.0 100.0 150.0 Integrated precipitation depth (mm)

Salinity at the coastal sites (psu) Fig. 5. Relationships of (a) DGM concentration and (b) salinity between the more open sea sites and the coastal sites. Error bars show standard deviation.

Fig. 6. Relationships between the concentration of (a) total Hg and (b) MMHg in the seawater at the coastal sites and integrated precipitation depth during the 24, 72 and, 120 h before the sampling day.

K. Marumoto, S. Imai / Marine Chemistry 168 (2015) 9–17

In contrast, the positive correlation between MMHg and solar radiation observed in this study contradicts previous suggestions that MMHg is photochemically broken down and converts into inorganic Hg (Mason et al., 2012; Lehnherr et al., 2011). Some researchers have reported that MMHg is produced by photochemical reactions in laboratory experiments (Akagi and Sakagami, 1972; Akagi and Takabatake, 1973; Yin et al., 2012). However, we have no additional data to show any evidence for the photochemical production of MMHg in surface seawater. Thus, the mechanism producing this pattern at the coastal sites of Minamata Bay remains unclear.

Baltic Sea and the Yellow Sea were lower than those in Minamata Bay. Negative flux (deposition flux) of Hg(0) was observed in winter in the Yellow Sea (Ci et al., 2010) as well as in the South China Sea (Tseng et al., 2013). Tseng et al. (2013) suggested that the South China Sea acts as a sink of atmospheric Hg(0) in winter due to low surface temperature (around 23 °C), enhanced vertical mixing and higher atmospheric Hg(0) concentration transported from the Asian continent. In Minamata Bay, the TGM concentration was slightly higher in winter than in summer. However, saturation was more than 100% even in winter, indicating that the bay is an Hg emission source throughout the year. Even though the uncertainty of estimation is large due to the spatial distribution of DGM concentration and Hg evasion flux in Minamata Bay, the annual Hg evasion flux could be roughly estimated at 47 ± 56 μg m− 2 year−1 using the mean hourly evasion flux based on all data over the observation period. Thus, the total amount of Hg released from the surface of Minamata Bay (surface area, 3.82 km2; Kumamoto Prefecture, 1998) was calculated at 0.18 ± 0.21 kg year−1.

3.3. Hg evasion flux Time series of Hg evasion fluxes at the more open sea sites and the coastal sites are shown in Fig. 7. The highest flux values were observed on 18 October 2012 when the wind speed was about 10 m s−1 due to the approach of typhoon No. 21; DGM concentration on that day was almost the same for the other days in the same season. A summary of DGM, %DGM, TGM in air, wind speed at the height of 10 m above sea level (u10), saturation and Hg flux over the entire sampling period in Minamata Bay and in other sea areas is presented in Table 3. DGM concentrations in Minamata Bay were relatively higher than in other sea areas, except for San Francisco Bay (Conaway et al., 2003). The mean value (± SD) of S, 1830 ± 1650% at the bay was also higher than those in other sea areas such as the South China Sea, the North Sea and the Mediterranean Sea (Andersson et al., 2007, 2011; Rolfhus and Fitzgerald, 2001; Wangberg et al., 2001; Ci et al., 2010). Hg evasion fluxes ranged from 0.1 to 33 ng m−2 h−1 (mean ± SD, 5.4 ± 6.3 ng m− 2 h−1). The mean flux at the coastal sites (6.2 ± 4.8 ng m− 2 h− 1) was twice that of the more open sea sites (2.6 ± 2.4 ng m−2 h− 1), except for observations conducted during the approaching typhoon. The fluxes were the highest in fall, although DGM concentrations and saturation were higher in summer. In Minamata Bay, the mean u10 showed seasonal variation with stronger winds in the fall, spring and winter compared to in the summer. Following Eqs. (2) and (3), Hg evasion flux was calculated using the two-layer gas exchange model and was largely influenced by wind speed. The Hg evasion fluxes in Minamata Bay were comparable with those in the Mediterranean Sea (Gardfeldt et al., 2003; Andersson et al., 2007), Long Island Sound (Rolfhus and Fitzgerald, 2001), San Francisco Bay (Conaway et al., 2003) and Tokyo Bay (Narukawa et al., 2006), but they were higher than those in the North Atlantic Ocean (Andersson et al., 2011), the Baltic Sea (Wangberg et al., 2001) and the Yellow Sea (Ci et al., 2010). DGM concentrations in the North Atlantic Ocean, the

