Exchange of dimethyl sulfide (DMS) between the surface ocean and the lower atmosphere was examined by using proton transfer reaction‐mass spectrometry coupled with the gradient flux (PTR‐MS/GF) system. We deployed the PTR‐MS/GF system and observed vertical gradients of atmospheric DMS just above the sea surface in the subtropical and transitional South Pacific Ocean and the subarctic North Pacific Ocean. In total, we obtained 370 in situ profiles, and of these we used 46 data sets to calculate the sea‐to‐air flux of DMS. The DMS flux determined was in the range from 1.9 to 31 μmol m⁻² d⁻¹ and increased with wind speed and biological activity, in reasonable accordance with previous observations in the open ocean. The gas transfer velocity of DMS derived from the PTR‐MS/GF measurements was similar to either that of DMS determined by the eddy covariance technique or that of insoluble gases derived from the dual tracer experiments, depending on the observation sites located in different geographic regions. When atmospheric conditions were strongly stable during the daytime in the subtropical ocean, the PTR‐MS/GF observations captured a daytime versus nighttime difference in DMS mixing ratios in the surface air overlying the ocean surface. The difference was mainly due to the sea‐to‐air DMS emissions and stable atmospheric conditions, thus affecting the gradient of DMS. This indicates that the DMS gradient is strongly controlled by diurnal variations in the vertical structure of the lower atmosphere above the ocean surface.

1 Introduction

The ocean is one of the main sources of sulfur‐containing substances including COS, CS₂, H₂S, and dimethyl sulfide (DMS) in the atmosphere [Bates et al., 1992; Watts, 2000]. DMS has been highlighted as one of the essential compounds that regulate the Earth's climate [Charlson et al., 1987]. Produced from phytoplankton, DMS is ubiquitous in the surface ocean. Once DMS is emitted into the marine atmosphere, it is oxidized to sulfur dioxide, sulfuric acid, and sulfate aerosols. These acidic compounds influence atmospheric chemistry and contribute to the formation of cloud condensation nuclei (CCN) as well as to the growth of existing particles. The particles derived from DMS affect the Earth's radiation budget by directly scattering sunlight and indirectly influencing cloud physics and albedo.

Over the past several decades, advancements in field observations and computer modeling have contributed to better understanding of the role of DMS and to reevaluation of its importance in the Earth's system. Some studies indicate that the radiation budgets and the climate system are strongly affected by particles arising from human activity, sea salts, and organic compounds other than DMS [Murphy et al., 1998; Quinn and Bates, 2011] and that the sensitivity of the global CCN distribution to the DMS variability is low [Woodhouse et al., 2010]. On the other hand, DMS is one of the main precursors of CCN in the pristine marine atmosphere including the Southern Ocean [Bates et al., 1992; Chin et al., 2000]. Modeling study suggests that sensitivities of CCN formation to regional changes in DMS emission are different [Woodhouse et al., 2013]. Hence, the importance of DMS emission can be different depending on the region, and the investigation of the regional features of sea‐to‐air DMS flux can further advance our knowledge of regional CCN dynamics.

The sea‐to‐air flux of DMS can be estimated by the difference in DMS concentrations between the ocean surface and the lower atmosphere, multiplied by gas transfer velocity (k). Based on reevaluated climatology of sea surface DMS concentrations, Lana et al. [2011] estimated the mean annual emissions of DMS to be 28.1 Tg S a⁻¹. Land et al. [2014] recently calculated the DMS emission to be 19.6 Tg S a⁻¹, using the DMS concentration in surface seawater from Lana et al. [2011] combined with temperature and wind speed observed from satellites. The difference of these DMS emissions is due to the choice of k, many of which are parameterized with wind speed, with the relationships between k and wind speed being variable. One recent study has used the transfer velocity derived from the measurement of DMS flux [Land et al., 2014], while the calculation of the global DMS emission has often used the transfer velocity of other gases, such as CO₂, He, and SF₆ [e.g., Kettle et al., 1999; Lana et al., 2011].

Recently, efforts have been made to determine the gas transfer velocity of DMS with the eddy covariance (EC) technique coupled with fast detection (> 1 Hz) by using online mass spectrometry which is atmospheric pressure ionization‐chemical ionization mass spectrometry (API‐CIMS) [Bell et al., 2013, 2015; Huebert et al., 2004, 2010; Marandino et al., 2007, 2009; M. Yang et al., 2009; Yang et al., 2011]. These EC observations reported the k value for DMS on short time scales ranging from 10 min to 1 h, indicating a difference in the gas transfer velocity between DMS and insoluble gases. Measurements of gas exchange using insoluble gases have suggested that the relationship between k and wind speed is nonlinear [Ho et al., 2006; Nightingale et al., 2000; Wanninkhof, 1992]. In contrast, the majority of direct DMS flux measurements suggest a linear relationship between k and wind speed [Marandino et al., 2009; Yang et al., 2011]. Blomquist et al. [2006] suggest that the differences in these relationships might be due to the disproportionate influence of bubbles on the flux of insoluble gases.

While these observations made by independent approaches are useful to better determine the k values for DMS, many of them currently rely on the EC technique. The gradient flux (GF) technique combined with gas chromatographic (GC) detection for air DMS concentrations has been used for measurements of DMS flux and for determination of k [Hintsa et al., 2004; McGillis et al., 2001; Zappa et al., 2007; Zemmelink et al., 2002, 2004]. The GF observations have shown a nonlinear relationship between k and wind speed in the coastal waters and Southern Ocean, which is different from the linear relationship determined by the EC observations. The GF method requires assumptions about the scalar profile in the surface layer, and thus a less direct method than the EC method. Further observations with the GF technique which are independent of the EC will help provide more insights to improve our knowledge of gas transfer. However, the GF technique combined with GC has advantages and disadvantages of its air‐sampling method. The GC/GF method takes in large volumes of air samples at different heights simultaneously over ~30 min to measure the profiles of DMS concentrations [Hintsa et al., 2004; Zappa et al., 2007; Zemmelink et al., 2002, 2004]. One advantage of simultaneous sampling is that the measured DMS concentration represents the average at each profile. On the other hand, one disadvantage of long‐time sampling is that the DMS concentrations might be influenced by the temporal atmospheric variability due to meteorological factors such as horizontal transport in a short time.

Our group recently developed a technique combining the GF technique with the proton transfer reaction‐mass spectrometry (PTR‐MS) system (PTR‐MS/GF) and succeeded in measuring the DMS and acetone fluxes from the platform of the research vessel (R/V) Hakuho Maru [Tanimoto et al., 2014]. The PTR‐MS/GF system continuously measures the profiles of DMS concentrations in the marine atmosphere just above the ocean surface. The DMS measurements at different heights are not simultaneous but are successively switched at 1 min intervals. Successive measurements allow the determination of the vertical DMS profiles in the short time scale (i.e., each 7 min cycle). If the DMS profiles are influenced by the temporal atmospheric variability, the DMS mixing ratio in the profiles might be varied without gradient. The short‐time measurements of the profiles using the PTR‐MS/GF system can evaluate whether each profile is influenced by the temporal atmospheric variability or not. Moreover, as PTR‐MS can detect various volatile organic compound (VOC) species, such as isoprene, methanol, and acetone, with high sensibility and high time resolution, the PTR‐MS/GF system has a great potential to measure the fluxes of multiple VOCs between the ocean and the atmosphere. At present, there is only one air‐sea flux data set obtained by the PTR‐MS/GF system in the subtropical North Pacific Ocean [Tanimoto et al., 2014]. Further, flux observations using the PTR‐MS/GF system are needed to help improve our understanding of the air‐sea fluxes of VOCs in the open ocean.

In this paper we describe vertical profiles of DMS above the sea surface observed using the PTR‐MS/GF system in three ocean regions: the subtropical and transitional South Pacific and the subarctic North Pacific Ocean. Then we present sea‐to‐air fluxes of DMS and analyze the gas transfer velocity of DMS under different environmental conditions in wind speed (1 ~ 10 m s⁻¹), atmospheric boundary layer stability and biological activity. The gas transfer velocity of DMS determined by the PTR‐MS/GF technique was compared to that derived by the EC technique, the NOAA/COARE model, and other GF observations.

2 Materials and Methods

2.1 Study Sites

The observations using the PTR‐MS/GF system were made at eight sites during three cruises by R/V Hakuho Maru, operated by the Japan Agency for Marine‐Earth Science and Technology (JAMSTEC), between December 2011 and January 2012 in the eastern South Pacific Ocean (KH‐11‐10), between August–October 2012 in the subarctic North Pacific Ocean (KH‐12‐4) and between December 2013 and January 2014 in the South Pacific Ocean (KH‐13‐7) (Figure 1 and Table 1). In general, the PTR‐MS/GF observations at each site were conducted during 2–3 h when the ship was stationary either during the day or at night. In order to examine diurnal variations in the DMS profiles in the subtropical region, the PTR‐MS/GF observations were made during both daytime and nighttime at a single station, Station 21 on the KH‐11‐10 cruise. In addition, measurements of DMS concentrations in the surface seawater were made while the ship was in motion.

