1. Introduction
Potential temperature θ and salinity S variation at a fixed density is referred to by oceanographers as spiciness (e.g., Veronis 1972; Munk 1981). Warm and salty water is considered “spicy” and described by high spiciness. Subsurface density compensated θ–S anomalies (i.e., subsurface spiciness anomalies) are believed to be mainly produced in the formation regions of thermocline water masses in response to changes of air–sea fluxes (Bindoff and McDougall 1994). Subsequent to the subduction of water masses, spiciness anomalies propagate as passive tracers with subsurface flows in the ocean. Equatorward migration of subsurface spiciness anomalies conveys midlatitude signals to low latitudes, which is believed to be an important regime for the low-frequency modulations of the tropical climate (Gu and Philander 1997; Zhang et al. 1998).
In the Pacific Ocean, the shallow-layer meridional overturning circulation, involving subduction in subtropics, equatorward subsurface advection, upwelling near the equator, and poleward Ekman flow, is often called the Pacific subtropical cell (STC; McCreary and Lu 1994; Liu 1994). An important feature of the Pacific STC is the asymmetry of pathways between the two hemispheres. Although the South Pacific subtropics is directly connected to the equator through a broad interior communication route, most subtropical water in the North Pacific has to flow to the western boundary before moving equatorward, which is due to the blocking effect of the high potential vorticity barrier (Liu 1994; Lu and McCreary 1995; Johnson and McPhaden 1999) under the intertropical convergence zone (ITCZ) in the tropical North Pacific. Such flow pattern highlights the importance of the Philippine Sea (PS), which is located in the far western North Pacific (Fig. 1).
(a) AP (m2 s−2) relative to 2500 dbar and (b) salinity (psu) between 23.5 and 24.5 σθ from annual climatology of WOA09 (Antonov et al. 2010; Locarnini et al. 2010). The colored map shows the bathymetry (m) based on General Bathymetric Chart of the Oceans (GEBCO; Smith and Sandwell 1997).
Citation: Journal of Physical Oceanography 42, 6; 10.1175/JPO-D-12-06.1
Subsurface signal propagation from midlatitude interior North Pacific to the PS has been reported in previous studies based on historical hydrographic data (Deser et al. 1996; Schneider et al. 1999). Surface-forced decadal thermal anomalies are subducted in the central and eastern North Pacific and move southwestward to the western tropical Pacific in the thermocline along mean potential vorticity contours (Schneider et al. 1999). Using Argo float measurements from 2001 through 2008, Sasaki et al. (2010) described persistent southwestward propagation of density-compensated spiciness anomalies at 25.0–25.5 σθ across the Pacific basin. These anomalies are of pronounced amplitude (~0.15 psu for salinity) at the origin in the northeastern Pacific and reduced to ~0.043 psu when reaching the western Pacific (145°–170°E) because of along-path dispersion. Results from GCMs also confirmed this pathway of propagation and suggested a further advection along the western boundary to the equator (e.g., Pierce et al. 2000; Yeager and Large 2004; Luo et al. 2005). Both observations and model results also indicate that these anomalies have been obscure when reaching the western Pacific and therefore inefficient in coupling midlatitude and the equatorial Pacific Ocean (Schneider et al. 1999; Nonaka and Xie 2000; Giese et al. 2002). Despite these findings, the process of spiciness anomaly evolution near the western boundary remains unclear because of the shortage of observations.
In the western tropical North Pacific, advection of thermocline water from interior Pacific is primarily through the North Equatorial Current (NEC). Figure 1a shows the acceleration potential (AP) field at 23.5–24.5 σθ of World Ocean Atlas 2009 (WOA09). The NEC, which is represented by quasi-zonal AP contours between 9° and 18°N, bifurcates upon approaching the Philippine coast, into the equatorward-flowing Mindanao Current (MC) and the poleward-flowing Kuroshio (Toole et al. 1990; Qiu and Lukas 1996). Hence, thermocline water carried by the NEC can either flow back to subtropics via the Kuroshio or farther to the tropics via the MC. At the depths of the thermocline, the NEC carries mainly the warm/salty North Pacific Tropical Water (NPTW; Tsuchiya 1968), which exists as a subsurface salinity maximum. The intrusion of NPTW high-salinity tongue, with S > 34.9 psu between 23.5 and 24.5 σθ, creates two salinity fronts on its boundaries (Fig. 1b): the southern front is narrower, approximately zonal, and situated at around 12°N in the southern PS and the northern front is much broader and situated in southwest–northeast direction in the northern PS. The two fronts are potentially important for generating spiciness anomalies locally through cross-front advection mechanism (Schneider 2000; Kilpatrick et al. 2011). Using a coupled model, Schneider (2000) showed that spiciness anomalies in the PS are generated by wind-driven anomalous advection across mean spiciness fronts and transported subsequently to the equator via the MC, which can significantly influence sea surface temperature (SST) and trade winds in the equatorial Pacific. This decadal spiciness mode in the tropical Pacific involves no midlatitude influence, providing a possible mechanism for the observed decadal variations in tropical Pacific. However, this hypothesis was based on model simulation. Whether local generation dominates over propagation from midlatitudes in the tropical spiciness variability and whether those anomalies can impose significant effect on equatorial SST require examining with observations.
Because of paucity of subsurface salinity measurements, spiciness variations of thermocline water masses were mainly monitored at repeated hydrographic sections in previous studies (e.g., Kessler 1999; Suga et al. 2000; Lukas 2001). Among them, Suga et al. (2000) described interdecadal salinity changes of NPTW during 1967–95 at 137°E using repeated measurements of the Japan Meteorology Agency (JMA). The authors found that salinity changes of NPTW are mainly related with the variations of the NEC transport and cannot be well explained by atmospheric forcing in the formation region of NPTW. Suga et al. (2000) also noted the difference in the variability of southern and northern parts of NPTW (divided by 15°N), which suggests that the low-frequency subsurface salinity variations in the PS may be spatial dependent. Apparently, analysis at a single section cannot reveal the exact spatial pattern of salinity variability.
In the present study, we have three objectives. The first is to investigate the spatial–temporal characteristics of subsurface spiciness variability in the PS with newly available observations accumulated in the past decade. In the results, pronounced interannual variations with out-of-phase salinity changes in the northern and southern PS are detected. The second objective is to explore the underlying mechanism of the observed interannual salinity variations. This is pursued by seeking connection to current changes and combing satellite altimeter data. Our third objective is to examine the possible impact of these spiciness anomalies on downstream regions, especially equatorial Pacific. With these efforts, we can evaluate the role of the PS in coupling the extratropical North Pacific and the equatorial Pacific Ocean.
