Distributions and fluxes of methane were determined during two surveys in March–May 2001 in the Yellow Sea and the East China Sea. Methane concentrations in the surface and bottom waters range from 2.52 to 5.48 and 2.81 to 8.17 nM, respectively. The distributions of methane are influenced obviously by the Yangtze River effluent and Kuroshio water. CH4 input via the Yangtze River is estimated to be 3.17 mol/s, of which a considerable part may be lost by air-sea exchange during estuarine mixing. Net CH4 flux exported from the shelf to the Kuroshio is about 1.84 mol/s. Methane enrichments in bottom waters occur widely, which reveals sediment sources of CH4. However, the CH4 input from the sediments of the studied region in spring is lower than other shelf regions due to low organic carbon in the sediments and high O2 contents in the water column. The sea-to-air methane fluxes are estimated to be 1.36 ± 1.45 and 2.30 ± 2.36 μmol m−2 d−1 using Liss and Merlivat  and Wanninkhof  relationships, respectively, and the estimated spring emission rate of methane ranges from 9.32 × 10−3 to 15.7 × 10−3 Tg CH4 yr−1. However, these estimations suffer from the neglect of seasonal variability and should be taken as a low limit. Therefore more measurement campaigns should be carried out to enhance our understanding of this particular oceanic region.
 Methane is an important atmospheric trace gas, which currently accounts for ∼20% of the radiative forcing due to increase of greenhouse gases [Houghton et al., 2001] and plays a critical role in the atmospheric chemistry and air-sea interaction [Cicerone and Oremland, 1988]. The atmospheric concentration of methane has increased by 1060 ppb (151%) since 1750 and continues to increase [Houghton et al., 2001]. On the basis of the inventory of major sources and sinks of methane, the prediction of methane impact on the world's climate can be made. For instance, it is demonstrated that the world's ocean emits about 11–18 Tg CH4 yr−1 to the atmosphere, which contributes only about 4% to the total global emissions and has a modest effect in the global budget of atmospheric methane [Crutzen, 1991; Bange et al., 1994]. However, emission of CH4 from the ocean changes considerably in time and space. The surface waters of the vast oligotrophic oceanic areas are usually near gas-exchange equilibrium with the atmosphere [Bates et al., 1996], and therefore make only a small contribution to the total oceanic emissions, whereas continental shelves and coastal regions, which occupy a small part (∼16%) of the world ocean, appear to be responsible for ∼75% of the oceanic CH4 emission [Bange et al., 1994]. Nevertheless, owing to the high variability in hydro-chemical properties of world oceans, particularly shelf and coastal waters, and the sparse data coverage, the global inventory of CH4 remains largely uncertain and the oceanic CH4 budget is poorly documented.
 The East China Sea (ECS) and the Yellow Sea (YS) together form an important marginal area of the northwest Pacific Ocean that is one of the largest continental shelves in the world, with a total surface area of 1.2 × 106 km2, of which ∼75% has water depth less than 50–100 m. The hydrographic characters of this region are greatly influenced by a circulation system that includes the Kuroshio, Tsushima Warm Current, Yellow Sea Warm Current on the eastern boundary of the shelf, the Yellow Sea and the East China Sea Coastal Currents on the western side and the Taiwan Warm Current dispersing out in the middle shelf [Su, 1998; Lee et al., 2000]. The Kuroshio, carrying oligotrophic waters northward with low concentrations of nutrients and chlorophyll, and high temperature and salinity, is found to be close to equilibrium with the atmospheric CH4 [Tsurushima et al., 1996; Rehder and Suess, 2001], and may influence the CH4 distribution in ECS shelf through water mass exchange [Liu et al., 2003]. The Yangtze River (Changjiang), one of the largest rivers in the world with huge amounts of freshwater discharge (9.24 × 1011 m3/yr), sediment load (4.86 × 108 t/yr), and abundant nutrients, is merged into the ECS and the YS from the west [Tian et al., 1993; Zhang, 1996]. Previous studies show that riverine waters usually have high concentrations of dissolved methane, which range from 4.8 to 5000 nM [Scranton and McShane, 1991; De Angelis and Lilley, 1987; De Angelis and Scranton, 1993] on the world. High riverine loads of nutrients and organic materials are expected to provide conditions favorable for the microbial production of CH4 in the estuarine and coastal areas. Therefore the Yangtze River is also expected to affect the methane budget and hence the CH4 distribution in the YS and ECS. There are, however, very limited data available for dissolved methane in the YS and ECS, which are obtained from one or two transects across the shelf [Tsurushima et al., 1996; Zang, 1998]. In this study we present data of methane from seawater samples covering almost whole area of the ECS and the YS (Figure 1). The objective of this study is to understand the spatial distributions of methane in the YS and ECS, to estimate the CH4 emission into the atmosphere, and to evaluate the contribution from this region to the global sea-air flux of CH4.
