1 INTRODUCTION

Dissolved oxygen (DO) in marine systems, which mainly affected by a complex interaction between physical processes (air-sea exchange and convection) and biochemical processes (photosynthesis, respiration and remineralization), has a great influence on the distribution, transport, and transformation of elements and compounds, and is essential for marine organisms (Song et al., 1996; Gao and Song, 2008; Rovelli et al., 2016). Changes in DO concentration, distribution, supply, and consumption, therefore, have considerable effects on local marine ecosystems. For instance, hypoxia (DO < 2 mg/L or < 30% saturation) would perturb geochemical dynamics, alters the structure and function of ecosystems (Rabalais et al., 2010), and create a heavy biological stress on marine organisms (Diaz and Rosenberg, 1995).

The East China Sea (ECS), a marginal sea area under the interaction of terrestrial rivers (e.g. Changjiang (Yangtze) River) and the Kuroshio Current (Fig. 1a), is noted for its high primary production, abundant fishery resources, and rich biodiversity (Song, 2010; Hung et al., 2013). In recent years, the ECS has undergone increasing anthropogenic environmental changes, and the global warming is exerting a subtle influence on the ocean currents (Deutsch et al., 2006; Feely et al., 2008; Lui et al., 2015; Rho et al., 2016). Furthermore, the ECS coastal water is suffering from frequent eutrophication (Liu et al., 2000; Chen et al., 2003; Li et al., 2014; Xing et al., 2016) and serious hypoxia in spring and summer (Li et al., 2002; Chen et al., 2007; Wei et al., 2007; Wang et al., 2012; Zhu et al., 2016). Since hypoxia begins to take place in spring of recent years (Guo et al., 2014), the DO distribution in the ECS during spring has become a topic of science.

The Kuroshio Current (KC) is a western boundary current of the Pacific Ocean, which originates from the Philippine Sea east of the Lüzon Strait and flows along the outer edge of the ECS shelf. Many Research programs such as the China-Japan Joint Research Program on the Kuroshio (JRK, 1986–1992) and the Kuroshio Edge Exchange Processes (KEEP, 1989–2000) have strengthened the evidence of the Kuroshio's impact on adjacent sea areas (Lin et al., 1995; Wong et al., 2000). There is no doubt that the Kuroshio intrusion has affected the hydrological conditions, the distribution of biota, and the components of seabed sediments in the marginal seas of China, particularly the ECS. Recently, some researchers have reported the effects of the Kuroshio intrusion on the nutrients, inorganic carbon, and trace elements in the ECS (Zhang et al., 2007; Guo et al., 2012; Lu et al., 2016; Liu et al., 2017). However, very few reports concern the effect of the Kuroshio intrusion on the DO in the ECS.

In this study, we investigated the spatial distribution of DO in the ECS and its adjacent Kuroshio waters during spring. Based on the spatial distribution of DO with low oxygen zone in the coastal area, we searched for evidence to prove the impact of Kuroshio on the DO in the ECS.

2 SAMPLING AND ANALYTICAL METHOD

A field investigation was carried out in the ECS and Kuroshio Current regions on the R/V Kexue No.1 from 18 May to 13 June 2014 (Fig. 1b). At each station, seawater samples were collected from various depths using a rosette sampler with 10-liter Niskin sampling bottles. Water samples in the ECS were collected at depths of 2 (surface layer), 10, 20, 30, 50, 75 m, 100 m, and 3 m above the seafloor (bottom layer), respectively, according to the water depth of each station; while samples in the Kuroshio area were collected at depths of 5 (surface layer), 30, 50, 75, 100, 150, 200, 300, 500, 800, 1 000, 1 500, and 2 000 m, respectively. Data of environmental conditions, such as depth, temperature (T), salinity (S), and density (sigma-t, σ_t), were obtained at each station using a conductivity-temperature-depth (CTD) sensor (SBE-911 plus).

