1. Introduction
In winter, the deepest mixed layer (ML) in the North Pacific subtropical gyre is distributed in a region south of the Kuroshio Extension (KE), with a depth of a few hundred meters (Fig. 1: e.g., Ohno et al. 2009). The deep ML south of the KE forms the North Pacific Subtropical Mode Water (STMW) characterized by a thick thermostad layer (Masuzawa 1969; Hanawa 1987), controlling the ventilation process (Oka and Qiu 2012; Oka et al. 2015). The mixed layer temperature (MLT), which is equivalent to sea surface temperature (SST), affects the lower atmosphere around the KE in winter (Taguchi et al. 2009; Tokinaga et al. 2009; Kwon et al. 2010; Tanimoto et al. 2011).
Climatological map of the mixed layer depth (MLD) (m) in March (color), calculated using temperature profiles from Argo floats of Oka et al. (2007): the MLD is defined as the depth at which the temperature difference between 10 m and the bottom of mixed layer (ML) is less than 1°C; this is averaged for each 1° × 1° grid box using profiles from a 3° × 3° grid box centered on the 1° × 1° grid box with a weighting function d−2 [d is the distance in degrees from the center of the 1° × 1° grid box], for observation points where d is >1°, using the methodology of Oka et al. (2012). Contours represent the sea surface height (SSH), reconstructed from satellite-derived altimetry data and mean dynamic topography data from AVISO (http://www.aviso.altimetry.fr), with an interval of 10 cm. The thick black contour represents the SSH = 90 cm contour, regarded as the Kuroshio Extension axis (see text for details). The red rectangle represents the region south of the KE (30°–37°N, 141°–155°E). The white circles show repeat hydrographic stations along 137°E conducted by the Japan Meteorological Agency.
Citation: Journal of Climate 29, 3; 10.1175/JCLI-D-15-0206.1
Climatological map of the mixed layer depth (MLD) (m) in March (color), calculated using temperature profiles from Argo floats of Oka et al. (2007): the MLD is defined as the depth at which the temperature difference between 10 m and the bottom of mixed layer (ML) is less than 1°C; this is averaged for each 1° × 1° grid box using profiles from a 3° × 3° grid box centered on the 1° × 1° grid box with a weighting function d−2 [d is the distance in degrees from the center of the 1° × 1° grid box], for observation points where d is >1°, using the methodology of Oka et al. (2012). Contours represent the sea surface height (SSH), reconstructed from satellite-derived altimetry data and mean dynamic topography data from AVISO (http://www.aviso.altimetry.fr), with an interval of 10 cm. The thick black contour represents the SSH = 90 cm contour, regarded as the Kuroshio Extension axis (see text for details). The red rectangle represents the region south of the KE (30°–37°N, 141°–155°E). The white circles show repeat hydrographic stations along 137°E conducted by the Japan Meteorological Agency.
Citation: Journal of Climate 29, 3; 10.1175/JCLI-D-15-0206.1
Climatological map of the mixed layer depth (MLD) (m) in March (color), calculated using temperature profiles from Argo floats of Oka et al. (2007): the MLD is defined as the depth at which the temperature difference between 10 m and the bottom of mixed layer (ML) is less than 1°C; this is averaged for each 1° × 1° grid box using profiles from a 3° × 3° grid box centered on the 1° × 1° grid box with a weighting function d−2 [d is the distance in degrees from the center of the 1° × 1° grid box], for observation points where d is >1°, using the methodology of Oka et al. (2012). Contours represent the sea surface height (SSH), reconstructed from satellite-derived altimetry data and mean dynamic topography data from AVISO (http://www.aviso.altimetry.fr), with an interval of 10 cm. The thick black contour represents the SSH = 90 cm contour, regarded as the Kuroshio Extension axis (see text for details). The red rectangle represents the region south of the KE (30°–37°N, 141°–155°E). The white circles show repeat hydrographic stations along 137°E conducted by the Japan Meteorological Agency.
Citation: Journal of Climate 29, 3; 10.1175/JCLI-D-15-0206.1
Three major factors controlling variations in wintertime mixed layer depth (MLD) south of the KE have been suggested in previous studies. The first factor is the large wintertime oceanic buoyancy loss driven by the East Asian winter monsoon (Suga and Hanawa 1990; Bingham 1992; Suga and Hanawa 1995; Yasuda and Hanawa 1997). The second factor is the seasonal thermocline intensity resulting from surface heating during the preceding warm season (e.g., Kako and Kubota 2007; Tomita et al. 2010); hence, if the seasonal thermocline is weak (strong) during the warm season, a deeper (shallower) ML develops during the following winter. The third factor is the subsurface stratification intensity from the seasonal thermocline to the main thermocline during the preceding warm season (e.g., Qiu and Chen 2006; Iwamaru et al. 2010), as strong subsurface stratification is an unfavorable condition for winter ML development. Subsurface stratification is intensified by a shallowing of the main thermocline, which propagates from the central North Pacific as a baroclinic Rossby wave (Qiu et al. 2007; Sugimoto and Hanawa 2010), and by southward transport of high potential vorticity water from the Kuroshio–Oyashio confluence region across the KE due to strong eddy activity (Qiu and Chen 2006; Qiu et al. 2007; Kouketsu et al. 2012; Oka et al. 2012, 2014) associated with the shallowing of the main thermocline and the resultant convoluted KE path (i.e., an unstable state of the KE path) (Qiu and Chen 2005, 2010). Iwamaru et al. (2010) implied that the cause of variation in the winter MLD changed around 1990: wintertime buoyancy loss influenced the winter MLD before this date, whereas subsurface stratification intensity has been a dominant factor since this time and the influence of buoyancy loss on the winter MLD was comparable to that of surface stratification intensity during the previous warm season.
The winter MLT south of the KE ranges approximately between 16° and 19°C on interannual and longer time scales (e.g., Hanawa and Kamada 2001). On interannual time scales, negative anomalies of MLT tend to form in years with a stronger winter monsoon due to large amounts of heat released from the ocean to the atmosphere and increased southward Ekman transport (Suga and Hanawa 1995; Yasuda and Hanawa 1997, 1999; Taneda et al. 2000; Hanawa and Kamada 2001; Hanawa and Yoritaka 2001). On longer time scales, the gyre spinup associated with the intensification of the westerlies induces an increase in warm water advection by the Kuroshio (Deser et al. 1999; Yasuda and Kitamura 2003; Sugimoto et al. 2010), resulting in positive temperature anomalies in the winter ML of the nearby KE (Qiu 2000; Vivier et al. 2002; Yasuda and Kitamura 2003). In addition to these mechanisms, the decadal variations in the strength of subsurface stratification induced by the oceanic Rossby wave generated by the meridional movement of the Aleutian low (AL) in the central North Pacific (Ceballos et al. 2009; Sugimoto and Hanawa 2010; Seo et al. 2014) and by changes in path state of KE (Qiu and Chen 2005, 2010; Seo et al. 2014) would strongly influence winter MLD variations on low-frequency time scales, which would in turn affect the formation of temperature anomalies in the winter ML through vertical entrainment.
A long-term time series is needed to detect a cause of wintertime MLD variation and evaluate its influence on the MLT. Iwamaru et al. (2010) produced monthly time series of MLD south of the KE for a long period of 1971–2007 using temperature profiles: the MLDs were first averaged in each 1° (latitude) × 1° (longitude) subregion in each year/month, and then they were averaged over the region south of the KE. The time series showed the deep winter MLD in the mid-1980s and mid-2000s, the period of which was estimated roughly as an interdecadal time scale (~20 yr). But, the value in this method might be affected by observation points; if the observations are performed in shallow MLD regions, the shallow MLD as a monthly mean value is calculated inevitably. The removal of the influence of observation points is required for accurate estimates of monthly mean MLD. In this study, we apply a new approach to remove the influence of observation points properly and produce even more robust time series of MLD south of the KE (30°–37°N, 141°–155°E; blue rectangle in Fig. 1) based on the historical temperature profiles during a long period of 1968–2014. Then we reveal the process that determines the wintertime MLD, using both observational data and simulation outputs of a one-dimensional turbulence closure model. Additionally, we quantitatively assess the contribution of MLD variation (i.e., vertical entrainment) on the formation of winter MLT anomalies by performing a ML heat budget analysis. The remainder of this paper is organized as follows. Section 2 outlines the dataset and processing procedures, and section 3 investigates the cause of long-term wintertime MLD variations south of the KE, using both observational data and simulation outputs. Section 4 presents a quantitative assessment of the role of vertical entrainment in the formation of temperature anomalies in the winter ML, performing a ML heat budget analysis based on observational data. Finally, a summary and the conclusions are given in section 5.
2. Data, processing procedures, and definitions
We use temperature profiles archived in the World Ocean Database 2013 (WOD13; Boyer et al. 2013) and at the Japan Oceanographic Data Center (JODC; http://www.jodc.go.jp) and profiles from Argo floats (Oka et al. 2007), from January 1968 to December 2014. To control data quality, we first removed profiles duplicated in the different data sources. For each profile, the measured data were then compared with all values measured in the same month within a 1° (latitude) × 1° (longitude) grid box centered on the observation point, and data were excluded if they fell outside of three standard deviations of the mean. Profiles with large temperature inversions (dT/dz < −0.1°C m−1) anywhere were also removed. After quality control procedures, temperature profiles were vertically interpolated into 1-m intervals following Akima (1970). Because we focus on the region south of the KE, we use only profiles with T > 16°C at 200 m, because T = 15°C at 200 m is regarded as a good indicator of the KE axis (Kawai 1972): this is to done to minimize the risk of identifying subarctic profiles coming from north of the KE.
The MLD is defined as the depth at which the temperature difference between 10 m and the bottom of the ML is less than 1°C (Qiu 2000). We checked that the results were insensitive to the threshold value of the MLD, revealing that almost identical results were obtained using other threshold values such as 0.5°C. Figure 2 displays the number of profiles in which the MLD was detected in the region south of the KE, revealing that more than 30 profiles are available for most months.
Number of profiles in which the MLD was detected, per month, in the region south of the KE (30°–37°N, 141°–155°E). Note that profiles north of the KE are excluded (see text for details). The white circles represent months when profiles were <5.
Citation: Journal of Climate 29, 3; 10.1175/JCLI-D-15-0206.1
Number of profiles in which the MLD was detected, per month, in the region south of the KE (30°–37°N, 141°–155°E). Note that profiles north of the KE are excluded (see text for details). The white circles represent months when profiles were <5.
Citation: Journal of Climate 29, 3; 10.1175/JCLI-D-15-0206.1
Number of profiles in which the MLD was detected, per month, in the region south of the KE (30°–37°N, 141°–155°E). Note that profiles north of the KE are excluded (see text for details). The white circles represent months when profiles were <5.
