Impact of mesoscale eddies on Kuroshio intrusion ...oomg.meas.ncsu.edu/wp-content/uploads/2017/04/Yin-etal_JGR_201… · RESEARCH ARTICLE 10.1002/2016JC012263 Impact of mesoscale

RESEARCH ARTICLE10.1002/2016JC012263

Impact of mesoscale eddies on Kuroshio intrusion variabilitynortheast of TaiwanYuqi Yin1,2 , Xiaopei Lin2,3 , Ruoying He2,4 , and Yijun Hou1,2

1Key Laboratory of Ocean Circulation and Waves, Institute of Oceanography, Chinese Academy of Sciences, Qingdao,China, 2Qingdao National Laboratory for Marine Science and Technology, Qingdao, China, 3Physical OceanographyLaboratory, Ocean University of China, Qingdao, China, 4Department of Marine, Earth and Atmospheric Sciences, NorthCarolina State University, Raleigh, North Carolina, USA

Abstract Abundant mesoscale eddies propagate to the east of Taiwan and affect the Kuroshio Currentand its onshore intrusion northeast of Taiwan. But the dynamical process of how eddy activities modulatethe Kuroshio intrusion is still unclear. In this study, in situ ocean velocity observations, drifter trajectories,tide gauge measurements, satellite sea surface height data, and assimilative Hybrid Coordinate Ocean Mod-el (HYCOM) reanalysis outputs were used to analyze a cyclonic eddy-induced large Kuroshio onshore intru-sion during late autumn 2008. Along with nine anticyclonic eddy events and seven cyclonic eddy eventsextracted from long-term HYCOM reanalysis outputs, the variability of the eddy-induced intrusion wasquantified. The impacts of the cyclonic eddy and the anticyclonic eddy were opposite. Under the impact ofcyclonic (anticyclonic) eddy, the total onshore intrusion between 1228E and 1248E across the 200 m isobathswas decreased (increased) by 15.7% (7.3%), while the onshore intrusion west of 122.68E and the northeast-ward current on the outer shelf were increased (decreased) by 30.6% (4.7%) and 31% (10.1%), respectively.The arrival of cyclonic (anticyclonic) eddy brought positive (negative) potential vorticity (PV) flux and modu-lated the local ocean vertical stratification, which weakened (enhanced) the cross-slope PV gradient and theshelf slope constraint on the upper layer current and, therefore, favored (inhibited) the Kuroshio onshoreintrusion.

1. Introduction

Off the northeastern coast of Taiwan, the Kuroshio Current (KC) collides with the continental shelf of theEast China Sea (ECS) then veers northeastward along the shelf break (Figure 1). Observations [Uda, 1936;Hsueh et al., 1992] suggest that only a small part of the Kuroshio water flows onto the shelf northeast of Tai-wan. The long-term mean onshore Kuroshio intrusion is estimated to be 1.4 Sv along the whole ECS shelf,compared to the Kuroshio volume transport (KVT) in the East Taiwan Channel of �24 Sv [Tang et al., 2000;Liang et al., 2003; Guo et al., 2006; Isobe, 2008]. Traditionally, the Kuroshio intrusion northeast of Taiwan wasthought to be a quasi-steady phenomenon (especially in the winter time). It is defined as the KuroshioBranch Current (KBC) by previous studies. The KBC, together with the Taiwan Strait Warm Current (TSWC), isthe origin of the Taiwan Warm Current [Yuan and Su, 1984; Su and Pan, 1987; Guan and Fang, 2006], whichsupplies the Tsushima Warm Current [Fang et al., 1991; Isobe, 2008]. The KBC can even reach 308N offChangjiang River mouth from time to time [Kondo, 1985; Yang et al., 2012; Wang and Oey, 2016]. This cross-shelf Kuroshio intrusion provides an important way for heat and mass exchange between the ECS and thewestern Pacific Ocean [Chen et al., 1995; Matsuno et al., 2009] and plays an important role in controlling thephysical and biochemical environments on the ECS shelf [Chen, 1996; Gong et al., 1997; Liu et al., 2010; Sassaand Tsukamoto, 2010].

Because of the constraint of potential vorticity (PV) conservation, the Kuroshio generally flows along thesteep ECS slope without external forcing. Studies have proposed mechanisms to explain how the Kuroshioovercomes this strong topographic PV barrier to intrude onto the shelf region.

First, the collision of Kuroshio and the zonally running slope were considered as the primary cause for theintrusion [Chao, 1990; Hsueh et al., 1992]. Due to the zonal pressure gradient and the absent of lateralboundary northeast of Taiwan, the inertia drives the Kuroshio water to flow along the pressure gradient and

Key Points:� Moorings provide direct observations

of the cyclonic eddy-inducedKuroshio intrusion� The variability of the eddy-induced

Kuroshio intrusion is revealed andquantified� Dynamical interpretations of the

eddy-induced Kuroshio intrusion areproposed

Correspondence to:X. Lin,[email protected]

Citation:Yin, Y., X. Lin, R. He, and Y. Hou (2017),Impact of mesoscale eddies onKuroshio intrusion variability northeastof Taiwan, J. Geophys. Res. Oceans, 122,doi:10.1002/2016JC012263.

Received 20 AUG 2016

Accepted 20 FEB 2017

Published online 23 FEB 2017

VC 2017. American Geophysical Union.

All Rights Reserved.

YIN ET AL. EDDY ON KUROSHIO INTRUSION 1

Journal of Geophysical Research: Oceans

PUBLICATIONS

http://dx.doi.org/10.1002/2016JC012263

http://orcid.org/0000-0002-1732-3063

http://orcid.org/0000-0001-6413-4479

http://orcid.org/0000-0001-6158-2292

http://onlinelibrary.wiley.com/journal/10.1002/(ISSN)2169-9291/

http://publications.agu.org/

intrude onto the ECS shelf [Liang andSu, 1994]. But without any external forc-ing, PV conservation will require theintruded water back to the Kuroshioand has little contribution to the cross-shelf exchange. Yang et al. [2012] sug-gested that the vortex-tube stretching,which is induced by the steep slope onthe northern side of the I-Lan Ridge,leads to the Kuroshio water onshoreintrusion below the thermocline.

Second, the local atmospheric forcingcould steer the Kuroshio into the shelfregion. Wind-driven Ekman transport[Chao, 1990; Chen, 1996] and windstress PV input [Guo et al., 2006] havebeen suggested to account for theintrusion in the winter time. Chuangand Liang [1994] showed a 1 month lagbetween the wind field change and theintrusion. Using an idealized oceanmodel, Oey et al. [2010] reported theseasonal Kuroshio intrusion by surfaceheat flux forcing. Their PV budget indi-cates that the enhanced intrusion inwinter is accounted for by the jointeffect of baroclinicity and bottom relief(JEBAR) [Huthnance, 1984] term, whichis associated with strong sea surfacecooling in winter. The thermocline inthe offshore side tilts under the surfacecooling and a large JEBAR is generated,which could weaken the cross-shelf PV

gradient and increase the Kuroshio intrusion. This result is consistent with Guo et al. [2006], who suggestedthat the JEBAR term balances the advection of the geostrophic PV term and induces the Kuroshio intrusionalong the 200 m isobath. Furthermore, the anomalous positive wind stress curl and surface cooling north-east of Taiwan produce a deep upper layer and strong tilting of thermocline along the Kuroshio, whichgives rise to JEBAR and modulates the interannual variability of the Kuroshio intrusion [Wang and Oey,2014]. In addition, the sea surface heat flux gradient shows higher correlation than wind stress with thestronger intrusion events northeast of Taiwan [Wu et al., 2014].