2. Data and processing
Accumulation of Argo profiles (Roemmich et al. 2009) over the past decade has provided us not only a tool to monitor variations of subsurface water masses but also a means to explore the dynamics underlying the observed variability. Argo float data have been widely used in recent studies of spiciness anomalies (e.g., Johnson 2006; Yeager and Large 2007; Sasaki et al. 2010). Gridded datasets constructed mainly with Argo float data, such as Grid Point Value of the Monthly Objective Analysis (MOAA GPV) by Hosoda et al. (2008), are preferred material for the investigation of subsurface spiciness variations. The MOAA GPV is a global 1° × 1° grid dataset of monthly temperature and salinity constructed with data from profiling floats of the Argo array, Triangle Trans-Ocean Buoy Network (TRITON) buoy measurements, and conductivity–temperature–depth (CTD) casts of research cruises. For each month, data from various sources are vertically interpolated onto selected standard pressure levels from 10 to 2000 dbar using the Akima spline (Akima 1970) and then gridded into 1° × 1° fields using the two-dimensional optimal interpolation (OI) method. Details of data processing, gridding, and error estimation of MOAA GPV are described comprehensively in Hosoda et al. (2008). The final dataset provides monthly temperature and salinity fields starting from January 2001 at the launching of Argo project to the present. In recent researches, MOAA GPV has been widely used to explore not only global variations (Hosoda et al. 2009) but also physics in dynamically complicated regions, such as the Mindanao Dome (Kashino et al. 2011) and North Pacific Subtropical Countercurrent (Kobashi and Xie 2012) regions.
Over the complete 10-yr span of the MOAA GPV dataset (Fig. 2), there is a steady increase of profile number in our interested region, from <3000 per month in 2001 to >10 000 per month after 2007. Corresponding to the increase in observations is a decrease in mean interpolation errors of salinity S and temperature T, which is estimated in the OI process and provided with the dataset of MOAA GPV (for details of error calculation, see Hosoda et al. 2008). The S and T errors have been <0.03 psu and <0.5°C after 2003 and <0.01 psu and <0.2°C after 2006, respectively. For our analysis, we mainly use the data between January 2003 and December 2010, and the data in the periods of July–December 2002 and January–June 2011 are also used occasionally for low-pass filtering to ensure the 8-yr length of the data. Spiciness variations with a peak-to-peak difference of O(0.1 psu) in salinity could be regarded significant with respect to the 0.01–0.03-psu interpolation error. Salinity of MOAA GPV in pressure coordinates is transformed into 0.1 kg m−3 binned potential density levels through a linear interpolation. In this study, subsurface spiciness is represented by salinity value between 23.5 and 24.5 σθ, which corresponds to the standard pressure levels of 75, 100, 125, 150, and 200 dbar in the MOAA GPV dataset. Subsurface spiciness anomaly in each month is thus defined as the deviation of salinity from the 2003–10 mean in this density range.
Number of profiles in each month (dotted black) and mean interpolation error for salinity S (black thin; psu) and temperature T (gray thin; °C) between 23.5 and 24.5 kg m−3 of MOAA GPV dataset in the region of 120°E–180°, 5°–25°N.
Citation: Journal of Physical Oceanography 42, 6; 10.1175/JPO-D-12-06.1
Throughout the paper, individual data profiles are also used occasionally for a comparison with the MOAA GPV fields. Argo float profiles up to December 2010 are obtained from the Global Ocean Data Assimilation Experiment (GODAE). For our study, we use only those with a quality flat of 1 or 2, which indicates “good” or “probably good” quality. Some of the profiles are of obvious erroneous records because of sensor drifting. They are detected and excluded by hand. Also used in this study are CTD measurements at 137°E section by the JMA. This transect starts from 34°N south of Japan and ends at 3°N near the New Guinea, cutting across the entire mid- and low-latitude northwestern Pacific Ocean. In situ measurements are repeated annually since 1967, semiannually (winter and summer) since 1973, and seasonally since 1994, with stations spaced in 1° latitude bins. There are also other repeated CTD sections in the western North Pacific Ocean: for example, those near 18° and 8°N and 130°E conducted during the 1980s and 1990s under the framework of Tropical Ocean and Global Atmosphere (TOGA) and World Ocean Circulation Experiment (WOCE). Because these sections are not comparable with JMA 137°E section in temporal continuity and length, we use only the JMA 137°E section in this study.
Another important dataset available these years is the high-quality altimeter height product by the Archiving, Validation, and Interpretation of Satellite Oceanographic data (AVISO) project, which merges sea surface height (SSH) measurements of Ocean Topography Experiment (TOPEX)/Poseidon and European Remote Sensing Satellite (ERS) for a better spatial coverage. In this study we use the SSH product of AVISO (Ducet et al. 2000), which is with a 7-day temporal resolution and a 1/4° × 1/4° spatial resolution in the period from October 1992 through April 2011.
To clarify the role of wind forcing in the generation of the oceanic variations in the PS, our analysis in section 4 also involves the monthly wind stress data of the European Centre for Medium-Range Weather Forecasts (ECMWF) reanalysis (Balmaseda et al. 2008), which is of a 1° × 1° spatial resolution and a temporal span from 1959 to 2009. For climate indices, the Southern Oscillation index (SOI) is taken from the Climate Prediction Center (CPC) of the National Oceanic and Atmospheric Administration (NOAA).
3. Interannual spiciness changes
We first examine salinity changes at 137°E section of the JMA. Figure 3a shows the salinity variations between 23.5 and 24.5 σθ from 1994 to 2009, during which CTD measurements are repeated seasonally. To damp seasonal fluctuation, the data are low passed with a 1.25-yr (involving five cruises) running mean before plotting. The main body of NPTW is highlighted by S > 34.925 psu values. The latitudinal displacement of high-salinity water is clearly discernible, with northerly positions in 1994–95, 1998, and 2003–05 and southerly positions in 1996, 2000–01, and 2008–09. Correspondingly, salinity anomaly field shows a seesaw pattern between the northern and southern parts of NPTW (Fig. 3b). Northward (southward) migration of NPTW high-salinity water creates positive salinity anomalies in the northern (southern) PS. The peak-to-peak salinity difference exceeds 0.15 psu in both ends. These results indicate that, during the past 1–2 decades, low-frequency spiciness variability in the thermocline of the PS is dominated by interannual signals that appear to be highly spatial dependent. Then, what is the spatial pattern of subsurface salinity variations in the PS? Analysis at a single section is far behind clarifying it.
(a) Salinity (psu) and (b) salinity anomaly (psu) between 23.5 and 24.5 σθ as functions of time (yr) and latitude along 137°E from the JMA repeated hydrographic surveys during 1994–2010. The 34.925-psu contour in (a) and the 0-psu contour in (b) are highlighted with thick black lines. Black triangles indicate dates of the surveys. Both salinity and its anomaly are low passed by a 1.25-yr running mean (five cruises) before plotting.
Citation: Journal of Physical Oceanography 42, 6; 10.1175/JPO-D-12-06.1
We thus turn to the MOAA GPV dataset for more information. The 2003–10 mean fields are shown in Fig. 4. In this dataset, the salty NPTW mainly exists north of 10°N, with is high-salinity core (S > 34.9 psu) situated between 23.5 and 24.5 σθ (Fig. 4a). In that density range, mean fields of salinity and AP (Fig. 4b) are well consistent with the WOA09 annual climatology (recall Fig. 1b). Here, AP is calculated relative to 2000 dbar, which is deepest pressure level in MOAA GPV. The two salinity fronts produced by westward intrusion of NPTW, which are potentially important for local spiciness anomaly generation, are also well resolved by this dataset. Before describing the salinity variations in this layer, it is also necessary to check its depth distribution (Fig. 4c). The depth of this layer exhibits a zonal ridge (>160 m) between 13° and 17°N and decreases rapidly as approaching both northward and southward, reflecting the strong baroclinicity of the NEC. This layer has been shallower than 140 m north of 20°N because of the surface atmospheric cooling and south of 10°N because of the upwelling feature of the North Pacific tropical gyre.
