Surface water hydrography of the Kuroshio Extension during the Pliocene–Pleistocene climate transition

Marine Micropaleontology 101 (2013) 106–114

Contents lists available at SciVerse ScienceDirect

Marine Micropaleontology journal homepage: www.elsevier.com/locate/marmicro

Surface water hydrography of the Kuroshio Extension during the Pliocene–Pleistocene climate transition Nicholas L. Venti ⁎, Katharina Billups University of Delaware, School of Marine Science and Policy, 700 Pilottown Road, Lewes, DE 19958, United States

a r t i c l e

i n f o

Article history: Received 29 February 2012 Received in revised form 10 January 2013 Accepted 9 February 2013 Keywords: Pliocene Pleistocene North Pacific Kuroshio Extension Planktic foraminifera Globigerinoides ruber Oxygen isotopes Sea surface hydrography Orbital-scale ODP Site 1208

a b s t r a c t The Pliocene–Pleistocene climate transition offers an opportunity to study the effect of glaciation on the ocean– climate system. We present a Globigerinoides ruber δ18O record from Ocean Drilling Program Site 1208 (Kuroshio Current Extension; KCE). This exclusively (sub)tropical foraminifer, a summer/fall mixed-layer dweller at the KCE, affords the first long (3.0 Ma to 1.8 Ma) orbital-scale (2.5-kyr time step) account of the sea surface in this area. The section's temperature-corrected benthic foraminiferal δ18O record constrains global changes in ice volume, yielding a Δδ18O record that primarily reflects summer/fall KCE hydrography (temperature and salinity). A 0.3‰ decrease in Δδ18O values at 2.7 Ma coincides with the onset of Northern Hemisphere glaciation, indicating as much as 1.5 °C warming during the summer/fall to suggest that the subtropical North Pacific sea surface provided heat and moisture for expanding ice sheets. On the orbital scale, the 41-kyr cycle that dominates high-latitude climate is absent from the Δδ18O record, indicating a stable surface water hydrographic regime on this time scale. Rather, the Δδ18O record varies at and is coherent with the 19-kyr precessional component of the regional insolation curve, supporting a direct response to subtropical insolation and insensitivity to extra-regional forcing factors, such as ice sheets. © 2013 Elsevier B.V. All rights reserved.

1. Introduction The early Pleistocene offers an attractive laboratory for understanding natural variability of the ocean–atmosphere system in two important ways. First, it was colder than the preceding Late Pliocene, this later interval's climate characterized by the 2.7-Ma advent of widespread Northern Hemisphere glaciation (NHG; Jansen and Sjøholm, 1991; Haug et al., 1995; Balco et al., 2005; Harris, 2005; Lisiecki and Raymo, 2005). Second, development of an east–west sea surface temperature (SST) gradient in the early Pleistocene (~2 Ma), which characterizes the modern equatorial Pacific, marks the end of perennial El Niño-like conditions (Cannariato and Ravelo, 1997; Chaisson and Ravelo, 2000; Wara et al., 2005; Lawrence et al., 2006; Ravelo et al., 2006, 2007; Etourneau et al., 2010). The Kuroshio Current system, the western boundary current of the North Pacific subtropical gyre, is an integral part of the North Pacific climate system. This warm-water jet efficiently transports heat from the western Pacific warm pool to the cool mid-latitude atmosphere. Changes in meridional heat transport potentially explain obliquity's dominant pacing of ice volume during this time (Philander and Fedorov, 2003; Raymo and Nisancioglu, 2003), perhaps providing