Third, the interannual variability of the Kuroshio onshore intrusion along the ECS shelf break is also relatedto the KVT [C. Liu et al., 2014; X. Liu et al., 2014]. Northeast of Taiwan, the onshore intrusion increases whenthe KVT decreases. This phenomenon is attributed to the Kuroshio’s inertia, and the intrusion depends onthe balance between the inertia and the b-effect [Sheremet, 2001; Yuan and Wang, 2011]. However, this the-ory ignores the effect of topography, which requires further investigation.

Fourth, abundant frontal eddies and instability waves around the ECS shelf break [Qiu et al., 1990; Jameset al., 1999] contribute to the water exchange between the ECS shelf and the Kuroshio. Based on analysis ofwater mass, Guo and Wan [1995] showed that one frontal eddy outputs 0.44 Sv of shelf water into the Kur-oshio and draws 0.04 Sv of Kuroshio water onto the shelf. Using high-resolution model, nonlinear frontalwaves were excited around the shelf break, which results in more water exchange between the Kuroshioand the ECS shelf [Isobe and Beardsley, 2006].

Mesoscale eddies can contribute to the water exchange around the shelf break. In the Mid-Atlantic Bight offthe northeastern United States, large eddies and rings developed from the Gulf Stream could either

Figure 1. Winter circulation around Taiwan Island, overlaid with the location ofADCP mooring (268300N, 1228360E, black triangle), Keelung (258090N, 1218450E),and Ishigaki (248200N, 1248100E) stations, the I-Lan Ridge (blue line), and the Ptransect (green line, normal to the 200 m isobath). The thick, thin, and dashedred arrows represent the KC, the KBC, and the TSWC. Also shown are the 200 m(thick black contour) and 2000 m (thin contour) isobaths.

Journal of Geophysical Research: Oceans 10.1002/2016JC012263


transport cold shelf water offshore [Joyce, 1984; Cherian and Brink, 2016] or transport warm and salty wateronto the shelf [Oey and Zhang, 2004; Zhang and Gawarkiewicz, 2015]. Owing to the baroclinic instability ofthe North Pacific Subtropical Countercurrent (STCC)-North Equatorial Current (NEC) system, abundant meso-scale eddies are generated and propagate to the western boundary, impacting the Kuroshio’s path and vol-ume transport [Johns et al., 2001; Zhang et al., 2001; Kuo and Chern, 2011; Chang and Oey, 2011; Hsin et al.,2013; Soeyanto et al., 2014; Tsai et al., 2015]. Gawarkiewicz et al. [2011] suggested that the onshore intrusionnortheast of Taiwan is associated with low KVT, a cyclonic eddy pattern, and offshore meandering east ofTaiwan. In contrast, little onshore intrusion is observed during the period with high KVT and an anticycloniceddy pattern east of Taiwan. V�elez-Belch�ı et al. [2013] also showed that cyclonic eddies from the westernPacific are well correlated with the observed large Kuroshio intrusion northeast of Taiwan. Using the sameidealized model as Hsueh et al. [1992], V�elez-Belch�ı et al. [2013] suggested that it is the incident angle of theKuroshio, modified by eddies, that determines the strength of the intrusion. The shoreward migration ofthe Kuroshio axis [Ichikawa et al. 2008] and thermal front [Hsin et al., 2011] takes place northeast of Taiwanwhen the KVT weakens, and vice versa, especially at period of 120 days, which is attributed to the interac-tion between the westward propagating open-ocean mesoscale eddies and the Kuroshio east of Taiwan.Mesoscale eddies are also suggested to have a role in modulating the seasonal variability of the Kuroshiointrusion northeast of Taiwan [Yin et al., 2014].

While these studies have provided important information on eddies’ impact on the variability of the Kur-oshio intrusion, in situ observations and a dynamical explanation of eddy activities leading to the Kuroshiointrusion are still lacking. The objective of this study is to investigate the impact of eddies on the Kuroshioand its intrusion specifically in the subsurface ocean and give the dynamical interpretation. A velocity timeseries observed by an acoustic Doppler current profiler (ADCP) mooring northeast of Taiwan and concurrenthigh-resolution HYCOM reanalysis fields were utilized to study a cyclonic eddy-induced large Kuroshioonshore intrusion during late autumn 2008, which reveal the intrusion variability. The dynamical interpreta-tions were made based on the PV budget analysis. Furthermore, the generality of eddy-induced intrusionand the applicability of the dynamical interpretations were verified by multiple strong eddy eventsextracted from 9 year HYCOM reanalysis data.

2. Data

This study used a unique in situ observation opportunity near the edge of the continental shelf northeast ofTaiwan. An ADCP mooring was deployed at 113 m (see Figure 1) from 21 September 2008 to 5 April 2009.Velocity and bottom temperature were sampled every 10 min using a bottom-mounted 300 kHz ADCP anda thermistor. With 8 m vertical bins, the ADCP provided velocity records in 13 vertical layers of the water col-umn. The top two bins were excluded in our analysis because the data were contaminated by sidelobereflection. Because we focus on subtidal circulation, a 36 h low-pass filter was applied to the remaining 11vertical bins, ranging from 14 to 94 m, to remove tidal signal.

Other in situ data include daily sea level observations at Keelung and Ishigaki (see Figure 1), which wereused to compute the difference between the stations’ sea level anomalies (relative to the 2008–2009 meansea level) as an indicator of Kuroshio volume transport [John et al., 2001]. The 10 m wind and sea surfaceheat flux fields used in this paper are from the ERA interim data set [Dee et al., 2011].

We used a 9 year (2004–2012) realistic ocean reanalysis output generated by the data assimilative globalHybrid Coordinate Ocean Model (HYCOM) [Chassignet et al., 2009] to depict the impact of eddies on theKuroshio and its intrusion. HYCOM applied the Navy Coupled Ocean Data Assimilation system [Cummingsand Smedstad, 2013], which assimilates available satellite altimeter and sea surface temperature observa-tions, in situ sea surface temperature, and vertical temperature and salinity profiles from XBTs, Argo floats,and moored buoys. HYCOM reanalysis fields are available daily with 1/128 horizontal resolution and 33z-levels.

We also used an eddy census database provided by http://cioss.coas.oregonstate.edu/eddies/ [Cheltonet al., 2011], satellite-based AVISO geostrophic velocity maps, and drifter data from the International ArgoProgram [Argo, 2000; Lumpkin et al., 2012] to trace mesoscale eddies east of Taiwan and quantify theircharacteristics.



http://cioss.coas.oregonstate.edu/eddies/

3. Results and Discussion

3.1. A Cyclonic Eddy-Induced Large Kuroshio Intrusion EventV�elez-Belch�ı et al. [2013] showed that there was a cyclonic eddy-induced large Kuroshio intrusion event inOctober 2008 (their Figure 6). To highlight the intrusion, we designated September as the period beforeeddy impact and October as the period during eddy impact. Both the AVISO-derived and HYCOM-simulatedmonthly mean surface velocity fields showed that the KC northeast of Taiwan (Figure 2) turned from north-eastward in September to nearly northward in October and intruded further onto the midshelf. During thisperiod, the KC interacted with a mesoscale cyclonic eddy propagated from the open ocean, and meanderedoffshore east of Taiwan. Trajectories data from 18 drifters that passed through the I-lan Ridge transectdepicted paths of the KC during our study period (Figure 3A). All drifters flowed along the slope in Septem-ber. Then in October most drifters moved directly north once they flowed over the I-lan Ridge and penetrat-ed to the 200 m isobath, consistent with the simulated drifters’ trajectories from the AVISO and HYCOMdata (Figures 3B and 3C).