Citation: Journal of Climate 29, 3; 10.1175/JCLI-D-15-0206.1
Based on temperature profiles, the intensity of the seasonal thermocline is defined as the local vertical maximum of the vertical temperature gradient between the sea surface and 200-m depth. Here, 200-m depth is selected to adequately cover the seasonal thermocline depth. To calculate the intensity of the subsurface stratification, we prepared two indicators: 1) the magnitude of the vertical temperature gradient between 100- and 300-m depth (Iwamaru et al. 2010) and 2) the MTD taken as the depth of the T = 12°C isotherm (Uehara et al. 2003), which is located in the middle of the main thermocline in the western part of the North Pacific subtropical gyre. In the following sections, we show only the results based on the MTD because almost identical results were obtained using the magnitude of the vertical temperature gradient.
We use three SST datasets: 1) the monthly National Oceanic and Atmospheric Administration (NOAA) Extended Reconstructed SST, version 3b, dataset (ERSST.v3b; Smith et al. 2008) with a spatial resolution of 2° (latitude) × 2° (longitude), the product of which is generated by using in situ data from the International Comprehensive Ocean–Atmosphere Dataset (ICOADS) and is available monthly throughout the analysis period; 2) Monthly Merged Satellite and In Situ Data from the Global Daily Sea Surface Temperatures (MGDSST; Kurihara et al. 2006) dataset, analyzed by the Japan Meteorological Agency (JMA) on 1/4° (latitude) × 1/4° (longitude) spatial resolution, the product of which is available since 1985; and 3) daily high-resolution SST dataset blended with the Moderate Resolution Imaging Spectroradiometer (MODIS) infrared data and the Advanced Microwave Scanning Radiometer for Earth Observing System (AMSR-E) and Windsat microwave data, with a spatial resolution of 1/10° (latitude) × 1/10° (longitude) (Hosoda and Sakaida 2015). Diurnal SST warming was removed for adjusting measured SSTs to daily minimum SSTs by an empirical method (Hosoda 2013) based on satellite remote sensing data of solar radiation and sea surface wind, and an optimal interpolation (OI) technique was applied for blending infrared and microwave SSTs, using decorrelation scales derived by Hosoda (2015).
We use the sea surface height (SSH) dataset reconstructed by adding the satellite-derived SSH anomalies complied by the Archiving, Validation, and Interpretation of Satellite Oceanographic Data (AVISO) to the mean topography (http://www.aviso.altimetry.fr). The SSH data are available from January 1993 onward, with daily temporal resolution and 1/4° (latitude) × 1/4° (longitude) spatial resolution. In this study, we define the KE axis as the 90-cm SSH contour (thick line in Fig. 1) because this is consistently located at or near the position of the maximum north–south gradient of SSH. We also use the AVISO sea surface geostrophic velocity dataset with daily temporal resolution and 1/4° (latitude) × 1/4° (longitude) spatial resolution (http://www.aviso.altimetry.fr).
We use summertime repeat hydrographic observations along 137°E (white circles in Fig. 1), conducted by the JMA. Summer cruise data along 137°E are available since 1972, except for 2009 when adverse weather due to a typhoon prevented data collection. Data at each observation point are vertically interpolated into 1-m intervals following Akima (1970), and are horizontally interpolated into intervals every 0.5° latitude using the objective analysis scheme proposed by Levitus (1982) and Qiu and Joyce (1992). In this study, the STMW is defined as a layer with a vertical temperature gradient of less than 1.5°C (100 m)−1 (Hanawa and Suga 1995), with temperatures of 16°–19°C (Oka 2009; Sugimoto and Hanawa 2014) and thickness >50 m. A local vertical minimum in the vertical temperature gradient is labeled as the core of STMW.
Monthly atmospheric variables of net surface heat flux (NHF; i.e., the sum of net surface longwave radiation, net surface shortwave radiation, latent heat flux, and sensible heat flux) and wind stress are from the Japanese 55-year Reanalysis Project (JRA-55; Kobayashi et al. 2015) with a spatial resolution of 1.25° (latitude) × 1.25° (longitude). The wind stress curl (WSC) is calculated as the spatial derivative of the wind stress. As the derivative operation tends to exaggerate small-scale features of the WSC filed compared with the atmospheric pressure field, we highlight the large-scale features, by smoothing the WSC field using a Gaussian filter with an e-folding scale of 300 km. Sugimoto and Hanawa (2009) showed that the meridional movement of the AL is linked to the West Pacific (WP) teleconnection pattern, identified as the second mode of an empirical orthogonal function (EOF) analysis of the WSC field over the North Pacific (e.g., Ishi and Hanawa 2005; Ceballos et al. 2009); the first mode represents a change in magnitude of the AL associated with the Pacific–North American (PNA) teleconnection pattern. We use the WP index from the NOAA/Climate Prediction Center (NOAA/CPC; http://www.cpc.ncep.noaa.gov), as an indicator of the AL meridional movement, given as rotated empirical orthogonal functions (REOFs) of monthly 700-hPa geopotential height anomalies, obtained using the methodology of Barnston and Livezey (1987).
We use a one-dimensional turbulent closure model (Noh and Kim 1999; Kako and Kubota 2007) to elucidate the cause of MLD variations. This model is run with a vertical resolution of 2 m from the sea surface to 700-m depth and a time step of 600 s. The 6-hourly data (NHF, wind stress, and freshwater flux) from JRA-55 are used as the surface forcing for the model, and the initial ocean conditions are set using the above temperature profiles and salinity profile from the World Ocean Atlas 2013 (WOA13; Zweng et al. 2013).
We use a 10% significance level for all correlation coefficients, based on Student’s two-sided t test in which the degrees of freedom are estimated by dividing the data length by the integral time scale, calculated following Davis (1976).
3. Decadal variations in MLD
We produce the monthly mean time series of MLD south of the KE, by applying a below procedure to remove the influence of observation points properly. First we calculate the monthly MLD climatology averaged on a 1° (latitude) × 1° (longitude) grid using MLDs from 1968 to 2014. Next we calculate the MLD “anomaly” for each year and month, subtracting the monthly MLD climatology at the nearest grid point from the MLD of each profile. Finally, we obtain the monthly MLD time series by adding the monthly MLD climatology averaged for the region south of the KE to the MLD anomalies averaged in each year and month.
We especially focus on March, which is the month with the deeper ML (Takeuchi and Yasuda 2003; Ohno et al. 2009). Figure 3a (black line) displays the MLD in March, which has low-frequency variations. Interestingly, the time series clearly shows another deep MLD period in the mid-1990s in addition to the deep periods in the mid-1980s and mid-2000s by Iwamaru et al. (2010). We examined each profile to check the MLD in the mid-1990s. Figure 4 displays a distribution of profiles in 1995, which is the year with the deepest MLD in the 1990s. This shows deep MLDs greater than 300 m in most profiles. The reason of the difference between the two time series is possibly because of updated database: Iwamaru et al. (2010) used the World Ocean Database 2005 (WOD05; Boyer et al. 2006), while we used the WOD13. Consequently, it is found that the winter MLD south of the KE has significant decadal variation (Fig. 3b) due to the detection of deep MLD period in the mid-1990s.
(a) Time series of MLD (m) in March, for the region south of the KE (30°–37°N, 141°–155°E) (black line). The gray shading denotes one standard error estimated from anomaly values for months with at least five profiles; March 1997 was the only month with less than five profiles. The red line indicates the decadal MLD time series (smoothed with a 1–3–4–3–1 filter). (b) The Morlet wavelet transform coefficient for the normalized raw MLD time series in (a). Shading indicates the amplitude of the real part of the wavelet coefficient. The black line represents the local wavelet spectra, which are defined as the square of the absolute wavelet transform coefficient, and are significant at the 10% significance level. The significance level of the wavelet amplitude is evaluated using a Monte Carlo simulation based on a red noise (AR-1) model for the observed lag−1 correlation coefficient using a 10 000-point surrogate time series. The curved line represents a cone of influence.
Citation: Journal of Climate 29, 3; 10.1175/JCLI-D-15-0206.1
(a) Time series of MLD (m) in March, for the region south of the KE (30°–37°N, 141°–155°E) (black line). The gray shading denotes one standard error estimated from anomaly values for months with at least five profiles; March 1997 was the only month with less than five profiles. The red line indicates the decadal MLD time series (smoothed with a 1–3–4–3–1 filter). (b) The Morlet wavelet transform coefficient for the normalized raw MLD time series in (a). Shading indicates the amplitude of the real part of the wavelet coefficient. The black line represents the local wavelet spectra, which are defined as the square of the absolute wavelet transform coefficient, and are significant at the 10% significance level. The significance level of the wavelet amplitude is evaluated using a Monte Carlo simulation based on a red noise (AR-1) model for the observed lag−1 correlation coefficient using a 10 000-point surrogate time series. The curved line represents a cone of influence.
Citation: Journal of Climate 29, 3; 10.1175/JCLI-D-15-0206.1
(a) Time series of MLD (m) in March, for the region south of the KE (30°–37°N, 141°–155°E) (black line). The gray shading denotes one standard error estimated from anomaly values for months with at least five profiles; March 1997 was the only month with less than five profiles. The red line indicates the decadal MLD time series (smoothed with a 1–3–4–3–1 filter). (b) The Morlet wavelet transform coefficient for the normalized raw MLD time series in (a). Shading indicates the amplitude of the real part of the wavelet coefficient. The black line represents the local wavelet spectra, which are defined as the square of the absolute wavelet transform coefficient, and are significant at the 10% significance level. The significance level of the wavelet amplitude is evaluated using a Monte Carlo simulation based on a red noise (AR-1) model for the observed lag−1 correlation coefficient using a 10 000-point surrogate time series. The curved line represents a cone of influence.
Citation: Journal of Climate 29, 3; 10.1175/JCLI-D-15-0206.1
Distribution of profiles in which the MLD was detected in March 1995. Color-filled circles show the MLD. The dashed rectangle represents the region south of the KE (30°–37°N, 141°–155°E).
Citation: Journal of Climate 29, 3; 10.1175/JCLI-D-15-0206.1
Distribution of profiles in which the MLD was detected in March 1995. Color-filled circles show the MLD. The dashed rectangle represents the region south of the KE (30°–37°N, 141°–155°E).
Citation: Journal of Climate 29, 3; 10.1175/JCLI-D-15-0206.1
Distribution of profiles in which the MLD was detected in March 1995. Color-filled circles show the MLD. The dashed rectangle represents the region south of the KE (30°–37°N, 141°–155°E).