**Figure 1**
Open in figure viewer PowerPoint

Ship tracks (gray lines) and flux observation stations (circles) during the KH‐10‐1 (light blue), subtropical sites (green), and transitional site (orange) on KH‐11‐10, KH‐12‐4 (blue), and KH‐13‐7 (pink) cruises made in the subtropical and subarctic Pacific Ocean. Black lines represent 7 day back trajectories originating from the flux observation sites at the height of 500 m.

Table 1. Summary of Observation Sites on Three Cruisesa^a Standard deviations are in parenthesis.

Cruise	Station	Site	Region	Date	Wind Speed (m/s)		Temperature (°C)		Sea Surface Temperature (°C)		Chl a (mg/m³)	Slope		Sea DMSb^b Surface seawater DMS concentrations were measured at a depth of 5 m. (nM)		Air DMSc^c Surface air DMS mixing ratios were measured at a height of 14 m. (ppb)		DMS Flux (μmol/m²/d)
KH‐11‐10	15	23°S, 120°W	Subtropical	2012/1/9	9.2	(0.6)	24.3	(0.3)	25.6	(0.0)	0.06	−0.025	(0.002)	3.29	(0.00)	0.16	(0.02)	12.1	(1.2)
	18	30°S, 107°W	Subtropical	2012/1/14	3.2	(0.5)	21.7	(0.2)	22.5	(0.3)	0.02	−0.033	(0.008)	2.34	(0.34)	0.40	(0.04)	5.2	(1.6)
	21	23°S, 100°W	Subtropical	2012/1/18–19	4.1	(2.1)	23.7	(1.0)	24.5	(0.1)	0.02	−0.029	(0.007)	1.66	(0.09)	0.36	(0.04)	6.6	(3.4)
	22	20°S, 100°W	Transitional	2012/1/21	7.3	(0.4)	23.6	(0.1)	24.7	(0.1)	0.09	−0.071	(0.010)	6.82	(0.19)	0.68	(0.07)	26.1	(3.5)
KH‐12‐4	9	47°N, 170°E	Subarctic	2012/9/2	6.2	(0.3)	13.1	(0.1)	13.5	(0.0)	0.56	−0.024	(0.014)	3.17	(0.26)	0.48	(0.03)	7.4	(4.4)
	11	47°N, 180°W	Subarctic	2012/9/7	8.4	(0.1)	13.2	(0.0)	12.7	(0.1)	0.82	−0.025	(0.009)	2.74	(0.21)	0.88	(0.02)	9.8	(3.4)
	17	43°N, 132.6°W	Subarctic	2012/9/28	7.9	(0.2)	18.0	(0.0)	18.1	(0.1)	0.13	−0.041	(0.013)	4.00	(0.31)	0.78	(0.08)	14.8	(4.6)
KH‐13‐7	U	33°S, 175°W	Subtropical	2014/1/16	7.5	(3.4)	21.1	(0.8)	22.5	(0.0)	0.08	−0.029	(0.006)	2.37	(0.25)	0.49	(0.12)	11.1	(2.8)

^a Standard deviations are in parenthesis.
^b Surface seawater DMS concentrations were measured at a depth of 5 m.
^c Surface air DMS mixing ratios were measured at a height of 14 m.

2.2 Gradient Flux Technique Combined With PTR‐MS

2.2.1 Gradient Flux Technique

For a gas emitted from the ocean into the atmosphere, there is a mean gradient in the lower atmosphere, decaying with elevation above the ocean surface. Vertical mixing by turbulent eddies causes upward transport of a gas. Using the Monin‐Obukhov similarity theory, the flux (F) can be expressed as

$urn:x-wiley:2169897X:media:jgrd53925:jgrd53925-math-0001$ (1)

where u_* is the friction velocity, к is the von Karman constant (assigned a value of 0.4), C is the concentration of the gas, z is the height above the mean sea surface, and ϕ (z/L) is the Monin‐Obukhov stability function. Empirical relationships for stability functions have been proposed by many authors [e.g., Businger et al., 1971; Dyer and Hicks, 1970; Paulson, 1970]. In this study, we used the following relations [Garratt, 1992; Fairall et al., 1996]:

$urn:x-wiley:2169897X:media:jgrd53925:jgrd53925-math-0002$ (2)

where z is the measurement height and L is the Monin‐Obukhov length:

$urn:x-wiley:2169897X:media:jgrd53925:jgrd53925-math-0003$ (3)

where T_v is the virtual temperature, g is the acceleration due to gravity, and F_v is the buoyancy flux; and

$urn:x-wiley:2169897X:media:jgrd53925:jgrd53925-math-0004$ (4)

where H and E are the sensible and latent heat flux, respectively, and c_p is the specific heat of air at constant pressure, ρ is the air density, and T is air temperature. We calculated u*, H, and E using the COARE 3.0 algorithm [Fairall et al., 2003]. Thus, the flux can be determined based on the observed parameters: the slope of the gradient from the ocean surface, and meteorological data in air and ocean such as wind velocity, ambient temperature, sea surface temperature, and humidity [Iwata et al., 2004].

2.2.2 Measurements of DMS Profiles in the Air Above the Ocean Surface

The profiles of DMS concentrations in the air just above the ocean surface were obtained using a floating buoy platform [see also Tanimoto et al., 2014, Figures 1 and 2]. In brief, the floating buoy was placed approximately 25 m upwind of the ship by the ship's boom, to minimize the influence of turbulence due to the existence of the ship and/or contamination from the ship's facilities/laboratories. The buoy was equipped with six sampling ports at elevations of 1, 5 (2 or 3), 10 (or 20), 35 (or 50), 110, and 245 cm above sea level. The sampling at less than 5 cm was made for a technical testing purpose. An additional sampling inlet was placed on the upper deck of the ship (14 m above sea level (asl)). Membrane filters (Advantec‐Toyo Ltd.) were placed at the individual inlets to prevent seawater droplets from being sucked into the sampling tubes. Ambient air sampled at the individual inlets was continuously sucked at a flow rate of 0.7 standard liters per minute toward a switching unit by a perfluoroalkoxy‐Teflon tube (3 mm inner diameter, 30–40 m long). All the seven sampling lines were continuously pumped from the inlets to the switching unit by individually pumps to avoid ambient air being stagnant in the sampling lines. Sample streams were then switched by automated valves in the switching unit at 1 min intervals, and only the selected stream was directed to PTR‐MS in a sequential manner. Simultaneously, the mixing ratios of H₂O at each height were measured by a temperature/humidity sensor (Visala, model HMP45A) to obtain the latent heat (water vapor) flux on the KH‐11‐10 and KH‐12‐4 cruises. The latent heat flux of the KH‐13‐7 cruise was not determined due to failure of the sensor. Fundamental meteorological parameters including wind velocity, wind direction, and temperature and humidity in ambient air were monitored by custom‐built sensors (Nippon Electric Instrument, Inc.) installed on the upper deck of the ship. Surface seawater temperature was measured by a thermometer.

For detection of DMS we used a commercially available PTR‐MS instrument (PTR‐MS‐hs, IONICON Analytik). The ion signal at m/z = 63 was obtained as DMS at 5 s intervals. The detection sensitivity under dry conditions was determined by dynamic dilution of a gravimetrically prepared standard gas balanced with ultrapure N₂ (5 ppm, Japan Fine Products Co.). The detection sensitivity depended on the humidity of the sample air. The DMS concentration at each height was determined by the detection sensitivity and humidity correction factor [Kameyama et al., 2009].

In order to obtain reliable data sets for flux calculations with minimum errors, we selected the data with the following two criteria. The first criterion is a wind direction. The position of the buoy relative to the ship was always fixed on the port side of the ship, because the buoy was supported by a wire from the ship's boom. To obtain DMS profiles without any effect of the ship, we selected the data obtained with a wind direction of −20° to −60° relative to the ship, which was from the port side of the ship's front. The second is the determination coefficient (R²) being more than 0.5 in the gradient of the vertical DMS profiles for each 7 min cycle. In the GF technique, because the DMS profiles should be the log profiles predicted by theory, they were validated by the R² criterion.

For the flux calculation, we determined slopes of the data from 3 to 4 consecutive cycles (21–28 min) that met the above two criteria, resulting in one single flux data. We calculated the uncertainties in the slope (95% confidence interval) according to a previous study [Zemmelink et al., 2004]. The uncertainty in the slope on the three cruises was around 40%. The uncertainty of the sampling height due to wave motions in the determination of the DMS flux was estimated to be less than 10%. Overall, our conservative estimate of the total uncertainty in the determination of the DMS flux by the PTR‐MS/GF method was approximately 50%.

2.3 Measurement of the Sea Surface DMS Concentration

The concentrations of DMS dissolved in the surface seawater were measured with the Equilibrator Inlet‐PTR‐MS (EI‐PTR‐MS) technique before and after the GF observation [Kameyama et al., 2009, 2010; Omori et al., 2013]. The surface seawater sampled by the ship's pumping system from a depth of approximately 5 m was continuously flowed into a glass‐equilibrator (volume of 10 L) at a flow rate of 1 L min⁻¹. Ultrapure nitrogen was injected from the bottom inlet of the equilibrator at a flow rate of 120 cm³ min⁻¹ STP, and the gas‐phase samples were introduced to PTR‐MS. The PTR‐MS measurements were conducted in MID mode at a 5 s integration time for each mass per cycle to obtain the mass signals at 1 min intervals. The detection sensitivity for VOCs under dry conditions and its humidity dependence were determined following previous work [Kameyama et al., 2009]. Given that DMS is in equilibrium between the gas and liquid phases, DMS concentrations in the seawater samples were calculated from the concentrations in the sample gas extracted from the equilibrator by the Henry's Law solubility [Kameyama et al., 2009].