Mean fields from MOAA GPV: (a) vertical salinity section (psu) superimposed by potential density (black contours; in kg m−3) at 130.5°E; (b) salinity (psu) and (c) depth (m) fields superimposed by AP (m2 s−2) between 23.5 and 24.5 σθ. The isohalines of 34.8, 34.9, 35.0, and 35.1 psu in (b) and the isobaths of 120, 140, and 160 m are highlighted by thick white curves.
Citation: Journal of Physical Oceanography 42, 6; 10.1175/JPO-D-12-06.1
One effective way to capture the prime features of subsurface spiciness variations is to conduct an empirical orthogonal function (EOF) analysis of the salinity field, where the salinity field is low passed with a cutoff period at 100 days to damp intraseasonal and other high-frequency fluctuations. Figure 5 shows the spatial pattern and temporal evolution of the first EOF mode, which accounts for 28.1% of the total salinity variance. This mode is in agreement with the variation identified at 137°E (Fig. 3b), showing out-of-phase variations between the northern and southern PS. In the northern PS, the magnitude of variance increases with latitude. Multiplied by the mode amplitude (Fig. 5b), the peak-to-peak difference of this mode exceeds 0.15 psu northeast of Luzon. In the southern PS, there are two maximum areas: the broader one resides between 128° and 138°E east of Mindanao and the smaller one locates at around 152°E, with both centered at ~10°N and with amplitudes exceeding 0.1 psu. In the south of the PS, another band of variance is between 2° and 6°N, with signals weaker and in opposite sign to that in the southern PS. With regard to temporal evolution, this mode switches from positive phase in 2003 to negative phase in 2009, followed seemingly by a rebound in 2010. The second EOF mode, which explains 20.2% of the total salinity variance, is a seasonal cycle reflecting the renewal of NPTW high-salinity water through advection of NEC from its subtropical origin (not shown). If we damp this annual cycle by lengthening the cutoff period of the low-pass filter to 370 days, the first EOF mode, which is identical to that described above, can explain 42.6% of the total variance, except a slight decrease in mode amplitude (Figs. 5c,d).
The first EOF mode of salinity (psu) between 23.5 and 24.5 σθ, which explains 28.1% of the total variance: (a) the spatial pattern and (b) the corresponding time series. The salinity field is low passed with a 100-day cutoff period before the EOF analysis. (c),(d) As in (a),(b), but for the salinity field low passed with a 370-day cutoff period.
Citation: Journal of Physical Oceanography 42, 6; 10.1175/JPO-D-12-06.1
Evolution of subsurface salinity in the northern and southern PS can be better illustrated in time–pressure plots at typical sites (Fig. 6), where isopycnal salinity anomalies are transformed back to pressure space. Sites are chosen at 20.5°N, 130.5°E for the northern PS and at 10.5°N, 130.5°E for the southern PS, both near the center of variance maxima in EOF mode 1. Superimposed on a weak seasonal cycle, the most striking feature in the layer of 23.5–24.5 σθ is the strongly contrasting salinity anomalies between the periods of 2003–05 and 2008–10. Positive (negative) anomalies with magnitude exceeding 0.1 psu appears during 2003–05 (2008–10) in the northern PS (Fig. 6a). The reverse can also be seen at the southern PS site (Fig. 6b). Except an overall out-of-phase relationship, there are also some detailed differences. For instance, anomalies in the northern PS can gradually penetrate to deeper isopycnals, and the deeper layer (σθ > 25.0 kg m−3) in the southern PS shows weaker but distinct salinity changes that seem independent on those in NPTW layer.
Pressure–time sections of isopycnal salinity anomaly (psu) of MOAA GPV at (a) 20.5°N, 130.5°E and (b) 10.5°N, 130.5°E. Black contours denote the evolution of potential density σθ (kg m−3).
Citation: Journal of Physical Oceanography 42, 6; 10.1175/JPO-D-12-06.1
The accordance of patterns between MOAA GPV and JMA 137°E datasets brings also qualm, because the JMA 137°E profiles are also involved in the construction of MOAA GPV fields. Considering the paucity of Argo data in the initial years (<6000 per month in 2003 and 2004; Fig. 3a), we cannot help suspecting that the pattern revealed in Fig. 5 might be just reflecting signals at 137°E. Examining the robustness of the revealed variability can be achieved simply by a direct comparison of salinity diagrams of Argo profiles between the two periods. This is conducted in two small boxes representing the conditions in the northern and southern PS (Fig. 7): box A (19°–22°N, 123°–127°E) (left panel) and box B (9.5°–11°N, 152°–160°E) (right panel), which are well sampled by Argo profiles during 2003–04 and 2009–10. Salinity difference between the two periods is clearly discernible in both boxes, especially in the subsurface layer. At 24.0 σθ, average salinity during 2009–10 is lower (higher) than that during 2003–04 in box A (B) by at least 0.15 psu. Note that the salinity difference in deeper layers (σθ > 25.0 kg m−3) is in opposite sign to that in subsurface salinity maximum (NPTW) layer. All these features are consistent with patterns in Figs. 5 and 6. This result provides concrete evidence suggesting that the variances revealed by MOAA GPV are an authentic phenomenon.
A θ–S scatterplot from Argo float profiles during 2003–04 (black dots) and 2009–10 (gray dots) in (left) box A (19°–22°N, 123°–127°E) and (right) box B (9.5°–11°N, 152°–160°E), with gray lines denoting potential density σθ (kg m−3). Geographic distributions of these Argo profiles are also inserted in the panels.
Citation: Journal of Physical Oceanography 42, 6; 10.1175/JPO-D-12-06.1
It is also noticeable that the salinity difference is mainly confined in subsurface layer in the northern PS (box A), whereas it is equally large in both surface and subsurface layers in the southern PS (box B). Figure 6b also shows that salinity anomalies are largest in the layer of 22.0–23.0 σθ, which is quite close to the mixed layer. Because of the prevailing isopycnal ascend at low latitudes, the upper thermocline is rather shallow (around 80–150 dbar) near 10° and 20°N (recall Figs. 4c, 6). Salinity variations in the mixed layer, which are controlled by atmospheric fluxes, might penetrate downward to the thermocline through diapycnal mixing, producing the observed anomalies in NPTW. This spice injection effect (Yeager and Large 2004, 2007) will be discussed in section 5. Instead, in the following section, we first examine the role of current variability, which turns out to be a plausible cause of the observed spiciness variations.