⁎ Corresponding author. Tel.: +1 413 687 3515. E-mail address: [email protected] (N.L. Venti). 0377-8398/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.marmicro.2013.02.004

moisture for the formation of the Northern Hemisphere (NH) ice sheets (e.g., Haug et al., 2005). Furthermore, changes in heat transport in the Kuroshio Current system may have been closely related to the cooling in the eastern equatorial Pacific and the end of persistent El Niño-like conditions during the early Pleistocene (Philander and Fedorov, 2003). Here, we explore the applicability of a planktic foraminiferal δ18O record from Ocean Drilling Program (ODP) Site 1208, located at the Kuroshio Current Extension (KCE; Fig. 1), to reconstruct surface-water hydrography spanning the Plio/Pleistocene climate transition. The planktic foraminifer, Globigerinoides (Gs.) ruber, represents the first long surficial record (3.00–1.76 Ma) from the region to span the Pliocene– Pleistocene boundary at the orbital timescale (2.5-kyr time step). We use the section's benthic foraminiferal δ18O record, smoothed to minimize deep-water temperature changes (after Sosdian and Rosenthal, 2009), as a measure of global ice-volume changes. By subtracting it from the Gs. ruber δ18O record we obtain a Δδ18O record that primarily reflects surface-water hydrography (temperature and salinity). We argue that the residual Δδ18O record is primarily a surface water temperature signal. This is because the evaporation/precipitation balance, which affects surface water δ18O values, remains relatively constant throughout the year in the modern ocean (e.g., salinity varies only between 34.4 and 34.5, Fig. 2b). Spatially, too, regional surfacewater δ18O values are relatively constant across a broad area of the northwestern Pacific (Schmidt, 1999; Bigg and Rohling, 2000). Thus,

N.L. Venti, K. Billups / Marine Micropaleontology 101 (2013) 106–114

107

50°N

Sea of Okhotsk

Subarctic water mass

nt

re

45°N

hio

as Oy

Sea of Japan

40°N

nt

rre

30°N

Northern Branch

Sendai

Tokyo

35°N

r Cu

u io C

KCE

1208

Main Branch

Recirculation Shatsky Rise

Gyre

h ros

Ku

Subtropical water mass

25°N

20°N 130°E

140°E

150°E

160°E

170°E

180°

Summer (July-September) sea surface temperature (°C) Fig. 1. ODP Site 1208 on Shatsky Rise (Shipboard Scientific Party, 2002). Surface currents are sketched after Mizuno and White (1983), Sainz-Trapaga et al. (2001), Qiu (2002), and Yasuda (2003). The base map shows summer (July–September) sea surface temperatures averaged in a 1° × 1° grid (Locarnini et al., 2010) generated with the web application at http://iridl.ldeo.columbia.edu/SOURCES/.NOAA/.NODC/.WOA09/.

we do not anticipate large changes in the hydrographic regime to have introduced large changes in surface water δ18O values. Specifically, we propose that these temperature changes approximate conditions during the summer/fall season. Plankton tows indicate that Gs. ruber is restricted to waters warmer than 15 °C and prefers temperatures of >20 °C (Bradshaw, 1959), i.e., the mixed layer in the northwest Pacific (Kuroyanagi and Kawahata, 2004; Fig. 2a). Regional sediment-trap studies confirm that this species inhabits the KCE predominantly during summer and fall (Eguchi et al., 2003; Mohiuddin et al., 2004). Thus, the Gs. ruber Δδ 18O record resolves the response of summer/fall mixed-layer hydrography to the growth of NHG during the Plio/Pleistocene climate transition.

a

2. Regional setting The warm Kuroshio Current system transports a large volume of water (as much as 45 Sverdrup; Sv), with an additional 90 Sv in the associated recirculation gyre (Wijffels et al., 1998), forming the northwestern boundary of the North Pacific anticyclonic subtropical gyre (Reid, 1997; Fig. 1). The northeastward-flowing Kuroshio Current converges with the southwestward-flowing subarctic Oyashio Current in the offshore region east of Japan, northeast of Tokyo (35.7°N) and southeast of Sendai (38.1°N; Yasuda, 2003). The high kinetic energy of mesoscale eddies in this region introduces meanders into the Kuroshio (Wyrtki et al., 1976) as it continues east between 33°N and 37°N as the

b Gs. ruber

Fig. 2. Temperature (a) (Levitus and Boyer, 1994) and salinity (b) (Levitus et al., 1994) versus depth at Site 1208 during winter (January, February, and March) and summer (July, August, and September). The shaded area in panel (a) indicates Globigerinoides ruber's preferred habitat.