2.4 Calculation of Transfer Velocity of DMS

The transfer velocities of DMS were calculated using the following equation:

$urn:x-wiley:2169897X:media:jgrd53925:jgrd53925-math-0005$ (5)

where k_DMS is the DMS transfer velocity, C_w is the DMS concentration in the surface seawater, C_a is the atmospheric DMS concentration, and α is the Henry's law solubility for DMS from Kameyama et al. [2009]. The α in Kameyama et al. [2009] was the average of previous studies [Cline and Bates, 1983; Dacey et al., 1984; De Bruyn et al., 1995; Przyjazny et al., 1983; Wong and Wang, 1997]. We used C_w at the depth of 5 m and C_a at the height of 14 m. k_DMS results also from waterside transfer velocity (k_w) and airside transfer velocity (k_a):

$urn:x-wiley:2169897X:media:jgrd53925:jgrd53925-math-0006$ (6)

Airside resistance to the exchange of sparingly soluble gases is small and usually neglected. For DMS, however, as gas solubility increases with decreasing temperature, the air resistance becomes more important. We normalized the transfer velocity of DMS for temperature dependence in the airside resistance as suggested by McGillis et al. [2000] and Yang et al. [2011]:

$urn:x-wiley:2169897X:media:jgrd53925:jgrd53925-math-0007$ (7)

where γ is the airside gradient factor affecting the flux. The value of γ was calculated from the ratio of the water and airside transfer velocity: γ = 1/(1 + k_a/αk_wc). k_a was calculated with the following equation: k_a = 8814u_* + 6810u_*², following Yang et al. [2013]. U_10N is neutral wind speed at the height of 10 m, which was calculated using the COARE 3.0 algorithm [Fairall et al., 2003]. k_wc is waterside transfer velocity calculated using the formulation from the NOAA/COARE gas transfer model:

$urn:x-wiley:2169897X:media:jgrd53925:jgrd53925-math-0008$ (8)

Here ρ_w and ρ_a are the density of water and air, respectively, δ_w is the waterside molecular sublayer thickness, ĸ the von Karman constant (0.4), and h_w = 13.3/(AΦ). In h_w, A is an empirical constant (1.3) [Blomquist et al., 2006] and Φ accounts for surface buoyancy flux enhancement of the transfer which only becomes important in wind speeds less than 2 m s⁻¹. The white cap fraction ƒ_wh = 3.84 10⁻⁶ U_10N^3.14 [Woolf, 1997], and G = α⁻¹[1 + (14αS_c^–1/2)^–1/1.2]^–1.2. B is an empirical constant, 1 [Blomquist et al., 2006]. The k_w was normalized by the Schmidt number (S_c) to yield k_w660: k_w660 = k_w (S_c/660)^1/2. The airside transfer was normalized to 27.2°C (S_c = 660 for DMS). The total normalized transfer velocity was determined by k₆₆₀ = (1/k_w660 + α₆₆₀/k_a)^–1. α₆₆₀ is the solubility of DMS at 27.2°C.

3 Results and Discussion

3.1 Environmental Conditions

The observations of the air‐sea DMS flux were made in the subtropical South Pacific (Stations 15, 18, and 21 on the KH‐11‐10 cruise and Station U on the KH‐13‐7 cruise), transitional zone between subtropical and upwelling waters (Station 22 on the KH‐11‐10 cruise) and subarctic North Pacific waters (Stations 9, 11, and 17 on the KH‐12‐4 cruise) (Figure 1 and Table 1). The subtropical region was characterized by high sea surface temperatures around 22–25.6°C and low chlorophyll a concentrations. In the transitional zone, the chlorophyll a concentration was 1.5–4 times higher than that in the subtropical region, although other environmental parameters in both of the transitional and subtropical areas were similar. The subarctic region was characterized by low sea surface temperature ranging from 14 to 18°C and high chlorophyll a levels. The atmospheric boundary layer stability was close to neutral (|z/L| < 0.05) in all regions except Stations 18 and 21 on the KH‐11‐10 and KH‐13‐7 cruises (Table S1 in the supporting information). The atmospheric instability at Station 21 (z/L = −0.05 to −0.57) was due to the higher temperature in the sea than in the air and a low wind speed (< 4 m s⁻¹).

The sea surface DMS concentrations in the subtropical region ranged from 1.5 to 3.3 nM (average: 2.1 ± 0.5 nM, Table 1), which is consistent with previous work reporting approximately 2.15 nM [Bates et al., 1987], and those for this area compiled in the Lana et al. [2011] database. In the transitional zone (Station 22 on the KH‐11‐10 cruise) the seawater DMS concentrations were approximately 2 times higher than those in the subtropical region due to high biological activity (Table 1). This is in good agreement with past observations in the transitional zone in the South Pacific (10–20°S), reporting the DMS concentrations of 7 to 8 nM [Bates and Quinn, 1997; Marandino et al., 2009]. The seawater DMS concentrations in the subarctic zone ranged from 2.7 to 4.3 nM (with an average of 3.7 ± 0.6 nM), which is consistent with previous reports [Aranami et al., 2002; Asher et al., 2011; Kameyama et al., 2009; Wong et al., 2005].

3.2 DMS Profiles in the Air Above the Ocean Surface

Figure 2 shows the vertical profiles of DMS mixing ratios in the air above the ocean surface up to 14 m asl. From the three cruises. In total, 370 profiles of the DMS vertical gradient were obtained with the PTR‐MS/GF system. Of these 46 profiles were selected according to the criteria of wind direction and the determination coefficient of the DMS gradient (slope of the height versus concentration) (R² > 0.5).

**Figure 2**
Open in figure viewer PowerPoint

Typical vertical profiles of atmospheric DMS mixing ratios above the ocean surface in the (a) subtropical ocean (Station 15, KH‐11‐10), (b) subarctic ocean (Station 17, KH‐12‐4), and (c) subtropical ocean (Station U, KH‐13‐7). Gray thin lines indicate individual measurements with red thick lines being the mean profiles.

As seen in Table 1, the atmospheric DMS mixing ratios observed at the height of 14 m were low (on average 0.37 ± 0.09 ppb) in the subtropical region, compared to those in other regions. Higher DMS levels at this height were measured in the transitional zone (0.68 ppb) and the subarctic region (0.48–0.89 ppb, average 0.74 ± 0.15 ppb). These regional differences in the atmospheric DMS concentrations are consistent with those previously observed in the south Pacific Gyre [Marandino et al., 2009]. However, the DMS levels in air observed in the South Pacific (on average 0.49 ppb, at Station U on the KH‐13‐7 cruise) were higher than those in previous observations (up to 0.3 ppb) [Marandino et al., 2009], although the seawater DMS level at this site was in good agreement with previous data. Presumably, air masses at this site were influenced by air masses from the Southern Ocean that was associated with high biological activity. The 7 day back trajectories indicated that air masses arrived at Station U from the Southern Ocean over New Zealand (Figure 1). In addition, high levels of methanesulfonic acid (MSA) derived mainly from DMS were detected around Station U (M. Uematsu, University of Tokyo, personal communication, 2015), indicating that the cause of high atmospheric DMS levels observed at Station U was long‐range transport from the Southern Ocean, not local emissions from the sea surface water nearby.

The vertical profiles of DMS above the ocean surface show a negative gradient against the height (Figure 2), demonstrating that DMS is emitted from the sea surface to the lower atmosphere [Hintsa et al., 2004; McGillis et al., 2001; Tanimoto et al., 2014; Zappa et al., 2007; Zemmelink et al., 2002, 2004]. The gradients in the DMS profiles as denoted as the slope of the DMS mixing ratios against the logarithmic‐scaled (or semilogarithmic)‐scaled height were −0.03 ± 0.01 in the subtropical and subarctic regions and −0.07 ± 0.01 in the transitional zone at Station 22 on the KH‐11‐10 cruise (see Table 1). In the subtropical and subarctic regions, there is no relationship between the slopes and environmental parameters including the oceanic and atmospheric DMS concentrations, wind speed, and temperature. The slope in the transitional zone (Station 22) shows the strongest gradient in DMS in air, due mainly to high DMS emissions from the sea surface water. The high DMS level in the sea surface water in the transitional zone seems to contribute to the DMS pool in the air just over the sea surface, resulting in the strong gradient in the DMS profiles.

3.3 Flux Obtained by the PTR‐MS/GF System

Figure 3 illustrates the comparison of the latent heat (water vapor) fluxes determined by the gradient flux method and those calculated by the NOAA/COARE 3.0 gas transfer model [Fairall et al., 2003]. Based on the reduced‐major‐axis (RMA) regression [Ayers, 2001], there is a high correlation between the two data sets of the fluxes with the regression line being generally close to a 1:1 correspondence (slope = 0.96 ± 0.05, intercept = −4.0 ± 4.8, r² = 0.89). This correspondence supports the validity of the GF technique used in this study.