3.4. Hg exchange across the air–sea interface Wet Hg deposition fluxes have also been observed in Minamata Bay area since September 2008 (Marumoto and Matsuyama, 2014). The mean volume weight concentration of total Hg in wet deposition collected during the period from June 2012 to May 2013 was 7.0 ng L−1, which is about three times higher than the average concentration of Hg in seawater. The annual wet Hg deposition flux was 15.4 μg m−2 year−1. Atmospheric Hg is also deposited by dry deposition processes such as diffusion of gaseous Hg and settling of particulate Hg. However, dry deposition fluxes of Hg were not observed in this area. From the findings on wet and dry deposition fluxes of Hg at 8 non-urban sites in Japan (Sakata and Marumoto, 2005), the annual dry deposition fluxes of Hg were significantly correlated with the annual wet deposition fluxes (N = 8, r = 0.84, P b 0.01). The relationship is described by the following regression line: “Dry Hg” ¼ 0:387  “Wet Hg” þ 2:59: Assuming that this relationship can be applied to the Minamata Bay area, the annual dry Hg deposition flux in the coastal area was estimated to be 8.5 μg m− 2 year− 1, and the total atmospheric deposition flux (wet + dry) was estimated to be about 24 μg m−2 year−1, corresponding to 0.092 kg year−1 of the total amount of Hg deposited from atmosphere to the surface of Minamata Bay. The evasion flux and deposition flux were the same order of magnitude in air–sea interface of Minamata Bay. Table 4 shows evasion and atmospheric deposition fluxes of Hg in

More open sea sites

Coastal sites

18 October 2012*

40

Hg Flux (ng m-2 h-1)

15

30 20 10 0 6/1/2012

8/1/2012

10/1/2012

12/1/2012

1/31/2013

4/2/2013

6/2/2013

Date Fig. 7. Time series of Hg evasion fluxes at more open sea sites and coastal sites in Minamata Bay. * Sampling could not be conducted at the open sea sites on 18 October 2012 due to the strong winds from an approaching typhoon.

16

K. Marumoto, S. Imai / Marine Chemistry 168 (2015) 9–17

Table 3 Survey of DGM concentration, saturation, mercury flux and wind speed reported in the literature and in this study. Area

Sampling period

N

DGM (pg L−1)

TGM in air (ng m−3)

Saturation (%)

Mercury flux (ng m−2 h−1)

Wind speed u10 (m/s)

Water temperature (°C)

Literature

Mediterranean Sea Mediterranean Sea

Jul.–Aug. 2000 Summer, 2003 Spring, 2004 Autumn, 2004 Jul. 2005 Aug., 2008 Sep., 2008 Jun., 2009 Sep., 2009 Oct., 2009 Aug., 2010 Jul. 1997 Mar. 1998 Oct. 2008 Jan. 2009 Apr.–May 2009 Aug. 2009 Feb. 1996 May–97 Aug. 1995 Oct. 1996 Oct. 1997 Apr. and Jul. 1999 Oct. 2004–Jan. 2005 All Summer Fall Winter Spring

23 C⁎ C⁎ C⁎ C⁎

28 ± 17 38 24 38 12 ± 2 32.0 ± 12.0 25.6 ± 4.0 24.0 ± 1.4 22.2 ± 2.8 18.0 ± 4.6 39.2 ± 6.8 17.6 17.4 27.0 ± 16.4 16.0 ± 6.0 23.0 ± 8.7 69.0 ± 23.3 47 ± 7 30 ± 8 85 ± 22 66 ± 20 20 ± 8 180 ± 180 53 ± 26 116 ± 76 218 ± 84 88 ± 23 78 ± 37 86 ± 42

