Trace gas measurements over the northwest Pacific during the 2002 IOC cruise



[1] The R/V Melville cruised from Osaka (Japan) on 1 May and reached Hawaii on 5 June on a project for the Intergovernmental Oceanographic Commission (IOC) in 2002. During this cruise, the concentrations of atmospheric trace gases (O3, CO, DMS, hydrocarbons, and halocarbons) were measured. Air at high latitudes and low latitudes exhibited starkly different characteristics regarding their chemical composition. The concentrations of anthropogenic species clearly decreased from high latitude to low latitude. On the other hand, biogenic species such as DMS and alkenes were highly abundant at lower latitudes. Backward air trajectories show that the northwestern continental air mass was dominant at higher latitudes and the eastern marine air mass was dominant at lower latitudes. However, the long-range transport of pollutants to clean regions near Hawaii was also observed. The ratios of ethane to CO decreased from high latitude to low latitude. On the basis of a VOC ratio analysis, the benzene concentration is relatively higher at low latitudes. DMS concentrations and wind speed at low latitudes have good correlation. This indicates that at low latitudes, the DMS concentration at the ocean surface is roughly uniform.

1. Introduction

[2] East Asia, particularly China, is one of the most interesting regions with regard to air pollution because of a rapid increase in economic development. In this region, regional as well as local air pollution will become a serious situation in the near future [Akimoto, 2003]. The pollutants emitted in east Asia are primarily transported to the northwest Pacific by the westerly winds [Jaffe et al., 1997; Kato et al., 2001, 2004]. During some sporadic events, the polluted air is transported to the west coast of the United States [Jaffe et al., 2005]. This type of long-range transport of pollutants has been recently referred to as “inter-continental transport.” From the viewpoint of hemispheric or global atmospheric environment, regional pollution as well as international pollution should be monitored. Observations from islands in the Pacific Ocean would provide important information about the transport of polluted air from the east Asian region [Kato et al., 2001, 2004]. However, these islands are geographically limited. The lack of available observation locations in the Pacific can be overcome by performing measurements on ships. Some atmospheric trace gas measurements have been performed on ships in the northwest Pacific [Saito et al., 2002; Laurier et al., 2003].

[3] Further, it is noteworthy to observe the atmospheric trace species over the ocean. The interaction between the atmosphere and the ocean is a frequently debated issue. For example, iron input from land could increase the biogenic activity in the subarctic northwestern Pacific region [Tsuda et al., 2003]. This process would increase the vertical transport of carbon from the sea surface to deep sea and would contribute to suppressing an increase in atmospheric CO2. Further, dimethylsulfide (DMS) is produced by specific planktons in the ocean and eventually emitted from the sea surface to the atmosphere [Lovelock et al., 1972; Aranami and Tsunogai, 2004]. DMS is oxidized to produce sulfur aerosol; this could change the cloud properties. This increase in DMS emission could affect the solar radiation budget over the Earth. Since the DMS concentration in the ocean is variable geographically and seasonally, the collection of observation data at various places and seasons is valuable. Other trace species that are emitted from the ocean influence the reactions in the oceanic atmosphere. In this paper, the results of the atmospheric trace species observed during the Intergovernmental Oceanographic Commission (IOC) cruise conducted in 2002 are discussed with respect to long-range transport of pollutants over the Pacific and emission of the oceanic biogenic species to the atmosphere.

2. Experimental Method

2.1. Sampling Locations

[4] The IOC 2002 research cruise was conducted on the R/V Melville operated by the Scripps Institution of Oceanography, USA. The cruise track is shown in Figure 1. Each point plotted in this figure corresponds to the sampling location of canisters used for volatile organic compounds (VOC) measurements (one sample per day). The ship departed Osaka, Japan, on 1 May and reached to near Tokyo, Japan, on 8 May; it arrived at the highest latitude point (50°N, 167°E) on 15 May. The ship then headed south along the 170°E line up to 24°N and turned east to arrive at Hawaii on 5 Jun.

Figure 1.

Cruise track of IOC 2002. The sampling locations for VOC analysis, which were taken once a day, are plotted. The categorized areas (JAPAN, NORTH, SOUTH) are shown.

