Hypoxia disaster in waters adjacent to the Changjiang estuary

Hypoxia disaster in waters adjacent to the Changjiang estuary Xiaofan Luo1 * , Hao Wei2 * , Renfu Fan2, Zhe Liu1, Liang Zhao1 1 College of Marine Science and Engineering, Tianjin University of Science & Technology, Tianjin, 300457, China 2 School of Marine Science and Technology, Tianjin University, Nankai District, Tianjin, 300072, China Correspondence to: Hao Wei (weihao@ouc.edu.cn) 5 *These authors contributed equally to this work and should be considered co-first authors.


Introduction
Dissolved oxygen (DO) is important for marine life.Low concentration of DO (e.g., <3.0 mg L -1 ) causes hardship and even threats for most species living in the ocean.In coastal waters, increasing occurrence of hypoxia (DO <2.0 mg L -1 ) is becoming a global environmental issue (Diaz and Rosenberg, 2008;Conley et al., 2009).One of the main goals of regional environment management is to reduce the area and volume of hypoxia, and this requires understanding and prediction of low-oxygen evolution (Feng et al., 2012).It is generally agreed that stratification and organic matter degradation are main reasons for the formation of hypoxia.However, stratification is influenced by different physical processes; and for hypoxia in estuarine regions, there have been continuous debates on the roles played by the riverine nutrient loads and freshwater discharge (Bianchi et al., 2010).
Since the middle of last century, seasonal survey has started for hypoxia in the lower water column adjacent to t he Changjiang estuary (Fig. 1), an area with a high primary production (Gu, 1980;Chen et al., 1988;Tian et al., 1993;Zhao et Nat. Hazards Earth Syst. Sci. Discuss., doi:10.5194/nhess-2016-59, 2016 Manuscript under review for journal Nat.Hazards Earth Syst.Sci.Published: 28 April 2016 c Author(s) 2016.CC-BY 3.0 License.and from August to September in 2009, Zhu et al. (2015) emphasized the influence of pycnocline on the spatial variation of hypoxia.Through combining reanalysis of data from 2006 to 2007 with results of an ocean circulation model, Wei et al. (2015b) noted possible influence of seasonal circulation condition on the position of hypoxia off the Changjiang estuary.
Previous studies have reached a consensus that the movement of water masses and changes of stratification induced by ocean circulation variations are all responsible for the formation and development of hypoxia.However, there lacks agreement on which factor, among the TWC, the CDW, fronts and the Yellow Sea Water (YSW), plays the dominant role.It is possible that each previous study captured one aspect of hypoxia formation and variation in a specific area, but the conclusion was limited by the particular dataset analyzed.For instance, the hypoxia center in summer of 2006 was located at a northern position relative to 1959, but did this northward shift occur frequently in recent years?Is it generally true tha t from beginning to the end of summer, the hypoxia center appeared successively at the southern, northern, then back to southern locations?What is the exact evolution of hypoxia from formation to decay?Besides the blocking of oxygen exchange by stratification, fronts and ocean eddies, how are DO distribution and hypoxia evolution influenced by the lateral transport of the TWC accompanied with the Kuroshio Subsurface Water (KSW)?What role does the YSW play?
From a synthesis of observations of hypoxia reported in literatures and made within the recent decade (Fig. 1 and Table 1), significant year-to-year variations in the location and evolution of hypoxia can be identified.This study attempts to further explore the influence of changes in hydrodynamic condition on the evolution of hypoxia through analyzing new observations carried out in June, August and October in 2012 and monthly from May to September in 2013.According to hydrodynamic condition and topography features, we divide the waters adjacent to the Changjiang estuary into the southern and northern sub-regions (Fig. 1).The southern sub-region includes the offshore water of Zhejiang, the submarine river canyon and the middle shelf of the ECS.The northern region is the Changjiang Bank.The low-oxygen/hypoxia centers in the two years are mapped out from the observed bottom DO distribution.Distributions of stratification and hydrography are described.By linking the DO distribution to variations of hydrodynamic condition, we analyze the relative contributions of the TWC, the CDW and the YS tidal front in the two sub-regions, with a special attention paid on the influence of the lateral advection of the KSW.

