RESEARCH ARTICLE10.1002/2017JC012754

Impact of typhoons on the Changjiang plume extension in theYellow and East China SeasJoon Ho Lee1 , Il-Ju Moon2 , Jae-Hong Moon1 , Sung-Hun Kim3, Yeong Yun Jeong2 , andJun-Ho Koo4

1Deparment of Earth and Marine Sciences, Jeju National University, Jeju, South Korea, 2Typhoon Research Center, JejuNational University, Jeju, South Korea, 3National Typhoon Center, Korea Meteorological Administration, Jeju, South Korea,4National Institute of Fisheries Science, Gijang, Busan, South Korea

Abstract It is well known that river discharges, winds, ocean currents, and tides are major dynamical fac-tors that determine the distribution and extension of the Changjiang plume (CP) in the Yellow and EastChina Seas (YECS). Using observations and numerical experiments, this study demonstrates that, in additionto these factors, typhoons in the YECS also play a crucial role in the extension of the CP during the summerseason. The hydrographic data observed at the Ieodo Ocean Research Station (IORS) and by a research ves-sel during the period of Typhoons Ewiniar (0603) and Dianmu (1004) showed that the typhoon-inducedstrong vertical mixing modified spatial distribution of the CP significantly, resulting in the delay of the CP’sextension by as much as up to 20 days. A series of numerical experiments for Typhoon Dianmu also showedthat the typhoon plays a blocking role for the extension of CP for up to 17 days through the vertical mixingprocess and the change of background winds. In particular, it is found that the delay due to Dianmu in 2010contributed to the avoidance of potential mass mortality of marine life by preventing the low-salinity waterfrom spreading to the aquaculture regions near Jeju Island.

1. Introduction

The Changjiang (also known as the Yangtze River) transports a large amount of fresh water into the Yellowand East China Seas (YECS). It is a major nutrient source for the YECS, which is one of the most favorablefishing grounds in the world [Chen et al., 2010]. By mixing with the ambient saline water, the discharge fromthe Changjiang forms a water type of the Changjiang plume (CP), known as Changjiang diluted water [Lieet al., 2003]. The distribution pattern of the CP shows a remarkable seasonal variation; in winter, it hugs thecoast of China southwestward, while it extends northeastward to the southern Yellow Sea and northernEast China Sea in summer, adopting a tongue-shaped pattern (Figure 1). Occasionally in the summer,extremely low-salinity water below 25.0 psu extends up to the vicinity of Jeju Island and the southwesterncoastal area of Korea (Figure 2), causing an abrupt change in the local oceanographic conditions and seriousdamage to fishery environments due to mass mortality [Seo et al., 1999]. To reduce the possible damagedue to a similar event in these regions, it is important to understand and predict the behavior of the CP’sextension.

It is known that the distribution of the seasonal CP is influenced by monsoon winds, the discharge of theChangjiang, tides, and ocean currents such as the Kuroshio Current and Taiwan-Tsushima Warm Current[Bang and Lie, 1999; Beardsley et al., 1985; Chang and Isobe, 2003; Chen et al., 2008; Lie et al., 2003; Moonet al., 2009; Oh and Park, 2004; Wu et al., 2014]. In particular, the summertime extension of the CP is mainlycontrolled by the discharge and the surface winds [Lie et al., 2003]. The CP begins to extend eastward inJune (red contour in Figure 1), when moderate southerly winds begin (red arrow in Figure 1), with a rapidincrease in the discharge (bar chart in Figure 1). The offshore extension toward Jeju Island in a tongue-shaped pattern is fully developed during July–August (blue and green, respectively) with strong southerlywinds prevailing and the discharge at a maximum. In September, when southerly winds vary with relativelystrong northeasterly winds, the plume retreats westward to the Changjiang mouth (black).

On the other hand, it is well known that strong winds during the passage of the typhoons produce the verti-cal mixing and upwelling in the upper ocean and the resultant intense surface cooling [Chiang et al., 2011;

Special Section:Oceanic Responses andFeedbacks to TropicalCyclones

Key Points:� Typhoons in the Yellow and East

China Seas play a crucial role in theChangjiang plume extension duringsummer� Typhoons contributed to the delay of

the Changjiang plume extension nearJeju Island up to 20 days� Typhoons prohibited the Changjiang

plume from spreading to the coastalcultivating region

Correspondence to:I.-J. Moon,ijmoon@jejunu.ac.kr

Citation:Lee, J. H., I.-J. Moon, J.-H. Moon,S.-H. Kim, Y. Y. Jeong, and J.-H. Koo(2017), Impact of typhoons on theChangjiang plume extension in theYellow and East China Seas, J. Geophys.Res. Oceans, 122, 4962–4973,doi:10.1002/2017JC012754.

Received 31 JAN 2017

Accepted 28 APR 2017

Accepted article online 4 MAY 2017

Published online 21 JUN 2017

VC 2017. American Geophysical Union.

All Rights Reserved.

LEE ET AL. IMPACT OF TYPHOONS ON CHANGJIANG PLUME 4962

Journal of Geophysical Research: Oceans

PUBLICATIONS

Lin et al., 2008; Moon, 2005; Moonand Kwon, 2011]. Since the CPflows in the surface layer, the ver-tical mixing can influence the dis-tribution and extension of the CP.Few studies have examined thepossible effects of typhoons onthe CP. You et al. [2011], for exam-ple, investigated the impact oftyphoon-induced vertical mixingon the behavior of the CP byusing a numerical model, butthey focused on qualitative inter-pretation without quantitativeanalysis.

