Access the full text.
Sign up today, get DeepDyve free for 14 days.
References for this paper are not available at this time. We will be adding them shortly, thank you for your patience.
MAY 2022 XU E T A L . 873 a a,b a,b a a a a,b QINBO XU, CHUN ZHOU, WEI ZHAO, QIANWEN HU, XIN XIAO, DONGQING ZHANG, FAN YANG, a,b a,b XIAODONG HUANG, AND JIWEI TIAN Frontier Science Center for Deep Ocean Multispheres and Earth System and Physical Oceanography Laboratory/Sanya Oceanographic Institution, Ocean University of China, Qingdao/Sanya, China Qingdao National Laboratory for Marine Science and Technology, Qingdao, China (Manuscript received 15 September 2021, in final form 27 January 2022) ABSTRACT: Intraseasonal fluctuation with periods of ∼90 days in the South China Sea (SCS) basin is investigated based on an array of seven subsurface moorings. In the deep layer, the 90-day fluctuation is revealed to contribute significantly to the variability in the current, accounting for ∼69% of the subinertial variance. This fluctuation propagates westward along the mooring section with a phase speed of ∼4.6 cm s . In the upper layer, the fluctuation also propagates westward with a similar phase speed, but with opposite phase to that of the deep layer. These results suggest that the 90-day fluctuation regulating the abyssal SCS should be the first mode baroclinic Rossby wave. A set of experiments based on a two-layer dynamic model reveal that both the local wind stress curl and the flow originating from the North Pacific through the Luzon Strait contribute to drive the 90-day fluctuation in the deep SCS, while the latter plays the dominant role. KEYWORDS: Ocean; Abyssal circulation; Rossby waves 1. Introduction Strait, with the deepest sill at approximately 2400 m, is the only deep connection between the SCS and the northwestern Pacific. The deep layer of the ocean has long been improperly 23 2 21 As a result of strong diapyncal mixing (∼10 m s )inthe regarded as inactive. Until recent decades, this concept was nulli- deep SCS (Tian et al. 2009; Yang et al. 2016), there is a persis- fied by multiple deep-ocean observations, revealing considerable tent pressure gradient driving a deep overflow from the Pacific motions thousands of meters below the global ocean surface. For into the SCS (Liu and Liu 1988; Qu et al. 2006b; Tian et al. example, in the North Atlantic, low-frequency variability in the 2006; Chang et al. 2010; Zhao et al. 2014; Zhou et al. 2014; deep current with periods of several days to hundreds of days Zhao et al. 2016; Ye et al. 2019). The cold dense water from the has been revealed (Thompson and Luyten 1976; Hogg 1981; Pacific sinks to the deep SCS after crossing the Luzon Strait Hamilton 1990; Pickart and Watts 1990; Pen ˜a-Molino et al. 2012; (Wyrtki 1961), which then drives basin-scale cyclonic circulation Hamilton et al. 2019); the deep western boundary current in the in the deep SCS (Wang et al. 2011; Lan et al. 2013; Shu et al. Southern Pacific is modulated by variability with a period of 2014; Zhou et al. 2017; Gan et al. 2016). It then upwells and 10–50 days (Whitworth et al. 1999); and in the Indian Ocean’s leaves the SCS through a number of shallow passages (Qu et al. Mascarene Basin, deep water has large velocity fluctuations of 2005). Shu et al. (2014) suggested that one possible mechanism 40–60 days (Quadfasel and Swallow 1986; Scha ¨r and Davies for the upwelling in the abyssal SCS is the interaction between 1988; Warren et al. 2002). In addition, a recent study in the South the continental slope-trapped waves and the planetary Rossby China Sea (SCS) reported that the temporal variability in the waves. Moreover, bottom topography interacted with strong deep western boundary current is dominated by intraseasonal low-frequency periodic abyssal flows could trigger internal fluctuations by a period of approximately 90 days and that the wave and catalyze internal wave energy cascade toward abyssal amplitude of the deep western boundary current variability mixing (Nikurashin and Ferrari 2010; Liang and Thurnherr exceeds its mean value (Zhou et al. 2017). 2012; Hu et al. 2020). Because the SCS circulation acts as a heat The SCS, with a maximum water depth of over 5000 m and and freshwater conveyor between the Pacific and Indian Oceans 6 2 an area of approximately 3.5 3 10 km , is the largest marginal (Qu et al. 2006a, 2009; Gordon et al. 2012), the temporal varia- sea in the northwestern PacificOcean (Fig. 1a). The SCS is a tions of the deep SCS circulation may play a potentially impor- semienclosed basin surrounded by Asian landmasses, the tant role in regulating the regional thermohaline circulation and Philippine Islands, and the Great Sunda Islands. The Luzon climate. In recent years, the temporal variations in the deep SCS have attracted increasing attention. Based on current meas- Denotes content that is immediately available upon publica- urements from six year-long current meter moorings, Zhou tion as open access. et al. (2017) found that the temporal variability in the deep western boundary current in the SCS is dominated by intra- Supplemental information related to this paper is available seasonal fluctuations at an approximately period of 90 days at the Journals Online website: https://doi.org/10.1175/JPO-D-21- with a mean westward phase speed of ∼2.9 cm s along the 0207.s1. mooring section. They further suggested that the 90-day fluc- tuation cannot be attributed to the topographic Rossby wave Corresponding author: Chun Zhou, chunzhou@ouc.edu.cn because the direction of the phase propagation for the 90-day DOI: 10.1175/JPO-D-21-0207.1 Ó 2022 American Meteorological Society. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses). 874 J OUR N A L O F P HY SI C A L O C E A N OGR A P HY VOLUME 52 FIG. 1. Bottom topography of the SCS and locations of the moorings (Smith and Sandwell 1997). (a) Topography of the SCS; (b) regional bathymetry indicated by solid rectangular in (a). The red stars denote the mooring locations. The red arrow in (b) denotes the mean current at HAB-120 m of the D1–D7. fluctuation is at odds with the theoretical topographic Rossby on the intraseasonal variability and the correlation analysis of wave propagation direction. Zhou et al. (2020) revisited the the observations are presented in section 3. The driving mecha- nism of the intraseasonal variability is discussed in section 4. intraseasonal variability in the deep western boundary current through direct observations from three current moorings, and Finally, the results are summarized in section 5. they attributed this variability to the intraseasonal variability in the Luzon Strait overflow or the eddies in the deep SCS 2. Data and model forced by the stable Luzon Strait overflow. Based on long- a. Mooring data term and large-scale observations, Zheng et al. (2021) sug- gested the existence of low-frequency topographic Rossby As part of the SCS deep circulation experiment, an array of waves with a period of nearly 65 days in the northern and seven moorings (D1–D7) was deployed in the middle of the western boundaries of the