Field observations and data processing
Concentrations of DO and hydrographic data are obtained from 10 cruises carried out in June, August and October i n 2012, and monthly from May to September in 2013 (Table 2).The data for May and July in 2013 is a merge of observations made from two ships in each month.RBR 620-CTD with Seapoint sensor of DO was used in all the 10 cruises to measure profiles of water temperature, salinity, DO concentration.SBE-CTD was deployed in 8 cruises excluding the survey on Subei Shoal in 2013 and that within the Changjiang estuary in July of 2013.When both RBR 620-CTD and SBE-CTD took measurements, the profiles of water temperature and salinity from SBE-CTD were used because of the higher sampling Data recorded during descending profiling are used.For each profile, the inverse pressure calibration was applied and spikes were removed as the basic data quality control.For each cruise that covered the whole observational grids, about 100 water samples from surface, middle and bottom layers were taken from stations along selected sections.Salinity of these samples was measured by the SYA2-2 salinometer in laboratory to calibrate the salinity from the CTD measurement in each cruise.The DO sensor on RBR620-CTD was calibrated prior to each cruise as the following.It was firstly put into pure water aerated by oxygenate pump for 60 minutes to get the voltage for DO reaching 100 % saturation, and was then put into the anoxia water (i.e., saturated sodium sulfite solution) to get the voltage for zero DO reading.The DO concentration was determined according to the linear scaling between these two voltages.During some cruises, the chemistry group also used the Winkler Titration method to measure the DO of water samples from different layers at selected stations.This enabled an alternative calibration of the DO measured by the DO senor.Data was sampled at high frequencies, for instance, 24 Hz for the SBE-CTD and 8 Hz for the RBR-620 CTD.The raw data was averaged over 0.2 m vertical intervals.The buoyant frequency (N 2 ) was calculated from water temperature and salinity.The maximum N 2 in a profile was used to represent the stratification intensity of the water column at a particular station.
Stations to the south of 33° N are chosen for analyzing the hypoxia off the Changjiang estuary (Fig. 1).Sampling grids for each cruise and sections to be discussed in following sections are shown in Fig. 2. According to Su et al. (1996), salinity can be used to identify water mass in the ECS, i.e., S <30.0 representing the diluted water, 34.5> S >34.0 for the TWC water, and S >34.5 for the KSW.The area with bottom temperature in the range of 13.0-17.0℃ stands for the summer tidal front of the YS (Zhao, 1985).According to China's 1992 National Marine Investigation Standard by the State Bureau of Technical Supervision, the phycnocline is defined to exist if the vertical gradient of density is larger than 0.1 kg m -4 for the water depth less than 200 m.This equals to the maximum N 2 of a vertical profile larger than 1×10 -3.0 s -2 .

Bottom DO concentration in June, August and October, 2012
There was no hypoxia in June of 2012 (Fig. 3a).The bottom DO concentration was larger than 4.0 mg L -1 , with higher values in the eastern part and lower values in the western part of survey area.A band of relatively low DO concentration of 4.0 -5.0 mg L -1 was found along the 30 m isobath to the north of 29° N.
In August, there were only a few stations located in the southern area (Fig. 3b).The DO concentration was 2.0 -3.0 mg L -1 lower than that in June.Over the western side of the Changjiang Bank, i.e. near the Subei Bank, DO <3.0 mg L -1 was observed.A hypoxia center with DO = 1.0 mg L -1 appeared in the northwestern survey field (station I1, at 33° N, 122.5°E).
to the south of the Changjiang estuary though hypoxia disappeared.These two centers were located near the northern end of the submarine river canyon with DO = 4.8 mg L -1 (station PN1), and the offshore waters of Zhejiang with DO <4.0 mg L -1 , about 2.0 mg L -1 lower than that in the surrounding waters.
Overall, in summer of 2012 hypoxia off the Changjiang estuary was not severe.There was no hypoxia in June and October though the low-oxygen centers existed.The hypoxia center was located in the northern sub-region in August.

