Growing emissions of anthropogenic carbon dioxide (CO2) and other greenhouse gases are causing warming, acidification, and deoxygenation of the ocean globally (IPCC 2013). To understand the physical and biogeochemical changes and variations in the ocean as well as their controlling mechanisms and impacts/feedbacks on climate and ecosystems, it is imperative that the oceanographic community continues observing the ocean for a prolonged period of time through multidisciplinary approaches. The 137°E repeat hydrographic section, which has been maintained by the Japan Meteorological Agency (JMA), marked its 50th anniversary in 2016. It traverses the Kuroshio, the subtropics, and the tropics in the western North Pacific between Japan and New Guinea (Fig. 1). There are other long-term time series stations and lines such as Station P/Line P in the eastern subarctic North Pacific (Whitney and Freeland 1999), the Hawaii Ocean Time-series program in the central subtropical North Pacific (Karl and Lukas 1996), and the Bermuda Atlantic Time-series Study in the western subtropical North Atlantic (Michaels and Knap 1996). However, none of them is comparable to the 137°E section in terms of the meridional and vertical extent and the variety of measured variables (Bingham et al. 2002).

Fig. 1
figure 1

The 137°E section (thick black line) and the associated surface and subsurface currents (solid and dashed white lines). KE Kuroshio extension, nSTCC (sSTCC) northern (southern) STCC (Kobashi et al. 2006), NEC North equatorial current, MC Mindanao current, MUC Mindanao undercurrent, ITF Indonesian through flow, NECC North Equatorial Countercurrent, NSCC North Subsurface Countercurrent, EUC Equatorial Undercurrent, NGCUC New Guinea Coastal Undercurrent. nNLM and tLM denote a nearshore non-large-meander path and a typical large-meander path of the Kuroshio, respectively (Kawabe 1995). Note that the width of currents is not considered, particularly for the NEC (see Fig. 4). Thin black contours with color indicate isobaths at an interval of 500 m, based on ETOPO2v2 (National Geophysical Data Center 2006)

The 137°E section was initiated in 1967 as part of the Cooperative Study of the Kuroshio and Adjacent Regions (CSK) supported by the Intergovernmental Oceanographic Commission, UNESCO under the leadership of Dr. Jotaro Masuzawa (Photo 1), who was a Senior Scientific Officer of the JMA in 1956–1971 and the Director-General later in 1980–1983 (Masuzawa 1967; Kuroda 2017). Enlightened by Prof. Raymond B. Montgomery during his stay at the Johns Hopkins University in 1962–1964 (Masuzawa and Nagasaka 1975), Masuzawa aimed to “establish an observational framework for ocean monitoring, at least in the western North Pacific as the Japanese area of responsibility”, “in order to investigate common variability in phenomena on as large a scale as possible” (Masuzawa 1978). Impressively, his foresight and vocation emerged long before the 1980s, during which time the world shared the recognition that it is of crucial existential importance for human society to accurately understand the changes in the ocean that play an essential role in regulating climate and ecosystems, followed by the establishment of the Global Ocean Observing System in 1991. Masuzawa chose the 137°E meridian “to observe major currents and water masses as undisturbed by islands and submarine mountains as possible” (Mazuzawa 1967; Fig. 1). It is worth noting that at that time the research community had not identified or described some of the major currents/water masses appearing in the 137°E section, such as the Subtropical Countercurrent (STCC; Uda and Hasunuma 1969), the Subtropical Mode Water (STMW; Masuzawa 1969), and the North Subsurface Countercurrent (NSCC; Tsuchiya 1975). In addition, while global warming as a result of anthropogenic CO2 emission and its partial absorption by the ocean had been surmised (Revelle and Suess 1957), the impact of the resultant carbonate system changes in seawater on marine ecosystems as well as the decline of dissolved oxygen (O2) had not been recognized.

The first historical survey of the 137°E section was conducted in January 12–24, 1967 over a distance of 3860 km between 34°N and 0°45′S in the first cruise of R/V Ryofu-maru II (Mazuzawa 1967). Since then, the observational program has not been interrupted for more than 50 years, in particular during the earlier period when little scientific achievement was made (Fig. 2). A decade after the launch, Masuzawa (1978) expressed his concern: “There has been always a doubt in my mind about what we can reveal from yearly observations, even for examining large-scale, long-term variations.” “Although the section has weathered hardships for more than 10 years, my concern that it can be all pain and no gain has not gone away. While I secretly have an irreverent idea that the value of the section should be judged in 30 years, the responsibility of planning still weighs heavily on my shoulders.” Despite his doubt, more than 100 papers using the 137°E data have been published during the past 50 years (Supplementary Table S1) with increasing frequency (Fig. 2) and with increasing significance. Among them, seven (Midorikawa et al. 2005, 2010; Nakano et al. 2007; Ren and Riser 2010; Ishii et al. 2011; Purkey and Johnson 2012; Qiu and Chen 2012) were quoted by Chapter 3 of the Intergovernmental Panel on Climate Change (IPCC) Fifth Assessment Report on ocean observation (Rhein et al. 2013). For the half-century-long contributions to the advancement of marine science through monitoring and data management, the 9th POMA Award was given from the North Pacific Marine Science Organization (PICES) to the 137°E section in 2016.

Fig. 2
figure 2

Cumulative number of papers using the 137°E section data published in all journals and books (Supplementary Table S1; circles) and those excluding three in-house journals of the JMA (Geophysical Magazine, Oceanographical Magazine, and Weather Service Bulletin; dots), plotted against year from 1967 through 2016

In this review, we summarize scientific achievements accomplished through analyses of the 137°E data. In addition to these studies, many others have used the data in combination with other observations for various targets, e.g., to investigate surface thermohaline variability in the tropical Pacific (Ando and McPhaden 1997), to estimate global air–sea fluxes of CO2 (Takahashi et al. 1997 and its updates), and to calculate nitrate transport by the Kuroshio from the East China Sea to the south of the Kuroshio (Guo et al. 2013). Modelling studies have also used the 137°E data for initial conditions and for validations, particularly in recent years (e.g., Luo and Yamagata 2003; Fujii et al. 2009; Shibano et al. 2011; Wada et al. 2011; Yara et al. 2012; Qiu et al. 2014b; Nishikawa et al. 2015; Chen et al. 2016; Tseng et al. 2016; Toyoda et al. 2017).

Section 2 introduces the observation history of the 137°E section. Section 3 describes the typical distribution of currents and water masses in the 137°E section. Sections 4 and 5 present scientific achievements in physical and biogeochemical oceanography, respectively. Section 6 demonstrates achievements in the other research fields—specifically, in terms of biological properties and marine pollution. Section 7 summarizes the achievements from the 137°E section, evaluates the extent to which Dr. Masuzawa’s aims were attained, and discusses implications for future research and directions.

Observation history

In the earlier years, the 137°E section was visited from 34°N to 1°S in the latter half of January (the time period and southward progression of winter observations have been unchanged for 50 years) with a typical station interval of 40′ north of 32°N, 1° at 8°N–32°N, and 30′ south of 8°N (Masuzawa 1967, 1968; Akamatsu and Sawara 1969; Masuzawa et al. 1970; Nagasaka and Sawara 1972; Shuto 1996; Fig. 3). At each station, temperature (T) was measured using reversing thermometers and water samples were collected by Nansen bottles (Photo 2) down to 1250 m depth, except for 4000 m at every 5° between 30°N and the equator. Depth was determined by wire angle at depths less than 100 m and by the thermometric method using a pair of pressure-protected and unprotected thermometers at depths greater than 100 m. Water samples were analyzed for a wide variety of parameters, such as salinity (S), O2, nutrients (Masuzawa et al. 1970; Sagi 1970), ammonia (Sagi 1969b), pH and total alkalinity (Akiyama et al. 1968), calcium (Sagi 1969a), trace metals such as iron and manganese, chlorophyll-a and pheophytin (Kawarada and Sano 1969), phytoplankton and zooplankton (Kawarada et al. 1968), among others. In the equatorial region, current velocity from the surface to 500 m depth relative to 600–900 m was measured by two TS-II (Roberts-type) current meters from the drifting vessel to supplement geostrophic calculations (Bessho 1995). Maritime meteorological observations were also carried out routinely, eight times a day (Hayashi et al. 1968, 1970).

Fig. 3
figure 3

Latitude of hydrographic observations in winter and summer at the 137°E section, plotted against year from 1967 through 2016. The color of dots indicates the maximum depth of standard layers at each station

In 1972, in response to the domestic enforcement of the Act on Prevention of Marine Pollution and Maritime Disaster in the previous year, the JMA started the Observation for Monitoring Background Marine Pollution targeting heavy metals (mercury and cadmium), and made the 137°E section biannual by adding summer observations around JulyFootnote 1 (Sagi et al. 1974; Kamiya 2013; Fig. 3). The observations south of 3°N had been continued until summer 1986, but were discontinued in summer 1988; subsequently, the 137°E section has been occupied between 3°N and 34°N. A conductivity–temperature–depth profiler (CTD) mounted with Niskin bottles and a bottom-mounted acoustic Doppler current profiler (ADCP) were introduced in summer 1988, enabling high-resolution profiling of T and S with respect to pressure, water sampling at precise depths, and measurements of subsurface current velocity (Shuto 1996).