deep SCS basin. In addition to the deep SCS basin (Fig. 1) during the period from early July temporal variation in the deep western boundary current, the 2015 to early June 2016. The numbering scheme is such that topographic Rossby waves near seamounts and continental D1 is farthest to the east, located in the western area of the slopes has also been studied. Shu et al. (2016) reported ener- Stewart Seamount, and D7 is farthest to the west, located in getic topographic Rossby waves with periods of 9–14 days in the western area of the Xianbei Seamount. the southern part of the SCS based on 5 years of direct meas- At D2 and D5, one upward-looking 75-kHz acoustic Doppler urements. Wang et al. (2019) observed topographic Rossby current profiler (ADCP) was mounted on the mooring at a waves with peak spectral energy at ∼14.5 days around the nominal depth of ∼550 m. Thirty-three current meters, includ- Dongsha Islands by deep moorings. Both Shu et al. (2016) and ing Aanderaa RCM Seaguard and NORTEK Deep Water Wang et al. (2019) identified the energy source of topographic Aquadopp current meters, were mounted on the moorings at Rossby waves associated with mesoscale eddies in the upper nominal depths of around 1000, 2000, 2500, 3000, 3500, and ocean. Quan et al. (2021) further investigated the topographic 120 m above the bottom (HAB). Each NORTEK Deep Water Rossby waves in the abyssal SCS and suggested two possible Aquadopp current meters were assembled with temperature energy sources: mesoscale perturbation in the upper layer and and pressure sensors. In addition, SBE 37-SM conductivity– large-scale background circulation in the deep layer. temperature–depths (CTDs) were mounted on the moorings to At present, studies on the temporal variability in the deep record the hydrographic characteristics of the deep water. Cur- SCS mainly focus on boundary areas with complex topography. rent meters and CTDs have standard calibrations from the For the inner region of the deep SCS basin, characteristics of manufacturers. The accuracy of the velocity measurements is 21 21 intraseasonal fluctuations and its dynamic mechanism are 0.15 cm s for the Aanderaa RCM Seaguard, 0.5 cm s for vague due to shortage of direct observations. Based on direct NORTEK Aquadopp and 0.1%S 6 5mm s for the ADCP observations and dynamic models, this paper reveals the ener- (S is the water velocity relative to the ADCP). The accuracies getic intraseasonal fluctuations in the deep basin of the SCS, of the sensors are 0.0028C(0.18C) for the temperature of CTD and intends to present a new mechanism to explain the dynam- (NORTEK current meter), 0.003 mS cm for conductivity, ics, which is helpful for a comprehensive understanding of the and 7 m (15 m) for the pressure of CTD (NORTEK). All of the temporal variability in the deep SCS. In section 2, we describe instruments returned high-quality data. The mooring configura- the data and model applied to the analysis in detail. The results tions are shown in Table 1 and Fig. 2. MAY 2022 XU E T A L . 875 TABLE 1. Design of moorings shown in Fig. 1. Moorings Depth (m) Period Location Instrument Instrument depth (m) Sample interval (min) D1 4000 2 Aug 2015–7 Jun 2016 118.508E, 17.058N RCM 3880, 3480, 2980, 1980 60 NORTEK 2480 60 CTD 3885 30 D2 3970 2 Aug 2015–7 Jun 2016 118.208E, 17.288N ADCP 515 5 RCM 3850, 3550, 2040, 1020 60 NORTEK 2550 60 CTD 3855, 2045, 1025, 520 30 D3 3990 2 Aug 2015–7 Jun 2016 117.918E, 16.918N RCM 3870, 3460, 2940, 1950 60 NORTEK 2450 60 CTD 3875, 3465, 2945, 1955 30 D4 4000 4 Aug 2015–7 Jun 2016 117.618E, 16.848N RCM 3880, 3480, 1980 60 NORTEK 2480 60 CTD 3885, 3485, 1985 30 D5 4000 5 Aug 2015–6 Jun 2016 117.318E, 16.778N ADCP 570 5 RCM 3880, 3580, 3080, 2070, 1070 60 NORTEK 2580 60 CTD 3885, 3085, 2075, 1075, 575 30 D6 4050 2 Jul 2015–6 Jun 2016 117.028E, 16.708N RCM 3930, 2520, 2010 60 CTD 3935, 3525, 2015 30 D7 4060 5 Aug 2015–6 Jun 2016 116.438E, 16.538N RCM 3940, 3540, 3050, 2040 60 NORTEK 2550 60 CTD 3945, 3545, 3055, 2045 30 b. CTD profiles The hydrographic data used for this study consist of 47 temper- ature/salinity profiles in the SCS. Historical hydrographic data, including large numbers of CTD profiles from the World Ocean Database 2013 (WOD13; c. AVISO SSH data and CCMP wind vector Boyer et al. 2013)and ourown field experiments sampled by analysis product Sea-Bird SBE 9 plus, were also used in this study. Profiles that extend shallower than 1500-m depth and those not passing stan- To help interpret the mooring observations, the 1/48 daily dard deviation checks or flagged as outliers were eliminated. gridded altimeter and sea level anomaly (SLA) data during FIG. 2. Section view of topography from southwest to northeast and locations of the current meters, CTDs, and upward-looking ADCPs. 876 J OUR N A L O F P HY SI C A L O C E A N OGR A P HY VOLUME 52 the observation period were also used in this study. Addi- Medium-Range Weather Forecasts analyses (Uppala 2005), tionally, a 1/48 6-hourly cross-calibrated multiplatform was also used. (CCMP) gridded surface vector wind dataset (Atlas et al. d. Numerical model 2009), produced via a variational analysis method combin- A two-layer, nonlinear, and primitive equation model is used ing extensive cross-calibrated multiple satellite datasets to study the relevant dynamics. The model is governed by the fol- with ship and buoy observations and European Centre for lowing set of equations: u u u h l t 1 1 1 x 1 4 () 1 u 1 y 2 fy 2g 2 A= u 2 1 2 F || u u 1 1 1 1 1 1 1 t x y x h r h 1 1 y y y h l t 1 1 1 1 4 () 1 u 1 y 1 fu 2g 2 A= y 2 1 2 F || y y 1 1 1 1 1 1 1 t x y y h r h 1 1 h () h u () h y () h u () h y 1 1 1 1 1 2 2 2 2 1 1 1 1 0 t x y x y (1) u u u h Dr h l 2 2 2 1 2 4 1 u 1 y 2 fy 2g 2 g 2 A= u 2 || u u 2 2 2 2 2 2 t x y x r x h 2 2 y y y h Dr h l 2 2 2 4 1 2 1 u 1 y 1 fu 2g 2 g 2 A= y 2 || y y 2 2 2 2 2 2 t x y y r y h h () h u () h y 2 2 2 2 1 1 0 t x y where (u , y )and h are the velocity and layer thickness in the C grid and a small time step of 10 s. The equation is solved by n n n nth layer (n = 1 and 2 for the upper and lower layers, respec- the central difference scheme in space and the two-step Mat- tively), h is the sea surface height, h is the interface height 1 2 suno difference scheme in time. anomaly, f is the Coriolis parameter, and t and t are the zonal In this paper, a set of sensitivity experiments were con- x y ducted. Each experiment is driven by CCMP wind stress or and meridional surface wind stresses. We choose the smallest vis- flow through the Luzon Strait. The flow through the Luzon cosity and bottom drag just to maintain the stability of calcula- 11 4 21 Strait was obtained from a north–south mooring array consist- tions. A =1 3 10 m s is the biharmonic viscosity, l =0.01is ing of five full-depth moorings deployed on the western side a quadratic bottom drag coefficient. The bottom stress is applied of the Luzon Strait (along the ∼119.98E longitude), which are to the upper layer when the lower layer vanishes and the change described in detail in Sun et al. (2020) denoted as NS1–NS5. of bottom stresses is handled by F (F =1if h . 