Citation: Journal of Climate 29, 3; 10.1175/JCLI-D-15-0206.1
In the following analysis, the variables are smoothed using a 1–3–4–3–1 filter with a half-power point at 8.4 yr to remove the shorter time scales (year-to-year and interannual time scales). These smoothed variables are labeled simply as “decadal” in this study.
a. Impact of surface cooling and subsurface stratification intensity on the winter MLD
We explore the cause for the decadal MLD variations, focusing on surface cooling, the intensity of preexisting seasonal thermocline, and the intensity of preexisting subsurface stratification. The decadal NHF (Fig. 5a) shows large variance throughout the analysis period (Table 1), while variance of the decadal seasonal thermocline intensity (Fig. 5b) is quite small over the period. Variance of the decadal MTD (Fig. 5c), regarded as an indicator of subsurface stratification intensity, tends to increase in later years, as variance in the most recent 15-yr period is 10 times that in the first 15 years of the present study period (Table 1). It is therefore expected that on decadal time scales, different factors are responsible for control on the winter MLD during different periods.
Time series of (a) decadal downward NHF (W m−2) during winter (December–February; the values are plotted as the calendar year for midwinter conditions, e.g., a winter value for December 1999–February 2000 is plotted as the year 2000), (b) decadal seasonal thermocline intensity (°C m−1) during the preceding warm season (July–September; the values are plotted as the following year, e.g., a summer value for 1999 is plotted as the year 2000), and (c) decadal MTD (m) during the preceding warm season (July–September), in the region south of the KE (30°–37°N, 141°–155°E). The time series in (b) and (c) are produced by the same procedure as for the MLD time series. Note that profiles north of the KE were excluded for the time series calculations, and that the number of profiles was >30 in the warm season of each year. The dashed line in each panel represents the time series of decadal MLD (m) in March south of the KE (the same as the red line in Fig. 3a).
Citation: Journal of Climate 29, 3; 10.1175/JCLI-D-15-0206.1
Time series of (a) decadal downward NHF (W m−2) during winter (December–February; the values are plotted as the calendar year for midwinter conditions, e.g., a winter value for December 1999–February 2000 is plotted as the year 2000), (b) decadal seasonal thermocline intensity (°C m−1) during the preceding warm season (July–September; the values are plotted as the following year, e.g., a summer value for 1999 is plotted as the year 2000), and (c) decadal MTD (m) during the preceding warm season (July–September), in the region south of the KE (30°–37°N, 141°–155°E). The time series in (b) and (c) are produced by the same procedure as for the MLD time series. Note that profiles north of the KE were excluded for the time series calculations, and that the number of profiles was >30 in the warm season of each year. The dashed line in each panel represents the time series of decadal MLD (m) in March south of the KE (the same as the red line in Fig. 3a).
Citation: Journal of Climate 29, 3; 10.1175/JCLI-D-15-0206.1
Time series of (a) decadal downward NHF (W m−2) during winter (December–February; the values are plotted as the calendar year for midwinter conditions, e.g., a winter value for December 1999–February 2000 is plotted as the year 2000), (b) decadal seasonal thermocline intensity (°C m−1) during the preceding warm season (July–September; the values are plotted as the following year, e.g., a summer value for 1999 is plotted as the year 2000), and (c) decadal MTD (m) during the preceding warm season (July–September), in the region south of the KE (30°–37°N, 141°–155°E). The time series in (b) and (c) are produced by the same procedure as for the MLD time series. Note that profiles north of the KE were excluded for the time series calculations, and that the number of profiles was >30 in the warm season of each year. The dashed line in each panel represents the time series of decadal MLD (m) in March south of the KE (the same as the red line in Fig. 3a).
Citation: Journal of Climate 29, 3; 10.1175/JCLI-D-15-0206.1
Variance of the decadal MLD in March (red line in Fig. 3a), the decadal wintertime NHF (solid line in Fig. 5a), the decadal preexisting seasonal thermocline intensity (solid line in Fig. 5b), the decadal preexisting MTD (solid line in Fig. 5c), the decadal MLD in March simulated with full forcing (dashed line in Fig. 7a), the decadal MLD in March in the ocean preconditioning (OP) run (solid line in Fig. 7b), the decadal MLD in March in the atmospheric forcing (AF) run (solid line in Fig. 7c), the monthly MTD anomaly smoothed by a 37-month running-mean filter in the Rossby wave generation region (31°–33°N, 180°–160°W), the monthly WSC anomaly smoothed by a 37-month running-mean filter around the Rossby wave generation region (30°–35°N, 180°–140°W), and the normalized decadal wintertime (December–February) WP index for the first 15-yr (1970–84) and last 15-yr (1998–2012) periods. Values in parentheses represent the correlation coefficients of the decadal MLD in March for both the OP run and the AF run vs the decadal MLD in March simulated with full forcing; boldface typeface represents values exceeding the 10% significance level.
We explore the period dependency for the relationship between the MLD and the three factors (wintertime surface cooling, preexisting surface stratification intensity, and preexisting subsurface stratification intensity) by performing a “running” correlation analysis using a 15-yr window. Results show that surface cooling is the dominant control on the MLD in winter in the late 1970s and 1980s, whereas the strength of subsurface stratification is the main control after 1990 (Fig. 6), as pointed out by Iwamaru et al. (2010). On decadal time scales, the intensity of the seasonal thermocline has no impact on the MLD in the following winter.
Time series obtained by a running correlation analysis of decadal MLD in March (Fig. 3a) vs decadal NHF in winter (Fig. 5a) (black), decadal preexisting seasonal thermocline intensity (Fig. 5b) (blue), and decadal preexisting MTD (Fig. 5c) (red), using a window of 15 years. The correlation coefficient for a given year is calculated, for instance, the 1980 value shown by a cross represents a correlation coefficient between decadal MLD and decadal MTD for 1980–94, and the reference year is shifted by one year each time. White circles indicate significant values exceeding the 10% significance level.
Citation: Journal of Climate 29, 3; 10.1175/JCLI-D-15-0206.1
Time series obtained by a running correlation analysis of decadal MLD in March (Fig. 3a) vs decadal NHF in winter (Fig. 5a) (black), decadal preexisting seasonal thermocline intensity (Fig. 5b) (blue), and decadal preexisting MTD (Fig. 5c) (red), using a window of 15 years. The correlation coefficient for a given year is calculated, for instance, the 1980 value shown by a cross represents a correlation coefficient between decadal MLD and decadal MTD for 1980–94, and the reference year is shifted by one year each time. White circles indicate significant values exceeding the 10% significance level.
Citation: Journal of Climate 29, 3; 10.1175/JCLI-D-15-0206.1
Time series obtained by a running correlation analysis of decadal MLD in March (Fig. 3a) vs decadal NHF in winter (Fig. 5a) (black), decadal preexisting seasonal thermocline intensity (Fig. 5b) (blue), and decadal preexisting MTD (Fig. 5c) (red), using a window of 15 years. The correlation coefficient for a given year is calculated, for instance, the 1980 value shown by a cross represents a correlation coefficient between decadal MLD and decadal MTD for 1980–94, and the reference year is shifted by one year each time. White circles indicate significant values exceeding the 10% significance level.
Citation: Journal of Climate 29, 3; 10.1175/JCLI-D-15-0206.1
Of note, the surface cooling has had only a small influence on the MLD in recent years (Fig. 6) despite having almost the same variance throughout the analysis period (Table 1). This result indicates that surface cooling induces only minor variations in the winter MLD. In fact, in the late 1970s and 1980s when the decadal NHF was correlated significantly with the decadal MLD, the amplitude of the decadal MLD was smaller than that after ~1990 (Fig. 3; Table 1). To quantitatively assess the contribution of both surface cooling and subsurface stratification intensity to wintertime MLD, we use a one-dimensional turbulent closure model that is run separately for each year, beginning in late summer (15 September): the initial ocean conditions are set using temperature profile, produced using the same procedure as for the MLD time series described above, and the salinity profile from the WOA13. We checked that the results were insensitive to the start date and found that almost identical results were obtained when beginning in July, August, and October. In this simulation, the MLD is regarded as the depth at which the temperature difference between 10 m and the bottom of the ML is less than 0.1°C due to the formation of a vertically well-mixed layer, using mean values for March. In Fig. 7a, the simulated decadal MLD accurately reproduces the observed MLD (R = 0.83, which exceeds the 10% significance level). We also perform two runs to show the difference between the influences of surface cooling and preexisting subsurface stratification intensity on the wintertime MLD: 1) an ocean preconditioning (OP) run that is initialized with temperature profiles in September of each year and forced by the 6-hourly NHF climatology, and 2) an atmospheric forcing (AF) run that is initialized with a temperature profile climatology for September and forced by the 6-hourly NHF (Table 2). The decadal MLD in the OP run (Fig. 7b) shows a small amplitude before ~1990 and a larger amplitude after ~1990, corresponding to the behavior of the decadal MTD in Fig. 5c: As expected, the decadal MLD in the OP run matches well with the full forcing results (Fig. 7a) in the last 15-yr period in terms of variance and phase (Table 1). The variance of the OP run is about 55% of that of the decadal MTD (Table 1). This result indicates that a vertical displacement of 30 m in the MTD induces a displacement of ~23 m in the MLD on decadal time scales. On the other hand, the amplitude of the decadal MLD in the AF run (Fig. 7c) is roughly estimated as 10 m throughout the analysis period, which is similar to the decadal MLD with full forcing before ~1990. The simulation results show that on decadal time scales, the winter MLD is determined by a combination of surface cooling and ocean preconditioning (i.e., subsurface stratification intensity). The influence of ocean preconditioning is especially dominant if the vertical displacement of the MTD is greater than 15 m.
(a) Time series of observed (solid line) and simulated (dashed line) decadal MLD (m) in March, in the region south of the KE (30°–37°N, 141°–155°E). (b),(c) As in (a), but for the decadal MLD in March in the OP run [solid line in (b)] and the AF run [solid line in (c)].
Citation: Journal of Climate 29, 3; 10.1175/JCLI-D-15-0206.1
(a) Time series of observed (solid line) and simulated (dashed line) decadal MLD (m) in March, in the region south of the KE (30°–37°N, 141°–155°E). (b),(c) As in (a), but for the decadal MLD in March in the OP run [solid line in (b)] and the AF run [solid line in (c)].
Citation: Journal of Climate 29, 3; 10.1175/JCLI-D-15-0206.1
(a) Time series of observed (solid line) and simulated (dashed line) decadal MLD (m) in March, in the region south of the KE (30°–37°N, 141°–155°E). (b),(c) As in (a), but for the decadal MLD in March in the OP run [solid line in (b)] and the AF run [solid line in (c)].