1.8 ± 0.4 1.8 1.7 1.7 1.7 ± 0.1 1.4 ± 0.2 1.5 ± 0.1 1.4 ± 0.1 NA⁎⁎ 1.4 ± 0.2 NA⁎⁎ 1.7 1.4 2.36 ± 0.66 3.06 ± 0.95 2.83 ± 0.94 1.98 ± 0.97 2.8 3.0 3.3 3.0 3.0 2.0 1.9 ± 0.6 2.05 ± 0.36 1.77 ± 0.19 2.15 ± 0.21 2.23 ± 0.35 2.05 ± 0.49

590 ± 380 790 321 728 150 ± 30 628 ± 232 617 ± 122 524 ± 59 527 ± 71 368 ± 129 934 ± 159 250 130 350 ± 260 96 ± 39 200 ± 98 1100 ± 592 283 ± 66 200 ± 54 898 ± 244 664 ± 198 219 ± 101

6.8 ± 8.4 5.2 0.8 3.6 0.4 ± 0.3 4.3 ± 3.4 3.0 ± 2.9 4.7 ± 3.7 2.1 ± 0.7 2.2 ± 1.7 6.8 ± 5.1 1.6 0.8 0.89 ± 1.84 −0.06 ± 0.64 0.32 ± 0.71 0.88 ± 1.38 1.8 2.8 4.3 4.4 0.6 2.5–46 5.6 ± 5.0 5.0 ± 5.8 4.1 ± 2.6 9.6 ± 8.8 1.7 ± 1.6 5.3 ± 6.1

7.4 ± 3.0 5.1 3.6 5.1

24.7 ± 1.0

Gardfeldt et al., 2003 Andersson et al., 2007

11 20.2 ± 3.1 27.7 ± 0.3 24.4 ± 0.2 28.1 ± 0.5 19.6 ± 4.3 28.3 ± 0.4 17.4 3.3 18.8 ± 0.46 2.4 ± 0.4 8.0 ± 0.5 19.0 ± 0.9

Andersson et al., 2011 Soerensen et al., 2013

North Atlantic Ocean West Atlantic Ocean

Baltic Sea Yellow Sea

Long Island Sound in U.S.A.

San Francisco Bay Tokyo Bay Minamata Bay

11 9

8 10 6 4 14 19 22 75 18 21 18 18

800 ± 550 1800 ± 1600 4150 ± 1770 1230 ± 330 830 ± 440 1200 ± 670

6.3 5.1 7.4 5.2 6.6 6.2 6.2 8.5 3.8 2.8 3.2 1.5 4.5 6.4 3.1 4.5 3.6

± ± ± ± ± ±

5.5 3.0 1.6 4.8 2.2 3.2

± ± ± ± ± ±

± ± ± ±

3.2 2.3 3.3 0.9 2.3 2.4

1.9 2.1 2.0 1.0

Wangberg et al., 2001 Ci et al., 2010

Rolfhus and Fitzgerald, 2001

2.6 2.0 0.7 2.3 1.2 1.7

15.3 20.1 26.7 22.6 13.8 16.8

± ± ± ± ± ±

3.6 5.3 1.6 2.2 1.5 2.3

Conaway et al., 2003 Narukawa et al., 2006 This study

⁎ Continuous monitoring. ⁎⁎ NA = not available.