2. Methods and Materials
 Two cruises were conducted in the ECS and the YS, including 26 March to 23 April 2001 by R/V Beidou (973-I), and 30 April to 17 May 2001 by R/V Dong Fang Hong 2 (973-II), respectively. The sample locations are shown in Figure 1. The 973-I cruise contains 10 cross-shelf transects (A–J), in which we measured the horizontal distribution of methane in the surface, middle, and bottom waters. In the 973-II cruise, vertical distributions of methane were studied at stations covering different hydrographic regions with N4 off the Yangtze River Estuary, P1 and P2 on the slope of ECS, N6 and L20 on the mid-shelf, N5 in the upwelling region, and N2 at the Yellow Sea cold water region.
 Water samples from various depths were collected using 2-L Niskin bottles mounted to a Seabird CTD Rosette (General Oceanics Co.). Samples for methane determination were transferred from Niskin bottles into 135-mL glass bottles using rubber connecting tubes with glass pipette ends. After overflow of approximately 1.5- to 2-fold of bottle volume, 1 mL of saturated solution of HgCl2 was added to inhibit microbial activity, then the sample bottle was sealed with a rubber stopper pierced with a 5-gauge hypodermic needle to exclude the excessive water. After removing the hypodermic needle, the stopper and the bottle are then wrapped tightly and stored upside down in a dark box and kept at 4°C. All the water samples were analyzed after returning to the shore laboratory within 40 days of collection. A small headspace (usually <0.3 mL) formed in the sample bottle after collection and was corrected for the partitioning of CH4 between the sample water and the headspace as described by Tilbrook and Karl . To test reliability of the storage procedure used in this study, experiments were done using in situ seawaters, and the results showed that the decrease of CH4 concentration in sample bottles during a period of 50 days is less than 3%, and the effect of storage on the sample concentration appears to be small. Data of temperature, salinity, and dissolved oxygen were obtained from the shipboard CTD profiles.
 In the laboratory, dissolved methane was measured by gas chromatography (GC) using a gas-stripping method [Swinnerton and Linnenbom, 1967]. Ultra-pure N2, purified by passing it through a MS 5A trap maintained in liquid nitrogen, was used as the purging gas. Dissolved methane and other gases were purged from about 135 mL of seawater by bubbling with ultra-pure N2 at the flow rate of 80 mL min−1 for 5 min. The extracted gases were allowed to pass through K2CO3 and Ascarite® traps to remove water vapors and carbon dioxide, respectively, and then through a U-shaped stainless steel trap packed with 80/100 mesh Porapak Q to concentrate methane. During stripping, the Porapak Q trap was immersed in the liquid nitrogen. After completion of the stripping procedure, the Porapak Q trap was transferred from liquid nitrogen to a boiling water bath, and the released methane was injected into a Shimadzu GC-14B gas chromatography. In the chromatography, methane was separated on a 3 m × 3 mm in diameter stainless steel column packed with 80/100 mesh Porapak Q and detected by a flame ionization detector (FID). The CH4 retention time is about 2 min. Calibration of the FID response was done by injection of a certain volume of standard gas of 49.6 ppmv CH4/N2 (Research Institute of China National Standard Materials) into the stripper filled with blank seawater. The methane of the blank seawater was previously stripped out, together with other dissolved gases, by ultra-pure N2. After injection, the blank seawater is subsequently analyzed by the same procedure used for unknown samples. Calibration is carried out every 2 hours of GC operation. The method detection limit (MDL) for methane analysis in this study was 0.06 nM (MDL is defined as CH4 concentration in a 135-mL seawater sample corresponding to 2 standard deviations of seven replicates of the blank), and the precision of repeated analysis of water samples was better than 3% in routine sample analysis.
 In the 973-I cruise, discrete air samples at nine stations were collected from the bow of the ship using 100-mL glass bottles similar to those used for seawater sampling. Analysis of air samples is carried out through the stripper by the same procedure used for seawater determination so that no separate calibration procedure is required.