Citation: Journal of Climate 29, 3; 10.1175/JCLI-D-15-0206.1
Combinations of data used for simulations in the different model runs.
b. Cause of increased variance in subsurface stratification intensity in recent years
The variance of subsurface stratification intensity south of the KE has increased in recent years. Numerous studies have shown that the decadal MTD variation (representing the subsurface stratification intensity) south of the KE is caused by large-scale oceanic Rossby waves formed by a meridional movement of the AL (Ceballos et al. 2009; Sugimoto and Hanawa 2010; Seo et al. 2014). We explore the long-term behavior of Rossby waves from 1970 to 2012, a period that begins more than 10 years earlier than that in past works based on observational data. Figure 8a displays the MTD anomaly averaged over a meridional band of 31°–33°N, which is south of the KE. It is apparent that after ~1990, decadal-scale variations are observed in the western region with most signals being traced from the eastern region around 170°W. This result indicates that the MTD anomaly closely follows the behavior of oceanic Rossby waves. Intriguingly, the variance of the MTD anomaly in the Rossby wave generation region has also increased in recent years: the variance during the most recent 15-yr period is several times greater than that during the first 15-yr period (Table 1).
(a) Longitude–time diagram of the MTD anomaly (m) averaged over a meridional band of 31°–33°N, smoothed by a 37-month running-mean filter. Before calculating the anomaly, the MTD was gridded at 1° (latitude) × 1° (longitude) by applying a Gaussian filter with an e-folding scale of 300 km and 5 months. Positive (negative) values represent deep (shallow) anomalies. (b) Time series of WSC (kg m−2 s−2) around the Rossby wave generation region (30°–35°N, 180°–140°W), smoothed by a 37-month running-mean filter (red line). The black line represents the normalized decadal wintertime (December–February) WP index.
Citation: Journal of Climate 29, 3; 10.1175/JCLI-D-15-0206.1
(a) Longitude–time diagram of the MTD anomaly (m) averaged over a meridional band of 31°–33°N, smoothed by a 37-month running-mean filter. Before calculating the anomaly, the MTD was gridded at 1° (latitude) × 1° (longitude) by applying a Gaussian filter with an e-folding scale of 300 km and 5 months. Positive (negative) values represent deep (shallow) anomalies. (b) Time series of WSC (kg m−2 s−2) around the Rossby wave generation region (30°–35°N, 180°–140°W), smoothed by a 37-month running-mean filter (red line). The black line represents the normalized decadal wintertime (December–February) WP index.
Citation: Journal of Climate 29, 3; 10.1175/JCLI-D-15-0206.1
(a) Longitude–time diagram of the MTD anomaly (m) averaged over a meridional band of 31°–33°N, smoothed by a 37-month running-mean filter. Before calculating the anomaly, the MTD was gridded at 1° (latitude) × 1° (longitude) by applying a Gaussian filter with an e-folding scale of 300 km and 5 months. Positive (negative) values represent deep (shallow) anomalies. (b) Time series of WSC (kg m−2 s−2) around the Rossby wave generation region (30°–35°N, 180°–140°W), smoothed by a 37-month running-mean filter (red line). The black line represents the normalized decadal wintertime (December–February) WP index.
Citation: Journal of Climate 29, 3; 10.1175/JCLI-D-15-0206.1
We explore WSC variations, which generate the oceanic Rossby waves. The WSC anomaly around the Rossby wave generation region has an opposite sign to the WP index (Fig. 8b), showing negative (positive) WSC anomalies associated with the northward (southward) movement of the AL. Variation in both the WSC and the related AL meridional movement (i.e., the WP index) also appears to have increased in recent years (Fig. 8b), as pointed out recently by Pak et al. (2014): the ratio of variance during the last 15-yr period to that in first 15-yr period in both the WSC and the WP index is similar to that of the underlying MTD variation (Table 1). After ~1990, the increased variance of the WSC, which results from the marked AL meridional movement, induces large variation in the MTD in the central North Pacific. As a result, the MTD signals propagate westward and reach the region south of the KE after a delay of a few years, resulting in increased variance of subsurface stratification intensity.
Previous studies have reported that strong eddy activity associated with an unstable state of the KE path leads to intensified subsurface stratification (Qiu and Chen 2006; Qiu et al. 2007; Oka et al. 2012, 2014). We examine the relationship between the KE path state and subsurface stratification intensity using a long-term dataset that begins more than 10 years prior to the period in the past works. Figure 9 displays the decadal time series of the KE path length, as compiled by Seo et al. (2014): a longer (shorter) path length represents the unstable (stable) state of the KE path. As suggested by recent works (e.g., Oka et al. 2012, 2014), the wintertime MLD tends to be shallower during the unstable state of the KE path (R = −0.66 for the two time series, which exceeds the 10% significance level). The long-term time series in Fig. 9 shows that the KE path in an unstable state has tended to become longer (more convoluted) in recent years, which would result in increased variance in the intensity of subsurface stratification south of the KE.
Time series of wintertime decadal KE path length (km) from 141° to 155°E, produced by Seo et al. (2014) (solid line); the KE (strictly, KE northern boundary) is detected based on the position of a strong winter SST gradient, using the NOAA Optimum Interpolation SST (OISST) product, which is based on the Advanced Very High Resolution Radiometer (AVHRR) infrared satellite data (AVHRR-only product; Reynolds et al. 2007). The dashed line represents the decadal MLD (m) in March (the same as the red line in Fig. 3a).
Citation: Journal of Climate 29, 3; 10.1175/JCLI-D-15-0206.1
Time series of wintertime decadal KE path length (km) from 141° to 155°E, produced by Seo et al. (2014) (solid line); the KE (strictly, KE northern boundary) is detected based on the position of a strong winter SST gradient, using the NOAA Optimum Interpolation SST (OISST) product, which is based on the Advanced Very High Resolution Radiometer (AVHRR) infrared satellite data (AVHRR-only product; Reynolds et al. 2007). The dashed line represents the decadal MLD (m) in March (the same as the red line in Fig. 3a).
Citation: Journal of Climate 29, 3; 10.1175/JCLI-D-15-0206.1
Time series of wintertime decadal KE path length (km) from 141° to 155°E, produced by Seo et al. (2014) (solid line); the KE (strictly, KE northern boundary) is detected based on the position of a strong winter SST gradient, using the NOAA Optimum Interpolation SST (OISST) product, which is based on the Advanced Very High Resolution Radiometer (AVHRR) infrared satellite data (AVHRR-only product; Reynolds et al. 2007). The dashed line represents the decadal MLD (m) in March (the same as the red line in Fig. 3a).
Citation: Journal of Climate 29, 3; 10.1175/JCLI-D-15-0206.1
The above results showed that the decadal subsurface stratification intensity south of the KE results from the combined effect of large-scale oceanic Rossby waves and mesoscale eddy activity associated with the path state of the KE, and indicated that the increased variance in subsurface stratification intensity after ~1990 reflects both enhanced variance in the WSC in the central North Pacific and a more convoluted KE path in recent years.
4. Decadal variations in MLT
a. ML heat budget analysis
As in Fig. 3, but for MLT, defined as the temperature at a depth of 10 m. The time series is produced using the same process as for the monthly MLD time series.
Citation: Journal of Climate 29, 3; 10.1175/JCLI-D-15-0206.1
As in Fig. 3, but for MLT, defined as the temperature at a depth of 10 m. The time series is produced using the same process as for the monthly MLD time series.
Citation: Journal of Climate 29, 3; 10.1175/JCLI-D-15-0206.1
As in Fig. 3, but for MLT, defined as the temperature at a depth of 10 m. The time series is produced using the same process as for the monthly MLD time series.
Citation: Journal of Climate 29, 3; 10.1175/JCLI-D-15-0206.1
The decadal temperature tendency term (black line in Fig. 11a) has large positive peaks in the late 1980s, late 1990s, and late 2000s, and large negative peaks in the early 1980s, early 1990s, and early 2000s. This temporal behavior is consistent with the model outputs (brown line in Fig. 11a), and a significant correlation coefficient is obtained between the two time series (R = 0.69). We investigate the relative contributions of the four terms (i.e., air–sea heat exchange, Ekman advection, vertical entrainment, and geostrophic advection) to the decadal temperature tendency. Results show that the decadal vertical entrainment term (red line in Fig. 11b) has a large variance and is comparable to the decadal temperature tendency term (black line in Fig. 11a); a significant correlation coefficient is obtained between the two time series (R = 0.76). The amplitude of the decadal air–sea heat exchange term (blue line in Fig. 11b) is also large, but the temporal behavior is completely different from that of the decadal temperature tendency term (R = −0.06). The variances of the decadal Ekman advection term (green line in Fig. 11b) and the decadal geostrophic advection term (purple line in Fig. 11b) are relatively small, and neither time series is significantly correlated with the decadal temperature tendency term (R = 0.04 for the Ekman advection term and R = −0.07 for the geostrophic advection term). We examine the period dependency for the relationship between the decadal temperature tendency term and the above four terms. Results show that the vertical entrainment is the dominant control on the temperature tendency throughout the analysis period (Fig. 12), except for the early 1970s. That is, the vertical entrainment is primarily responsible for the formation of decadal wintertime MLT south of the KE.
(a) Time series of the decadal anomaly of the temperature tendency term (°C yr−1) (black line), defined as MLT in March minus MLT in March of the previous year, in the region south of the KE (30°–37°N, 141°–155°E). The brown line represents a time series of the decadal anomaly (°C yr−1), which is the sum of the air–sea heat exchange term, the Ekman advection term, the vertical entrainment term, and the geostrophic advection term; each term is defined as the sum of values from April of the previous year to March, in the region south of the KE. (b) As in (a), but for the decadal anomaly of the vertical entrainment term (°C yr−1) (red line), the air–sea heat exchange term (°C yr−1) (blue line), the Ekman advection term (°C yr−1) (green line), and the geostrophic advection term (°C yr−1) (purple line). (c) As in (a), but for the residual term (orange line).
Citation: Journal of Climate 29, 3; 10.1175/JCLI-D-15-0206.1
(a) Time series of the decadal anomaly of the temperature tendency term (°C yr−1) (black line), defined as MLT in March minus MLT in March of the previous year, in the region south of the KE (30°–37°N, 141°–155°E). The brown line represents a time series of the decadal anomaly (°C yr−1), which is the sum of the air–sea heat exchange term, the Ekman advection term, the vertical entrainment term, and the geostrophic advection term; each term is defined as the sum of values from April of the previous year to March, in the region south of the KE. (b) As in (a), but for the decadal anomaly of the vertical entrainment term (°C yr−1) (red line), the air–sea heat exchange term (°C yr−1) (blue line), the Ekman advection term (°C yr−1) (green line), and the geostrophic advection term (°C yr−1) (purple line). (c) As in (a), but for the residual term (orange line).
