Observation of Bottom-Trapped Topographic Rossby Waves to the West of the Luzon Strait, South China Sea

Hua Zheng aSchool of Oceanography, Shanghai Jiao Tong University, Shanghai, China
bState Key Laboratory of Satellite Ocean Environment Dynamics, Second Institute of Oceanography, Ministry of Natural Resources, Hangzhou, China

Search for other papers by Hua Zheng in
Current site
Google Scholar
PubMed
Close
,
Xiao-Hua Zhu aSchool of Oceanography, Shanghai Jiao Tong University, Shanghai, China
bState Key Laboratory of Satellite Ocean Environment Dynamics, Second Institute of Oceanography, Ministry of Natural Resources, Hangzhou, China
cSouthern Marine Science and Engineering Guangdong Laboratory (Zhuhai), Zhuhai, China

Search for other papers by Xiao-Hua Zhu in
Current site
Google Scholar
PubMed
Close
,
Juntian Chen bState Key Laboratory of Satellite Ocean Environment Dynamics, Second Institute of Oceanography, Ministry of Natural Resources, Hangzhou, China

Search for other papers by Juntian Chen in
Current site
Google Scholar
PubMed
Close
,
Min Wang aSchool of Oceanography, Shanghai Jiao Tong University, Shanghai, China
bState Key Laboratory of Satellite Ocean Environment Dynamics, Second Institute of Oceanography, Ministry of Natural Resources, Hangzhou, China

Search for other papers by Min Wang in
Current site
Google Scholar
PubMed
Close
,
Ruixiang Zhao bState Key Laboratory of Satellite Ocean Environment Dynamics, Second Institute of Oceanography, Ministry of Natural Resources, Hangzhou, China

Search for other papers by Ruixiang Zhao in
Current site
Google Scholar
PubMed
Close
,
Chuanzheng Zhang bState Key Laboratory of Satellite Ocean Environment Dynamics, Second Institute of Oceanography, Ministry of Natural Resources, Hangzhou, China

Search for other papers by Chuanzheng Zhang in
Current site
Google Scholar
PubMed
Close
,
Ze-Nan Zhu bState Key Laboratory of Satellite Ocean Environment Dynamics, Second Institute of Oceanography, Ministry of Natural Resources, Hangzhou, China

Search for other papers by Ze-Nan Zhu in
Current site
Google Scholar
PubMed
Close
,
Qiang Ren dKey Laboratory of Ocean Circulation and Waves, Institute of Oceanology, Chinese Academy of Sciences, Qingdao, China
eCenter for Ocean Mega-Science, Chinese Academy of Sciences, Qingdao, China

Search for other papers by Qiang Ren in
Current site
Google Scholar
PubMed
Close
,
Yansong Liu dKey Laboratory of Ocean Circulation and Waves, Institute of Oceanology, Chinese Academy of Sciences, Qingdao, China
eCenter for Ocean Mega-Science, Chinese Academy of Sciences, Qingdao, China
fMarine Dynamic Process and Climate Function Laboratory, Pilot National Laboratory for Marine Science and Technology (Qingdao), Qingdao, China

Search for other papers by Yansong Liu in
Current site
Google Scholar
PubMed
Close
,
Feng Nan dKey Laboratory of Ocean Circulation and Waves, Institute of Oceanology, Chinese Academy of Sciences, Qingdao, China
eCenter for Ocean Mega-Science, Chinese Academy of Sciences, Qingdao, China
fMarine Dynamic Process and Climate Function Laboratory, Pilot National Laboratory for Marine Science and Technology (Qingdao), Qingdao, China

Search for other papers by Feng Nan in
Current site
Google Scholar
PubMed
Close
, and
Fei Yu dKey Laboratory of Ocean Circulation and Waves, Institute of Oceanology, Chinese Academy of Sciences, Qingdao, China
eCenter for Ocean Mega-Science, Chinese Academy of Sciences, Qingdao, China
fMarine Dynamic Process and Climate Function Laboratory, Pilot National Laboratory for Marine Science and Technology (Qingdao), Qingdao, China
gUniversity of Chinese Academy of Sciences, Beijing, China

Search for other papers by Fei Yu in
Current site
Google Scholar
PubMed
Close
Free access

Abstract

Topographic Rossby waves (TRWs) play an important role in deep-ocean dynamics and abyssal intraseasonal variations. Observational records from 15 current- and pressure-recording inverted echo sounders (CPIESs) and two moorings deployed in the northern Manila Trench (MT), South China Sea (SCS), for over 400 days were utilized to analyze the widely existing near-21-day bottom-trapped TRWs in the trench. The TRWs were generally generated in winter and summer, dominated by perturbations in the upper ocean. Kuroshio intrusion and its related variabilities contributed to the perturbations in winter, whereas the perturbations generated north of Luzon Island dominated in summer. Eddies north of Luzon propagated northwestward in the summer of 2018; however, these eddies caused the Kuroshio meanderings in the Luzon Strait (LS) in the summer of 2019. The variations in the Kuroshio path and the Kuroshio-related eddies induced TRWs in the deep ocean in regions with steep topography. However, the spatiotemporal distributions of TRWs were complex owing to the propagation of the waves. Some cases of TRWs showed no relation to the local upper-layer perturbations but propagated from adjacent regions. Some of these TRWs were induced by perturbations in the upper ocean in adjacent regions, and propagated anticlockwise in the MT with shallow water to their right, while others may be related to the intraseasonal variations in deep-water overflow in the LS and propagated northward. This study suggests that the Kuroshio and Kuroshio-related eddies significantly contribute to the dynamic processes associated with intraseasonal variations in the deep SCS through the generation of TRWs.

Significance Statement

Topographic Rossby waves (TRWs) are fluctuations generated when water columns travel across sloping topography under potential vorticity conservation. Based on observations of 15 current- and pressure-recording inverted echo sounders (CPIESs) and two moorings in the northern Manila Trench (MT) in the South China Sea (SCS), TRWs with periods of approximately 21 days were observed and analyzed. This study describes the generation, propagation, and spatiotemporal distribution of TRWs west of the LS and confirms that regional Kuroshio meanderings and upper eddies play important roles in the dynamic processes associated with intraseasonal variations in the deep SCS; the study may thus contribute to knowledge on the dynamic response of the abyssal current to mesoscale perturbations in the upper ocean.

© 2022 American Meteorological Society. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).

Production note: The School of Oceanography, Shanghai Jiao Tong University, State Key Laboratory of Satellite Ocean Environment Dynamics, Second Institute of Oceanography, Ministry of Natural Resources, and the Southern Marine Science and Engineering Guangdong Laboratory (Zhuhai) contributed equally to the work and should be regarded as co–first institutional affiliations.

Corresponding author: Xiao-Hua Zhu, xhzhu@sio.org.cn

Abstract

Topographic Rossby waves (TRWs) play an important role in deep-ocean dynamics and abyssal intraseasonal variations. Observational records from 15 current- and pressure-recording inverted echo sounders (CPIESs) and two moorings deployed in the northern Manila Trench (MT), South China Sea (SCS), for over 400 days were utilized to analyze the widely existing near-21-day bottom-trapped TRWs in the trench. The TRWs were generally generated in winter and summer, dominated by perturbations in the upper ocean. Kuroshio intrusion and its related variabilities contributed to the perturbations in winter, whereas the perturbations generated north of Luzon Island dominated in summer. Eddies north of Luzon propagated northwestward in the summer of 2018; however, these eddies caused the Kuroshio meanderings in the Luzon Strait (LS) in the summer of 2019. The variations in the Kuroshio path and the Kuroshio-related eddies induced TRWs in the deep ocean in regions with steep topography. However, the spatiotemporal distributions of TRWs were complex owing to the propagation of the waves. Some cases of TRWs showed no relation to the local upper-layer perturbations but propagated from adjacent regions. Some of these TRWs were induced by perturbations in the upper ocean in adjacent regions, and propagated anticlockwise in the MT with shallow water to their right, while others may be related to the intraseasonal variations in deep-water overflow in the LS and propagated northward. This study suggests that the Kuroshio and Kuroshio-related eddies significantly contribute to the dynamic processes associated with intraseasonal variations in the deep SCS through the generation of TRWs.

Significance Statement

Topographic Rossby waves (TRWs) are fluctuations generated when water columns travel across sloping topography under potential vorticity conservation. Based on observations of 15 current- and pressure-recording inverted echo sounders (CPIESs) and two moorings in the northern Manila Trench (MT) in the South China Sea (SCS), TRWs with periods of approximately 21 days were observed and analyzed. This study describes the generation, propagation, and spatiotemporal distribution of TRWs west of the LS and confirms that regional Kuroshio meanderings and upper eddies play important roles in the dynamic processes associated with intraseasonal variations in the deep SCS; the study may thus contribute to knowledge on the dynamic response of the abyssal current to mesoscale perturbations in the upper ocean.

© 2022 American Meteorological Society. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).

Production note: The School of Oceanography, Shanghai Jiao Tong University, State Key Laboratory of Satellite Ocean Environment Dynamics, Second Institute of Oceanography, Ministry of Natural Resources, and the Southern Marine Science and Engineering Guangdong Laboratory (Zhuhai) contributed equally to the work and should be regarded as co–first institutional affiliations.

Corresponding author: Xiao-Hua Zhu, xhzhu@sio.org.cn

1. Introduction

The South China Sea (SCS) is a semienclosed basin that connects the Pacific and Indian Oceans. The Luzon Strait (LS), whose largest sill depth is approximately 2400 m, is the only channel that connects the deep SCS and its surrounding oceans. The SCS has a unique basin-scale three-layer circulation system of cyclonic, anticyclonic, and cyclonic structures in the upper, middle, and deep ocean, respectively, which is associated with the “sandwich” structure of the water transport between the SCS and the Pacific Ocean through the LS (Yuan 2002; Gan et al. 2016; D. Wang et al. 2019; Zhu et al. 2019).

The abundant multiscale dynamic processes in the SCS establish a bond between the abyssal circulation and the upper and middle circulation. On the one hand, perturbations generated in the deep ocean related to the upper ocean or propagated from the LS significantly contribute to large abyssal temporal variances (e.g., Lan et al. 2015; Zhang et al. 2016; Xiao et al. 2016; Zhou et al. 2020; Wang et al. 2021). On the other hand, internal waves and deep eddies result in enhanced diapycnal mixing (∼10−3 m2 s−1) in the deep SCS, which sustain deep-water overflow through the LS and further intensify the abyssal circulation and overturning circulation (e.g., Tian et al. 2009; Yang et al. 2014; Quan and Xue 2019; Wang et al. 2017). In addition, deep water in the SCS rises near the continental shelf and islands with rough topographies and further contributes to the upper circulation (Shu et al. 2014; Wang et al. 2012).

The temporal variances of the deep currents were greater than their temporal averages (Zhou et al. 2017). Internal tides generated near the LS can propagate into the SCS and strongly influence diapycnal mixing (Alford et al. 2015). The abyssal circulation intensifies in summer but weakens in winter because of the variability in the deep-water overflow in the LS (Lan et al. 2015; Gan et al. 2016; Zhou et al. 2017; Zheng et al. 2021a); the overflow also potentially contributes to the intraseasonal variability of the deep circulation through transport variances or related deep eddies (Zhou et al. 2020). Deep eddies can also be caused by interaction between the cyclonic–anticyclonic eddy pair and the seamount topography (Shu et al. 2022). In addition, baroclinic Rossby waves also contribute to abyssal variances (Xu et al. 2022).