Bottom DO concentration from May to September, 2013
In May, the vertical and horizontal distributions of DO were fairly homogenous (Fig. 3d, vertical distribution was not shown) .
DO values were 8.0-9.0 mg L -1 in the near shore region and 7.0-8.0mg L -1 in the offshore region.The minimum DO value, located at station kb, was 6.7 mg L -1 .
In June, DO was lower than 4.0 mg L -1 in the southern sub-region.A hypoxia center appeared near the slope along the 30-50 m isobaths offshore of Zhejiang.DO values at most stations to the north of the Changjiang estuary were less than 5.0 mg L -1 (Fig. 3e).Along the 122.5°E section, the DO concentration in the lower water column over the Changjiang Bank was substantially larger than that to the south of 31° N. In the southern area, the hypoxia water had a thickness of about 10 m in June (Fig. 4b).DO in June of 2013 was about 3.0 mg L -1 lower than that in June of 2012 (Figs. 3a,e and Figs. 4a,b), and hypoxia occurred earlier in 2013.The occurrence of hypoxia in June was only reported in 2003 in the submarine river canyon near 30.45°N (Xu, 2005).
In July, bottom DO was larger than 7.0 mg L -1 in the Changjiang estuary and Hangzhou Bay, except that DO <4.0 mg L - 1 was found at stations X5 and X6 near the northern corner of the Changjiang river mouth (Fig. 3f).Over the Changjiang Bank, at all stations with water depth larger than 30 m, bottom DO was less than 3.0 mg L -1 ; severe hypoxia occurred within 32-32.4°N, 122.4-123.5°E over the western part of the bank.The huge area with DO <3.0 mg L -1 extended to southwest of Jeju Island.DO values were 3.0-5.0mg L -1 over the middle shelf of the ECS.Overall, the DO concentrations in the southern and eastern parts of the survey area were higher than that in the middle part.
In August, the survey area was smaller than that in July.Over the Changjiang Bank, the area of hypoxia was still large, and extended northward and eastward compared with July (Fig. 3g).In the southern sub-region, a low-oxygen center with DO <3.0 mg L -1 was located in the northern end of the submarine river canyon.DO distribution along section K (at 32° N) showed that the hypoxia water reached 30 m thickness from stations K2 to K6 over the Changjiang Bank, extending about 100 km in length (Fig. 4d).DO <2.0 mg L -1 was observed at five stations and hypoxia area was separated into several patches extending northward to 33° N and eastward to 125° E.Over the whole bank, DO was less than 4.0 mg L -1 , much lower than that in August 2012 (Fig. 4c).
In September, the DO concentration over the western bank increased quickly, while DO <3.0 mg L -1 was still found in the outer edge of the bank and hypoxia occurred near 33° N (Fig. 3h).Another area with DO <3.0 mg L -1 was in the river canyon to the south of the Changjiang estuary.In 2013, hypoxia first appeared in the southern sub-region in June and was sustained over the Changjiang Bank in July, August and September.Multiple low-oxygen centers appeared from June to September.In the southern sub-region, the hypoxia centers were stably located in the coastal water near the 30-50 m isobaths.In the northern sub-region, the positions of hypoxia centers changed around the bank.Overall, compared with historical reports, in the summer of 2013 hypoxia appeared earlier, occupied a larger area, with a larger thickness, and was maintained longer over the Changjiang Bank.