As a number of the hydrographic sections of the JMA in the western North Pacific were selected as the repeat sections of the World Ocean Circulation Experiment Hydrographic Program (WHP) during 1990–1994, the JMA further improved its observation skill while coordinating these sections and carried out the 137°E section in summer 1994 as the WHP P9 one-time section, which was characterized by full-depth, high-resolution, and multi-parameter observations (Kaneko 1998; Kaneko et al. 1998). The standard observation depth of the regular 137°E section extended down to 2000 m in summer 1995 in conjunction with the launch of R/V Ryofu-maru III (Kaneko 2002; Fig. 3). The observations in the Kuroshio region were further augmented in summer 2001, resulting in the present station interval of 20′ north of 31°N, 30′ at 30°N–31°N, and 1° south of 30°N. In addition to the winter and summer observations shown in Fig. 3, those in spring and fall were conducted between 1992 and 2009 (e.g., Qiu and Chen 2010a, 2012), although the observations by R/V Keifu-maru I before 2001 had some shortcomings, such as lower-quality S data, larger station spacing, and smaller observational depth range.

To develop a globally coordinated network of sustained hydrographic sections after the completion of WHP, the Global Ocean Ship-based Hydrographic Investigations Panel (GO-SHIP), which is now part of the Global Climate Observing System and the Global Ocean Observing System, was established in 2007 with distinguished coordinating services by the International Ocean Carbon Coordination Project and the Climate Variability and Predictability Program (CLIVAR) (Hood et al. 2009). After conforming to the higher data quality standard of GO-SHIP, the JMA launched high-accuracy hydrographic observations by R/Vs Ryofu-maru III and Keifu-maru II in spring 2010, and conducted P9 revisit observations in summer 2010 and summer 2016 (Kamiya 2013) covering all the level-1 variables of GO-SHIP ( In the regular 137°E section cruises with longer station intervals and limited sampling depths, most of the level-1 variables including many of Essential Ocean Variables for physics and biogeochemistry ( have been measured. Since summer 2010, dissolved O2 profiles have been collected continuously by a rapid-response optical sensor RINKO-III (JFE Advantech Co., Ltd.) mounted on the CTD and reported after calibrating with the dissolved O2 data from the Winkler titration method (Uchida et al. 2011; Sasano et al. 2011).

The data obtained by shipboard observations of the JMA, including measurements along the 137°E section, are publicly available online ( These include the cruise summary, the station information, and data of hydrography, CTD and expendable CTD, expendable and digital bathythermograph, water sampling, greenhouse gases, petroleum hydrocarbon and heavy metals, floating pollutants, tar ball, maritime meteorology, and aerology. Among these data sets, those from cruises that meet GO-SHIP criteria have been submitted to the CLIVAR and Carbon Hydrographic Data Office ( All datasets from cruises with ocean interior high-precision carbonate system measurements (Sect. 5) have also been stored in the Pacific Ocean Interior Carbon Database (PACIFICA; Suzuki et al. 2013; and the Global Ocean Data Analysis Project Version 2 (GLODAPv2; Olsen et al. 2016; Data of underway CO2 measurements in surface seawater since the early 1980s (Sect. 5) have also been available in the Global Surface pCO2 (LDEO) database (Takahashi and Sutherland 2017; and the Surface Ocean CO2 Atlas (SOCAT; Bakker et al. 2016;

Typical distribution of currents and water masses in the 137°E section

The 137°E section crosses the western part of the North Pacific subtropical gyre and the North and South Pacific tropical gyres, and numerous zonal currents that are components of these gyres (Fig. 1). The Kuroshio, which is the western boundary current of the North Pacific subtropical gyre, usually flows eastward at 33°N along the southern coast of Japan (Fig. 4), but frequently shifted offshore as far south as 30°N forming a large-meander path from the mid-1970s to the early 1990s (Qiu and Joyce 1992; Kawabe 1995; horizontal bars at the bottom of each panel in Fig. 8). To the south of the Kuroshio exists weak westward flow associated with the recirculation gyre of the Kuroshio, which has been often called the Kuroshio Countercurrent. Farther south, the North Equatorial Current (NEC) flows westward broadly between 25° and 7°N. Its northern part north of 15°N reaches a depth of 1000 m with a subsurface current core at 200–300 m, while the southern part south of 15°N has a thickness of 300–500 m and is strongest at the surface (Masuzawa 1967). Embedded in the northern NEC between 18°N and 24°N, multiple subtropical fronts and the associated STCCs flow eastward in the shallow layer (Masuzawa 1967; Kimura 1982), corresponding to the northward shoaling of isotherms in the upper 200 m in contrast with the northward deepening in the underlying layers (Fig. 7a). As a result of the baroclinic instability associated with the reversal of the current direction between the STCC and the NEC, the STCC region exhibits large mesoscale eddy variability, being second only to the Kuroshio Extension region in the North Pacific (Qiu 1999; Fig. 5). At 2°N–7°N, the North Equatorial Countercurrent (NECC) flows eastward in the shallow layer above the thermocline (at ~ 200 m; Fig. 4). Its core shifts southward with increasing depth and connects to that of the eastward-flowing NSCC in the subthermocline layers on the long-term average (Qiu and Joyce 1992; Gouriou and Toole 1993), while the two cores are observed separately in some individual sections (Guan 1986; Bingham and Lukas 1995; Fig. 6).

Fig. 4
figure 4

(after Qiu et al. 2017)

a Distribution of eastward geostrophic velocity (cm s−1) relative to 1000 dbar in the winter 137°E section, averaged between 1967 and 1987. Number on the ordinate denotes depth in meters. Negative values are hatched (after Shuto 1996). b Distribution of eastward velocity measured by ADCP (color) and potential density (kg m−3) obtained by CTD (black contours) in the 137°E section, averaged between 2004 and 2016

Fig. 5
figure 5

© Copyright 2010 American Meteorological Society (AMS). Permission to use figures, tables, and brief excerpts from this work in scientific and educational works is hereby granted provided that the source is acknowledged. Any use of material in this work that is determined to be “fair use” under Section 107 of the U.S. Copyright Act or that satisfies the conditions specified in Section 108 of the U.S. Copyright Act (17 USC §108) does not require the AMS’s permission. Republication, systematic reproduction, posting in electronic form, such as on a website or in a searchable database, or other uses of this material, except as exempted by the above statement, requires written permission or a license from the AMS. All AMS journals and monograph publications are registered with the Copyright Clearance Center ( Questions about permission to use materials for which AMS holds the copyright can also be directed to the AMS Permissions Officer at Additional details are provided in the AMS Copyright Policy statement, available on the AMS website (

Root-mean-squared sea surface height variability in the North Pacific based on high-pass-filtered satellite altimeter data from October 1992 to April 2009. Regions where the variability exceeds 12 cm are indicated by thin black contours. White contours denote the mean sea surface height field (cm) by Niiler et al. (2003) (after Qiu and Chen 2010a).

Fig. 6
figure 6

(after Guan 1986)

Distributions of eastward velocity (cm s−1) measured by two current meters in the 137°E section in a winter and b summer of 1981

Underneath the NEC and the NECC, the recently identified North Equatorial Undercurrents (NEUCs; Qiu et al. 2013a) flow eastward at depths greater than 300 dbar at 5°N, 9°N, 13°N, and 18°N (Fig. 4b). The existence of quasi-stationary NEUCs across the tropical North Pacific was first suggested by geostrophic velocity distributions in the 137°E section referred to 2000 dbar after 1995 (Fig. 3) that, unlike those referred to shallower depths before 1995 (e.g., Fig. 4a; Qiu and Joyce 1992), captured these jets in the intermediate layer (Bo Qiu, personal communication). It was then confirmed (Qiu et al. 2013a) by the subsurface velocity distribution in the North Pacific based on T/S measurements by Argo profiling floats (Roemmich et al. 2001). The NEUCs may be a part of quasi-zonal jet-like structures, which are ubiquitous over the global ocean (e.g., Maximenko et al. 2005) and whose formation mechanism has been actively investigated (see Chen et al. 2015 for reviews and extensive references). Based on analyses of an eddy-resolving ocean general circulation model and a 1.5-layer reduced gravity model, the companion paper (Qiu et al. 2013b) also offered a formation mechanism; the NEUCs are formed by breaking of wind-forced baroclinic Rossby waves, which are generated annually at the eastern boundary of Pacific basin, due to nonlinear triad interactions, and the subsequent eddy–eddy interactions.

In the equatorial region, direct current measurements by two current meters until 1986 (Sect. 2) yielded a consistent structure (Masuzawa 1967, 1968, 1970; Akamatsu and Sawara 1969; Nagasaka and Sawara 1972; Guan 1986; Bingham and Lukas 1995; Fig. 6). The eastward-flowing Equatorial Undercurrent (EUC; Cromwell et al. 1954), which begins north of New Guinea and appears as a strong subsurface jet on the equator in the central to eastern Pacific (e.g., Wyrtki and Kilonsky 1984), is observed as a relatively weak jet at 100–300 m, 0.5°N–1.5°N in the 137°E section. Between the equator and 1°S, the monsoonal New Guinea Coastal Current (NGCC; Wyrtki 1961) flows eastward in boreal winter and westward in boreal summer along the northern coast of New Guinea at the surface, while the New Guinea Coastal Undercurrent (NGCUC; Lindstrom et al. 1987; Tsuchiya et al. 1989), which is the low-latitude western boundary current of the tropical South Pacific, flows westward at depths greater than 100 m throughout the year. A majority portion of the NGCUC turns east near and east of the 137°E section to feed the EUC (Tshuchiya et al. 1989; Gouriou and Toole 1993; Bingham and Lukas 1995), and some of the remaining portion turns east in the downstream regions to feed the NSCC and the NECC (Tsuchiya 1991; Talley et al. 2011).