Mesoscale perturbations in the upper ocean have a significant impact on deep circulation. On the one hand, near-inertial waves induced by the monsoon and typhoons were observed in the deep ocean (Xiao et al. 2016); on the other hand, some mesoscale eddies directly reached the bottom of the ocean (Zhang et al. 2016; Shu et al. 2018). In addition, upper perturbations can induce topographic Rossby waves (TRWs) on sloping topographies through potential vorticity (PV) adjustment (Shu et al. 2016; Q. Wang et al. 2019; Quan et al. 2021a,b; Zheng et al. 2021b; Wang et al. 2021; Shu et al. 2022).

As a dominant oscillating pattern of intraseasonal variability in the deep SCS, the TRWs with periods ranging from approximately 10 days to dozens of days have been observed in recent years. TRWs were first observed in the SCS near the Nansha Islands by Shu et al. (2016); persistent energetic TRWs with periods of 9–14 days—whose amplitudes were approximately one order of magnitude higher than the average current—were induced by variabilities in the upper-layer current. TRWs observed in the northeast SCS—with periods of approximately 14.5 days—contributed to more than 40% of the bottom current fluctuations (Q. Wang et al. 2019); upper mesoscale fluctuations of 10–20 days were considered as their possible energy sources. The propagation of TRWs in the SCS was first captured by 11 current- and pressure-recording inverted echo sounders (CPIESs); large-scale near-65-day TRWs were observed in the deep basin, with the fluctuations showing a westward phase speed among the array, and the energy being related to the surface perturbations southwest of Taiwan Island (Zheng et al. 2021b). Based on five moorings deployed in the northern SCS, Wang et al. (2021) presented the spatial distribution of TRWs in terms of frequency and intensity. TRWs with longer periods (30–80 days) were observed west of the Dongsha Islands, whereas TRWs near the Dongsha Islands and in the LS indicated shorter periods (8–25 days). The generation of these TRWs was mainly related to perturbations in the upper layer. Modeling studies have also indicated the wide existence of TRWs in the deep SCS. Quan et al. (2021a) indicated that 5–60-day TRWs can account for over 40% of the deep kinetic energy variability, with the possibility of reaching 70% over the slopes. Upper mesoscale perturbations and deep large-scale circulation induce TRWs through pressure work and baroclinic instability, respectively, which further modulate the abyssal circulation (Quan et al. 2021b).

Regions west of the LS are characterized by complex topography, three-layer SCS–Pacific water exchange, and multiple upper perturbations, providing an environment conducive to the generation of TRWs. Previous studies have indicated that TRWs play an important role in the energy pathway from the Loop Current to the Gulf of Mexico basin (e.g., Hamilton 2009; Hamilton et al. 2019). Hence, it is reasonable to speculate that the strong western boundary current, the Kuroshio, can be an important energy source for TRWs in the deep SCS. Although previous observations have revealed that perturbations in the upper layer can induce TRWs in the deep SCS, knowledge about TRWs associated with the Kuroshio variabilities (Kuroshio meandering, Kuroshio intrusion, and Kuroshio-related eddies) near the LS is limited to modeling studies due to the paucity of in situ observations. Apart from the Kuroshio variabilities, the LS region has an abundance of mesoscale eddies locally generated by wind or propagated from the North Pacific, which may also contribute to the generation of TRWs. Knowledge of the energy pathway between the abundant upper-layer mesoscale perturbations and the deep ocean in this region is limited, and how TRWs are generated and propagate west of the LS remains unclear.

An array including 28 CPIESs and two moorings was deployed west of the LS and SCS, providing an opportunity to reveal the connection between the upper-layer perturbations and the TRWs. The remaining part of this paper is organized as follows. The observation data and methods are described in section 2, the observations of the near-21-day TRWs are described in section 3, the generation and propagation of TRWs are discussed in section 4, and section 5 summarizes the study.

2. Data and methods

a. Data

Two moorings and 28 CPIESs were deployed west of the LS for over 400 days between June 2018 and July 2019 to clarify the spatiotemporal evolution of the current system near the LS in the SCS. Of these, 27 CPIESs were successfully recovered, excluding C19. Only 15 CPIESs and two moorings deployed in the northern Manila Trench (MT) were used in this study of the near-21-day TRWs (Fig. 1) as oscillations were not observed at sites farther west (not shown).

Fig. 1.
Fig. 1.

Map of the study region in the South China Sea. Current- and pressure-recording inverted echo sounder (CPIES) sites and mooring sites are indicated by black dots and black triangles, respectively. Numbers below site names indicate depths (m).

Citation: Journal of Physical Oceanography 52, 11; 10.1175/JPO-D-22-0065.1

1) CPIES observations

The CPIES is a pressure-recording inverted echo sounder (PIES) attached to the seafloor and equipped with an Aanderaa Doppler Current Sensor positioned approximately 50 m above the PIES; the round-trip acoustic travel time (τ), bottom pressure (Pbot), and near-bottom zonal and meridional currents (u, υ) were measured.

Two columns of CPIESs that included 16 instruments (C13–C28) were deployed west of the LS at 119.8° and 120.5°E (Fig. 1). C21 and C23 were deployed in December 2017, while the others were deployed in June 2018. Most of the instruments were recovered in June 2019, except for C19. Fundamental properties including τ, u, υ, and Pbot were used in this study. τ was recorded 24 times each hour—the standard deviation (STD) of a 24-ping sample was typically less than 2.2 m s. The absolute accuracy of the current sensor is ±0.15 cm s−1, and the observations were carried out at hourly intervals. High-quality data were recorded at most sites; however, the current sensor of C24 stopped working in February 2019. Pbot records with a sampling interval of 10 min, a resolution of 0.001 dbar, an absolute accuracy of ±0.01%, and a full-scale range of 6000 dbar were documented by pressure sensors. Following Kennelly et al. (2007), all records measured by CPIESs were preprocessed (e.g., despiked, detided, and dedrifted), resampled hourly, and converted to coordinated universal time. Finally, all records were 72-h low-pass filtered and subsampled at 12-h intervals.

2) Mooring observations

Two moorings were deployed west of the LS with approximate water depths of 3132 and 4271 m (Fig. 1); the moorings used in the study were equipped with instruments whose details are listed in Table 1. M01 was deployed in December 2017, and M02 was deployed in June 2018; both were recovered in July 2019. The two moorings were equipped with RDI Workhorse 75-kHz upward- and downward-looking acoustic Doppler current profilers (ADCPs) at approximately 400 m from the surface. Both upward- and downward-looking ADCPs observed 74 bins with a vertical resolution of 8 m, covering the depth from ∼60 to ∼1000 m. Moreover, three current meters (Aquadopp from Nortek) were deployed at 983, 1999, and 2912 m at M01, and two at M02 at 2148 and 4151 m from the surface. The deepest current meters at M01 and M02 were positioned approximately 220 and 120 m above the ocean bottom, respectively. The sampling interval of the ADCPs and current meters was 1 h. All data were detided and 72-h low-pass filtered. Finally, the records from the ADCPs and current meters were subsampled at one-day and half-day intervals, respectively.

Table 1

Mooring information.

Table 1

3) Other data

For the calibration of τ, conductivity–temperature–depth (CTD) profiles obtained near each CPIES during the observation period were used. To convert τ to τ1200 and establish an empirical relationship between τ1200 and the sea surface height (SSH) anomaly, 635 CTD profiles and 961 Argo profiles in the study region were used. In addition, historical profiles with depths greater than 2500 dbar were used to obtain the Brunt–Väisälä frequency (N) in the deep ocean.

Furthermore, absolute dynamic topography (ADT) and surface geostrophic currents from the Copernicus Marine Environment Monitoring Service (CMEMS) were used to analyze the generation of near-21-day TRWs. The topography was obtained from the Earth topography 1-min grid (ETOPO1) developed by the National Centers for Environmental Information.

b. Methods

1) Dispersion relation of TRWs

The TRWs are waves with periods ranging from several to hundreds of days, induced by the conservation of PV when water columns travel across a sloping topography. The characteristics of TRWs can be described by the dispersion relation (Pickart 1995):
λυ2=(k2+l2+βkω)(Nf)2,
λυtanh(λυH)=N2ωf(kHylHx),
where 1/λυ is the vertical trapping scale, (k, l) are the zonal and meridional wavenumbers, ω is the frequency, f is the Coriolis parameter, β is the planetary beta effect, H is the water depth, and (Hx, Hy) are the zonal and meridional gradients of depth.
To filter out small-scale features of topography, the ETOPO1 data were smoothed at 1/3° × 1/6°. A series of B-splines were further fitted to the bottom depth to obtain the topography gradients ∇H = (Hx, Hy) following Meinen et al. (1993). The topography beta effect ( βTopo=f|H|/H ) was much more significant than the planetary beta effect (β) at the mooring and CPIES sites that we focused on. The ratio βTopo/β shows the lowest values at C22 (6.2), C23 (7.9), and C26 (9.7), and is greater than 10 at other sites, indicating that β can be ignored at CPIES sites and the dispersion relation can be simplified as follows:
ω=N|H|sin(θ)coth(|K|NHf),
where K = (k, l), and θ is the clockwise angle K made with the ∇H. Under the assumption of a 30-km wavelength at periods less than 30 days used in previous studies (Q. Wang et al. 2019; Wang et al. 2021), and combining with the typical N value of approximately 1 × 10−3 s−1 for the focused depth, coth(|K|NH/f)1, the equation becomes
ω=N|H|sin(θ).

The direction of the group velocity cg=(ω/k,ω/l) is perpendicular to the direction of the wavenumber vector. When θ is within (0, π/2), the TRWs exhibit downslope wavenumber vector and upslope group velocity, while they propagate with upslope wavenumber vector and downslope group velocity when θ is within (π/2, π). The simplified dispersion relation [Eq. (4)] was further used to calculate the directions of the wavenumber vector and group velocity at the observation sites in section 3.

2) Ray-tracing model

A ray-tracing model is used to trace the rays of the TRWs and to identify their characteristics during propagation. To accommodate the condition when rays reach regions where βTopo is comparable to β, the original dispersion relation [Eqs. (1) and (2)] was used in the model. The variances of the wave-group locations and wavenumbers over time are expressed as follows (Oey and Lee 2002):
dxidt=cgi+ui,
dkidt=m=1nωγmγmxikjujxi,
where i, j represent the two horizontal dimensions of parameters, d/dt=(/t)+(cg+U) is a derivation moving with the wave group, ∇ is the Hamiltonian operator, U = (u1, u2) is horizontal background current, K = (k1, k2) is horizontal wavenumber vector, and X = (x1, x2) represents the location of the wave group. Environmental parameters including H, ∇H, and N are represented by γm. The temporal interval of operation of the model was 1 h.