other sea areas. In other areas, the Hg evasion fluxes were also comparable to atmospheric Hg deposition. Hg evasion is an important process in the Hg cycle of Minamata Bay. However, Yano (2013) suggested that annual transport of particulate total Hg from Minamata Bay to the Yatsushiro Sea, which is connected to the bay, was estimated at 6.0 kg. This process seems to be more important for Hg removal in sediments and seawater in Minamata Bay. It is possible that the contribution of Hg exchange at the sea surface as a removal process of Hg from Minamata Bay is small compared to the outflow of sediments containing Hg. Further research, including more accurate estimation of Hg evasion flux that takes into account Hg inflow through runoff water and ground water, is required in order to conduct a detailed mass balance study. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.marchem.2014.09.007. Acknowledgments We wish to thank Dr. Shinichiro Yano and Dr. Akira Tai of Kyushu University, Dr. Akihide Tada of Nagasaki University, Dr. Akito Matsuyama of National Institute for Minamata Disease (NIMD) and the fisheries cooperative association of Minamata for their cooperation in the Hg survey in Minamata Bay. We are also grateful to A. Kubo for her assistance

with sampling and Hg analysis. Finally, we would like to thank the two reviewers for their constructive and helpful comments which greatly improved our manuscript. References Akagi, H., Sakagami, Y., 1972. Studies on photochemical alkylation of inorganic mercury. –Alkylation of inorganic mercury in water by irradiation with sunlight or black light. J. Hyg. Chem. 18 (6), 358–362 (in Japanese). Akagi, H., Takabatake, E., 1973. Photochemical formation of methylmercuric compounds from mercuric acetate. Chemosphere 2, 131–133. Amos, H.M., Jacob, D.J., Streets, D.G., Sunderland, E.M., 2013. Legacy impacts of all-time anthropogenic emissions on the global mercury cycle. Glob. Biogeochem. Cycles 27 (410–421), 2013. http://dx.doi.org/10.1002/gbc.20040. Amyot, M., Gill, G.A., Morels, F.M.M., 1997. Production and loss of dissolved gaseous mercury in coastal seawater. Environ. Sci. Technol. 31, 3606–3611. Andersson, M.E., Gardfeldt, K., Wangberg, I., Sprovieri, F., Pirrone, N., Lindqvist, O., 2007. Seasonal and daily variation of mercury on and off shore sites from the Mediterranean Sea. Mar. Chem. 107 (1), 104–116. Andersson, M.E., Gardfeldt, K., Wangberg, I., Stromberg, D., 2008a. Determination of Henry's law constant for elemental mercury. Chemosphere 73, 587–592. Andersson, M.E., Gardfeldt, K., Lindqvist, O., 2008b. Enhanced concentrations of dissolved gaseous mercury in the surface waters of the Arctic Ocean. Mar. Chem. 110, 190–194. Andersson, M.E., Sommar, J., Gardfeldt, K., 2011. Air–sea exchange of volatile mercury in the North Sea. Mar. Chem. 125, 1–7. Bauer, J.E., Bianchi, T.S., 2011. Dissolved organic carbon cycling and transformation. In: Wolanski, E., McLusky, D.S. (Eds.), Treatise on Estuaries and Coastal Science vol. 5. Academic Press, Waltham, pp. 7–67.

Table 4 Survey of Hg evasion from the sea and atmospheric deposition in Minamata Bay reported in the literature and in this study. Area

Evasion (μg m−2 year−1)

Deposition (μg m−2 year−1)

(Wet)

Pettaquamscutt Estuary, Northeastern U.S.A. Baltic Sea Bermuda Tokyo Bay, Japan Minamata Bay, Japan

11 9.5 30 51 47

22 10 9.6 39 24

12 7.2 19 15.4

(Dry) 9 2.4 20 8.5⁎

Evasion/deposition

Literature

0.50 0.95 3.1 1.3 2.0

Mason et al., 1993 Wangberg et al., 2001 Gichuki and Mason, 2014 Narukawa et al., 2006 This study

⁎ Rough estimate based on annual wet Hg deposition flux data at Minamata and the relationship between annual wet and dry Hg deposition observed at eight sites in Japan.