Citation: Journal of Climate 29, 3; 10.1175/JCLI-D-15-0206.1
(a) Time series of the decadal anomaly of the temperature tendency term (°C yr−1) (black line), defined as MLT in March minus MLT in March of the previous year, in the region south of the KE (30°–37°N, 141°–155°E). The brown line represents a time series of the decadal anomaly (°C yr−1), which is the sum of the air–sea heat exchange term, the Ekman advection term, the vertical entrainment term, and the geostrophic advection term; each term is defined as the sum of values from April of the previous year to March, in the region south of the KE. (b) As in (a), but for the decadal anomaly of the vertical entrainment term (°C yr−1) (red line), the air–sea heat exchange term (°C yr−1) (blue line), the Ekman advection term (°C yr−1) (green line), and the geostrophic advection term (°C yr−1) (purple line). (c) As in (a), but for the residual term (orange line).
Citation: Journal of Climate 29, 3; 10.1175/JCLI-D-15-0206.1
As in Fig. 6, but for the decadal temperature tendency term (black line in Fig. 11a) vs decadal air–sea heat exchange term (blue line in Fig. 11b) (blue), decadal vertical entrainment term (red line in Fig. 11b) (red), decadal Ekman heat advection term (green line in Fig. 11b) (green), and decadal geostrophic advection term (purple line in Fig. 11b) (purple).
Citation: Journal of Climate 29, 3; 10.1175/JCLI-D-15-0206.1
As in Fig. 6, but for the decadal temperature tendency term (black line in Fig. 11a) vs decadal air–sea heat exchange term (blue line in Fig. 11b) (blue), decadal vertical entrainment term (red line in Fig. 11b) (red), decadal Ekman heat advection term (green line in Fig. 11b) (green), and decadal geostrophic advection term (purple line in Fig. 11b) (purple).
Citation: Journal of Climate 29, 3; 10.1175/JCLI-D-15-0206.1
As in Fig. 6, but for the decadal temperature tendency term (black line in Fig. 11a) vs decadal air–sea heat exchange term (blue line in Fig. 11b) (blue), decadal vertical entrainment term (red line in Fig. 11b) (red), decadal Ekman heat advection term (green line in Fig. 11b) (green), and decadal geostrophic advection term (purple line in Fig. 11b) (purple).
Citation: Journal of Climate 29, 3; 10.1175/JCLI-D-15-0206.1
The vertical entrainment process is mainly determined by two variables, as shown in Eq. (3): MLD (we and H) and temperature (Tm and Td). To evaluate the relative contributions of the two variables to vertical entrainment, we perform a simple experiment using two sets of data: monthly raw data and monthly climatological data. Using these data, we model the vertical entrainment term based on the dataset combinations shown in Table 3, with each combination designated as a “run.” The decadal MLD run (red line in Fig. 13) resembles the decadal vertical entrainment term (red line in Fig. 11b; R = 0.74, which exceeds the 10% significance level), and a large portion (90.1%) of variance in the vertical entrainment term is attributed to the MLD run. On the other hand, the decadal temperature run (blue line in Fig. 13) is not significantly correlated with the vertical entrainment term (R = 0.43), and the proportion of variance explained by this run is much smaller (39.3%). That is, MLD variation is the main control on the vertical entrainment process, which determines the temperature tendency in the winter ML and results in decadal MLT variation. In fact, the decadal MLD in March is strongly correlated with the decadal temperature tendency term (R = −0.73, which exceeds the 10% significance level). Based on Eq. (3), it is reasonably acceptable that the vertical entrainment process is dominant in a cold season when the ML deepens. Actually, a comparison between strong and weak entrainment periods clearly shows a rapid deepening of the ML in January and February during strong entrainment periods (Fig. 14).
Combinations of data used for vertical entrainment calculations in the different model runs.
Time series of the decadal anomaly of two runs (°C yr−1), in the region south of the KE (30°–37°N, 141°–155°E): the MLD run (red line) and the temperature run (blue line). The dashed line represents the decadal anomaly of the vertical entrainment term (°C yr−1) (the same as the red line in Fig. 11b).
Citation: Journal of Climate 29, 3; 10.1175/JCLI-D-15-0206.1
Time series of the decadal anomaly of two runs (°C yr−1), in the region south of the KE (30°–37°N, 141°–155°E): the MLD run (red line) and the temperature run (blue line). The dashed line represents the decadal anomaly of the vertical entrainment term (°C yr−1) (the same as the red line in Fig. 11b).
Citation: Journal of Climate 29, 3; 10.1175/JCLI-D-15-0206.1
Time series of the decadal anomaly of two runs (°C yr−1), in the region south of the KE (30°–37°N, 141°–155°E): the MLD run (red line) and the temperature run (blue line). The dashed line represents the decadal anomaly of the vertical entrainment term (°C yr−1) (the same as the red line in Fig. 11b).
Citation: Journal of Climate 29, 3; 10.1175/JCLI-D-15-0206.1
A composite of monthly MLD (m): (red line) strong entrainment periods (1991–93, 2003–05, and 2011/12) and (blue line) weak entrainment periods (1987–89, 1997–99, and 2007–09); strong (weak) entrainment periods are defined as periods during negative (positive) peaks of the decadal vertical entrainment term (red line in Fig. 11b).
Citation: Journal of Climate 29, 3; 10.1175/JCLI-D-15-0206.1
A composite of monthly MLD (m): (red line) strong entrainment periods (1991–93, 2003–05, and 2011/12) and (blue line) weak entrainment periods (1987–89, 1997–99, and 2007–09); strong (weak) entrainment periods are defined as periods during negative (positive) peaks of the decadal vertical entrainment term (red line in Fig. 11b).
Citation: Journal of Climate 29, 3; 10.1175/JCLI-D-15-0206.1
A composite of monthly MLD (m): (red line) strong entrainment periods (1991–93, 2003–05, and 2011/12) and (blue line) weak entrainment periods (1987–89, 1997–99, and 2007–09); strong (weak) entrainment periods are defined as periods during negative (positive) peaks of the decadal vertical entrainment term (red line in Fig. 11b).
Citation: Journal of Climate 29, 3; 10.1175/JCLI-D-15-0206.1
We examine the residual term because it has a large measurable amplitude with a low-frequency time scale (orange line in Fig. 11c). Figure 15 displays a snapshot in 2009, which is the year with the large residual term. In the mid-March 2009, the KE was characterized by warm SSTs and a large southward meander around 146°E (Fig. 15a). Later in this month, the cold eddy detached southward from the KE, and then the ambient flow of the cold eddy advected warm water (Fig. 15b), moving westward in the region south of the KE. It is expected that the residual term can be largely explained by the heat transport due the ambient flow of eddy. We investigate eddy activity south of the KE in terms of eddy kinetic energy (EKE) and the number of cold eddies detached southward from the KE as indictors representing the eddy heat transport. The number of detached cold eddies that pass through the region south of the KE shows a low-frequency variation, with larger numbers around 2000 and in the late 2000s, and smaller numbers in the early 1990s and early 2000s (Fig. 16a). The decadal EKE (Fig. 16b) also has a low-frequency variation, with larger values around 2000 and in the late 2000s, and smaller values in the early 1990s and early 2000s. Temporal features of both the eddy number and the EKE are similar to the behavior of the residual term. This suggests that the heat transport due to the ambient flow of cold eddies affects the MLT south of the KE. Recent studies have pointed out that the eddy activity around the KE is related to the path state of the KE (e.g., Qiu and Chen 2010; Oka et al. 2012). In fact, during periods with an unstable KE path, the number of eddies tends to be greater and the EKE levels tend to be higher (Fig. 16).
Snapshots taken at (a) 18 Mar 2009 and (b) 23 Apr 2009 using satellite-derived blended SST of Hosoda and Sakaida (2015) (°C) (color), sea surface velocity vectors of AVISO (cm s−1) (arrows; larger than 50 cm s−1), and the SSH = 90 cm contour of AVISO (thick gray line), regarded as the KE axis.
Citation: Journal of Climate 29, 3; 10.1175/JCLI-D-15-0206.1
Snapshots taken at (a) 18 Mar 2009 and (b) 23 Apr 2009 using satellite-derived blended SST of Hosoda and Sakaida (2015) (°C) (color), sea surface velocity vectors of AVISO (cm s−1) (arrows; larger than 50 cm s−1), and the SSH = 90 cm contour of AVISO (thick gray line), regarded as the KE axis.
Citation: Journal of Climate 29, 3; 10.1175/JCLI-D-15-0206.1
Snapshots taken at (a) 18 Mar 2009 and (b) 23 Apr 2009 using satellite-derived blended SST of Hosoda and Sakaida (2015) (°C) (color), sea surface velocity vectors of AVISO (cm s−1) (arrows; larger than 50 cm s−1), and the SSH = 90 cm contour of AVISO (thick gray line), regarded as the KE axis.
Citation: Journal of Climate 29, 3; 10.1175/JCLI-D-15-0206.1
(a) Number of cold eddies south of the KE (30°–37°N, 141°–155°E), detached southward from the KE (red bar, left axis), from April of the previous year to March. Here, the eddies are defined as a closed contour of SSH = 90 cm, and we used only eddies that could be tracked over 4 weeks, as recommended by Itoh and Yasuda (2010). The blue line represents the time series of the residual term in Eq. (1) (°C yr−1) (the same as the orange line in Fig. 11c). The black dashed line represents the time series of wintertime decadal KE path length (the same as solid line in Fig. 9). (b) As in (a), but for the time series of the EKE (m2 s−2), calculated from high-passed sea surface geostrophic velocity data with time scales shorter than 300 days, based on a methodology of Qiu and Chen (2005), averaged for the region south of the KE from April of the previous year to March; raw time series (red dashed line, left axis) and decadal time series (red solid line, left axis).
Citation: Journal of Climate 29, 3; 10.1175/JCLI-D-15-0206.1
(a) Number of cold eddies south of the KE (30°–37°N, 141°–155°E), detached southward from the KE (red bar, left axis), from April of the previous year to March. Here, the eddies are defined as a closed contour of SSH = 90 cm, and we used only eddies that could be tracked over 4 weeks, as recommended by Itoh and Yasuda (2010). The blue line represents the time series of the residual term in Eq. (1) (°C yr−1) (the same as the orange line in Fig. 11c). The black dashed line represents the time series of wintertime decadal KE path length (the same as solid line in Fig. 9). (b) As in (a), but for the time series of the EKE (m2 s−2), calculated from high-passed sea surface geostrophic velocity data with time scales shorter than 300 days, based on a methodology of Qiu and Chen (2005), averaged for the region south of the KE from April of the previous year to March; raw time series (red dashed line, left axis) and decadal time series (red solid line, left axis).