0, and F =0if The five moorings were deployed for three years from June h = 0). Theinitial layer interface is set at 1000 m in the deep 2014 to June 2017. To obtain the daily mean velocity section basin and at the seafloor wherever the depth is shallower than of Luzon Strait, the daily mean velocities obtained from the 1000 m. According to the mean potential density profile r(z) mooring array are linearly interpolated to north/south bound- (see online supplemental material Fig. S1) estimated over ary of the Luzon Strait by setting zero velocity at the sidewalls. 158–188N and 1168–1198E based on hydrographic data afore- Then, we vertically averaged the velocities above 1000 m and mentioned, we calculate the mean potential density above below 1000 m to obtain the upper and lower flow through the 23 23 1000 m (1026.3 kg m ) and below 1000 m (1027.6 kg m )as Luzon Strait in the model. The daily mean wind stress vector the upper density r and lower density r in the model. and velocity across the Luzon Strait are linearly interpolated 1 2 to each calculation time step. The model is spun up from rest. Accordingly, the density difference between two layers (Dr)is Each experiment is integrated for 2.5 years from July 2014 to 1.3 kg m . The Rossby deformation radius at 168N in the two- December 2016, and the last 2 years are used for analyses. layer model is approximately 70 km for h = 1000 m and h = 1 2 3000 m, which is similar to the estimate based on observation (described in section 3). The model has a resolution of 0.28 in 3. Results the domain from 48 to 238N in the meridional direction and We begin with a brief introduction to the mean state of the 8 8 from 108 to 121 E in the zonal direction. The model uses deep current. By averaging the raw velocity time series ETOPO2 bathymetric data (Smith and Sandwell 1997). No nor- observed by the current meters over the whole observation mal flow or nonslip boundary conditions are applied. All lateral period, we obtain the mean current at all moorings. Below boundaries are closed except the Luzon Strait which is opened 1000 m, the mean currentis weak (generally not exceeding or closed for different experiments. The model uses a staggered 1cm s ) and does not vary significantly in the vertical MAY 2022 XU E T A L . 877 FIG. 3. The 15-day low-passed (a)–(g) y and (h)–(n) u time series observed at D1–D7 with different colors indicating the observation at different depths. direction. The mean currents at HAB-120 m of D1–D7 are smaller than the amplitudes of the subinertial variability, shown in Fig. 1b. Except for D6, the mean current pointed to which will be shown below. the northwest at all moorings, which is consistent with the a. Observed intraseasonal variations cyclonic circulation in the deep SCS basin proposed by previ- ous studies (Wang et al. 2011; Lan et al. 2013; Shu et al. 2014; To focus on the intraseasonal signals, we applied a 15-day Zhou et al. 2017; Gan et al. 2016). The deep current at D6 Butterworth low-pass filter to the raw time series of velocity to points to the southeast, which is different from the current at exclude the effects of tides and the spring–neap cycle. Figure 3 other moorings. This difference may be related to the Xianbei shows the 15-day low-pass time series for the northward (y)and Seamount on the west side of D6. Numerous studies have eastward (u) velocity components at D1–D7. Notably, the intra- found that anticyclonic flows with different mechanisms can seasonal variations are obvious at all the moorings with meridi- form around isolated seamounts (Eriksen 1991; Freeland onal amplitudes of ∼5cm s , and the zonal amplitudes are 1994; Codiga and Eriksen 1997; Lavelle et al. 2003). The raw somewhat smaller. Vertically, the velocity exhibits a coherent velocity time series (see supplemental material Figs. S2–S8) phase and similar amplitude below 1000 m. However, as seen shows notable tidal signals, which are ubiquitous in the ocean. from sites D2 and D5, the variations in the velocity at a depth A Butterworth bandpass filter is applied to the velocity time of 500 m are significantly different from those below 1000 m. series to resolve the semidiurnal and diurnal tidal signals (cut- A spectral analysis of the low-pass meridional velocity com- off frequencies are [1.7 2.3] cycles per day for the semidiurnal ponent y below 1000 m shows that the subinertial variability is tidal signal and [0.85 1.15] cycles per day for the diurnal tidal dominated by an approximately 90-day fluctuation at all of the signal). Obviously, diurnal tides are stronger than semidiurnal moorings (Fig. 4), which is approximate to the features of the tides, with amplitudes reaching ∼4and ∼3cm s ,respectively. deep western boundary current at similar latitudes (Zhou et al. These amplitudes are stronger than the mean current but are 2017, 2020). The peak period of u is usually similar to the peak 878 J OUR N A L O F P HY SI C A L O C E A N OGR A P HY VOLUME 52 FIG.4. (a)–(g) Spectrum analysis of the low-passed y time series at D1–D7. The bold lines show the power spectra of the y time series at different depths. The 95% confidence interval is indicated by the vertical bar. The frequency bands of 90-day fluctuation for the bandpass filtering are shaded. period of y, but the time series of u are generally muddier and into the corresponding distances between moorings gives a mean much less energetic than the time series of y,so we will focus westward phase speed of ∼4.6 cm s along the mooring section, on the meridional velocity y hereinafter. and the corresponding wavelength L is ∼360 km. The 90-day To study the characteristics of the 90-day fluctuation, a But- fluctuation at the deep western boundary has similar propagation terworth bandpass filter with a time window of 70–110 days is characteristics, suggesting that they may be driven by the same applied to all the raw time series of the meridional velocity y mechanism (Zhou et al. 2017). Although the phase propagation below 1000 m (see Fig. S9). By comparing the standard devia- is clearly westward, the energy propagation has a sense of east- tion of the bandpassed y (the mean value of all current obser- ward (Fig. 5a). This propagation characteristics are quite similar vations is approximately 0.98 cm s ) with the standard to the typical short Rossby waves, a candidate for intraseasonal deviation of the low-pass y time series (the mean value of all variability in the deep ocean. current meter observations is approximately 1.42 cm s ), we Below the depth of 1000 m, the 90-day fluctuation exhibits a find that the 90-day fluctuation accounts for ∼69% of the subi- coherent phase vertically and has westward propagation in the nertial variance. The 90-day fluctuation exhibits a coherent phase horizontal direction. As for the upper layer, the spectra of the vertically from 1000 m down to the bottom (Fig. 3). However, 15-day low-pass y for D2 and D5 have similar peaks to those horizontally, the fluctuation is out of phase at different moorings. in the deep ocean (Figs. 6a,b), both of which have 90-day Figure 5a shows the Hovmo ¨ller diagram of bandpassed y along peaks. The bandpassed velocities at D2 (Fig. 6c)and D5 the mooring section at a depth of 2000 m, indicating notable (Fig. 6d) show that the 90-day fluctuation has a consistent westward propagation of the 90-day fluctuation. A lag correlation phase above 500 m and gradually intensifies from 500 m to the analysis for the bandpassed y at 2000 m suggests that significant sea surface. However, the phase of the 90-day fluctuation in lags exist between the western and eastern moorings (Fig. 5b). the upper layer is basically opposite to the phase in the deep At mooring D4, the relative lead times for D1–D7 are 23, 21, 4, layer. Similarly, the 90-day fluctuation signals in the upper 0, 212, 228, and 236 days, respectively. Dividing these times layer also propagate westward. The depth-averaged y above MAY 2022 XU E T A L . 