K. Marumoto, S. Imai / Marine Chemistry 168 (2015) 9–17 Berlin, M., Zalups, R.K., Flowler, B.A., 2007. Handbook on the Toxicology of Metals, In: Nordberg, G., Fowler, B.A., Nordberg, M., Friberg, L.T. (Eds.), 3rd edition Elsevier, California, p. 675. Black, F.J., Paytan, A., Knee, K.L., Sieyes, N.R., Ganguli, P.M., Gray, E., Flegal, R.A., 2009. Submarine groundwater discharge of total mercury and monomethylmercury to Central California coastal waters. Environ. Sci. Technol. 43, 5652–5659. Ci, Z., Zhang, X., Wang, Z., 2010. Elemental mercury in coastal seawater of Yellow Sea, China: temporal variation and air–sea exchange. Atmos. Environ. 45 (1), 183–190. Conaway, C.H., Squire, S., Mason, R.P., Flegal, A.R., 2003. Mercury speciation in the San Francisco Bay estuary. Mar. Chem. 80, 199–225. Costa, M., Liss, P.S., 1999. Photoreduction of mercury in sea water and its possible implications for Hg0 air–sea fluxes. Mar. Chem. 68, 87–95. Costa, M., Liss, P.S., 2000. Photoreduction and evolution of mercury from seawater. Sci. Total Environ. 261, 125–135. Gardfeldt, K., Sommar, J., Ferrara, R., Ceccarini, C., Lanzillotta, E., Munthe, J., Wangberg, I., Lindqvist, O., Pirrone, N., Sprovieri, F., Pesenti, E., Stromberg, D., 2003. Evasion of mercury from coastal and open waters of the Atlantic Ocean and the Mediterranean Sea. Atmos. Environ. 37 (Supplement No.1), S73–S84. Gichuki, S.W., Mason, R.P., 2014. Wet and dry deposition of mercury in Bermuda. Atmos. Environ. 87, 249–257. Heimbürger, L.-.E., Cossa, D., Marty, J.-.C., Migon, C., Averty, B., Dufour, A., Ras, J., 2010. Methyl mercury distributions in relation to the presence of nano- and picophytoplankton in an oceanic water column (Ligurian Sea, North-western Mediterranean). Geochim. Cosmochim. Acta 74, 5549–5559. Horvat, M., Kotnik, J., Logar, M., Fajon, V., Zvonaric, T., Pirrone, N., 2003. Speciation of mercury in surface and deep-sea waters in the Miditerranean Sea. Atmos. Environ. 37, 93–108. Jähne, B., Münnich, K.O., Bosinger, R., Dutzi, A., Huber, W., Libner, P., 1987. Journal of. Geophys. Res. 92 (C2), 1937–1949. Kida, A., Sakai, S., 2005. Preliminary estimation of mercury emission inventories for Japan's air. Waste Manag. Res. 16 (4), 191–203 (In Japanese, abstract in English). Kindaichi, M., Matsuyama, A., 2005. Variation of total mercury concentration in fish caught in Minamata Bay over 26 years. J. Jpn Soc. Water Environ. 