Four iodocarbons, chloroiodomethane (CH2ClI), diiodomethane (CH2I2), methyl iodide (CH3I), and ethyl iodide (C2H5I), were measured with an automated preconcentration gas chromatography–mass spectrometry system at two remote marine sites, Hateruma Island (24.05°N, 123.8°E) in the East China Sea and Cape Ochiishi (43.15°N, 145.5°E) on the eastern coast of Hokkaido, for 17 months, giving the first full-year high-frequency data sets for all these compounds. CH2ClI and CH2I2, which are highly photolyzed, showed remarkable diurnal variation in all seasons, with lower concentrations in the daytime, whereas CH3I and C2H5I showed no significant diurnal changes. At Cape Ochiishi, all of the iodocarbons showed clear seasonal variations and were highest in summer and autumn, which are characterized by algal blooms in the adjacent ocean. At Hateruma Island, which is surrounded by subtropical oligotrophic waters, the seasonal variations were not significant, but C2H5I and CH2I2 showed lower mixing ratios in summer. The nighttime mixing ratio of CH2ClI was strongly and positively correlated with wind speed throughout the observation period at Hateruma Island, suggesting the ubiquitous presence of CH2ClI sources, probably nonbiogenic, in the subtropical ocean. CH3I and C2H5I mixing ratios occasionally increased at Hateruma Island concurrently with winter and spring pollution events from the Asian continent, indicating possible continental anthropogenic sources or emissions from seawater off the continental coast. A rise in the CH3I baseline mixing ratio was also observed at Hateruma Island, when the air mass origin became more southerly, compared with the earlier, easterly source area, suggesting higher concentrations in equatorial regions.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
 The iodocarbons are lost through photolysis in the atmosphere, producing iodine atoms that can decompose tropospheric ozone catalytically, and may contribute to new particle formation [McFiggans et al., 2000; Jimenez et al., 2003; Pechtl et al., 2006]. Therefore, iodocarbons can affect global radiative forcing [O'Dowd et al., 2002]. Their photolysis reactivity is in the order CH3I < C2H5I < C3H7I < CH2ClI < CH2BrI < CH2I2, with CH2I2 having a lifetime of several minutes (in the daytime) and CH3I having one of several days [Roehl et al., 1997; Mössinger et al., 1998]. The most reactive compounds, CH2ClI, CH2BrI, and CH2I2, have been proposed as the cause of new particle formation observed along the coasts of Europe, although recent studies have shown that molecular iodine, rather than iodocarbons, released from exposed seaweed may be the more important source of iodine atoms in coastal areas [McFiggans et al., 2004; Palmer et al., 2005]. We have little knowledge of the importance of iodocarbons as an iodine source over the open ocean, mostly because of sparse measurement data there.
 Clearly, more extensive studies are needed to evaluate the importance of iodocarbons in the marine atmosphere, including intensive observation of their atmospheric mixing ratios in the marine boundary layer, which should indicate the balance between the sea-to-air flux and the atmospheric removal rate of each iodocarbon. Here, we report hourly measurements of CH2ClI, CH2I2, C2H5I, and CH3I in the atmosphere at Hateruma Island (24.05°N, 123.8°E) in the East China Sea and at Cape Ochiishi (43.15°N, 145.5°E) on the eastern coast of Hokkaido, and discuss their sources and sinks, focusing particularly on the subtropical marine data from Hateruma Island.
2. Experimental Methods
 Atmospheric CH2ClI, CH2I2, CH3I, and C2H5I were measured hourly with an automated preconcentration gas chromatography–mass spectrometry (GC–MS, Agilent 6890/5973) system at ground stations on Hateruma Island and at Cape Ochiishi as a part of the National Institute for Environmental Studies (NIES) halocarbon monitoring project. Hateruma Island is a small island in the pathway of the Kuroshio Current in the East China Sea. It is 250 km east of Taiwan and 500 km southwest of Okinawa Island. The prevailing winds are southeasterly in summer and northwesterly in winter. The Cape Ochiishi monitoring station is situated on a cliff at the southern tip of Cape Ochiishi, which projects southward from the eastern coast of Hokkaido into the Oyashio Current of the western North Pacific. Prevailing winds at Cape Ochiishi are also southeasterly in summer and northwesterly in winter. Coastal areas near Cape Ochiishi are rich in seaweed and plankton, while Hateruma Island is surrounded by fringing coral reefs with poor seaweed growth.