3. Results and Discussion
3.1. Horizontal Distributions of Methane
 In the 973-I cruise, the methane concentrations in surface and bottom waters vary from 2.52 to 5.48 nM and 2.81 to 8.17 nM, respectively, throughout the ECS and YS. The horizontal distributions of methane in the surface and bottom waters are shown in Figure 2, which illustrates the gradients of methane concentration decreasing gradually offshore, particularly in the inner and middle shelf regions. In the eastern side of the ECS shelf, methane concentrations in the surface waters are quite low, influenced by Kuroshio water with methane of 2.0–2.5 nM. From Figure 2, it can be easily identified that the methane distribution shows a high methane plume dispersed off the Yangtze River Estuary, which extends northeastward along the axis to the Cheju Island (Figure 2). Hydrographic data indicate that the Yangtze River Diluted Water disperses into the northern ECS and the southeast YS, and its front reaches the southwest of Cheju Island (125°E) [Liu et al., 1992; Hur et al., 1999]. Since dissolved methane in the surface waters of downstream Yangtze River ranges from 112 to 190 nM in spring (G. L. Zhang, unpublished data, 2003), which is almost 2 orders of magnitude higher than the methane concentration of 2–3 nM found in the open ocean, the high CH4 plume could be explained by the influence of the Yangtze River inflows.
 During our survey, CH4 enrichment in the near-bottom waters is found at most stations, including those off the Yangtze River Estuary. This finding underlines that there exist important sources of methane in deep waters, presumably from the bottom sediments. However, at stations N1, C3, D2, and E1, located along the YS coast with depth less than 40 m, high levels of CH4 are found at the surface due to the influence of the YS Coastal Current with rich methane, and CH4 distributes uniformly in the middle and near-bottom waters.
3.2. Distribution of Methane Across the Shelf
 The vertical distribution of methane along the PN line in the ECS is shown in Figure 3, which indicates an obvious onshore-to-offshore gradient with the CH4 concentrations decreasing with distance off the Yangtze River Estuary. The gradient supports the influence of the Yangtze River effluents with high methane level in the surface and near-bottom waters in the shallow shelf region. At station P2 at the edge of continental shelf, high methane of 9 nM is found from the near-bottom water, and Tsurushima et al.  and Zang  also reported methane higher than 6 nM in the near-bottom water at the same region. This suggests a year-round near-bottom or sediment source in the shelf-edge region. Previous studies showed that the subsurface and intermediate water of Kuroshio could intrude on the shelf along the bottom of the shelf edge and provide a large amount of additional nutrients to ECS shelf waters [Zang, 1998]. However, a low methane level of 2.14–2.67 nM occurs in the depth range of 80–400 m at station P1 in this study, which is much lower than the methane concentrations found from the near-bottom water in the shelf-edge region and hence overrules the support of the Kuroshio waters to the accumulation of methane in the near-bottom waters. Tsurushima et al.  excluded another possible source from bottom sediment and concluded that high CH4 might be due to the accumulation of organic matter flowing from the coast to the oceanic zone via the near-bottom waters of the shelf or due to the southward flow of old water mass from the northern ECS along the shelf edge. Since no methane concentrations higher than 9 nM are observed in the bottom waters of northern ECS in this study, the former explanation seems more reasonable than the latter.
3.3. Vertical Profiles of Dissolved Methane in the YS and the ECS
 Vertical profiles of methane at eight stations in the YS and the ECS are presented in Figure 4, which shows the fingerprint of the CH4 profile from various subregions of the study area. Since the geography and hydrography of the YS and the ECS show complex features, the continent shelf is strongly influenced by the wind and land-source input, particularly the shallow shelf regions near the coast; the vertical profiles of methane are highly variable in time and space.
 At station N1 with low water temperature of 10°C–11°C in northwest YS (Figure 4a), CH4 saturations fall with depth from 95% to 88% in the water column. Salinity and dissolved O2 are vertically uniform because of the strong vertical mixing in coastal waters in early spring [Wang et al., 1998]. At station N2 in the YS Cold Water, water temperature decreases sharply in the top 40 m, while CH4 saturation tends to increase gradually from 91% in the surface to 111% in the near-bottom waters, corresponding to a reduction of dissolved oxygen by 20% (Figure 4b).