Citation: Journal of Climate 29, 3; 10.1175/JCLI-D-15-0206.1
(a) Number of cold eddies south of the KE (30°–37°N, 141°–155°E), detached southward from the KE (red bar, left axis), from April of the previous year to March. Here, the eddies are defined as a closed contour of SSH = 90 cm, and we used only eddies that could be tracked over 4 weeks, as recommended by Itoh and Yasuda (2010). The blue line represents the time series of the residual term in Eq. (1) (°C yr−1) (the same as the orange line in Fig. 11c). The black dashed line represents the time series of wintertime decadal KE path length (the same as solid line in Fig. 9). (b) As in (a), but for the time series of the EKE (m2 s−2), calculated from high-passed sea surface geostrophic velocity data with time scales shorter than 300 days, based on a methodology of Qiu and Chen (2005), averaged for the region south of the KE from April of the previous year to March; raw time series (red dashed line, left axis) and decadal time series (red solid line, left axis).
Citation: Journal of Climate 29, 3; 10.1175/JCLI-D-15-0206.1
b. Influence of decadal MLT south of the KE on the temperature field in the western part of the subtropical gyre
The region south of the KE corresponds to the main formation area of STMW (Hanawa 1987; Hanawa and Talley 2001; Oka and Qiu 2012; Sugimoto and Hanawa 2014). The STMW formed south of the KE is advected westward–southwestward with the Kuroshio Countercurrent (Bingham 1992; Suga and Hanawa 1995) and is then distributed widely over the western part of the North Pacific subtropical gyre (see the review by Oka and Qiu 2012). Thick thermostad layer, corresponding to the STMW, is detected in the downstream region of the Kuroshio Countercurrent (Fig. 17a). Figure 17b displays the long-term core layer temperature (CLT) of STMW, averaged for 25°–30°N, along 137°E in the western part of the subtropical gyre. This record shows a marked decadal-scale variation, as noted previously (e.g., Taneda et al. 2000). The temporal behavior of the CLT closely resembles that of the decadal wintertime MLT south of the KE (R = 0.85, which exceeds the 10% significance level). The decadal MLT south of the KE strongly influences the temperature field in the western part of the subtropical gyre, in association with the advection of STMW.
(a) STMW distribution in July (m) (shading), represented by a region where the thermostad layer thickness of T = 16°–19°C is >100 m, and the geopotential anomaly at 250 m relative to 1000 m in July (m2 s−2) (contours), calculated from WOA13 (Locarnini et al. 2013; Zweng et al. 2013). Thick (thin) contours are drawn with an interval of 1.0 (0.2) m2 s−2. The gray rectangle indicates the region south of the KE (30°–37°N, 141°–155°E) and the red dotted line shows repeat hydrographic sections along 137°E conducted by the JMA. (b) Summertime STMW CLT (°C) along 137°E, averaged for 25°–30°N; raw time series (red dashed line) and decadal time series (red solid line). The black line indicates the time series of decadal MLT in March in the region south of the KE (the same as the red line in Fig. 10a).
Citation: Journal of Climate 29, 3; 10.1175/JCLI-D-15-0206.1
(a) STMW distribution in July (m) (shading), represented by a region where the thermostad layer thickness of T = 16°–19°C is >100 m, and the geopotential anomaly at 250 m relative to 1000 m in July (m2 s−2) (contours), calculated from WOA13 (Locarnini et al. 2013; Zweng et al. 2013). Thick (thin) contours are drawn with an interval of 1.0 (0.2) m2 s−2. The gray rectangle indicates the region south of the KE (30°–37°N, 141°–155°E) and the red dotted line shows repeat hydrographic sections along 137°E conducted by the JMA. (b) Summertime STMW CLT (°C) along 137°E, averaged for 25°–30°N; raw time series (red dashed line) and decadal time series (red solid line). The black line indicates the time series of decadal MLT in March in the region south of the KE (the same as the red line in Fig. 10a).
Citation: Journal of Climate 29, 3; 10.1175/JCLI-D-15-0206.1
(a) STMW distribution in July (m) (shading), represented by a region where the thermostad layer thickness of T = 16°–19°C is >100 m, and the geopotential anomaly at 250 m relative to 1000 m in July (m2 s−2) (contours), calculated from WOA13 (Locarnini et al. 2013; Zweng et al. 2013). Thick (thin) contours are drawn with an interval of 1.0 (0.2) m2 s−2. The gray rectangle indicates the region south of the KE (30°–37°N, 141°–155°E) and the red dotted line shows repeat hydrographic sections along 137°E conducted by the JMA. (b) Summertime STMW CLT (°C) along 137°E, averaged for 25°–30°N; raw time series (red dashed line) and decadal time series (red solid line). The black line indicates the time series of decadal MLT in March in the region south of the KE (the same as the red line in Fig. 10a).
Citation: Journal of Climate 29, 3; 10.1175/JCLI-D-15-0206.1
5. Summary and conclusions
We investigated the long-term behavior of the wintertime MLD and MLT south of the KE (30°–37°N, 141°–155°E), using monthly time series produced based on historical temperature profiles archived in the WOD13 and JODC and profiles from Argo floats during 1968–2014; more than 30 profiles were available for most months. The MLD in March had low-frequency variations and showed significant decadal variations after the late 1980s. Observational data and simulation outputs from a one-dimensional turbulent closure model revealed that the cause of MLD variation was different before and after ~1990, as pointed out by Iwamaru et al. (2010). In the late 1970s and 1980s, the MLD was influenced by surface cooling, whereas after this the main influence was the intensity of subsurface stratification, induced by the combined effect of large-scale oceanic Rossby waves generated in the central North Pacific and the path state of the KE. It was found that this change at about 1990 was due to a dramatic increase in the variance of subsurface stratification intensity in recent years. In contrast, the variance of surface cooling was largely invariant throughout the analysis period. We propose that this increase in the variance of subsurface stratification intensity is attributable to larger WSC variation around the Rossby wave generation region, with a marked meridional movement of the AL, and a more convoluted path of the KE during an unstable state of the KE path.
The MLT in March, which is equivalent to SST, south of the KE also showed significant decadal variations after the late 1980s, the period of which was similar to the wintertime MLD. The ML heat budget analysis from observational data revealed that the entrainment is primarily responsible for the decadal MLT throughout the analysis period, except for the early 1970s. During deeper (shallower) periods of winter MLD, the strong (weak) vertical entrainment process, resulting from a rapid (slow) deepening of the ML in January and February, formed the negative (positive) anomaly of temperature tendency. This results in the formation of decadal temperature variations in the winter ML. The decadal wintertime MLT south of the KE has a strong effect on the temperature field in the western part of the North Pacific subtropical gyre, associated with the advection of STMW. To further our understanding of temperature variations in the western part of the North Pacific subtropical gyre, the subduction and advection processes of STMW should be further explored. The satellite-derived SST map with a high spatial resolution provided an interesting perspective that the MLT south of the KE can be affected by the eddy heat transport due to the ambient flow of cold eddies detached southward from the KE. Recent studies by eddy-resolving model outputs also reported that the ambient flow of cold eddies detached southward from the KE can transport warm water (e.g., Aoki et al. 2013). A quantitative assessment of eddy-induced heat transport associated with the KE path state is needed to reveal the variation in decadal MLT, which would provide new insight into the decadal temperature variations in the North Pacific subtropical gyre.
The decadal wintertime MLT south of the KE is controlled mainly by vertical entrainment. This implies that the MLT (i.e., SST) influences the amount of upward heat release to the overlying atmosphere. Figure 18 displays a spatial distribution of correlation coefficients between the decadal SST and decadal upward turbulent heat flux (THF; i.e., the sum of latent heat flux and sensible heat flux) at each grid point in late winter (February–March) after 1990. Interestingly, positive correlation coefficients are found in the region south of the KE although the coefficients are only significant in some areas. This result indicates that upward heat release tends to be larger over areas of warm SST south of the KE on the decadal time scales after 1990. In addition, significant correlation coefficients are widespread in the region north of the KE. This positive relationship is a result of the upward heat release induced by warm SSTs due to “warm eddies” detached northward from the KE during a period of instability in the KE path (Sugimoto and Hanawa 2011). That is, the path state of the KE significantly influences oceanic conditions (e.g., MLD and MLT) in regions both north and south of the KE. Therefore, the cause of changes in the path state of the KE should be revealed for further understanding of decadal variations in both oceanic and atmospheric fields.
Map of the correlation coefficients between decadal SST and decadal upward THF at each grid point of the objectively analyzed air–sea fluxes (OAFlux) dataset (Yu et al. 2008) in late winter (February–March) of 1990–2012. The yellow lines indicate regions exceeding the 10% significance level. The black contours represent the upward THF climatology in late winter (200, 250, 300, and 350 W m−2), and the blue rectangle indicates the region south of the KE (30°–37°N, 141°–155°E).
Citation: Journal of Climate 29, 3; 10.1175/JCLI-D-15-0206.1
Map of the correlation coefficients between decadal SST and decadal upward THF at each grid point of the objectively analyzed air–sea fluxes (OAFlux) dataset (Yu et al. 2008) in late winter (February–March) of 1990–2012. The yellow lines indicate regions exceeding the 10% significance level. The black contours represent the upward THF climatology in late winter (200, 250, 300, and 350 W m−2), and the blue rectangle indicates the region south of the KE (30°–37°N, 141°–155°E).
Citation: Journal of Climate 29, 3; 10.1175/JCLI-D-15-0206.1
Map of the correlation coefficients between decadal SST and decadal upward THF at each grid point of the objectively analyzed air–sea fluxes (OAFlux) dataset (Yu et al. 2008) in late winter (February–March) of 1990–2012. The yellow lines indicate regions exceeding the 10% significance level. The black contours represent the upward THF climatology in late winter (200, 250, 300, and 350 W m−2), and the blue rectangle indicates the region south of the KE (30°–37°N, 141°–155°E).