3) Estimation of sea surface height variability from CPIES

SSH variability was estimated using τ and Pbot records, following Baker-Yeboah et al. (2009) and Park et al. (2012). The contribution from mass loading is given by
ηbt=Pbotρbg,
where ρb is the typical deep-seawater density from historical hydrocasts and g = 9.8 m s−1 is the gravitational acceleration. The contribution of the upper baroclinic height, named steric height, is given by
ηbc=ϕg=1gPbot01ρdp,
where ϕ is the dynamic height and ρ is the density. Here, a reference level pref = 1200 dbar was chosen instead of the Pbot because the baroclinic variance below 1200 dbar is less than 5% of the total in this region and historical hydrocasts deeper than 1200 dbar were limited.
To monitor SSH variability, τ was calibrated and converted to τ1200 which indicates the round-trip travel time from a reference level (1200 dbar here) to the surface, following Donohue et al. (2010). During the conversion, τ was calibrated by CTD profiles taken near CPIES sites during the observations, the contribution of mass loading was removed, the latitudinal-independent dynamic τ was obtained and converted to τ1200 based on historical profiles, and the seasonal variations in τ1200 were removed. Detailed information on the calibration and conversion processes is provided by Zheng et al. (2022). A seasonal model for ϕ/g was considered and an empirical relationship between ϕ/g and deseasonal τ1200 was established based on historical hydrocasts (Figs. 2a,b). The steric height variability is given by
ηbc=ϕg+ϕseasonalg,
where the prime indicates mean-removed fluctuations. Finally, the total SSH variability is calculated as η=ηbt+ηbc. The root-mean-square (RMS) error between the SSH anomaly from CPIES and ADT varied from 4.8 to 9.7 cm among sites, with an average of 7.4 cm. The SSH anomaly at C21 over one and a half years was displayed as an example; the SSH anomaly estimated from CPIES was similar to that from ADT (Fig. 2c). However, the RMS error between the two series reached 6.1 cm; this was mainly because the CPIES estimation included more high-frequency variabilities which were not captured by satellites owing to low temporal resolution. Hence, the SSH anomaly from the CPIES was used in the further analysis of the relationship between upper-layer perturbations and TRWs in the deep ocean for the near-21-day period.
Fig. 2.
Fig. 2.

(a) Empirical relationship between deseasoned τ1200 and steric height (ϕ/g). Gray dots are values from historical hydrocast profiles. Black line is fitted from gray dots. (b) Seasonal model for steric height (ϕ/g). Gray dots in (b) indicate deviation between gray dots and fitted line in (a) arranged by days in a year. The black line is fitted from gray dots. (c) Comparison of sea surface height variability at C21 obtained from CPIES (black) and ADT (gray).

Citation: Journal of Physical Oceanography 52, 11; 10.1175/JPO-D-22-0065.1

3. Results

a. Near-21-day fluctuations

The variance-preserving spectra of deep currents observed by current meters at M01 and M02 are presented to identify the dominant periods (Figs. 3a,b and 4a,b). Significant intraseasonal fluctuations with periods of approximately 21 days were found at both M01 and M02. At M01, the fluctuations at 1999 m were weaker than those near the bottom (2912 m) and in the upper layer (983 m) (Figs. 3a,b). The meridional component (υ) of the current was much more energetic near the 21-day fluctuations than the zonal component (u) at M02. The fluctuations at 4151 m were significantly stronger than those at 2148 m (Figs. 4a,b). A bottom-enhanced vertical structure was observed at both M01 and M02. Wavelet analysis was used to determine the temporal variance of fluctuations. The near-21-day fluctuations occurred from July to March in two years at 2912 m at M01 (Figs. 3e,f), while they occurred throughout the observation period—except in the autumn of 2018—near the bottom at M02 (Figs. 4e,f). However, the fluctuations were weaker and less durable when leaving the bottom, at both M01 and M02 (Figs. 3c,d and 4c,d).

Fig. 3.
Fig. 3.

Variance-preserving spectra of (a) u and (b) υ, at 983 m (brown), 1999 m (blue), and 2912 m (red) at M01. Dashed lines indicate 95% confidence level of red noise. Wavelet analysis of (c) u and (d) υ at 1999 m at M01. Wavelet analysis of (e) u and (f) υ at 2912 m at M01. Black contours indicate 95% confidence level. Color indicates log2-scaled variance of wavelet transform of normalized value. Green shadows indicate periods of 18–28 days.

Citation: Journal of Physical Oceanography 52, 11; 10.1175/JPO-D-22-0065.1

Fig. 4.
Fig. 4.

As in Fig. 3, but for u and υ at 2148 m (blue) and 4151 m (red) at M02.

Citation: Journal of Physical Oceanography 52, 11; 10.1175/JPO-D-22-0065.1

A third-order Butterworth filter with cutoff periods of 18–28 days (as indicated by green shadows in Figs. 3 and 4) was further applied to obtain near-21-day fluctuations (u21, υ21). The contributions of u21 (υ21) to the variance of the current—which were determined by the STD ratios of u21 (υ21) to the hourly records—were 18% (24%) at 2912 m and 10% (10%) at 1999 m at M01. For M02, the fluctuation of u21 (υ21) accounted for 13% (18%) of the variance at 4151 m, and 7% (8%) of the variance at 2148 m. The contributions of u21 (υ21) near the bottom were 19% (31%) of the tidal contributions on average; the ratio was substantial because the LS characterized by complex topography is a region where energetic internal tides are generated. More substantially, for 72-h low-pass filtered currents in which high-frequency including tides and near-inertial waves were removed, the contribution of u21 (υ21) reached 72% (53%) at 2912 m and 46% (34%) at 1999 m at M01, and 30% (39%) at 4151 m and 17% (21%) at 2148 m at M02.

The feature of the bottom trap follows the characteristics of TRWs, which are a candidate for intraseasonal fluctuations in the deep SCS; and further analysis was performed to confirm whether the near-21-day fluctuations in our observations can be considered as TRWs as well. The process of the identification of TRWs was 1) the theoretical direction of wavenumber vector and cg was obtained from Eq. (4) and indicated by green and purple arrows in Fig. 5, respectively; 2) the major axes of STD of near-21-day fluctuations, which represents the direction of energy propagation, were obtained from current observations; and 3) the major axes of standard deviation ellipses were compared with the purple arrows. As shown in Fig. 5, all the major axes presented a southwest–northeast orientation, indicating the direction of energy propagation. Although both acute and obtuse angles were obtained when Eq. (4) was used, the downslope cg exhibited fewer deviations from the major axes of the ellipses. The deviations of the major axes from the theoretical values were small at M02 (3.2° at 4151 m and 11.5° at 2148 m). The cg based on both the observation and the dispersion relation was generally along the isobaths. However, M01 showed a much greater deviation (44.0° at 2912 m and 31.7° at 1999 m). The mooring was located in the northern corner of the MT, with 3000-m isobaths on three sides of it. Although cg (nearly westward) from the dispersion relation was nearly perpendicular to ∇H (nearly southward), the gradient provided by the smoothed topography may have been inaccurate. This indicates that the near-21-day fluctuations at M02 follow the dispersion relation of TRWs, while the significant deviation at M01 may be due to the complex topography at the location.

Fig. 5.
Fig. 5.

Standard deviation ellipses of the near-21-day variance from moorings. Blue ellipses indicate the variance at 1999 m of M01 and 2148 m of M02. Red ellipses indicate the variance at 2912 m of M01 and 4151 m of M02. Directions of downslope (∇H, black arrows), group velocity (cg, purple arrows), and wavenumber vector (K, green arrows) are indicated. Deviation of group velocity direction and ellipse major axes are listed. Isobaths are indicated by gray lines.

Citation: Journal of Physical Oceanography 52, 11; 10.1175/JPO-D-22-0065.1

To describe the spatial distribution of near-21-day fluctuations and to further confirm whether the fluctuations were TRWs, recordings of near-bottom currents from the CPIESs were analyzed. According to the results of wavelet analysis, all observations of currents in the deep MT showed near-21-day fluctuations, except for C20 which was on the north side of the array (Fig. 6); however, CPIESs west of the MT did not capture the fluctuations (not shown). This signifies that near-21-day fluctuations were common but trapped in the MT. Although fluctuations with wider bands were identified from Fig. 6, only 18–28-day fluctuation was analyzed as fluctuations with other periods were not widely existent in the MT. The meridional component (υ) showed more energetic and longer-lasting fluctuations in general. The fluctuations had complex temporal distributions, and energetic fluctuations appeared at different times among the sites. Specifically, the site closest to M01 (C21) showed a temporal distribution similar to that at M01. The energetic fluctuations during October–January and April–July at M02 also appeared at the site closest to it (C27). However, the fluctuations during January–April at M02 were not significantly observed at C27. Energetic fluctuations at the northern and eastern sites (C13–C18, C21–C24, and C25) mostly occurred during August–March, but were completely absent during April–May. However, the fluctuations were significant in the spring, at C27 and C28.

Fig. 6.
Fig. 6.

Wavelet analysis of near-bottom u and υ from CPIES observations. Black contours indicate 95% confidence level. Color indicates log2-scaled variance of wavelet transform of normalized value. Green shadows indicate periods of 18–28 days. Subgraphs are arranged following spatial distribution of CPIESs.

Citation: Journal of Physical Oceanography 52, 11; 10.1175/JPO-D-22-0065.1

The near-21-day fluctuations observed by CPIESs accounted for 5%–30% (10%–44%) of the zonal (meridional) component of the abyssal current, with an average of 12% (18%); this was 15% (27%) of the tidal contributions. For 72-h low-pass-filtered currents, the average contributions reached 43% (44%). The fluctuations were identified as TRWs following the process used for mooring observations introduced above. Most 18–28-day STD ellipses from the CPIES observations presented major axes parallel to the isobaths, as displayed in Fig. 7. The directions of ∇H are also shown; the directions of K and cg were calculated using the dispersion relation. Similar to the mooring observations, the downslope cg exhibited fewer deviations from the major axes of the ellipses. Although the average deviation was approximately 21°, the median deviation was only 11.9°, and the deviations at most sites were less than 20°. This indicates that the near-21-day fluctuations followed the characteristics of TRWs at most sites. The energy of the fluctuations propagated anticlockwise along the isobaths in the MT, as indicated by the direction of cg.

Fig. 7.
Fig. 7.

Standard deviation ellipses of the near-21-day variance from CPIES observations. Directions of downslope (∇H, black arrows), group velocity (cg, purple arrows), and wavenumber vectors (K, green arrows) are indicated. Deviation of group velocity direction and ellipse major axes are listed. Isobaths are indicated by gray lines.

Citation: Journal of Physical Oceanography 52, 11; 10.1175/JPO-D-22-0065.1

Although C16, C25, and C28 showed extremely large deviations, the ellipses at C16 and C25 were almost circular, and their major axes were not reliable. Furthermore, large deviations there were possibly related to small-scale regional topographical characteristics. Figure 7 shows a cross-slope major axis at C25 and winding 3000-m isobaths around C16. While the 4000-m isobaths near C28 were parallel to the major axis, the 3000-m isobaths were bent. It may be noted that C21 was only 12 km north of M01, and although the fluctuations showed similar temporal distributions (Figs. 3e,f and 6), the directions of the STD ellipses were quite different (Figs. 5 and 7). C21 indicated a northwest–southeast major axis following the dispersion relation of TRWs, while the southwest–northeast major axis at M01 deviated from the theoretical direction. This implies that the regional variance in the near-21-day fluctuations was conspicuous there, which might be due to the complex topography with abrupt changes at the corner of the MT.

In summary, the energetic, bottom-trapped, near-21-day fluctuations widely observed by moorings and CPIESs in the deep northern MT followed the dispersion relation of TRWs. However, the observations at several sites show large deviations when compared with the dispersion relation owing to the complex and small-scale local topography. Near-21-day TRWs propagated anticlockwise along the isobaths in the northern MT, with slightly downslope group velocities and upslope wavenumber vectors. However, the near-21-day TRWs were trapped in the MT owing to the downslope group velocities and energy dissipation.

b. Relationship between upper-layer perturbations and TRW generation

Perturbations in the upper layer interacting with complex topography are considered as an important factor in the generation of TRWs (e.g., Hamilton 2009; Shu et al. 2016; Ku et al. 2020; Wang et al. 2021). TRWs are generated through energy transfer from the upper-layer perturbations by potential vorticity adjustment to changing depths of the bottom and the interface between layers (Hamilton 2009). As shown in Figs. 3a, 3b, and 9b, although the energy of the near-21-day fluctuations was strong at 2912 m and weak at 1999 m, that at 983 m was significantly strong, indicating the possible energy source of the near-21-day bottom-trapped TRWs and implying that upper perturbations cannot directly reach the bottom but might influence the deep ocean through PV adjustment.