**Figure 3**
Open in figure viewer PowerPoint

Comparison of latent heat (water vapor) fluxes observed by the gradient flux method against those calculated by NOAA COARE 3.0 gas transfer model, with the best fit line (bold solid line) obtained by the reduced‐major‐axis (RMA) regression (y = 0.96 ± 0.05x − 4 .0 ± 4.8, r² = 0.89, n = 40) with 95% confidence interval (solid lines) and the 1:1 correspondence line (dashed line).

In Figure 4 the sea‐to‐air DMS fluxes obtained by the PTR‐MS/GF system are plotted as a function of 10 m neutral wind speed (U_10N). After the selection by wind direction relative to the ship, the fluxes were in the range from −9.7 to 35 μmol m⁻² d⁻¹ (Figure 4a). Then, we selected the profiles associated with high determination coefficient (R² > 0.5) for the successive 3–4 cycles (> 20 min) to exclude the data apart from the logarithmic curves. The slopes of the profiles with lower R² (less than 0.5) tended to be lower than that with higher R² (Figure S2). The DMS fluxes that met with the above two criteria ranged from 1.9 to 31 μmol m⁻² d⁻¹.

**Figure 4**
Open in figure viewer PowerPoint

Sea‐to‐air fluxes of DMS as a function of neutral wind speed at 10 m height (U_10N) measured at all the stations during the four cruises. (a) All flux data that have passed the wind direction criteria (gray) and fluxes satisfying both the wind direction and the determination coefficient (R²) criteria (orange). (b) The same flux data (using the data passing the two criteria) as in Figure 4a but color coded by the cruise number. Error bars represent the 95% confidence interval of the slope of the DMS profiles. Dashed lines are shown for the sea‐to‐air DMS flux of 0 μmol m⁻² d⁻¹.

The DMS flux selected by two criteria enhanced with increasing wind speed (Figure 4b). The fluxes obtained from previous observations using the same PTR‐MS/GF system in the western North Pacific Ocean in May 2010 (KH‐10‐1) [Tanimoto et al., 2014] are also plotted. In the subtropical South Pacific, the DMS fluxes were in the range from 1.9 to 14 μmol m⁻² d⁻¹ with the mean flux being 7.2 ± 3.6 μmol m⁻² d⁻¹. This is similar to 5.7 ± 6.6 μmol m⁻² d⁻¹ reported by Marandino et al. [2009] in the same latitudinal region of the South Pacific Ocean in January 2006 using the EC technique. The mean DMS flux in our study in the subarctic North Pacific Ocean was 13 ± 4.9 μmol m⁻² d⁻¹, also being consistent with previous observations reporting 9.4 and 24 μmol m⁻² d⁻¹ in August 2007 and 2008, respectively, in the eastern Pacific subarctic gyre [Asher et al., 2011], and 11.3 ± 16.9 and 20.1 ± 10.4 μmol m⁻² d⁻¹ in the western and eastern North Pacific, respectively, in July–September 1997 [Aranami et al., 2002]. In contrast, the DMS flux in the transitional zone was the highest, with an average of 26 ± 3.5 μmol m⁻² d⁻¹ (Table 1 and Figure 4), in good agreement with previous values of 0–36 μmol m⁻² d⁻¹ in the upwelling and transitional regions in the South Pacific [Bates et al., 1992; Marandino et al., 2009]. This high DMS flux is due to the high concentration of sea surface DMS that leads to a high level of DMS in the air, resulting in a large gradient over the sea surface (Figure S1).

3.4 Diurnal Variations of DMS Profiles Above the Ocean

In the PTR‐MS/GF observations we saw diurnal variations in the DMS profiles at Station 21 (KH‐11‐10) in the subtropical South Pacific (Figure 5). The observations from 1:00 to 13:00 UTC on August 18 showed a change in DMS profiles during night. While the DMS mixing ratios just over the sea surface (< 1 m asl) were around 0.6 ppb during early night, they decreased to 0.30 ppb at night. As a result, the vertical gradients in the DMS profiles became weak from early to late night (Figure 5a). DMS concentration in the sea surface water also decreased from 1.9 nM to 1.6 nM as the night progressed (Figure 5b). However, the seawater DMS decrease was not able to explain the air DMS reduction by half. Since the sea surface temperature was lower than the air temperature during daytime, the stronger vertical gradients observed at early night are likely due to stable atmospheric stratification (as shown in Figure 5c). As the night progressed, the air temperature became lower than the sea surface temperature (Figure 5c). This induced vertical mixing in the atmosphere above the ocean surface, contributing to the dilution of DMS over the sea surface, thus resulting in a weaker vertical gradient relative to that during daytime.

**Figure 5**
Open in figure viewer PowerPoint

Time series of (a) the atmospheric DMS mixing ratios and slope in the DMS profiles, (b) the DMS concentration in seawater at the depth of 5 m, and (c) air temperature, SST, wind speed, and solar radiation during night (18 January 2012 1:00–13:00 UTC) and from early morning to daytime (19 January 2012 13:00–22:00 UTC) at Station 21 in the subtropical South Pacific.

In contrast, from early morning toward daytime (19 August 13:00–22:00 UTC) the DMS levels over the sea surface gradually increased (Figure 5a). This DMS increase was caused by the emissions of DMS from the sea surface, and the stable atmospheric stability contributed to gradually stronger DMS gradients during the day. The gradients of DMS mixing ratios varied diurnally with changes in the air structure. However, the diurnal variations of the DMS flux were not determined here, because the DMS profiles during the morning on 18 January were not the log profile (R² of slope < 0.5) due to unstable conditions of the atmosphere.

The diurnal cycle of the DMS mixing ratios in the surface marine air (approximately 10–15 m asl) is typically characterized by a maximum in the early morning and minimum during the late afternoon, as observed in the subtropical and equatorial Pacific Ocean [Marandino et al., 2007; Yvon and Saltzman, 1996]. This cycle is due mainly to photochemical losses of DMS during daytime. In the present work the opposite feature was observed in the subtropical South Pacific. The DMS levels in the surface marine air (< 1 m asl) were at their minimum at late night and increased from early morning to daytime. This diurnal change can be explained by vertical mixing in the air at night and stable atmospheric stratification during daytime. Therefore, we suggest that the DMS levels and profiles in the surface air over the ocean in the subtropical regions are strongly influenced by air structure in addition to photochemical reactions.

During daytime (19 August, 17:00–19:30 UTC), there was local rain that dramatically reduced the air temperature, and the atmospheric condition changed to become unstable for a short time (Figure 5c). However, the gradient in DMS did not show clear changes (Figure 5a). This suggests that local and sudden changes in atmospheric conditions do not affect air DMS profiles immediately. Recent observations of DMS flux by the EC technique suggest that shipboard flux measurements are not sensitive to the in situ seawater DMS concentrations, but changes in seawater DMS upwind of the ship were [Bell et al., 2015]. Therefore, it is possible that micrometeorological observations on board often reflect air conditions far from the observation sites.

3.5 Gas Transfer Velocity

The gas transfer velocity, k₆₆₀, as determined from the PTR‐MS/GF measurements are plotted against 10 m neutral wind speed (U_10N) (Figure 6). The k₆₆₀ values obtained from previous observations in the western North Pacific Ocean (KH‐10‐1) [Tanimoto et al., 2014] are also plotted. Figure 6a shows the data that have only passed the wind direction criterion, ranging from −21.3 to 79.9 cm h⁻¹, along with the data that have passed both the wind direction and the gradient (R² > 0.5) criteria, ranging from 1.4 to 35 cm h⁻¹. It is seen that the k₆₆₀ data satisfying the two criteria are scattered less than the data satisfying only the wind direction criterion.

**Figure 6**
Open in figure viewer PowerPoint

Plots of k₆₆₀ as a function of U_10N from the four cruises. Plots are (a) all k₆₆₀ data that have passed the wind direction criteria (gray) and k₆₆₀ satisfying both the wind direction and the determination coefficient (R²) criteria (orange). (b) The same k₆₆₀ (using the data passing the two criteria) as in Figure 6a but color coded by the site and cruise number. Error bars represent the uncertainty in k₆₆₀ determined by the 95% confidence interval of the slope of the DMS profiles. (c) The k₆₆₀ color coded by the atmospheric boundary layer stability (z/L). The two curves in Figures 6b and 6c indicate the k₆₆₀ predicted by parameterizations based on insoluble gases [*Nightingale et al.*, 2000, N00; *Ho et al.*, 2006, H06].