4. Anomalous across-front advection
To seek the cause of the observed spiciness changes, it is of interest to examine the relationship between salinity anomalies and other variables. In Fig. 8, we draw the yearly maps in 2003, 2005, 2007, and 2009. Throughout these years, large salinity anomalies and velocity anomalies are both confined west of 160°E, indicating a correspondence between them. In 2003, at the positive phase of the salinity EOF mode 1 (recall Fig. 5), the PS is dominated by an anomalous cyclonic circulation. In the northern half of this circulation, westward/northwestward velocities transport more than normal salty water from high-salinity NPTW tongue across the northern front to the northern PS, producing positive salinity anomalies there. In its southern half, northward/northeastward velocities carry freshwater (<34.8 psu) in the tropical gyre (5°–10°N) across the quasi-zonal southern front and suppress southwestward transport of NPTW to the southern PS, resulting in a decrease of salinity between 7° and 14°N in the southern PS. In 2005, as the anomalous cyclone decays, salinity anomalies get greatly weakened correspondingly. The year 2007 seems at a transition phase. Salinity and velocity anomalies are both weak. Large salinity anomalies can be seen south of 10°N, which are also accompanied by large velocity anomalies. In 2009, an anticyclonic circulation is fully developed in the PS. Eastward/southeastward velocities in its northern flank suppress the spreading of salty NPTW into the northern PS, producing negative salinity anomalies there, whereas southward/southwestward velocities in its southern half bring more NPTW to the southern PS and result in a salinity elevation there. The year-to-year salinity changes can be well explained by concurrent geostrophic velocity variations. Anomalous cross-front geostrophic advection is likely to be the probable cause for the observed interannual salinity changes, supporting the regime proposed by Schneider (2000). However, it should be kept in mind that the salinity and current variations and their relationship are derived mainly from the 8-yr records of MOAA GPV. Does the interannual variability in PS circulation exist in the pre-Argo era? What in turn is the cause of the interannual variations in the PS circulation? Clarifying these two problems demands subsurface velocity records with a longer time span.
Maps of yearly-mean salinity (gray contours; psu), salinity anomaly (color shading; psu), and geostrophic velocity anomaly (vectors; m s−1) of MOAA GPV between 23.5 and 24.5 σθ in 2003, 2005, 2007, and 2009. Note that only velocity anomalies with a magnitude > 0.01 m s−1 and north of 4.5°N are drawn.
Citation: Journal of Physical Oceanography 42, 6; 10.1175/JPO-D-12-06.1
(a) Scatterplot of AP (m2 s−2) relative to 2000 dbar between 23.4 and 24.5 σθ from MOAA GPV vs synchronous SSH (cm) of AVISO altimeter data between 5° and 20°N and between 125°E and 180°. The red curve denotes the least squares fit. (b) AP of MOAA GPV (blue) and that conversed from altimeter SSH Ah (red) averaged in the region 8°–18°N, 125°–160°E. The linear correlation between the two reaches 0.80.
Citation: Journal of Physical Oceanography 42, 6; 10.1175/JPO-D-12-06.1
As has been done to salinity field, we apply an EOF analysis to the derived Ah. Focusing on changes of interannual and longer time scales, Ah time series is low passed with a cutoff length of 370 days ahead. As expected, the dominant EOF mode, which accounts for 51.5% of the total variance, reflects exactly the anomalous cyclonic/anticyclonic circulation in the PS (Fig. 10a). The high (low) Ah center of this cyclone (anticyclone) is situated around 12°–13°N, in accordance with the AP anomaly pattern in MOAA GPV (recall Fig. 8). It means that this interannual variation in PS circulation is prominent not only during the 8-yr span of MOAA GPV but also in the AVISO era ranging from 1993 to 2010. The salinity variances seen in MOAA GPV can be well explained by the negative-to-positive phase transition of Ah during 2003–10. The correlation between its time series and that of the salinity EOF mode 1 in Fig. 5d reaches −0.69. It is especially interesting to conduct a lead–lag correlation analysis. The largest correlation of r = −0.89 occurs when Ah leads the salinity by 12 months, which provides concrete evidence for attributing the salinity variations to anomalous advection. A close look at the time series of EOF mode 1 (Fig. 10b) reveals that this mode is at positive phase (anticyclonic anomalies) during 1995–96, 1999–2001, 2005–06, and 2007–09 and at negative phase (cyclonic anomalies) during 1993–94, 1997–98, and 2002–04, showing great resemblance to ENSO signals. Its linear correlation with the 3-month running mean Southern Oscillation index (SOI) reaches r = 0.75. Past studies have pointed out the close relationship between low-frequency variations of the PS circulation and ENSO events (e.g., Qiu and Lukas 1996; Kim et al. 2004; Kashino et al. 2009; Qiu and Chen 2010). Then how are the variations of subsurface currents and water masses in the far western North Pacific involved in ENSO variability, which signals mainly in the central and eastern equatorial Pacific?
The first (explains 51.5% of the total variance) EOF mode of Ah (m2 s−2), which is low passed with a 370-day cutoff period: (a) the spatial pattern and (b) corresponding temporal evolution (thick solid line; m2 s−2) and its linear trend (thin solid line; m2 s−2). Also shown in (b) are the 3-month running mean SOI (gray bars) and time series of EOF mode 1 of the 370-day low-passed salinity field from MOAA GPV (thick dashed line; psu; as in Fig. 5d).
Citation: Journal of Physical Oceanography 42, 6; 10.1175/JPO-D-12-06.1
To clarify this relationship, we compare the temporal evolutions of Ah and wind stress curl anomaly (WSCA) calculated with ECMWF reanalysis within the latitude band of 8°–18°N (Fig. 11) where the anomalous circulation mainly develops. During strong ENSO events, large WSCA appears between 8° and 18°N in the western and central Pacific (130°E–170°W), which is produced mainly by the prominent changes of zonal wind at the equator [see Fig. 12 in Kim et al. (2004) and Fig. 11 in Qiu and Chen (2010)]. As the oceanic response to the wind stress curl changes, subsurface AP anomalies are evoked in the upper thermocline at these longitudes (Fig. 11b) and propagate westward as Rossby waves. AP signals are further enhanced in the western Pacific because of local Ekman pumping. Combined effects of remote and local wind forcing result in the observed subsurface AP changes in the PS (Figs. 11a,b). For instance, during strong El Niño events of 1997–98 and 2002–04, positive WSCA emerges in the western and central Pacific with magnitudes of 3–5 × 10−8 N m−3. As a response, negative AP anomalies are evoked and transmitted westward. The propagation speed of AP signals, estimated coarsely through dividing the distance from 130° and 170°E at 14°N by 0.5 yr, is ~0.25 m s−1, which agrees well with the observational speed of first-order baroclinic Rossby waves averaged within the 8°–18°N latitude range (Kessler 1990; Chelton et al. 1998). In the PS, AP anomalies can reach negative magnitudes > 1 m2 s−2, characterizing the anomalous cyclonic gyre in the negative phase in Ah EOF mode 1. During strong La Niña events of 1999–2001 and 2007–09, negative WSCA appears in western and central Pacific, resulting in positive AP anomalies > 1 m2 s−2 and thus the anomalous anticyclonic gyre in the PS. It is interesting that, during weak ENSO episodes (e.g., 1995–96 and 2005–06 La Niña periods), relatively weak negative WSCAs appear only in the central Pacific. Though Rossby waves can also be seen propagating westward, these AP signals have been rather weak in the PS because of canceling effect of positive WSCA in the PS. Therefore, low-frequency variability of subsurface circulation in the PS and resultant salinity changes show pronounced quasi-decadal signals, which are primarily oceanic responses to strong and long-lasting El Niños and La Niñas. Similar variations are also evident in the time series of NEC bifurcation (Qiu and Chen 2010).