During the same period, low Kuroshio transport was reflected by both tidal gauge and HYCOM-simulatedsea level differences between Ishigaki and Keelung (Figure 4A). Our 7 month long ADCP mooring deploy-ment northeast of Taiwan took concurrent measurements during this eddy-Kuroshio interaction period (Fig-ures 4B and 4C).

The depth-averaged subtidal current (36 h low-pass filtered) was decomposed into along-shelf (458

from north) and cross-shelf (3158 from north) components. Mean and maximum along-shelf velocitieswere 0.21 and 0.56 m s21, respectively. Spectrum analyses showed a �2–5 day period in along and

Figure 2. Mean geostrophic velocity (A) derived from AVISO sea surface height data and HYCOM model-simulated mean 15 m velocity(B) in September (top) and October (bottom). Gray contours are the 200 m isobaths.



cross-shelf velocity, and an 11 day period in cross-shelf velocity and bottom temperature. A high corre-lation between velocity and sea surface wind over a �2–4 day period was found over the ECS continen-tal slope northeast of Taiwan [Chuang and Liang, 1994]. The 11 day oscillation can be attributed to theKuroshio baroclinic instability waves that were observed near the ECS slope [James et al., 1999]. With a20 day low-pass filter, the correlation coefficient between along and cross-shelf velocities was 20.68,the correlation coefficient between along-shelf velocity and the bottom temperature was 0.67, and thecorrelation coefficient between cross-shelf velocity and the bottom temperature was 20.45. Thus, thealong-shelf velocity indicated the Kuroshio intrusion, which raised the bottom temperature by deliver-ing warmer water.

There was a persistent strong along-shelf flow in October, with a mean of 0.35 m s21, twice as fast as duringthe other 6 month observations (0.18 m s21). Observed bottom temperature also showed a significantincrease (mean 1.38C, maximum 3.58C) in October. Together, the velocity and temperature observations pro-vide direct evidence for the Kuroshio intrusion. Although the eddy-induced strong intrusion only lasted for1 month, it accounted for more than 30% of the total transport observed over the 7 month period, repre-senting a significant contribution to cross-shelf water exchange.

In addition, there was another strong subtidal current from December to February (Figure 4B), which shouldbe associated with the KBC induced by surface cooling [Oey et al., 2010]. The KVT also decreased and wasassociated with the impact of another weaker cyclonic eddy located southeast of Ishigaki (not shown here).Therefore, the Kuroshio intrusions in December–February were potentially initiated by both surface coolingand cyclonic eddy effects.

To better understand the Kuroshio intrusion in October and the corresponding vertical structure of the shelfcurrent, the subtidal current at the mooring station was decomposed into three components: Ekman cur-rent, barotropic current, and baroclinic current. Using a one-dimensional model and the ERA-interim dataset, the wind-induced Ekman current was calculated according to

Figure 3. (A) Trajectories of observed surface drifters, (B) simulated drifters with the surface geostrophic velocity from AVISO data, and (C) the 15 m velocity from HYCOM simulationresults in September (top) and October (bottom). Dots represent the starting locations of drifters. Six drifters were observed in September and 12 in October. The simulated drifters werereleased every other day in each month and tracked by the fourth-order Runge-Kutta integration scheme.



f 3~u5t@2~u@z2

; (1)

where f is the local Coriolis parameter,~u is the velocity vector, and t is the viscosity coefficient. The surfaceboundary condition is

q0t@~u@z

5~s; (2)

where q0 is the constant density (1025 kg m23) and ~s is the surface wind stress. Here t is 0.01 m2 s21

according to Price and Sundermeyer [1999]. After removing the Ekman current, the depth-averagedcurrent was taken as the barotropic current, and the remaining current was taken as the baroclinic current(Figure 5). Compared with the barotropic and baroclinic currents, the Ekman currents were very small (e.g.,�5 cm/s at 14 m) and thus negligible. Except in November, the magnitude of baroclinic current was onlyabout 5–20% of that of barotropic current. This quasi-barotropic current should be largely related to stron-ger mixing associated with wind and surface cooling in late fall to early spring [e.g., Oey et al., 2010].

Figure 4. (A) Observed and modeled sea level anomaly difference (unit: m) between Ishigaki and Keelung. Both time series are 36 h low-pass filtered. Period H (15 August to 15 September) is defined as the high (before eddy interaction) Kuroshio transport phase, and Period L(1 October to 31 October) is defined as the low (during eddy interaction) Kuroshio transport phase. (B) Depth-averaged (36 h low-pass fil-tered) along and cross-shelf velocities (unit: m s21) measured by the mooring. Positive values are northeastward (shoreward) for thealong-shelf (cross-shelf) velocities. Dashed line is the mean value (0.21 m s21) of the along-shelf velocity. (C) Bottom temperature (36 hlow-pass filtered, unit: 8C) measured by the mooring. The dashed line indicates the mean temperature (18.458C).



Figure 5. Vertical profiles of (A) observed along-shelf sub-tidal current velocity (color, unit: m s21), and (B) its barotropic and (C) barocliniccomponents. All three time series are 36 h low-pass filtered.

Figure 6. Comparisons between mooring observations and HYCOM simulation results at two adjacent grid points (124.58E, 268N) and(124.58E, 26.58N). (A) U and V are the depth-averaged along (458 from north) and cross (3158 from north) shelf velocity (unit: m s21), respec-tively. Uadcp and Vadcp are from mooring observations. U26N and V26N are model-simulated depth-averaged velocity at (124.58E, 268N), andU26.5N and V26.5N are model-simulated depth-averaged velocity at (124.58E, 26.58N). (B) Tadcp is the bottom temperature (unit: 8C) frommooring observations. T26N and T26.5N are the model-simulated bottom temperature at (124.58E, 268N) and (124.58E, 26.58N), respectively.All time series are filtered by the 10 day running mean.



To fill the spatial and temporal gaps in mooring observations and satellite data, the high-resolution data-assimilative HYCOM outputs for this period were used for further diagnostic analyses. At the sea surface,comparisons of the surface current (Figure 2), drifter trajectories (Figure 3) and sea level anomaly difference(Figure 4A) showed good performance by HYCOM. The deficiency was that both the extent of onshoreintrusion and its westward migration along the 200 m isobath were weaker than observed (Figure 6A). How-ever, the simulated bottom temperature at the mooring station matched well with the observation (Figure6B). In general, the HYCOM outputs resolved the cyclonic eddy and its interaction with Kuroshio reasonably.

To focus on the eddy’s impact, Period H (15 August to 15 September) was defined as the high (before eddyimpact) Kuroshio transport phase, and Period L (1 October to 31 October) was defined as the low (duringeddy impact) Kuroshio transport phase (Figure 4A). The P transect and the 200 m isobath northeast of Tai-wan (see Figure 1) were selected to estimate the variability of both Kuroshio and its intrusion in the subsur-face ocean, respectively. The mean normal velocities and their differences at the two locations are shown inFigure 7, and the volume transports across the two transects are shown in Table 1. With the impact of thecyclonic eddy, the onshore volume transport across the 200 m isobath from 1228E to 1248E decreased by41.7%. It increased (37.9%) west of 122.68E and decreased (73%) east of 122.68E, showing a distinct west-ward migration from 1238E to 1228E–122.58E. A weakened Kuroshio and an intensified (31.3%) northeast-ward current on the outer shelf can be seen along the P transect. The southwestward flow at the bottomalong the slope disappeared completely. There was a remarkable shoreward axis migration, especially inthe upper layer. The temperature, Brunt-V€ais€al€a frequency, and their differences at the two transects areshown in Figure 8. Usually, the cyclonic eddy traps cold water and propagates westward. The arrival of thecyclonic eddy uplifted isopycnals in the offshore direction and depressed isopycnals around the sloperegion. As a result, in the main intrusion region (west of 123.258E), the vertical stratification was weakened(enhanced) in the upper (lower) layer around the shelf break region and enhanced (weakened) in the upper(lower) layer in the deep ocean side. So the cyclonic eddy changed the thermocline tilting across the 200 misobath northeast of Taiwan and weakened cross-shelf PV gradient in the upper layer, which played the

Figure 7. Profiles of mean normal velocities (unit: m s21) at the 200 m isobath northeast of Taiwan (positive values are onshore, top) and at the P transect (positive values are northeast-ward, bottom) during (A) Period H, (B) Period L, and (C) the difference (B-A). The contour interval is 0.1 m s21.



similar role with the anomalous positive wind stress curl and surface cooling as shown in Wang and Oey[2014, 2016].