In general, the k₆₆₀ values derived by the PTR‐MS/GF technique show an increase with increasing wind speed but are associated with a wide range depending on differences in observation sites or regions (Figure 6b). For wind speeds up to 7 m s⁻¹, the k₆₆₀ for the subtropical South Pacific (Stations 18 and 21 on the KH‐11‐10 cruise and Station U on the KH‐13‐7 cruise) exhibited higher values than those in other regions (KH‐10‐1, Stations 15 and 22 on the KH‐11‐10 and KH‐12‐4 cruises). In particular, the k₆₆₀ at a wind speed lower than 4 m s⁻¹ at a subtropical site (Station 21 on the KH‐11‐10 cruise) showed a mean value of ~10 cm h⁻¹ (in a range between 4.9 and 25 cm h⁻¹), which was threefold higher than that in the subtropical North Pacific (KH‐10‐1 cruise) (Figure 6b). A high k₆₆₀ at weak wind speed was observed under unstable conditions with atmospheric stability lower than −0.05 (z/L = −0.05 to −0.57, Table S1 and Figure 6c). The unstable condition in the atmosphere was due to a sea surface temperature higher than the air temperature and a low wind speed. The physical process controlling surface turbulence under strongly unstable and low wind speed conditions is buoyancy, resulting in high k₆₆₀ at low wind speed. Previous studies of CO₂ flux observations in the equatorial ocean also showed that the buoyancy causes high levels of the CO₂ flux and k₆₆₀ of CO₂ at low wind speed [McGillis et al., 2004]. Therefore, k₆₆₀ values when z/L < −0.05 are omitted hereafter. This filtering removed nine data from the subtropical observations (Stations 18 and 21 on the KH‐11‐10 and KH‐13‐7 cruises).

Figure 7 shows that the k₆₆₀ values at Stations 18 and 21 on the KH‐11‐10 cruise in the subtropical South Pacific were about 2–3 times higher than those at other observation sites. The relationship between k₆₆₀ and U_10N at other observation sites (except for Stations 18 and 21 on the KH‐11‐10) is similar to that reported in previous studies [Ho et al., 2006; Nightingale et al., 2000]. Nightingale et al. [2000, N00] and Ho et al. [2006, H06] showed parameterizations of k using a quadratic function of wind speed from artificial injections of two volatile tracers (³He and SF₆) in the North Sea and the Southern Ocean, respectively.

**Figure 7**
Open in figure viewer PowerPoint

Plot of k₆₆₀ as a function of U_10N at Stations 18 and 21 on the KH‐11‐10 in the subtropical region and at other sites in the subtropical, transitional, and subarctic regions. The k₆₆₀ data were not used when z/L < −0.05. Error bars represent the uncertainty in k₆₆₀ determined by the 95% confidence interval of the slope of the DMS profiles. Curves indicate k values predicted by parameterizations with insoluble gases [*Nightingale et al.*, 2000, N00; *Ho et al.*, 2006, H06], the best fit with eddy covariance (EC) observations [*Yang et al.*, 2011, Yang11], and the NOAA/COARE model (A = 1.3, B = 1.0) [*Blomquist et al.*, 2006, Blomquist06].

The k₆₆₀ values at Stations 18 and 21 on the KH‐11‐10 cruise in the subtropical South Pacific were higher than in N00 and H06 (Figure 7). The high k₆₆₀ may reflect that the real gas exchange is higher than that estimated from wind speed and the DMS concentration in the surface water, although there is no obvious reason why the k₆₆₀ is high in the subtropical ocean. One possible factor is direct contribution of the high DMS levels in the surface microlayer to the DMS flux. The seawater DMS concentrations in the near surface water (at a depth of approximately 5 m) sampled by the ship's underway pumping system are generally used as the sea surface concentrations [e.g., Kettle et al., 1999; Lana et al., 2011]. However, there might be gradients of DMS in the surface seawater that can arise from a variety of factors, including biological production/degradation and photochemical loss [Kieber et al., 1996], ventilation from sea to air [Zemmelink et al., 2005], and the existence of a sea surface microlayer [Yang et al., 2005]. It was reported that the DMS level in the surface microlayer was 0.64–2.9 times higher than that in the bulk surface water [Yang et al., 2005; G. P. Yang et al., 2009]. In oligotrophic regions such as the subtropical South Pacific Ocean, surfactant enrichment in the microlayer is greater than in more productive waters, and it was reported that the microlayer is stable enough to exist even at a wind speed of 6.6 m s⁻¹ [Wurl et al., 2011]. Therefore, we interpret that the sea surface DMS levels are likely to be higher than the bulk DMS concentrations in the subtropical Pacific Ocean, even at moderate wind speeds (3–7 m s⁻¹), resulting in the overestimation of the gas transfer velocity.

3.6 Comparison of the Gas Transfer Velocity

Several research groups have made direct measurements of the DMS fluxes by fast mass spectrometry‐based EC methods and determined k values. For example, Marandino et al. [2009] showed DMS fluxes and their k values in the equatorial to southern region of the eastern South Pacific Ocean during January 2006. Yang et al. [2011] summarized the k values obtained from five observations in the Equatorial East Pacific, Sargasso Sea, Northeast Atlantic, Southern Ocean, and Southeast Pacific Ocean. Bell et al. [2013] measured the DMS flux in the North Atlantic bloom region in June/July 2011. In addition, Blomquist et al. [2006] estimated the relationship between wind speed and k for DMS using the COARE gas transfer model (A = 1.3, B = 1.0). These studies indicated that the gas transfer velocity of DMS is lower than those of insoluble trace gases determined by the dual tracer experiments [Ho et al., 2006; Nightingale et al., 2000], because of reduced bubble‐mediated exchange and more airside resistance [Yang et al., 2011].

The k₆₆₀ values determined by the GF technique were compared with those of other methods. Figure 7 shows two curves of the k₆₆₀‐to‐U_10N relationship derived from the EC observations [Yang et al., 2011] and the COARE model [Blomquist et al., 2006], together with the N00 and H06 parameterizations, plotted with the observed data here. The k₆₆₀ determined by PTR‐MS/GF agreed with those determined by the EC technique and the dual tracer experiments at wind speeds lower than 6 m s⁻¹ except for those in the subtropical South Pacific (Stations 18 and 21 on the KH‐11‐10 cruise). At wind speeds higher than 6 m s⁻¹, the range of k₆₆₀ values determined by PTR‐MS/GF became wide. The difference in k₆₆₀, determined by the EC technique and the dual tracer experiments, also increases with wind speed. The similarities between k₆₆₀ determined by our GF technique and other techniques depended on the observation sites. The k₆₆₀ determined at subtropical and transitional sites on the KH‐11‐10 was close to that of the EC technique, and other k₆₆₀ at subarctic and subtropical sites was close to that of insoluble gas in the dual tracer experiments. The PTR‐MS/GF observations in various regions show the possibility that the difference in between the k₆₆₀ of DMS determined by the EC technique and that of the insoluble gases in the dual tracer experiments was due to not only the solubility but also other environmental factors.

The curves of k₆₆₀ obtained by other GF methods are shown in Figure 8 [Hintsa et al., 2004; Zemmelink et al., 2004]. These authors determined k₆₆₀ from open oceanic observations in the subtropical and equatorial Pacific using the GF technique with gas chromatography. The k values for DMS determined by these two works are rather different: one is clearly threefold higher and the other is similar to the k values determined by the EC technique, and those in the present work lie in between. These results suggest a large uncertainty in determining the k₆₆₀ of DMS, in particular by GF methods, and strongly highlight the need of making simultaneous measurements by both the GF and EC techniques and comparing the results, to narrow down the uncertainty of k values for DMS, or at least to identify the methodological differences. This would ultimately lead to the scientific or technical advancement of the research community.

**Figure 8**
Open in figure viewer PowerPoint

Comparison of k₆₆₀ between the GF and EC methods. The k values (filled circles) and the k binned at 1 m s⁻¹ intervals (open circles) are determined at all sites other than Stations 18 and 21 on the KH‐11‐10. Error bars correspond to the standard deviations for the mean values within each bin. Curves indicate k values predicted by parameterizations with insoluble gases [*Nightingale et al.*, 2000, N00; *Ho et al.*, 2006, H06], the EC observations [*Yang et al.*, 2011, Yang11], the NOAA/COARE model (A = 1.3, B = 1.0) [*Blomquist et al.*, 2006, Blomquist06], and the GF observations [*Hintsa et al.*, 2004, Hintsa04; *Zemmelink et al.*, 2004, Zemmelink04].

Tanimoto et al. [2014] summarized the advantages and disadvantages of the PTR‐MS/GF method. Its advantages include that (1) the GF system is simple and its operation is not complicated, relative to the EC technique that requires correction of wind data for ship motion and distortion; (2) the system can be deployed in the open ocean off the side of “any type” of research vessel; and (3) a time resolution of the order of seconds (< 1 Hz) can be used. Thus, observations by the PTR‐MS/GF technique seem to be easier than observations by the EC technique and to measure a greater range of gas fluxes that cannot be measured fast enough for covariance. The present study has demonstrated an additional advantage of the PTR‐MS/GF method, that is, the ability to capture temporal and spatial variations in the DMS profiles in the air above the ocean surface as well as the DMS flux. This contributes to the advancement of our understanding of the factors controlling the variations of atmospheric DMS over the ocean surface. One of the disadvantages of the GF techniques is that the GF technique requires a stability function. This makes the GF technique less direct than the EC technique. Other disadvantages of the PTR‐MS/GF method are that the GF observations can be accomplished only under relatively calm conditions and deployed only when the ship is stationary. Therefore, it is difficult to collect a large amount of DMS flux data by the GF method. The EC measurements can be made at wind speeds higher than 12 m s⁻¹ and when the ship is in motion [e.g., Bell et al., 2013; Yang et al., 2011]. With these advantages of the EC approach, a large quantity of DMS flux data can be accumulated under various conditions. On the other hand, turbulent eddies are not developed under low wind speed or stable conditions, which significantly increase the uncertainties in the EC measurements. The conditions might be an area where the GF technique can be very useful [Edson et al., 2004].