(a) Time series of Ah EOF mode 1 (m2 s−2; as in Fig. 10b). Time–longitude maps of (b) 370-day cutoff low-passed Ah (m2 s−2) and (c) WSCA (×10−8 N m−3) from ECMWF averaged in the zonal band of 8°–18°N. (d) The 3-month running mean SOI.
Citation: Journal of Physical Oceanography 42, 6; 10.1175/JPO-D-12-06.1
5. Discussion
In the preceding section, we have proposed a plausible mechanism for the observed interannual salinity variations. However, there are some problems demanding further clarification. For example, are other mechanisms also at work in generating these subsurface salinity anomalies? Can the salinity variations in the PS influence downstream region through the transports of western boundary currents? How does spiciness vary on deeper layers or at longer time scales? These points are addressed in this section.
a. The possible role of spice injection
As has been pointed out in section 3, the layer of 23.5–24.5 σθ is rather shallow, especially in boreal winter, in the PS. Although most of the interannual salinity changes can be well explained by current variability, we cannot still exclude the influence of diapycnal mixing processes. Yeager and Large (2004) proposed a mechanism of spiciness anomaly generation on non-outcropped isopycnals, in which warm/salty anomalies are produced under weak winter density stratification and a destabilizing salinity gradient at the base of the mixed layer. Such process requires decreasing salinity with depth and generates solely positive anomalies. Because of heavy rainfall, sea surface salinity in the PS is low (Fig. 12). The increasing salinity with depth does not meet the conditions of Yeager and Large’s (2004) mechanism. However, anomaly injection, either positive or negative, can be achieved by the interannual changes in diapycnal mixing term. Neglecting the temporal changes in eddy diffusion coefficient, the effect of turbulent and double-diffusion mixing is primarily determined by vertical gradient of salinity. Interannual variations in surface layer salinity Ssur, as has been illustrated by Delcroix and Hénin (1991), are large in western tropical Pacific because of rainfall variability in response to the zonal displacement of the ascending branch of Walker and Hadley cells in response to ENSO cycle. In the MOAA GPV dataset, Ssur (calculated between 10 and 20 dbar) shows a year-to-year difference of 0.4 psu during 2003–10 (not shown), which is much larger than that in the subsurface layer [O(0.1 psu)]. For the surface/subsurface salinity difference (Fig. 12), ΔS = Ssur − Ssub, where Ssub is defined as the mean salinity between 100 and 200 dbar. The patterns of ΔS and its anomaly ΔS′ show great resemblance to those of Ssur. It means that the variances of ΔS reflect mainly Ssur changes.
Maps of yearly-mean surface/subsurface salinity difference, ΔS = Ssur − Ssub, from MOAA GPV (thick black contours; psu) and its anomaly ΔS′ (thin contours and shading; psu) in 2003, 2005, 2007, and 2009. Here, Ssur and Ssub are calculated by averaging salinity between 10 and 20 dbar and between 100 and 200 dbar, respectively.
Citation: Journal of Physical Oceanography 42, 6; 10.1175/JPO-D-12-06.1
The effect of ΔS on Ssub may be evident east 150°E in the southern PS, where subsurface salinity anomalies are less related with current changes (recall Fig. 8) and can be explained by ΔS changes. However, if Ssub variations are controlled by diapycnal salinity flux, the spatial patterns of Ssub variances and ΔS′ should be more or less in accordance. In the region east of 150°E where ΔS′ is large in magnitude, subsurface salinity variances are weak. Although the maximum of ΔS variances, with a peak-to-peak difference of 0.4 psu, is centered at 13°–15°N, subsurface variances are near zero at these latitudes (recall Fig. 5). Most importantly, ΔS variations cannot explain the out-of-phase variances in Ssur between the southern and northern PS. For example, ΔS′ is positive at all latitudes near the Philippine coast in 2003, whereas subsurface salinity anomaly is negative in the southern PS in that year (recall Fig. 8). Therefore, though diapycnal mixing may be also at work in generating some of the subsurface salinity anomalies, variations in geostrophic current is the controlling factor in determining subsurface salinity variability.
b. On deeper isopycnals
In Fig. 6b, we have noticed that salinity variations on deeper isopycnals in the PS are independent of those in NPTW layer. Pronounced spiciness variability around 25.0 σθ in the Pacific Ocean has been found in both observations and model results (e.g., Giese et al. 2002; Yeager and Large 2004, 2007; Luo et al. 2005; Sasaki et al. 2010). Using Argo float data, Sasaki et al. (2010) described the propagation of prominent spiciness anomalies at 25.0–25.5 σθ from the eastern North Pacific basin to the western tropical Pacific during 2001–08. For a better understanding the physics of spiciness variability, it is of interest to compare features of anomalies in NPTW layer and deeper layers. Figure 13 provides the yearly fields at 25.0–25.5 σθ in the western North Pacific. Comparing with those in NPTW layer, both salinity and current variances are weaker at 25.0–25.5 σθ. North of 10°N, salinity variations are mainly due to anomaly propagation from the east. The positive and negative anomalies arriving at 170°E in 2004 and 2008, which have been described by Sasaki et al. (2010), are eroded quickly in the western Pacific, resulting in weak salinity variations in the PS.
Yearly-mean salinity (black contours; psu), salinity anomaly (color shading; psu), and geostrophic velocity anomaly (vectors; m s−1) of MOAA GPV between 25.0 and 25.5 σθ during 2003–10. Note that only velocity anomalies with a magnitude > 0.01 m s−1 and north of 4.5°N are drawn.
Citation: Journal of Physical Oceanography 42, 6; 10.1175/JPO-D-12-06.1
The difference of salinity variations between NPTW layer and 25.0–25.5 σθ layer in the PS may be due to the contrasting circulation patterns and water mass distributions in the two layers. Neglecting the changes of mixing terms, isopycnal salinity anomalies are predominantly produced by changes in isopycnal salinity advection term U ⋅ 
c. Equatorward propagation
To a large extent, clarifying the downstream effect of the observed spiciness variability, especially on equatorial SST, is of equal importance to understanding their generation mechanisms. Schneider (2000) suggested that subsurface anomalies in the PS can be subsequently advected to lower latitudes and influence equatorial air–sea interaction. In our results, large salinity anomalies can also be seen south of 7°N in both 23.5–24.5 σθ and 25.0–25.5 σθ layers (recall Figs. 8, 13). These anomalies are also accompanied by large U′, primarily associated with the highly variable NECC. Interannual changes in the strength and latitudinal position of the NECC (Qiu and Joyce 1992) can result in prominent variations in subsurface water mass properties (Kashino et al. 2011). In the western equatorial Pacific, variances of current field seem also to be the dominant factor in deciding the subsurface salinity modulations. Salinity variations in the PS are not likely to have a significant influence in the equatorial changes.