Climatological evolution of hydrodynamic conditions
The evolution of hydrodynamic conditions in the Yellow and East China Seas is related to monsoon, Kuroshio, runoff of the Changjiang River, and bathymetry of the region (Su, 2001).
In each year, stratification starts to develop in May corresponding to surface heating.The tidal front in the YS app ears in May and gets intensified gradually in the following three months.This front can extend southeast-ward across the Changjiang Bank.In summer, the current along the western coast of the YS (within 20 m isobath) flows northward as a response to the wind (Liu and Hu, 2009).Both tidal front and northward coastal current are main obstacles for supplement of the YSW to the low-oxygen water over the Changjiang Bank.
The CDW extension is mainly influenced by the advancement and recession of the TWC, especi ally its inner branch (Weng and Wang, 1985;Wei et al., 2015a).When the TWC intrudes further north over the continental shelf, the CDW tends to veer northeastward substantially.A strong halocline over the Changjiang Bank is created by the CDW and the water in lower layer.The combination of halocline and thermal stratification blocks the oxygenate aeration of the lower layer from the surface.
The TWC consists of Taiwan Strait water in the upper layer characterized with high temperature and moderately hig h salinity, and the KSW in the lower layer with low temperature and high salinity (Weng and Wang, 1985).The pycnocline is formed in the region passing through by the TWC, and is also influenced by the coastal fresh water to the south of the Changjiang estuary.The KSW can be clearly identified by the characteristic oxygen concentration (4.5 mg L -1 ).Hence the KSW provides the relatively low and high oxygen water before and after the formation of low -oxygen center (DO <3.0 mg L -1 ), respectively.The TWC has two branches over the continental shelf.In May, the TWC is reinforced by the southerly wind.The near shore branch of the TWC flows northward along the 50-60 m isobath offshore of Zhejiang into the submarine river canyon, and in July it can reach the northern corner of the Changjiang river mouth.This branch upwells to the sea surface in the coastal area to the south of the Changjiang estuary.The outer branch of the TWC can reach the Changjiang Bank along the middle shelf of the ECS.It flows along the edge of the bank in May and June, across the bank in July and August, along the edge again in September and October.Eventually, the TWC runs into the Tsushima Strait (Su, 2001).
In summary, the hydrodynamic condition, especially the TWC and the CDW, plays a leading role in the evolution,

Bottom DO, KSW intrusion and stratification in the southern sub-region
In the southern sub-region, the appearance of low-oxygen centers was in consistent with the distribution of strong stratification (Fig. 5), and the evolution of DO concentration can be influenced by the lateral transport of the KSW with DO = 4.5 mg L -1 .
A low-oxygen center often first occurs in the confluent area of coastal current and the TWC, which is usually located between 30 m and 50 m isobaths where stratification is stronger and more persistent than that in the middle shelf of the ECS.
In June and August of 2012, because of the intensive KSW intrusion toward the northern corner of the mouth of the Changjiang (Figs. 6a, b), hypoxia did not happen despite of the presence of strong stratification (N 2 >10 -2.5 s -2 ) (Figs. 5a, b).
In October of 2012, owing to the TWC recession and enhanced monsoon, stratification was overall weak and DO increased (Figs. 3c,5c,6c and Table 3).
From May to September of 2013, the southern coastal area was strongly stratified (Figs.5d-h and Table 3).In May, the persistence of stratification was insufficient for oxygen consumption, resulting the presence of relatively high DO concentration.In June, the southern branch of the CDW expanded more widely relative to June 2012 (Figs.7a, e), and the TWC reached the latitude of 30.5°N (Fig. 6e).This led to a large N 2 values (N 2 >10 -2.0 s -2 ) in the coastal water to the south of the Changjiang estuary (Fig. 5e).With the presence of stratification, DO was rapidly consumed and a large low -oxygen area formed.In the meanwhile, the KSW was located to the south of 28° N (Figs.6e, 8e) and could not provide the relatively DOrich water via lateral transportation.As a consequence, the low-oxygen area developed into hypoxia (Fig. 3e).In July, with the KSW intruding northward and occupying the majority of bottom water in the southern sub-region (Fig. 6f), hypoxia faded away and did not re-appear afterward.Thus, the timely replenishment of the KSW could prevent the evolution of lowoxygen center into hypoxia.
The scatter plot of the bottom salinity versus DO concentration in the southern sub-region (Fig. 9a) shows that hypoxia did not occur when S >34.5; and when hypoxia occurred, S must be less than 34.5.This is consistent with our above analysis.
Evidences supporting this conclusion can be found in previous studies.Hypoxia occurred in the coastal area to south of the Changjiang estuary in August of 1959August of , 1976August of -85, 1981August of , 1999August of , 2002August of and 2006;;June of 2003;and September and October of 2006 (Table 1).Among these years, in 2006 hypoxia was maintained for three months while the KSW intruded at a southern location (Zou et al., 2008;Zhou et al., 2010;Wang et al., 2012), and in August 1959 hypoxia occurred in the northern end of the canyon where bottom salinity was 33.0 (Liu et al., 2012).No input of DO-rich water from upstream led to the generation of hypoxia in this sub-region.For the rest hypoxia cases, there was insufficient hydrological data to identify the locations of