This study aims to use quantita-tive analysis to investigate theCP’s response to typhoons andto demonstrate that typhoonsin the YECS, which have beenneglected in most CP studies, arealso a crucial factor affecting sum-mertime CP distribution. For this,we first present observational evi-dence, based on measurementsfrom an ocean research stationand vessel during the passage of

two typhoons, Ewiniar in 2006 and Dianmu in 2010, over the CP. In particular, we demonstrate from bothobserved salinity data and theoretical estimates that the typhoons induced a significant delay in the extension ofthe CP toward Jeju Island.

Since the analyses from the in situ measurements have limits in space and time, we conduct additionalnumerical experiments. First, we simulate the CP’s responses to Typhoon Dianmu (1004) using realisticatmospheric/oceanic conditions and compare the results with observations. Then, the quantitative effectsof the typhoon-induced vertical mixing and the changed atmospheric circulations on salinity distributionare investigated using freshwater thickness and volume from a series of numerical experiments with/with-

out typhoon forcing.

Details of the numerical experi-ments and the data set used aredescribed in the following sec-tions. The observed responses ofthe CP to typhoons are presentedin section 3. The numerical resultsare described in section 4. Finally,conclusions and summaries aregiven in section 5.

2. Descriptions of Dataand NumericalSimulations

During the period of TyphoonsEwiniar (0603) and Dianmu (1004),a unique data set was obtained

Figure 1. Monthly distributions of 30 psu isopleths from June to September in the YECS,adopted from climatological surface salinity maps of Lie et al. [2003]. Arrows in left-bottomcorner represent monthly averaged wind vectors within the CP areas (29–348N, 121–1278E). The bar chart in the bottom-right corner represents monthly mean discharges ofthe Changjiang for 1993–2013. Light gray lines represent tracks of typhoons passing overthe CP. A red cross symbol represents the location of IORS. Thick black lines with filled andopen circles represent the tracks of Ewiniar (0603) and Dianmu (1004), respectively.

Figure 2. Sea surface salinity observed in the middle of August 1996. Lower than 24 psusalinity reached the west coast of Jeju Island.

Journal of Geophysical Research: Oceans 10.1002/2017JC012754

LEE ET AL. IMPACT OF TYPHOONS ON CHANGJIANG PLUME 4963

from the Ieodo Ocean Research Station(IORS), which is located in the middle ofthe northern East China Sea and in theCP’s extension region in summer (Figure1). The observed variables at the IORS arewater temperature and salinity at fivelayers (8, 16, 24, 32, and 40 m depths),maximum wave height, and air pressure,humidity, wind direction, and windspeed (10 min averaged) at a 44 m alti-tude [Moon et al., 2010]. Hydrographicmeasurements were also made by aresearch vessel near Jeju Island using aconductivity-temperature-depth (CTD)sensor before and after the passage ofTyphoon Dianmu.

The information on the typhoons’tracks was obtained from the besttrack archives of the Regional Special-ized Meteorological Center (RSMC),Tokyo Typhoon Center. The data setsconsist of the typhoons’ names, longi-tude and latitude positions, minimumsurface central pressures, and maxi-mum sustained wind speeds with 6 hintervals for the period 1951–2015 (65

years). The monthly mean wind vectors (1979–2015) in the YECS are calculated using the National Centersfor Environmental Prediction-National Center for Atmospheric Research (NCEP–NCAR) analysis data [Kalnayet al., 1996], which have a horizontal resolution of 2.58 3 2.58 latitude-longitude. The monthly mean dis-charge (1992–2013) of the Changjiang is obtained from the Datong station.

Simulations for oceanic responses to Typhoon Dianmu (1004) are performed using the Regional OceanModeling System (ROMS) over the YECS, which was specially designed for predicting the behavior ofChangjiang diluted water [Lee, 2014; Lee et al., 2011] and named as YECS model in this study. The YECSmodel is a three-dimensional ocean model that is applied with free surface primitive equations usingorthogonal curvilinear and stretched terrain-following coordinates. The domain of the YECS model uses therotated rectangle (blue box in Figure 3). The horizontal and vertical resolution is about 8 km and 20 layers,respectively. The bottom topography was extracted from a combination of two topographic data sets,ETOPO5 from the National Geophysical Data Center (NGDC) and digital 30 s gridded bathymetric data overthe Korean marginal seas [Seo, 2008]. Bottom stress is parameterized using a quadratic bottom friction law,in which the drag coefficient is 0.0025 [Lee and Beardsley, 1999]. The coefficients for horizontal diffusion andviscosity are 20 and 50 m2 s21, respectively. The boundary conditions of the YECS model are provided bythe North Western Pacific (NWP) model, which has a resolution of 25 km for the horizontal and 20 layers forthe vertical [Lee, 2014]. The initial and boundary conditions for the temperature and salinity in the NWPmodel were obtained from the World Ocean Atlas 2001 (WOA2001). The NWP model was stabilized for 10years with monthly mean Cooperative Ocean and Atmosphere Dataset (COADS) surface forcing data.