Citation: Journal of Climate 29, 3; 10.1175/JCLI-D-15-0206.1
In this study, we have especially focused on the western area (west of 155°E) south of the KE. As shown in Fig. 1, relatively deep winter ML is distributed in the eastern region between 160° and 170°E. We examine the relationship between the decadal wintertime SST in the western and eastern areas. The SST in the eastern area (black line in Fig. 19a) also shows a marked decadal variation, but the behavior is different to that in the western area (red line in Fig. 19a): there is no significant correlation between the two time series (R = 0.31). In fact, most of the significant correlations with the decadal SST in the eastern area are restricted to the region east of 155°E (Fig. 19c), while the decadal SST in the western area is significantly correlated with the SST in the region west of 155°E (Fig. 19b). Interestingly, high correlation coefficients with the decadal SST in the eastern area (Fig. 19c) are obtained along the KE axis from 141° to 155°E, an area that represents the north–south movement of the KE. Therefore, the decadal SST east of 155°E might be formed by transport variations associated with changes in the strength of the KE, which is related to the north–south movement of the KE (e.g., Qiu and Chen 2005, 2010). The STMW formed in the deep winter ML east of 155°E is advected into the southern–southwestern part (south of 25°N) of the North Pacific subtropical gyre (Oka 2009). To more completely understand the decadal-scale variations in the North Pacific subtropical gyre, it is necessary to explore the processes that generate SST anomalies in the eastern region (east of 155°E) south of the KE, as well as the advection of such anomalies in the subsurface layer.
(a) Time series of the decadal SST (°C) in March of MGDSST, in an area south of the KE: (red line) western area at 32°N, 148°E [red circle in (b)] and (black line) eastern area at 33°N, 165°E [black circle in (c)]. (b) Distribution of correlation coefficients between the decadal SST in the western area [red line in (a)] and the decadal SST field in March. The yellow and purple lines indicate the regions exceeding the 10% significance level. The contours represent the SSH climatology of AVISO (contour interval is 10 cm); the thick black line represents the SSH = 90 cm contour, regarded as the KE axis. (c) As in (b), but for the eastern area [black line in (a)].
Citation: Journal of Climate 29, 3; 10.1175/JCLI-D-15-0206.1
(a) Time series of the decadal SST (°C) in March of MGDSST, in an area south of the KE: (red line) western area at 32°N, 148°E [red circle in (b)] and (black line) eastern area at 33°N, 165°E [black circle in (c)]. (b) Distribution of correlation coefficients between the decadal SST in the western area [red line in (a)] and the decadal SST field in March. The yellow and purple lines indicate the regions exceeding the 10% significance level. The contours represent the SSH climatology of AVISO (contour interval is 10 cm); the thick black line represents the SSH = 90 cm contour, regarded as the KE axis. (c) As in (b), but for the eastern area [black line in (a)].
Citation: Journal of Climate 29, 3; 10.1175/JCLI-D-15-0206.1
(a) Time series of the decadal SST (°C) in March of MGDSST, in an area south of the KE: (red line) western area at 32°N, 148°E [red circle in (b)] and (black line) eastern area at 33°N, 165°E [black circle in (c)]. (b) Distribution of correlation coefficients between the decadal SST in the western area [red line in (a)] and the decadal SST field in March. The yellow and purple lines indicate the regions exceeding the 10% significance level. The contours represent the SSH climatology of AVISO (contour interval is 10 cm); the thick black line represents the SSH = 90 cm contour, regarded as the KE axis. (c) As in (b), but for the eastern area [black line in (a)].
Citation: Journal of Climate 29, 3; 10.1175/JCLI-D-15-0206.1
Acknowledgments
The authors thank the members of the Physical Oceanography Group at Tohoku University for useful discussions. Dr. Eitarou Oka provided useful and constructive comments that improved the manuscript. The comments of two anonymous reviewers were helpful in revising the manuscript. The first author (SS) was supported by funds from the Japan Society for Promotion of Science [Grant-in-Aid for Young Scientists (B) 15K17756], and from the Ministry of Education, Culture, Sports, Science and Technology [Grant-in-Aid for Scientific Research on Innovative Areas 25106702, “A ‘hot spot’ in the climate system: Extra-tropical air–sea interaction under the East Asian monsoon system”].
REFERENCES
Akima, H., 1970: A new method of interpolation and smooth curve fitting based on local procedures. J. Assoc. Comput. Mach., 17, 589–602, doi:10.1145/321607.321609.
Aoki, K., S. Minobe, Y. Tanimoto, and Y. Sasai, 2013: Southward eddy heat transport occurring along southern flanks of the Kuroshio Extension and Gulf Stream in a 1/10° global ocean general circulation model. J. Phys. Oceanogr., 43, 1899–1910, doi:10.1175/JPO-D-12-0223.1.
Barnston, A. G., and R. E. Livezey, 1987: Classification, seasonality, and persistence of low-frequency atmospheric circulation patterns. Mon. Wea. Rev., 115, 1083–1126, doi:10.1175/1520-0493(1987)115<1083:CSAPOL>2.0.CO;2.
Bingham, F. M., 1992: Formation and spreading of subtropical mode water in the North Pacific. J. Geophys. Res., 97, 11 177–11 189, doi:10.1029/92JC01001.
Boyer, T. P., and Coauthors, 2006: World Ocean Database 2005. NOAA Atlas NESDIS 60, 190 pp. [Available online at ftp://ftp.nodc.noaa.gov/pub/WOD05/DOC/wod05_intro.pdf.]
Boyer, T. P., and Coauthors, 2013: World Ocean Database 2013. NOAA Atlas NESDIS 72, 209 pp. [Available online at http://data.nodc.noaa.gov/woa/WOD13/DOC/wod13_intro.pdf.]
Ceballos, L. I., E. Di Lorenzo, C. D. Hoyos, N. Schneider, and B. Taguchi, 2009: North Pacific Gyre Oscillation synchronizes climate fluctuations in the eastern and western boundary systems. J. Climate, 22, 5163–5174, doi:10.1175/2009JCLI2848.1.
Davis, R., 1976: Predictability of sea surface temperature and sea level pressure anomalies over the North Pacific Ocean. J. Phys. Oceanogr., 6, 249–266, doi:10.1175/1520-0485(1976)006<0249:POSSTA>2.0.CO;2.
Deser, C., M. A. Alexander, and M. S. Timlin, 1999: Evidence for a wind-driven intensification of the Kuroshio Current Extension from the 1970s to the 1980s. J. Climate, 12, 1697–1706, doi:10.1175/1520-0442(1999)012<1697:EFAWDI>2.0.CO;2.
Hanawa, K., 1987: Interannual variations of the wintertime outcrop area of Subtropical Mode Water in the western North Pacific Ocean. Atmos.–Ocean, 25, 358–374, doi:10.1080/07055900.1987.9649280.
Hanawa, K., and T. Suga, 1995: A review on the subtropical mode water in the North Pacific (NPSTMW). Biogeochemical Process and Ocean Flux in the Western Pacific, H. Sakai and Y. Nozaki, Eds., Terra Scientific Publishing Company, 613– 627.
Hanawa, K., and J. Kamada, 2001: Variability of core layer temperature (CLT) of the North Pacific subtropical mode water. Geophys. Res. Lett., 28, 2229–2232, doi:10.1029/2000GL011716.
Hanawa, K., and L. D. Talley, 2001: Mode waters. Ocean Circulation and Climate, G. Siedler, J. Church, and W. J. Gould, Eds., Academic Press, 373–386.
Hanawa, K., and H. Yoritaka, 2001: North Pacific subtropical mode waters observed in long XBT cross sections along 32.5°N line. J. Oceanogr., 57, 679–692, doi:10.1023/A:1021276224519.
Hosoda, K., 2013: Empirical method of diurnal correction for estimating sea surface temperature at dawn and noon. J. Oceanogr., 69, 631–646, doi:10.1007/s10872-013-0194-4.
Hosoda, K., 2015: Global space-time scales for day-to-day variations of daily-minimum and diurnal sea surface temperatures: Their distinct spatial distribution and seasonal cycles. J. Oceanogr., doi:10.1007/s10872-015-0327-z, in press.
Hosoda, K., and F. Sakaida, 2015: Development of daily-minimum sea surface temperature data set based on microwave and visible/infrared measurements from space. Proc. 30th Int. Symp. on Space Technology and Science, Kobe, Japan, JAXA, 2015-n-37. [Available online at https://www.meeting-schedule.com/ists30/pdf/2015-n-37.pdf.]
Ishi, Y., and K. Hanawa, 2005: Large-scale variabilities of winter- time wind stress curl field in the North Pacific and their relation to atmospheric teleconnection patterns. Geophys. Res. Lett., 32, L10607, doi:10.1029/2004GL022330.
Itoh, S., and I. Yasuda, 2010: Characteristics of mesoscale eddies in the Kuroshio–Oyashio extension region detected from the distribution of the sea surface height anomaly. J. Phys. Oceanogr., 40, 1018–1034, doi:10.1175/2009JPO4265.1.
Iwamaru, H., F. Kobashi, and N. Iwasaka, 2010: Temporal variations of the winter mixed layer south of the Kuroshio Extension. J. Oceanogr., 66, 147–153, doi:10.1007/s10872-010-0012-1.
Jerlov, N. G., 1968: Optical Oceanography. Elsevier, 194 pp.
Kako, S. I., and M. Kubota, 2007: Variability of mixed layer depth in Kuroshio/Oyashio Extension region: 2005–2006. Geophys. Res. Lett., 34, L11612, doi:10.1029/2007GL030362.
Kawai, H., 1972: Hydrography of the Kuroshio Extension. Kuroshio: Its Physical Aspects, H. Stommel and K. Yoshida, Eds., University of Tokyo, 235–352.
Kobayashi, S., and Coauthors, 2015: The JRA-55 Reanalysis: General specifications and basic characteristics. J. Meteor. Soc. Japan, 93, 5–48, doi:10.2151/jmsj.2015-001.
Kouketsu, S., H. Tomita, E. Oka, S. Hosoda, T. Kobayashi, and K. Sato, 2012: The role of meso-scale eddies in mixed layer deepening and mode water formation in the western North Pacific. J. Oceanogr., 68, 63–77, doi:10.1007/s10872-011-0049-9.
Kurihara, Y., T. Sakurai, and T. Kuragano, 2006: Global daily sea surface temperature analysis using data from satellite microwave radiometer, satellite infrared radiometer and in-situ observations (in Japanese). Weather Bull., 73, s1–s18.
Kwon, Y.-O., M. A. Alexander, N. A. Bond, C. Frankignoul, H. Nakamura, B. Qiu, and L. A. Thompson, 2010: Role of the Gulf Stream and Kuroshio–Oyashio systems in large-scale atmosphere–ocean interaction: A review. J. Climate, 23, 3249–3281, doi:10.1175/2010JCLI3343.1.
Levitus, S., 1982: Climatological Atlas of the World Ocean. NOAA Prof. Paper 13, U.S. Govt. Printing Office, Washington, DC, 173 pp.
Locarnini, R. A., and Coauthors, 2013: Temperature. Vol. 1, World Ocean Atlas 2013, NOAA Atlas NESDIS 73, 40 pp. [Available online at http://data.nodc.noaa.gov/woa/WOA13/DOC/woa13_vol1.pdf.]