To test this hypothesis, we compare in Fig. 14 the salinity variations in upstream and downstream regions of the MC. As the number of Argo profiles increased rapidly in recent years, the measurements in 2009–10 have been comparable in amount to those in the period 2003–08. In the upstream region of the MC (represented by a 1.5° × 2.5° box covering 10.5°–12°N, 129°–131.5°E), near the variance maximum of the southern PS, a salinity difference of O(0.1 psu) is clearly discernible in the subsurface layer between the periods of 2003–08 and 2009–10. The interannually varying thermocline water in the southern PS can be carried to the equator via the MC. At the southern tip of Mindanao, North Pacific Water first flows into the Celebes Sea before feeding the NECC (Lukas et al. 1991; Fine et al. 1994). Thus, the downstream box is chosen at 3°–4.5°N, 123.5°–126°E in the Celebes Sea (Fig. 14a). Within this box, the mean subsurface salinity is larger during 2009–10 than during 2003–08 by 0.01–0.05 psu (Fig. 14c), much smaller than the 0.10–0.15-psu value in the upstream region (Fig. 14b). The reduction of anomaly magnitude may be due to various reasons. First, the subsurface salty NPTW is freshened from 34.9 to 35.05 psu in the upstream box to ~34.8 psu in the Celebes Sea, indicating a strong effect of diapycnal mixing, which acts to erode the NPTW salinity maximum. Elevated mixing level has been observed in both the PS (Jing et al. 2011) and Indonesian Sea (Ffield and Gordon 1992) and shown to be mainly associated with complicated topography in these regions (Fig. 14a). Second, in the Celebes Sea thermocline, water masses of different origins and ages are brought together (Fine et al. 1994; Kashino et al. 1996). Stirring by eddies and coastal waves, isopycnal mixing is also intensive among these waters with contrasting properties. The third effect comes from variations in current field. The pronounced changes in the PS circulation can modify the water source of the MC by shifting the bifurcation point of the NEC, which can also influence the salinity variations in the Celebes Sea. Because of intensive mixing and prominent variations in current field, spiciness anomalies dissipate rapidly in the MC and Celebes Sea.
(a) Topography (color shading) with blue (red) dots show geographic distributions of Argo profiles during 2003–08 (2009–10) in two 2.5° × 1.5° boxes: 10.5°–12°N, 129°–131.5°E and 3°–4.5°N, 123.5°–126°E. (b),(c) Mean salinity (dotted curves; in psu), standard deviation (error bars), and maximum/minimum value profiles (dashed curves) with respect to potential density σθ (kg m−3) calculated using the Argo profiles in the two boxes are shown.
Citation: Journal of Physical Oceanography 42, 6; 10.1175/JPO-D-12-06.1
If we transform the salinity anomaly through αθ′ + βS′ = 0, where α and β are thermal expansion and haline contraction coefficients, respectively, the 0.01–0.05-psu salinity difference is equivalent to a potential temperature difference of only 0.03°–0.15°C. Remember that the water has to experience further mixing when converging with South Pacific Water brought by the New Guinea Coastal Undercurrent (Lindstrom et al. 1987) before joining the NECC. Anomalies with such magnitude in the Celebes Sea are not likely to impose significant influence on equatorial SST. The western boundary of the tropical North Pacific seems quite efficient in dumping extratropical thermal signals.
d. At longer time scales
Preceding analysis in section 3 relates interannual subsurface salinity variances to current changes. Evolution of Ah revealed in section 3 shows that such variability in PS circulation is prominent throughout the period of 1992–2010. A linear fitting in Fig. 10b reveals that there is also a positive trend of ~0.0076 m2 s−2 yr−1 in the time series of Ah EOF mode 1, indicating an overall elevation of 0.14 m2 s−2 from 1993 to 2010. This difference is comparable to the amplitude of interannual variations. The circulation in the PS is slowly shifting to a more anticyclonic pattern. These results are also consistent with the observed sea level rise and southward migration of the NEC bifurcation point in the PS during this period (Qiu and Chen 2012). According to the connection between subsurface salinity and current variances derived in section 4, we can expect a decreasing (increasing) trend of salinity in the northern (southern) PS. To test this hypothesis, we compare in Fig. 15a the yearly salinity anomaly series at 9°–11°N (representing the southern PS; 
(a) Time series of salinity anomaly (dotted curves; psu) at JMA 137°E section at 23.5–24.5 σθ averaged between 9° and 11°N (black) and between 19° and 21°N (gray). (b) South/north salinity anomaly difference 
Citation: Journal of Physical Oceanography 42, 6; 10.1175/JPO-D-12-06.1
6. Summary
The PS is considered as an important region for the study of subtropical/tropical interaction in the Pacific Ocean. Thermocline water masses subducted in subtropical North Pacific are taken westward into the PS before spreading to the tropics. Low-frequency subsurface spiciness variability is potentially important in coupling the subtropical and tropical oceanic variations. Because of a shortage of observations, previous researches of subsurface water variations are mainly confined at repeated sections (e.g., Qiu and Joyce 1992; Suga et al. 2000) or based on model results (e.g., Schneider 2000; Luo et al. 2005). The specific spatiotemporal features of subsurface spiciness variability are not fully clarified. In this study, we combine existing measurements from Argo profiling floats, satellite altimeters, and research surveys to explore subsurface spiciness variations (represented by salinity changes at 23.5–24.5 σθ) in the PS. Our research is mainly focused on signals at an interannual time scale, which is of the largest amplitude and most interesting spatial pattern. Observed characteristics of the variances are comprehensively described, and the underlying driving mechanism is also explored. The main findings are summarized as follows:
Using repeated CTD measurements of JMA cruises at 137°E, we find that, in the PS, high-salinity water of NPTW shows an evident latitudinal displacement during 1994–2009, which results in out-of-phase interannual salinity variances between the southern and northern PS. It is suggested that low-frequency subsurface spiciness variability in the PS is highly spatially dependent.
Specific spatial and temporal features of salinity changes are then explored with MOAA GPV dataset. Consistent with the patterns at 137°E, low-frequency salinity variability during 2003–10 is dominated by interannual signals. The leading EOF mode of subsurface salinity is an out-of-phase change between the northern and southern PS, which evolves from positive phase in 2003 (high salinity in northern PS and low salinity in southern PS) to negative phase in 2009. The peak-to-peak amplitude of this mode exceeds 0.1 psu in both northern and southern PS. This interannual variability is confirmed by comparing individual Argo profiles in typical zones during the two periods of 2003–04 and 2009–10.
By analyzing yearly maps of salinity and geostrophic velocity anomalies at 23.5–24.5 σθ, the connection between salinity and current variations is revealed. In 2003, a cyclonic anomalous circulation develops near the Philippine coast, transporting more (less) than normal high-salinity NPTW to the northern (southern) PS. Anomalous cross-front salinity advection produces positive (negative) salinity anomalies in the northern (southern) PS. A scenario of opposite sign happens in 2009, when an anomalous anticyclone appears in the PS, which produces negative (positive) salinity anomalies in northern (southern) PS. Therefore, the observed interannual subsurface salinity variances in the PS are mainly caused by variations in the PS circulation.
Through a function relationship deduced by least squares fitting, we conversed altimetric SSH of AVISO from 1992 to 2010 to subsurface AP. The conversed AP (Ah) field shows a pronounced interannual changes in the PS, which is identical to that recorded in MOAA GPV. The leading mode of Ah variations are well correlated with salinity EOF mode 1 in MOAA GPV, with the correlation coefficient reaching −0.89 when Ah variances lead the salinity variances by 1 yr. Further analysis suggests that the changes in PS circulation are closely associated with ENSO index (r = 0.75 with 3-month running mean SOI). During strong El Niño (La Niña) episodes, positive (negative) WSCA appears between 8° and 18°N in the western and central Pacific because of large changes in equatorial zonal winds, evoking westward-propagating upwelling (downwelling) Rossby waves in the central Pacific and anomalous upward (downward) Ekman pumping velocities in the western Pacific, resulting in the anomalous cyclonic (anticyclonic) circulation in the PS. Because weak ENSO events have reduced oceanic expression in the western Pacific, these variations exhibit mainly quasi-decadal signals.