Initial conditions of the YECS model are obtained from sea surface height, temperature, salinity, and velocityfields for January, produced by 10 year integration of the NWP model and interpolated into the resolutionof the YECS model. The open boundary conditions for the free surface and depth-averaged momentum aregiven by the Chapman [1985] and Flather [1976] formulations. Three-dimensional fields of velocity aretreated with an Orlanski-type radiation condition [Marchesiello et al., 2001]. Surface boundary conditions aregiven by 6 hourly NCEP FNL (FiNal anaLysis) data, which include wind, air temperature, air pressure, humid-ity, and precipitation data with a horizontal resolution of 0.28. Air-sea heat and momentum fluxes are calcu-lated by the bulk formulae [Fairall et al., 1996].

Figure 3. Model domain (blue box) and bottom topography (gray line). Threeblack boxes (A, B, and C) represent the southern Yellow Sea, the South Sea ofKorea, and the East China Sea regions, respectively. Line S, in red, represents thepathway of the CP inflow toward Jeju Island.

Journal of Geophysical Research: Oceans 10.1002/2017JC012754

LEE ET AL. IMPACT OF TYPHOONS ON CHANGJIANG PLUME 4964

Warner et al. [2005b] incorporated a number of turbulence closure schemes into ROMS. In the present study,the vertical mixing is parameterized using the Generic Length Scale (GLS) mixing scheme of K-Kl, in which thebackground diffusivity was chosen as 1025 m2 s21. This scheme has been successfully applied to coastal andestuarine areas [Banas et al., 2009; MacCready et al., 2008; Warner et al., 2005a]. Barotropic and baroclinic timesteps were set at 300 and 20 s, respectively. Real-time tides were also considered and were calculated afternodal correction using 10 tidal components (M2, S2, N2, K2, K1, O1, P1, Q1, Mf, and Mm) provided by the TPXO(TOPEX/POSEIDON) 7-Atlas. As a freshwater source, the YECS model considers the daily freshwater dischargefrom the Changjiang River for realistic salinity simulations using discharge data at the Datong station.

3. Observed CP Responses to Typhoons in the YECS

For the analysis of upper ocean responses to typhoons, we used the IORS data for air pressure, wind speed,maximum wave height, and water temperature/salinity at an 8 m depth as well as CTD measurement data

Figure 4. Time series of atmospheric (10 m wind speed and surface pressure) and oceanic variables (maximum wave height and 10 mwater temperature and salinity) observed at the IORS during the passages of (a, b) Ewiniar and (c, d) Dianmu.

Journal of Geophysical Research: Oceans 10.1002/2017JC012754

LEE ET AL. IMPACT OF TYPHOONS ON CHANGJIANG PLUME 4965

near Jeju Island. Here, air pres-sure and wind speed data atIORS were converted into valuesat the surface and 10 m priorto the analysis. The observedatmospheric and oceanic varia-bles were then paired and plot-ted during the period of Ewiniarand Dianmu (Figure 4).

When Typhoon Ewiniar passedby 60 km to the east of theIORS (Figure 1), the observedsurface pressure, 10 m windspeed, and maximum waveheight recorded up to 971 hPa,29.1 m s21, and 9 m, respec-tively (Figure 4a and Table 1).The track of Typhoon Dianmuwas very similar to that of Ewi-niar, but the surface pressure(989 hPa), wind speed (17 ms21), and wave height (8 m)were not as strong as those of

Ewiniar, though the center of Dianmu passed directly over the IOR (Figure 4c and Table 1). At that time, thetemperature and salinity sensor installed at the 8 m depth of the IORS within the mixed layer (ML) recordeda significant temperature drop (�108C and 48C) and salinity surge (� 9.3 and 3.6 psu) for Ewiniar and Dia-nmu, respectively (Figure 4b). While such a large upper ocean temperature decrease due to typhoons has

Figure 5. Spatial distributions of surface salinity and vertical sections of salinity measured by a research vessel (a, c) before (3–7 August2010) and (b, d) after (12–13 August 2010) the passage of Dianmu near Jeju Island. Solid lines with open circles represent the track of Dia-nmu. The vertical sections of salinity are plotted along Line A (red line).

Table 1. Comparison of the Best Estimates of Variables and Parameters From TyphoonsEwiniar and Dianmua

Variable or Parameter Ewiniar (0603) Dianmu (1004)

Observed max. 10 m wind speed 29.1 m s21 17 m s21

Observed min. surface pressure 971 hPa 989 hPaTC approach time and date 01:50 UTC

10 Jul 200615:20 UTC

10 Aug 2010Distance from storm center 60 km 5 kmTranslation speed, UH 10 m s21 7 m s21

Burger number, B5g0h1

4f 2 R2max

0.013 0.007Mach number, C5 UH

c 5 3.5Nondimensional storm speed, S5 pUH

4fRmax1.9 1.0

Rossby number for ML current, Q5 sqo h1 UH f 0.3 0.1

Max. temperature cooling 108C 48CMax. salinity increase 9.3 psu 3.6 psuBackground wind speed, U10_background 7.0 m s21 6.5 m s21

Storm’s radius, R30kt 148 km 185 kmObserved salinity RT �20 days �17 daysTheoretically estimated salinity RT 18.3 days 24.6 days

aHere, we define h1 5 10 m, f 5 7.7 3 1025 s21, and c 5 2 m s21. For Ewiniar and Dia-nmu, we define Rmax 5 55 km and 70 km, s 5 22.5 Pa and 5.3 Pa, and g05 0.09 and0.08 m s22, respectively, based on the best track data and Price et al. [1994]. U10_background

is obtained by averaging the SSE (direction toward IORS) component of 10 m windspeed observed at the IORS during the RT after the storm’s passage. R30k is estimatedusing the shortest radius of 30 kt winds since the CP is located along the left of trackwhere wind speeds are low.