2.2. Ozone, Carbon Monoxide, and VOC

[5] Ozone (O3) and carbon monoxide (CO) in the atmosphere were continuously measured with commercial instruments provided by Thermo Environmental Instruments (TEI) Inc. The O3 analyzer (TEI model 49C, UV absorption method) was calibrated by a standard O3 calibrator (TEI model 49CPS) before and after the cruise. The precision of the instruments was about 1 ppbv. It is known that the CO analyzer (TEI model 48C, NDIR absorption method) is strongly influenced by water vapor concentration, and the zero level drift (background level) of this analyzer tends to increase gradually. To remove the influence of zero drift, zero gas generated by a heated Pt catalyst was periodically measured. Zero gas as measured for 15 min followed by ambient air measurements for 45 min. To achieve the same water concentration as that of ambient air, the zero gas was generated only by the Pt catalyst. The concentration of CO was calibrated by using standard gas (1.8 ppm, Nippon Sanso, Japan) before and after the cruise. The precision of the CO analyzer was 6.0 ppb deduced from the standard deviation during zero air measurement. The data obtained from these instruments were logged in a PC every minute.

[6] During the cruise, ambient air was sampled for the VOC analyses. The air was compressed by PFA bellows pump (Iwaki, Japan) into a canister (Restek) whose inner surface was coated with fused silica in order to stabilize the trace components for longer storage periods. One canister sample was prepared everyday and a total of 35 samples were collected. The sample canisters were analyzed after the cruise by GC-FID (HP6890) and GC-MS (HP5973) at the laboratory of Tokyo Metropolitan University [Kato et al., 2004, 2001]. 500 mL of the sampled air was concentrated by a three-stage concentrator (Entech 7000). The first stage comprised glass beads trapped at −160°C and desorbed at 20°C. The second stage comprised tenax trapped at −60°C and desorbed at 180°C. The third stage comprised cryofocus with a silcosteel tube. It was trapped at −165°C and heated rapidly to 100°C. Then, the concentrated VOC was injected into the GC. The flow of the GC-FID carrier gas (He) was maintained at 4 mL min−1. The temperature for the GC was maintained at −50°C for 8 min. Then, it was heated to 40°C at the rate of 5°C min−1. Subsequently, it was heated to 150°C at the rate of 15°C min−1. The GC column was an HP-1 column (length: 60 m, diameter: 320 μm, and the film thickness: 1 μm). The identification and quantification were performed by using a mixed standard containing 56 gas species (Enviro-Mat, Matheson). Selected hydrocarbons observed in the oceanic atmosphere are listed in Table 1. Some halocarbons, listed in Table 1, were analyzed using the GC-MS. The GC column and oven temperature control were the same as those used in the GC-FID measurement, except the carrier gas flow. The flow of He was maintained constant at 1 mL min−1. The electron impact (EI) method was used for ionization. Identification and concentration calibration of the measured species were performed using the standard gas (TO-14, Takachiho, Japan). After checking the retention time in the scan mode, each halocarbon peak was monitored using the selected ion monitoring (SIM) mode. DMS was analyzed by both GC-FID and GC-MS and calibrated with the standard gas (Takachiho, Japan). Carbonyl sulfide (COS) was also analyzed by GC-MS in this cruise. However, the absolute concentration was not determined, as it monitored relative to the intensity of CFC-11 (CCl3F).

Table 1. Average Concentrations of VOCs, CO, and O3 for Each Areaa
SpeciesJAPAN (n = 5)bNORTH (n = 11)SOUTH (n = 16)
AverageStd. Dev.AverageStd. Dev.AverageStd. Dev.
  • a

    Units: pptv for VOC and ppbv for CO and O3.

  • b

    Very polluted air near Osaka and Tokyo in JAPAN area is excluded.

Anthropogenic VOC      
Biogenic VOC      
Long-life halocarbons (CFC)      
CO and O3      

3. Results and Discussions

3.1. Study Area

[7] The observed data was categorized into three areas for explanation, as shown in Figure 1. The “JAPAN” area was near Japan, and the air would be influenced by the pollutants emitted from Japan. The “NORTH” area corresponded to that above the 35°N latitude, while the “SOUTH” area corresponded to that below this latitude, including the final leg of the cruise to Hawaii. The borderline of the categories at 35°N is chosen because the path of the Kuroshio stream is almost along this latitude.