 At station N4 off the Yangtze River Estuary (Figure 4c), temperature and dissolved O2 have almost uniform vertical profiles. The increase of salinity with depth within the top 10 m demonstrates that the surface water is affected by riverine effluent. The methane saturation increases steadily with depth from 694% in the surface to 1230% at 10 m, and then becomes relatively uniform until bottom, which indicates great CH4 losses by air-sea exchange in the surface and an important CH4 source from the bottom sediments at N4. At stations N5, N6, and L20, weak pycnoclines can be found. At the ECS coastal upwelling region (i.e., N5), high CH4 saturation of 362% is found from the subsurface water at the top of halocline. This methane maximum can probably be explained by both in situ biological production and the supply from upwelled deep water rich in methane. Closer to the bottom, the CH4 is high, with a maximum saturation of 383%, though data are scattered (Figure 4d). The vertical distribution of dissolved O2 shows a reduction of concentration beneath the halocline, implying favorable conditions for in situ CH4 production there. The CH4 distribution at station N6 (Figure 4e) shows a methane maximum in the near-bottom waters, indicating that methane could be produced from the sediments and diffused rapidly into the overlying water. The existence of the pycnocline prevents the upward transport of methane into waters of the upper mixing layer, thus reducing the sea-air exchange flux of methane on the shelf. At station L20 (Figure 4f), the stratification is observed at about 50–60 m, temperature and salinity are uniformly distributed in the upper 50 m, and water properties show rapid change across the pycnocline. With regards to CH4, the saturation is 130–137% at surface mixing layer; it then increases with depth to a maximum of 269% below the pycnocline, which is twice as much as that in the surface waters. Across the pycnocline, dissolved O2 falls by ∼30% in saturation (Figure 4f).
 Low surface methane concentrations of 2.18 and 2.24 nM corresponding to saturations of 113% and 115%, with temperature of 25.2°C–25.7°C and salinity of 34.4‰ is found at stations P1 and P2 (Figures 4g and 4h). Rehder and Suess  observed about 2 nM of CH4 in the surface waters of the Kuroshio; therefore the low surface methane concentrations of station P1 and P2 probably represent the CH4-poor Kuroshio surface waters. At station P2, a methane maximum with saturation of 427% is found at 140 m (Figure 4h), suggesting a CH4 source in near-bottom waters; then CH4 decreases again to the bottom, although the dissolved oxygen decreases gradually beneath the surface mixing layer. At station P1, a high level of CH4 occurs at about 40 m, corresponding to the existence of the pycnocline; this phenomenon of subsurface methane maximum is observed in various oceanic provinces and is attributed to the in situ biological production [Burke et al., 1983; Cynar and Yayanos, 1992; Tilbrook and Karl, 1995; Patra et al., 1998; Upstill-Goddard et al., 1999]. The second but deeper methane maximum is found at about 400 m at station P1. The advective supply from the methane-rich bottom water at shallower region and the in situ production of CH4 via degradation of organic materials in combination are potential mechanisms to maintain this high methane level. Below the two CH4 maximums, methane saturation falls with depth to 48.8% at 700 m in accordance with the lowest salinity, which reveals the character of Kuroshio intermediate waters. Below 700 m, methane saturation increases again with depth, suggesting a bottom sediment source. Lin et al.  have found the organic carbon concentrations in bottom sediments increased across the shelf break and showed the highest level in sediments at water depths of 1000–1500 m in the ECS, which supports the hypothesis that CH4 in near-bottom waters is maintained by the inputs from the organic-rich sediments.
3.4. CH4 Input Via the Yangtze River and Estuarine CH4 Removal
 CH4 concentrations in the lower reaches of the Yangtze River range from 112 to 190 nM with an average of 137 nM in spring (G. L. Zhang, unpublished data, 2003), which falls within but toward the low end of the reported CH4 concentrations in worldwide rivers [Upstill-Goddard et al., 2000]. We estimate the input of methane from the Yangtze River to the ECS and YS by multiplying the mean fresh water CH4 concentration of 137 nM by the mean discharge of 23148 m3/s in April 2001 [Changjiang Water Resources Commission, 2002], which yields a methane flux of 3.17 mol/s. Although the Yangtze River discharge is subject to seasonal variations, it appears that the period of our survey (April/May) is a time of transition from the low discharge of winter to the high water flow of summer [Changjiang Water Resources Commission, 2002]. Therefore the results of our observation represent the mean value, and indicate that Yangtze River is an important source for CH4 in the ECS and YS.