Journal of Geophysical Research: Oceans 10.1002/2017JC012754

LEE ET AL. IMPACT OF TYPHOONS ON CHANGJIANG PLUME 4966

been reported by earlier studies [Lin et al., 2003; Moon and Kwon, 2011], the sudden typhoon-induced salin-ity increase has not been explored so far.

This unique observation demonstrates that the typhoon-induced strong vertical mixing and upwelling notonly can change the upper ocean thermal structure but also modify its salinity distribution. Figure 4b also

Figure 6. Comparisons of horizontal surface distributions (a, b) and vertical sections (c, d) for the simulated salinity between (a, c) before(8 August, 2010) and (b, d) after (14 August, 2010) the passage of Dianmu near Jeju Island. The vertical sections of salinity are plotted alongLine A in (a). White solid lines with filled circles in (b) represent the track of Dianmu.

Figure 7. Comparisons of area mean (124.5–126.58E, 31.5–32.5100588N) wind stress vectors between (a) Typ. EXP and (b) Clm. EXP. Thegray shading represents the period of storm passage.

Journal of Geophysical Research: Oceans 10.1002/2017JC012754

LEE ET AL. IMPACT OF TYPHOONS ON CHANGJIANG PLUME 4967

shows that it took about 20 days forthe surface water to recover its originalsalinity value, suggesting that atyphoon can lead to significant delayof the CP’s extension in this region.Similar oceanic responses to TyphoonDianmu, i.e., a sudden temperaturedrop (�48C) and salinity surge (� 3.5psu), were also observed at the IORS(Figure 4d and Table 1), although thetemperature and salinity changes werenot as large as in Ewiniar’s case. ForDianmu, the recovery time (RT) ofsalinity at the IORS was about 17 days.

Analysis of nondimensional parameters(Table 1) during the two typhoon eventsreveals that the case of Ewiniar was dis-tinguished by a large Rossby numberfor ML current (the ratio of horizontaladvection of momentum to the Coriolisforce), which leads to enhanced nonlo-cal effects by horizontal advection dur-ing the forced stage response [Priceet al., 1994]. In the case of Dianmu, asmall Mach number (the ratio of stormtranslation speed to the gravest modeinternal wave phase speed) and smallnondimensional storm speed (the ratioof the local inertial period to thetyphoon residence time) provided rela-tively favorable conditions for upwelling

and inertial motions during the relaxation stages, mainly due to the slower storm translation speed than forEwiniar [Dickey et al., 1998]. This implies that the large temperature cooling and salinity increase for Ewiniarwere mainly induced by the strong forced stage response caused by the relatively high maximum wind speed.

For Ewiniar and Dianmu, the CP’s RTs were estimated theoretically using the storm’s radius (R30kt) and thesurface current speeds (us) induced by after-storm background winds (U10_background). Here, we assume thatthe salinity recovery is made only by a direct advection from the outer regions of the storm’s radius throughthe shortest way and that there are no background currents. The surface current speeds [Pond and Pickard,1983] can be estimated as

us50:0127 U10 backgroundffiffiffiffiffiffiffiffiffiffiffiffiffiffi

ðsin /Þp (1)

where / is latitude (328N). Using U10_background 5 7.0 m s21 (Ewiniar) and 6.5 m s21 (Dianmu) andR30kt 5 148 km (Ewiniar) and 185 km (Dianmu), the estimated RT (5 R30kt/us) was 18.3 days and 24.6 days,respectively (see Table 1). These values were generally similar to those observed (about 20 and 17 days forEwiniar and Dianmu, respectively). The difference between the theoretical estimations and the observationsmight be from the effects of tidal mixing and background current, which were not considered in the equa-tion (1), as well as the uncertainty of storm’s radius.

Considering that the range of the CP’s eastward extension in summer is very sensitive to monsoon winds’variation over time, the delay of the extension due to the typhoon, on a time scale of about 2–4 weeks, mayresult in producing a different CP pattern from that expected without the typhoon’s effect. This also sug-gests that the typhoon’s effect can even contribute to preventing serious damage to coastal environmentsfrom low-salinity water. Typhoon Dianmu (1004), which passed by Jeju Island, is a good example to

Figure 8. Horizontal distribution of (a) surface salinity simulated from Clm. EXPand the difference between the two experiments (Clm. EXP–Typ. EXP; Figures 7aand 6b) on 14 August 2010. In (b), blue colors represent the increase in surfacesalinity (i.e., weakening of the CP) due to the inclusion of Typhoon Dianmu. Sym-bol § indicates the track of Dianmu.