These currents transport various water masses originating from the North and South Pacific to the 137°E section (Fig. 7). The subsurface S maximum centered at 150 m, 15°N is the North Pacific Tropical Water (NPTW; Cannon 1966; Tsuchiya 1968), also referred to as the North Pacific Subtropical Underwater. NPTW is formed in the central part of the North Pacific subtropical gyre where evaporation dominates precipitation and is transported westward by the NEC. The S minimum at 500–900 m north of 17°N along the isopycnal of σ θ  = 26.8 kg m−3 (θ is potential temperature and σ θ is potential density) is the North Pacific Intermediate Water (NPIW; Sverdrup et al. 1942; Reid 1965). NPIW originates in the Okhotsk Sea, enters the North Pacific subtropical gyre through the mixed water region east of Japan, and is then advected westward by the Kuroshio recirculation and the northern NEC (Talley 1993; Yasuda 2004; Fujii et al. 2013). The S minimum in the 137°E section extends southward of 17°N, beyond which its σ θ decreases from 26.8 to 26.5–26.3 kg m−3. The S minimum at 11°N–17°N at σ θ  < 26.6 kg m−3 is the Tropical Salinity Minimum (TSM; Yuan and Talley 1992), which is formed in the eastern North Pacific through merging of NPIW and the Shallow Salinity Minimum formed in the California Current region (Reid 1973) and is transported westward by the lower part of the NEC.

Fig. 7
figure 7

(after Masuzawa 1967)

Distributions of a T and b S in the 137°E section in winter 1967. Dots indicate the observation locations. Dotted lines in b indicate thermosteric anomaly contours of 175, 125, and 80 centiliters per ton (cl/t), which correspond to σt = 26.26, 26.79, and 27.26 kg m−3 (σt is potential density calculated using T instead of θ)

The S maximum centered at ~ 150 m, south of 6°N is the South Pacific Tropical Water (Cannon 1966), while the S minimum at ~ 700 m, south of 11°N at σ θ  = 27.2 kg m−3 is the Antarctic Intermediate Water (Sverdrup et al. 1942; Reid 1965). These waters formed in the South Pacific are transported westward to the eastern coast of Australia by the South Equatorial Current, then northward through the Coral and Solomon Seas by the Great Barrier Reef Undercurrent and the North Queensland Current, and finally northwestward through the Vitiaz Strait (Fig. 1) to the western equatorial Pacific by the NGCUC (Tsuchiya et al. 1989; Tsuchiya 1991; Qu and Lindstrom 2002). Subsequently, the South Pacific Tropical Water is transported eastward by the EUC and the NECC (Masuzawa 1972; Tsuchiya et al. 1989). Some Antarctic Intermediate Water is advected eastward by the NSCC, while most of it is transported northward along the eastern coast of Mindanao by the Mindanao Undercurrent (Tsuchiya 1991; Hu and Cui 1991; Fine et al. 1994).

Another feature seen in the 137°E section is the weak stratification near the top of the thermocline at 16–18 °C south of the Kuroshio (Fig. 7a). This structure was termed STMW by Masuzawa (1969) because it produces a mode in the volume distribution on the T–S diagram for the subtropical gyre. Subsequently, the moniker mode water has been applied to all similar water masses characterized by weak stratification or low potential vorticity (PV) in the world oceans (Hanawa and Talley 2001).

Scientific achievements in physical oceanography

The data from the 137°E section are best applied to investigations of temporal variations of the meridional–vertical structures and their integrated values (e.g., dynamic height, cross-sectional area, and volume transport) in relation to climate variations. During the first two decades, they were analyzed to reveal interannual variations of thermohaline structures and currents. This was mainly considered in the tropical region in relation to the El Niño-Southern Oscillation (ENSO) (Masuzawa and Nagasaka 1975; Guan 1986; Andow 1987; Saiki 1987), as well as to typhoon occurrence (Nagasaka 1981), weather in Japan (Kurihara 1984, 1985), and the Indian summer monsoon (Yasunari 1990). The 137°E data were also combined with other meridional transects to demonstrate that the decorrelation length scale of and the necessary sampling density for subsurface temperature variability in the western North Pacific is 3° (lat.)/3° (lon.) at 17.5°N–30°N and 6°/10° at 5°N–17.5°N (White et al. 1982), which rationalized expansion of a regional expendable bathythermograph observation network of the North Pacific Experiment to the entire Pacific in 1980 (White 1995).

Since Hanawa et al. (1988) examined winter mixed layers south of the Kuroshio in relation to ENSO, a number of studies have also used the 137°E data to explore the subtropics. Moreover, extension of the time series has enabled us to describe not only interannual but also decadal variability (Shuto 1996), particularly in relation to the Pacific Decadal Oscillation (PDO; Mantua et al. 1997), and now to separate decadal variability and long-term trends (Qiu and Chen 2012; Oka et al. 2017). In the following part of this section, we review scientific achievements in physical oceanography from the 137°E section, mainly over the last 30 years in terms of thermohaline structures (Sect. 4.1) and large to mesoscale currents (Sect. 4.2). In addition, the 137°E data have contributed to studies on submesoscale and microscale phenomena in recent years (Sect. 4.3). These studies on smaller scales have benefitted from the CTD and ADCP measurements repeated along the section since summer 1988, which have facilitated a reduction in the error bars of estimations and clarified temporal variability of the phenomena.

Thermohaline structures

After its discovery (Masuzawa 1969), STMW had been left unexplored for two decades, until its formation process associated with deep winter mixed layers south of the Kuroshio and the Kuroshio Extension (KE) was investigated by Hanawa (1987) and Hanawa and Hoshino (1988). Following these studies, Suga et al. (1989) examined STMW in the 137°E section in the winters spanning 1967–1987 and the summers spanning 1972–1986. By using apparent oxygen utilization (AOU), they clarified that a half (1)-year-old STMW appeared in the summer (winter) section as far south as 26°N (23°N), suggesting differences in the formation region and the subsequent southwestward advection by the Kuroshio recirculation that were verified later by using climatological data (Suga and Hanawa 1990, 1995a) and Argo profiling float data (Oka 2009).

As for year-to-year variability, Suga et al. (1989) demonstrated that less STMW was advected from the east during large-meander periods of the Kuroshio (Fig. 8a), during which anomalously warm STMW was formed locally around 137°E, that is, in an isolated recirculation gyre west of the Kuroshio large meander (Hayashi 2008; Sugimoto and Hanawa 2014). Furthermore, Suga and Hanawa (1995b) analyzed the 137°E section up to 1991 to indicate that during non-large-meander periods of the Kuroshio, STMW with lower PV and lower AOU tended to be formed in winters with stronger monsoon expression, suggesting the importance of surface cooling. On the other hand, Soga et al. (2005) examined the summer occupation of the 137°E section up to 2004 to point out that thinner and warmer STMW was formed around 2000 despite moderately strong cooling. Supporting their result, Qiu and Chen (2006) analyzed T profiles south of the KE during 1993–2004 to demonstrate that the STMW formation is largely controlled by PDO-related decadal variability of the KE, its southern recirculation gyre, and the associated eddy fieldFootnote 2 (Qiu and Chen 2005; Qiu et al. 2007, 2014a). Specifically, the STMW formation weakens (strengthens) during unstable (stable) periods of the KE (Fig. 8a), likely because high (low) eddy activity transports more (less) high PV water north of the KE to the STMW formation region to hinder (facilitate) deepening of winter mixed layers. Such importance of oceanic control on the STMW formation and subduction was confirmed by analyses of decade-long Argo profiling float data (Rainville et al. 2014; Oka et al. 2015; Cerovečki and Giglio 2016), although it does not preclude or deny the importance of atmospheric control prior to 1991. An analysis of historical T profiles over 1968–2014 indicated that the STMW formation was controlled primarily by surface cooling during the late 1970s and 1980s and by oceanic processes after ~ 1990 (Sugimoto and Kako 2016). Furthermore, an analysis of the 137°E section over winters spanning 1967–2016 demonstrated that for the last 25 years, S of STMW decreased (increased) during unstable (stable) KE periods, which can be also explained by the decadal variability in water exchange between the saline subtropics and the fresher mixed water region across the KE (Oka et al. 2017). Interestingly, such decadal S variability of STMW was almost in phase along the 137°E section against our expectation from the climatological STMW circulation (Bingham et al. 2002), suggesting the importance of horizontal eddy mixing.

Fig. 8
figure 8

Time series of cross-sectional area of a STMW, b NPTW, and c NPIW in the 137°E section during 1967–2016, calculated from optimally interpolated T/S data of Nakano et al. (2007). Here, STMW is defined as regions of PV < 2.0 × 10−10 m−1 s−1 and θ = 16–19.5 °C (e.g., Suga et al. 1989), NPTW as regions of S > 34.9 north of 7°N (e.g., Suga et al. 2000), and NPIW as regions of S < 34.2 at depths greater than 200 dbar (e.g., Shuto 1996). Dots (circles) represent winter (summer) observations. Red (green) plots in b and c indicate the area north (south) of 18°N and 25°N, respectively. Horizontal bars at the bottom of each panel denote large-meander periods of the Kuroshio. Solid (dotted) bars at the top of each panel indicate stable (unstable) periods of the KE after October 1992

Studies on low-frequency variability of NPTW based on the 137°E section data have been advanced with the development of other hydrographic data and surface flux data. Shuto (1996) analyzed NPTW in the 137°E section spanning 1967–1987, although its relation to surface freshwater fluxes in the formation region, he mentioned, could not be evaluated because of the absence of evaporation and precipitation data. By using available wind stress data, he demonstrated that on decadal time scales, the cross-sectional area of NPTW (S > 35.0) corresponded to the absolute maximum of negative wind stress curl in the subtropical region with a time lag of 0–2 years. A subsequent study using the 137°E data during 1967–1995 (Suga et al. 2000) showed that the area and S of NPTW (S > 34.9) increased remarkably in association with the 1976/77 regime shift (Nitta and Yamada 1989; Trenberth 1990; Fig. 8b). This change was explained only partly by the thermohaline forcing estimated from available evaporation and wind stress data, but was well explained by the Ekman pumping calculated from wind stress data.