 However, the relationship between CH4 and salinity for the adjacent marine waters off the Yangtze River Estuary (e.g., data from E, F, and G transects to the west of 126°E) shows a nonlinear character and can be simulated by a second-order polynomial (Figure 5). Such a deviation from conservative behavior has also been found in other estuarine and coastal waters, including the North Sea and the Black Sea [Upstill-Goddard et al., 2000; Amouroux et al., 2002], and is ascribed to substantial CH4 removal through microbial oxidation and air-sea gas exchange during estuarine mixing. Studies in estuaries of the North Sea showed that CH4 removal, which is significant in the initial stages of mixing and essentially completed at salinity 5‰ of the upper estuary, can reach ∼90% and strongly decrease the influx of CH4 to the North Sea [Upstill-Goddard et al., 2000]. Owing to a lack of data from the Yangtze River Estuary, the potential removal of CH4 in estuarine mixing cannot be determined directly. Assuming that 90% of CH4 can be removed in the Yangtze River Estuary with a partition of 93/7 between air-sea exchange and oxidation as derived by Upstill-Goddard et al. , the CH4 removal can be tentatively estimated here, which yields an additional atmospheric CH4 flux of 2.65mol/s in the Yangtze River Estuary. This will increase the atmospheric CH4 flux from ECS and YS (see section 3.7.2) by 8–13%. Therefore the significance of such CH4 removal in the estuaries has to be considered when evaluating the CH4 budget in marine environments like ECS and YS. Clearly, detailed studies of CH4 distributions in the Yangtze River Estuary in the low-salinity area are prerequisites to assess the CH4 flux at land-sea boundary and hence the CH4 inventory for the northwest Pacific.
3.5. CH4 Exchange Between ECS Shelf Water and Kuroshio
 Since ECS and YS represent one of the largest continental margin seas, and act as a transition zone between land and ocean, it is therefore important to investigate the marine processes transporting materials from the continental shelf to the open ocean [Wong et al., 2000; Hung et al., 2003]. As the Kuroshio travels along the northwestern slope of the Okinawa Trough, the warm, saline, and CH4-poor Kuroshio Surface Water is imported into the shelf while the fresher, colder, and CH4-rich ECS Shelf Water is exported to the Kuroshio through frontal processes. A simple estimate of the net CH4 loss through interactions between ECS Shelf Water and Kuroshio can be obtained from the average volume transport between ECS shelf and Kuroshio, and an estimate of the methane concentrations associated with the inflow and outflow. Here the water budget used by Wong et al.  is applied to our estimation, namely about 0.5 Sv of Kuroshio Surface Water, and 0.5 Sv of the upwelling Kuroshio Subsurface Water is imported into the East China Sea Shelf while 1.1 Sv of the ECS Shelf Water is exported to the Kuroshio and the Japan Sea. At Station P1, low methane of about 2.28 and 2.14 nM is found at the Kuroshio surface (0–50 m) and subsurface waters (50–150 m), respectively. The average methane concentration of about 3.68 nM is observed at the water column of the ECS shelf. Assuming the concentrations of 2.28 and 2.14 nM to be representative for the inflow water from the Kuroshio surface and subsurface water, respectively, and 3.68 nM for the outflow of ECS Shelf Water, net CH4 exported from the Shelf to the Kuroshio is therefore about 1.84 mol/s, which accounts for ∼60% of the total riverine input via the Yangtze River. Therefore the methane distribution of CH4 in the ECS continental shelf area will be influenced greatly by the Kuroshio shelf exchange processes.
3.6. CH4 Supply From Sediments
 Previous studies showed that methane can be produced through bacterial degradation of organic materials in the coastal sediments, followed by release into the overlying near-bottom waters through sediment water exchange [Martens and Klump, 1980; Ivanov et al., 2002]. Seepages of thermogenic methane from the sediments also may occur in some specific shelf areas such as the North Sea and the UK continental shelf [Judd et al., 1997; Rehder et al., 1998].