 The NIES monitoring project collects air samples from the top of a tower (36.5 m above ground and 46.5 m above sea level at Hateruma Island, and 51 m above ground and 96 m above sea level at Cape Ochiishi) and analyzes more than 30 species of natural and anthropogenic halocarbons by GC–MS (selected ion monitoring) once each hour. After every five air analyses, a gravimetrically prepared standard gas (including most of the target compounds but not CH2ClI, CH2I2, or C2H5I; Taiyo Nissan Co., Ltd.) is analyzed by the same procedure for quantification of the results. Details of the sampling and analytical methods have been published elsewhere [Enomoto et al., 2005; Yokouchi et al., 2006].
 Selected ions for quantification were m/z (mass to charge ratio) 268 for CH2I2, m/z 176 for CH2ClI, m/z 142 for CH3I, and m/z 156 for C2H5I. The three iodocarbons other than CH3I were quantified on the basis of their sensitivity relative to tetrachloroethylene (C2Cl4) (monitored ion, m/z 166), which, like CH3I, is a component of the working standard. The sensitivities relative to C2Cl4 were determined by analysis of a vaporized liquid standard: Methanol solution (0.5 μL) containing known concentration of (160–330 ng/μL) of halocarbons was injected into a 5 mL glass vial whose contents were flushed with helium purge gas into the preconcentration trap. The reagents (purity > 97%) were purchased from Tokyo Kasei Kogyo Co., Ltd.
 To evaluate stability of the relative sensitivities throughout the observation period, we calculated the relative standard deviation of the ratios of three components of the standard gases, C2Cl4, CH3I, and CH2Br2 (monitored ion, m/z 174) throughout a period of 17 months in the Hateruma Island observation data. The relative standard deviation of the sensitivity ratio of CH2Br2/C2Cl4 was about 4%, and that of CH3I/C2Cl4 was 6.5%. The retention time and monitored ion (m/z) of CH2Br2 were closer, compared with CH3I, to those of C2Cl4, which can explain the somewhat smaller deviation of the CH2Br2/C2Cl4 sensitivity ratio compared with that of CH3I/C2Cl4. We can predict similar deviations of the sensitivity ratios of C2H5I/C2Cl4 and CH2ClI/C2Cl4 to that of CH2Br2/C2Cl4, because the retention times of C2H5I and CH2ClI (16.6 min and 18.4 min, respectively) are similar to that of C2Cl4, and their monitored ions (m/z) (176 and 156) are close to that of C2Cl4. Taking into account the large monitored ion (m/z) and long retention time of CH2I2, we would expect a little larger variability of the sensitivity ratio CH2I2/C2Cl4 in this study. Detection limits were approximately 0.01 ppt for CH2ClI and CH3I, 0.005–0.01 ppt for CH2I2, and 0.02 ppt for C2H5I.
3. Results and Discussion
 During the period between August 2008 and January 2010, more than 9000 and 7000 data sets of the four targeted iodocarbons in the atmosphere were obtained at Hateruma Island and Cape Ochiishi, respectively (Figures 1 and 2). These are the first full-year high-frequency data sets for all these compounds. Previously, only the variation of CH3I had been investigated year round, by semimonthly sampling and measurement [Yokouchi et al., 2001, 2008], and the variation of CH2ClI had been measured weekly in the Arctic [Yokouchi et al., 1996]. The data for 15 September to 28 October 2009 are shown in detail, together with meteorological data (solar radiation, wind direction, and wind speed) during the same period, in Figure 3. The means, medians, and ranges of the data are listed in Table 1. The correlation coefficients of the four iodocarbons for the complete data set and for selected periods are also listed in Table 2 (for Hateruma Island) and Table 3 (for Cape Ochiishi).
Table 1. Atmospheric Mixing Ratios of CH2ClI, CH2I2, CH3I, and C2H5I Observed at Hateruma Island and Cape Ochiishia
Observation period for Hateruma Island is August 2008 to January 2010; observation periods for Cape Ochiishi are September 2008 to December 2008 and March 2009 to January 2010; nd denotes not detectable.
nd to 0.99
nd to 2.1
nd to 0.07
nd to 0.4
0.43 to 5.2
0.12 to 5.2
nd to 0.91
nd to 0.78
Table 2. Correlation Coefficient Among the Iodocarbons Observed at Hateruma Island
For the whole data between August 2008 and January 2010.