A decade later, accumulation of Argo profiling float data and satellite altimeter sea surface height data as well as development of various surface flux data enabled us to fully examine variations in the NPTW formation and circulation with the underlying mechanisms. Katsura et al. (2013) analyzed Argo float data spanning 2003–2011 to demonstrate that the most saline portion of NPTW around 15°N in the 137°E section was formed in the western part of the sea surface S maximum near 25°N, 180°. In this part of the formation region, excess evaporation over precipitation seemed to be balanced by eddy diffusion, whose strength varied in association with the PDO (Qiu and Chen 2013; Sect. 4.2). As a result, the maximum S of NPTW in the 137°E during 1992–2012 lagged the PDO index by 3 years, which was considered to be the sum of 1 year for the eddy activity adjustment to the PDO forcing and 2 years for the NPTW advection. More recently, Nakano et al. (2015) analyzed the 137°E measurements spanning 1967–2011 to reveal that the decadal variability of NPTW area (S > 34.9) differed between its southern (and the most saline) part at 10°N–18°N and its northern part at 18°N–25°N, the latter of which dominated variability of the total area (Fig. 8b). The decadal variability of northern NPTW area was positively correlated with both evaporation over precipitation and Ekman pumping in the formation region west of 180°, and also with the PDO-related eddy kinetic energy in the STCC region (Qiu and Chen 2013). It might be also related to the decadal KE variability, as the area (Nakano et al. 2015; Fig. 8b) and S (Nan et al. 2015) of northern NPTW exhibited a similar decadal variability to the S of STMW mentioned above, at least after ~ 1990 (Oka et al. 2017).

Low-frequency variability of NPIW and its cause have also been investigated, commonly assuming that the distribution of NPIW in the 137°E section depends primarily on its westward advection in the subtropical gyre. Qiu and Joyce (1992) analyzed the 137°E data over 1967–1988 to demonstrate that the cross-sectional area of NPIW (S < 34.25) decreased during large-meander periods of the Kuroshio (Fig. 8c), as a result of weakening of the Kuroshio recirculation, as is the case with STMW. Shuto (1996), without considering the Kuroshio path state, indicated that decadal variability of NPIW area (S < 34.2) corresponded to that of the absolute maximum of negative wind stress curl with a time lag of 3–6 years. Nakano et al. (2005) used the 137°E data up to 2000 to exhibit that the NPIW area (S < 34.2), which had dominant time scales of 10 and 3–4 years, was correlated with the North Pacific index (Trenberth and Hurrell 1994) representing the strength of the Aleutian Low with a lag of about 11 years on decadal time scales, and with wind stress curl in the central North Pacific with a time lag of 4 years on interannual time scales. More recently, Sugimoto and Hanawa (2011) analyzed the 137°E section during 1972–2008 to demonstrate that the decadal variability of NPIW area (S < 34.2) was dominated by that of its northern part north of 25°N (Fig. 8c), which corresponded to the intensity of the Kuroshio recirculation that fluctuated as the first-mode baroclinic response to the Aleutian Low activity.

Generally speaking, it is becoming increasingly clear when reviewing recently published studies (Sugimoto and Hanawa 2011; Nakano et al. 2015; Oka et al. 2015, 2017) that decadal variability of the KE system commonly influences characteristics of STMW, NPTW, and NPIW, whose cross-sectional area in the 137°E section simultaneously increases (decreases) during stable (unstable) KE periods characterized by strengthened (weakened) recirculation gyre and low (high) eddy activity. In Fig. 8, such synchronized decadal variability seems to exist to some extent for the last 25 years since 1992, during which the KE path has been monitored by satellite altimeter measurements and the Kuroshio has mostly taken a non-large-meander path. This is an interesting relation that needs quantification in a future study because if true, the decadal variability of representative water masses in the 137°E section is controlled by oceanic processes rather than local air–sea processes, and is remotely controlled by the PDO-related atmospheric forcing in the central North Pacific.

The 137°E section has also contributed to clarifying long-term T and S changes. In the North Pacific subtropical gyre, a freshening trend has been widely observed in the main thermocline/halocline and the underlying NPIW and TSM (e.g., Lukas 2001; Wong et al. 2001; Ren and Riser 2010) as a part of global change (e.g., Wong et al. 1999; Hosoda et al. 2009; Durack and Wijffels 2010). Nakano et al. (2007) analyzed the 137°E section during 1967–2005 to report an isobaric freshening trend reaching − 0.0015 year−1 from the lower thermocline/halocline to NPIW/TSM. On the other hand, Qiu and Chen (2012) and Nan et al. (2015) used more recent time series during 1993–2009 and 1992–2009, respectively, to present a freshening trend several times larger than Nakano et al.’s (2007) from the upper thermocline/halocline to the surface layer, which suggests interdecadal changes. The latest analysis of 50-year time series during 1967–2016 (Oka et al. 2017; Fig. 9) clarified that rapid freshening began in mid-1990s and persisted for the last 20 years in the upper halocline corresponding to STMW and a lighter variety of Central Mode Water (CMW; Nakamura 1996; Suga et al. 1997; Oka and Suga 2005). Additional analyses of the repeat hydrographic section along 144°E maintained by the Japan Coast Guard demonstrated that the freshening trend originated in the winter mixed layer in the KE region, although the mechanism of surface freshening in that region remains unclear. In association with this long-term freshening, the cross-sectional area of NPTW exhibits a trend toward decrease after the 1976/77 regime shift (Nakano et al. 2015; Fig. 8b), while that of NPIW shows a trend toward increase throughout the 50-year period (Fig. 8c).

Fig. 9
figure 9

(modified after Oka et al. 2017)

Average (black contour; units °C in a) and linear trend (white contours with color; units °C year−1 in a and year−1 in b) of a θ and b S with respect to pressure in the 137°E section during 1967–2016. Linear trend of S during c 1967–1996 and d 1997–2016, otherwise following b

In contrast with the freshening subtropics, T and S in the upper tropical ocean presented a large trend toward increase centered at 12°N, 150 dbar (Fig. 9a, b). This is primarily related to southward shift of NEC, whose thermosteric effect resulted in a local sea-level rise exceeding 10 mm year−1 that is more than three times faster than the global average (Qiu and Chen 2012; Sect. 4.2).

Large to mesoscale currents

Qiu and Joyce (1992) analyzed the 137°E section in winters of 1967–1988 and summers of 1972–1988 to examine variations of various currents and water masses. They argued that for El Niño years relative to non-El Niño years, surface dynamic height in the tropical region was lower, as previously reported for summer 1972 and winter 1973 (Masuzawa and Nagasaka 1975), geostrophic volume transport of both the NEC and the NECC was larger (71 and 69 Sv versus 57 and 42 Sv; 1 Sv ≡ 106 m3 s−1), and the boundary between the NEC and the NECC and the southern boundary of the NECC both shifted southward by 1°. They also demonstrated that the volume transport of the two currents agreed well in both amplitude and phase with the Sverdrup transport estimated from wind stress data.

Two decades later, bifurcation of the NEC at the eastern coast of the Philippines was examined by using satellite altimeter sea surface height data during 1993–2009 (Qiu and Chen 2010b). Interannual variation of the bifurcation latitude, which was much larger than its seasonal variation, corresponded well with the Niño-3.4 index, being higher in El Niño years with larger volume transport of both the NEC and the NECC. The satellite altimeter data and the quarterly 137°E section data during 1993–2009 analyzed in a subsequent study (Qiu and Chen 2012) consistently demonstrated that over the 17 years the NEC bifurcation location and the boundary between the NEC and the NECC shifted southward by 2° and 1°, and the volume transport of the NEC and the NECC increased by 8 and 4 Sv, respectively (Fig. 10). Using a 1.5-layer reduced gravity model, Qiu and Chen (2012) concluded that the southward migration and the strengthening of the NEC/NECC accompanied by the rapid sea-level rise near the NEC bifurcation latitude (Sect. 4.1) were caused by the surface wind stress of the recently strengthened atmospheric Walker circulation. In support of this finding, Hu and Hu (2014) used ADCP data in the upper 200 m between 8°N and 18°N of the 137°E section during 1993–2008 to demonstrate that the volume transport of the NEC corresponded to the Niño-3.4 index with a time lag of 6 months and increased by 5 Sv over the 16 years. For longer time scales, Zhai et al. (2013) analyzed the 137°E section during 1972–2008 to indicate that the NEC transport exhibited decadal variability with a time scale of ~ 10 years in association with vertical movements of the sea surface and the permanent pycnocline in the southern part of the NEC, which was explained by decadal variability in the wind stress forcing. Such interannual to decadal variations and long-term changes of the NEC and the associated surface thermohaline structure likely impact the spawning of the Japanese eel (Anguilla japonica) near seamounts around 14°N, 143°E in the West Mariana Ridge (Fig. 1) and the subsequent larval transport to the growth habitats in East Asia (Tsukamoto 1992; Kimura et al. 1994, 2001; Kimura and Tsukamoto 2006; Tsukamoto 2006; Kim et al. 2007; Zenimoto et al. 2009; Hsu et al. 2017).