879 FIG. 5. Westward propagation of the 90-day fluctuation. (a) Hovmoller diagram of the bandpassed y (cm s )at depth of 2000 m; black dashed lines are contours of zero velocity. (b) Lag correlation coefficient of bandpassed y between D4 and D1–D7 at depth of 2000 m. 500 m at D2 leads that at D5 by 24 days, and the correspond- b. Dynamic of the 90-day fluctuation ing westward phase velocity is very similar to that of the deep Although previous studies have found a similar fluctuation at layer, approximately 4.7 cm s along the section. the deep western boundary (Zhou et al. 2017, 2020), the dynamic FIG. 6. (a) Spectrum analysis of the 15-day low-passed y time series at D2. The bold lines show the power spectra of the y time series at different depths. The 95% confidence interval is indicated by the vertical bar. The frequency bands of 90-day fluctuation for the bandpass filtering are shaded. (b) As in (a), but for D5. (c) The bandpassed velocity time series for D2 at different depths (red arrows). The bandpassed velocity time series at the bottom (gray arrows) is over- lapped for comparison. (d) As in (c), but for D5. 880 J OUR N A L O F P HY SI C A L O C E A N OGR A P HY VOLUME 52 FIG. 7. (a) Eigenfunctions of the first three baroclinic modes. (b) Depth-integrated kinetic energy of the lowest three modes of the 90-day fluctuation at D2. (c) As in (b), but for D5. of the fluctuation is unclear. Zhou et al. (2017) suggested that the i R , (4) f() u 90-day fluctuation cannot be attributed to the topographic Rossby wave, and the other candidates include barotropic and where f =2Vsinu is the Coriolis parameter for Earth’srotation baroclinic Rossby waves. Here, by analyzing the velocity data rate V and latitude u. With the background Brunt–Vais ¨ al ¨ af ¨ re- obtained from the observations, we find that the 90-day fluctua- 2 quency N (z) calculated from the mean potential density pro- tion propagates westward at all depths and that the phase in the file r(z) (see supplemental material Fig. S9) estimated over upper layer is basically opposite to the phase in the deep layer. 158–188N and 1168–198E mentioned above, we calculate the All ofthe characteristics ofthe 90-day fluctuation are quite simi- first baroclinic Rossby radii of deformation R = 75 km here. lar to those of the first mode baroclinic Rossby wave. The west- Although the propagation direction may be somewhat different ward phase speed of the first baroclinic Rossby wave is from the section, we roughly assume that fluctuation propagate westward along the section and calculate K =2p/L with 211 21 21 c 2 , (2) px 2 2 L = 360 km. Using b =2.2 3 10 m s and R =75km K 1 1=R here, we find that the phase speed for the first baroclinic Rossby wave is 4.6 cm s , suggesting that the 90-day fluctuation should where K is the wavenumber, b is the derivative of the Coriolis be the first barocline Rossby wave. parameter, and R is the first baroclinic Rossby radius of defor- ¨ ¨ ¨ Additionally, the eigenfunctions of the first three baroclinic mation. If the background Brunt–Vaisala frequency N (z)is modes of the 90-day fluctuation are presented in Fig. 7a.When provided, the baroclinic Rossby radii of deformation can be choosing different N (z) near the observed area, the eigenfunc- obtained by solving the Sturm–Liouville eigenvalue problem tions change only marginally. The depth of the first mode node is for the vertical structure of the vertical velocity (Chelton et al. approximately 1000 m, which may account for the weak 90-day 1998), which can be written in the following form: fluctuation signal at 1000-m depth (Fig. 4). The baroclinic veloc- 2 2 F N () z ity [u , y ] associated with the 90-day fluctuation (computed by 1 F 0, 2 2 z c (3) subtracting the depth-averaged current at each time from the bandpassed current) at D2 and D5 can be expressed by the F 0at z 0, 2 H, superposition of discrete modes [] u ,y () u ˆ ,y ˆ F z i i i() i1 where c is the separation constant and H is the local mean (Alford 2003; Nash et al. 2005). Parameters u ˆ and y ˆ are the i i water depth. There is a countable set of decreasing nonnega- time-varying ith modal magnitudes of zonal velocity and meridio- tive eigenvalues c and corresponding eigenfunctions F (z). i i nal velocity, respectively. Here, the 90-day fluctuation is decom- The subscript i denotes the baroclinic mode number. The posed into the first three modes. The shaded areas in Figs. 7b eigenvalues and eigenfunctions are estimated by discretizing and 7c denote the depth-integrated kinetic energy the continuously stratified Eq. (3) and solving the resulting system of equations numerically. Outside of the tropics, the [1=2 r u ˆ 1 y ˆ dz]of the first three modes at moorings D2 i i 2H Rossby radius of deformation can be defined (Gill 1982)as MAY 2022 XU E T A L . 881 FIG.8.(a) Hovmoller diagram of bandpassed SLA along mooring section during the observation period. (b)–(e) Bandpassed SLA pat- terns on 20 Nov 2015, 20 Dec 2015, 20 Jan 2016, and 20 Feb 2016. The black solid line represents the 1000-m isobaths, and the black stars denote the mooring locations. (f) Time–latitude plots of bandpassed SLA along 1178E during the observation period. The solid line indi- cated the bandpassed time series of the ADCP measured y averaged above 100 m at D5. and D5, respectively. The kinetic energy of the first mode days is applied to the SLA time series. Figure 8a shows the accounts for approximately 80% of the total kinetic energy at Hovmoller ¨ diagrams of the bandpassed SLA along the moor- both D2 and D5, further indicating that the 90-day fluctuation ing section during the observation period of the mooring. A should be the first baroclinic Rossby wave. westward propagation fluctuation at an approximately 90-day period is clearly revealed, and the phase speed is estimated to be c. Evidence from satellite altimetry ∼4.4 cm s . Based on the SLA patterns at different times, the To better understand the 90-day fluctuation in the deep distribution of the SLA is banded along the north–south direc- SCS basin, we explored its spatial structure and temporal evo- tion (Figs. 8b–e). As these banded structures propagate westward lution using AVISO SLA data. Similar to the mooring data, a and pass the mooring