For the data between 15 September and 28 October 2009.
 Different patterns of seasonal and diurnal variation are apparent between two groups, CH2ClI and CH2I2, and CH3I and C2H5I (Figures 1 and 2). Archer et al.  reported similar seasonal variation differences in the seawater concentrations between these two groups in the western English Channel. It is likely that different production and removal processes control the concentrations of the compounds in each group. Therefore, we grouped the iodocarbons into two categories, (1) CH2ClI and CH2I2, and (2) CH3I and C2H5I, to examine the variations in their concentrations.
3.1. Atmospheric Concentrations of CH2ClI and CH2I2 at Hateruma Island and Cape Ochiishi
 The mean mixing ratios of CH2ClI—0.12 pptv at Hateruma Island and 0.18 pptv at Cape Ochiishi—are within the range of previously reported values. They are close to those from Mace Head (Ireland) of 0.11 pptv in spring [Carpenter et al., 1999] and 0.16 pptv in summer [Carpenter et al., 2000], as well as to that from Christmas Island (Kiribati) of 0.10 pptv in March [Varner et al., 2008]. They are lower than mean values from the Atlantic and Southern oceans (0.32 pptv in September–October) [Chuck et al., 2005], from Appledore Island (Maine, USA; 0.68 pptv in summer), [Varner et al., 2008], and from the western North Pacific Ocean (0.27 pptv in spring) [Kurihara et al., 2010], whereas they are much higher than those from Oahu Island (Hawaii; 0.04 pptv in September) [Varner et al., 2008] and Alert (Arctic; 0.01 pptv year round) [Yokouchi et al., 1996].
 However, the mean mixing ratios of CH2I2—0.008 pptv at Hateruma Island and 0.03 pptv at Cape Ochiishi—are at the low end of previously reported data, which are mostly from coastal regions: <0.08 to 1.02 pptv from Spitzbergen [Schall and Heumann, 1993], <0.02 to 0.36 pptv and <0.02 to 0.46 pptv from Mace Head [Carpenter et al., 1999; Carpenter, 2003], 0.11–19.8 pptv from Lilia (French Atlantic Coast) [Peters et al., 2005], 0.3–3.1 pptv from Dagebüll (German North Sea coast) [Peters et al., 2005], 0.01–0.07 pptv from Roscoff (northwest coast of France) [Jones et al., 2009], and <0.009 to 0.02 pptv from the temperate and tropical Atlantic Ocean [Jones et al., 2010].
 Mixing ratios of both CH2ClI and CH2I2 showed high temporal variability, which we attributed mostly to diurnal variation (Figure 3). They were much higher at night than in the daytime all days in the record, and CH2I2 in particular was rarely detected at midday. This finding can be explained by their short atmospheric lifetimes (several hours for CH2ClI and several minutes for CH2I2) during the day due to photolysis [e.g., Rattigan et al., 1997; Mössinger et al., 1998; Roehl et al., 1997]. Varner et al.  observed a very similar diurnal variation of CH2ClI in the coastal North Atlantic region and at two sites in the remote Pacific, Christmas Island and Oahu Island. Other researchers [Carpenter et al., 1999; Peters et al., 2005] have reported that tidal change is a main cause of the diurnal variation of short-lived dihalomethanes in the atmosphere at coastal sites. However, we detected no evidence of tidal dependence in either the Hateruma Island or Cape Ochiishi data set in this study, perhaps because the waters around Hateruma Island are not rich in algae, and Cape Ochiishi station is situated on a cliff (∼45 m high), and macroalgae in its vicinity are growing in the deeper water than those in the previous studies.
 Seasonal variations of the CH2ClI and CH2I2 mixing ratios differed considerably between Cape Ochiishi and Hateruma Island. At Cape Ochiishi, both compounds were much more abundant in summer/autumn than in winter/spring, and their concentrations were well correlated (e.g., R2 = 0.55, 0.48, 0.55, and 0.54, in August, September, October, and November 2009, respectively). Considering that macroalgae are abundant in the marine waters surrounding Cape Ochiishi, and algal blooms develop in summer and autumn, algae are likely the main source of CH2ClI and CH2I2 there. At Hateruma Island, in contrast, CH2ClI and CH2I2 mixing ratios were highest in autumn and lowest in summer. Moreover, CH2ClI increased significantly during and around typhoon events (29–30 September 2008; 7–11 August and 6–7 October 2009). The median monthly mixing ratio of CH2I2 at Hateruma Island was lower than that at Cape Ochiishi throughout the year, although the difference was small in late winter and early spring. This result reflects Hateruma Island's location in subtropical oligotrophic waters without warm-season algal blooms, as well as the greater degree of CH2I2 photolysis at that site compared to Cape Ochiishi.