After sampling, DO concentrations (mg/L) were immediately determined on board the vessel, using the traditional Winkler titration method at precision of 7×10^-5 mg/L (Bryan et al., 1976). The pH (the total hydrogen-ion concentration scale) of the water samples were measured at 25±0.1℃ by a pH meter (Thermo Orion 5-star) with a combination electrode (REX E-201-D), which was calibrated by buffers (Tris and 2-aminopyridine) prepared at a salinity 35. The precision of pH measurement was ±0.002 pH unit. Then chlorophyll a (Chl a) samples were collected by filtering 200‒500 mL seawater samples through cellulose-acetate filters (acid-cleaned, poresize: 0.45 μm), and the filters were stored at -20℃ in darkness before further analysis.

In laboratory, Chl a on the filters were extracted in 10-mL N, N-dimethylformamide, and then the extractions were determined using a fluorescence spectrophotometer (Hitachi F-4600), with a precision better than 2% (Suzuki and Ishimaru, 1990).

In addition, dissolved oxygen saturation (DO%) and apparent oxygen utilization (AOU, which can be used as reference for oxygen deficit) were calculated using following equations (Weiss, 1970; Riley and Skirrow, 1975):

where DO is the oxygen concentration in the water and DO_s is the corresponding saturation concentration.

3 RESULT AND DISCUSSION 3.1 Hydrographic characteristics

As a result of interactions among monsoon activities, submarine topography, river discharge, and the Kuroshio intrusion, the ECS is a quite complex seasonal circulation system (Chen, 2009; Yuan and Hsueh, 2010). When spring comes after winter, owing to the weakened northeast monsoon and increased rainfall in the eastern China mainland, the Changjiang Diluted Water (CDW) starts strengthening and turns southeastward, the Zhe-Min Coastal Current (ZMCC) still flows southward, and the Taiwan Strait Water (TSW) starts flowing through the Taiwan Strait (Fig. 1a) (Yang et al., 2013). Therefore, the Taiwan Warm Current (TWC) flowing northward is composed by the TSW and the Kuroshio Branch Current (KBC) originated from the upwelling Kuroshio water that ascends into the ECS shelf.

Some physical parameters (T, S, and σ_t) were strongly influenced by the aforementioned currents, and their distributions (Fig. 2–4) show specific hydrodynamic characteristics in the study area. For example, in the vertical profiles of T, S, and σ_t in the Kuroshio water east of Taiwan Island (Fig. 4), T decreased and σ_t increased with depth, while S decreased after a short increased and then increased again with depth. The remarkably similar variations of T, S, and σ_t between the stations located in the Kuroshio indicate that the Kuroshio water had layered structure. As shown in the T-S diagram (Fig. 2), the Kuroshio waters can be categorized into 4 types according to the changing trend of T/S/σ_t: the Kuroshio Surface Water (KSW, 0–75 m), with the highest T, high S and the lowest σ_t (averaged T, S, and σ_t in this water mass was 25.68℃, 34.62 and 22.85 kg/m³, respectively, in the same manner hereinafter); the Kuroshio Subsurface Water (KSSW, 100–300 m), with relatively lower T, the highest S and low σ_t (19.12℃, 34.75 and 24.74 kg/m³); the Kuroshio Intermediate Water (KIW, 400–800 m), with much lower T, the lowest S and high σt (7.65℃, 34.33 and 26.77 kg/m³); and the Kuroshio Deep Water (KDW, 1 000–2 000 m), with the lowest T, low S and the highest σ_t (2.86℃, 34.53 and 27.52 kg/m³) (Ichikawa and Chen, 2000; Liu et al., 2000; Zuo et al., 2016). These different water masses, which were categorized in terms of the indexes of temperature and salinity, also had different levels of DO (Fig. 2).

For the ECS continental shelf area (depth≤200 m), the horizontal contour graphs of T, S, and σ_t (Fig. 3) exhibit evident regional distribution characteristics. In the surface layer, the low values for T, S, and σ_t were observed in the north and nearshore part, while high values (T>24℃, S>34, σ_t>22 kg/m³) were observed in the southeastern part. In the 30-m layer, low T and S, and high σ_t were observed in the marginal zone of the study area, and high T, S, and low σ_t water (T>22℃, S>34, σ_t < 23.5 kg/m³) was found in the south and middle zones. In the bottom layer, a cold, saline, and dense water (T < 20℃, S>34.2, σ_t>24.5 kg/m³) occupied the major part, while water with higher T, lower S and σ_t was only observed in the rest coastal zone. Besides, a low T and high σ_t center in the northeast of Taiwan Island were observed in the 30-m layer contour map. Such regional distributions indicated that the cold, fresh, and less density riverine water had a strong influence in the coastal area, but the influence was gradually reduced as the distance away from the coast and the depth increased. Whereas the KSW, which had higher T, S, and σ_t than the riverine water, intruded northwestward into half a part of the ECS continental shelf in the surface layer; and the upwelling Kuroshio deeper water, which had lower T and higher σ_t than surrounding waters, combined with the KSW to form the KBC intruding into the ECS. The DO contents also varied with the different hydrographic characteristics of water (See details in the following Section 3.2).