section, a corresponding 90-day fluctuation Butterworth bandpass filter with a time window of 70–110 signal is observed at the mooring sites. Furthermore, the 882 J OUR N A L O F P HY SI C A L O C E A N OGR A P HY VOLUME 52 FIG. 9. Wavelet analysis of (a) wind stress curl and (b) SLA averaged over a 38 3 38 (latitude 158–188N, longitude 1168–1198E) box centered at the mooring section from 2012 to 2016. The color-filled contours are the variance of the wavelet transform of normalized values. The black solid contours are the 95% confidence level. The black dashed line displays the period of 90 days, and the blue rectangles indicate the mooring observation period. time–latitude plots of bandpassed SLA from 128 to 208N along winter (Fig. 9b). The similarity between Figs. 9a and 9b sug- 1178E show no obvious south–north propagation for the 90-day gests that the 90-day fluctuation in the ocean appears to have fluctuation, and the strong zonal gradients of positive/negative a certain correlation with the wind stress curl. SLA are associated with strong geostrophic meridional currents In addition to the local wind forcing, the disturbances (Fig. 8f). The solid line in Fig. 8f indicates the bandpassed time originating from the Luzon Strait should also play a role in series of the ADCP measured y averaged above 100 m at D5, the modulation of the Rossby wave in the SCS. Previous and a larger amplitude of y matches well with the boundaries studies have found that mesoscale eddies and Rossby waves between the positive and negative SLA bands. In addition, the pass through the Luzon Strait from the Pacific to the SCS north–south banded structure of the SLA ranges from 128 to (Zhang et al. 2013; Xie et al. 2016), and the deep overflow in 8 the Luzon Strait also has energetic intraseasonal variations 20 N, meaning that 90-day fluctuation exist throughout the deep with a period of 20–100 days (Zhou et al. 2014; Zhao et al. SCS basin. The 90-day fluctuation observed at the deep western 2016; Ye et al. 2019). Zhou et al. (2020) suggested that the boundary current in the SCS also has similar characteristics intraseasonal variability in the deep western boundary current (Zhou et al. 2017), which further confirmsthisconjecture. in the SCS may be related to the intraseasonal variability in the Luzon overflow. These results indicate that the disturban- 4. Driving mechanism ces from the Luzon Strait can enter the SCS and may be The existence of Rossby waves requires initial disturbances another energy source for the 90-day fluctuation. and energy sources. We therefore investigate the driving To explore the driving mechanism of the 90-day fluctuation mechanism for the 90-day fluctuation. The most immediate in the SCS, we conducted three sensitivity experiments. In the driving force is the local wind forcing. Previous studies have first experiment (EXP1), all straits connecting the SCS with found that the wind in the SCS not only motivates obvious surrounding seas/oceans are closed, and the model is forced seasonal variability (Wyrtki 1961; Qiangen et al. 1986) but by daily CCMP wind stress to study the effects of the local also triggers intraseasonal (10–100 days) variability (Kemball- wind forcing. In the second experiment (EXP2), the wind field Cook and Wang 2001; Mao and Chan 2005; Wang et al. 2009). is closed, and the Luzon Strait is opened. The model is driven Driven by the local wind, the circulation in the SCS has dra- by the flow through the Luzon Straits. The third experiment matic seasonal variability, and its adjustment to the local wind (EXP3) is driven by both the local wind and flow through the can be understood in terms of Rossby wave dynamics (Yang Luzon Strait. and Liu 2003; Wu et al. 2008; Xu and Oey 2015; Yang et al. The time series of y and y obtained from EXP1 at D2 and 1 2 2017; Jan et al. 2021). Wavelet analysis following the method D5 are presented in Figs. 10a–d. For a comparison with the of Torrence and Compo (1998) is conducted to investigate observations, the bandpassed time series of the mooring-mea- how the averaged wind stress curl over a 38 3 38 (i.e., sured y averaged above 1000 m and averaged below 1000 m at 158–188N, 1168–1198E) box centered at the mooring section D2 and D5 are overlapped in the picture. There is a small varies in amplitude and frequency from 2012 to 2016. difference in the phase of the 90-day fluctuation between the Figure 9a shows that the variation in the wind stress curl at model and the observations. However, the amplitudes between approximately 90 days is energetic here, which is seasonally them are obviously different, with the model outputs being much dependent and enhanced in boreal winter, lasting for several smaller than that of the observed outputs. This finding indicates months. Similarly, the corresponding averaged SLA time that the energy from the local wind cannot fully supply the series from 2012 to 2016 exhibit the same characteristics, with 90-day fluctuation in the SCS. Therefore, we need to look for an obvious 90-day signal and an enhanced signal in boreal other sources of energy for the 90-day fluctuation. The results of MAY 2022 XU E T A L . 883 FIG. 10. (a) Time series of y (red line) obtained from EXP1 at D2. The bandpassed time series of the mooring measured y averaged above 1000 m (light red line) at D2 is overlapped for comparison. (b) Time series of y (blue line) obtained from EXP1 at D2. The band- passed time series of the mooring measured y averaged below 1000 m (light blue line) at D2 is overlapped for comparison. (c),(d) As in (a) and (b), but for D5. (e)–(h)Asin (a)–(d), but for EXP2. (i)–(l) As in (a)–(d), but for EXP3. the experiment driven by flow through the Luzon Strait (EXP2) material Fig. S10), suggesting that the temporal variations in are shown in Figs. 10e–10h. LikeEXP1, thephase of the90-day the inner region and the western boundary of the SCS may fluctuation in EXP2 is similar to observation, but the amplitude have a common driving mechanism. Moreover, by comparing of the 90-day fluctuation is still smaller than the observed result. the standard deviations of y during the observation period Notably, the phase of the 90-day fluctuationinEXP1 and EXP2 (Fig. 11d), we find the 90-day fluctuation in EXP2 is stronger is approximate, which means that the combination of the two than that in EXP1, which indicates that the disturbance from will produce stronger fluctuation. Figures 10i–10l shows the the Luzon Strait provides more energy for the 90-day fluctua- results of the experiment driven by both the local wind and flow tion in the SCS than the local wind. through the Luzon Strait (EXP3). Different from EXP1 and EXP2, the velocity in EXP3 at D2 and D5 is very close to the 5. Discussion and summary observation in both phase and amplitude. The correlation Based on long-term observations from seven moorings located coefficients of the meridional velocity in the upper (lower) in the middle of the SCS, we investigated the characteristics and layer between simulation and observation are 0.76 (0.57) at D2 and 0.96 (0.85) at D5. Moreover, the 90-day fluctuation