108

N.L. Venti, K. Billups / Marine Micropaleontology 101 (2013) 106–114

Kuroshio Current Extension (KCE) in the northwestern Pacific (Mizuno and White, 1983). Because the anticyclonic gyre extends to depths great enough to enclose the Shatsky Rise at 158°E (at least 4000 m; Reid, 1997; Qiu, 2002), the bathymetric structure bifurcates the KCE (Mizuno and White, 1983). One quarter of the KCE's volume flows northeast around the central and northern highs of Shatsky Rise to 38–40°N before continuing east; the remainder continues east between the central and southern highs of Shatsky Rise (35°N; Mizuno and White, 1983; Sainz-Trapaga et al., 2001). Cool mid-latitude westerly winds enhanced by the cold, dry, winter East Asian monsoon efficiently remove heat (>100 W/m2 annually) from the Kuroshio (Da Silva et al., 1994). Due to this extensive heat loss, KCE sea surface temperatures (SSTs) decrease from 24 °C in the summer to 14 °C in the winter, nearly erasing the thermocline (Levitus and Boyer, 1994; Locarnini et al., 2010; Fig. 2a). Increased evaporation accompanies the severe winter heat loss such that sea surface salinity increases slightly from 34.4 in summer to 34.5 in winter (Levitus et al., 1994; Antonov et al., 2010; Fig. 2b). The seasonal salinity change is dampened by winter storms that bring excessive rainfall to the KCE (Xie and Arkin, 1997). The mean latitude, volume, and intensity of the geostrophic current, however, do not undergo regular seasonal changes (e.g., Mizuno and White, 1983). This is because the Kuroshio Current system is fundamentally a return flow, compensating within the ocean's physical limits for the wind-driven subtropical circulation (Qiu, 2002). As such, variations in KCE flow are determined primarily by wind stress curl (e.g., Yasuda et al., 1985; Miller et al., 1998; Deser et al., 1999; Sainz-Trapaga et al., 2001; Qiu, 2002) and sea surface height in the subtropics (e.g., Imawaki et al., 2001). Though seasonal cycles are too short to affect these controls, decadal-scale variability characterizes KCE flow and SSTs. Because El Niño–Southern Oscillation involves changes in the Walker circulation and thus wind stress curl, KCE flow increases toward the La Niña phase (Sainz-Trapaga et al., 2001). KCE SSTs, on the other hand, correspond to the Pacific Decadal Oscillation (e.g., Latif and Barnett, 1996; Zhong and Liu, 2009). The relationship between the El Niño–Southern Oscillation and the Pacific Decadal Oscillation is less clear (Zhang et al., 1997; Newman et al., 2003; Schneider and Cornulle, 2005). Nevertheless, changes in the accumulation of pollen (e.g., Kawahata and Ohshima, 2002), and the diversity of foraminifer assemblages (Venti, 2006) suggest subtle shifts in the strength and position of the KCE on timescales of 104–10 6 years. 3. Material and methods ODP Site 1208 (36.1°N, 158.5°E, 3346 m water depth) on the central high of Shatsky Rise offers an ideal location to reconstruct surface-water hydrography in the KCE (Fig. 1). Cored in a carbonate-rich drift deposit (Shipboard Scientific Party, 2002), the continuous Pliocene–Pleistocene section offers high sedimentation rates averaging 4 cm/kyr (Venti and Billups, 2012). The section's benthic foraminiferal δ18O record provides orbital-scale age control via the Lisiecki and Raymo (2005) global δ18O stack and indexes glacial conditions (Venti and Billups, 2012). Continuous presence of dissolution-susceptible foraminiferal tests, such as Globigerina bulloides and Gs. ruber, (e.g. Vincent and Berger, 1981), throughout the Pliocene–Pleistocene section suggest relatively good carbonate preservation (Venti, 2006). Orbital-scale surface ocean hydrography is reconstructed by sampling 2–15 tests of white sensu lato Gs. ruber (Wang, 2000), the dominant variety at this site, to minimize morphology-related δ18O variability (Kawahata, 2005). Tests were selected from the 212–355-μm sediment fraction to minimize size-related stable-isotope variations. This selection strategy should reflect the warmest waters available at this location while minimizing δ18O variability among individual tests because it avoids juveniles, which largely account for the species' occurrence in cooler upper thermocline waters (Bradshaw, 1959), and because forms associated exclusively with tropical SSTs—pink and