 Although the absence of a summer increase of CH2ClI at Hateruma Island is consistent with low bioactivity in the surrounding ocean, the median monthly mixing ratio of CH2ClI at Hateruma Island often exceeded that at Cape Ochiishi in autumn and winter. As nonbiogenic sources of iodocarbons, reactions between photochemically produced methyl radicals and iodine atoms in seawater [Moore and Zafiriou, 1994; Happell and Wallace, 1996], and reactions between marine dissolved organic matter (DOM), dissolved iodide, and ozone [Martino et al., 2009] have been proposed for CH3I, and for CH2I2, CHClI2 and CHI3, respectively. As DOM and dissolved iodide are ubiquitous at the sea surface, such reactions might also provide a ubiquitous source of CH2ClI to the marine atmosphere, being the major source of the CH2ClI observed at Hateruma Island in this study.
 An outstanding feature of the CH2ClI mixing ratio in the atmosphere at Hateruma Island, other than the diurnal variation related to solar radiation, was its close correlation with wind speed. When we examined the relationships between CH2ClI and wind speed data in year round (14 August 2008 to 31 January 2010), autumn (26 September to 31 October 2008), and winter/spring (7 February to 29 March 2009) data sets, further categorizing them into whole day, nighttime (0:00–5:00 local time, LT), and daytime (10:00–17:00 LT) data (Figures 4a–4c), we found that the highest correlation (R2 = 0.58) was for autumn nighttime data. Varner et al.  also reported a good correlation between wind speed and CH2ClI mixing ratios in summer data from Appledore Island, except when wind speed was high (>9 m s−1), which occurred mostly during the daytime.
 The sea-to-air flux of a substance (F) is expressed as
where Ca and Cw are the concentrations of CH2ClI in atmosphere and seawater, respectively, H is the Henry's law constant, and Kw is the air-sea exchange coefficient. Since the air-sea exchange coefficient (Kw) is approximately proportional to the square of the wind speed [e.g., Wanninkhof, 1992], a significant correlation between the atmospheric mixing ratio and wind speed suggests that the sea-to-air flux is well reflected in the atmospheric mixing ratio, and also that the air-sea concentration difference (Cw − Ca/H) is fairly constant throughout the observation period. In fact, the daytime lifetime of CH2ClI is so short (a few hours) that the nighttime mixing ratio should reflect mostly the buildup of oceanic emissions after sunset, thus minimizing the influence of long-range transport, which could obscure the nearby emission strength. Therefore, nighttime atmospheric CH2ClI may be a good tracer of its flux, and the good year-round correlation of nighttime atmospheric CH2ClI with wind speed (R2 = 0.37) implies that the air-sea concentration difference of CH2ClI is likely fairly constant around Hateruma Island. Under highly supersaturated conditions (Cw > > Ca/H), the seawater concentration (Cw) can be considered the air-sea concentration difference. A steady concentration of CH2ClI in the surface seawater would support the presence of ubiquitous nonbiotic sources, as described above.
 We estimated the CH2ClI flux in the Hateruma Island area by two methods, both based on a simple box model, using data from 23 to 30 September 2009, when wind speed was typically 6–10 m s−1: (1) The average accumulation rate of atmospheric CH2ClI in the nighttime (1900 to 0200 LT) was about 0.07 ± 0.05 pptv h−1, and the average boundary layer height at 0200 LT was about 500 m (calculated using METEX program developed by Zeng et al. ). If we assumed full mixing in the boundary layer, then the sea-to-air flux of CH2ClI was 28 ± 20 ng m−2 h−1, or 3.8 ± 2.7 nmol m−2 d−1. (2) The average midday mixing ratio of CH2ClI was about 0.05 pptv, and the boundary layer height in the daytime was about 750 m (650–850 m). If we assumed its atmospheric lifetime due to photolysis to be in the range of 3–10 h, then the calculated flux was 30–100 ng m−2 h−1 (4–13 nmol m−2 d−1). These estimates are comparable to the reported mean CH2ClI fluxes of 5.5 nmol m−2 d−1 in the Atlantic Ocean (temperate and tropical regions) [Chuck et al., 2005], 19 nmol m−2 d−1 around Christmas Island, and 5.8 nmol m−2 d−1 around the Hawaiian Islands [Varner et al., 2008]. They are also close to the flux estimate based on the seawater measurements near Hateruma Island in May 2009, 3.5 nmol m−2 d−1 (unpublished). It is thus likely that CH2ClI is emitted from the tropical/subtropical ocean with a flux of several to over 10 nmol m−2 d−1.