To identify the relationship between the upper layer and deep ocean at near-21-day periods, the squared coherences between the near-bottom kinetic energy (2912 m at M01 and 4151 m at M02) and the upper-layer kinetic energy (average above 800 m from ADCPs) were calculated (Fig. 8). The near-bottom kinetic energy values at M01 and M02 were conspicuously coherent with the upper-layer variations at the periods under focus. The squared coherences reached 0.83 at approximately 23 days at M01 and 0.55 at approximately 21 days at M02—significantly higher than the 95% confidence levels (0.50 for M01 and 0.43 for M02)—indicating a possible connection between the deep fluctuations and the upper-layer perturbations. The near-bottom variations corresponding to the upper-layer variations revealed an approximate time lag of half a day at M01 and 10 days at M02.

Fig. 8.
Fig. 8.

(a) Squared coherence and (b) phase difference, between near-bottom kinetic energy and upper-layer kinetic energy at M01. The 95% confidence level (horizontal lines) and 18–28-day period (gray shadows) are shown. (c),(d) As in (a) and (b), but at M02. Upper kinetic energy is the average of values above 800 m from ADCPs; near-bottom kinetic energy is from the deepest current meters on both moorings.

Citation: Journal of Physical Oceanography 52, 11; 10.1175/JPO-D-22-0065.1

As shown by the time series of 18–28-day bandpass-filtered eddy kinetic energy (EKE) in the deep ocean, two and three energetic near-21-day TRWs events occurred at M01 and M02 during the observation interval, respectively (Fig. 9). It may be noted that different y axes were used to represent the EKE near (red) and far from (blue) the bottom; a bottom-enhanced feature appeared in all five events. As indicated by the orange boxes, energetic events occurred more frequently at M02 (July 2018–August 2018, November 2018–February 2019, and April 2019–May 2019) while occurred only in winter (December 2017–March 2018 and November 2018–January 2019) at M01. The EKE at M01 at 2912 m was approximately 6 times greater than that at 1999 m. The energy at 4151 m was 2–4 times greater than that at 2148 m for the first two events at M02 (events C and D), while fluctuations were not observed at 2148 m in Event E, which might be due to the different energy source as discussed below.

Fig. 9.
Fig. 9.

Time series of 18–28-day bandpass-filtered EKE (a) in upper layer and (b) at 983 m (brown), 1999 m (blue), and 2912 m (red) at M01. (c),(d) As in (a),(b), but for M02 and for different depths. Bandpass-filtered EKE at 983 m is also presented at the bottom of (a). Significant durations of near-21-day TRWs are indicated by orange boxes.

Citation: Journal of Physical Oceanography 52, 11; 10.1175/JPO-D-22-0065.1

The time series of 18–28-day EKE in the upper layer from the ADCPs equipped on M01 and M02 was used to determine the possible energy sources (Figs. 9a,c). Deep near-21-day TRWs generally correspond to energetic, local, upper-layer perturbations. However, perturbations in the upper ocean do not always induce TRWs in the deep ocean. At M01, the strong upper perturbations in Event A reached a depth of approximately 1000 m and were captured by the current meter at 983 m, whereas the perturbations in Event B indicated the highest energy levels in the uppermost 300 m, with the perturbations reaching a depth of over 600 m. The deeper but weaker upper perturbations in Event A caused much more energetic TRWs in the deep ocean than the stronger but shallower upper perturbations in Event B. Although the upper perturbations were stronger in April–June 2018, they were shallower than in Event A and weaker than in Event B. The upper perturbations were energetic in September 2018 and deep in March 2019, and TRWs were not generated owing to the shallow depth or weak surface energy. This suggests that only deep upper perturbations with strong energy induced strong TRWs at M01. The TRWs at M02 were much weaker than those at M01, which might be due to weaker upper-layer perturbations or greater bottom depth. The TRWs in the winter of 2018 (Event D) corresponded well with the strong and deep upper-layer fluctuations at M02. Although upper-layer perturbations were also present in events C and E, they were weak and not as deep as in the other events. Near-21-day bottom energy at these two events may have resulted from the superposition of TRWs generated locally and TRWs propagated from elsewhere, which will be further discussed in the next section.

Large-scale in situ measurements of upper-layer currents around our study region during the observation period are few. However, the time series of τ recorded by the CPIESs—the variations of which were dominated by variations of the thermohaline structure—can be used to monitor the upper-layer perturbations. Although not as high as the coherences between near-bottom and upper currents from moorings, near-bottom currents (either u or υ at most sites) and τ showed coherences greater than the 95% confidence level at 18–28-day periods during the durations with abundant TRWs (July 2018–April 2019, as indicated in Fig. 6), especially in the eastern column of CPIESs (Fig. 10). Compared with other periods, fluctuations with 18–28-day periods were the only intraseasonal variances that widely showed coherence with upper-layer fluctuations at most sites. This indicates that abyssal TRWs are generally in coherence with upper-layer perturbations.

Fig. 10.
Fig. 10.

Squared coherence between near-bottom currents (u, dashed lines; υ, solid lines) and τ from CPIESs during July 2018–April 2019. The 95% confidence level (horizontal lines) and 18–28-day periods (gray shadows) are indicated. Subgraphs are arranged following spatial distribution of CPIESs.

Citation: Journal of Physical Oceanography 52, 11; 10.1175/JPO-D-22-0065.1

The time series of τ was converted to SSH variability through an empirical relationship (Fig. 2). The amplitude of the bandpass filtered SSH anomaly varied from 3.34 (C27) to 8.90 cm (C15) among the sites. The near-21-day SSH variabilities were significant at the center of the LS (C13–C18, C23–C25), whereas they were weak in the northwest and southwest of the array (bottom-right panel of Fig. 11). Here, we found that most near-21-day TRWs in the deep ocean (defined as TRWs with EKE greater than the mean plus STD) were accompanied by energetic upper-layer perturbations—indicated by the magenta boxes in Fig. 11—signifying their energy sources.

Fig. 11.
Fig. 11.

Time series of 18–28-day bandpass-filtered near-bottom EKE (black) and SSH anomaly (blue) from CPIES observations. Near-bottom EKE at M01 and M02 (red) are also shown at C21 and C27, respectively. All time series were normalized, and normalized coefficients are shown. Light cyan and pink shadows indicate winter and summer, respectively. Magenta boxes indicate energetic TRWs related to local upper-layer perturbations, and blue boxes indicate energetic TRWs that have no relation to the local upper layer. Blue dashed lines show upper envelope lines, and horizontal lines indicate mean plus STD values. Subgraphs are arranged following spatial distribution of CPIESs. STDs of 18–28-day SSH anomaly from June 2018 to July 2019 are shown in the bottom-right panel. The red box in the bottom-right panel is the region used in Fig. 12.

Citation: Journal of Physical Oceanography 52, 11; 10.1175/JPO-D-22-0065.1

Near-bottom EKE at M01 and M02 is indicated by red lines at their closest sites (C21 and C27), and mooring observations show variations similar to those of the CPIES observations. It may be noted that the near-bottom currents recorded by CPIESs (∼50 m above the seafloor) were closer to the bottom than those recorded at M01 and M02 (∼220 and ∼120 m above the seafloor, respectively), and the former had higher energy than the latter as shown in the figure, indicating bottom enhancement. The energetic upper perturbations observed by ADCPs at M02, and τ at C27 in the winter of 2018 corresponded to the bottom-trapped TRWs observed at M02 (Fig. 9d). However, the near-bottom variations in the winter of 2018 did not reach the mean plus STD levels (Fig. 11). They were much weaker than TRWs at other times when the upper perturbations were not as strong as those in the winter of 2018. This suggests that TRWs at M02 and C27 in the summer of 2018 and spring of 2019 were not completely related to perturbations in the upper ocean. Special cases were also noticed at other sites; although the upper ocean currents showed energetic variability in winter 2018 at C24–C27, only C26 showed energetic TRWs in the deep ocean. A possible reason for this may be that the upper perturbations were too shallow to induce a deep-ocean response. In some other cases, TRWs in the deep ocean were not related to local upper ocean fluctuations. For example, although the local upper-layer variabilities in September 2018 and March 2019 were not significant at C23–C25, TRWs in the deep ocean were energetic.

As the TRWs were generated through PV adjustment and generally propagated along the isobaths with shallow water to their right, the TRWs observed by CPIESs whose cause cannot be attributed to local upper-layer perturbations may be related to propagation. TRWs observed at different times in different regions; TRWs related to local upper-layer perturbations are indicated by magenta boxes, while those related to propagation are shown by blue boxes in Fig. 11. The TRWs in summer 2019 were not as significant as the energetic TRWs during other durations probably because the amplitude at the end of the records was distorted during the filtration process. However, consistent peaks appeared at most sites during June. We have divided the months into seasons as April–May (spring), June–September (summer), October–early November (autumn), and November–March (winter), following Zhao and Zhu (2016); TRWs mostly appeared in winter and summer (indicated by light cyan and pink shadows in Fig. 11, respectively).

Near-21-day TRWs showed complex spatiotemporal distributions in the MT, as described above. There were chaotic variations in the amplitude of TRWs among sites because of the combined effect of upper perturbations, topographical gradients, water depths, energy dissipation, and energy superposition. On the one hand, upper-layer perturbations can induce bottom-enhanced TRWs locally; on the other hand, TRWs generated in the deep ocean can propagate along isobaths. In some cases, TRWs may be a result of the superposition of locally generated TRWs and TRWs propagated from elsewhere. Locally generated TRWs dominated the eastern and northern parts of the array, while the southwestern region presented more adjacently propagated TRWs. This may be related to differences in the distribution of upper perturbations and water depth; the near-21-day SSH anomaly showed a low amplitude at C26–C28, and the water depths were considerable at these locations.

4. Discussion

a. Coupling of upper and lower layer

Although there was a relationship between near-21-day TRWs and upper-layer perturbations according to Figs. 811, the dynamic process of TRW generation was not revealed based on the analysis above. As most cases of TRWs obtained energy from the upper-layer perturbations through PV adjustment, the coupling of the upper and lower layer was examined. A two-layer model was assumed in the study region with the interface of the 8°C isotherm, which refers to the interface chosen in the Loop Current region (Hamilton 2009) and Kuroshio depth in the study region (Long et al. 2021). The depth of the 8°C isotherm is obtained based on the empirical relationship from historical profiles as shown in Fig. 12a, the process of which is same as calculating steric height from CPIES observations as introduced in section 2. However, the seasonal model was not considered for the 8°C isotherm because it can be ignored below 300 dbar.

Fig. 12.
Fig. 12.

(a) Empirical relationship between deseasoned τ1200 and the 8°C isotherm. Gray dots are values from historical hydrocast profiles. Black line is fitted from gray dots. (b) Average depth of the 8°C isotherm among C23–C25. (c) Normalized potential vorticity anomaly (NPVA) in upper (red) and lower (blue) layers in red box in the bottom-right panel of Fig. 11. Light cyan and pink shadows indicate winter and summer, respectively. (d) As in (c), but for 18–28-day bandpass-filtered records. (e) Squared coherence and (f) phase difference between NPVA in upper and lower layers during July 2018–April 2019. Horizontal and vertical dashed lines indicate 95% confidence level and 21-day period, respectively. Gray shadows indicate the 18–28-day period.