 In YS and ECS, a band-type distribution along the coast south of the Yangtze River Delta with high Corg concentrations of 0.4–0.9% in the inner shelf, low concentrations (<0.2%) of Corg in the outer shelf sediments, and Corg concentrations of less than 0.4% for the majority of continental shelf sediments was observed in ECS [Lin et al., 2002]. The Corg concentrations ranging from 0.0% to 1.1% with an average of 0.3% are observed in the surface sediments of the southeastern YS [Cho et al., 1999]. In general, the Corg concentration in the YS and ECS sediments is lower than the average value of 0.75% for the world shelf [Berner, 1982] and those observed in other shelf sediments, such as the northwestern Black Sea shelf [Galimov et al., 2002]. However, the methanogenesis rates depend strongly on the availability of organic matter [Heyer et al., 1990], and a considerable part of the methane produced in sediments can be oxidized in the upper horizons of bottom sediments and the overlying water column [Ivanov et al., 2002]. In this study, high oxygen contents (4–11 mg/L) in the water column and low Corg concentrations in the sediments of YS and ECS are unfavorable for the production and subsequent transfer of methane from the sediments to the overlying water. Although surveys conducted on the ECS Shelf indicated the potential of abundant fossil fuel and natural gas exploration, and large oil-gas fields, for example, Pinghu, are currently under exploitation [Gu, 1996], no obvious hydrocarbon seepage has been observed yet. As a result, sediment release of CH4 into the ECS and YS water column results in methane enrichment of 3–8 nM in the near-bottom waters compared to surface waters in spring. However, in the North Sea with shallow gas deposits and active sediment CH4 seepages [Rehder et al., 1998], and the Black Sea with anoxic and organic-rich sediments [Reeburgh et al., 1991; Ivanov et al., 2002], high CH4 concentrations of some dozens to several hundreds nanomoles per liter in the near-bottom waters are found. Considering the big difference in CH4 concentrations in bottom waters related to the North Sea and the Black Sea, however, it is reasonable to predict that sediment input of CH4 to the water column of ECS and YS in spring should be 1 to 2 orders of magnitude lower and contribute limitedly to the atmospheric CH4 fluxes.
 However, studies in coastal sediments show that CH4 production rates may exhibit seasonal variations with a maximum in summer and considerably lower in spring and winter [Bange et al., 1994; Ivanov et al., 2002]. In the studied area, strong O2 depletion occurs in the bottom waters off the Yangtze River Estuary in summer [Li et al., 2002]. In this O2-deficient region centering at 122°30′E and 30°50′N the minimum O2 level is about 1 mg/L and the area with O2 content of less than 2 mg/L covers 13700 km2, and the area with O2 content of 2.0–3.5 mg/L can extend southeastward to around the 100-m isobath [Li et al., 2002]. In combination with the arrival of large amounts of allochthonous and autochthonous organic matter in shallow-water bottom sediments off the Yangtze River Estuary, it is reasonable to deduce that methane production in the bottom sediments of the O2-deficient region should be much greater in summer than in spring. Tsurushima et al.  found that CH4 concentrations in the bottom 30-m layer of the ECS continental shelf area increased three-fold during the stratified period in summer to autumn compared with the cold seasons, and CH4 concentration was inversely correlated fairly well with the dissolved oxygen. These results support the hypothesis of higher sediment input of CH4 from ECS in warm seasons and indicate that CH4 input from the sediments of ECS and YS should have an obvious seasonal nature. Therefore the contribution of the ECS shelf area to the atmospheric CH4 may be underestimated if derived only from the direct observation of CH4 in the surface waters in spring.
3.7. Sea-Air CH4 Exchange
3.7.1. Variations in Air Measurements and Surface Saturations
 In the 973-I cruise, the atmospheric methane over the YS and the ECS ranged from 1.70 to 1.82 ppmv with an average of 1.76 ± 0.05 ppmv (n = 9), which is very close to the mean of 1.759 ppmv between Kyushu and the northern tip of Taiwan reported by Rehder and Suess . Expected equilibrium concentrations of methane in surface seawaters are calculated using the mean atmospheric CH4 (i.e., 1.76 ppmv) and the in situ temperature and salinity of given water samples using the equation of Wiesenburg and Guinasso . The saturation refers to the measured methane concentration divided by the estimated equilibrium concentration in surface seawater. As shown in Figure 6, methane is supersaturated in the surface waters throughout the study areas in 973-I cruise, and the saturations range from 110 to 222%, with an average of (134 ± 22)%. From Figure 6 it is clear that the ECS shows higher variability and higher values of CH4 saturations than the YS. This feature might result from the nature of heterogenic hydrodynamics, for example, the complicated circulation and multiple water masses, of the ECS relative to the YS. On average, the observed methane saturation in the ECS is higher than the previous reported value of 116% [Rehder and Suess, 2001] for the eastern part of the ECS influenced by the Kuroshio. Tsurushima et al.  reported mean saturations of 113%, 154%, 135%, and 138% for February and October 1993 and June and August 1994, respectively, at the surface water along the PN section, which are comparable to the data of this study. Our results for ECS and YS in spring fall toward the low end of the ranges of CH4 saturation ratios in coastal and shelf areas worldwide (Table 1). However, the observed saturations higher than 100% indicate that the YS and ECS are sources of atmospheric methane.