sensu stricto—are absent and very rare, respectively, in the Site 1208 section. Sampling intervals follow continuous 10-cm spacing between 86.44 mbsf (1.760 Ma) and 135.95 mbsf (3.000 Ma), yielding an average time step of 2.5 kyr—sufficient temporal resolution to capture precessional-scale cycles. Stable-isotope analyses were conducted at the University of Delaware with the GV Instruments IsoPrime dual-inlet mass spectrometer. The instrument is equipped with a Multiprep peripheral device and is used to perform the automated reaction of up to 40 samples in individual vials with hot (90 °C) phosphoric acid. Oxygen and carbon isotopic values have been corrected to the Peedee Belemnite standard using NBS-19 and Carrara Marble as an in-house standard. For δ18O values, analytical precision is better than 0.08‰ in the size range of the samples (~20–150 μg). To assess external reproducibility, we replicated 7% of the measurements. Duplicate (or triplicate for some intervals) values are on average b0.25‰ (n= 38) for δ18O, suggesting that these measurements are highly replicable and that variability among individual tests from the same sample is low (Fig. 3b). Gs. ruber δ18O measurements will be made available online in SedDB (www.seddb.org) and at the NOAA Paleoclimatology archive (http://www.ncdc.noaa.gov/ paleo/paleocean.html). Though the Gs. ruber δ18O values are generally reproducible, some intervals remained undersampled after double- and triple-checking for appropriate material. The large majority (>2/3) of our measurements are based on at least 8 tests, a common threshold for down-core paleoceanographic reconstructions (e.g., Oppo et al., 2001; Hagen and Keigwin, 2002; Weirauch et al., 2008). However, 7% of the δ18O measurements are based on groups of 2–4 tests, while another 22% are based on 5–7 tests. Although including δ18O values based on small groups reduces accuracy slightly, capturing a complete orbital-scale spectrum of variability requires high-resolution sampling. Reassuringly, δ18O measurements composed of only a few (2–4) individuals align with those based on higher numbers of tests, further supporting low δ18O variability among individual tests from the same sample (Fig. 3b). Insufficient numbers of Gs. ruber tests prevented a measurement in 83 of 523 (16%) sampled intervals, leaving gaps in the record. Random distribution of these gaps with respect to glacial versus interglacial intervals provides evidence against a temperature-induced sampling bias. Time-series analysis is performed applying standard Blackman– Tukey techniques (Jenkins and Watts, 1968). Cross-spectral analyses are generated using the Arand software (Howell et al., 2006). All records are linearly detrended and interpolated at 2.5-kyr intervals to approximate the sampling resolution of the Gs. ruber record. Lags are set to 150. As the record is virtually complete with only one gap greater than 10 kyr, we interpolate across missing data points. Thus, to the extent that our sampling is continuous and consistent, spectral analyses should reflect robust patterns; under-sampling of the signal should introduce noise, not regularity. As noted above (Section 1), to isolate a surface-water hydrographic signal (temperature and salinity), we subtract a basic index for global-mean seawater δ18O from the Gs. ruber δ18O record. The seawater δ18O approximation is a 10-kyr running mean of the benthic (pure Planulina wuellerstorfi) δ18O record at Site 1208 (Venti and Billups, 2012). Smoothing the benthic record with a 10-kyr running mean decreases obliquity-scale variability by about 50% to reflect the estimate for glacial/interglacial-scale seawater δ 18O variability of Sosdian and Rosenthal (2009) based on Mg/Ca-derived Atlantic bottom water temperature estimates for this interval of time. We recognize that this generalized approximation of ice-volume-derived δ18O changes is less suitable in the Pacific, across individual marine isotope stages, and at higher temporal resolution than the obliquity scale. Despite these limitations, we emphasize that the seawater δ 18O approximation is more realistic than attributing the entire raw benthic foraminifer δ18O record to ice-volume change. Using a marine isotopic reservoir index from the Site 1208 section rules out any uncertainties due to potential inter-site age-model mismatches.