Thus, the GF and EC techniques can be complements to each other, and simultaneous observations using both techniques may help us to understand the atmospheric DMS dynamics and to make accurate evaluations of the DMS emissions from the ocean. In terrestrial environments, Park et al. [2013, 2014] studied the temporal changes in VOC fluxes using the EC method with high time resolution and examined VOC dynamics in the atmosphere upper plants using the GF method. Using this dual observation method, they succeeded in clarifying the diurnal variation of VOC flux and its controlling factors. The application of the GF and EC methods for simultaneous observations will be useful to evaluate the dynamics and flux of DMS and other VOCs in marine environments.

4 Conclusion

We measured the DMS flux between the surface ocean and the lower atmosphere, using the PTR‐MS/GF system in subtropical, transitional, and subarctic regions in the South and North Pacific Ocean. The PTR‐MS/GF system was successfully deployed, revealing regional variations in the DMS flux. The strongest gradient in the DMS profile in the air was observed at a site with high biological productivity, resulting in the highest DMS flux among the three oceanic regions. In the subtropical ocean, the GF observations captured diurnal variations in the profiles of atmospheric DMS. The atmospheric DMS mixing ratios decreased by vertical mixing during night and increased during daytime due to the oceanic emissions and stable atmospheric conditions. The diurnal cycle of the DMS mixing ratios in the marine boundary layer at the height of 10–20 m is typically characterized by a maximum in the early morning and minimum during the late afternoon mainly due to photochemical losses of DMS during the daytime [Marandino et al., 2007; Yvon and Saltzman, 1996]. In addition, the present study showed the opposite cycle for the atmospheric DMS mixing ratios in the surface marine air (< 1 m asl), pointing to the important role of atmospheric structure over the ocean.

Based on the DMS flux observed using the PTR‐MS/GF system, the sea‐to‐air transfer velocity of DMS was calculated. The gas transfer velocity of DMS determined by the PTR‐MS/GF system showed a similarity to either that determined by the EC technique or that of the insoluble gases in the dual tracer experiment, depending on the observation sites. This indicates that the difference in the environmental factors influences the gas transfer velocity of DMS, although it is difficult to identify the factors in this study. The similarity of the gas transfer velocity highlights the need to further discuss the environmental factors to affect the gas transfer velocity of DMS, by comparing these two techniques on the same ship platforms.

Acknowledgments

We express our sincere thanks to the captain and crew of the R/V Hakuho Maru and all the scientists on board for their support during the cruise, in particular, Atsushi Tsuda (University of Tokyo) for providing in situ biological data. The information on the cruises of KH‐11‐10 and KH‐12‐4 were obtained from the University of Tokyo (http://cesdweb.aori.u‐tokyo.ac.jp/~database/oceandb/cruise.cgi) and GEOTRACES International Data Assembly Centre (http://www.bodc.ac.uk/geotraces/data/inventories/), respectively. Any requests concerning the data in this paper should be directed to the corresponding author (Yuko Omori; omori.yuko.ft@u.tsukuba.ac.jp). We are grateful to Mingxi Yang (Plymouth Marine Laboratory), Tim Lueker (Scripps Institution of Oceanography), and other anonymous reviewers for valuable comments and suggestions improving the paper. Financial support was given by multiple grants: Grant‐in‐Aid for Scientific Research in Priority Areas “Western Pacific Air–Sea Interaction Study (W‐PASS)” (1867001); Grant‐in‐Aid for Scientific Research (B) (23310016 and 16H02967) and Grant‐in‐Aid for Scientific Research (A) (24241010 and 15H01732) from the Ministry of Education, Culture, Sports, Science and Technology, Japan; the Global Environment Research Fund (RFa‐1102) of the Ministry of the Environment, Japan; the Global Environment Research Account for National Institutes by the Ministry of the Environment, Japan, Japanese Association for Marine Biology (JAMBIO) (24‐02, 25‐44, 26‐09, and 27‐34); and Asahi Breweries Foundation. This research is a contribution to Surface Ocean Lower Atmosphere Study (SOLAS) and International Global Atmospheric Chemistry (IGAC) projects of the ex‐International Geosphere‐Biosphere Programme (IGBP).