Bottom DO, CDW spreading and stratification in the northern sub-region
In the northern sub-region, the low-oxygen centers were located where stratification was strong, and the pattern of hypoxia was influenced by the extension of the CDW.
In June 2012, both strong stratification and relative low-oxygen occurred in the western part of the Changjiang Bank (Figs.3a, 5a); and till August, strong stratification was maintained and DO decreased to less than 3.0 mg L -1 (Figs.3b, 5b).
The CDW veered northeastward substantially (Fig. 7b), and below it a salinity front was produced as the TWC encountered the coastal fresh water (Fig. 6b).To the east of salinity front, there existed a tidal front (Fig. 8b).The DO-rich YSW was transported to the eastern bank along this tidal front.Due to the lack of DO replenishment, the low-oxygen area to the north of salinity front in the western bank developed into hypoxia (Fig. 3b).In October, the CDW switched back to flowing southward along the coast of Zhejiang.As stratification disappeared (N 2 <10 -3.5 s -2 ), bottom DO increased to 7.0 mg L -1 (Figs.3c, 5c).
In summer of 2013, the low-oxygen center and hypoxia in the bank all showed correspondence with strong stratification (N 2 >10 -2.0 s -2 ) (Fig. 3 and Fig. 5).In July, the CDW expanded eastward, covering a zone spanning from the Changjiang estuary to Jeju Island.This facilitated the formation of a large low-oxygen area with DO <3.0 mg L -1 over the whole bank under strong stratification (Figs. 3f,5f,7f).In the meanwhile, the KSW and the YSW along the tidal front relieved the hypoxia in the southern and eastern bank, respectively (Figs. 6f, 8f).Hence, hypoxia events only happened in the western bank without DO supplement (Fig. 3f).In August, the KSW did not approach the mouth of the Changjiang River, but retreated southward (Fig. 6g).Meanwhile, the northward migration of the tidal front (Fig. 8g) limited the DO replenishment by the YSW.As a consequence, a sever hypoxia evolved in the area with low-oxygen and sustained stratification (Fig. 3g, 5g and Table 3).In September, over the Changjiang Bank the CDW shrunk, stratification was broken almost everywhere (Figs. 5h, 7h), and DO increased accordingly (Fig. 3h and Table 3).At the outer edge of the bank, however, stratification ( N 2 >10 -3.0 s -2 ) was still present because of the vertical temperature differences, low DO (<3.0 mg L -1 ) was maintained along the outer edge (stations I4, I5 and K6), and hypoxia occurred at the northeastern location (station I5).
The bottom DO concentration was in negative correlation with the strength o f stratification over the Changjiang Bank (Fig. 9b).A regression relationship was obtained as DO = -1.67 × Log10N 2 + 0.43，with r = -0.67 (significant at the 0.05 confidence level).Values of DO were always larger than 4.0 mg L -1 when stratification was very weak (N 2 <10 -3.0 s -2 ).
Hypoxia only happened under strong stratification, though high DO concentration could also exist under the same situation.
Thus, stratification is a necessary condition for hypoxia formation over the bank.
In summer, the CDW will turn onto the bank but the direction of its extension may vary.For instance, the CDW spread