Journal of Geophysical Research: Oceans 10.1002/2017JC012754

LEE ET AL. IMPACT OF TYPHOONS ON CHANGJIANG PLUME 4968

demonstrate this effect. As shown in Figure 5a, the CP, with very low salinity (up to 23 psu), approachedJeju Island on 10 August 2010, just before the arrival of Dianmu, which was a serious threat to people culti-vating near Jeju Island. Fortunately, however, the plume was mixed and diminished by the passage ofTyphoon Dianmu (Figure 5b). The vertical salinity section (Figures 5c and 5d) along Line A (Figure 5b) showsthat the typhoon-induced mixing was limited to the upper layer but effectively mixed the low-salinity waterfloating in the upper 15 m with deeper high-saline water. Indeed, it turned out that the typhoon-induceddelay of the CP’s extension toward Jeju Island played a critical role in preventing the low-salinity water fromspreading to the large cultivating area of the coastal region of Jeju Island. In the following section, we willperform a quantitative analysis to investigate the effect of Typhoon Dianmu on the CP’s extension near JejuIsland using the YECS model.

4. Simulated Impacts of Typhoons on the CP’s Extension

To examine the effect of typhoon on the CP’s extension, first we simulated oceanic variations during thepassage of Typhoon Dianmu in 2010, using realistic typhoon wind forcing. The model simulations revealthat, before the arrival of Dianmu (8 August), the CP extended to the west of Jeju Island, in which the 25psu contour line had a tongue shape and reached 100 km from west coast of Jeju Island, heading toward

Figure 9. Spatial distributions for freshwater thickness (a) before the passage of Typhoon Dianmu (8 August 2010), (b) after the passage(14 August 2010), and (c) their differences, estimated from Typ. EXP. Item (d) shows the same as for Figure 9b, but from Clm. EXP, using cli-matological wind instead of typhoon forcing (14 August 2010). Item (e) shows the differences between Figures 9d and 9b. Symbol § indi-cates the track of Dianmu.

Journal of Geophysical Research: Oceans 10.1002/2017JC012754

LEE ET AL. IMPACT OF TYPHOONS ON CHANGJIANG PLUME 4969

Jeju Island (Figure 6a), which was con-sistent with observations (Figure 5a).However, after Dianmu passed bythe west of Jeju Island at 15 UTC 10August, the shape and magnitude ofthe CP changed significantly: thetongue-like CP turned to the north,and the low-salinity water below26 psu disappeared (Figure 6b). Inparticular, the vertical structure ofsalinity at the upper layers changeddramatically (Figures 6c and 6d). Alens-like CP structure at the upperlayers disappeared, and the isoha-lines of 30 and 33 psu at 125.88Eshallowed (from a 5 m depth to thesurface) and deepened (from 12 to16 m), indicating vertical mixing atthe upper layers. These patterns arecommonly found in the vertical struc-

ture of temperatures during the passage of typhoons [Lin et al., 2003; Moon and Kwon, 2011] and match wellwith those from measurements (Figures 5c and 5d).

If Typhoon Dianmu had not passed this region, how much further would the CP have extended to the eastand toward Jeju Island? To address this question, we carried out an additional experiment (Clm. EXP), inwhich the model setup was the same as the original experiment (Typ. EXP) but replacing the wind datafrom 2010 that contained the typhoon vortex for the period of the typhoon’s passage (7–14 August) withclimatological winds. The time series of the wind vector during the passage of the typhoon shows strongwinds and drastic changes for Typ. EXP but relatively weak winds and smooth variations for Clm. EXP (Fig-ure 7). The results from Clm. EXP reveal that the CP extended further to the east, branching off: one branchheaded to the southern coast of Korea, and the other approached the western coast of Jeju Island (compareFigure 8a with Figure 6b). The difference in surface salinity between the two experiments (Clm. EXP – Typ.EXP) clearly reveals that Dianmu prevented the extension of the CP, resulting in an extensive increase (upto �9 psu) in surface salinity along the original CP as well as in the western coastal regions of Jeju Island(see Figure 8b).

To estimate the impact of the typhoon on the CP extension quantitatively, we calculated freshwater thick-ness (df ), which is the equivalent depth of freshwater relative to the reference salinity (s0). Freshwater thick-ness makes it possible to visualize the horizontal pattern of plume dispersal more readily and is defined asthe vertical integral of the salinity anomaly [Choi and Wilkin, 2007],

df 5

ðg

2h

s02ss0

dz (2)

where s is the salinity of the water column, g is sea water level, h is bottom depth, and s0 is determinedusing the maximum value of simulated salinity to ensure that freshwater thickness (df ) is always positive.

The estimated spatial distribution of the freshwater thickness (Figure 9a) matches well with that of surfacesalinity (Figure 6a), and it is clear that the freshwater thickness became shallower after the storm’s passage(Figure 9b). The difference (Figure 9c) between before and after the storm passage also reveals that the df

mostly shallowed along the storm track and the original CP regions (blue shades) due to the direct effect oftyphoon-induced mixing, while the df thickened near the mouth of the Changjiang (red shades around1238E) and the Kyushu regions (around 1288E), which seems to have been caused by the advection of low-salinity water due to cyclonic winds from Dianmu. A comparison of the differences (Figure 9e) in the df dis-tribution between Typ. EXP (Figure 9b) and Clm. EXP (Figure 9d) on 14 August further supports the beliefthat Typhoon Dianmu contributed to a significant reduction in freshwater thickness along the CP, i.e., block-ing the CP’s extension.

Figure 10. Comparison of freshwater volumes at Line S between Clm. EXP and Typ.EXP for August 2010. Symbol solid inverted triangle represents the arrival time ofDianmu near Jeju Island. A dashed arrow represents the period when the freshwatervolume from Typ. EXP returned to the value of Clm. EXP.