Fig. 10
figure 10

(after Qiu and Chen 2012) © Copyright 2012 American Meteorological Society (AMS). Permission to use figures, tables, and brief excerpts from this work in scientific and educational works is hereby granted provided that the source is acknowledged. Any use of material in this work that is determined to be “fair use” under Section 107 of the U.S. Copyright Act or that satisfies the conditions specified in Section 108 of the U.S. Copyright Act (17 USC §108) does not require the AMS’s permission. Republication, systematic reproduction, posting in electronic form, such as on a website or in a searchable database, or other uses of this material, except as exempted by the above statement, requires written permission or a license from the AMS. All AMS journals and monograph publications are registered with the Copyright Clearance Center ( Questions about permission to use materials for which AMS holds the copyright can also be directed to the AMS Permissions Officer at Additional details are provided in the AMS Copyright Policy statement, available on the AMS website (

Time series of a transport of the northern part of NEC circulating in the subtropical gyre, b transport of the southern part of NEC circulating in the tropical gyre, c the center latitude of the tropical gyre, and d the NECC transport based on the satellite altimeter data (left) and the quarterly 137°E section data (right). In the left (right) panels, gray line denotes the monthly (quarterly) values. In all panels, black line denotes the low-pass-filtered time series, and dashed line denotes the linear trend. Note that the increasing trend (4.1 Sv during 1993–2009) of the NECC transport detected by satellite altimeter data was not observed in the 137°E section data, presumably because of the local decreasing trend north of New Guinea

Since the STCC was discovered by Uda and Hasunuma (1969), its generation mechanism has been explored by many studies (Kobashi and Kubokawa 2012). Among them, Kubokawa and Inui (1999) and Kubokawa (1999) demonstrated using an idealized ocean general circulation model that low-PV waters with various densities subduct from the mixed layer depth front in the northern part of the subtropical gyre and pile up in the southern part of the gyre, and an eastward near-surface countercurrent is generated along the southern edge of this low-PV pool. In response to their numerical result, Aoki et al. (2002) analyzed four meridional repeat hydrographic sections including 137°E and the WHP one-time sections to examine the relation between two STCCs in the western North Pacific (Hasunuma and Yoshida 1978; Fig. 1) and mode waters (Oka and Qiu 2012). They concluded that the northern STCC was located at the southern edge of STMW, while the southern STCC was located at the southern edge of low-PV waters in a wide density range including CMW. Such correspondence between STCCs, including a third one in the eastern North Pacific, and mode waters was further demonstrated using climatology constructed from T profiles over the North Pacific (Kobashi et al. 2006).

The intensity of the STCC varies seasonally with a maximum in spring and a minimum in fall (White et al. 1978), largely because of the seasonal variation in the wind stress forcing (Takeuchi 1986). This alters the strength of baroclinic instability between the STCC and the NEC, resulting in the eddy kinetic energy in the STCC region (Fig. 5) being maximum in April–May and minimum in December–January (Qiu 1999; Kobashi and Kawamura 2002). To further reveal year-to-year variability of STCC’s mesoscale eddy field, Qiu and Chen (2010a) analyzed satellite altimeter sea surface height data over 1993–2008 and demonstrated that eddy kinetic energy was higher in 1996–1998 and 2003–2008 and lower in the other years. They also examined the quarterly 137°E data during the same period to show that the vertical shear between the STCC and the NEC was larger and more favorable for baroclinic instability in eddy-rich years than eddy-weak years. Their subsequent study (Qiu and Chen 2013) revealed that decadal variability of eddy kinetic energy was highly correlated with the PDO index with a lag of 6 months and was due to the decadal variability of the vertical shear, which was controlled by that of surface heat flux forcing rather than wind stress forcing through convergence of Ekman heat flux.

Interannual variability in the volume transport of the Kuroshio and its relation to the path variations south of Japan (Fig. 1; horizontal bars at the bottom of each panel in Fig. 8) have also been the subject of broad interest. Qiu and Joyce (1992) analyzed the 137°E section over 1967–1988 to demonstrate that the net transport of the Kuroshio, which was defined as the eastward transport of the Kuroshio minus the westward transport of the Kuroshio Countercurrent, was larger during large-meander periods (39 Sv) than non-large-meander periods (29 Sv) on average. This supported observational results in the East China Sea (Kawabe 1980; Saiki 1982) and numerical results (Chao 1984; Yoon and Yasuda 1987; Akitomo et al. 1991), while recent sensitivity experiments using a data-assimilation model (Usui et al. 2013) indicated that smaller transport is favorable for maintenance of the large meander, as pointed out by earlier observational studies (e.g., Nan’niti 1960; Nitani 1975). For longer-term variations, Sugimoto et al. (2010) analyzed the 137°E section over 1972–2007 to demonstrate that the net Kuroshio transport varied on a time scale of ~ 10 years in association with vertical movements of the permanent pycnocline in the southern part of the Kuroshio, which fluctuated as the first-mode baroclinic response to the Aleutian Low activity. They also indicated that the transport of the Kuroshio recirculation also fluctuated decadally, affecting the cross-sectional area of NPIW in the 137°E section (Sugimoto and Hanawa 2011; Sect. 4.1).

Several studies have analyzed the 137°E section to examine abyssal circulation in the Philippines Sea, where deep inflow through the channel at the junction of the Yap and West Mariana Ridges at 12°N, 139°E first enters the West Mariana Basin and then spreads northward to the Shikoku Basin and westward to the Philippines Basin (Mantyla and Reid 1983; Uehara and Taira 1990; Kawabe 1993; Fig. 1), as confirmed by the WHP full-depth observations including P9 (Kaneko et al. 1998, 2001). Sudo (1986) analyzed θ and dissolved O2 data at 4000 m depth at 15, 20, 25, and 30°N during 1976–1983 and argued that property anomalies propagated from 15°N to 30°N in 1 year at a speed of 5–6 cm s−1, while Fukasawa et al. (1995) subsequently reinterpreted the same θ time series as indicating that the anomalies propagated from 15°N to 30°N in 3 years at a speed of 2 cm s−1. Further analyses of deep CTD casts continued at every 5° and the GO-SHIP P9 revisits in 2010 and 2016 (Fig. 3) are expected to clarify long-term variations of deep water properties and the abyssal circulation in the Philippines Sea.

Smaller-scale phenomena

Turbulent diapycnal mixing and its spatiotemporal variations, which control transport of heat, salt, and nutrients and maintain the ocean stratification, have drawn increasing attention in recent years (e.g., Wunsch and Ferrari 2004; Alford et al. 2016). Jing and Wu (2010) applied the Thorpe-scale (Thorpe 1977) and fine-scale parameterization (Kunze et al. 2006) methods to available CTD profiles at the 137°E section during 1998–2007 to examine the time-mean structure and seasonal variation of diapycnal mixing. Time-mean diapycnal diffusivity (K) exceeded 10−4 m2 s−1 around 8°, 25°, and 34°N over the rough topography, and was 7.4 × 10−5 m2 s−1 when averaged in the section at depths greater than 300 dbar, being much higher than the values of O (10−5 m2 s−1) obtained by microstructure measurements in the mid-latitude ocean interior (Gregg 1987). In addition, K in the Kuroshio region fluctuated seasonally with a maximum in winter and a minimum in spring and summer and decreased in amplitude with depth, presumably because of seasonally varying surface wind stress forcing. A subsequent study using CTD profiles at the JMA’s three repeat sections including 137°E and the fine-scale parameterization method (Qiu et al. 2012) demonstrated that time-mean K at the 137°E section was high at 8°N–10°N, 22°N–24°N, 25°N–29°N, and 32°N–34°N (Fig. 11). The high K at the first two latitude bands over the rough topography was enhanced with depth, while that at 25°N–29°N over the relatively featureless topography, which was also observed at the other two sections, was vertically uniform and was attributable to parametric subharmonic instability. Lag-correlation analyses with external forcing indicated that the time-varying K at 25°N–29°N corresponded to spring-neap modulated semidiurnal tidal current with a time lag of 6 days and to surface wind forcing with no time lag.

Fig. 11
figure 11

(after Qiu et al. 2012)

a (Top) Distribution of time-mean diapycnal diffusivity (K) inferred from 57 CTD surveys at the 137°E section during 1997–2010. Gray shade denotes the bathymetry of Smith and Sandwell (1997). (Bottom) Red curve shows the roughness of bathymetry defined as the root-mean-square bathymetric height variance in a 32 km × 32 km square. b Distribution at the ASUKA section southeast of Shikoku, Japan, otherwise following a

Submesoscale processes have also drawn increasing attention in recent years (e.g., Thomas et al. 2008; McWillams 2016), and are expected to be further clarified by the onset of the Surface Water and Ocean Topography (SWOT) satellite mission planned for 2021, whose horizontal resolution (~ 15 km; Fu and Ubelmann 2014) is one order smaller than that of ongoing satellite observations (~ 150 km; Ducet et al. 2000). As satellite altimeter measurements detect signals of both balanced geostrophic flow that is dominant at meso and larger scales and unbalanced internal waves that are dominant at smaller scales, it is important for the SWOT mission to identify the length scale at which geostrophic flow loses its dominance and is overtaken by internal waves. To answer this question, Qiu et al. (2017) recently analyzed ADCP current velocity data in 33 surveys at the 137°E section to estimate the transition length scale in four latitude bands. Kinetic energy levels for internal waves were similar among the four bands, and those for geostrophic flow primarily determined the transition scale in each band. The estimated transition scale differed greatly among latitude bands, being 15 km in the eddy-abundant Kuroshio band (28°N–34°N), 50 and 80 km in the moderately unstable STCC (14°N–28°N) and NECC (3°N–9°N) bands, respectively, and 250 km in the stable NEC band (9°N–14°N).