3.2. Dynamical Interpretations of the Cyclonic Eddy-Induced Kuroshio Intrusion VariabilityHow does the cyclonic eddy modulate the Kuroshio and enable it to overcome the PV barrier caused by thesteep slope and intrude onto the ECS shelf? To answer this question, we followed the PV budget analysisadopted by Guo et al. [2006] and Oey et al. [2010] and used the HYCOM reanalysis circulation fields to exam-ine the role of the density field in the eddy-induced Kuroshio intrusion in terms of the depth-averaged vor-ticity balance during Periods H and L. After vertically averaging the primitive momentum equations andthen cross differentiating, the vorticity equation can be written as

@

@trz3

MH

� �1M � r f

H

� �5Jðv;H21Þ1rz3

ss2sb

q0H

� �2rz3

AH

� �1rz3

DH

� �; (3)

Figure 8. Profiles of mean Brunt-V€ais€al€a frequency (color, unit: 1023 s21) at the 200 m isobath northeast of Taiwan (top) and P transect (bottom) during (A) Period H, (B) Period L, and(C) the difference (B-A). Contours are potential density (unit: kg m23) during Period H (black) and Period L (green).

Table 1. Volume Transports Across the 200 m Isobath and P Transect During Each Time Perioda

200 m Isobath P Transect

Time Period 1228E–122.68E 122.68E–1248E 1228E–1248E 25.78N–26.58N

Period L 2.95 Sv 37.9% 1.47 Sv 273% 4.42 Sv 241.7% 3.27 Sv 31.3%Period H 2.14 Sv 5.44 Sv 7.58 Sv 2.49 SvAnticyclonic Eddies 2.03 Sv 24.7% 4.75 Sv 13.4% 6.78 Sv 7.3% 1.96 Sv 210.1%JA 2.13 Sv 4.19 Sv 6.32 Sv 2.18 SvCyclonic Eddies 2.82 Sv 30.6% 1.63 Sv 247.8% 4.45Sv 215.7% 3.0 Sv 31%SO 2.16 Sv 3.12 Sv 5.28 Sv 2.29 Sv

aThe percentage is the volume transport anomaly during the impact of mesoscale eddies.



where JðA; BÞ5@A=@x@B=@y2@A=@y@B=@x and rz3ðCÞ5@Cy=@x2@Cx=@y; M5Ð 02H~udz is a volume trans-

port vector with~u being the velocity; Jðv;H21Þ is the JEBAR term with v5Ð 02H zgq=q0dz; ss is the wind stress;

sb5CDj�ubj�ub is the bottom stress, CD52:231023 is the drag coefficient, and �ub represents the averaged-velocity in the bottom layer (10 m above the seafloor); and two vectors, A5Ax~i1Ay~j and D5Dx~i1Dy~j , arethe nonlinear advection term and the vertically intergrated horizontal diffusion term, respectively. TheJEBAR term is the only term related to the density field, which was used to diagnose the contribution of thevertical stratification change to the Kuroshio intrusion across the ECS shelf break [Mertz and Wright, 1992].

The mean values of each PV term in equation (3) during Periods H and L were estimated along the 200 misobath northeast of Taiwan. All other terms were smaller (not shown here) and negligible, except the APV(advection of the geostrophic potential vorticity, second term in the left-hand side of equation (3)) andJEBAR terms (Figures 9A and 9B). The positive and negative APV terms correspond to onshore and offshoretransport, respectively. In both periods, the JEBAR term basically balanced the APV term along the 200 misobath, and the variability of the two terms coincided well with that of the Kuroshio intrusion. These twoterms increased west of 122.68E and decreased east of 122.68E from Period H to Period L. These results sug-gest that JEBAR is the main term determining the cross-shelf advection. In other words, the change in localstratification, which appears as the JEBAR term, is responsible for the Kuroshio intrusion variability. A similarexplanation was proposed by Oey et al. [2010] and Wang and Oey [2014], who attributed enhanced Kur-oshio intrusion to an increased JEBAR term along the ECS shelf break introduced by strong sea surface windcurl and cooling.

In addition to the impact of the cyclonic eddy, surface heat flux also has an effect on the density field, whichmay account for part of the JEBAR term. It is hard to separate the contribution of surface heat flux to theJEBAR term. The mean net heat flux (from the ERA-interim data set) on the shelf (1228E–1248E, 25.58N–278N)was 48.88 and 226.78 W m22 during Periods H and L, respectively. Oey et al. [2010] produced increased

Figure 9. Mean APV and JEBAR terms (unit: s22) in a meridional range of 1/38 centered on the 200 m isobath northeast of Taiwan during (A) Period H and (B) Period L. (C) Depth-averaged PV (unit: s21 m21) in the upper 100 m at the P transect during Period H and L. Mean PV (color, unit: s21 m21) profiles during (D) Period H, (E) Period L, and (F) the difference(E-D). Contours are potential density (unit: kg m23) during Period H (black) and Period L (green).



onshore velocity of about 30% in winter with a seasonal heat flux of 6400 W m22. If we use this result as areference, the onshore velocity anomaly induced by the surface heat flux during Period L will be about 3%and thus negligible. Therefore, the local ocean vertical stratification change during Period L was mainlycaused by the cyclonic eddy.

The above diagnostic analyses indicate that the strong Kuroshio intrusion in October 2008 is mainly due tothe large JEBAR term from eddy-induced local density field change. But how could the density change helpthe Kuroshio overcome the topographic PV barrier in the slope region? The constraint of bottom topogra-phy on the current in the upper layer is determined by the vertical structure of density field. In the case of abarotropic fluid, the current can fully feel the topography variation when it collides with a slope. However,the baroclinicity of the fluid could reduce the bottom topography PV constraint to the upper layer. Takingthe ocean as a two-layer fluid and assuming that fluid in the lower layer is nearly stagnant, the topographyvariation in the lower layer does not affect the upper layer current. As mentioned earlier, in the lower oceanlayer around the shelf-break, the vertical stratification was enhanced and would eliminate the pressure gra-dient, which made the southwestward flow at the bottom disappear completely during Period L (Figures 7and 8). This weakens the bottom topography PV constraint and leads the upper layer Kuroshio to intrudeinto the shelf region.

The PV during the two periods (Figures 9D and 9E) was calculated using the equation q5 f 1fq �

@q@z , where q is

PV, f is the local Coriolis parameter, f5 @v@x 2 @u

@y is relative vorticity, and q is density. In general, due to the dif-ference of vertical stratification, the PV in the upper layer is largest on the shelf and decreases gradually off-shore, which generates a strong PV gradient across the slope and blocks the Kuroshio water intrusion.During the strong intrusion Period L, the cyclonic eddy carried positive PV flux and increased PV in the off-shore direction by lifting the isopycnal layers (Figure 9E). The depth-averaged PV gradient in the upper100 m across the slope (around 25.58N) was weakened from 2.89 3 1029 s21 m21 to 0.34 3 1029 s21 m21

per degree (Figure 9C).