of mechanism of the 90-day fluctuation in the deep SCS basin. The deep current in all three experiments show westward propaga- velocity time series of the deep flow in the SCS shows significant tion along the mooring section and have a similar phase speed temporal variability on intraseasonal time scales (Fig. 3), and the dominant time scale of this variability is approximately 90 days with the observation (Figs. 11a–c). Compared to the other two experiments, the results from EXP3 fits with the observations (Fig. 4). Theamplitudeofdeep flow variability is much greater remarkably well, indicating that the 90-day fluctuation in the than the mean value, and the 90-day fluctuation accounts for SCS, especially for the deep layer, is dominated by both local approximately 69% of the total variance in the subinertial cur- wind and flow through the Luzon Strait. The westward 90-day rent. The Hovmo ¨ller diagram shows that the 90-day fluctuation fluctuation at the deep western boundary mentioned in Zhou tends to propagate westward along the mooring section (Fig. 5a) et al. (2017) is also well simulated in EXP3 (see supplemental with a phase speed of ∼4.6 cm s (Fig. 5b). The temporal 884 J OUR N A L O F P HY SI C A L O C E A N OGR A P HY VOLUME 52 FIG. 11. (a) Hovmoller diagram of y (cm s ) derived from EXP1 along mooring section during the observation period; black dashed lines are contours of zero velocity. (b) As in (a), but for EXP2. (c) As in (a), but for EXP3. (d) STD of y along mooring section during the observation period. evolution of the full-depth y at D2 and D5 revealed that there is wavenumber l. There are two distinct groups of Rossby also a strong 90-day fluctuation in the upper layer (Fig. 6a), and waves, long ones carry the energy to the west while short ones its phase is opposite to that in the deep layer (Figs. 6b,c). The 90- send the energy to the east. Accoding to Pedlosky (1987), 2 2 day fluctuation in the upper layer also propagates westward, and when zonal wavenumber k . l 1 1=R , the Rossby waves the corresponding phase speed (∼4.7 cm s along the mooring are short in the x direction and the energy goes eastward. section) is very similar to that of the deep layer. These character- Remarkably, the wavenumber of the 90-day fluctuation 25 21 25 21 istics of the 90-day fluctuation are similar to those of the first K (1.75 3 10 m ) is larger than 1/R (1.33 3 10 m ), mode baroclinic Rossby wave. Further analysis shows that the combined with small meridional wavenumber, implying the 90-day fluctuation propagates westward at the speed of the first 90-day fluctuation may be short Rossby wave, which corre- baroclinic Rossby wave. Moreover, we find that the kinetic sponds to the energy appear to eastward propagation found energy of the first mode accounts for approximately 80% of the by observations and model results (Figs. 5a, 8a, 11). Com- total kinetic energy by decomposing the velocity into the first pared with long Rossby waves, the energy of short Rossby three modes at both D2 and D5. These results suggest that the waves is dissipated more easily, which may be the reason for 90-day fluctuation in the SCS basin should be dominated by the the weakening of the 90-day fluctuation on the east side of the first mode baroclinic Rossby wave. mooring section (D1–D3). This view is supported by the fact The result of satellite altimeter is also consistent with the that the 90-day fluctuation on the western boundary is more mooring observation, and it is found that the 90-day fluctua- energetic than that on the interior of the SCS deep basin tion propagates westward with a north–south band structure (Zhou et al. 2017). Moreover, a spectral peak approximately (Fig. 8). There is no obvious north–south propagation, indi- 60 days appeared at D2 and D3, and the same spectral peak cating that the 90-day fluctuation has a small meridional also appeared at the upper of D2. This may be related to the MAY 2022 XU E T A L . 885 strong 60-day oscillation of the local wind, which requires fur- can be accessed at http://www.remss.com/measurements/ ther observation or numerical model study. ccmp/. The mooring data used in this study are available from The existence of Rossby waves requires driving forces. The the corresponding author upon reasonable request. variation in the local wind stress curl corresponds well with the variation in the local SLA, indicating that there is a certain REFERENCES correlation between the 90-day fluctuation in the ocean and the wind (Fig. 9). Moreover, the SCS is a semiclosed basin Alford, M. H., 2003: Redistribution of energy available for ocean connected to the Pacific Ocean through the Luzon Strait. Pre- mixing by long-range propagation of internal waves. Nature, vious studies have shown that energetic intraseasonal variation 423,159–162, https://doi.org/10.1038/nature01628. signals enter the SCS through the Luzon Strait and could have Atlas,R., R. N. Hoffman,J.Ardizzone,M.Leidner,and J. C. Jusem, 2009: Development of a new cross-calibrated multiplatform an impact on the temporal variations in the SCS (Zhou et al. (CCMP) ocean surface wind product. 13th Conf. on Integrated 2014; Zheng et al. 2014; Xu and Oey 2015; Xie et al. 2016). Observing and Assimilation Systems for Atmosphere, Oceans, Thus, both local wind and disturbances originating from the and Land Surface (IOAS-AOLS), Phoenix, AZ, Amer. Meteor. Luzon Strait can drive the 90-day fluctuation in the SCS. With Soc., 4B.1, https://ams.confex.com/ams/89annual/techprogram/ the help of a two-layer model, we simulated the westward paper_145957.htm. 90-day fluctuation in the deep SCS well. Sensitivity experi- Boyer, T. P., and Coauthors, 2013: World Ocean Database 2013. ments verify that the 90-day fluctuation driven by both local S. Levitus and A. Mishonov, Eds., NOAA Atlas NESDIS 72, wind and the disturbances originating from the Luzon Strait, 209 pp., https://doi.org/10.7289/V5NZ85MT. while the latter play a more important role than the former Chang, Y.-T., W.-L. Hsu, J.-H. Tai, T.-Y. Tang, M.-H. Chang, and S.