Citation: Journal of Physical Oceanography 52, 11; 10.1175/JPO-D-22-0065.1

The PV anomaly is defined as (Hamilton 2009)
PVA=ζ+fhfH,
where h is the layer thickness, H is the average thickness of the layer, and ζ=(υ/x)(u/y) is the relative vorticity. Then normalizing by H/f,
NPVA=Hhζf+(Hh1).
The currents in upper and deep layers during the calculation of ζ were mapped from CPIES observations, the detailed process is presented by Zheng et al. (2022).

Considering that C23–C25 showed energetic TRWs at similar depths, normalized potential vorticity anomaly (NPVA) in the box in bottom-right panel of Fig. 11 was calculated. Layer thickness was estimated by the average depth of the 8°C isotherm at C23–C25 (Fig. 12b). The upper and deep NPVA were estimated from surface currents and deep currents at 3000 dbar, respectively, as shown in Fig. 12c. The upper and lower PVA generally showed a significant negative correlation, indicating the PV adjustment. The squared coherences between upper and lower NPVA during July 2018–April 2019, the durations which had abundant TRWs, reached 0.81 at 21 days, significantly higher than the 95% confidence levels (0.43) (Fig. 12e). The phase difference was 142° (Fig. 12f), indicating that the phases in the upper and lower layers were nearly opposed. For 18–28-day bandpass-filtered NPVA, a negative correlation was observed in summer 2018 and winter 2018 (Fig. 12c), corresponding to the observation of TRWs in the deep ocean. Although significant coherence was found at ∼40 days, it might be a regional variance as only C23–C26 showed consecutive energetic fluctuations at 40-day period (Fig. 6). When negative vorticity perturbations dominated the upper layer, greater thickness of the upper layer and a negative NPVA appeared. Meanwhile, the thickness of the lower layer decreased, the lower layer showed a positive NPVA, and the water columns traveled across the isobaths under PV conservation. TRWs were generated during the cross-slope traveling of water columns.

b. Dynamics of near-21-day upper-layer perturbations

As most cases of energetic near-21-day TRWs were associated with upper-layer perturbations, especially at the center of the LS, the ADT was analyzed to diagnose the dynamics of the upper-layer fluctuations. The TRWs generally occurred in winter and summer; therefore, four cases, i.e., during winter 2017 (Case 1), summer 2018 (Case 2), winter 2018 (Case 3), and summer 2019 (Case 4) when the upper-layer perturbations were energetic were scrutinized to analyze the upper-layer dynamics (Fig. 13).

Fig. 13.
Fig. 13.

(a) Daily maps of Case 1 in winter 2017. ADT (color) and surface geostrophic currents (arrows) are shown. The 18–28-day bandpass-filtered SSH anomaly (SSHA) is indicated by colored dots. (b)–(d) As in (a), but for Case 2 in summer 2018, Case 3 in winter 2018, and Case 4 in summer 2019, respectively. Blue boxes in (a) and (d) indicate regional Kuroshio path meanders. Blue boxes in (b) indicate the regions of eddies. Blue boxes in (c) indicate the regions at center of LS with significant ADT variances.

Citation: Journal of Physical Oceanography 52, 11; 10.1175/JPO-D-22-0065.1

Although most instruments were not deployed in the winter of 2017, significant upper-layer fluctuations were observed at M01, C21, and C23. Kuroshio intrusion with a looping path occurred in the winter, and daily maps of ADT and surface geostrophic currents from 21 January to 15 February were extracted at 5-day intervals to identify upper-layer perturbations (Fig. 13a). The Kuroshio meandered during the intrusion with a period of approximately 21 days (the meandering is indicated by blue boxes). The path was in the middle of the box on 21 January but moved to the northeast on 26 January. It moved back to the middle on 31 January and then meandered further southwest on 5 February. After 5 February, the path shifted again, moving to the northeast of the box on 15 February. The near-21-day SSH anomaly at C21 and C23 indicated a negative value on 26 January, when the Kuroshio path was located at the northeast of the box. In contrast, a positive value occurred on 5 February, with the southwest shift of the Kuroshio path. Thus, Case 1 is related to the regional meandering of the Kuroshio at the center of the LS during the intrusion.

The Kuroshio generally presented a leaping path in the summer of 2018; however, near-21-day-period variabilities related to eddies were found north of Luzon which propagated northwestward (Fig. 13b). A cyclonic eddy—shown in deep blue surrounded by light blue, and indicated by the blue box—was observed north of Luzon on 2 July. It then propagated northwestward and merged into the large low-ADT region southwest of Taiwan on 7 July. An anticyclonic eddy was observed on 12 and 17 July at the center of the LS, as indicated by the blue boxes. Meanwhile, a cyclonic eddy was generated southeast of the anticyclonic eddy north of Luzon. The eddy pair moved northwestward from 12 to 27 July, covering the northeastern part of the observation area. The bandpass-filtered SSH anomaly clearly showed the propagation of the near-21-day upper variability. C18 first attained the largest negative (positive) value when the cyclonic (anticyclonic) eddy reached it on 2 and 22 July (July 12); C13–C16 and C22–C25, which were on the northwest of C18, showed negative (positive) on 7 and 27 July (July 17) as the eddy propagated northwestward. The generation of these eddies was related to the local wind stress curl north of Luzon, the details of which have been presented by Zhao et al. (2022).

The Kuroshio intrusion also occurred in the winter of 2018; however, both looping and leaking paths appeared, which made the upper-layer current system much more complex (Fig. 13c). The anticyclonic eddy at C20 and C21 on 1 December fell out of the Kuroshio looping path in mid-November, and a cyclonic eddy south of the anticyclonic eddy was generated, accompanied by the looping intrusion (not shown). The ADT at the center of the LS (indicated by blue boxes) increased (1–11 December) when the cyclonic eddy moved westward. The Kuroshio leaked into the SCS and turned counterclockwise southwest of Taiwan on 21–26 December, which again decreased the ADT in the area within the boxes. The near-21-day SSH anomaly showed a negative value when the cyclonic eddy dominated the area within the boxes (1–6 December), and became positive when the ADT in the area within the boxes increased as the eddy moved westward (11–16 December). It reverted to a negative value when the ADT decreased again owing to the anticlockwise leaking path (21–26 December). The variances at other sites were observed to follow those at C17 and C18, indicating possible northwest propagation of the perturbations. Hence, Case 3 is resulted from the combined effect of the Kuroshio-induced eddy and Kuroshio leaking.

In the summer of 2019, the leaping path of the Kuroshio was more stable than in the summer of 2018; however, regional meandering occurred in the LS, as indicated by the blue boxes (Fig. 13d). It was obvious that the Kuroshio meandered southwestward during 2–7 June and on 27 June, but meandered northeastward during 12–17 June. The near-21-day SSH anomalies at C17 and C18 located in the meander zone led to significant anomalies at other sites. The anomalies were positive (negative) at C17 and C18 when the Kuroshio meandered southwestward (northeastward). A possible cause of the Kuroshio meander was the variability generated north of Luzon, which was similar to that generated in the summer of 2018. As shown in Fig. 13d, a significantly low ADT appeared north of Luzon on 2 June and propagated northwestward along the western edge of the Kushiro path on the following days, which further made the Kuroshio meander.

The energetic near-21-day TRWs were mostly related to the upper ocean; the perturbations in the upper ocean with near-21-day periods were induced by the Kuroshio path variabilities and eddies. Variations with similar periods in Kuroshio intrusion in the LS have been observed at mooring from 2007 to 2011 (Yuan et al. 2017). The upper perturbations occurred in the center of the LS in all the cases discussed above; however, there were differences in the dynamic processes among the cases. Kuroshio intrusion occurred during the winter both in 2017 and 2018. The looping Kuroshio path showed regional meandering with periods of approximately 21 days in the winter of 2017. The upper currents in winter 2018 were complex, as both looping and leaking paths were found, accompanied by near-21-day surface variabilities. Kuroshio leaped the LS in summer; variabilities generated north of Luzon were considered the possible causes of the upper-layer variabilities. However, the detailed dynamic processes differed between the two years. The variabilities generated north of Luzon propagated northwestward and caused the meandering of the Kuroshio path in summer 2019, but propagated northwestward as eddies that covered the study region in summer 2018, possibly because the Kuroshio in summer 2019 was stronger and more stable than that in the summer of 2018.

The near-21-day variabilities related to the Kuroshio path meanderings and eddies were distinguishable and persistent in the gridded ADT distributions for winter 2017, summer 2018, and summer 2019. The upper current system was variable in the winter of 2018, which manifested as the looping Kuroshio path, the leaking Kuroshio path, and their related eddies. Although the near-21-day perturbations were not perceivably presented by the gridded ADT, they did exist in the complex upper currents according to the CPIES-measured SSH anomaly. Considering the vertical structure of the Kuroshio, surface perturbations related to the Kuroshio can be deep. Although the perturbations in the summer of 2018 were induced by eddies, the anticyclonic eddies generated there carried the Kuroshio water can reach a considerable depth. Combined with the steep topography, the deep upper-layer perturbations in winter and summer in the LS were more likely to induce TRWs.

c. Propagation of near-21-day TRWs

Near-21-day TRWs were induced by local upper-layer perturbations most of the time; however, there were many cases that had no relation to local perturbations in the upper ocean, especially at the western sites of the array. A ray-tracing model was used to trace the propagation path of the energy, as shown in Fig. 14a. As the propagation of TRWs could be affected by background currents, rays with (dark red lines) and without (red lines) consideration of background currents were obtained, respectively. According to CPIES observations, the deep MT showed a cyclonic circulation along the slope. The currents were weak around C28 in spring when energetic TRWs were observed there, and showed low velocity in all months at the north corner of MT; therefore, background currents were considered as 1 cm s−1 along the slope in the ray-tracing model. Although a slightly faster group speed appeared when considering the background currents, the rays were similar to those without background currents. This indicated that the effect of background currents was small as the group velocity of TRWs was significantly faster than background currents, which is consistent with the TRWs modeling study in the SCS (Quan et al. 2021a). Rays with background currents showed a smaller angle with the isobaths due to the along-slope background currents.

Fig. 14.
Fig. 14.

(a) Wave rays (red and dark red lines) and wavenumber vectors (blue arrows) from ray-tracing model. Arrows are plotted at 5-day intervals. Isobaths are indicated by gray lines. Red and dark red lines are rays without and with consideration of background currents. (b)–(g) Time series of 18–28-day bandpass-filtered EKE at CPIES and mooring sites. Magenta boxes indicate TRWs related to local upper-layer perturbations, and blue boxes indicate TRWs related to propagation. Horizontal lines indicate mean plus STD values. Red arrows indicate propagation of TRWs.

Citation: Journal of Physical Oceanography 52, 11; 10.1175/JPO-D-22-0065.1

The rays traced backward from C23 reached the regions around C14 and C15, and the rays traced backward from C25 reached regions around C23, indicating the possibility of energy propagation between these sites. The time series of 18–28-day bandpass-filtered EKE at these sites confirmed the propagation (Figs. 14bg). Energetic TRWs propagated from C15 reached C23 in January 2019 with a time lag of approximately 25 days, similar to the time indicated by the ray-tracing model (∼38 days). It took 20–40 days for the TRWs at C23 to propagate to C25. The time taken for propagation between these two sites in the ray-tracing model was approximately 32 days without background currents and approximately 25 days with background currents, which was close to the observed time. In addition, the rays traced backward from M02 across to C28, the observed time lag between M02 and C28 was approximately 30 days, close to that obtained from the ray-tracing model without background currents (∼30 days) and with background currents (∼20 days). The energetic TRWs between January and April 2019 at C28 propagated to M02 and further reached C27, as shown in Figs. 14eg.