Comparison of hypoxia development between the water adjacent to the Changjiang estuary and the Gulf of Mexico
Worldwide, the frequent occurrence of hypoxia is related to eutrophication (Conley et al., 2009).For Gulf of Mexico, a statistical model for hypoxia prediction has been established based on nutrient loads from Mississippi River (Turner et al. , 2006).For Chesapeake Bay, the prediction of hypoxia volume is related to river runoff (Scully, 2010).However, for waters adjacent to the Changjiang estuary, hypoxia can be influenced by many complicated factors.Here hypoxia is not directly related to runoff from river.In summer of 2006 the runoff was extremely low (Fig. 10) but severe hypoxia occurred (Zou et al., 2008;Zhou et al., 2010;Wang et al., 2012).In summer of 2012 the runoff was among the highest in the recent decade but only a small area of hypoxia existed in August (Figs.3a-c and Fig. 10).Compared with 2012, in 2013 the runoff was far more less but hypoxia occupied extremely large areas and sustained much longer (Figs.3d-h and Fig. 10).Clearly, the area or volume of hypoxia adjacent to the Changjiang estuary cannot be predicted based on river runoff.
Over the Texas-Louisiana shelf in the northern Gulf of Mexico, the formation and destruction of hypoxia is primarily a local vertical process, and the biological processes are responsible for producing hypoxia change from east to west (Hetland and DiMarco, 2008;Bianchi et al., 2010).Hypoxia is predominantly caused by water column respiration in the eastern region with steep shelf break, and by benthic respiration over the wide shelf to the west.Similarly, the hypoxia area adjacent to the Changjiang estuary can be divided into an area on the wide bank and an area with steep slope according to bathymetry.
Over the Changjiang Bank, low-oxygen is related to the presence of the CDW, while the CDW's position is influenced by the TWC (Wei et al., 2015a).Strong tidal mixing and monsoon may both induce the detachment of freshwater from the CDW (Xuan et al., 2012).This results in stronger changes of the river plume on the Changjiang Bank, compared with that on the west part of the Texas-Louisiana shelf where tides and wind are relatively weak.Although the low -oxygen centers to the south of the Changjiang estuary are constrained by bathymetry, the bottom DO concentration is significantly influenced by the shelf circulation including the TWC and shoreward intrusion of the Kuroshio.The lateral transport and vertical processes both control the hypoxia adjacent to the Changjiang estuary.The prediction of low -oxygen needs information about salinity evolution at the bottom to the south of the Changjiang estuary and at the surface over the Changjiang Bank, that defines the extensions of the KSW and the CDW, respectively.

Conclusions
As one of the major causes of ecological disasters in the coastal seas, especially in large river estuaries, hypoxia is becoming a global environmental issue and has attracted wide attention in recent decades.Base on new observational data from 10 cruises carried out in 2012 and 2013, the distribution of dissolved oxygen and evolution of hypoxia in waters adjacent to the Changjiang estuary are studied.The linkage of summer hypoxia with hydrodynamic conditions, including variations of water with bathymetric feature, fronts exist throughout of the year in the northern end and block the DO exchange with water to the north.Thus lateral transportation is mainly from the south.The KSW (S >34.5) carried by the near shore branch of the TWC can substantially affect the bottom DO concentration in this region.Hypoxia is difficult to form if the KSW appears early in the middle shelf of the ECS and intrudes further northward in summer.In this region, hypoxia does not occur at locations with bottom S >34.5; and when hypoxia does occur, there must be bottom S < 34.5.b) In the northern sub-region, i.e. over the Changjiang Bank, there is a high possibility that hypoxia occurs from July to September.The distribution of bottom DO corresponds to the intensity of stratification, the two being in negative correlation with a regression relationship of DO = -1.67 × Log10N 2 + 0.43.DO is larger than 4.0 mg L -1 when stratification is weak (N 2 <10 -3.0 s -2 ).Over the bank, low-oxygen area is mostly related to the CDW spreading.The combination of the Subei coastal current (northward in summer), tidal front of the YS and the bathymetry of the bank restrains DO replenishment from surrounding waters.Under the condition of sustained stratification, a low-oxygen center may involve into hypoxia.Generally, the coastal current transports the YSW to the Changjiang Bank after the transition of monsoon in autumn, and in the meanwhile stratification has already disappeared.Thus, water from the YS may play a minor role in the relief of hypoxia in summer.
In summary, based on new observations in 2012 and 2013, especially in 2013 with monthly data from May to September, it is evident that hypoxia in waters adjacent to the Changjiang estuary is substantially influenced by shelf circulation, including the CDW plume, the TWC and the KSW intrusion.The hypoxia evolution in 2013 was distinctly different from that in 2006 and 1958-1959.Due to the lack of monthly observation covering the same region in different years, credible model simulations of inter-annual variations of hydrodynamic conditions are needed for further investigating the linkage of hypoxia with the KSW intrusion and the CDW spreading.It is hopeful that the prediction of hypoxia can be made according to the evolution of surface and bottom salinity.September in 2013).The layout is designed to ease comparison between 2012 and 2013 for June and August.Black, green 435 and blue bold lines denote section at 122.5°E and section K in Fig. 4, and section PN in Fig. 5, respectively.
Nat. Hazards Earth Syst.Sci.Discuss., doi:10.5194/nhess-2016-59,2016 Manuscript under review for journal Nat.Hazards Earth Syst.Sci.Published: 28 April 2016 c Author(s) 2016.CC-BY 3.0 License.duration and intensity of stratification in waters adjacent to the Changjiang estuary.According to laboratory experiments, Liu et al. (2012) proposed that the main reason for hypoxia formation was not oxygen consumption b ut the lack of replenishment after consumption.Hence hypoxia may occur in any area if stratification is sustained over a considerable duration without DO replenishment.
northward in June 2012, and eastward in July 2013.The largest values of N 2 were reached when the CDW lay over salty waters.Overall, hypoxia usually formed in the northwestern Changjiang Bank when the CDW extended northwestward, and Nat.Hazards Earth Syst.Sci.Discuss., doi:10.5194/nhess-2016-59,2016 Manuscript under review for journal Nat.Hazards Earth Syst.Sci.Published: 28 April 2016 c Author(s) 2016.CC-BY 3.0 License. in the eastern part when the CDW spread eastward substantially.
Nat. Hazards Earth Syst.Sci.Discuss., doi:10.5194/nhess-2016-59,2016 Manuscript under review for journal Nat.Hazards Earth Syst.Sci.Published: 28 April 2016 c Author(s) 2016.CC-BY 3.0 License.mass and stratification associated with circulation, is explored.The study area can be divided into the southern (south of Changjiang estuary) and northern (the Changjiang Bank) sub-regions.The mechanisms dominating the evolution of stratification and DO distribution are different in these two regions.a) In the southern sub-region, a confluent area of coastal current and TWC near the 30-50 m isobaths is where a lowoxygen center usually exists, and hypoxia may occur from June to October accompanied by strong stratification.Associated