3.2. Backward Trajectories

[8] To reveal the relationship between air mass origin and the change in concentration of the measured trace gases, backward trajectories were calculated using the NOAA ARL HYSPLIT4 model ( The backward trajectories from the VOC sampling points for ten days are plotted in Figure 2. The altitude was 500 m at the beginning of the calculations. In the JAPAN and NORTH areas, the air flowed from the continent (northern Japan, Siberia, or Kamchatka). In some trajectories, the air from the continent traveled over the northwest Pacific for a long duration. In the SOUTH area, air flowed from the marine areas (central or eastern Pacific). However, in two cases, air originated from the Southeast Asian region, as indicated by the green lines. Near the borderline of the NORTH and SOUTH areas, the air mass traveled over the northwest Pacific; the characteristics of this air mass were intermediate of the continental and maritime ones.

Figure 2.

Backward trajectories from VOC sampling points. Air masses from Southeast Asia are shown with green lines.

3.3. CO and O3

[9] The CO and O3 concentrations are shown in Figure 3 with the area categories indicated (all CO and O3 data are listed in Table S1). Universal time (UT) was used for all data. In the JAPAN area, large fluctuations in the concentration were observed, particularly for O3. These fluctuations are attributed to the transportation of polluted air from urban areas by the movement of a low pressure system. The concentration of O3 also decreases due to the reaction of O3 with NO that is present in highly polluted air near cities. In the NORTH area, the concentration of CO and O3 was relatively higher, similar to the air at higher latitudes. A decrease in the concentrations of both CO and O3 from the NORTH area to the SOUTH area is evident. This latitudinal gradient is produced by polluted air in high latitudes and clean oceanic air in low latitudes. The average concentrations of CO and O3 for each region are shown in Table 1. These concentrations do not include the data corresponding to the significantly polluted area near Japan. The data obtained from the JAPAN/NORTH area and the SOUTH area differs significantly. In the case of O3, 35.5 ± 4.1 for JAPAN, 40.6 ± 4.8 for NORTH, and 13.0 ± 10.4 for SOUTH.

Figure 3.

Concentration variations of CO, O3, and ethane. Area categories are also shown. These concentrations clearly decreased from the NORTH area to the SOUTH area.

3.4. Volatile Organic Compound (VOC)

[10] VOC is the generic name for hydrocarbons, halocarbons, etc. The average observed concentrations and standard deviations of the VOCs for each area are summarized in Table 1. The variations of the selected VOC concentration are shown in Figure 4 (all VOC data are listed in Table S2).

Figure 4.

Concentration variations of (a) anthropogenic hydrocarbons and (b) relatively short lifetime halocarbons. They are high in the JAPAN/NORTH area and low in the SOUTH area. Very high concentrations around 8 May were influenced by the Tokyo bay area.

3.4.1. Hydrocarbons

[11] Hydrocarbons are plotted on a log scale in Figure 4a. These are mainly emitted by human activities and their concentrations were higher near Japan and lower over the open Pacific Ocean. Since hydrocarbons and CO are emitted from similar sources (urban areas), they exhibited similar changes in the concentration. Ethane is also plotted with CO in Figure 3; their changes in the concentration are very similar. At around 8 May, fairly high peaks of pentane, butane, and benzene were observed near Tokyo. Very high concentrations were not observed for all the hydrocarbons. On the basis of this difference, the polluted air was suspected to originate from petroleum chemical plants located along the bay area near Tokyo, rather than combustion sources. On 28 May, when the ship was near Hawaii, higher concentrations of CO, O3, and VOCs were observed. The backward trajectories, as shown in Figure 2, indicate that this high concentration was caused by the long-range transport of pollutants from the mid latitudes or from Southeast Asia.

3.4.2. Halocarbons

[12] Selected halocarbons are shown on a log scale in Figure 4b. The concentration of very stable halocarbons CFC (Chlorofluorocarbon) was almost constant for all the air samples; this is because the emission of CFC is regulated and the removal process of CFC from the atmosphere is very slow. Therefore the results of the CFCs are listed only in Table 1, but not shown in a figure. Halocarbons, which contain hydrogen in their molecules, are removed relatively faster due to the reaction with OH radicals in the atmosphere. They are usually emitted from urban and industrial areas. Therefore it is reasonable to observe higher concentrations in the JAPAN area and lower concentrations in the SOUTH area. The concentrations of C2Cl4 and CH2Cl2 were very high near Tokyo, where the air was influenced by the industrial areas along the coast.