Analyzing the rays from the ray-tracing model, most cases that have no relation to local upper-layer variabilities can be explained as TRWs propagated from other regions (Fig. 14a). The TRWs of nonlocal origin at C23 and C25 propagated from their upstream, where TRWs were generated related to upper perturbations there. However, the TRWs during March–April 2019 at C28 were traced back to the western gaps of the LS. Previous studies have indicated intraseasonal variations with periods of approximately 30 days in the LS, which were strong in spring (Zhou et al. 2014; Ye et al. 2019). Hence, the source of TRWs of nonlocal origins at C28 may be related to intraseasonal variations in the LS. One possibility is that the TRWs generated in the LS propagated into the SCS along the isobaths, another possibility being that the high-PV deep-water overflow with intraseasonal variations intruded the SCS and generated TRWs near the western gaps of the LS.

Although the existence of TRWs in the deep SCS induced by upper-layer perturbations has been reported by previous studies (Shu et al. 2016; D. Wang et al. 2019; Zheng et al. 2021b; Wang et al. 2021) and TRWs with similar periods have been observed west of the LS (Wang et al. 2021), the observations from this study provide unprecedented coverage of TRWs in the northeastern SCS, and reveal the connection between the Kuroshio variabilities (Kuroshio meandering, Kuroshio intrusion, and Kuroshio-related eddies) and the TRWs based on observations for the first time. TRWs related to the loop current of the Gulf Stream exist in the northern half of the Gulf of Mexico basin, accounting for 80%–90% of the low-frequency motion (Hamilton 2009). Similarly, TRWs act as a bond of energy propagation between the Kuroshio variabilities and intraseasonal variations of the abyssal circulation in the SCS. The near-21-day TRWs propagated anticlockwise in the MT with shallow water to their right. However, the TRWs were trapped in the trough owing to the topographical characteristics and downslope group velocities; the contribution of the near-21-day TRWs was much lower than that in the Gulf of Mexico because the deep LS is a region where energetic internal waves and deep eddies are present. TRWs with other periods or generated west of the MT may propagate to the interior SCS and significantly contribute to abyssal circulation, as reported by Zheng et al. (2021b).

5. Summary

TRWs with periods of approximately 21 days were detected in the northern MT, west of the LS, SCS based on observations extending over 400 days from June 2018 to July 2019 utilizing 15 CPIESs and two moorings. The TRWs exhibited complex spatiotemporal distribution and energy sources. Some cases of TRWs were locally generated owing to upper-layer perturbations, whereas others originated in adjacent regions. Locally generated TRWs dominated the study region; however, several cases of TRWs propagated from adjacent regions were observed in the western part of the region.

Locally generated TRWs were induced by upper-layer perturbations—which showed significant seasonal variations—through PV adjustment. The near-21-day upper perturbations were widely generated in winter and summer but scarce in spring and autumn; however, different cases of energetic perturbations corresponded to different upper-layer dynamic processes. The energetic events in the winter of 2017 were related to the meandering of the looping Kuroshio path during the intrusion, whereas those in the winter of 2018 were accompanied by looping and leaking Kuroshio paths. The upper-layer perturbations in summer resulted from the variabilities north of Luzon. The eddies generated there propagated and covered the CPIES array during the summer of 2018. The northwestward-propagating variabilities caused regional Kuroshio meandering in the LS in the summer of 2019. Propagated TRWs in the northern part of the array were induced by the upper-layer fluctuations at adjacent regions, while those on the south originated from the west gap of the LS, probably related to the intraseasonal variations in the LS.

This study describes the generation, propagation, and spatiotemporal distribution of TRWs in the northern SCS and confirms that the Kuroshio and upper eddies play important roles in the dynamic processes associated with intraseasonal variations in the deep SCS. The study may therefore contribute to our knowledge of the dynamic responses of the abyssal current to mesoscale perturbations in the upper ocean.

Acknowledgments.

This study was sponsored by the National Natural Science Foundation of China (Grants 41920104006 and 41906024), the Scientific Research Fund of Second Institute of Oceanography, MNR (Grants JZ2001 and QNYC2102), the Project of State Key Laboratory of Satellite Ocean Environment Dynamics, Second Institute of Oceanography (SOEDZZ2106 and SOEDZZ2207), the Oceanic Interdisciplinary Program of Shanghai Jiao Tong University (Project SL2021MS021), the Innovation Group Project of Southern Marine Science and Engineering Guangdong Laboratory (Zhuhai) (311020004), and the Global Climate Changes and Air-sea Interaction Program (GASI-02-PAC-ST-Wwin).

Data availability statement.

Bathymetry data were obtained from ETOPO1 (doi:10.7289/V5C8276M). Surface geostrophic currents and absolute dynamic topography were obtained from CMEMS (http://marine.copernicus.eu/). For access to the analyzed mooring data, contact the corresponding author.

REFERENCES

  • Alford, M. H., and Coauthors, 2015: The formation and fate of internal waves in the South China Sea. Nature, 521, 6569, https://doi.org/10.1038/nature14399.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Baker-Yeboah, S., D. R. Watts, and D. A. Byrne, 2009: Measurements of sea surface height variability in the eastern South Atlantic from pressure sensor–equipped inverted echo sounders: Baroclinic and barotropic components. J. Atmos. Oceanic Technol., 26, 25932609, https://doi.org/10.1175/2009JTECHO659.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Donohue, K. A., D. R. Watts, K. L. Tracey, A. D. Greene, and M. Kennelly, 2010: Mapping circulation in the Kuroshio Extension with an array of current and pressure recording inverted echo sounders. J. Atmos. Oceanic Technol., 27, 507527, https://doi.org/10.1175/2009JTECHO686.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Gan, J., Z. Liu, and C. R. Hui, 2016: A three-layer alternating spinning circulation in the South China Sea. J. Phys. Oceanogr., 46, 23092315, https://doi.org/10.1175/JPO-D-16-0044.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hamilton, P., 2009: Topographic Rossby waves in the Gulf of Mexico. Prog. Oceanogr., 82, 131, https://doi.org/10.1016/j.pocean.2009.04.019.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hamilton, P., A. Bower, H. Furey, R. Leben, and P. Pérez-Brunius, 2019: The loop current: Observations of deep eddies and topographic waves. J. Phys. Oceanogr., 49, 14631483, https://doi.org/10.1175/JPO-D-18-0213.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kennelly, M. A., K. L. Tracey, and D. R. Watts, 2007: Inverted echo sounder data processing manual. University of Rhode Island GSO Tech. Rep. 2007-02, 90 pp., https://digitalcommons.uri.edu/cgi/viewcontent.cgi?article=1001&context=physical_oceanography_techrpts.

    • Crossref
    • Export Citation
  • Ku, A., Y. H. Seung, C. Jeon, Y. Choi, E. Yoshizawa, K. Shimada, K.-H. Cho, and J.-H. Park, 2020: Observation of bottom-trapped topographic Rossby waves on the shelf break of the Chukchi Sea. J. Geophys. Res. Oceans, 125, e2019JC015436, https://doi.org/10.1029/2019JC015436.

  • Lan, J., Y. Wang, F. Cui, and N. Zhang, 2015: Seasonal variation in the South China Sea deep circulation. J. Geophys. Res. Oceans, 120, 16821690, https://doi.org/10.1002/2014JC010413.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Long Y., X.-H. Zhu, X., Guo, F. Ji, and Z. Li, 2021: Variations of the Kuroshio in the Luzon Strait revealed by EOF analysis of repeated XBT data and sea-level anomalies. J. Geophys. Res. Oceans, 126, e2020JC016849, https://doi.org/10.1029/2020JC016849.

    • Crossref
    • Export Citation
  • Meinen, C., E. Fields, R. Pickart, and D. R. Watts, 1993: Ray tracing on topographic Rossby waves. University of Rhode Island GSO Tech. Rep. 1993-05, 49 pp., https://digitalcommons.uri.edu/cgi/viewcontent.cgi?article=1009&context=physical_oceanography_techrpts.

    • Crossref
    • Export Citation
  • Oey, L.-Y., and H.-C. Lee, 2002: Deep eddy energy and topographic Rossby waves in the Gulf of Mexico. J. Phys. Oceanogr., 32, 34993527, https://doi.org/10.1175/1520-0485(2002)032<3499:DEEATR>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Park, J.-H., D. R. Watts, K. A. Donohue, and K. L. Tracey, 2012: Comparisons of sea surface height variability observed by pressure-recording inverted echo sounders and satellite altimetry in the Kuroshio Extension. J. Oceanogr., 68, 401416, https://doi.org/10.1007/s10872-012-0108-x.

    • Search Google Scholar
    • Export Citation
  • Pickart, R. S., 1995: Gulf Stream–generated topographic Rossby waves. J. Phys. Oceanogr., 25, 574586, https://doi.org/10.1175/1520-0485(1995)025<0574:GSTRW>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Quan, Q., and H. Xue, 2019: Influence of abyssal mixing on the multi-layer circulation in the South China Sea. J. Phys. Oceanogr., 49, 30453060, https://doi.org/10.1175/JPO-D-19-0020.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Quan, Q., Z. Cai, G. Jin, and Z. Liu, 2021a: Topographic Rossby waves in the abyssal South China Sea. J. Phys. Oceanogr., 51, 17951812, https://doi.org/10.1175/JPO-D-20-0187.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Quan, Q., Z. Liu, S. Sun, Z. Cai, Y. Yang, G. Jin, Z. Li, and X. S. Liang, 2021b: Influence of the Kuroshio intrusion on deep flow intraseasonal variability in the northern South China Sea. J. Geophys. Res. Oceans, 126, e2021JC017429, https://doi.org/10.1029/2021JC017429.

    • Crossref
    • Export Citation
  • Shu, Y., H. Xue, D. Wang, F. Chai, Q. Xie, J. Yao, and J. Xiao, 2014: Meridional overturning circulation in the South China Sea envisioned from the high-resolution global reanalysis data GLBa0.08. J. Geophys. Res. Oceans, 119, 30123028, https://doi.org/10.1002/2013JC009583.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Shu, Y., and Coauthors, 2016: Persistent and energetic bottom-trapped topographic Rossby waves observed in the southern South China Sea. Sci. Rep., 6, 24338, https://doi.org/10.1038/srep24338.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Shu, Y., J. Chen, S. Li, Q. Wang, J. Yu, and D. Wang, 2018: Field-observation for an anticyclonic mesoscale eddy consisted of twelve gliders and sixty-two expendable probes in the northern South China Sea during summer 2017. Sci. China Earth Sci., 62, 451458, https://doi.org/10.1007/s11430-018-9239-0.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Shu, Y., and Coauthors, 2022: Deep-current intraseasonal variability interpreted as topographic Rossby waves and deep eddies in the Xisha Islands of the South China Sea. J. Phys. Oceanogr., 52, 14151430, https://doi.org/10.1175/JPO-D-21-0147.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Tian, J., Q. Yang, and W. Zhao, 2009: Enhanced diapycnal mixing in the South China Sea. J. Phys. Oceanogr., 39, 31913203, https://doi.org/10.1175/2009JPO3899.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wang, D., and Coauthors, 2019: Advances in research of the mid-deep South China Sea circulation (in Chinese). Sci. China Earth Sci., 62, 19922004, https://doi.org/10.1007/s11430-019-9546-3.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wang, G., R. X. Huang, J. Su, and D. Chen, 2012: The effects of thermohaline circulation on wind-driven circulation in the South China Sea. J. Phys. Oceanogr., 42, 22832296, https://doi.org/10.1175/JPO-D-11-0227.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wang, J., and Coauthors, 2021: Observed variability of bottom‐trapped topographic Rossby waves along the slope of the northeastern South China Sea. J. Geophys. Res. Oceans, 126, e2021JC017746, https://doi.org/10.1029/2021JC017746.