Supporting Information

References

Aranami, K., S. Watanabe, S. Tsunogai, A. Ohki, and K. Miura (2002), Chemical assessment of oceanic and terrestrial sulfur in the marine boundary layer over the Northern North Pacific during summer, J. Atmos. Chem., 41, 49– 66.
Crossref CAS Web of Science®Google Scholar
Asher, E. C., A. Merzouk, and P. D. Tortell (2011), Fine‐scale spatial and temporal variability of surface water dimethylsufide (DMS) concentrations and sea–air fluxes in the NE Subarctic Pacific, Mar. Chem., 126(1–4), 63– 75, doi:10.1016/j.marchem.2011.03.009.
Crossref CAS Web of Science®Google Scholar
Ayers, G (2001), Comment on regression analysis of air quality data, Atmos. Environ., 35(13), 2423– 2325, doi:10.1016/S1352‐2310(00)00527‐6.
Crossref CAS ADS Web of Science®Google Scholar
Bates, T. S., and P. K. Quinn (1997), Dimethylsulfide (DMS) in the equatorial Pacific Ocean (1982 to 1996): Evidence of a climate feedback?, Geophys. Res. Lett., 24(8), 861– 864.
Wiley Online Library CAS ADS Web of Science®Google Scholar
Bates, T. S., J. D. Cline, R. H. Gammon, and S. R. Kelly‐Hansen (1987), Regional and seasonal variations in the flux of oceanic dimethylsulfide to the atmosphere, J. Geophys. Res., 92(C3), 2930– 2938.
Wiley Online Library CAS ADS Web of Science®Google Scholar
Bates, T. S., B. K. Lamb, A. Guenther, J. Dignon, and R. E. Stoiber (1992), Sulfur emissions to the atmosphere from natural sources, J. Atmos. Chem., 14, 315– 337.
Crossref CAS ADS Web of Science®Google Scholar
Bell, T. G., W. De Bruyn, S. D. Miller, B. Ward, K. Christensen, and E. S. Saltzman (2013), Air–sea dimethylsulfide (DMS) gas transfer in the North Atlantic: Evidence for limited interfacial gas exchange at high wind speed, Atmos. Chem. Phys., 13(21), 11073– 11087, doi:10.5194/acp‐13‐11073‐2013.
Crossref CAS ADS Web of Science®Google Scholar
Bell, T. G., W. De Bruyn, C. A. Marandino, S. D. Miller, C. S. Law, M. J. Smith, and E. S. Saltzman (2015), Dimethylsulfide gas transfer coefficients from algal blooms in the Southern Ocean, Atmos. Chem. Phys., 15(4), 1783– 1794, doi:10.5194/acp‐15‐1783‐2015.
Crossref ADS Web of Science®Google Scholar
Blomquist, B. W., C. W. Fairall, B. J. Huebert, D. J. Kieber, and G. R. Westby (2006), DMS sea‐air transfer velocity: Direct measurements by eddy covariance and parameterization based on the NOAA/COARE gas transfer model, Geophys. Res. Lett., 33, L07601, doi:10.1029/2006GL025735.
Wiley Online Library CAS ADS PubMed Web of Science®Google Scholar
Businger, J. A., J. C. Wyngaard, Y. Izumi, and E. F. Bradley (1971), Flux‐profile relationships in the atmospheric surface layer, J. Atmos. Sci., 28, 181– 189.
Crossref ADS Web of Science®Google Scholar
Charlson, R. J., J. E. Lovelock, M. O. Andreae, and S. G. Warren (1987), Oceanic phytoplankton, atmospheric sulfur, cloud albedo and climate, Nature, 326, 655– 661.
Crossref CAS ADS Web of Science®Google Scholar
Chin, M., R. B. Rood, S. J. Lin, J. F. Müller, and A. M. Thompson (2000), Atmospheric sulfur cycle simulated in the global model GOCART: Model description and global properties, J. Geophys. Res., 105(D20), 24671– 24687, doi:10.1029/2000JD900384.
Wiley Online Library CAS ADS Web of Science®Google Scholar
Cline, J. D., and T. S. Bates (1983), Dimethyl sulfide in the equatorial Pacific Ocean: A natural source of sulfur to the atmosphere, Geophys. Res. Lett., 10(10), 949– 952.
Wiley Online Library CAS ADS Web of Science®Google Scholar
Dacey, J. W. H., G. Wakeham, and B. L. Howes (1984), Henry's law constants for dimethylsulfide in freshwater and seawater, Geophys. Res. Lett., 11(10), 991– 994.
Wiley Online Library CAS ADS Web of Science®Google Scholar
De Bruyn, W. J., E. Swartz, J. H. Hu, J. A. Shorter, P. Davidovits, D. R. Worsnop, M. S. Zahniser, and C. E. Kolb (1995), Henry's law solubilities and Śetchenow coefficients for biogenic reduced sulfur species obtained from gas‐liquid uptake measurements, J. Geophys. Res., 100, D4, 7245– 7251, doi:10.1029/95JD00217.
Wiley Online Library ADS Web of Science®Google Scholar
Dyer, A. J., and B. B. Hicks (1970), Flux‐gradient relationships in the constant flux layer, Q. J. R. Meteorol. Soc., 96, 715– 721.
Wiley Online Library ADS Web of Science®Google Scholar
Edson, J. B., C. J. Zappa, J. A. Ware, W. R. McGillis, and J. E. Hare (2004), Scalar flux profile relationships over the open ocean, J. Geophys. Res., 109, C08S09, doi:10.1029/2003JC001960.
Wiley Online Library ADS Web of Science®Google Scholar
Fairall, C. W., E. F. Bradley, D. P. Rogers, J. B. Edson, and G. S. Young (1996), Bulk parameterization of air‐sea fluxes for Tropical Ocean‐Global Atmosphere Coupled‐Ocean Atmosphere Response Experiment, J. Geophys. Res., 101, 3747– 3764.
Wiley Online Library ADS Web of Science®Google Scholar
Fairall, C. W., E. F. Bradley, J. E. Hare, A. A. Grachev, and J. B. Edson (2003), Bulk parameterization of air‐sea fluxes: Updates and verification for the COARE algorithm, J. Clim., 16(4), 571– 591, doi:10.1175/1520‐0442(2003)016<0571:BPOASF>2.0.CO;2.
Crossref ADS Web of Science®Google Scholar
Garratt, J. R. (1992), The Atmospheric Boundary Layer, 316 pp., Cambridge Univ. Press, Cambridge, U. K.
Google Scholar
Hintsa, E. J., J. W. H. Dacey, W. R. McGillis, J. B. Edson, C. J. Zappa, and H. J. Zemmelink (2004), Sea‐to‐air fluxes from measurements of the atmospheric gradient of dimethylsulfide and comparison with simultaneous relaxed eddy accumulation measurements, J. Geophys. Res., 109, C01026, doi:10.1029/2002JC001617.
Wiley Online Library ADS Web of Science®Google Scholar
Ho, D. T., C. S. Law, M. J. Smith, P. Schlosser, M. Harvey, and P. Hill (2006), Measurements of air‐sea gas exchange at high wind speeds in the Southern Ocean: Implications for global parameterizations, Geophys. Res. Lett., 33, L16611, doi:10.1029/2006GL026817.
Wiley Online Library CAS PubMed Web of Science®Google Scholar
Huebert, B. J., B. W. Blomquist, J. E. Hare, C. W. Fairall, J. E. Johnson, and T. S. Bates (2004), Measurement of the sea‐air DMS flux and transfer velocity using eddy correlation, Geophys. Res. Lett., 31, L23113, doi:10.1029/2004GL021567.
Wiley Online Library CAS ADS Web of Science®Google Scholar
Huebert, B. J., B. W. Blomquist, M. X. Yang, S. D. Archer, P. D. Nightingale, M. J. Yelland, J. Stephens, R. W. Pascal, and B. I. Moat (2010), Linearity of DMS transfer coefficient with both friction velocity and wind speed in the moderate wind speed range, Geophys. Res. Lett., 37, L01605, doi:10.1029/2009GL041203.
Wiley Online Library ADS Web of Science®Google Scholar
Iwata, T., K. Yoshikawa, K. Nishimura, Y. Higuchi, T. Yamashita, S. Kato, and E. Ohtaki (2004), CO₂ flux measurements over the sea surface by eddy correlation and aerodynamic techniques, J. Oceanogr., 60, 995– 1000, doi:10.1007/s10872‐005‐0007‐5.
Crossref CAS Web of Science®Google Scholar
Kameyama, S., H. Tanimoto, S. Inomata, U. Tsunogai, A. Ooki, Y. Yokouchi, S. Takeda, H. Obata, and M. Uematsu (2009), Equilibrator inlet‐proton transfer reaction‐mass spectrometry (EI‐PTR‐MS) for sensitive, high‐resolution measurement of dimethyl sulfide dissolved in seawater, Anal. Chem., 81, 9021– 9026, doi:10.1021/ac901630h.
Crossref CAS PubMed Web of Science®Google Scholar
Kameyama, S., H. Tanimoto, S. Inomata, U. Tsunogai, A. Ooki, S. Takeda, H. Obata, A. Tsuda, and M. Uematsu (2010), High‐resolution measurement of multiple volatile organic compounds dissolved in seawater using equilibrator inlet proton transfer reaction‐mass spectrometry (EI‐PTR‐MS), Mar. Chem., 122, 59– 73, doi:10.1016/j.marchem.2010.08.003.
Crossref CAS Web of Science®Google Scholar
Kettle, A. J., et al. (1999), A global database of sea surface dimethylsulfide (DMS) measurements and a procedure to predict sea surface DMS as a function of latitude, longitude, and month, Global Biogeochem. Cycles, 13, 399– 444, doi:10.1029/1999GB900004.
Wiley Online Library CAS ADS Web of Science®Google Scholar
Kieber, D. J., J. F. Jiao, R. P. Kiene, and T. S. Bates (1996), Impact of dimethylsulfide photochemistry on methyl sulfur cycling in the equatorial Pacific Ocean, J. Geophys. Res., 101, 3715– 3722, doi:10.1029/95JC03624.
Wiley Online Library CAS ADS Web of Science®Google Scholar
Lana, A., T. G. Bell, R. Simó, S. M. Vallina, J. Ballabrera‐Poy, A. J. Kettle, J. Dachs, L. Bopp, E. S. Saltzman, J. Stefels, J. E. Johnson, and P. S. Liss (2011), An updated climatology of surface dimethylsulfide concentrations and emission fluxes in the global ocean, Global Biogeochem. Cycles, 25, GB1004, doi:10.1029/2010GB003850.
Wiley Online Library CAS ADS Web of Science®Google Scholar
Land, P. G., J. D. Shutler, T. G. Bell, and M. Yang (2014), Exploiting satellite earth observation to quantify current global oceanic DMS flux and its future climate sensitivity, J. Geophys. Res. Oceans, 119, 7725– 7740, doi:10.1002/2014JC010104.
Wiley Online Library CAS ADS Web of Science®Google Scholar
Marandino, C. A., W. J. De Bruyn, S. D. Miller, and E. S. Saltzman (2007), Eddy correlation measurement of the air/sea flux of dimethylsulfide over the North Pacific Ocean, J. Geophys. Res., 112, D03301, doi:10.1029/2006JD007293.
Wiley Online Library ADS Web of Science®Google Scholar