3.2 DO distribution and the Kuroshio influence

DO concentrations in seawater can be influenced by the joint effect of many factors, such as temperature, salinity, photosynthesis, respiration, consumption of chemical reactions, internal transport of seawater and exchange with the atmosphere. Although no hypoxia (DO < 2 mg/L) was observed in the ECS shelf area during the survey done in late spring, a small low-DO (DO < 3 mg/L) zone was found in the northwest corner of the study area (Fig. 4). Based on the horizontal and vertical distributions of DO in the ECS and its adjacent Kuroshio waters, we discussed the corresponding influencing factors in the study area and tried to find out how the Kuroshio water can influence the hypoxia zone off the Changjiang River estuary.

3.2.1 Vertical profiles of DO in the Kuroshio

The KC water column east of the Taiwan Island was characterized by a permanent stabilized Oxygen Minimum Zone (OMZ) at ~800–1 000 m (Fig. 4). Water column stratification weakens from the KC to Station TW0-1 (Fig. 4), as a result of the decreasing influence of upwelling. The OMZ also decreases with depth from stations in the KC (~800–1 000 m) to Station TW0-1 (500 m). As stated before, the KC water has nearly homogeneous water properties and could be divided into 4 types. The density and AOU of Kuroshio water generally increased with depth through the whole water column, while DO concentration distributed in a different way. The highest DO concentrations were found in the KSW, varied from 6.32 to 7.17 mg/L, with an average of 6.72 mg/L, owing to the exchange process across the sea-air interface, diffusion process in the water, and photosynthesis process (Zhang et al., 2010). The concentrations of DO in the KSSW were slightly lower than the KSW, ranged from 5.59 to 6.97 mg/L, with an average of 6.27 mg/L. Under the KSSW, DO concentration decreased sharply until around 800-1 000 m depth with a minimum concentration of around 2.61 mg/L, which mainly because of the consumption of organic matter decomposition process and the lack of DO supplement from the upper layers. As for the following increase tendency of DO under the 1 000 m, it was related to the global thermohaline circulation (Rho et al., 2016). DO concentrations in Station TW0-1 (122.59°E, 25.49°N) were relatively lower than that in corresponding layers of Kuroshio seawater, strongly influenced by the upwelling KC water.

3.2.2 Horizontal distributions of DO

In the surface layer, DO concentration had the highest values (DO>8.2 mg/L) in the ECS coastal area (T < 22℃, S < 31 and σ_t < 21 kg/m³), where also featured by oxygen oversaturation (DO%>110%) and low AOU values (AOU < -30 μmol/kg) (Fig. 5), indicating that intensively photosynthetic activity took place in this area and released large amounts of oxygen. Whereas the relatively low DO, DO% (DO: 6.6–7.4 mg/L, DO%: 95%–105%) and high AOU (-15–15 μmol/kg) were observed in the southeastern part (T>24℃, S>34, σ_t>22 kg/m³), where was strongly influenced by the KSW, indicating that relatively weak photosynthetic process and/or relatively strong DO consumption processes took place in this area.

The content and distribution of DO in the 30-m layer were different from those in the surface layer. High DO, DO%, and low AOU (DO>6.6 mg/L, DO%>95% and AOU < 15 μmol/kg) in this layer, a result of weak photosynthetic and DO consumption processes, were observed in the warm, saline water tongue and the Kuroshio water area. While the relatively low DO, DO%, and high AOU (DO < 5.8 mg/L, DO% < 80% and AOU>60 μmol/kg) in the coastal area represented strong DO consumption process. Besides, a low-value center (DO < 6.5 mg/L) was found in the northeast of Taiwan Island, exactly the Kuroshio upwell zone.