28, 529–533 (In Japanese, abstract in English). Kumamoto Prefecture, 1998. An outline of operation on environmental restoration in Minamata Bay. Report of Kumamoto Prefecture, Japan (In Japanese). Lee, Y., Rahman, M.D.M., Kim, G., Han, S., 2011. Mass balance of total mercury and monomethylmercury in coastal embayment of a volcanic island: significance of submarine groundwater discharge. Environ. Sci. Technol. 45, 9891–9900. Lehnherr, I., St. Louis, V.L., Hintelmann, H., Kirk, L., 2011. Methylation of inorganic mercury in polar waters. Nat. Geosci. 4, 298–302. Liss, P.W., Slater, P.G., 1974. Flux of gases across the air–sea interface. Nature 247, 181–184. Logar, M., Horvat, M., Akagi, H., Pihlar, B., 2002. Simultaneous determination of inorganic mercury and methylmercury compounds in natural waters. Anal. Bioanal. Chem. 374, 1015–1021. Loux, N.T., 2004. A critical assessment of elemental mercury air/water exchange parameters. Chem. Speciat. Bioavailab. 16 (4), 127–138. Marumoto, K., Matsuyama, A., 2014. A. Mercury speciation in wet deposition samples collected from a coastal area of Minamata Bay. Atmos. Environ. 86, 220–227. Marumoto, K., Sakata, M., 2005. Mercury emission fluxes from soil surfaces and factors affecting their variations. Chikyukagaku (Geochem.) 39, 183–196 (In Japanese, abstract in English). Marumoto, K., Imai, S., Takeda, T., Sakugawa, H., 2012. Development of seawater sampler for collection of dissolved gaseous mercury preventing vaporization loss and influence of the loss on estimating mercury flux. Jpn Soc. Anal. Chem. 61, 1063–1072 (In Japanese, abstract in English). Mason, R.P., 2013. Trace Metals in Aqueous Systems. Wiley-Blackwell, UK, p. 203. Mason, R.P., Fitzgerald, W.F., Hurley, J., Hanson Jr., A.K., Donaghay, L., Sieburth, J.M., 1993. Mercury biogeochemical cycling in a stratified estuary. Limnol. Oceanogr. 38 (6), 1227–1241. Mason, R.P., Rolfhus, K.R., Fitzgerald, W.F., 1995. Methylated and elemental mercury cycling in surface and deep-ocean waters of the North Atlantic. Water Air Soil Pollut. 80, 665–667. Mason, R.P., Choi, A.L., Fitzgerald, W.F., Hammerschmidt, C.R., Lamborg, C.H., Soerensen, A.L., Sunderland, E.M., 2012. Mercury biogeochemical cycling in the ocean and policy implications. Environ. Res. 119, 101–117. Merlivat, L., Memery, L., 1983. Gas exchange across an air–water interface: experimental results and modeling of bubble contribution to transfer. J. Geophys. Res. 88 (C1), 707–724. Ministry of Health and Welfare, Japan, 1973. The Interim Regulatory Standard Value for Mercury in Fish and Shellfish. Notification No.99 by Director of environmental health in the Ministry of Health and Welfare, Japan (in Japanese).