Scientific achievements in biogeochemical oceanography

During and after the International Geophysical Year (1957–1958), it was found that CO2 in surface seawater was in general not in equilibrium with that in marine boundary air (e.g., Torii et al. 1959; Takahashi 1961; Keeling 1968). To demonstrate the spatiotemporal variations of air–sea CO2 exchange over the western North Pacific, the JMA measured the partial pressure of CO2 in marine boundary air (pCO\( _{2}^{\text{air}} \)) and that in an aliquot of air being equilibrated with surface seawater (pCO\( _{2}^{\text{sw}} \)) continuously along the 137°E section using a non-dispersive infrared gas analyzer and a showerhead-type equilibrator (Inoue 2000) for several winters from 1968 (Akiyama 1968, 1969; Masuzawa et al. 1970). After an interruption during the 1970s, the Meteorological Research Institute of the JMA, which had explored pCO2 distributions in the Pacific, Indian, and Southern Oceans from the late 1960s to the early 1970s onboard R/V Hakuho-maru of the University of Tokyo (Miyake and Sugimura 1969; Miyake et al. 1974), resumed the pCO2 measurements along the winter 137°E section for research purposes in 1981 (Fushimi 1987; Inoue et al. 1987). In the light of the importance of greenhouse gases in global warming (IPCC 1990), the JMA took over these measurements in 1990, by initiating biannual operational measurements of CO2 in surface seawater and in the air as part of the Observation for Monitoring Background Marine Pollution (Hirota et al. 1991, 1992, 1993; Fushimi et al. 1993).

Measurements during the earlier years revealed that pCO\( _{2}^{\text{sw}} \) in the subtropical region exhibits large seasonal variations being low in winter and high in summer, and that this region serves as a strong sink of CO2 on the annual average, which are now well known to the ocean carbon researchers. Along the winter 137°E section, pCO\( _{2}^{\text{sw}} \) is lowest at ~ 30°N in the subtropical region south of the Kuroshio and tends to be higher in the lower latitudes, while pCO\( _{2}^{\text{air}} \) shows smaller spatial variation, being nearly constant south of 15°N and increasing slightly northward north of 15°N (Fushimi 1987; Inoue et al. 1987; Fig. 12). Correspondingly, the difference between pCO\( _{2}^{\text{sw}} \) and pCO\( _{2}^{\text{air}} \)pCO2 = pCO\( _{2}^{\text{sw}} \)pCO\( _{2}^{\text{air}} \)) is largely negative, reaching a minimum of – 60 μatm at ~ 30°N, and increases southward to around zero or even positive in the equatorial region. In summer, by contrast, pCO\( _{2}^{\text{sw}} \) is comparable to or higher by up to 50 μatm than pCO\( _{2}^{\text{air}} \), which slightly decreases northward in the subtropical region (Hirota et al. 1991). The large seasonal variations of pCO\( _{2}^{\text{sw}} \) in the subtropics were attributed mainly to the thermodynamic effect of seasonal variation of sea surface temperature (SST) that is partly compensated for by the changes in carbon chemistry due to biological production in summer and by entrainment of CO2-rich subsurface water in winter (Inoue et al. 1987, 1995; Inoue and Sugimura 1988a; Murata and Fushimi 1996; Murata et al. 1998; Fig. 13a). The net air–sea CO2 flux (positive upward), which depends on ΔpCO2 and the gas transfer velocity expressed as a function of wind speed, fluctuates seasonally in this region between − 16 mmol m−2 day−1 in winter, associated with large negative ΔpCO2 and strong wind, and 4 mmol m−2 day−1 in summer (Inoue et al. 1995; Fig. 13b). For the annual average, the CO2 flux is at a minimum of − 8 mmol m−2 day−1 south of the Kuroshio and increases southward, changing its sign at 5°N–10°N. In the equatorial region at 137°E located in the western equatorial Pacific warm pool, the annual-mean CO2 flux range is 0.2–0.7 mmol m−2 day−1, which indicates a weak CO2 source. This source is much weaker than that in the upwelling region in the central and eastern equatorial Pacific known as the largest source in the world oceans (Tans et al. 1990).

Fig. 12
figure 12

CO2 concentration (xCO2) in surface seawater (cross, winter; plus, summer) and marine boundary air (dot, winter; circle, summer) along the 137°E section in winter and summer 1990, plotted against latitude (modified after Hirota et al. 1992). xCO2 is related to pCO2 as pCO2 = xCO2 (1 μatm − pH2O), where pH2O (μatm) is saturated water vapor pressure, which increases with temperature. As a result of the existence of pH2O, pCO2 values in μatm are 1–4% lower than xCO2 values in ppm

Fig. 13
figure 13

(after Inoue et al. 1995)

Seasonal variation of a ΔpCO2 (μatm) and b air–sea CO2 flux (mmol m−2 day−1) in the western North Pacific between 132°E and 142°E in 1990

In the equatorial region, pCO\( _{2}^{\text{sw}} \) and ΔpCO2 fluctuate interannually in relation to ENSO. The pCO2 measurements along the winter 137°E section from 1981 to 1985 revealed that in winter 1983 during the 1982/83 El Niño event, pCO\( _{2}^{\text{sw}}, \) and ΔpCO2 between 1°S and 7°N considerably increased in association with SST decrease, presumably because of shallowing of the thermocline, while those between 7°N and 17°N decreased with SST (Fushimi 1987; Inoue et al. 1987; Fig. 14). This contrasted with the central and eastern equatorial Pacific during the same period where pCO\( _{2}^{\text{sw}} \) decreased to the values almost equivalent to those of pCO\( _{2}^{\text{air}} \) as a result of the cessation of upwelling (Feely et al. 1987). Subsequent measurements repeated in the western and central Pacific during 1987–1989 using several Japanese vessels demonstrated that pCO\( _{2}^{\text{sw}} \) in the central equatorial Pacific decreased in January and February 1987 during the 1986/88 El Niño event and increased in January and February 1989 during the 1988/89 La Niña event, while pCO\( _{2}^{\text{sw}} \) at the 137°E section in the western equatorial Pacific warm pool exhibited relatively small variations, compared to those during the strong 1982/83 El Niño event (Inoue and Sugimura 1988b, 1992). Furthermore, extensive pCO2 measurements conducted over the entire equatorial Pacific during the 1990s as part of the Joint Global Ocean Flux Study showed that during the strong El Niño events, pCO\( _{2}^{\text{sw}} \) increased in the western equatorial Pacific but decreased to the near-equilibrium values in the central and eastern equatorial Pacific, resulting in the reduction of annual-mean CO2 emission from the equatorial Pacific to 0.2–0.4 Pg C year−1 that was much smaller than 0.8–1.0 Pg C year−1 during non-El Niño years (Feely et al. 2002).

Fig. 14
figure 14

(after Fushimi 1987)

Interannual variation of pCO\( _{2}^{\text{sw}} \) (left) and ΔpCO2 (right) along the 137°E section in winter during 1981–1985

In addition to its seasonal and interannual variations, a long-term trend toward pCO\( _{2}^{\text{sw}} \) increase in response to the pCO\( _{2}^{\text{air}} \) increase associated with anthropogenic CO2 emissions has been observed at 137°E. On the basis of the decade-long record from the winter 137°E sectionFootnote 3 from 1984 to 1993, Inoue et al. (1995) demonstrated that pCO\( _{2}^{\text{sw}} \) was increasing at a rate of + 1.8 ± 0.6 μatm year−1 at 15°N–30°N, which was equal to the rate of pCO\( _{2}^{\text{air}} \) increase, and + 0.5 ± 0.7 μatm year−1 at 3°N–14°N. Apart from a few studies indicating a long-term pCO\( _{2}^{\text{sw}} \) increase using two observations a decade apart (e.g., Inoue and Sugimura 1988a), this was the first study that corroborated the trend toward pCO\( _{2}^{\text{sw}} \) increase based on time-series data from shipboard observations. A decade later, Midorikawa et al. (2005) reconfirmed the persistent trend toward pCO\( _{2}^{\text{sw}} \) increase; its rate for the two decades of 1984–2003 averaged between 3°N and 34°N at the winter 137°E section was + 1.7 ± 0.2 μatm year−1 and was comparable to that of pCO\( _{2}^{\text{air}} \) (+ 1.60 ± 0.03 μatm year−1), which suggests that the western North Pacific Ocean responds rapidly to the increase of CO2 in the atmosphere.