According to previous literature, the KSW, KSSW, and KIW could ascend up along the continental shelf, but the KDW could not intrude into the ECS shelf (Chen et al., 1995; Zhou et al., 2015). Therefore, data of the core layer in the KIW (500 m) were used to analyze the bottom distributions in the study area. There were two low-DO centers in the bottom layer. One was in the Kuroshio water (DO < 4.6 mg/L, DO% < 45% and AOU>160 μmol/kg), indicating that the upwelling of the Kuroshio water is strong in spring. In addition, another was in the northeast coastal area with a minimum value of 2.63 mg/L at Station DH5-1a. Since the salinity at DH5-1a bottom water was 34.49, which must be influenced by the Kuroshio intruded water, this low-DO center probably was influenced by the Kuroshio to a certain extent.

3.2.3 Influence of the Kuroshio on DO distribution in the ECS

In order to analysis the influence mechanism and degree of the Kuroshio intrusion on the DO distribution in the ECS, some typical sampling stations were chosen for the following discussion. According to the results of Liu et al. (2017), which used Ba as a tracer to show the intrusion of Kuroshio water onto the ECS shelf, the Kuroshio subsurface water can ascend and flow northwestward along the middle shelf bottom till as far as the Qiantang River estuary during the period of investigation. Therefore, five stations (DH5-1, DH4-1, DH7-2, DH9-4, and TW2-2) within the region of Kuroshio intrusion were chosen to discuss the vertical distribution of DO and the impacts of the Kuroshio water (Fig. 6), with depth of 34 m, 55 m, 70 m, 90 m, and >2 000 m, respectively. Station TW2-2 represented the typical character of the Kuroshio water, with near-constant values of DO, DO%, and AOU in the upper water shallower than 100 m. The low Chl a value, the DO% around 100% and the AOU around 0 μmol/kg indicated that the photosynthetic and DO consumption processes in the Kuroshio upper water were relatively weaker than those in the ECS water.

For Station DH9-4 and DH7-2, seawater in the upper layers was influenced by warm, saline and lowdensity KSW which was featured by warm, saline, and low density, and in the lower layer were influenced by the cold, high salinity and density intruded Kuroshio water (Fig. 6). The differences of the vertical profiles of T, S, and σ_t between the upper and lower layers induced slight stratification structure in the water column. Except for the difference observed in temperature and salinity profiles, the values of DO, AOU, DO%, Chl a, and pH in Station DH9-4 and DH7-2 were basically same as that in Station TW2-2, which indicates that the waters at Stations DH9-4 and DH7-2 were highly similar to that of Station TW2-2. In other words, waters at Station DH9-4 and DH7-2 mainly come from the Kuroshio, which consistent with the results of Liu et al. (2017). His calculation showed that more than 85% of water in DH9-4 and DH7-2 station came from the Kuroshio.

For Station DH4-1 and DH5-1, their DO contents represent the DO characteristics in the ECS coastal water, which could suggest the influencing factors for the appearance of low DO content in the coastal water off the Changjiang River estuary. First of all, the vertical profiles of T, S, and σ_t indicated that the water column was strongly stratified, with surface water influenced by low S, σ_t river water and bottom water influenced by high S, σ_t seawater originated from the intruded Kuroshio water (Chen, 2008; Yang et al., 2013). The stratification structure restricted water exchange in the vertical direction, and at the same time restricted oxygen supply from the surface layer and air-sea interface to the bottom water (Chen et al., 2007; Zhang et al., 2010; Zhu et al., 2016). Secondly, the high S, σ_t seawater with high DO and low AOU values from the KSW and KSSW intruded into the ECS and advanced to nearshore along the seafloor (Fig. 3). At last, the over-loading nutrients discharged by the Changjiang River and the rich nutrients transported by the intruded Kuroshio water results in increased primary productivity in the coastal water, followed by the increased organic matter in the deep part of the water column (Zhang et al., 2010). The oversaturated and AOU-negative surface water indicated intensive primary productivity took place in the nearshore surface water (Wang et al., 2017), then a large quantity of organic matter was transported to the bottom and this organic matter consumed plenty oxygen through biochemical processes. Therefore, DH5-1 and DH4-1 had high values of DO in the surface layer and had low values in the bottom later. In other words, by means of stratification, low background DO value, and rich nutrients, the intruded Kuroshio water made a great contribution to the appearance of low DO content in the coastal water off the Changjiang River estuary.