Journal of Geophysical Research: Oceans 10.1002/2017JC012754

LEE ET AL. IMPACT OF TYPHOONS ON CHANGJIANG PLUME 4970

To quantify the blocking and delay ofthe CP’s extension near Jeju Island,freshwater volume was estimated atLine S (see Figure 3), which is a path-way of the CP’s inflow toward JejuIsland, as well as at three regions (A, B,and C in Figure 3), representing thesouthern Yellow Sea, the South Sea ofKorea, and the East China Sea regions,respectively. Freshwater volume (Vf ) isa useful index for monitoring the tem-poral changes (gain or loss) of freshwa-ter in a certain region, defined by

Vf 5

ðððs02s

s0dV (3)

where s is the salinity of the water col-umn and s0 is the maximum value ofsalinity. We first compared the timeseries of the Vf between Typ. EXP andClm. EXP along Line S for August 2010(Figure 10). In the Clm. EXP (black line),the Vf increased dramatically during7–8 August before the arrival (10August) of Dianmu, which implies thata strong CP could extend to Jeju Islandif there were no typhoon events andwind conditions were similar to the cli-matological pattern. This pattern isvery different from that of Typ. EXP;that is, the Vf slowly increased during7–10 August and suddenly decreasedwhen Dianmu arrived around JejuIsland on 10 August. The large differ-ence in Vf (�1.3 3 109 m3) betweenthe two experiments for the period of8–9 August (a few days before anarrival of Dianmu) suggests that thechange of atmospheric circulation dueto the approach of the typhoon indi-rectly influenced the constraint of theCP’s extension toward Jeju Island. Infact, the prevailing wind direction overthe regions from the Changjiang

mouth to Jeju Island changed from southerly for Clm. EXP to northeasterly for Typ. EXP during 7–9 August(Figure 7), which contributed to blocking the CP’s extension. On 24 August, the Vf of the two experimentsbecame identical, suggesting that the effect of the typhoon lasted for 17 days, which is, interestingly, verysimilar to the observed RT (�17 days) for Dianmu in Table 1.

Comparing the temporal variations for the total Vf from the two experiments at three regions (A, B, and C)further supports the view that Dianmu influenced the change in freshwater volume at wide regions for along period. In fact, the difference in the Vf between the two experiments at regions A and B (Figures 11aand 11b) lasted until the end of August (never recovering to climatological conditions), which means that atyphoon event could influence the CP’s migration not only for the typhoon-passage period (9–11 August)but also for entire summer period over the southern Yellow Sea (region A) and the South Sea of Korea,

Figure 11. Comparison of time series for total freshwater volumes at threeregions (A, B, and C) between Clm. EXP and Typ. EXP for August 2010. Symbolsolid inverted triangle represents the arrival time of Dianmu near Jeju Island.

Journal of Geophysical Research: Oceans 10.1002/2017JC012754

LEE ET AL. IMPACT OF TYPHOONS ON CHANGJIANG PLUME 4971

including Jeju Island (region B). This also emphasizes that the winds during 7–9 August played a critical rolein the migration of the CP to regions A and B. In region C (the East China Sea region), in contrast to theother two regions, recovery from the initial shock (i.e., the decrease of the Vf ) due to Dianmu occurred inabout 17 days (Figure 11c).

5. Conclusions and Summary

The Changjiang is known to play a crucial role in forming productive biological environments as a majornutrient source in the YECS. Occasionally, however, extremely low-salinity water, originating in the Chang-jiang, extends in the summer up to the vicinity of Jeju Island and the southwestern coastal area of Korea,causing serious damage to the fishery and cultivating industry. Consequently, understanding and predictingthe dispersal behavior of the CP in the summer is a very important research subject in these regions.

It is known that the summertime distribution of the CP is controlled by winds, discharge from the Chang-jiang, tides, and ocean currents. In addition to these factors, this study suggests that typhoons also play animportant role in the extension and distribution of the CP in summer. A unique data set observed at theIORS, which is located in the CP’s extension areas, as well as hydrographic measurements by a vessel duringtwo typhoons’ passages, allowed us to investigate the CP’s response to typhoons.

The most interesting findings obtained from the observations were that the typhoon-induced strong verti-cal mixing and upwelling not only changed the upper ocean thermal structure [Moon and Kwon, 2011] butalso modified its salinity distribution significantly, which lasted for up to 2–3 weeks. During the passage oftwo typhoons, Ewiniar (0603) and Dianmu (1004), the sudden surges of upper ocean salinity up to 9.3 and3.6 psu, respectively, were observed at the IORS, and it took about 20 and 17 days, respectively, to restorethe salinity to the pretyphoon conditions. This suggests that the typhoon can lead to a significant delay inthe CP’s extension. A simple estimate of the CP’s RT, which is based on the storm’s size and the surface cur-rent speed due to after-storm background winds, was also generally consistent with observation. In particu-lar, in the case of Dianmu, the typhoon-induced delay in the CP contributed to preventing low-salinitywater from spreading to the coastal region of Jeju Island and thereby avoiding potential mass mortality.