The carbonate system in seawater is characterized by four measurable parameters: pCO2, dissolved inorganic carbon (DIC), total alkalinity (TA), and pH. In principle, data of any two of these parameters and those of T and S can be used to derive the other carbonate system parameters. The JMA has measured DIC from discrete water samples at the 137°E section using a CO2 extraction-coulometry method (Ishii et al. 1998) in summer 1994 (WHP P9), winter 1997, and summer 2000, and every cruise since 2003. On the basis of the data from five meridional transects in the western North Pacific during 1994–1997 including the above two at 137°E, Ishii et al. (2001) demonstrated that surface TA normalized at S = 35 (nTA), which was calculated from pCO\( _{2}^{\text{sw}} \), surface normalized DIC (nDIC), SST, and sea surface salinity (SSS), was almost invariable in space and time. They also indicated that the seasonal variations in pCO\( _{2}^{\text{sw}} \) were ascribed mainly to those of SST and nDIC; specifically, surface nDIC decreased (increased) from winter (summer) to summer (winter), compensating for ~ 30% of pCO\( _{2}^{\text{sw}} \) increase (decrease) due to SST increase (decrease). Furthermore, Ishii et al. (2001) estimated the monthly surface nDIC distribution in the western subtropical gyre from pCO\( _{2}^{\text{sw}} \) and SST of Inoue et al. (1995) (Fig. 13a) assuming constant nTA. They diagnosed that the nDIC decrease from winter to summer was due to biological production and was partly compensated for by the air-to-sea CO2 flux, while its increase from summer to winter was explained by entrainment of subsurface water associated with mixed layer deepening, and that the effect of the horizontal DIC transport appears minor. Their analysis was substantially applied to the measurements along the winter 137°E section over 1983–2003 by Midorikawa et al. (2006). On interannual time scales, the effects of SST and nDIC variations on pCO\( _{2}^{\text{sw}} \) variations in the subtropical region have been compensating for each other through entrainment, thereby allowing the long-term trend toward pCO\( _{2}^{\text{sw}} \) increase to be observed distinctly. This was also the case for the warm pool of the equatorial region where SST tends to be higher and DIC tends to be lower when stratification was enhanced by the formation of barrier layers (Lukas and Lindstrom 1991), although the relationship between SST change and DIC change was not so robust there (Ishii et al. 2009).

The JMA also started high-precision measurements of pH in 2003 (Saito et al. 2008) and TA in 2010 using indicator dye-spectrophotometry methods. Data of surface pH during 2003–2008 calculated from pCO\( _{2}^{\text{sw}} \), SST, and SSS assuming a constant nTA agreed well with those measured spectrophotometrically when using dissociation constants of carbonic acid given by Lueker et al. (2000). This allowed the pH time series to be extended back to 1983 on the basis of pCO\( _{2}^{\text{sw}} \) measurements (Midorikawa et al. 2010). The calculated surface pH at 3°N–33°N of the 137°E section in winter (summer) during 1983–2007 (1987–2007) exhibited a long-term trend toward decrease at a rate of − 0.0018 ± 0.0002 year−1 (− 0.0013 ± 0.0005 year−1) on average at ambient SST (Fig. 15) and that of − 0.0015 ± 0.0003 year−1 (− 0.0014 ± 0.0004 year−1) when normalized to 25 °C. These trends were comparable to those observed at the time-series stations in the central subtropical North Pacific near Hawaii and those in the subtropics of the North Atlantic (Bates 2007; González-Dávila et al. 2007; Dore et al. 2009). A similar pH trend of − 0.0020 ± 0.0007 year−1, together with a trend of − 0.012 ± 0.005 year−1 for the aragonite saturation state, was determined for 1994–2008 at the northernmost stations (33°40′N–34°N) of the 137°E section located off the southern coast of Japan and inshore of the Kuroshio (Ishii et al. 2011).

Fig. 15
figure 15

(after Midorikawa et al. 2010)

Time series of surface pH at six latitudes at the 137°E section in a winter and b summer. Red circles, 3°N; violet triangles, 10°N; orange diamonds, 15°N; green squares, 20°N; light blue triangles, 25°N; and blue circles, 30°N

Midorikawa et al. (2011) subsequently suggested the acceleration of ocean acidification after the 1980s by demonstrating that, when the trend toward surface pH decrease determined at the winter 137°E section over 1983–2008 was extrapolated back to 1969–1970, the extrapolated pH value at each latitude was higher than that estimated from the pCO\( _{2}^{\text{sw}} \) measurements made in the winter of these years along 138°E and 158°E (Miyake et al. 1974; Inoue et al. 1999). On the other hand, Midorikawa et al. (2012) pointed out that in spite of the acceleration of pCO\( _{2}^{\text{air}} \) increase over the long term, the mean rate of pCO\( _{2}^{\text{sw}} \) increase at the winter 137°E section was lower for 1999–2009 than for the earlier period of 1984–1997 at most latitudes between 3°N and 33°N. This was particularly the case at 10°N–20°N where the southward migration of NEC (Qiu and Chen 2012; Sect. 4.2) possibly led to thickening of the warm and CO2-poor surface water and the reduction of entrainment of CO2-rich subsurface water. Thus, the rate of pCO\( _{2}^{\text{sw}} \) increase was likely to be decadally variable, and its mechanism needs to be pursued in a future study.

As a result of subduction from the winter mixed layer with increasing pCO\( _{2}^{\text{sw}} \), DIC in the ocean interior has been increasing. In the latest issue of the annual Marine Diagnosis Report (Japan Meteorological Agency 2017), it is shown that nDIC at the 137°E section exhibits a long-term trend toward increase on isopycnals down to σ θ  = 27.2 kg m−3 for the last two decades (Fig. 16). In the water column between the sea surface and σ θ  = 27.5 kg m−3 at a depth of ~ 1300 m, anthropogenic CO2 has been accumulated at a mean rate of 4.1–12.3 gC m−2 year−1 over 1994–2016 (Fig. 17). The accumulation rate was highest at 30°N, where STMW that transports CO2 from the surface layer to the interior is thickest. In addition to this long-term change, PDO-related decadal variability of STMW formation and subduction (Sect. 4.1) influences biogeochemical structure in the upper part of the permanent thermocline in the western subtropical gyre. By analyzing biogeochemical data at 25°N in the 137°E section, Oka et al. (2015) detected consistent changes during the stable KE period after 2010 that pH and aragonite saturation state (AOU, nitrate, and nDIC) in the STMW layer increased (decreased), particularly against long-term trends for pH, aragonite saturation state, and nDIC, in association with the increasing advection of STMW from the east (Fig. 18). This result indicates a new mechanism by which climate variability impacts biogeochemical properties in the ocean’s interior, and potentially also impacts the surface ocean acidification trend along with biological productivity.

Fig. 16
figure 16

(after Japan Meteorological Agency 2017)

Time series of preformed nDIC on the isopycnals of σ θ  = 25.0 (red), 26.0 (orange), 26.8 (green), 27.2 (blue), and 27.5 kg m−3 (purple) at a 10°N, b 20°N, and c 30°N of the 137°E section. Lines indicate the least-square fit. Values in parentheses denote the approximate depth of the isopycnals

Fig. 17
figure 17

(Replotted from Japan Meteorological Agency (2017)

Increasing rates of water column inventories of CO2 between the sea surface and σ θ  = 27.5 kg m−3 at 10°N–30°N of the 137°E section. Vertical bars indicate the 95% confidence interval

Fig. 18
figure 18

(after Oka et al. 2015)

Time series of a AOU, b salinity-normalized dissolved inorganic carbon (nDIC), c preformed nDIC, d pH, e aragonite saturation state index (Ωarag), and f nitrate (NO3) concentration on the isopycnal of σ θ  = 25.0 kg m−3, based on observations at 25°N of the 137°E section

Along with the ocean acidification, decrease of O2 concentration in seawater, i.e., deoxygenation, associated with the ocean warming is another potential threat to marine ecosystems. Takatani et al. (2012) analyzed the 137°E section over 1967–2010 to investigate the trend of dissolved O2 on isopycnals. They found marked decreases of O2 at 20°N–25°N in the subtropical gyre after the mid-1980s, with a rate of − 0.28 ± 0.08 μmol kg−1 year−1 on σ θ  = 25.5 kg m−3 (corresponding to the denser limit of STMW), − 0.36 ± 0.08 μmol kg−1 year−1 on 26.8 kg m−3 (NPIW), and − 0.23 ± 0.04 μmol kg−1 year−1 on 27.3 kg m−3 (O2 minimum layer). They demonstrated that the O2 decrease in the upper pycnocline (< 26.0 kg m−3) was due mainly to the isopycnal deepening and the decline of O2 solubility associated with upper ocean warming, while that in the layers below (> 26.0 kg m−3) was attributable primarily to AOU changes; specifically, the O2 decrease in the NPIW originated from the NPIW formation region, and that in the O2 minimum layer was explained by intensification of westward transport of low O2 water from the eastern North Pacific.

Scientific achievements in other research areas

Biological oceanography

Phytoplankton and zooplankton sampling along the 137°E section was conducted from 1967 through 2006. Phytoplankton in water samples from Nansen or Niskin bottles, mainly diatoms, were concentrated by decanting and centrifugalizing, and were identified and counted using a light microscope (Kawarada et al. 1968). Macrozooplankton was collected by vertically towing a 0.33-mm-mesh NORPAC net with a 45-cm-diameter aperture and a 180-cm length from 150 m depth to the sea surface, and chaetognaths and some copepods were identified and counted using a stereomicroscope (Kawarada et al. 1968; Kitou 1974; Kawashima and Nagai 1990; Nagai et al. 2015). In the CSK cruise in January–March 1967 including the first 137°E section (Masuzawa 1967), a new species of calanoid Copepoda, Calanoides philippinensis (Kitou and Tanaka 1969), was described.