Masuzawa, J., 1969: Subtropical mode water. Deep-Sea Res., 16, 463–472, doi:10.1016/0011-7471(69)90034-5.
Noh, Y., and H. J. Kim, 1999: Simulation of temperature and turbulence structure of the oceanic boundary layer with the improved near-surface process. J. Geophys. Res., 104, 15 621–15 634, doi:10.1029/1999JC900068.
Ohno, Y., N. Iwasaka, F. Kobashi, and Y. Sato, 2009: Mixed layer depth climatology of the North Pacific based on Argo observations. J. Oceanogr., 65, 1–16, doi:10.1007/s10872-009-0001-4.
Oka, E., 2009: Seasonal and interannual variation of North Pacific subtropical mode water in 2003–2006. J. Oceanogr., 65, 151–164, doi:10.1007/s10872-009-0015-y.
Oka, E., and B. Qiu, 2012: Progress of North Pacific Mode Water research in the past decade. J. Oceanogr., 68, 5–20, doi:10.1007/s10872-011-0032-5.
Oka, E., L. D. Talley, and T. Suga, 2007: Temporal variability of winter mixed layer in the mid- to high-latitude North Pacific. J. Oceanogr., 63, 293–307, doi:10.1007/s10872-007-0029-2.
Oka, E., B. Qiu, S. Kouketsu, K. Uehara, and T. Suga, 2012: Decadal seesaw of the central and subtropical mode water formation associated with the Kuroshio Extension variability. J. Oceanogr., 68, 355–360, doi:10.1007/s10872-011-0098-0.
Oka, E., and Coauthors, 2014: Synoptic observation of Central Mode Water in its formation region in spring 2003. J. Oceanogr., 70, 521–534, doi:10.1007/s10872-014-0248-2.
Oka, E., and Coauthors, 2015: Decadal variability of subtropical mode water subduction and its impact on biogeochemistry. J. Oceanogr., 71, 389–400, doi:10.1007/s10872-015-0300-x.
Pak, G., Y.-H. Park, F. Vivier, Y.-O. Kwon, and K.-I. Chang, 2014: Nonstationary relationship between the East Asian winter monsoon and North Pacific Oscillation. J. Climate, 27, 8185–8204, doi:10.1175/JCLI-D-13-00500.1.
Paulson, C. A., and J. J. Simpson, 1977: Irradiative measurements in the upper ocean. J. Phys. Oceanogr., 7, 952–956, doi:10.1175/1520-0485(1977)007<0952:IMITUO>2.0.CO;2.
Qiu, B., 2000: Interannual variability of the Kuroshio Extension system and its impact on the wintertime SST field. J. Phys. Oceanogr., 30, 1486–1502, doi:10.1175/1520-0485(2000)030<1486:IVOTKE>2.0.CO;2.
Qiu, B., and T. M. Joyce, 1992: Interannual variability in the mid- and low-latitude western North Pacific. J. Phys. Oceanogr., 22, 1062–1079, doi:10.1175/1520-0485(1992)022<1062:IVITMA>2.0.CO;2.
Qiu, B., and K. A. Kelly, 1993: Upper ocean heat balance in the Kuroshio Extension region. J. Phys. Oceanogr., 23, 2027–2041, doi:10.1175/1520-0485(1993)023<2027:UOHBIT>2.0.CO;2.
Qiu, B., and S. Chen, 2005: Variability of the Kuroshio Extension jet, recirculation gyre, and mesoscale eddies on decadal timescales. J. Phys. Oceanogr., 35, 2090–2103, doi:10.1175/JPO2807.1.
Qiu, B., and S. Chen, 2006: Decadal variability in the formation of the North Pacific Subtropical Mode Water: Oceanic versus atmospheric control. J. Phys. Oceanogr., 36, 1365–1380, doi:10.1175/JPO2918.1.
Qiu, B., and S. Chen, 2010: Eddy-mean flow interaction in the decadally modulating Kuroshio Extension system. Deep-Sea Res., 57, 1098–1110, doi:10.1016/j.dsr2.2008.11.036.
Qiu, B., S. Chen, and P. Hacker, 2007: Effect of mesoscale eddies on subtropical mode water variability from the Kuroshio Extension System Study (KESS). J. Phys. Oceanogr., 37, 982–1000, doi:10.1175/JPO3097.1.
Reynolds, R. W., T. M. Smith, C. Liu, D. B. Chelton, K. S. Casey, and M. G. Schlax, 2007: Daily high-resolution-blended analyses for sea surface temperature. J. Climate, 20, 5473–5496, doi:10.1175/2007JCLI1824.1.
Seo, Y., S. Sugimoto, and K. Hanawa, 2014: Long-term variations of the Kuroshio Extension path in winter: Meridional movement and path state change. J. Climate, 27, 5929–5940, doi:10.1175/JCLI-D-13-00641.1.
Smith, T. M., R. W. Reynolds, T. C. Peterson, and J. Lawrimore, 2008: Improvements to NOAA’s historical merged land–ocean surface temperature analysis (1880–2006). J. Climate, 21, 2283–2296, doi:10.1175/2007JCLI2100.1.
Suga, T., and K. Hanawa, 1990: The mixed-layer climatology in the northwestern part of the North Pacific Subtropical Gyre and the formation area of subtropical mode water. J. Mar. Res., 48, 543–566, doi:10.1357/002224090784984669.
Suga, T., and K. Hanawa, 1995: The subtropical mode water circulation in the North Pacific. J. Phys. Oceanogr., 25, 958–970, doi:10.1175/1520-0485(1995)025<0958:TSMWCI>2.0.CO;2.
Sugimoto, S., and K. Hanawa, 2009: Decadal and interdecadal variations of the Aleutian Low activity and their relation to upper oceanic variations over the North Pacific. J. Meteor. Soc. Japan, 87, 601–614, doi:10.2151/jmsj.87.601.
Sugimoto, S., and K. Hanawa, 2010: Impact of Aleutian low activity on the STMW formation in the Kuroshio recirculation gyre region. Geophys. Res. Lett., 37, L03606, doi:10.1029/2009GL041795.
Sugimoto, S., and K. Hanawa, 2011: Roles of SST anomalies on the wintertime turbulent heat fluxes in the Kuroshio–Oyashio confluence region: Influences of warm eddies detached from the Kuroshio Extension. J. Climate, 24, 6551–6561, doi:10.1175/2011JCLI4023.1.
Sugimoto, S., and K. Hanawa, 2014: Influence of Kuroshio path variation south of Japan on formation of subtropical mode water. J. Phys. Oceanogr., 44, 1065–1077, doi:10.1175/JPO-D-13-0114.1.
Sugimoto, S., K. Hanawa, K. Narikiyo, M. Fujimori, and T. Suga, 2010: Temporal variations of the net Kuroshio transport and its relation to atmospheric variations. J. Oceanogr., 66, 611–619, doi:10.1007/s10872-010-0050-8.
Taguchi, B., H. Nakamura, M. Nonaka, and S.-P. Xie, 2009: Influences of the Kuroshio/Oyashio Extensions on air–sea heat exchanges and storm-track activity as revealed in regional atmospheric model simulations for the 2003/04 cold season. J. Climate, 22, 6536–6560, doi:10.1175/2009JCLI2910.1.
Takeuchi, E., and I. Yasuda, 2003: Wintertime shoaling of oceanic surface mixed layer. Geophys. Res. Lett., 30, 2152, doi:10.1029/2003GL018511.
Taneda, T., T. Suga, and K. Hanawa, 2000: Subtropical mode water variation in the northwestern part of the North Pacific subtropical gyre. J. Geophys. Res., 105, 19 591–19 598, doi:10.1029/2000JC900073.
Tanimoto, Y., T. Kanenari, H. Tokinaga, and S.-P. Xie, 2011: Sea level pressure minimum along the Kuroshio and its extension. J. Climate, 24, 4419–4434, doi:10.1175/2011JCLI4062.1.
Tokinaga, H., Y. Tanimoto, S.-P. Xie, T. Sampe, H. Tomita, and H. Ichikawa, 2009: Ocean frontal effects on the vertical development of clouds over the western North Pacific: In situ and satellite observations. J. Climate, 22, 4241–4260, doi:10.1175/2009JCLI2763.1.
Tomita, H., S. Kako, M. F. Cronin, and M. Kubota, 2010: Preconditioning of the wintertime mixed layer at the Kuroshio Extension Observatory. J. Geophys. Res., 115, C12053, doi:10.1029/2010JC006373.
Uehara, H., T. Suga, K. Hanawa, and N. Shikama, 2003: A role of eddies in formation and transport of North Pacific subtropical mode water. Geophys. Res. Lett., 30, 1705, doi:10.1029/2003GL017542.
Vivier, F., K. A. Kelly, and L. Thompson, 2002: Heat budget in the Kuroshio Extension region: 1993–99. J. Phys. Oceanogr., 32, 3436–3454, doi:10.1175/1520-0485(2002)032<3436:HBITKE>2.0.CO;2.
Yasuda, I., T. Tozuka, M. Noto, and S. Kouketsu, 2000: Heat balance and regime shifts of the mixed layer in the Kuroshio Extension. Prog. Oceanogr., 47, 257–278, doi:10.1016/S0079-6611(00)00038-0.
Yasuda, T., and K. Hanawa, 1997: Decadal changes in the mode waters in the midlatitude North Pacific. J. Phys. Oceanogr., 27, 858–870, doi:10.1175/1520-0485(1997)027<0858:DCITMW>2.0.CO;2.
Yasuda, T., and K. Hanawa, 1999: Composite analysis of North Pacific subtropical mode water properties with respect to the strength of the wintertime East Asian monsoon. J. Oceanogr., 55, 531–541, doi:10.1023/A:1007843525069.
Yasuda, T., and Y. Kitamura, 2003: Long-term variability of North Pacific subtropical mode water in response to spin-up of the subtropical gyre. J. Oceanogr., 59, 279–290, doi:10.1023/A:1025507725222.
Yu, L., X. Jin, and R. A. Weller, 2008: Multidecade global flux datasets from the objectively analyzed air–sea fluxes (OAFlux) project: Latent and sensible heat fluxes, ocean evaporation, and related surface meteorological variables. Woods Hole Oceanographic Institution OAFlux Project Tech. Rep. OA-2008-01, Woods Hole Oceanographic Institution, 64 pp.
Zweng, M. M., and Coauthors, 2013: Salinity. Vol. 2, World Ocean Atlas 2013, NOAA Atlas NESDIS 74, 39 pp. [Available online at http://data.nodc.noaa.gov/woa/WOA13/DOC/woa13_vol2.pdf.]