-Y. Chao, 2010: Cold deep water in the South China (Fig. 11). These results not only improve our understanding of Sea. J. Oceanogr., 66,183–190, https://doi.org/10.1007/s10872- the temporal variability in the deep SCS, but also provide a 010-0016-x. novel perspective for studying the poorly investigated but Chelton, D. B., R. A. Deszoeke, M. G. Schlax, K. E. Naggar, and energetic variability in the abyssal oceans of the world. N. Siwertz, 1998: Geographical variability of the first baroclinic Additionally, the advance of understanding of variability in Rossby radius of deformation. J. Phys. Oceanogr., 28,433–460, the deep SCS could provide clues on the unresolved enhanced https://doi.org/10.1175/1520-0485(1998)028,0433:GVOTFB.2. abyssal mixing, which drive the meridional overturning circula- 0.CO;2. tion and has important influence on the SCS throughflow and Codiga, D. L., and C. C. Eriksen, 1997: Observations of low-fre- Indonesia throughflow. Previous studies have found that the quency circulation and amplified subinertial tidal currents at enhanced abyssal mixing in the SCS is not only affected by the Cobb Seamount. J. Geophys. Res, 102,22993–23 007, https:// doi.org/10.1029/97JC01451. active internal tides (Yang et al. 2016), but also associated with Delorme, B. L., and L. N. Thomas, 2019: Abyssal mixing through the energy cascade catalyzed by the interaction between eddies critical reflection of equatorially trapped waves off smooth and the complex topography (Hu et al. 2020). However, the topography. J. Phys. Oceanogr., 49,519–542, https://doi.org/ influence of intraseasonal variability on enhanced abyssal mixing 10.1175/JPO-D-18-0197.1. in the SCS has not been considered. Recent studies have indi- }}, }}, P. Marchesiello, J. Gula, G. Roullet, and M. J. cated that energetic Rossby waves can enhance abyssal mixing Molemaker, 2021: Enhanced abyssal mixing in the equatorial and then drive diapycnal upwelling at low latitudes when the full Pacific associated with non-traditional effects. J. Phys. Ocean- Coriolis force and so-called nontraditional effects are taken into ogr., 51,1895–1914, https://doi.org/10.1175/JPO-D-20-0238.1. account (Delorme and Thomas 2019; Delorme et al. 2021). The Eriksen, C. C., 1991: Observations of amplified flows atop a large Rossby waves in the deep SCS, with a large amplitude of more seamount. J. Geophys. Res., 96, 15227, https://doi.org/10.1029/ 91JC01176. than 5 cm s and a large spatial scale covering the whole deep Freeland, H., 1994: Ocean circulation at and near Cobb Seamount. basin, could play an important role in the enhanced abyssal mix- Deep Sea Res. Part Oceanogr. Res. Pap., 41,1715–1732, https:// ing. Further research is needed in this direction. doi.org/10.1016/0967-0637(94)90069-8. Gan, J., Z. Liu, and C. R. Hui, 2016: A three-layer alternating Acknowledgments. The authors thank the two anonymous spinning circulation in the South China Sea. J. Phys. Ocean- reviewers for their constructive comments and suggestions, ogr., 46,2309–2315, https://doi.org/10.1175/JPO-D-16-0044.1. especially one of the reviewer’s ideas about the potential Gill, A., 1982: Atmosphere-Ocean Dynamics. International impact of intraseasonal variability on enhanced abyssal mix- Geophysics Series, Vol. 30, Academic Press, 680 pp. Gordon, A. L., B. A. Huber, E. J. Metzger, R. D. Susanto, H. E. ing in the SCS. This study was supported by the National Hurlburt, and T. R. Adi, 2012: South China Sea throughflow Natural Science Foundation of China (42076027, 91858203, impact on the Indonesian throughflow. Geophys. Res. Lett., 91958205), the major project from Sanya Yazhou Bay Science 39, L11602, https://doi.org/10.1029/2012GL052021. and Technology City Administration (SKJC-KJ-2019KY04). Hamilton, P., 1990: Deep currents in the Gulf of Mexico. J. Phys. Oceanogr., 20,1087–1104, https://doi.org/10.1175/1520-0485 Data availability statement. Bathymetric data are down- (1990)020,1087:DCITGO.2.0.CO;2. loaded from http://topex.ucsd.edu/marine_topo/. Sea level }}, A. Bower,H.Furey,R.Leben,and P. Pere ´ z-Brunius, 2019: anomalies are from CMEMS (http://marine.copernicus.eu/). The loop current: Observations of deep eddies and topographic The CTD profiles are from WOD13 (https://www.ncei.noaa. waves. J. Phys. Oceanogr., 49, 1463–1483, https://doi.org/10. gov/products/world-ocean-database). The CCMP wind data 1175/JPO-D-18-0213.1. 886 J OUR N A L O F P HY SI C A L O C E A N OGR A P HY VOLUME 52 Hogg, N., 1981: Topographic waves along 708W on the continental Quadfasel, D. R., and J. C. Swallow, 1986: Evidence for 50-day rise. J. Mar. Res., 39,627–649. period planetary waves in the South Equatorial Current of Hu, Q., and Coauthors, 2020: Cascade of internal wave energy the Indian Ocean. Deep Sea Res. Part Oceanogr. Res. Pap., catalyzed by eddy-topography interactions in the Deep South 33, 1307–1312, https://doi.org/10.1016/0198-0149(86)90037-3. China Sea. Geophys.Res.Lett., 47, e2019GL086510, https:// Quan, Q., Z. Cai, G. Jin, and Z. Liu, 2021: Topographic Rossby doi.org/10.1029/2019GL086510. waves in the abyssal South China Sea. J. Phys. Oceanogr., 51, Jan, S., M.-H. Chang, Y. J. Yang, C.-H. Sui, Y.-H. Cheng, Y.-Y. 1795–1812, https://doi.org/10.1175/JPO-D-20-0187.1. Yeh, and C.-W. Lee, 2021: Mooring observed intraseasonal Scha ¨r, C., and H. C. Davies, 1988: Quasi-geostrophic stratified flow oscillations in the central South China Sea during summer over isolated finite amplitude topography. Dyn. Atmos. Oceans, monsoon season. Sci. Rep., 11, 13685, https://doi.org/10.1038/ 11, 287–306, https://doi.org/10.1016/0377-0265(88)90003-6. s41598-021-93219-3. Shu, Y., H. Xue, D. Wang, F. Chai, Q. Xie, J. Yao, and J. Xiao, Kemball-Cook, S., and B. Wang, 2001: Equatorial waves and 2014: Meridional overturning circulation in the South China air–sea interaction in the boreal summer intraseasonal oscil- Sea envisioned from the high-resolution global reanalysis lation. J. Climate, 14,2923–2942, https://doi.org/10.1175/1520- data GLBa0.08. J. Geophys. Res. Oceans, 119, 3012–3028, 0442(2001)014,2923:EWAASI.2.0.CO;2. https://doi.org/10.1002/2013JC009583. Lan, J., N. Zhang, and Y. Wang, 2013: On the dynamics of the }}, and Coauthors, 2016: Persistent and energetic bottom- South China Sea deep circulation. J. Geophys. Res. Oceans, trapped topographic Rossby waves observed in the southern 118, 1206–1210, https://doi.org/10.1002/jgrc.20104. South China Sea. Sci. Rep., 6, 24338, https://doi.org/10.1038/ Lavelle, J. W., E. T. Baker, and G. A. Cannon, 2003: Ocean cur- srep24338. rents at Axial Volcano, a northeastern Pacific seamount. J. Smith, W., and D. T. S. Sandwell, 1997: Global sea floor topography Geophys. Res., 108, 3020, https://doi.org/10.1029/2002JC001305. from satellite altimetry and ship depth soundings. Science, 277, Liang, X., and A. M. Thurnherr, 2012: Eddy-modulated internal 1956–1962, https://doi.org/10.1126/science.277.5334.1956. waves and mixing on a midocean ridge. J. Phys. Oceanogr., Sun, Z., Z. Zhang, B. Qiu, X. Zhang, C. Zhou, X. Huang, W. Zhao, 42, 1242–1248, https://doi.org/10.1175/JPO-D-11-0126.1. and J. Tian, 2020: Three-dimensional structure and interannual Liu C.-T., and R.