N.L. Venti, K. Billups / Marine Micropaleontology 101 (2013) 106–114

109

a

b

c

d

NHG onset

Fig. 3. Comparison of Site 1208 Globigerinoides ruber δ18O measurements (b) to summer (June 21–July 20) insolation at 36°N (a; Laskar et al., 1993), Δδ18O, a measure of sea surface hydrography (c, see text for details), and the Site 1208 benthic δ18O record (d), the gray curve in (d) reflects the 10-kyr running mean (see text for details). Individual Gs. ruber δ18O measurements based on ≥5 tests plot as filled circles and measurements based on 2–4 tests plot as open gray circles (b); the line connects sample averages. Selected isotope stages are labeled in (d). The arrow in (d) indicates the onset of NHG. Gray bars highlight isotope stages 73, 87 and 97 and dashed gray lines highlight extreme insolation values.

4. Results 4.1. Gs. ruber δ 18O time series The local (36°N) summer insolation curve (Laskar et al., 1993) is dominated by precessional-scale variations that find counterparts in the Site 1208 Gs. ruber δ18O record (Fig. 3a and b, respectively). Many insolation maxima correspond to minima in the δ18O record and vice versa. Pronounced minima in the δ18O record correspond to intervals of particularly high insolation maxima, those associated with the eccentricitymodulated high-amplitude precessional variations (e.g., at 2.10 Ma, 2.60 Ma, and 2.93 Ma; dashed lines in Fig. 3a, b). However, there is no apparent long-term amplitude modulation in the Gs. ruber δ18O record that would correspond to the longer-term amplitude modulation of the insolation curve (Fig. 3a and b, respectively). Comparison of the Gs. ruber δ 18O record to the section's benthic foraminiferal δ18O record provides a first-order view of the relationship between summer/fall KCE surface hydrography and glacial–interglacial climate variability during the Plio/Pleistocene transition (Fig. 3b and d, respectively). Most notably, Gs. ruber δ18O values at Site 1208 display a negligible long-term trend across the Plio/Pleistocene climate transition (Fig. 3b). This is in contrast to the familiar increase in benthic foraminiferal δ18O values (0.4‰ at Site 1208 and elsewhere) that reflects cooling and glaciation at this time (Fig. 3d). The absence of a longterm trend suggests that the Gs. ruber δ18O record's summer/fall subtropical sea surface hydrographic component is insensitive to highlatitude cooling and ice-volume expansion. On the orbital scale, there are variations in the Gs. ruber δ18O record that correspond to the benthic foraminiferal δ18O record's more prominent features, e.g., interglacial isotope stages 73 near 1.9 Ma, 87 at 2.25 Ma, and 97 near 2.5 Ma (gray bars in Fig. 3d and b, respectively).