To investigate the responses of the CP to typhoons quantitatively, additional numerical experiments with/without Typhoon Dianmu were performed and the results were analyzed in terms of freshwater thicknessand volume. The results revealed that Dianmu contributed to a significant reduction in freshwater thicknessalong the CP, as found in the surface salinity distribution. On the basis of the freshwater volume at a path-way of the CP’s inflow toward Jeju Island, the estimated salinity RT (17 days) for Dianmu was consistentwith observations, and it was found that the change in atmospheric circulation due to the approach of Dia-nmu indirectly influenced the blocking of the CP’s extension toward Jeju Island before the arrival of Dianmu(7–9 August).

An analysis of the 65 year tropical cyclone (TC) best track data reveals that, on average, 1.3 typhoonsper year (total 86 typhoons) passed over the CP (Figure 1). Some of them have kept strong winds above26 m s21 during passage through the historical CP regions. In this study, evidence from both measurementsand model simulations demonstrates that in addition to river discharges, winds, ocean currents, and tides,typhoons in the YECS can be a crucial factor affecting the CP’s summertime distribution and extensionthrough a direct mixing process over the CP and indirect effects caused by changes in wind direction. How-ever, it should be noted that the CP’s response to typhoons depends on various factors, such as a typhoon’ssize, intensity, track, and translation speed, as well as the subsurface water structure. Further research isneeded to investigate the detailed processes involved in the CP’s responses to these factors.

ReferencesBanas, N. S., P. MacCready, and B. M. Hickey (2009), The Columbia River plume as cross-shelf exporter and along–coast barrier, Cont. Shelf

Res., 29, 292–301, doi:10.1016/j.csr.2008.03.011.Bang, I. K., and H. J. Lie (1999), A numerical experiment on the dispersion of the Changjiang River plume, J. Korean Soc. Oceanogr., 34, 185–199.Beardsley, R. C., R. Limeburner, H. Yu, and G. A. Cannon (1985), Discharge of the Changjiang (Yangtze River) into the East China Sea, Cont.

Shelf Res., 4, 57–76.Chang, P. H., and A. Isobe (2003), A numerical study on the Changjiang Diluted Water in the Yellow and East China Seas, J. Geophys. Res.,

108(C9), 3299, doi:10.1029/2002JC001749.

AcknowledgmentsThis research was supported by the2017 scientific promotion programfunded by Jeju National University.The authors appreciate twoanonymous reviewers for theircomments that contribute to improvethe manuscript. The historical westernNorth Pacific typhoon best track datais provided from the RegionalSpecialized Meteorological Center(RSMC), Tokyo Typhoon Center. Thedata are available at http://www.jma.go.jp/jma/jma-eng/jma-center/rsmc-hp-pub-eg/besttrack.html. Oceanclimatological temperature and salinitydata (WOA2001) were from NOAA’sNational Oceanographic Data Center(NODC). The data set is available athttp://www.nodc.noaa.gov/OC5/WOA01/qd_ts01.html. Theatmospheric surface wind data (NCEP/NCAR Reanalysis) were from NOAA/NCEP. The data are available at http://www.esrl.noaa.gov/psd/data/gridded/data.ncep.reanalysis.surface.html.Topographic data sets (ETOP5) werefrom the National Geophysical DataCenter (NGDC). The data are availableat https://www.ngdc.noaa.gov/mgg/global/relief/ETOPO5/. Monthly meanatmospheric data (COADS) were fromthe NOAA-CIRES Climate DiagnosticsCenter. The data are available at http://www.cdc.noaa.gov/. The monthlymean discharges for the Changjiangwere from Bureau of Hydrology, whichare available at http://www.cjh.com.cn.

Journal of Geophysical Research: Oceans 10.1002/2017JC012754

LEE ET AL. IMPACT OF TYPHOONS ON CHANGJIANG PLUME 4972

Chen, C., et al. (2008), Physical mechanisms for the offshore detachment of Changjiang Diluted Water in the East China Sea, J. Geophys.Res., 113, C02002, doi:10.1029/2006JC003994.

Chen, H., Z. Yu, Q. Yao, T. Mi, and P. Liu (2010), Nutrient concentrations and fluxes in the Changjiang Estuary during summer, Acta Oceanol.Sin., 29, 107–119.

Chapman, D. C. (1985), Numerical treatment of cross-shelf open boundaries in a barotropic coastal Ocean Model, J. Phys. Oceanogr., 15,1060–1075, doi:10.1175/1520-0485(1985)015< 1060:NTOCSO>2.0.CO;2.

Chiang, T.-L., C.-R. Wu, and L.-Y. Oey (2011), Typhoon Kai-Tak: An ocean’s perfect storm, J. Phys. Oceanogr., 41, 221–233.Choi, B.-J., and J. L. Wilkin (2007), The effect of wind on the dispersal of the Hudson River plume, J. Phys. Oceanogr., 37, 1878–1897.Dickey, T., D. Frye, J. McNeil, D. Manov, N. Nelson, D. Sigurdson, H. Jannasch, D. Siegel, T. Michaels, and R. Johnson (1998), Upper-ocean

temperature response to Hurricane Felix as measured by the Bermuda testbed mooring, Mon. Weather Rev., 126, 1195–1201.Fairall, C. W., E. F. Bradley, D. P. Rogers, J. B. Edson, and G. S. Young (1996), Bulk parameterization of air-sea fluxes for tropical ocean-global

atmosphere coupled-ocean atmosphere response experiment, J. Geophys. Res., 101(C2), 3747–3764.Flather, R. A. (1976), A tidal model of the northwest European continental shelf, M�em. Soc. R. Sci. Liege, 10, 141–164.Kalnay, E, et al. (1996), The NCEP/NCAR 40-year reanalysis project, Bull. Am. Meteorol. Soc., 77, 437–472.Lee, J. H. (2014), Analysis and prediction of Changjiang Diluted Water in the adjacent sea of Jeju Island, PhD thesis, Jeju Natl. Univ., Jeju,