 The correlation between the nighttime mixing ratio of CH2I2 and wind speed at Hateruma Island was less significant, even though this compound has a shorter photochemical lifetime than CH2ClI, and the atmospheric mixing ratio would thus be expected to reflect its sea-to-air flux. This result therefore suggests that the seawater concentration of CH2I2 is likely to be more variable either temporally or spatially than that of CH2ClI, probably because of its rapid photodecomposition in surface seawater [Martino et al., 2005]. For a preliminary estimate, we calculated the flux of CH2I2 using its rate of increase during the night of 5 October 2009 (0.0024 pptv h−1; from 0.005 pptv at 1900 LT to 0.034 pptv at 0700 LT the next morning), obtaining a flux of 14 ng m−2 h−1 (1.3 nmol m−2 d−1). This value is very similar to the flux estimated from seawater measurements in this region in May 2009 (∼1.3 nmol m−2 d−1, unpublished), and in the same order of magnitude as the average flux (3.5 nmol m−2 d−1) measured in the western English Channel [Archer et al., 2007].
3.2. CH3I and C2H5I
 The mean mixing ratio of CH3I, 1.2 pptv, at Hateruma Island was a little higher than the mean, 0.77 pptv, of semimonthly measurements made at the same site during 2003–2005, whereas the mean mixing ratio at Cape Ochiishi (0.81 pptv) was similar to the 2003–2005 mean (0.72 pptv) [Yokouchi et al., 2008]. Both values are in the range of typical background mixing ratios of CH3I (0.2–2 pptv) (Table 1) [Yokouchi et al., 2008], but they are much lower than those measured at some European coastal sites: 4.7–23.5 pptv at Dagebüll at the German North Sea Coast, and 7.6–1830 pptv at Lilia at the French Atlantic Coast of Brittany [Peters et al., 2005]. There are fewer measurements reported for C2H5I than CH3I. The mean mixing ratios of C2H5I, 0.15 pptv at Hateruma Island and 0.08 pptv at Cape Ochiishi, are close to those from Asian seas, <0.03 to 0.31 pptv [Yokouchi et al., 1997], and from Mace Head, Ireland, <0.02 to 0.21 pptv [Carpenter et al., 1999], but lower than those from Lilia, 2.22–96.9 pptv [Peters et al., 2005], and Roscoff, 0.21–0.82 pptv [Jones et al., 2009].
 CH3I and C2H5I measured at Hateruma Island and Cape Ochiishi showed little diurnal variation, in contrast to the remarkable diurnal variation of CH2ClI and CH2I2 described in section 3.1. This absence of diurnal variation is consistent with their longer atmospheric lifetimes, days for CH3I, and about a day for C2H5I [Roehl et al., 1997; Rattigan et al., 1997]. As a result, the daytime decrease in their mixing ratios is smaller, although they are also lost through photolysis. Short-term variations of CH3I and C2H5I were very similar with each other at both Hateruma Island and Cape Ochiishi, strongly suggesting that they have common sources and sinks [Yokouchi et al., 1997; Carpenter et al., 1999]. Because CH3I is probably produced in the open ocean by a light-dependent production pathway that is not directly dependent on biological activity [Richter and Wallace, 2004], the observed high correlation between CH3I and C2H5I suggests that similar nonbiogenic emissions are important C2H5I sources. Seasonal variations of CH3I and C2H5I, however, were quite different between the two sites. At Cape Ochiishi, the two compounds had higher mixing ratios in summer and autumn, showing a sinusoidal pattern similar to that observed for CH3I at the same site as well as at some other midlatitude Northern Hemisphere sites during 2003–2005 [Yokouchi et al., 2008]. This result is consistent with the positive correlation of the CH3I concentration in seawater with surface seawater temperature (SST) [Archer et al., 2007]. At Hateruma Island, however, the summer increase of atmospheric CH3I was less prominent, as Yokouchi et al.  reported previously, and C2H5I showed a substantial decrease in summer.