The amplitude of the orbital-scale variability is larger in the Gs. ruber δ18O record (total range of 1.5–2.5‰ throughout) than in the benthic δ18O record (amplitude of 0.1–0.3‰ from 3.0 to 2.7 Ma, then increasing to 0.2–0.6‰ between 2.7 and 1.8 Ma). Higher-amplitude variability in the planktic foraminiferal δ 18O record may reflect relatively large fluctuations in the KCE summer/fall surface-water hydrographic component (primarily temperature) that is superimposed on glacial/ interglacial-type variations. As explained in Section 3, a five-point (10-kyr) running mean of the benthic δ18O record approximates the seawater δ18O record (Fig. 3d). Thus, the difference between Gs. ruber δ18O and this simple seawater δ18O approximation ostensibly removes ice-volume-derived changes from the planktic record to isolate the subtropical summer/fall hydrographic component (temperature and salinity; Fig. 3c). By subtracting the benthic foraminiferal δ18O value from the planktic value, Δδ18O minima (most negative values) indicate a warming and/or decrease in salinity; Δδ18O maxima (less negative values) indicate a cooling (and/or increase in salinity) of the sea surface. In the Δδ18O record, variability corresponding to the interglacial isotope stages (73, 87, and 97) that are all prominent in the Gs. ruber δ18O record is diminished (gray bars in Fig. 3c). The lack of these features in the Δδ18O record provides first-order confirmation that the ice-volume changes have been minimized by subtracting the approximation of ice volume (Fig. 3c). On the other hand, the benthic δ18O record varies predominantly at the obliquity timescale rather than precessional frequencies. Thus, the Δδ18O record potentially emphasizes the more extreme values associated with times of high-amplitude precessional forcing as noted above (dashed lines in Fig. 3c), and it illustrates an apparent association between maxima in this record following precessionally-related insolation minima (Fig. 3a). On the finer scale, variability in the Δδ18O record is some 0.6–1.8‰ (Fig. 3c), which reflects a temperature change potentially on the order

110

N.L. Venti, K. Billups / Marine Micropaleontology 101 (2013) 106–114

of 3–9 °C, e.g., using the equation of Bemis et al. (1998). Low Δδ18O values tend to correspond to high insolation values (Fig. 3c and a, respectively), pointing to a precession-scale response of the summer/fall sea surface. At 2.75 Ma, the Δδ 18O record clearly shows a 0.3‰ decrease as benthic foraminiferal δ18O values increase (Fig. 3c and d, respectively). If salinity remained unchanged, this decrease represents a 1.5 °C increase in KCE summer/fall SSTs according to the paleotemperature equation for Gs. ruber δ18O values (Bemis et al., 1998). This indicates that the summer/fall KCE became warmer (and/or fresher) as NH glaciers grew larger. 4.2. Time-series analysis Time-series analysis of the Gs. ruber δ18O series confirms that this record is characterized by orbital-scale variability at both eccentricity bands, at the obliquity band, and at the 19-kyr period of precession (Fig. 4a). Gs. ruber δ18O cycles lag NH summer insolation cycles by 45° ±30° at the 1/41-kyr frequency (5 kyr± 3 kyr), and by 100°± 20° at the 1/19-kyr frequency (5 kyr ± 1 kyr; Fig. 4c). The time lags of 5 kyr at the obliquity frequency would be consistent with the time lag of the Lisiecki and Raymo (2005) age model, as this age model specifies a 5-kyr lag behind the obliquity forcing at 3.0 Ma, which increases to 15 kyr by 1.5 Ma. A 5-kyr lag in the precessional band, however, is too large to be accounted for in the same manner and may indicate a response of the subtropical sea surface to the regional insolation maximum delayed by orbital-scale climatic feedback processes, like how the heat capacity of water delays the SST response to insolation on the seasonal time scale. Spectral power at the eccentricity band has been