In conclusion, DO in the KC basically stays in a constant distribution pattern, while DO in the ECS is more changeable and strongly influenced by various factors including the Kuroshio intruded water.

3.3 Impacts of Kuroshio intrusion on DO budgets and variations in the ECS

Since DO contents in the KC were relatively constant than those in the ECS shelf water, we tried to evaluate the impact of the intruded Kuroshio water to the DO pool in the ECS shelf area.

For the intruding Kuroshio waters in spring, the ratio of the water flux among the KSW, KSSW, and KIW is 3:3:1, and KIW comprises a fairly low proportion (~14.3%) in the intruded Kuroshio water (Chen, 1998). According to our previous study about the water fluxes budgets in the spring for the ECS based on box models for the conservation of water and salt masses, we found that the water fluxes of the KSW, KSSW, and KIW are 0.781, 0.781, and 0.260 Sv (1 Sv=10⁶ m³/s), respectively (Zuo et al., 2016). As mentioned above, the DO in KSW and KSSW was close to saturation (Fig. 5). The high DO, low AOU, and high nutrients Kuroshio water (KSW and KSSW) could continuously transport 1.01×10¹⁰ mg/s DO into the East China Sea in spring, which accounted for 90.1% of the total input DO from the intruded Kuroshio water. Since the outflow water flux for the ECS shelf, which includes both the outflow flux through the Tsushima Strait (2.85 Sv) and the outflow flux across the ECS shelf slope, is 3.52 Sv (Zuo et al., 2016), the outflow water flux across the ECS shelf slope is 0.67 Sv. Then take the average value of the surface DO concentrations in the stations farthest from the shore as DO concentration of the outflow ECS water, the ECS exports 4.91×10⁹ mg/s DO to the open sea. Therefore, owning to the Kuroshio intruded water, 6.32×10⁹ mg/s DO is imported into the ECS across the ECS slope in spring.

According to the relationship between Carbon and Oxygen during photosynthetic process (106C–138O₂) (Stumm and Morgan, 1996) and the primary production in the ECS (Gong et al., 2003; Zuo et al., 2016), 1.39×10¹⁰ mg/s oxygen is produced through photosynthesis process in the ECS area. Besides, the O₂ fluxes across the sea-air interface during the cruise in spring 2014 were calculated according to O₂ exchange flux formulas in previous studies (Liss and Merlivat, 1986; Stigebrandt, 1991; Wanninkhof, 1992; Wanninkhof, 2014). Oxygen was generally released into the atmosphere through the sea-air interface during the survey, with an average flux of 0.03 g/(m²·d) (i.e. 1.91×10⁸ mg/s) in the ECS. In this case, the DO imported into the ECS by the Kuroshio intruded water plays an important role in the ECS DO pool.

More importantly, according to the study of Liu et al. (2017), the KSSW could ascend and northwestward along the middle shelf bottom till as far as the Qiantang River estuary, where the Kuroshio water still accounted for 65% of the water volume in the spring of 2014. This result fully illustrates that the KSSW could reach the hypoxic region of the East China Sea, and the dissolved oxygen in the KSSW might affect the DO content and the distribution of the hypoxic region.

Figure 5 showed that there was a low dissolved oxygen zone near the Hangzhou Bay during the period of investigation. The low-DO zone should be part of the hypoxic region in the East China Sea. The hypoxic region in the East China Sea is one of the focus problems in the current study of China's offshore waters, which have obtained many results about it. The hypoxic region began to form in the spring and reaches its peak in the summer. However, its position, region, and intensity fluctuated every year, and the reason for the fluctuation was still unsettled. Based on the result of this investigation, the hypoxic region is obviously affected by the invading water of the Kuroshio. At the station DH5-1, the DO concentration in the bottom water is 2.63 mg/L, and the Kuroshio intruded water was accounted about 55% of the water volume. If there is no dissolved oxygen from the Kuroshio invasion, the DO would be consumed to zero. Therefore, the dissolved oxygen from the Kuroshio invasion slowed down the intensity of the hypoxia. If there is no dissolved oxygen from the Kuroshio invasion, a large area of Hypoxia Zone will be formed outside the bay of Hangzhou Bay during a hypothetical investigation.