3.4.3. Biogenic VOC

[13] Some VOCs related to biogenic activities in the ocean are shown in Figure 5. Their concentration variations were completely different from those of anthropogenic hydrocarbons. The DMS concentration was fairly variable and tended to be higher at lower latitudes (the SOUTH area). Further, the concentrations of ethylene and propylene were higher at low latitudes and low in the JAPAN/NORTH area except for the air influenced by urban pollution in the JAPAN area. The increase of alkenes during storage was often suspected for the canister sampling. For our measurements, the lowest concentrations of ethylene and propylene were about 20 pptv and 10 pptv, respectively. When these lowest values were considered as the upper limit of the concentration increase during storage, the concentration fluctuation of alkenes over the ocean was clearly higher and the increasing trend toward lower latitude is not an artifact. Isoprene is emitted from marine biogenic activities, and can be observed in the oceanic atmosphere in some regions [Yokouchi et al., 1999; Matsunaga et al., 2002]. However, high concentrations of isoprene were not observed during this cruise. CH3Br is an important O3 layer depletion gas [Butler, 2000] and there is some data over the ocean [Yokouchi et al., 2000]. An increased concentration of CH3Br was observed during this cruise. Main sources of CH3Br are ocean, fumigation, and biomass burning [Butler, 2000]. The highest concentration (69.8 pptv) was observed with long-range transport event (May 28) and was probably related to the biomass burning or anthropogenic activities. The higher concentration of CH3Br was observed during May 14 – May 21. However, enhanced anthropogenic VOC was not observed this period, and oceanic emission over mid-high latitude could cause this enhanced CH3Br concentration.

Figure 5.

Concentration variations of biogenic VOCs.

3.4.4. VOC Ratios

[14] Anthropogenic VOCs are mostly emitted from urban areas; however, they exhibit different reactivities with OH radicals in the atmosphere; which is the main removal process for VOCs. The rate constants for reactions of some VOCs with OH are summarized in Table 2. The ratio of ethane to CO, as shown in Figure 6a, provides some information about their sources. CO and ethane have similar lifetimes in the atmosphere; however, the ratio [(ethane)/(CO)] clearly decreased in the SOUTH area. There are two possible explanations for this ratio change. Besides transportation from land sources, CO is also generated in clean oceanic atmosphere by the oxidation of CH4 and larger VOCs [Warneck, 2000]. On the other hand, production of hydrocarbons was almost absent over the ocean, and their relative concentration to CO decreased over the SOUTH. Another possible explanation is the latitudinal difference of CO and VOC emissions. Biomass burning occurs more often at low latitudes and the emission ratio is relatively low ((ethane)/(CO) = 0.0028 [Shirai et al., 2003]). Since the order of reactivity is propane > acetylene > ethane, a decrease in (propane)/(ethane) is more pronounced than (acetylene)/(ethane) (Figure 6b). In the case of benzene, the order of reactivity with OH radicals is benzene > propane > acetylene. However, (benzene)/(acetylene) was almost constant over the SOUTH area. Moreover, (benzene)/(propane) increased over the SOUTH area. These trends cannot be explained by the removal process by OH radicals, suggesting an enhanced emission of benzene at low latitudes. Again, one of the possible explanations is biomass burning since it occurs more often at lower latitudes. Since the longitudinal air transport is faster than latitudinal air transport, the air in low latitudes would be influenced by the emission from regional biomass burning even for clean oceanic air. In other words, “background air” which is free from local pollution like oceanic air, can be influenced constantly by biomass burning at low latitudes. The lifetime (inverse of k × [VOC]) of benzene can be estimated as 6.4 day when the OH concentration is assumed to be 1.5 × 106 radicals cm−3 at low latitude during spring-summer time [Spivakovsky et al., 2000]. This lifetime is not long enough to become uniform longitudinally, but the ratios of VOC could maintain the information of sources. The emission ratio from biomass burning in tropical regions indicates a relatively increased emission of benzene {[(benzene)/(acetylene)] = 0.28 and [(benzene)/(propane)] = 1.49} [Shirai et al., 2003] than the observed ratio in the JAPAN area, as shown in Figure 6b. Therefore biomass burning would be one source of the relatively enhanced emission of benzene at low latitudes.

Figure 6.

Ratios of (a) ethane to CO and (b) hydrocarbons. The [(ethane)/(CO)] ratio decreased in the SOUTH area because CO is also produced on the oceanic atmosphere from the oxidation of CH4. The [(benzene)/(acetylene)] and [(benzene)/(propane)] ratios increased in the SOUTH area, indicating more benzene sources in low latitudes.