 The most remarkable feature of the variations of CH3I and C2H5I (Figure 1) is the repeated sharp increases followed by gentle declines on a time scale of hours ∼ days in winter and spring, when the island is affected by the outflow of atmospheric pollutants from the southern Asian continent. We compared the variations of CH3I and C2H5I with those of a typical anthropogenic compound, HCFC-22 (difluorochloromethane), which has been widely used as a replacement cryogen for CFC-12 (difluorodichloromethane), during February–April 2009 (Figure 5). CH3I and C2H5I peaks coincided with HCFC-22 peaks, suggesting that the two iodocarbon compounds, like HCFC-22, have anthropogenic sources in the Asian continent. Because the iodocarbon peaks are often broader than the corresponding HCFC-22 peaks, for example, on 20–21 February, 6–8 March, and 25–26 April (Figure 5), their presumed anthropogenic sources might be more widely distributed than the HCFC-22 sources. Other possible sources of the enhanced CH3I and C2H5I concentrations include emissions from terrestrial ecosystems and from seawater in the immediate vicinity of the continental coast, both of which would result in an increase in the mixing ratios observed at Hateruma Island when air masses come from the Asian continent.
 In the observations of CH3I and C2H5I at Hateruma Island, we also noted some differences in their baseline variations. First, the C2H5I baseline was lower in summer, whereas the CH3I baseline was even higher in August–September 2008 than in other months of the study period (Figure 1). Second, the baseline CH3I mixing ratio sometimes increased when the baseline values of C2H5I and the anthropogenic gas HCFC-22 were stable at their lowest level or decreasing (e.g., Figure 5, the periods marked with arrows). Back-trajectory analyses showed that these baseline increases of CH3I were observed concurrently with the air mass origin becoming more southerly, compared with the earlier, easterly source area (Figure 6). This finding suggests that the atmospheric CH3I distribution may have a latitudinal gradient, with higher values at lower latitudes, where higher SST and solar radiation would cause higher emissions of CH3I. This is consistent with the finding that CH3I in surface seawater was well correlated with SST in the subtropical waters [Ooki et al., 2010]. In the case of C2H5I, its higher photolysis rate might have canceled out the higher emissions at lower latitudes, resulting in a lack of a latitudinal gradient in its atmospheric mixing ratio, although more studies are necessary to confirm this interpretation.
 This study reports the first precise, full-year data sets of four iodocarbons in the atmosphere at two remote marine sites, Hateruma Island and Cape Ochiishi, providing new insight into their possible sources and sinks. Diurnal variation of CH2ClI and CH2I2 was very prominent, and the highly photolabile CH2I2 was rarely observed in the atmosphere at midday. Nighttime atmospheric CH2ClI was strongly correlated with wind speed at Hateruma Island in the subtropical ocean, suggesting the ubiquitous presence of CH2ClI at a fairly constant concentration. This result is consistent with nonbiotic, rather than biogenic, production of this compound at least in subtropical oceans. Using box model calculations, we roughly estimated the hourly sea-to-air flux of CH2ClI to be a few tens of nanograms per square meter. However, the mixing ratios of the iodocarbons at Cape Ochiishi were greatly enhanced during the algal bloom season there, showing the importance of biogenic sources in the coastal Oyashio Current waters.
 CH3I and C2H5I showed very similar variations at both sites, suggesting common sources and sinks. Although they also showed a summer/autumn increase at Cape Ochiishi, their day-to-day variations were different from those of CH2ClI and CH2I2. Their variations were consistent with the previously proposed photochemical production of CH3I in seawater, although biogenic sources were not excluded. In winter and spring at Hateruma Island, CH3I and C2H5I mixing ratios increased concurrently with pollution events from the Asian continent, indicating possible anthropogenic sources, or emission from seawater in the immediate vicinity of the continental coast. Anthropogenic emissions of CH3I and C2H5I would cause some perturbation of natural iodine chemistry in the marine atmosphere. Variation of the baseline mixing ratio of CH3I coupled with air mass trajectories to Hateruma Island suggest that emissions of this compound are higher at lower latitudes in the subtropical ocean.
 Observations at Hateruma Island and Cape Ochiishi were supported by the Global Environment Fund (Ministry of the Environment of Japan). We thank Nobukazu Oda and Fujio Shimano of the Global Environmental Forum Foundation (GEFF) for their great help in the maintenance of the analytical systems and stations.