Kuroshio intrusion could be used to illustrate the fluctuations in the position, range, and intensity of the East China Sea Hypoxia Zone. Because the region and intensity of the Kuroshio intrusion were different in different seasons, dissolved oxygen from the Kuroshio invasion would be more or less. Since the influence of Kuroshio intrusion on the hypoxic region is not stable, which resulted in the position, region, and intensity of hypoxic region change every year. In winter, the front of the Kuroshio could extend to the continental shelf, a large amount of dissolved oxygen brought by the Kuroshio invasion replenish the consumed dissolved oxygen in time, which is not conducive to the formation of the low oxygen zone. In summer, the Kuroshio flows mainly along the edge of the shelf, little of the dissolved oxygen from the Kuroshio intrusion was transported to the interior of the shelf, so the intensity of hypoxic region is the strongest.

In addition, in coastal environments, high concentrations of nutrients, and warm environment lead to strong primary production and oxygen production. The Kuroshio intruded water just contains high-content phosphate and complemented to the riverine nitrate-surplus water. Then high primary productivity also induce low DO due to local or large scale oxygen demand from the microbial decay of sinking organic particles (Zhang et al., 2010). Moreover, warming environment leads to more severe stratification phenomenon in coastal water. Therefore, primary production and subsequent decomposition of organic matter naturally change DO and AOU. Strengthening stratification of seawater under global warming and increased nutrient fluxes from terrestrial sources are expected to exacerbate coastal hypoxia phenomenon (Lui et al., 2014). However, recent studies suggest that DO in deeper seawater declined as a consequence of Global Warming (Zhang et al., 2010; Mostofa et al., 2016). As for the Kuroshio water, compared with previous survey data (Chen et al., 1995; Lui et al., 2014; Kodama et al., 2015), the DO content in the Kuroshio waters has indeed decreased in the KIW but increased in the KSSW (Lui et al., 2014). Under such circumstances, the influence of the Kuroshio intrusion on the East China Sea Hypoxia Zone will be more complicated. More effort is needed to study the impact of the Kuroshio intrusion on the East China Sea Hypoxia Zone.

4 CONCLUSION

The horizontal and vertical distributions of DO concentration in the ECS and Kuroshio waters were investigated. In the Kuroshio Current, DO content was the highest in the euphotic layer, then decreased sharply with depth to about 1 000 m, and increased with depth gradually thereafter. While in the ECS continental shelf area, DO content had high values in the coastal surface water and low values in the nearbottom water. In general, DO content had high values in the coastal surface layer and low values in the Kuroshio deep water and the nearshore bottom water. Such spatial distribution was influenced by many factors, including mainly temperature, salinity, primary productivity, remineralization process, movement and mixing of water masses, and sea-air exchange.

A low-DO zone off the Changjiang River estuary was found in spring 2014, and it was formed under the combined influence of many factors. The influence factors, such as stratification of the water column, high deposition of organic matter along with high primary productivity, and mixing along with transport of oceanic current waters, can further cause hypoxia in the ECS coastal area. What is noteworthy is that compared with previous studies, the intrusion of KSW and KSSW brought a large amount of dissolved oxygen onto the continental shelf and to the coastal area, which could alleviate the oxygen deficit phenomenon in the ECS. That is, the Kuroshio intruded water with rich dissolved oxygen can impact the position, range, and intensity, thus the formation/ destruction of the ECS Hypoxia Zone. But more research is needed to quantificationally analyze the impact degree of the Kuroshio in future.

5 DATA AVAILABILITY STATEMENT

The datasets generated and/or analyzed during the current study are available from the corresponding author on reasonable request.

6 ACKNOWLEDGEMENT

The authors gratefully acknowledge LIU Yang for his assistance in the field. We also appreciate the valuable comments on the manuscript by the editors and anonymous reviewers.