Korea.Lee, S. H., and R. C. Beardsley (1999), Influence of stratification on residual tidal currents in the Yellow Sea, J. Geophys. Res., 104(C7), 15,679–

15,701, doi:10.1029/1999JC900108.Lee, J. H., I.-C. Pang, I.-J. Moon, and J.-H. Ryu (2011), On physical factors that controlled the massive green tide occurrence along the south-

ern coast of the Shandong Peninsula in 2008: A numerical study using a particle-tracking experiment, J. Geophys. Res., 116, C12036, doi:10.1029/2011JC007512.

Lie, H.-J., C.-H. Cho, J.-H. Lee, and S. Lee (2003), Structure and eastward extension of the Changjiang River plume in the East China Sea, J.Geophys. Res., 108(C3), 3077, doi:10.1029/2001JC001194.

Lin, I., W. T. Liu, C.-C. Wu, G. T. F. Wong, C. Hu, Z. Chen, W.-D. Liang, Y. Yang, and K.-K. Liu (2003), New evidence for enhanced ocean primaryproduction triggered by tropical cyclone, Geophys. Res. Lett., 30(13), 1718, doi:10.1029/2003GL017141.

Lin, I. I., C. C. Wu, I. F. Pun, and D. S. Ko (2008), Upper-ocean thermal Structure and the western North Pacific category-5 typhoons Part I:Ocean features and the category 5 typhoon’s intensification, Mon. Weather Rev., 136, 3288–3306.

MacCready, P., N. Banas, B. Hickey, E. Dever, and Y. Liu (2008), A model study of tide-and wind-induced mixing in the Columbia River estu-ary and plume, Cont. Shelf Res., 29(1), 278–291, doi:10.1016/j.csr.2008.03.015.

Marchesiello, P., J. C. McWilliams, and A. Shchepetkin (2001), Open boundary conditions for long-term integration of regional oceanic mod-els, Ocean Modell., 3, 1–20.

Moon, I.-J. (2005), Impact of a coupled ocean wave-tide-circulation system on coastal modeling, Ocean Modell., 8, 203–236.Moon, I.-J. and S. J. Kwon (2011), Impact of upper-ocean thermal structure on the intensity of Korean peninsula landfall typhoons, Prog.

Oceanogr., 105, 61–66.Moon J.-H, I.-C. Pang, and J.-H. Yoon (2009), Response of the Changjiang diluted water around Jeju Island to external forcings: A modeling

study of 2002 and 2006, Cont. Shelf Res., 29, 1549–1564.Moon, I.-J., Shim, J. S., Lee, D. Y., J. H. Lee, I.-K. Min, and K. C. Lim (2010), Typhoon studies using the Ieodo ocean research station: Part I.

Importance and present status of typhoon observation, Atmosphere, 20, 247–260.Oh, I. S., and T. Park (2004), A numerical study on the dispersion of the Yangtze River water in the Yellow and East China Seas, J. Korean

Soc. Oceanogr., 39, 119–135.Pond, S., and G. L Pickard (1983), Introductory Dynamical Oceanography, 2nd ed., 329 pp., Pergamon, Oxford, U. K.Price, J. F., T. B. Sanford, and G. Z. Forristall (1994), Forced stage response to a moving hurricane, J. Phys. Oceanogr., 24, 233–260.Seo, H. L., Y. G. Cho, H. Y. Seo, and D. H. Kim (1999), Mass mortality of macrobenthic animals around Jeju Isalnd in 1996: Roles of physical

factors, Korean J. Environ. Biol., 17, 175–182.Seo, S. N. (2008), Digital 30 sec gridded bathymetric data of Korea marginal seas-KorBathy30s (in Korean with English abstract), J. Korean

Soc. Coastal Ocean Eng., 20, 110–120.You, S. H., H. M. Eom, and S. B. Ryoo (2011), Numerical study of typhoon effects on the Changjiang diluted water using an operational

ocean forecasting system, J. Coastal Res., 64, 1927–1930.Warner, J. C, W. R. Geyer, and J. A. Lerczak (2005a), Numerical modeling of an estuary: A comprehensive skill assessment, J. Geophys Res.,

110, C05001, doi:10.1029/2004JC0 02691.Warner, J. C., C. R. Sherwood, B. Butman, H. Arango, and R. P. Signell (2005b), Performance of four turbulence closure models implemented

using a generic length scale method, Ocean Modell., 8, 81–113, doi:10.1016/j.ocemod.2003.12.003.Wu, H., J. Shen, J. Zhu, J. Zhang, and L. Li (2014), Characteristics of the Changjiang plume and its extension along the Jiangsu Coast, Cont.

Shelf Res., 76, 108–123.

Journal of Geophysical Research: Oceans 10.1002/2017JC012754

LEE ET AL. IMPACT OF TYPHOONS ON CHANGJIANG PLUME 4973