    • Crossref
    • Export Citation
  • Wang, Q., and Coauthors, 2019: Energetic topographic Rossby waves in the northern South China Sea. J. Phys. Oceanogr., 49, 26972714, https://doi.org/10.1175/JPO-D-18-0247.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wang, X., Z. Liu, and S. Peng, 2017: Impact of tidal mixing on water mass transformation and circulation in the South China Sea. J. Phys. Oceanogr., 47, 419432, https://doi.org/10.1175/JPO-D-16-0171.1.

    • Search Google Scholar
    • Export Citation
  • Xiao, J., and Coauthors, 2016: On the near-inertial variations of meridional overturning circulation in the South China Sea. Ocean Sci., 12, 335344, https://doi.org/10.5194/os-12-335-2016.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Xu, Q., and Coauthors, 2022: Dynamics of the baroclinic Rossby waves regulating the abyssal South China Sea. J. Phys. Oceanogr., 52, 873887, https://doi.org/10.1175/JPO-D-21-0207.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Yang, Q., J. Tian, W. Zhao, X. Liang, and L. Zhou, 2014: Observations of turbulence on the shelf and slope of northern South China Sea. Deep-Sea Res. I, 87, 4352, https://doi.org/10.1016/j.dsr.2014.02.006.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Ye, R., C. Zhou, W. Zhao, J. Tian, Q. Yang, X. Huang, Z. Zhang, and X. Zhao, 2019: Variability in the deep overflow through the Heng-Chun Ridge of the Luzon Strait. J. Phys. Oceanogr., 49, 811825, https://doi.org/10.1175/JPO-D-18-0113.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Yuan, D., 2002: A numerical study of the South China Sea deep circulation and its relation to the Luzon Strait transport. Acta Oceanol. Sin., 21, 187202.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Yuan, Y., C. Yang, Y.-H. Tseng, X.-H. Zhu, H. Wang, and H. Chen, 2017: Analysis of longer period variation of the Kuroshio Current intrusion into the Luzon Strait using rectified wavelet power spectra. Prog. Oceanogr., 156, 6177, https://doi.org/10.1016/j.pocean.2017.04.013.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Zhang, Z., J. Tian, B. Qiu, W. Zhao, C. Ping, D. Wu, and X. Wan, 2016: Observed 3D structure, generation, and dissipation of oceanic mesoscale eddies in the South China Sea. Sci. Rep., 6, 24349, https://doi.org/10.1038/srep24349.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Zhao, R., and X.-H. Zhu, 2016: Weakest winter South China Sea western boundary current caused by the 2015–2016 El Niño event. J. Geophys. Res. Oceans, 121, 76737682, https://doi.org/10.1002/2016JC012252.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Zhao, R., and Coauthors, 2022: Summer anticyclonic eddies carrying Kuroshio waters observed by a large CPIES array west of the Luzon Strait. J. Phys. Oceanogr., https://doi.org/10.1175/JPO-D-22-0019.1, in press.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Zheng, H., C. Zhang, R. Zhao, X.-H. Zhu, Z.-N. Zhu, Z.-J. Liu, and M. Wang, 2021a: Structure and variability of abyssal current in northern South China Sea based on CPIES observations. J. Geophys. Res. Oceans, 126, e2020JC016780, https://doi.org/10.1029/2020JC016780.

    • Crossref
    • Export Citation
  • Zheng, H., X.-H. Zhu, C. Zhang, R. Zhao, Z.-N. Zhu, and Z.-J. Liu, 2021b: Propagation of topographic Rossby waves in the deep basin of the South China Sea based on abyssal current observations. J. Phys. Oceanogr., 51, 27832791, https://doi.org/10.1175/JPO-D-21-0051.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Zheng, H., and Coauthors, 2022: Observation of abyssal circulation to the west of the Luzon Strait, South China Sea. J. Phys. Oceanogr., 52, 20912109, https://doi.org/10.1175/JPO-D-21-0284.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Zhou, C., W. Zhao, J. Tian, Q. Yang, and T. Qu, 2014: Variability of the deep-water overflow in the Luzon Strait. J. Phys. Oceanogr., 44, 29722986, https://doi.org/10.1175/JPO-D-14-0113.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Zhou, C., W. Zhao, J. Tian, X. Zhao, Y. Zhu, Q. Yang, and T. Qu, 2017: Deep western boundary current in the South China Sea. Sci. Rep., 7, 9303, https://doi.org/10.1038/s41598-017-09436-2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Zhou, M., G. Wang, W. Liu, and C. Chen, 2020: Variability of the observed deep western boundary current in the South China Sea. J. Phys. Oceanogr., 50, 29532963, https://doi.org/10.1175/JPO-D-20-0013.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Zhu, Y., J. Sun, Y. Wang, S. Li, T. Xu, Z. Wei, and T. Qu, 2019: Overview of the multi-layer circulation in the South China Sea. Prog. Oceanogr., 175, 171182, https://doi.org/10.1016/j.pocean.2019.04.001.

    • Crossref
    • Search Google Scholar
    • Export Citation
Save
  • Alford, M. H., and Coauthors, 2015: The formation and fate of internal waves in the South China Sea. Nature, 521, 6569, https://doi.org/10.1038/nature14399.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Baker-Yeboah, S., D. R. Watts, and D. A. Byrne, 2009: Measurements of sea surface height variability in the eastern South Atlantic from pressure sensor–equipped inverted echo sounders: Baroclinic and barotropic components. J. Atmos. Oceanic Technol., 26, 25932609, https://doi.org/10.1175/2009JTECHO659.1.

    • Search Google Scholar
    • Export Citation
  • Donohue, K. A., D. R. Watts, K. L. Tracey, A. D. Greene, and M. Kennelly, 2010: Mapping circulation in the Kuroshio Extension with an array of current and pressure recording inverted echo sounders. J. Atmos. Oceanic Technol., 27, 507527, https://doi.org/10.1175/2009JTECHO686.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Gan, J., Z. Liu, and C. R. Hui, 2016: A three-layer alternating spinning circulation in the South China Sea. J. Phys. Oceanogr., 46, 23092315, https://doi.org/10.1175/JPO-D-16-0044.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hamilton, P., 2009: Topographic Rossby waves in the Gulf of Mexico. Prog. Oceanogr., 82, 131, https://doi.org/10.1016/j.pocean.2009.04.019.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hamilton, P., A. Bower, H. Furey, R. Leben, and P. Pérez-Brunius, 2019: The loop current: Observations of deep eddies and topographic waves. J. Phys. Oceanogr., 49, 14631483, https://doi.org/10.1175/JPO-D-18-0213.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kennelly, M. A., K. L. Tracey, and D. R. Watts, 2007: Inverted echo sounder data processing manual. University of Rhode Island GSO Tech. Rep. 2007-02, 90 pp., https://digitalcommons.uri.edu/cgi/viewcontent.cgi?article=1001&context=physical_oceanography_techrpts.

  • Ku, A., Y. H. Seung, C. Jeon, Y. Choi, E. Yoshizawa, K. Shimada, K.-H. Cho, and J.-H. Park, 2020: Observation of bottom-trapped topographic Rossby waves on the shelf break of the Chukchi Sea. J. Geophys. Res. Oceans, 125, e2019JC015436, https://doi.org/10.1029/2019JC015436.

    • Crossref
    • Export Citation
  • Lan, J., Y. Wang, F. Cui, and N. Zhang, 2015: Seasonal variation in the South China Sea deep circulation. J. Geophys. Res. Oceans, 120, 16821690, https://doi.org/10.1002/2014JC010413.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Long Y., X.-H. Zhu, X., Guo, F. Ji, and Z. Li, 2021: Variations of the Kuroshio in the Luzon Strait revealed by EOF analysis of repeated XBT data and sea-level anomalies. J. Geophys. Res. Oceans, 126, e2020JC016849, https://doi.org/10.1029/2020JC016849.

    • Crossref
    • Export Citation
  • Meinen, C., E. Fields, R. Pickart, and D. R. Watts, 1993: Ray tracing on topographic Rossby waves. University of Rhode Island GSO Tech. Rep. 1993-05, 49 pp., https://digitalcommons.uri.edu/cgi/viewcontent.cgi?article=1009&context=physical_oceanography_techrpts.

    • Crossref
    • Export Citation
  • Oey, L.-Y., and H.-C. Lee, 2002: Deep eddy energy and topographic Rossby waves in the Gulf of Mexico. J. Phys. Oceanogr., 32, 34993527, https://doi.org/10.1175/1520-0485(2002)032<3499:DEEATR>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Park, J.-H., D. R. Watts, K. A. Donohue, and K. L. Tracey, 2012: Comparisons of sea surface height variability observed by pressure-recording inverted echo sounders and satellite altimetry in the Kuroshio Extension. J. Oceanogr., 68, 401416, https://doi.org/10.1007/s10872-012-0108-x.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Pickart, R. S., 1995: Gulf Stream–generated topographic Rossby waves. J. Phys. Oceanogr., 25, 574586, https://doi.org/10.1175/1520-0485(1995)025<0574:GSTRW>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Quan, Q., and H. Xue, 2019: Influence of abyssal mixing on the multi-layer circulation in the South China Sea. J. Phys. Oceanogr., 49, 30453060, https://doi.org/10.1175/JPO-D-19-0020.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Quan, Q., Z. Cai, G. Jin, and Z. Liu, 2021a: Topographic Rossby waves in the abyssal South China Sea. J. Phys. Oceanogr., 51, 17951812, https://doi.org/10.1175/JPO-D-20-0187.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Quan, Q., Z. Liu, S. Sun, Z. Cai, Y. Yang, G. Jin, Z. Li, and X. S. Liang, 2021b: Influence of the Kuroshio intrusion on deep flow intraseasonal variability in the northern South China Sea. J. Geophys. Res. Oceans, 126, e2021JC017429, https://doi.org/10.1029/2021JC017429.

    • Crossref
    • Export Citation
  • Shu, Y., H. Xue, D. Wang, F. Chai, Q. Xie, J. Yao, and J. Xiao, 2014: Meridional overturning circulation in the South China Sea envisioned from the high-resolution global reanalysis data GLBa0.08. J. Geophys. Res. Oceans, 119, 30123028, https://doi.org/10.1002/2013JC009583.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Shu, Y., and Coauthors, 2016: Persistent and energetic bottom-trapped topographic Rossby waves observed in the southern South China Sea. Sci. Rep., 6, 24338, https://doi.org/10.1038/srep24338.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Shu, Y., J. Chen, S. Li, Q. Wang, J. Yu, and D. Wang, 2018: Field-observation for an anticyclonic mesoscale eddy consisted of twelve gliders and sixty-two expendable probes in the northern South China Sea during summer 2017. Sci. China Earth Sci., 62, 451458, https://doi.org/10.1007/s11430-018-9239-0.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Shu, Y., and Coauthors, 2022: Deep-current intraseasonal variability interpreted as topographic Rossby waves and deep eddies in the Xisha Islands of the South China Sea. J. Phys. Oceanogr., 52, 14151430, https://doi.org/10.1175/JPO-D-21-0147.1.