17

Ministry of the Environment, Japan, 2004. Mercury Analysis Manual. The committee for the Mercury Analysis Manual. Nakagawa, R., 1999. Estimation of mercury emissions from geothermal activity in Japan. Chemosphere 38, 1867–1871. Narukawa, M., Sakata, M., Marumoto, K., Asakura, K., 2006. Air–sea exchange of mercury in Tokyo Bay. J. Oceanogr. 62, 249–257. Nightingale, P.D., Malin, G., Law, C.S., Watson, A.J., Liss, P.S., Liddicoat, M.I., Boutin, J., Upstill-Goddard, R.C., 2000. In situ evaluation of air–sea gas exchange parameterizations using novel conservative and volatile tracers. Glob. Biogeochem. Cycles 14 (1), 373–387. Nriagu, J.O., 1989. A global assessment of natural sources of atmospheric trace metals. Nature 338, 47–49. Ogawa, H., Tanoue, E., 2003. Dissolved organic matter in oceanic waters. J. Oceanogr. 59, 129–147. Pacyna, E.G., Pacyna, J.M., Steenhuisen, F., Wilson, S., 2006. Global anthropogenic mercury emission inventory for 2000. Atmos. Environ. 40, 4048–4063. Qureshi, A., O'Driscoll, N.J., Macleod, M., Meuhold, Y., Hungerbuhler, K., 2010. Photoreactions of mercury in surface ocean water: gross reaction kinetics and possible pathways. Environ. Sci. Technol. 44, 644–649. Rajar, R., Zegar, D., Cetina, M., Akagi, H., Yano, S., Tomiyasu, T., Horvat, M., 2004. Application of three-dimensional mercury cycling model to coastal seas. Ecol. Model. 171, 139–155. Rolfhus, K.R., Fitzgerald, W.F., 2001. The evasion and spatial/temporal distribution of mercury species in Long Island Sound, CT-NY. Geochim. Cosmochim. Acta 65 (3), 407–418. Rolfhus, K.R., Fitzgerald, W.F., 2004. Mechanisms and temporal variability of dissolved gaseous mercury production in coastal seawater. Mar. Chem. 90, 125–136. Sakata, M., Marumoto, K., 2005. Wet and dry deposition fluxes of mercury in Japan. Atmos. Environ. 39, 3139–3146. Sakata, M., Marumoto, K., Narukawa, M., Asakura, K., 2006. Mass balance and sources of mercury in Tokyo Bay. J. Oceanogr. 62, 767–775. Schwarzenbach, R.P., Gashwend, P.M., Imborden, D.M., 1993. Environmental Organic Chemistry. Wiley Inter-science, New York, p. 228. Soerensen, A.L., Sunderland, E.M., Holmes, C.D., Jacob, D.J., Yantosca, R.M., Skov, H., Christensen, J.H., Strode, S.A., Mason, R.P., 2010. An improved global model for air– sea exchange of mercury : high concentrations over the North Atlantic. Environ. Sci. Technol. 44, 8574–8580. Soerensen, A.L., Mason, R.P., Balcom, P.H., Sunderland, E.M., 2013. Drivers of surface ocean mercury concentrations and air–sea exchange in the West Atlantic Ocean. Environ. Sci. Technol. 47, 7757–7765. Streets, D.G., Devane, M.K., Lu, Z., Bond, T.C., Sunderland, E.M., Jacob, D.J., 2011. All-time releases of mercury to the atmosphere from human activities. Environ. Sci. Technol. 45 (24), 10485–10491. Szymczycha, B., Miotk, M., Pempkowiak, J., 2013. Submarine ground water discharge as a source of mercury in the Bay of Puck, the southern Baltic Sea. Water Air Soil Pollut. 224, 1542. http://dx.doi.org/10.1007/s11270-013-1542-0. Tomiyasu, T., Matsuyama, A., Eguchi, T., Fuchigami, Y., Oki, K., Horvat, M., Rajar, R., Akagi, H., 2006. Spatial variations of mercury in sediment of Minamata Bay, Japan. Sci. Total Environ. 368, 283–290. Tomiyasu, T., Matsuyama, A., Eguchi, T., Marumoto, K., Oki, K., Akagi, H., 2008. Speciation of mercury in water at the bottom of Minamata Bay, Japan. Mar. Chem. 112, 102–106. Tomiyasu, T., Takenaka, S., Noguchi, Y., Kodamatani, H., Matsuyama, A., Oki, K., Kono, Y., Kanzaki, R., Akagi, H., 2014. Estimation of the residual total mercury in marine sediments of Minamata Bay after a pollution prevention project. Mar. Chem. 159, 19–24. Tseng, C.M., Lamborg, C.H., Hsu, S.C., 2013. A unique seasonal pattern in dissolved elemental mercury in the South China Sea, a tropical and monsoon-dominated marginal sea. Geophys. Res. Lett. 40, 167–172. http://dx.doi.org/10.1029/2012GL054457. U.S. Environmental Protection Agency, 2002. Method 1631. Revision E: Mercury in Water by Oxidation, Purge and Trap, and Cold Vapor Atomic Fluorescence Spectrometry. EPA-821-R-02-019, United States. United Nations Environmental Programme (UNEP), 2013. Global Mercury Assessment 2013 — Sources, emissions, releases and environment transport. DTI/1636/GE. Wangberg, I., Schmolke, S., Schgar, P., Munthe, J., Ebinghaus, R., Iverfeldt, A., 2001. Estimates of air–sea exchange of mercury in the Baltic Sea. Atmos. Environ. 35, 5477–5484. Wanninkhof, R., 1992. Relationship between wind speed and gas exchange over the ocean. J. Geophys. Res. 97 (C5), 7373–7382. Yano, S., 2013. In-situ measurement of mercury transport in the seawater of Minamata Bay. Procedia Earth Planet. Sci. 6, 448–456. Yin, Y., Chen, B., Mao, Y., Wang, T., Liu, J., Cai, Y., Jiang, G., 2012. Possible alkylation of inorganic Hg(II) by photochemical processes in the environment. Chemosphere 88, 8–16.