Table 1. Summary of CH4 Measurements in Various Coastal and Shelf Areas in Literature
Number of Stations
Surface CH4, nM
Surface R, %
Flux, μmol m−2 d−1
Wind Speed, m/s
Kw was estimated by the LM86 equation.
Kw was estimated by the W92 equation.
Values are calculated according to the results shown in the references.
 The supersaturation of methane observed in the surface waters of the YS and the ECS supports net sea-to-air methane fluxes (F in mol m−2 d−1), which can be estimated by the following equation:
where Cobs and Ceq are observed and expected equilibrium concentrations of methane in mol/L in surface seawater, and kw is gas transfer coefficient in cm/h, which is the function of the wind speed and the Schmidt number (Sc). Several empirical relationships have been used to quantify the wind speed dependence of kw [Smethie et al., 1985; Liss and Merlivat, 1986; Wanninkhof, 1992; Erickson, 1993]. Here, we calculate kw using both the trilinear kw/wind speed relationship established by Liss and Merlivat  (LM86) and the quadratic kw/wind speed relationship established by Wanninkhof  (W92), respectively. These two relationships represent lower (LM86) and higher (W92) estimates in sea-air exchange parameterizations. The coefficient kw is adjusted by multiplying (600/Sc)n for LM86 (n = 1/2 for wind speeds >3.6 m/s and n = 2/3 for wind speeds <3.6 m/s), and (660/Sc)1/2 for W92. The Sc for methane in seawater is calculated using the equation given by W92. The major uncertainty in the assessment of sea-air gas fluxes is related to the estimation of the gas transfer coefficient, which depends on the type of wind data used. Morell et al.  found that fluxes computed using climatological wind speed data often exceed those using ship-based wind speed measurements by over 50%. In this work, we compute the gas transfer coefficients using shipboard in situ wind speed data; hence the obtained methane fluxes should be instantaneous and conservative. In the 973-I cruise, the measured wind speeds at about 10 m above the sea surface (u10) range from 1.8 to 10.3 m/s. On the basis of the in situ data of wind speed and the sea surface temperature, the sea-to-air flux for each station is estimated using the selected kw/u10 relationships, and the results are summarized in Table 2. From Table 2, it can be seen that using W92 leads to higher flux estimates than using LM86 by a mean factor of about 1.6. The discrepancy between the two sets of flux values indicates the uncertainty involved in using different sea-air exchange models. The difference of CH4 fluxes between YS and ECS can be seen from the horizontal distribution of methane fluxes shown in Figure 7, which indicates that the CH4 fluxes in the YS are relatively lower but show a more stable distribution than those in the ECS. From Figure 7, the influence of the Yangtze River Estuary on the atmospheric methane flux is quite clear; for example, the sea to air flux of CH4 in the area adjacent to the Yangtze River (∼4.0 μmol m−2 d−1) can be five- to ten-fold higher than the values (<0.5 μmol m−2 d−1) observed in the broad middle shelf farther offshore (Figure 7).
Table 2. Average Sea-Air Methane Fluxes and Extrapolated Spring Emission Rate From the YS and ECSa
Average Flux, μmol m−2 d−1
Spring Emission Rate, Tg CH4 yr−1
LM86: fluxes calculated using the trilinear k/u10 relationship established by Liss and Merlivat ; W92: fluxes calculated using the quadratic k/u10 relationship established by Wanninkhof . Numbers in the parentheses are the total station numbers.