-J. Liu, 1988: The deep current in the Bashi variability of the Kuroshio Loop Current in the northeastern Channel. Acta Oceanogr. Taiwan, 20, 107–116. South China Sea. J. Phys. Oceanogr., 50, 2437–2455, https://doi. Mao, J., and J. C. L. Chan, 2005: Intraseasonal variability of the org/10.1175/JPO-D-20-0058.1. South China Sea summer monsoon. J. Climate, 18, 2388–2402, Thompson, R., and J. R. Luyten, 1976: Evidence for bottom- https://doi.org/10.1175/JCLI3395.1. trapped topographic Rossby waves from single moorings. Nash, J. D., M. H. Alford, and E. Kunze, 2005: Estimating inter- Deep-Sea Res. Oceanogr. Abstr., 23,629–635, https://doi.org/ nal wave energy fluxes in the ocean. J. Atmos. Oceanic Tech- 10.1016/0011-7471(76)90005-X. nol., 22,1551–1570, https://doi.org/10.1175/JTECH1784.1. Tian,J., Q. Yang,X. Liang,L.Xie, D.Hu, F. Wang,and T. Qu, Nikurashin, M., and R. Ferrari, 2010: Radiation and dissipation of 2006: Observation of Luzon Strait transport. Geophys. Res. internal waves generated by geostrophic motions impinging Lett., 33, L19607, https://doi.org/10.1029/2006GL026272. on small-scale topography: Theory. J. Phys. Oceanogr., 40, }}, }}, and W. Zhao, 2009: Enhanced diapycnal mixing in the 1055–1074, https://doi.org/10.1175/2009JPO4199.1. South China Sea. J. Phys. Oceanogr., 39, 3191–3203, https:// Pedlosky, J., 1987: Geophysical Fluid Dynamics. Springer, 710 pp., doi.org/10.1175/2009JPO3899.1. https://doi.org/10.1007/978-1-4612-4650-3. Torrence, C., and G. P. Compo, 1998: A practical guide to wavelet Pena-Molino, B., T. M. Joyce, and J. M. Toole, 2012: Variability analysis. Bull. Amer. Meteor. Soc., 79,61–78, https://doi.org/ in the Deep Western Boundary Current: Local versus remote 10.1175/1520-0477(1998)079,0061:APGTWA.2.0.CO;2. forcing. J. Geophys. Res. Oceans, 117, 574–586, https://doi. Uppala, S. M., 2005: The ERA40 reanalysis. Quart. J. Roy. org/10.1029/2012JC008369. Meteor. Soc., 131, 2961–3012, https://doi.org/10.1256/qj.04.176. Pickart, R. S., and D. R. Watts, 1990: Deep Western Boundary Cur- Wang, B., H. Fei, Z. Wu, J. Yang, X. Fu, and K. Kikuchi, 2009: rent variability at Cape Hatteras. J. Mar. Res., 48, 765–791, Multi-scale climate variability of the South China Sea https://doi.org/10.1357/002224090784988674. monsoon: A review. Dyn. Atmos. Oceans, 47,15–37, Qiangen, Z., J. He, and P. Wang, 1986: A study of circulation dif- https://doi.org/10.1016/j.dynatmoce.2008.09.004. ferences between East-Asian and Indian summer monsoons Wang, G., S.-P. Xie, T. Qu, and R. X. Huang, 2011: Deep South with their interaction. Adv. Atmos. Sci., 3,466–477, https:// China Sea circulation. Geophys. Res. Lett., 38, L05601, doi.org/10.1007/BF02657936. https://doi.org/10.1029/2010GL046626. Qu, T., Y. Du, G. Meyers, A. Ishida, and D. Wang, 2005: Con- Wang, Q., and Coauthors, 2019: Energetic topographic Rossby necting the tropical Pacific with Indian Ocean through South waves in the northern South China Sea. J. Phys. Oceanogr., China Sea. Geophys. Res. Lett., 32, L24609, https://doi.org/10. 49, 2697–2714, https://doi.org/10.1175/JPO-D-18-0247.1. 1029/2005GL024698. Warren, B. A., T. Whitworth, and J. H. LaCasce, 2002: }}, }}, }}, and H. Sasaki, 2006a: South China Sea through- Forced resonant undulation in the deep Mascarene Basin. flow: A heat and freshwater conveyor. Geophys. Res. Lett., Deep-Sea Res. II, 49, 1513–1526, https://doi.org/10.1016/ 33, L23617, https://doi.org/10.1029/2006GL028350. }}, J. B. Girton, and J. A. Whitehead, 2006b: Deepwater over- S0967-0645(01)00151-5. Whitworth, T., III, B. A. Warren, W. D. Nowlin Jr., S. B. Rutz, flow through Luzon Strait. J. Geophys. Res., 111, C01002, https://doi.org/10.1029/2005JC003139. R. D. Pillsbury, and M. I. Moore, 1999: On the deep western- }}, Y. T. Song, and T. Yamagata, 2009: An introduction to the boundary current in the southwest Pacificbasin. Prog. Ocean- South China Sea throughflow: Its dynamics, variability, and ogr., 43,1–54, https://doi.org/10.1016/S0079-6611(99)00005-1. application for climate. Dyn. Atmos. Oceans, 47,3–14, https:// Wu, X., Q. Xie, Z. He, and D. Wang, 2008: Free and forced doi.org/10.1016/j.dynatmoce.2008.05.001. Rossby waves in the western South China Sea inferred from MAY 2022 XU E T A L . 887 Jason-1 satellite altimetry data. Sensors, 8, 3633–3642, https:// circulation. J. Geophys. Res. Oceans, 118, 6479–6494, https:// doi.org/10.3390/s8063633. doi.org/10.1002/2013JC008994. Wyrtki, K., 1961: Physical oceanography of the Southeast Asian Zhao, W., C. Zhou, J. Tian, Q. Yang, B. Wang, L. Xie, and T. waters. NAGA Rep. 2, 195 pp., https://escholarship.org/uc/ Qu, 2014: Deep water circulation in the Luzon Strait. J. item/49n9x3t4. Geophys. Res. Oceans, 119,790–804, https://doi.org/10. Xie, L., Q. Zheng, J. Tian, S. Zhang, Y. Feng, and X. Yi, 2016: 1002/2013JC009587. Cruise observation of Rossby waves with finite wavelengths Zhao, X., C. Zhou, W. Zhao, J. Tian, and X. Xu, 2016: Deepwater propagating from the Pacific to the South China Sea. J. Phys. overflow observed by three bottom-anchored moorings in the Oceanogr., 46, 2897–2913, https://doi.org/10.1175/JPO-D-16- Bashi Channel. Deep Sea Res. I, 110,65–74, https://doi.org/10. 0071.1. 1016/j.dsr.2016.01.007. Xu, F.-H., and L.-Y. Oey, 2015: Seasonal SSH variability of the Zheng, H., X.-H. Zhu, C. Zhang, R. Zhao, Z.-N. Zhu, and Z.-J. northern South China Sea. J. Phys. Oceanogr., 45,1595–1609, Liu, 2021: Propagation of topographic Rossby waves in the https://doi.org/10.1175/JPO-D-14-0193.1. deep basin of the South China Sea based on abyssal current Yang, H., and Q. Liu, 2003: Forced Rossby wave in the northern observations. J. Phys. Oceanogr., 51, 2783–2791, https://doi. South China Sea. Deep Sea Res. I., 50,917–926, https://doi. org/10.1175/JPO-D-21-0051.1. org/10.1016/S0967-0637(03)00074-8. Zheng, Q., and Coauthors, 2014: Standing wave modes observed }}, L. Wu, S. Sun, and Z. Chen, 2017: Selective response of the in the South China Sea deep basin. J. Geophys. Res. Oceans, South China Sea circulation to summer monsoon. J. Phys. Oce- 119,4185–4199, https://doi.org/10.1002/2014JC009957. anogr., 47, 1555–1568, https://doi.org/10.1175/JPO-D-16-0288.1. Zhou, C., W. Zhao, J. Tian, Q. Yang, and T. Qu, 2014: Variability Yang, Q., W. Zhao, X. Liang, and J. Tian, 2016: Three-dimen- of the deep-water overflow in the Luzon Strait. J. Phys. Ocean- sional distribution of turbulent mixing in the South China ogr., 44,2972–2986, https://doi.org/10.1175/JPO-D-14-0113.1. Sea. J. Phys. Oceanogr., 46, 769–788, https://doi.org/10.1175/ }}, }}, }}, X. Zhao, Y. Zhu, Q. Yang, and T. Qu, 2017: JPO-D-14-0220.1. Deep Western Boundary Current in the South China Sea. Ye, R., C. Zhou, W. Zhao, J. Tian, Q. Yang, X. Huang, Z. Zhang, Sci. Rep., 7, 9303, https://doi.org/10.1038/s41598-017-09436-2. and X. Zhao, 2019: Variability in the deep overflow through Zhou, M., G. Wang, W. Liu, and C. Chen, 2020: Variability of the the Heng-Chun ridge of the Luzon Strait. J. Phys. Oceanogr., observed deep western boundary current in the South China 49, 811–825, https://doi.org/10.1175/JPO-D-18-0113.1. Zhang, Z., W. Zhao, J. Tian, and X. Liang, 2013: A mesoscale Sea. J. Phys. Oceanogr., 50, 2953–2963, https://doi.org/10. eddy pair southwest of Taiwan and its influence on deep 1175/JPO-D-20-0013.1.
Journal of Physical Oceanography – American Meteorological Society
Published: May 22, 2022
You can share this free article with as many people as you like with the url below! We hope you enjoy this feature!
Read and print from thousands of top scholarly journals.
Already have an account? Log in
Bookmark this article. You can see your Bookmarks on your DeepDyve Library.
To save an article, log in first, or sign up for a DeepDyve account if you don’t already have one.
Copy and paste the desired citation format or use the link below to download a file formatted for EndNote
Access the full text.
Sign up today, get DeepDyve free for 14 days.
All DeepDyve websites use cookies to improve your online experience. They were placed on your computer when you launched this website. You can change your cookie settings through your browser.