So far, based on mooring observations and HYCOM assimilative data, the large intrusion has mainly beenattributed to the eddy-induced vertical stratification change and PV flux. Are these features and interpreta-tions universal and applicable during other cyclonic eddy events? And how about anticyclonic eddy events?In the following section, we will discuss these issues by using long-term HYCOM reanalysis data.

3.3. Kuroshio Intrusion Variability Induced by Mesoscale EddiesUsing the trajectories of mesoscale eddies detected by Chelton et al. [2011], we traced open ocean eddiesthat arrived eat of Taiwan (1228E–1258E, 228N–258N) from 2004 to 2012. There were 20 anticyclonic eddiesand 24 cyclonic eddies. The arrival of anticyclonic (cyclonic) eddy could rise up (bring down) the sea level atthe east side of Kuroshio and then enhance (weaken) the KVT across the I-Lan Ridge transect. Therefore, theKVT anomaly was used as an indicator of the impact of mesoscale eddies. The extreme positive KVT anoma-ly represents the influence of an anticyclonic eddy, while the extreme negative KVT anomaly represents theinfluence of a cyclonic eddy.

Based on the HYCOM reanalysis data from 2004 to 2012, the mean daily KVT across the I-Lan Ridge was26.54 Sv (anomalies are shown in Figure 10A). The KVT was filtered with a 2 years high-pass filter, and theextreme KVT was defined as events exceeding 1.2 standard deviations (std) of the demeaned time series.Over 9 years, summer and autumn stand out as the seasons having the greatest number of extreme KVTevents (Figure 10B). The total time of the extreme high KVT events in July and August (JA) was 165 days(red stars), and the total time of the extreme low KVT events in September and October (SO) was 122 days(blue stars). There were nine extreme high KVT events in JA with a mean KVT anomaly of 8.44 Sv, whilethere were seven extreme low KVT events in SO with a mean KVT anomaly of 29.46 Sv. In contrast, the KVTanomaly in JA and SO climatology (composited by the percentage of each month) was only 3.09 Sv and21.99 Sv, respectively. In this study, we focused on the extreme KVT events in JA and SO to investigate theimpact of mesoscale eddies on the Kuroshio and its intrusion.

We traced mesoscale eddies east of Taiwan for each of the 16 KVT anomaly events (Figure 10C). Amongthese eddies, the mean amplitude, rotation speed, and radius of the nine anticyclonic eddies were20.9 cm, 39.7 cm s21, and 139 km, respectively. The mean amplitude, rotation speed, and radius of theseven cyclonic eddies were 30.2 cm, 48.3 cm s21, and 139.1 km, respectively. Most of these eddies were



generated in the STCC region near the western boundary; only one anticyclonic eddy was tracked back toeast of 1508E.

HYCOM reanalysis fields during these extreme high KVT events in JA and low KVT events in SO were aver-aged to form composite velocity fields at 15 m and sea surface height anomalies to describe the impact ofanticyclonic and cyclonic eddies (Figures 10D and 10E), respectively. Similar to Gawarkiewicz et al. [2011],we find that during the impact of anticyclonic eddies, the sea surface height composite showed a positiveanomaly east of Taiwan and a weak negative anomaly at the southern edge of the ECS shelf. The Kuroshiowas intensified, but the onshore intrusion northeast of Taiwan was inconspicuous. In contrast, during theimpact of cyclonic eddies, the sea surface height composite showed a negative anomaly east of Taiwan and

Figure 10. (A)Two years high-pass filtered KVT anomalies (unit: Sv) derived from HYCOM simulation results. Dashed lines are 61.2 std (5.69 Sv) of KVT anomalies. The red stars corre-spond to the KVT anomalies that are higher than 1.2 std in JA, while the blue stars correspond to the KVT anomalies that are lower than 21.2 std in SO. (B) Monthly distribution of theextreme KVT anomalies which exceed the 1.2 std ranges of the demeaned time series. (C) Trajectories of nine eddies. Red and blue lines correspond to anticyclonic and cyclonic eddies,respectively. ‘‘1’’ and ‘‘o’’ corresponds to the position of eddy’s generation and end, respectively. (D) Composite fields of extreme high KVT events in JA and (E) low KVT events in SO:mean velocity at 15 m depth (arrows) and sea level anomaly (color, unit: m).



a larger positive anomaly at the southern edge of the ECS shelf. The Kuroshio was weaker but had a largermeander northeast of Taiwan. The onshore intrusion reached the midshelf, consistent with the statisticalresults of 20 year drifter tracks [V�elez-Belch�ı et al., 2013, Figure 8]. In addition, the directions of the flowthrough the Kerama Gap in the two composites were opposite, which should be associated with the impactof mesoscale eddies east of Taiwan.

The mean normal velocities and their anomalies during the impact of eddies at the 200 m isobaths and Ptransect were calculated, and the volume transports across them are shown in Table 1. In contrast to theresults of JA climatology, it shows an intensified Kuroshio and a weakened (10.1%) northeastward currenton the outer shelf with the impact of anticyclonic eddies (Figure 11). In addition, the southwestward flowat the bottom along the slope was enhanced. The onshore volume transport across the 200 m isobathfrom 1228E to 1248E increased by 7.3%. However, it decreased (4.7%) west of 122.68E and increased(13.4%) east of 122.68E. With the impact of cyclonic eddies, a weakened Kuroshio and an intensified (31%)northeastward current on the outer shelf were shown compared with the results of SO climatology(Figure 12). The southwestward flow at the bottom along the slope almost disappeared. There was a sig-nificant shoreward axis migration, especially in the upper layer. The onshore volume transport across the200 m isobath from 1228E to 1248E decreased by 15.7%. However, it increased (30.6%) west of 122.68Eand decreased (47.8%) east of 122.68E, which shows a conspicuous westward migration of Kuroshiointrusion.

The temperature, the Brunt-V€ais€al€a frequency, and their anomalies at the two transects during eddy impactwere also calculated. During the impact of anticyclonic eddies, isopycnals were uplifted around the shelf-break and depressed in the offshore region (Figure 13). The stratification was enhanced (weakened) in theupper (lower) ocean around the shelf-break. In the offshore region, the stratification was weakened in theupper layer and almost remained unchanged in the lower layer. The impact of cyclonic eddies on the strati-fication is opposite comparing to that of anticyclonic eddies (Figure 14).

Figure 11. Profiles of mean normal velocities (unit: m s21) at the 200 m isobath northeast of Taiwan (positive values are onshore, top) and P transect (positive values are northeastward,bottom) during (A) the impact of anticyclonic eddies, (B) JA climatology, and (C) the difference (A-B). The contour interval is 0.1, 0.1, and 0.05 m s21 in A, B, and C, respectively.