bFigure 1 .
Figure 1.Topography and schematic map of summer circulation in the Yellow Sea and East China Sea.Black arrows denote 425 circulation including the Yellow Sea Current (YSC), Subei Coastal Current (SBCC), Taiwan Warm Current (TWC), and Kuroshio intrusion (KBC: surface Kuroshio Branch Current; O-KBBC: offshore Kuroshio Bottom Branch Current; N-KBBC: Nearshore Kuroshio Bottom Branch Current, i.e.Kuroshio Subsurface Water).Contours in gray denote isobaths of 30, 50 70, 100, and 200 m.Various symbols represent sampling stations of different cruises labelled in the right pane that observed the occurrence of hypoxia.Red line marks the boundary between southern and northern sub-regions.430

Figure 3 .
Figure 3. Dissolved oxygen concentration (mg L -1 ) of bottom water in the same layout as Fig. 2. 440

Figure 6 .
Figure 6.Same as Fig. 3 except for bottom salinity (in psu).The contour of S = 30.0 in blue represents the extension of the Changjiang Diluted Water; S = 34.0 in magenta represents the Taiwan Warm Current; and S = 34.5 in red indicates the intrusion of Kuroshio Subsurface Water.455

Figure 8 .
Figure 8. Same as Fig. 3 except for bottom temperature (in °C).The 18°C contour characterizes the Kuroshio Subsurface Water and is emphasized in bold.Shading area denotes frontal zone.

Figure 9 .Figure 10 .
Figure 9. Dissolved oxygen concentration (in mg L -1 ) of bottom water versus (a) bottom salinity (in psu) in the southern subregion, and (b) maximum squared buoyancy frequency at logarithmic scale (N 2 : s -2 ) in the northern sub-region.

Table 2 .
Information of cruises analyzed in this study.

Table 3 .
Regional mean temperature at the bottom (Tbottom: in ℃), salinity in the surface (Ssurface: in psu), salinity at the bottom (Sbottom: in psu), maximum squared buoyancy frequency at logarithmic scale (N 2 max: in s -2 ), dissolved oxygen concentration of bottom water (DO: in mg L -1 ) and the number (N) of samples in areas with frequent occurrence of hypoxia.