Marandino, C. A., W. J. De Bruyn, S. D. Miller, and E. S. Saltzman (2009), Open ocean DMS air/sea fluxes over the eastern South Pacific Ocean, Atmos. Chem. Phys., 9, 345– 356, doi:10.5194/acp‐9‐345‐2009.
Crossref CAS ADS Web of Science®Google Scholar
McGillis, W. R., J. W. H. Dacey, N. M. Frew, E. J. Bock, and R. K. Nelson (2000), Water‐air flux of dimethylsulfide, J. Geophys. Res., 105, 1187, doi:10.1029/1999JC900243.
Wiley Online Library CAS ADS Web of Science®Google Scholar
McGillis, W. R., J. B. Edson, J. D. Ware, J. W. H. Dacey, J. E. Hare, C. W. Fairall, and R. Wanninkhof (2001), Carbon dioxide flux techniques performed during GasEx‐98, Mar. Chem., 75(4), 267– 280, doi:10.1016/S0304‐4203(01)00042‐1.
Crossref CAS Web of Science®Google Scholar
McGillis, W. R., et al. (2004), Air‐sea CO₂ exchange in the equatorial Pacific, J. Geophys. Res., 109, C08S02, doi:10.1029/2003JC002256.
Wiley Online Library CAS ADS Web of Science®Google Scholar
Murphy, D. M., J. R. Anderson, P. K. Quinn, L. M. McInnes, F. J. Brechtel, S. M. Kreidenweis, A. M. Middlebrook, M. Pósfai, D. S. Thomson, and P. R. Buseck (1998), Influence of sea‐salt on aerosol radiative properties in the Southern Ocean marine boundary layer, Nature, 392(6671), 62– 65, doi:10.1038/32138.
Crossref CAS ADS Web of Science®Google Scholar
Nightingale, P. D., P. S. Liss, and P. Schlosser (2000), Measurements of air‐sea gas transfer during an open ocean algal bloom, Geophys. Res. Lett., 27(14), 2117– 2120, doi:10.1029/2000GL011541.
Wiley Online Library CAS ADS Web of Science®Google Scholar
Omori, Y., H. Tanimoto, S. Inomata, S. Kameyama, S. Takao, and K. Suzuki (2013), Evaluation of using unfiltered seawater for underway measurement of dimethyl sulfide in the ocean by online mass spectrometry, Limnol. Oceanogr. Methods, 11, 549– 560.
Wiley Online Library CAS Web of Science®Google Scholar
Park, J. H., A. H. Goldstein, J. Timkovsky, S. Fares, R. Weber, J. Karlik, and R. Holzinger (2013), Eddy covariance emission and deposition flux measurements using proton transfer reaction‐time of flight‐mass spectrometry (PTR‐TOF‐MS): Comparison with PTR‐MS measured vertical gradients and fluxes, Atmos. Chem. Phys., 13(3), 1439– 1456, doi:10.5194/acp‐13‐1439‐2013.
Crossref ADS Web of Science®Google Scholar
Park, J. H., S. Fares, R. Weber, and A. H. Goldstein (2014), Biogenic volatile organic compound emissions during BEARPEX 2009 measured by eddy covariance and flux–gradient similarity methods, Atmos. Chem. Phys., 14(1), 231– 244, doi:10.5194/acp‐14‐231‐2014.
Crossref ADS Web of Science®Google Scholar
Paulson, C. A. (1970), The mathematical representation of wind speed and profile temperatures in the unstable atmospheric surface layer, J. Appl. Meteorol., 9, 857– 861.
Crossref ADS Google Scholar
Przyjazny, A., W. Janicki, W. Chrzanowski, and R. Staszewski (1983), Headspace gas chromatographic determination of distribution coefficients of selected organosulphur compounds and their dependence on some parameters, J. Chromatogr., 280, 249– 260, doi:10.1016/S0021‐9673(00)91567‐X.
Crossref CAS Web of Science®Google Scholar
Quinn, P. K., and T. S. Bates (2011), The case against climate regulation via oceanic phytoplankton sulphur emissions, Nature, 480(7375), 51– 56, doi:10.1038/nature10580.
Crossref CAS ADS PubMed Web of Science®Google Scholar
Tanimoto, H., S. Kameyama, T. Iwata, S. Inomata and Y. Omori (2014), Measurement of air‐sea exchange of dimethyl sulfide and acetone by PTR‐MS coupled with gradient flux technique, Environ. Sci. Technol., 48, 526– 533.
Crossref CAS ADS PubMed Web of Science®Google Scholar
Wanninkhof, R. (1992), Relationship between wind‐speed and gas‐exchange over the ocean, J. Geophys. Res., 97, 7373– 7382, doi:10.1029/92JC00188.
Wiley Online Library ADS Web of Science®Google Scholar
Watts, S. F. (2000), The mass budgets of carbonyl sulfide, dimethyl sulfide, carbon disulfide and hydrogen sulfide, Atmos. Environ., 34(5), 761– 779, doi:10.1016/S1352‐2310(99)00342‐8.
Crossref CAS ADS Web of Science®Google Scholar
Wong, C. S., S. E. Wong, W. A. Richardson, G. E. Smith, M. D. Arychuk, and J. S. Page (2005), Temporal and spatial distribution of dimethylsulfide in the subarctic northeast Pacific Ocean: A high‐nutrient–low‐chlorophyll region, Tellus, 578, 317– 331.
Google Scholar
Wong, P. K., and Y. H. Wang (1997), Determination of the Henry's law constant for dimethyl sulfide in seawater, Chemosphere, 35, 535– 544, doi:10.1016/S0045‐6535(97)00118‐5.
Crossref CAS ADS Web of Science®Google Scholar
Woodhouse, M. T., K. S. Carslaw, G. W. Mann, S. M. Vallina, M. Vogt, P. R. Halloran, and O. Boucher (2010), Low sensitivity of cloud condensation nuclei to changes in the sea‐air flux of dimethyl‐sulphide, Atmos. Chem. Phys., 10(16), 7545– 7559, doi:10.5194/acp‐10‐7545‐2010.
Crossref CAS ADS Web of Science®Google Scholar
Woodhouse, M. T., G. W. Mann, K. S. Carslaw, and O. Boucher (2013), Sensitivity of cloud condensation nuclei to regional changes in dimethyl‐sulphide emission, Atmos. Chem. Phys., 13, 2723– 2733, doi:10.5194/acp‐13‐2723‐2013.
Crossref ADS Web of Science®Google Scholar
Woolf, D. K. (1997), Bubbles and their role in gas exchange, in The Sea Surface and Global Change, edited by R. Duce and P. Liss, pp. 173– 205, Cambridge Univ. Press, New York, doi:10.1017/CBO9780511525025.007.
Crossref Google Scholar
Wurl, O., E. Wurl, L. Miller, K. Johnson, and S. Vagle (2011), Formation and global distribution of sea‐surface microlayers, Biogeosciences, 8(1), 121– 135, doi:10.5194/bg‐8‐121‐2011.
Crossref CAS ADS Web of Science®Google Scholar
Yang, G. P., M. Levasseur, S. Michaud, and M. Scarratt (2005), Biogeochemistry of dimethylsulfide (DMS) and dimethylsulfoniopropionate (DMSP) in the surface microlayer and subsurface water of the western North Atlantic during spring, Mar. Chem., 96(3–4), 315– 329, doi:10.1016/j.marchem.2005.03.003.
Crossref CAS Web of Science®Google Scholar
Yang, G. P., M. Levasseur, S. Michaud, A. Merzouk, M. Lizotte, and M. Scarratt (2009), Distribution of dimethylsulfide and dimethylsulfoniopropionate and its relation with phytoneuston in the surface microlayer of the western North Atlantic during summer, Biogeochemistry, 94(3), 243– 254, doi:10.1007/s10533‐009‐9323‐y.
Crossref CAS Web of Science®Google Scholar
Yang, M., B. W. Blomquist, and B. J. Huebert (2009), Constraining the concentration of the hydroxyl radical in a stratocumulus‐topped marine boundary layer from sea‐to‐air eddy covariance flux measurements of dimethylsulfide, Atmos. Chem. Phys., 9, 9225– 9236, doi:10.5194/acp‐9‐9225‐2009.
Crossref CAS ADS Web of Science®Google Scholar
Yang, M., B. W. Blomquist, C. W. Fairall, S. D. Archer, and B. J. Huebert (2011), Air‐sea exchange of dimethylsulfide in the Southern Ocean: Measurements from so GasEx compared to temperate and tropical regions, J. Geophys. Res., 116, C00F05, doi:10.1029/2010JC006526.
Wiley Online Library CAS ADS Web of Science®Google Scholar
Yang, M., P. D. Nightingale, R. Beale, P. S. Liss, B. W. Blomquist, and C. W. Fairall (2013), Atmospheric deposition of methanol over the Atlantic Ocean, PNAS, 110, 20034– 20039, doi:10.1073/pnas.1317840110.
Crossref CAS ADS PubMed Web of Science®Google Scholar
Yvon, S. A., and E. S. Saltzman (1996), Atmospheric sulfur cycling in the tropical Pacific marine boundary layer (12°S, 135°W): A comparison of field data and model results: 2. Sulfur dioxide, J. Geophys. Res., 101(D3), 6911– 6918, doi:10.1029/2007JG000640/abstract.
Wiley Online Library CAS ADS Web of Science®Google Scholar
Zappa, C. J., W. R. McGillis, P. A. Raymond, J. B. Edson, E. J. Hintsa, H. J. Zemmelink, J. W. H. Dacey, and D. T. Ho (2007), Environmental turbulent mixing controls on air‐water gas exchange in marine and aquatic systems, Geophys. Res. Lett., 34, L10601, doi:10.1029/2006GL028790.
Wiley Online Library CAS ADS Web of Science®Google Scholar
Zemmelink, H. J., W. W. C. Gieskes, W. Klaassen, H. W. de Groot, H. J. W. de Baar, J. W. H. Dacey, E. J. Hintsa, and W. R. McGillis (2002), Simultaneous use of relaxed eddy accumulation and gradient flux techniques for the measurement of sea‐to‐air exchange of dimethyl sulphide, Atmos. Environ., 36, 5709– 5717, doi:10.1016/S1352‐2310(02)00577‐0.
Crossref CAS ADS Web of Science®Google Scholar
Zemmelink, H. J., J. W. H. Dacey, E. J. Hintsa, W. R. McGillis, W. W. C. Gieskes, W. Klaassen, H. W. de Groot, and H. J. W. de Baar (2004), Fluxed and gas transfer rates of the biogenic trace gas DMS derived from atmospheric gradients, J. Geophys. Res., 109, C08S10, doi:10.1029/2003JC001795.
Wiley Online Library ADS Web of Science®Google Scholar
Zemmelink, H. J., L. Houghton, J. W. H. Dacey, A. P. Worby, and P. S. Liss (2005), Emission of dimethylsulfide from Weddell Sea leads, Geophys. Res. Lett., 32, L23610, doi:10.1029/2005GL024242.
Wiley Online Library ADS Web of Science®Google Scholar

Citing Literature

Volume122, Issue13

16 July 2017

Pages 7216-7231

Sea‐to‐air flux of dimethyl sulfide in the South and North Pacific Ocean as measured by proton transfer reaction‐mass spectrometry coupled with the gradient flux technique

Abstract

1 Introduction