We also performed PV budget analyses with the HYCOM reanalysis data to examine the role of the densityfield in the Kuroshio intrusion during four periods: the impact of anticyclonic eddies, JA climatology, theimpact of cyclonic eddies, and SO climatology. All other PV terms in equation (3) are smaller (not shownhere) and negligible along the 200 m isobath northeast of Taiwan, except the APV and JEBAR term. In thefour periods, the JEBAR term basically balances with the APV term along the 200 m isobath (Figure 15), andthe variability of the two terms coincides well with the Kuroshio intrusion. These two terms decreased(increased) west of 122.68E and increased (decreased) east of 122.68E with the impact of anticyclonic(cyclonic) eddies, revealing a large westward migration of the maximum with cyclonic eddies. These resultssuggest that the change in local vertical stratification is responsible for the intrusion variability. Both the sur-face net heat flux (from the ERA-interim data set) and its horizontal gradient during the eddy periodsshowed little change (Figure 16), so the contribution of surface heat flux forcing to the variation of JEBARterm is negligible. Thus, we concluded that the variability of ocean vertical stratification and Kuroshio intru-sion during these extreme KVT events was mainly determined by mesoscale eddies.

The dynamical interpretations of the Kuroshio intrusion variability, proposed from a previous single cycloniceddy event, were also verified. During the impact of anticyclonic eddies, a strengthened PV constraint ofbottom topography on the upper layer current from the joint effect of weakened vertical stratification andenhanced southwestward flow in the lower layer at the shelf break can be derived. The anticyclonic eddyinputs additional PV flux during its interaction with the Kuroshio, decreasing PV offshore (Figure 17C). As aresult, the depth-averaged PV gradient in the upper 50 m across the slope (around 25.58N) is increasedfrom 2.18 3 1029 s21 m21 to 2.94 3 1029 s21 m21 per degree (Figure 18A), which inhibits the Kuroshioonshore intrusion. During the impact of cyclonic eddies, a weakened PV constraint of bottom topographyon the upper layer current from the joint effect of intensified vertical stratification and weakened south-westward flow in the lower layer at the shelf-break region can also be derived. The cyclonic eddy inputsadditional PV flux during its interaction with the Kuroshio, increasing PV offshore (around 24.68N in Figure

Figure 12. Profiles of mean normal velocities (unit: m s21) at the 200 m isobath northeast of Taiwan (positive values are onshore, top) and P transect (positive values are northeastward,bottom) during (A) the impact of cyclonic eddies, (B) SO climatology, and (C) the difference (A-B). The contour interval is 0.1, 0.1, and 0.05 m s21 in A, B, and C, respectively.



17F). As a result, the depth-averaged PV gradient in the upper 100 m across the slope (around 25.58N) is sig-nificantly reduced from 1.19 3 1029 s21 m21 to 0.48 3 1029 s21 m21 per degree (Figure 18B), enabling theKuroshio to overcome the PV barrier and intrude onto the midshelf.

In this section, based on the 9 year HYCOM assimilative data, the variability of the Kuroshio, the onshoreintrusion, and surrounding ocean vertical stratification were investigated through nine anticyclonic (JA) andseven cyclonic (SO) eddy events. The variability induced by multiple cyclonic eddy events is consistent withthat of the single cyclonic eddy event in October 2008. The variability induced by anticyclonic eddies wasgenerally opposite to that of cyclonic eddies, and the onshore intrusion variability with cyclonic eddies waslarger in this study. The dynamical interpretations on how the current in the upper layer overcomes the ECSslope PV constraint are applicable to multiple strong eddy events.

4. Summary and Discussion

This study provides more information on the eddy-induced variability of both Kuroshio and its intrusionnortheast of Taiwan and also provides dynamical interpretations.

A cyclonic eddy-induced large Kuroshio onshore intrusion event during October 2008 northeast of Taiwanwas captured by multiple observations. ADCP mooring observations provided direct in situ current observa-tions for the cyclonic eddy-induced Kuroshio onshore intrusion. The concurrent HYCOM reanalysis datarevealed the variability of the Kuroshio, the onshore intrusion, and surrounding ocean vertical stratification.Based on the variability characteristics, the ECS slope’s PV constraint on the current in the upper layer andthe cross-slope PV barrier were proposed as dynamical interpretations. Further, nine anticyclonic eddyevents in JA and seven cyclonic eddy events in SO were extracted from long-term HYCOM reanalysis data,and the generality of intrusion and the applicability of the dynamical interpretations were verified.

Figure 13. Profiles of mean Brunt-V€ais€al€a frequency (color, unit: 1023 s21) at the 200 m isobath northeast of Taiwan (top) and P transect (bottom) during (A) the impact of anticycloniceddies, (B) JA climatology, and (C) the difference (A-B). Contours are potential density (unit: kg m23) during JA climatology (black) and the impact of anticyclonic eddies (green).



Figure 14. Profiles of mean Brunt-V€ais€al€a frequency (color, unit: 1023 s21) at the 200 m isobath northeast of Taiwan (top) and P transect (lower panels) during (A) the impact of cycloniceddies, (B) SO climatology, and (C) the difference (A-B). Contours are potential density (unit: kg m23) during SO climatology (black) and the impact of cyclonic eddies (green).

Figure 15. Mean APV and JEBAR terms (unit: s22) in a meridional range of 1/38 centered on the 200 m isobath northeast of Taiwan during(A) the impact of anticyclonic eddies, (B) JA climatology, (C) the impact of cyclonic eddies, and (D) SO climatology.



Under the impact of cyclonic eddies, the Kuroshio weakened, the northeastward current on the outer shelfintensified, and the Kuroshio intruded far onshore northeast of Taiwan. The onshore velocity increased westof 122.68E and decreased east of 122.68E along the 200 m isobath, which showed the distinct westwardmigration of the Kuroshio intrusion. Stratification was weakened in the upper layer and enhanced in thelower layer around the shelf break. In the offshore region, stratification was enhanced in the upper layer. Ingeneral, the impacts of the anticyclonic eddy were opposite to those of the cyclonic eddy. The variability ofKuroshio onshore intrusion northeast of Taiwan was also quantified. Under the impact of cyclonic (anticy-clonic) eddy, although the total onshore intrusion between 1228E and 1248E across the 200 m isobaths wasdecreased (increased) by 15.7% (7.3%), the onshore intrusion west of 122.68E and the northeastward currenton the outer shelf was increased (decreased) by 30.6% (4.7%) and 31% (10.1%), respectively. Most of intru-sion turned back into the offshore region immediately, while the decreased and westward migratoryonshore intrusion reached a more onshore region (Figure 10E), which could have more substantive impacton the physical and biochemical environment of the ECS.

It is worth noting that the variability induced by cyclonic eddies is larger than that induced by anticycloniceddies. This might be due to the mean intensity of cyclonic eddies being stronger, with larger amplitudeand rotation speed. Additionally, the selected cyclonic and anticyclonic eddies events are in SO and JA,respectively. The mean PV gradient across the ECS slope is weaker in SO than in JA (compare Figure 17Bwith Figure 17E), so it should be easier to modulate the intrusion with the same eddy intensity in SO.

The comprehensive diagnostic analyses showed that the Kuroshio onshore intrusion variability was mainlyattributed to the eddy-induced local vertical stratification variation, while the attributions of sea surfacewind and heat flux forcing were negligible. The dynamical interpretations were summarized as follows.Under the impact of cyclonic eddies, the joint effect of enhanced vertical stratification and weakened south-westward flow in the lower layer weakened the ECS slope’s PV constraint on the upper layer current, whichled to more Kuroshio onshore intrusion. The positive PV flux by cyclonic eddies resulted in a positive PVanomaly in the offshore direction around the shelf break, which weakened the cross-shelf PV gradient and

Figure 16. Mean sea surface net heat flux (unit: W m22) during (A) the impact of anticyclonic eddies, (B) JA climatology, (D) the impact of cyclonic eddies, and (E) SO climatology. Meansea surface net heat flux anomaly during the impact of (C) anticyclonic eddies (A-B) and (F) cyclonic eddies (D-E). The contour interval is 10 W m22. The thick black contours are thecoastline, and the gray contours are the 200 m isobaths.



favored the Kuroshio onshore intrusion. With anticyclonic eddies, the joint effect of weakened vertical strati-fication and enhanced southwestward flow in the lower layer strengthened the ECS slope’s PV constrainton the upper layer current, which inhibited Kuroshio onshore intrusion.