Another possible mechanism is by salinity flux from surface layer through mixing diapycnal processes. By examining yearly fields of surface/subsurface salinity difference ΔS, the role of this spice injection mechanism is proven to be secondary.
With regard to the subtropical-to-tropical spiciness signal propagation, our results show that, because the spiciness anomalies advected from higher latitudes have been obscure in the western Pacific, interannual-to-decadal spiciness variability in the PS is dominated by locally generated signals in response to tropical wind forcing. These signals might be advected to lower latitudes via the transport of the MC. However, because of the efficient damping effect of intensive mixing and current variability off the Philippine coast and in the Celebes Sea, they have been greatly reduced in magnitude and not likely to affect the equatorial SST.
Despite these findings, we cannot yet deny the climatic importance of subsurface spiciness anomalies in the PS. Two potentially important processes are also indicated in our results. The first is in the northern PS. It is noticeable that the layer of 23.5–24.5 σθ has been nearly outcropped north of 22°N. The subsurface temperature/salinity anomalies in the PS may propagate northward via the Kuroshio and mesoscale eddies, imposing a possible effect on the SST of subtropical northwestern Pacific and adjacent seas. Similar poleward propagation of spiciness anomalies has been suggested by Laurian et al. (2006) in the North Atlantic Ocean. Second, though interannual-to-decadal anomalies are not likely to survive the western boundary mixing damping with prominent power, pronounced multidecadal changes observed at 137°E suggest a possible role of subsurface spiciness anomalies in climate changes at longer time scales. These processes might be interesting themes for future researches.
Acknowledgments
This research is supported by the National Basic Research Program of China (Grant 2012CB417401) and the National Natural Science Foundation of China (Project 40890152). The authors are grateful to Bo Qiu, William Kessler, and Yan Du for the useful discussion with them. Insightful comments and suggestions from two anonymous reviewers are helpful in improving our work.
We would like also to thank Dr. Shigeki Hosoda and other researchers for their efforts in constructing, archiving, and freely providing the valuable MOAA GPV dataset (through http://www.jamstec.go.jp/ARGO/J_ARGOe.html). CTD data at JMA 137°E section are available online (at http://www.jma.go.jp/jma/indexe.html). Argo profiles are provided by GODAE (through http://www.godae.org/). Altimetric SSH data are downloaded from AVISO online (at http://www.aviso.oceanobs.com/).
REFERENCES
Akima, H., 1970: A new method for interpolation and smooth curve fitting based on local procedures. J. Assoc. Comput. Mech., 17, 589–602.
Antonov, J. I., and Coauthors, 2010: Salinity. Vol. 2, World Ocean Atlas 2009, NOAA Atlas NESDIS 69, 184 pp.
Balmaseda, M. A., A. Vidard, and D. L. T. Anderson, 2008: The ECMWF ocean analysis system: ORA-S3. Mon. Wea. Rev., 136, 3018–3034.
Bindoff, N. L., and T. J. McDougall, 1994: Diagnosing climate change and ocean ventilation using hydrographic data. J. Phys. Oceanogr., 24, 1137–1152.
Chelton, D. B., R. A. de Szoeke, M. G. Schlax, K. E. Naggar, and N. Siwertz, 1998: Geographical variability of the first baroclinic Rossby radius of deformation. J. Phys. Oceanogr., 28, 433–460.
Delcroix, T., and C. Hénin, 1991: Seasonal and interannual variations of sea surface salinity in the tropical Pacific Ocean. J. Geophys. Res., 96, 22 135–22 150.
Deser, C., M. A. Alexander, and M. S. Timlin, 1996: Upper-ocean thermal variations in the North Pacific during 1970–1991. J. Climate, 9, 1840–1855.
Ducet, N., P. Y. Le Traon, and G. Reverdin, 2000: Global high-resolution mapping of ocean circulation from TOPEX/Poseidon and ERS-1 and -2. J. Geophys. Res., 105, 19 477–19 498.
Ffield, A., and A. L. Gordon, 1992: Vertical mixing in the Indonesian thermocline. J. Phys. Oceanogr., 22, 184–195.
Fine, R. A., R. Lukas, F. M. Bingham, M. J. Warner, and R. H. Gammon, 1994: The western equatorial Pacific: A water mass crossroads. J. Geophys. Res., 99 (C12), 25 063–25 080.
Giese, B. S., S. C. Urizar, and N. S. Fučkar, 2002: Southern Hemisphere origins of the 1976 climate shift. Geophys. Res. Lett., 29, 1014, doi:10.1029/2001GL013268.
Gu, D., and S. G. H. Philander, 1997: Interdecadal climate fluctuations that depend on exchange between the tropics and extratropics. Science, 275, 805–807.
Hosoda, S., T. Ohira, and T. Nakamura, 2008: A monthly mean dataset of global oceanic temperature and salinity derived from Argo float observations. JAMSTEC Report of Research and Development, Vol. 8, Japan Agency for Marine-Earth Science and Technology, Yokohama, Japan, 47–59.
Hosoda, S., T. Suga, N. Shikama, and K. Mizuno, 2009: Global surface layer salinity change detected by Argo and its implication for hydrological cycle intensification. J. Oceanogr., 65, 579–586.
Jing, Z., L. Wu, L. Li, C. Liu, X. Liang, Z. Chen, D. Hu, and Q. Liu, 2011: Turbulent diapycnal mixing in the subtropical northwestern Pacific: Spatial-seasonal variations and role of eddies. J. Geophys. Res., 116, C10028, doi:10.1029/2011JC007142.
Johnson, G. C., 2006: Generation and initial evolution of a mode water θ–S anomaly. J. Phys. Oceanogr., 36, 739–751.
Johnson, G. C., and M. J. McPhaden, 1999: Interior pycnocline flow from the subtropical to the equatorial Pacific Ocean. J. Phys. Oceanogr., 29, 3073–3089.
Kashino, Y., M. Aoyama, T. Kawano, N. Hendiarti, Syaefudi, Y. Anantasena, K. Muneyama, and H. Watanabe, 1996: The water masses between Mindanao and New Guinea. J. Geophys. Res., 101 (C5), 12 391–12 400.
Kashino, Y., N. Espana, F. Syamsudin, K. J. Richards, T. Jensen, P. Dutrieux, and A. Ishida, 2009: Observations of the North Equatorial Current, Mindanao Current, and the Kuroshio Current system during the 2006/07 El Niño and 2007/08 La Niña. J. Oceanogr., 65, 325–333.
Kashino, Y., A. Ishida, and S. Hosoda, 2011: Observed ocean variability in the Mindanao Dome region. J. Phys. Oceanogr., 41, 287–302.
Kessler, W. S., 1990: Observation of long Rossby waves in the northern tropical Pacific. J. Geophys. Res., 95, 5183–5217.