Along the 137°E section, both diatoms and chaetognaths were abundant in the Kuroshio and the equatorial regions, while they were scarce in the NEC (Kawarada et al. 1968). The latest analysis for winter observations during 1967–1995 (Nagai et al. 2015) recorded 26 chaetognath species, of which 22 were epipelagic and 4 were mesopelagic, and classified 17 common epipelagic species into six groups according to their distributions: Zonosagitta nagae in the Japanese coast type, Mesosagitta minima in the Japanese coast-Kuroshio type, Krohnitta subtilis, Pseudosagitta lyra, Sagitta bipunctata, and Serratosagitta pseudoserratodentata in the Kuroshio-subtropical type, Aidanosagitta neglecta, Ferosagitta ferox, Fe. robusta, and Flaccisagitta enflata in the NEC-NECC type, A. regularis, K. pacifica, Z. bedoti, and Z. pulchra in the NECC type, and Fl. hexaptera, Pterosagitta draco, and Se. pacifica in the bimodal type characterized by a bimodal distribution with the minimum in the NEC.

As with physical and biogeochemical parameters, biological properties have exhibited long-term variations and trends. Sugimoto and Tadokoro (1998) analyzed the 137°E section during 1970–1992 to demonstrate that the chlorophyll-a (Chl-a) concentration in the upper 200 m decreased drastically around 1980 in both winter and summer in all regions at 30°N–32°N, 21°N–23°N, and 14°N–16°N corresponding to the Kuroshio and its counter current, the STCC, and the NEC, respectively, and that the nighttime mesozooplankton biomass in the upper 150 m also decreased simultaneously in the first two regions. On the basis of the longer 137°E time series during 1971–2000, Watanabe et al. (2005) demonstrated that the mixed layer σ θ as well as phosphate and nitrate averaged vertically in the mixed layer exhibited a trend toward decrease in both winter and summer and in all areas at 30°N–34°N, 15°N–30°N, and 3°N–15°N, suggesting a decrease in the supply of nutrients from the subsurface (Fig. 19). Consistently, the water column Chl-a and net community production also had a trend toward decrease in all areas (Fig. 20). Furthermore, phosphate, Chl-a, and net community production in all seasons and areas fluctuated bidecadally with a period of ~ 21 years, lagging behind the PDO index by 2 years (Figs. 19, 20). Vertical structure of biological properties is also changing; during 1972–1997, the proportion of Chl-a existing in the upper 75 m of water column (the depth of subsurface Chl-a maximum) in summer has decreased (increased) at a rate of − 0.4% year−1 (0.4 m year−1) in the subtropical region of the 137°E section (Ishida et al. 2009).

Fig. 19
figure 19

(after Watanabe et al. 2005)

Time series of the mixed layer σ θ (a, b) and phosphate (PO4; c, d) and nitrate (NO3; e, f) averaged in the mixed layer at 30°N–34°N (black circles), 15°N–30°N (blue triangles), and 3°N–15°N (red squares) of the 137°E section in winter (a, c, e) and summer (b, d, f). Vertical bars denote standard errors for 3-year running mean. Solid lines and dashed curves represent the linear regression line and the nonlinear fitting curve estimated by the Fourier sine expansion, respectively

Fig. 20
figure 20

(after Watanabe et al. 2005)

Time series of the water column Chl-a (CHL) in a winter and b summer and c net community production (NCP) at the 137°E section, otherwise following Fig. 19. d PDO index in winter

Marine pollution

To evaluate over the global ocean the degree of petroleum contamination at the sea surface, the Marine Pollution (Petroleum) Monitoring Pilot Project (MaPPMoPP) was implemented by the Intergovernmental Oceanographic Commission and the World Meteorological Organization under the framework of the Integrated Global Ocean Station System in 1975–1980. In response to the request from the MaPPMoPP, the JMA started visual observations of floating pollutants such as floating plastics (Suzuoki and Shirakawa 1979) and oil slicks in 1976 and measurements of floating particulate petroleum residues (tar balls; Sano et al. 1979) and dissolved/dispersed petroleum hydrocarbons (Shigehara et al. 1979) in 1977 in the western North Pacific including the 137°E section, as part of the Observation for Monitoring Background Marine Pollution (Ogawa and Takatani 1998; Takatani et al. 1999; Kamiya 2013). Observations during the earlier years demonstrated that tar balls, collected by Neuston net towed at the sea surface, and petroleum hydrocarbons, extracted from surface water using carbon tetrachloride and determined by fluorescence spectrophotometry, were abundant in the Kuroshio and its countercurrent, in the STCC, and north of New Guinea, while they were scarce in the NEC (Suzuoki and Matsuzaki 1983; Takatani et al. 1986). In addition, tar balls were sticky and shiny in the Kuroshio and became crumbly and dull to the south and the east, implying that they had been weathered while being transported from their likely source in the western margin of the western North Pacific where the traffic of oil tankers was busy. At 137°E, tar balls decreased drastically around 1980, and petroleum hydrocarbons have gradually decreased with interannual fluctuations, owing to the restriction of oil discharges from ships by Annex I of the International Convention for the Prevention of Pollution from Ships (MARPOL) 73/78 that entered into force in 1983 (Takatani et al. 1986, 1999; Japan Meteorological Agency 2013; Fig. 21). On the other hand, floating plastics do not exhibit a clear long-term trend, despite the ban of dumping plastics by Annex V of the MARPOL 73/78 Convention that came into force in 1988.

Fig. 21
figure 21

Replotted from Japan Meteorological Agency (2013) with additional data in 2013–2016 (Japan Meteorological Agency 2017)

Time series of annual-mean concentrations of a tar balls, b petroleum hydrocarbons as chrysene equivalent, and c floating plastics at the 137°E section.

Photo 1
figure 22

(after Nakano 2016)

Dr. Jotaro Masuzawa (1922–2000)

Photo 2
figure 23

(Japan Meteorological Agency, personal communication)

Water sampling from Nansen bottles and reading reversing thermometers onboard R/V Ryofu-maru II (around 1969)


Hydrographic observations along the 137°E section have been repeated by the JMA using R/Vs Ryofu-maru and Keifu-maru biannually in winter since 1967 and in summer since 1972. These operational observations that have persisted for more than 50 years were initiated as part of the CSK (Masuzawa 1967), and have been maintained with an aim to detect long-term variations of oceanic conditions in the western North Pacific connected with climate change and to monitor background marine pollution (Shuto 1996) and more recently to evaluate changes in the carbon cycle and the mechanistic drivers (Nakano 2013). The data collected along 137°E have been analyzed not only by the JMA officers but also by scientists from various countries to quantify and understand seasonal to decadal variations and long-term changes of thermohaline structures (Sect. 4.1), large to mesoscale currents (Sect. 4.2), biogeochemical and biological properties (Sects. 5, 6.1), and marine pollutants (Sect. 6.2). Despite the concerns expressed by Dr. Masuzawa during the early years regarding its sampling frequency (Sect. 1), the 137°E section has been quite useful because it captures zonal integration of westward propagating Rossby signals during the course of which seasonal and shorter-term variations are cancelled out (Qiu and Chen 2010b).

When the 137°E section was initiated 50 years ago, shipboard observations were the main tool of ocean research. Since then, significant development has been made with other observing platforms and tools, including satellites, Argo, various datasets of the ocean interior structure and surface fluxes, and numerical models. The great expansion of observing platforms for oceanography over the last several decades has in fact enhanced and enriched the scientific opportunities afforded by the 137°E section itself and thereby the scientific value of the section, as new generations of researchers are able to return to the historical measurement record with new processes and time scale questions. This is reflected in the accelerated increase in scientific publications that incorporate measurements from 137°E (Fig. 2). The continued extension and maintenance of the time series has also provided unique opportunities for evaluating variability and transients on decadal and longer time scales over which the ocean plays a dominant role in the climate system. Furthermore, the introduction of CTD and ADCP measurements in the late 1980s opened the door to explorations of submesoscale and microscale phenomena (Sect. 4.3). Thus, beyond Dr. Masuzawa’s expectation, the 137°E section has contributed to understanding of variability on a wide range of spatiotemporal scales, and has developed into an irreplaceable platform for integrated, multidisciplinary ocean observations.

Recent findings on decadal variability of water masses and currents related to the PDO (Sects. 4.1, 4.2) and on smaller-scale phenomena (Sect. 4.3) strongly suggest that the 137°E section continues to be an essential component of a multi-platform observing system for physical oceanography. In other research areas, it is obviously a “gold mine”. For example, the measurements along 137°E offer a unique opportunity for linking the more well-studied surface variations on ocean biogeochemistry with subsurface variability (Sect. 5). In addition, although the duration of biogeochemical and biological measurements is relatively short compared to physical oceanographic measurements, the mechanistic understanding of scales of variability inferred from the 137°E section will be important in the interpretation of repeat hydrographic measurements, where there are known issues with aliasing of high-frequency variability in the decadal sampling strategy. Thus, analyses of continued measurements along 137°E will be of value not only for clarifying long-term variations over the local region but also for clarifying scales of variability more generally. Deep measurements including the WHP/GO-SHIP P9 transects in 1994, 2010, and 2016 also need to be analyzed to examine decadal variability in the abyssal Philippines Sea. It is worth noting, finally, that the JMA has also maintained other repeat hydrographic sections, including the one along 165°E corresponding to the WHP/GO-SHIP P13 section (e.g., Kawabe and Taira 1998; Midorikawa et al. 2002; Oka and Suga 2005; Kouketsu et al. 2013; Qiu et al. 2012; Sasano et al. 2015). Integrated analyses of these sections will lead to comprehensive understanding of common, large-scale variability in the western North Pacific that Dr. Masuzawa targeted 50 years ago.