Figure 18. Depth-averaged PV (unit: s21 m21) in the upper 50 m during the impact of anticyclonic eddy and JA climatology (A), and in theupper 100 m during the impact of cyclonic eddy and SO climatology (B) at the P transect.

Figure 17. Mean PV (color, unit: s21 m21) profiles during (A) the impact of anticyclonic eddies, (B) JA climatology, (D) the impact of cyclonic eddies, and (E) SO climatology along the Ptransect. Mean PV anomaly during the impact of (C) anticyclonic eddies (A-B) and (F) cyclonic eddies (D-E). Contours are the potential density (unit: kg m23).



ReferencesArgo (2000), Argo float data and metadata from Global Data Assembly Centre (Argo GDAC), SEANOE, doi:10.17882/42182.Chang, Y. L., and L. Y. Oey (2011), Interannual and seasonal variations of Kuroshio transport east of Taiwan inferred from 29 years of tide-

gauge data, Geophys. Res. Lett., 38, L08603, doi:10.1029/2011GL047062.Chao, S. Y. (1990), Circulation of East China Sea, a numerical study, J. Oceanogr. Soc. Jpn., 46, 273–295.Chassignet, E. P., et al. (2009), US GODAE global ocean prediction with the hybrid coordinate ocean model (HYCOM), Oceanography, 22(2),

64–75, doi:10.5670/oceanog.2009.39.Chelton, D. B., M. G. Schlax, and R. M. Samelson (2011), Global observations of nonlinear mesoscale eddies, Prog. Oceanogr., 91, 167–216.Chen, C. T. A. (1996), The Kuroshio intermediate water is the major source of nutrients on the East China Sea continental shelf, Oceanol.

Acta, 19(5), 523–527.Chen, C. T. A., R. Ruo, S. C. Pai, C. T. Liu and G. T. F Wong (1995), Exchange of water masses between the East China Sea and the Kuroshio

off northeastern Taiwan, Cont. Shelf. Res., 15, 19–39.Cherian, D. A. and K. H. Brink (2016), Offshore transport of shelf water by deep-ocean eddies, J. Phys. Oceanogr., 46, 3599–3621.Chuang, W. S., and W. D. Liang (1994), Seasonal variability of intrusion of the Kuroshio water across the continental shelf northeast of Tai-

wan, J. Oceanogr., 50, 531–542.Cummings, J. A., and O. M. Smedstad (2013), Variational data analysis for the global ocean, edited by S. K. Park and L. Xu, Data Assimilation

for Atmospheric, Oceanic and Hydrologic Applications, vol. II, Springer, Berlin.Dee, D., et al. (2011), The ERA-Interim reanalysis: Configuration and performance of the data assimilation system, Q. J. R. Meteorol. Soc.,

137(656), 553–597.Fang, G., B. Zhao, and Y. Zhu (1991), Water volume transport through the Taiwan Strait and the continental shelf of the East China Sea

measured with current meters, in Oceanography of Asian Marginal Seas, pp. 345–348, edited by K. Takano, Elsevier, New York.Gawarkiewicz, G., et al. (2011), Circulation and intrusions Northeast of Taiwan: Chasing and predicting uncertainty in the cold dome,

Oceanography, 24(4), 110–121, doi:10.5670/oceanog.2011.99.Gong, G. C., F. K. Shiah, K. K. Liu, W. S. Chuang, and J. Chang (1997), Effect of the Kuroshio intrusion on the chlorophyll distribution in the

southern East China Sea during spring 1993, Cont. Shelf Res., 17, 79–94.Guan, B., and G. Fang (2006), Winter counter-wind currents off the southeastern China coast: A review, J. Phys. Oceanogr., 62, 1–24.Guo, B., and B. Wan (1995), The fluctuation and evolution of oceanic front in the East China Sea, J. Oceanogr. Huanghai Bohai Seas, 13(2),

1–10.Guo, X., Y. Miyazawa, and T. Yamagata (2006), The Kuroshio onshore intrusion along the shelf break of the East China Sea: The origin of the

Tsushima Warm Current, J. Phys. Oceanogr., 36(12), 2205–2231.Hsin, Y. C., T. L. Chiang, and C. R. Wu (2011), Fluctuations of the thermal fronts off northeastern Taiwan, J. Geophys. Res., 116, C10005, doi:

10.1029/2011JC007066.Hsin, Y. C., B. Qiu, T. L. Chiang, and C. R. Wu (2013), Seasonal to interannual variations in the intensity and central position of the surface

Kuroshio east of Taiwan, J. Geophys. Res. Oceans, 118, 4305–4316, doi:10.1002/jgrc.20323.Hsueh, Y., J. Wang, and C. S. Chern (1992), The intrusion of the Kuroshio across the continental-shelf northeast of Taiwan, J. Geophys. Res.,

97(C9), 14,323–14,330.Huthnance, J. M. (1984), Slope currents and Jebar, J. Phys. Oceanogr., 14(4), 795–810.Ichikawa, K., R. Tokeshi, M. Kashima, K. Sato, T. Matsuoka, S. Kojima, and S. Fujii (2008), Kuroshio variations in the upstream region as seen

by HF radar and satellite altimetry data, Int. J. Remote Sens., 29(21), 6417–6426.Isobe, A. (2008), Recent advances in ocean-circulation research on the Yellow Sea and East China Sea shelves, J. Oceanogr., 64(4), 569–584.Isobe, A., and R. C. Beardsley (2006), An estimate of the cross-frontal transport at the shelf break of the East China Sea with the finite vol-

ume coastal ocean model, J. Geophys. Res. Oceans, 111, C03012, doi:10.1029/2005JC003290.James, C., M. Wimbush, and H. Ichikawa (1999), Kuroshio meanders in the East China Sea, J. Phys. Oceanogr., 29(2), 259–272.Johns, W. E., T. N. Lee, D. X. Zhang, R. Zantopp, C. T. Liu, and Y. Yang (2001), The Kuroshio east of Taiwan: Moored transport observations

from the WOCE PCM-1 array, J. Phys. Oceanogr., 31(4), 1031–1053.Joyce, T. M. (1984), Velocity and hydrographic structure of a Gulf Stream warm-core ring, J. Phys. Oceanogr., 14(5), 936–947.Kondo, M. (1985), Oceanographic investigations of fishing grounds in the East China Sea and the Yellow Sea—I. Characteristics of the

mean temperature and salinity distributions measured at 50m and near the bottom, Bull. Seikai Reg. Fish. Res. Lab., 62, 19–66.Kuo, Y. C., and C. S. Chern (2011), Numerical study on the interactions between a mesoscale eddy and a western boundary current, J. Oce-

anogr., 67(3), 263–272.Liang, X. S, and J. L. Su (1994), A two-layer model for the circulation in the East China Sea, Acta Oceanol. Sin., 13(3), 45–72.Liang, W. D., T. Y. Tang, Y. J. Yang, M. T. Ko, and W. S. Chuang (2003), Upper-ocean currents around Taiwan, Deep Sea Res., Part II, 50(6–7),

1085–1105.Liu, C., F. Wang, X. Chen, and J. S. Storch (2014), Interannual variability of the Kuroshio onshore intrusion along the East China Sea shelf

break: Effect of the Kuroshio volume transport, J. Geophys. Res. Oceans, 119, 6190–6209, doi:10.1002/2013JC009653.Liu, H., G. Gong, and J. Chang (2010), Lateral water exchange between shelf-margin upwelling and Kuroshio waters influences phosphorus

stress in microphytoplankton, Mar. Ecol. Prog. Ser., 409, 121–130.Liu, X., C. Dong, D. Chen, and J. Su (2014), The pattern and variability of winter Kuroshio intrusion northeast of Taiwan, J. Geophys. Res.