4.2 × 105
0.81 ± 0.50
1.33 ± 0.76
1.99 × 10−3
3.25 × 10−3
7.7 × 105
1.63 ± 1.67
2.77 ± 2.71
7.33 × 10−3
12.5 × 10−3
11.9 × 105
1.36 ± 1.45
2.30 ± 2.36
9.32 × 10−3
15.7 × 10−3
 On a global scale, the CH4 fluxes from the YS and ECS are comparable to the Aegean Sea and Arabian Sea, but much lower than the Black Sea and North Sea (Table 1). From Table 1, it can be seen that the high atmospheric CH4 fluxes from the Black Sea mainly resulted from the high surface CH4 saturation, which is about 4–5 times higher than those observed in ECS and YS, and is attributed predominately to the abundant CH4 release from the anaerobic shelf bottom sediments [Amouroux et al., 2002; Ivanov et al., 2002]. For the North Sea, both the higher gas transfer coefficients derived from the higher observed wind speeds and the higher surface CH4 saturations account for the atmospheric CH4 fluxes 5–9 times higher than the ECS and YS (Table 1). From Table 1, a clear trend of highly seasonal variation is found in the atmospheric CH4 fluxes from coastal and shelf regions. As to ECS and YS, the dispersion of the Yangtze River effluent plume begins in April and reaches maximum in August when the freshwater discharge floods together with large input of allochthonous organic matter and methane, and the development of oxygen depletion in the near-bottom waters off Yangtze River Estuary in summer and autumn provides favorable conditions for the production of methane in either deep waters or bottom sediments or both. This implies that CH4 flux from ECS and YS should have a seasonal nature, which is, however, necessary to prove in the near future. Therefore, geographic, meteorologic, and seasonal variations can cause highly variable atmospheric CH4 fluxes in shelf regions, which can significantly influence the global CH4 budget.
 On the basis of the mean methane flux, the emission rate of methane from the YS and ECS can be preliminarily estimated, and the results are shown in Table 2. The results show that the spring methane emission rates from the YS and the ECS range from 9.32 × 10−3 to 15.7 × 10−3 Tg CH4 yr−1 (i.e., 18.5 ∼ 31.1 mol/s of CH4) taking the result by LM86 as a lower limit and the result by W92 as an upper limit. From the CH4 emission data, the contribution of the YS and ECS together (about 0.3% of the total ocean) to the net global sea to air CH4 flux can be estimated, and is about 0.1% relative to the global oceanic CH4 emission of 11–18 Tg yr−1 [Bange et al., 1994], and suggests that the YS and ECS seems to contribute a limited amount to the total oceanic CH4 flux. However, our estimation suffers from the neglect of seasonal variability of methane fluxes from ECS and YS, and should be taken as low limits. In order to address the annual CH4 flux in detail, seasonal investigations should be undertaken in this region in the near future. On the other side, the mean CH4 saturation of 134% and the gas transfer coefficient of 6 cm/h reported for ECS and YS in this study are much lower than the mean value of 395% and 13.7 cm/h for shelf region used by Bange et al.  to estimate the global oceanic CH4 emission. Since ECS and YS represent one of the largest shelf areas in the world, this result suggests that the contribution of shelf regions to global oceanic CH4 emissions may be overestimated due to inappropriate extrapolation of surface CH4 saturations and gas transfer coefficients for many shelf areas without observed data. Therefore, in order to estimate the global oceanic CH4 emissions and the contribution from shelf regions more accurately, more studies on spatial and seasonal variations of distribution of methane in various oceanic provinces would be required.
 The distributions of methane in the YS and ECS are influenced by the methane-rich Yangtze River effluent and methane-poor Kuroshio water. Total CH4 input via the Yangtze River is estimated to be 3.17 mol/s; however, the CH4 removal through air-sea exchange and microbial oxidation during estuarine mixing may reduce CH4 input to ECS and YS by 90%. The water exchanges between the shelf and the Kuroshio result in a net export of methane from the ECS shelf to the Kuroshio or pelagic sea, implying that the marginal seas are a significant net source of methane to the ocean interior. The CH4 input from sediments in ECS and YS in spring is much lower than shelf areas like the Black Sea and North Sea due to the low organic carbon concentration in the sediments and high oxygen saturation in the bottom water, and hence contributes limitedly to the sea to air CH4 fluxes. The estimated atmospheric CH4 fluxes from ECS and YS in spring are low relative to other coastal oceans, which means that we should reassess the relative importance of coastal and shelf waters compared to open waters. Further studies should be designated to identify the seasonal variation of CH4 in seawaters of ECS and YS, and net methane input via the Yangtze River and contribution of methane from sediment source should be quantified as well to establish the inventory of CH4 in the NW Pacific shelf region.
 The authors are grateful to the Associate Editor and two anonymous reviewers for their constructive suggestions, which greatly improved the manuscript. The authors also wish to thank the crew of R/V Beidou and R/V Dong Fang Hong 2 and colleagues from the Laboratory of Marine Biogeochemistry, Ocean University of China, for assistance in field sample collections. D. J. Huang and H. Wei are acknowledged for providing us with the hydrographic and climatologic data. This study is funded by the Ministry of Science and Technology of China through contracts G1999043705 and 2001CB409703, by National Science Foundation of China through contract 40490000, and by the Ministry of Education of China through contract 20020423001.