Table 2. Reaction Rate Constants With OH Radical at 298 K, 1 atma
Speciesk, cm3 molecule−1 s−1
CO2.4 × 10−13
Ethane2.5 × 10−13
Acetylene7.6 × 10−13
Propane1.1 × 10−12
Benzene1.2 × 10−12
Ethylene8.5 × 10−12
Propylene2.6 × 10−11

3.5. Variation of DMS and COS

[15] The DMS concentration obtained during this cruise was low at high latitudes and high at low latitudes in springtime (Figure 5). The atmospheric DMS concentration is determined not only from the dissolved DMS concentration in the ocean surface but also from the flux from the sea to the atmosphere and the removal/mixing processes in the atmosphere [Koga and Tanaka, 1999]. In Figure 7, the atmospheric DMS concentration co-varies with wind speed. This variation of atmospheric DMS concentration is related to wind speed, but the DMS concentrations of the JAPAN/NORTH area were different from those of the SOUTH area. In Figure 8, the DMS concentration is plotted against the wind speed according to the regions. In the SOUTH area, the DMS concentration is strongly correlated with the wind speed (R2 = 0.62); however, in the JAPAN/NORTH area, the correlation is not clear (R2 = 0.15). Generally, the DMS flux is controlled by the DMS concentration in the ocean surface and the wind speed [Aranami and Tsunogai, 2004]. This correlation only in the SOUTH area indicates that the DMS concentration in the ocean may be roughly uniform at low latitudes during this period.

Figure 7.

DMS concentration and wind speed. There is a clear difference between the JAPAN/NORTH area and the SOUTH area.

Figure 8.

Correlation plot of DMS and wind speed. The JAPAN/NORTH area and the SOUTH area are plotted with different symbols. There is clear correlation between DMS and wind speed in the SOUTH area.

[16] COS is a relatively stable species in the atmosphere and is an important source of S in the stratosphere [Crutzen, 1976]. It is produced by terrestrial as well as oceanic sources, and is also produced by the reaction of DMS and CS2 [Kettle et al., 2002; Blake et al., 2004]. The relative concentration of COS observed during the cruise exhibited some fluctuations. COS is plotted with the DMS variation in Figure 9. They exhibited similar concentration variations. COS is plotted against DMS in Figure 10 using different symbols for JAPAN/NORTH and SOUTH. The regression lines indicate that there is slight correlation between atmospheric COS and DMS concentrations. When omitting two samples of high DMS about 160 pptv, the correlation seems to be similar for JAPAN/NORTH and SOUTH. The correlation between COS and DMS can be attributed to two reasons. One is their simultaneous emission from the ocean, and the other is the production of COS from DMS oxidation. Similar to DMS, there exists a positive correlation between COS and the wind speed in the SOUTH area (R2 = 0.28); however, this correlation is not as good as the correlation of DMS with wind (R2 = 0.62). This indicates that the contribution to COS production from DMS oxidation would be greater than that from direct ocean emission. This is consistent with the estimation of Kettle et al. [2002] where COS production due to DMS oxidation and direct ocean emission were 8.7 × 109 g and 5.6 × 109 g, respectively, during May in the Northern Hemisphere. In this assumption, the deviation of two high DMS samples can be explained. During the very high DMS period, the air did not spend enough time to produce COS from DMS.

Figure 9.

Comparison with COS and DMS. COS and DMS show similar concentration variations.

Figure 10.

Correlation plot of DMS and COS. The JAPAN/NORTH area and the SOUTH area are plotted with different symbols.

4. Conclusion

[17] Atmospheric trace gas measurements were conducted over the northwestern Pacific region during the IOC 2002 cruise. It was observed that the trace gas concentrations at high latitudes were in clear contrast with those at low latitudes. The concentration of pollutants from human activities was higher at high latitudes, and the contribution by biogenic species emitted from the surface ocean was higher at low latitudes. From the analysis of the VOC ratios, benzene had a relatively higher concentration at low latitude. The influence of biomass burning would be one reason for this relatively higher benzene concentration at low latitudes. Biogenic species at low latitudes exhibited some interesting characteristics. The atmospheric DMS concentration had a good correlation with the wind speed at low latitudes. COS and DMS showed similar variations and the production of COS from DMS oxidation would be expected.


[18] The authors acknowledge the crew of the R/V Melville and researchers involved with the IOC 2002 cruise. This work was partly supported by Core Research for Evolutional Science and Technology (CREST) of the Japan Science and Technology Agency and by the Ministry of Education, Science, Sports, and Culture of Japan (Grant-in-Aid 15201004).