Kessler, W. S., 1999: Interannual variability of the subsurface high salinity tongue south of the equator at 165°E. J. Phys. Oceanogr., 29, 2038–2049.
Kilpatrick, T., N. Schneider, and E. Di Lorenzo, 2011: Generation of low-frequency spiciness variability in the thermocline. J. Phys. Oceanogr., 41, 365–377.
Kim, Y. Y., T. Qu, T. Jensen, T. Miyama, H. Mitsudera, H.-W. Kang, and A. Ishida, 2004: Seasonal and interannual variations of the North Equatorial Current bifurcation in a high-resolution OGCM. J. Geophys. Res., 109, C03040, doi:10.1029/2003JC002013.
Kobashi, F., and S. P. Xie, 2012: Interannual variability of the North Pacific Subtropical Countercurrent: Role of local ocean–atmosphere interaction. J. Oceanogr., 68, 113–126.
Laurian, A., A. Lazar, G. Reverdin, K. Rodgers, and P. Terray, 2006: Poleward propagation of spiciness anomalies in the North Atlantic Ocean. Geophys. Res. Lett., 33, L13603, doi:10.1029/2006GL026155.
Lindstrom, E., R. Lukas, R. Fine, E. Firing, S. Godfrey, G. Meyers, and M. Tsuchiya, 1987: The Western Equatorial Pacific Ocean Circulation Study. Nature, 330, 533–537.
Liu, Z., 1994: A simple model of the mass exchange between the subtropical and tropical ocean. J. Phys. Oceanogr., 24, 1153–1165.
Locarnini, R. A., A. V. Mishonov, J. I. Antonov, T. P. Boyer, H. E. Garcia, O. K. Baranova, M. M. Zweng, and D. R. Johnson, 2010: Temperature. Vol. 1, World Ocean Atlas 2009, NOAA Atlas NESDIS 68, 184 pp.
Lu, P., and J. McCreary, 1995: Influence of the ICTZ on the flow of thermocline water from the subtropical to the equatorial Pacific Ocean. J. Phys. Oceanogr., 25, 3076–3088.
Lukas, R., 2001: Freshening of the upper thermocline in the North Pacific subtropical gyre associated with decadal changes of rainfall. Geophys. Res. Lett., 28, 3485–3488.
Lukas, R., E. Firing, P. Hacker, P. L. Richardson, C. A. Collins, R. Fine, and R. Gammon, 1991: Observations of the Mindanao Current during the Western Equatorial Pacific Ocean Circulation Study. J. Geophys. Res., 96 (C4), 7089–7104.
Luo, Y., L. M. Rothstein, R.-H. Zhang, and A. J. Busalacchi, 2005: On the connection between South Pacific subtropical spiciness anomalies and decadal equatorial variability in an ocean general circulation model. J. Geophys. Res., 110, C10002, doi:10.1029/2004JC002655.
McCreary, J. P., and P. Lu, 1994: On the interaction between the subtropical and the equatorial oceans: The subtropical cell. J. Phys. Oceanogr., 24, 466–497.
McPhaden, M. J., and D. Zhang, 2002: Slowdown of the meridional overturning circulation in the upper Pacific Ocean. Nature, 415, 603–608.
McPhaden, M. J., and D. Zhang, 2004: Pacific Ocean circulation rebounds. Geophys. Res. Lett., 31, L18301, doi:10.1029/2004GL020727.
Munk, W., 1981: Internal waves and small-scale processes. Evolution of Physical Oceanography, B. A. Warren and C. Wunsch, Eds., MIT Press, 264–291.
Nonaka, M., and S. P. Xie, 2000: Propagation of North Pacific interdecadal subsurface temperature anomalies in an ocean GCM. Geophys. Res. Lett., 27, 3747–3750.
Pierce, D. W., T. P. Barnett, and M. Latif, 2000: Connections between the Pacific Ocean tropics and midlatitudes on decadal timescales. J. Climate, 13, 1173–1194.
Qiu, B., and T. M. Joyce, 1992: Interannual variability in the mid- and low-latitude western North Pacific. J. Phys. Oceanogr., 22, 1062–1079.
Qiu, B., and R. Lukas, 1996: Seasonal and interannual variability of the North Equatorial Current, the Mindanao Current and the Kuroshio along the Pacific western boundary. J. Geophys. Res., 101, 12 315–12 330.
Qiu, B., and S. Chen, 2010: Interannual-to-decadal variability in the bifurcation of the North Equatorial Current off the Philippines. J. Phys. Oceanogr., 40, 2525–2538.
Qiu, B., and S. Chen, 2011: Effect of decadal Kuroshio Extension jet and eddy variability on the modification of North Pacific Intermediate Water. J. Phys. Oceanogr., 41, 503–515.
Qiu, B., and S. Chen, 2012: Multidecadal sea level and gyre circulation variability in the northwestern tropical Pacific Ocean. J. Phys. Oceanogr., 42, 193–206.
Roemmich, D. H., and Coauthors, 2009: The Argo program: Observing the global ocean with profiling floats. Oceanography, 22, 34–43.
Sasaki, Y. N., N. Schneider, N. Maximenko, and K. Lebedev, 2010: Observational evidence for propagation of decadal spiciness anomalies in the North Pacific. Geophys. Res. Lett., 37, L07708, doi:10.1029/2010GL042716.
Schneider, N., 2000: Decadal spiciness mode in the tropics. Geophys. Res. Lett., 27, 257–260.
Schneider, N., A. J. Miller, M. A. Alexander, and C. Deser, 1999: Subduction of decadal North Pacific temperature anomalies: Observations and dynamics. J. Phys. Oceanogr., 29, 1056–1070.
Smith, W. H. F., and D. T. Sandwell, 1997: Global sea floor topography from satellite altimetry and ship depth soundings. Science, 277, 1956–1962.
Suga, T., A. Kato, and K. Hanawa, 2000: North Pacific Tropical Water: Its climatology and temporal changes associated with the climate regime shift in the 1970s. Prog. Oceanogr., 47, 223–256.
Toole, J., R. Millard, Z. Wang, and S. Pu, 1990: Observations of the Pacific North Equatorial Current bifurcation at the Philippine coast. J. Phys. Oceanogr., 20, 307–318.
Tsuchiya, M., 1968: Upper Waters of the Intertropical Pacific Oceans. Vol. 4, Johns Hopkins Oceanographic Studies, Johns Hopkins Press, 50 pp.
Veronis, G., 1972: On properties of seawater defined by temperature, salinity and pressure. J. Mar. Res., 30, 227–255.
Wong, A. P. S., N. L. Bindoff, and J. A. Church, 1999: Large-scale freshening of intermediate waters in the Pacific and Indian Oceans. Nature, 400, 440–443.
Yeager, S. G., and W. G. Large, 2004: Late-winter generation of spiciness on subducted isopycnals. J. Phys. Oceanogr., 34, 1528–1547.
Yeager, S. G., and W. G. Large, 2007: Observational evidence of winter spice injection. J. Phys. Oceanogr., 37, 2895–2919.
Zhang, R. H., L. M. Rothstein, and A. J. Busalacchi, 1998: Origin of upper-ocean warming and El Niño change on decadal scales in the tropical Pacific Ocean. Nature, 391, 879–883.