    • Search Google Scholar
    • Export Citation
  • Tian, J., Q. Yang, and W. Zhao, 2009: Enhanced diapycnal mixing in the South China Sea. J. Phys. Oceanogr., 39, 31913203, https://doi.org/10.1175/2009JPO3899.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wang, D., and Coauthors, 2019: Advances in research of the mid-deep South China Sea circulation (in Chinese). Sci. China Earth Sci., 62, 19922004, https://doi.org/10.1007/s11430-019-9546-3.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wang, G., R. X. Huang, J. Su, and D. Chen, 2012: The effects of thermohaline circulation on wind-driven circulation in the South China Sea. J. Phys. Oceanogr., 42, 22832296, https://doi.org/10.1175/JPO-D-11-0227.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wang, J., and Coauthors, 2021: Observed variability of bottom‐trapped topographic Rossby waves along the slope of the northeastern South China Sea. J. Geophys. Res. Oceans, 126, e2021JC017746, https://doi.org/10.1029/2021JC017746.

    • Crossref
    • Export Citation
  • Wang, Q., and Coauthors, 2019: Energetic topographic Rossby waves in the northern South China Sea. J. Phys. Oceanogr., 49, 26972714, https://doi.org/10.1175/JPO-D-18-0247.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wang, X., Z. Liu, and S. Peng, 2017: Impact of tidal mixing on water mass transformation and circulation in the South China Sea. J. Phys. Oceanogr., 47, 419432, https://doi.org/10.1175/JPO-D-16-0171.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Xiao, J., and Coauthors, 2016: On the near-inertial variations of meridional overturning circulation in the South China Sea. Ocean Sci., 12, 335344, https://doi.org/10.5194/os-12-335-2016.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Xu, Q., and Coauthors, 2022: Dynamics of the baroclinic Rossby waves regulating the abyssal South China Sea. J. Phys. Oceanogr., 52, 873887, https://doi.org/10.1175/JPO-D-21-0207.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Yang, Q., J. Tian, W. Zhao, X. Liang, and L. Zhou, 2014: Observations of turbulence on the shelf and slope of northern South China Sea. Deep-Sea Res. I, 87, 4352, https://doi.org/10.1016/j.dsr.2014.02.006.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Ye, R., C. Zhou, W. Zhao, J. Tian, Q. Yang, X. Huang, Z. Zhang, and X. Zhao, 2019: Variability in the deep overflow through the Heng-Chun Ridge of the Luzon Strait. J. Phys. Oceanogr., 49, 811825, https://doi.org/10.1175/JPO-D-18-0113.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Yuan, D., 2002: A numerical study of the South China Sea deep circulation and its relation to the Luzon Strait transport. Acta Oceanol. Sin., 21, 187202.

    • Search Google Scholar
    • Export Citation
  • Yuan, Y., C. Yang, Y.-H. Tseng, X.-H. Zhu, H. Wang, and H. Chen, 2017: Analysis of longer period variation of the Kuroshio Current intrusion into the Luzon Strait using rectified wavelet power spectra. Prog. Oceanogr., 156, 6177, https://doi.org/10.1016/j.pocean.2017.04.013.

    • Search Google Scholar
    • Export Citation
  • Zhang, Z., J. Tian, B. Qiu, W. Zhao, C. Ping, D. Wu, and X. Wan, 2016: Observed 3D structure, generation, and dissipation of oceanic mesoscale eddies in the South China Sea. Sci. Rep., 6, 24349, https://doi.org/10.1038/srep24349.

    • Search Google Scholar
    • Export Citation
  • Zhao, R., and X.-H. Zhu, 2016: Weakest winter South China Sea western boundary current caused by the 2015–2016 El Niño event. J. Geophys. Res. Oceans, 121, 76737682, https://doi.org/10.1002/2016JC012252.

    • Search Google Scholar
    • Export Citation
  • Zhao, R., and Coauthors, 2022: Summer anticyclonic eddies carrying Kuroshio waters observed by a large CPIES array west of the Luzon Strait. J. Phys. Oceanogr., https://doi.org/10.1175/JPO-D-22-0019.1, in press.

    • Search Google Scholar
    • Export Citation
  • Zheng, H., C. Zhang, R. Zhao, X.-H. Zhu, Z.-N. Zhu, Z.-J. Liu, and M. Wang, 2021a: Structure and variability of abyssal current in northern South China Sea based on CPIES observations. J. Geophys. Res. Oceans, 126, e2020JC016780, https://doi.org/10.1029/2020JC016780.

  • Zheng, H., X.-H. Zhu, C. Zhang, R. Zhao, Z.-N. Zhu, and Z.-J. Liu, 2021b: Propagation of topographic Rossby waves in the deep basin of the South China Sea based on abyssal current observations. J. Phys. Oceanogr., 51, 27832791, https://doi.org/10.1175/JPO-D-21-0051.1.

    • Search Google Scholar
    • Export Citation
  • Zheng, H., and Coauthors, 2022: Observation of abyssal circulation to the west of the Luzon Strait, South China Sea. J. Phys. Oceanogr., 52, 20912109, https://doi.org/10.1175/JPO-D-21-0284.1.

    • Search Google Scholar
    • Export Citation
  • Zhou, C., W. Zhao, J. Tian, Q. Yang, and T. Qu, 2014: Variability of the deep-water overflow in the Luzon Strait. J. Phys. Oceanogr., 44, 29722986, https://doi.org/10.1175/JPO-D-14-0113.1.

    • Search Google Scholar
    • Export Citation
  • Zhou, C., W. Zhao, J. Tian, X. Zhao, Y. Zhu, Q. Yang, and T. Qu, 2017: Deep western boundary current in the South China Sea. Sci. Rep., 7, 9303, https://doi.org/10.1038/s41598-017-09436-2.

    • Search Google Scholar
    • Export Citation
  • Zhou, M., G. Wang, W. Liu, and C. Chen, 2020: Variability of the observed deep western boundary current in the South China Sea. J. Phys. Oceanogr., 50, 29532963, https://doi.org/10.1175/JPO-D-20-0013.1.

    • Search Google Scholar
    • Export Citation
  • Zhu, Y., J. Sun, Y. Wang, S. Li, T. Xu, Z. Wei, and T. Qu, 2019: Overview of the multi-layer circulation in the South China Sea. Prog. Oceanogr., 175, 171182, https://doi.org/10.1016/j.pocean.2019.04.001.

    • Search Google Scholar
    • Export Citation
  • Fig. 1.

    Map of the study region in the South China Sea. Current- and pressure-recording inverted echo sounder (CPIES) sites and mooring sites are indicated by black dots and black triangles, respectively. Numbers below site names indicate depths (m).

  • Fig. 2.

    (a) Empirical relationship between deseasoned τ1200 and steric height (ϕ/g). Gray dots are values from historical hydrocast profiles. Black line is fitted from gray dots. (b) Seasonal model for steric height (ϕ/g). Gray dots in (b) indicate deviation between gray dots and fitted line in (a) arranged by days in a year. The black line is fitted from gray dots. (c) Comparison of sea surface height variability at C21 obtained from CPIES (black) and ADT (gray).

  • Fig. 3.

    Variance-preserving spectra of (a) u and (b) υ, at 983 m (brown), 1999 m (blue), and 2912 m (red) at M01. Dashed lines indicate 95% confidence level of red noise. Wavelet analysis of (c) u and (d) υ at 1999 m at M01. Wavelet analysis of (e) u and (f) υ at 2912 m at M01. Black contours indicate 95% confidence level. Color indicates log2-scaled variance of wavelet transform of normalized value. Green shadows indicate periods of 18–28 days.

  • Fig. 4.

    As in Fig. 3, but for u and υ at 2148 m (blue) and 4151 m (red) at M02.

  • Fig. 5.

    Standard deviation ellipses of the near-21-day variance from moorings. Blue ellipses indicate the variance at 1999 m of M01 and 2148 m of M02. Red ellipses indicate the variance at 2912 m of M01 and 4151 m of M02. Directions of downslope (∇H, black arrows), group velocity (cg, purple arrows), and wavenumber vector (K, green arrows) are indicated. Deviation of group velocity direction and ellipse major axes are listed. Isobaths are indicated by gray lines.

  • Fig. 6.

    Wavelet analysis of near-bottom u and υ from CPIES observations. Black contours indicate 95% confidence level. Color indicates log2-scaled variance of wavelet transform of normalized value. Green shadows indicate periods of 18–28 days. Subgraphs are arranged following spatial distribution of CPIESs.

  • Fig. 7.

    Standard deviation ellipses of the near-21-day variance from CPIES observations. Directions of downslope (∇H, black arrows), group velocity (cg, purple arrows), and wavenumber vectors (K, green arrows) are indicated. Deviation of group velocity direction and ellipse major axes are listed. Isobaths are indicated by gray lines.

  • Fig. 8.

    (a) Squared coherence and (b) phase difference, between near-bottom kinetic energy and upper-layer kinetic energy at M01. The 95% confidence level (horizontal lines) and 18–28-day period (gray shadows) are shown. (c),(d) As in (a) and (b), but at M02. Upper kinetic energy is the average of values above 800 m from ADCPs; near-bottom kinetic energy is from the deepest current meters on both moorings.

  • Fig. 9.

    Time series of 18–28-day bandpass-filtered EKE (a) in upper layer and (b) at 983 m (brown), 1999 m (blue), and 2912 m (red) at M01. (c),(d) As in (a),(b), but for M02 and for different depths. Bandpass-filtered EKE at 983 m is also presented at the bottom of (a). Significant durations of near-21-day TRWs are indicated by orange boxes.

  • Fig. 10.

    Squared coherence between near-bottom currents (u, dashed lines; υ, solid lines) and τ from CPIESs during July 2018–April 2019. The 95% confidence level (horizontal lines) and 18–28-day periods (gray shadows) are indicated. Subgraphs are arranged following spatial distribution of CPIESs.

  • Fig. 11.

    Time series of 18–28-day bandpass-filtered near-bottom EKE (black) and SSH anomaly (blue) from CPIES observations. Near-bottom EKE at M01 and M02 (red) are also shown at C21 and C27, respectively. All time series were normalized, and normalized coefficients are shown. Light cyan and pink shadows indicate winter and summer, respectively. Magenta boxes indicate energetic TRWs related to local upper-layer perturbations, and blue boxes indicate energetic TRWs that have no relation to the local upper layer. Blue dashed lines show upper envelope lines, and horizontal lines indicate mean plus STD values. Subgraphs are arranged following spatial distribution of CPIESs. STDs of 18–28-day SSH anomaly from June 2018 to July 2019 are shown in the bottom-right panel. The red box in the bottom-right panel is the region used in Fig. 12.

  • Fig. 12.

    (a) Empirical relationship between deseasoned τ1200 and the 8°C isotherm. Gray dots are values from historical hydrocast profiles. Black line is fitted from gray dots. (b) Average depth of the 8°C isotherm among C23–C25. (c) Normalized potential vorticity anomaly (NPVA) in upper (red) and lower (blue) layers in red box in the bottom-right panel of Fig. 11. Light cyan and pink shadows indicate winter and summer, respectively. (d) As in (c), but for 18–28-day bandpass-filtered records. (e) Squared coherence and (f) phase difference between NPVA in upper and lower layers during July 2018–April 2019. Horizontal and vertical dashed lines indicate 95% confidence level and 21-day period, respectively. Gray shadows indicate the 18–28-day period.

  • Fig. 13.

    (a) Daily maps of Case 1 in winter 2017. ADT (color) and surface geostrophic currents (arrows) are shown. The 18–28-day bandpass-filtered SSH anomaly (SSHA) is indicated by colored dots. (b)–(d) As in (a), but for Case 2 in summer 2018, Case 3 in winter 2018, and Case 4 in summer 2019, respectively. Blue boxes in (a) and (d) indicate regional Kuroshio path meanders. Blue boxes in (b) indicate the regions of eddies. Blue boxes in (c) indicate the regions at center of LS with significant ADT variances.