Oceans, 119, 5380–5394, doi:10.1002/2014JC009879.Lumpkin, R., N. Maximenko, and M. Pazos (2012), Evaluating where and why drifters die, J. Atmos. Oceanic Technol., 29(2), 300–308.Matsuno, T., J. S. Lee, and S. Yanao (2009), The Kuroshio exchange with the South and East China Seas, Ocean Sci., 5(3), 303–312.Mertz, G., and G. D. Wright (1992), Interpretations of the JEBAR term, J. Phys. Oceanogr., 22, 301–305.Oey, L. Y., and H. C. Zhang (2004), The generation of subsurface cyclones and jets through eddy–slope interaction, Cont. Shelf Res., 24(18),

2109–2131.Oey, L. Y., Y. C. Hsin, and C. R. Wu (2010), Why does the Kuroshio northeast of Taiwan shift shelfward in winter?, Ocean Dyn., 60(2), 413–426.Price, J. F., and M. A. Sundermeyer (1999), Stratified Ekman layers, J. Geophys. Res., 104(C9), 20,467–20,494.Qiu, B., T. Toda, and N. Imasato (1990), On Kuroshio front fluctuations in the East China Sea using satellite and in situ observational data, J.

Geophys. Res., 951(C10), 18,191–18,204.Sassa, C., and Y. Tsukamoto (2010), Distribution and growth of Scomber japonicus and S. australasicus larvae in the southern East China Sea

in response to oceanographic conditions, Mar. Ecol. Prog. Ser., 419, 185–199.Sheremet, V. A. (2001), Hysteresis of a western boundary current leaping across a gap, J. Phys. Oceanogr., 31, 1247–1259.

AcknowledgmentsWe are grateful to anonymousreviewers for their inspiring andconstructive comments, which arehelpful in improving the work. Theprocessed mooring data used toconstruct figures in this work are alsoavailable, and anyone who wants toget access to these data could contactthe corresponding author, Xiaopei Lin([email protected]). Y.Y. wassupported by the National NaturalScience Foundation of China(41506021, 41476018, and 41421005),Global Change and Air-Sea InteractionProject (GASI-IPOVAI-01-06), CASStrategic Priority Project(XDA11020101), and the PostdoctoralApplied Research Project of Qingdao(2015217). L.X. was supported byChina’s National Basic ResearchPriorities Program (2013CB956202).



http://dx.doi.org/10.17882/42182

http://dx.doi.org/10.1029/2011GL047062

http://dx.doi.org/10.5670/oceanog.2009.39

http://dx.doi.org/10.5670/oceanog.2011.99

http://dx.doi.org/10.1029/2011JC007066

http://dx.doi.org/10.1002/jgrc.20323

http://dx.doi.org/10.1029/2005JC003290

http://dx.doi.org/10.1002/2013JC009653

http://dx.doi.org/10.1002/2014JC009879

Soeyanto, E., X. Guo, J. Ono, and Y. Miyazawa (2014), Interannual variations of Kuroshio transport in the East China Sea and its relation tothe Pacific Decadal Oscillation and mesoscale eddies, J. Geophys. Res. Oceans, 119, 3595–3616, doi:10.1002/2013JC009529.

Su, J. L., and Y. Q. Pan (1987), On the shelf circulation north of Taiwan, Acta Oceanol. Sin., 6(supp. I), 1–20.Tang, T. Y., J. H. Tai, and Y. J. Yang (2000), The flow pattern north of Taiwan and the migration of the Kuroshio, Cont. Shelf Res., 20(4–5),

349–371.Tsai, C. J., M. Andres, S. Jan, V. Mensah, T. B. Sanford, R. C. Lien and C. M. Lee (2015), Eddy-Kuroshio interaction processes revealed by moor-

ing observations off Taiwan and Luzon, Geophys. Res. Lett., 42(19), 8098–8105.Uda, M. (1936). Results of simultaneous oceanographic investigations in the Japan Sea and its adjacent waters during October and Novem-

ber, 1933, J. Imp. Fish. Exp. Stn., 7, 91–151.V�elez-Belch�ı, P., Centurioni, Luca R., Lee, D.-K., Jan, Sen, and Niiler, Pearn P. (2013), Eddy induced Kuroshio intrusions onto the continental

shelf of the East China Sea, J. Mar. Res., 71(1–2), 83107.Wang, J., and L. Y. Oey (2014), Inter-annual and decadal fluctuations of the Kuroshio in East China Sea and connection with surface fluxes

of momentum and heat, Geophys. Res. Lett., 41, 8538–8546, doi:10.1002/2014GL062118.Wang, J., and L. Y. Oey (2016), Seasonal exchanges of the Kuroshio and shelf waters and their impacts on the shelf currents of the East Chi-

na Sea, J. Phys. Oceanogr., 46(5), 1615–1632.Wu, C. R., Y. C. Hsin, T. L. Chiang, Y. F. Lin, and I. Tsui (2014), Seasonal and interannual changes of the Kuroshio intrusion onto the East China

Sea Shelf, J. Geophys. Res. Oceans, 119, 5039–5051, doi:10.1002/2013JC009748.Yang, D., B. Yin, Z. Liu, T. Bai, J. Qi, and H. Chen (2012), Numerical study on the pattern and origins of Kuroshio branches in the bottom

water of southern East China Sea in summer, J. Geophys. Res., 117, C02014, doi:10.1029/2011JC007528.Yin, Y., X. Lin, Y. Li, and X. Zeng (2014), Seasonal variability of Kuroshio intrusion northeast of Taiwan Island as revealed by self-organizing

map, Chin. J. Oceanol. Limnol., 32(6), 1435–1442.Yuan, Y., and J. Su (1984), Numerical modeling of the circulation in the East China Sea, in Ocean Hydrodynamics of Japan and East China

Seas, edited by T. Ichiye, pp. 167–186, Elsevier, New York.Yuan, D., and Z. Wang (2011), Hysteresis and dynamics of a western boundary current flowing by a gap forced by impingement of meso-

scale eddies, J. Phys. Oceanogr., 41, 878–888.Zhang, D. X., T. N. Lee, W. E. Johns, C. T. Liu, and R. Zantopp (2001), The Kuroshio east of Taiwan: Modes of variability and relationship to

interior ocean mesoscale eddies, J. Phys. Oceanogr., 31(4), 1054–1074.Zhang, W. G., and G. G. Gawarkiewicz (2015), Dynamics of the direct intrusion of Gulf Stream ring water onto the Mid–Atlantic Bight shelf,

Geophys. Res. Lett., 42, 7687–7695, doi:10.1002/2015GL065530.



http://dx.doi.org/10.1002/2013JC009529

http://dx.doi.org/10.1002/2014GL062118

http://dx.doi.org/10.1002/2013JC009748

http://dx.doi.org/10.1029/2011JC007528

http://dx.doi.org/10.1002/2015GL065530

Documents

Impact of mesoscale eddies on Kuroshio intrusion ...oomg.meas.ncsu.edu/wp-content/uploads/2017/04/Yin-etal_JGR_201… · RESEARCH ARTICLE 10.1002/2016JC012263 Impact of mesoscale