observed in other studies of the western Pacific (Medina-Elizalde and Lea, 2010), and may thus be a common feature of low-latitude records. Gs. ruber and benthic foraminiferal δ18O records are significantly coherent (above the 95% CI) at the 100-kyr and the 41-kyr periods (Fig. 4b). Only at the 41-kyr period are the two in phase (Fig. 4c), exemplifying the dominant control of the ice-volume beat in both records. Gs. ruber δ18O is significantly coherent (80% CI) with regional summer insolation at the 41-kyr period and at the 19-kyr precessional period (Fig. 4b). Time-series analysis of the Δδ18O record illustrates presence of significant power remaining in the two eccentricity bands and at the 19-kyr precessional period, but no concentration of power at the obliquity band (Fig. 5a). Foremost, this analysis underscores that the residual Δδ18O primarily records subtropical surface water hydrography, a regional climatic response unrelated to global-seawater δ 18O changes. Coherence and phase lags with respect to insolation at the 19-kyr precessional period and benthic foraminiferal δ 18O values at the eccentricity period are similar to those observed in the Gs. ruber δ18O record (Figs. 5b, c and 4b, c, respectively). Hence, subtracting a temperaturecorrected benthic foraminiferal δ18O record from the Gs. ruber δ18O record suggests very little ice-volume-related variability at these bands. To test whether the observed 5-kyr lag at the 19-kyr periodicity is consistent with a process similar to the seasonal lag of hydrography after the June insolation maximum, we also conduct a cross-spectral analysis between Gs. ruber and the subtropical August/September insolation (Fig. 6). Both the Gs. ruber δ 18O and the Δδ 18O record are highly coherent (95%) and in phase with the precessional component of this late-summer insolation curve (Fig. 6b). This comparison illustrates that on the precessional scale, the Gs. ruber δ 18O and Δδ 18O

a

b

c

Fig. 4. Cross-spectral analysis between Globigerinoides ruber δ18O, peak summer (June 21–July 20) insolation at 36°N (Laskar et al., 1993), and the section's benthic foraminiferal δ18O record: power spectra (a), coherence (b), and phase (c).

N.L. Venti, K. Billups / Marine Micropaleontology 101 (2013) 106–114

values are consistent with the species' preferred calcification during this season. This comparison also shows a loss of power at the obliquity band and a gain in power at the eccentricity bands in the late-summer relative to the mid-summer insolation spectrum (Figs. 6a and 5a, respectively). At the longer-term eccentricity band (400 kyr), the late summer/fall insolation curve is now significantly coherent with the Gs. ruber δ18O (and Δδ18O) record to the 80% level (Fig. 6b). To the extent that a phase measurement of 3 potential 400-kyr cycles in the 1.2 Myr-long record is meaningful, the Gs. ruber δ18O and Δδ18O cycles lead eccentricity maxima by about a quarter cycle (110°–120°, or lag eccentricity minima by 60°–70°). Nonetheless, the relatively high power in the proxy records at the eccentricity band might reflect the species' one-sided sensitivity: i.e., to warm summer/fall subtropical waters but not temperature changes during the cold season. The coherence between late-summer insolation and the Gs. ruber Δδ18O record on eccentricity scales provides geologic evidence for inferences from energy-balance models that show enhanced sensitivity of low latitudes to nonlinear effects associated with orbital geometry (e.g., Short et al., 1991). 5. Discussion In sum, subtracting the temperature-corrected benthic foraminiferal δ18O record from the planktic foraminiferal δ 18O record minimizes the effect of global ice-volume changes, leaving a residual that largely reflects sea surface hydrography. This is evidenced by the lack of 41-kyr variability in the Gs. ruber Δδ18O record. Because Gs. ruber lives primarily during the summer/fall, we interpret the Δδ18O record to reflect sea surface hydrography in the KCE during this season. Decreasing Δδ18O values from a mean of −3.2‰ to −3.5‰ thus suggest a summer/

111

fall warming of this region by as much as ~1.5 °C (assuming no change in surface water salinity) with the onset of widespread NHG at 2.73 Ma. At the orbital scale, the proxy indicates that surface water hydrography follows precession-scale summer insolation cycles. 5.1. The summer/fall KCE through the Plio-Pleistocene transition While NHG expanded at 2.7 Ma, Gs. ruber Δδ 18O values decreased by ~ 0.3‰ (e.g., Fig. 3d, c, respectively) suggesting that the summer KCE became warmer by as much as 1.5 °C. Summer/fall warming in the subtropics is consistent with increased summer/fall warmth in the subarctic North Pacific, proposed as a potential moisture source for high-latitude glaciations (Haug et al., 2005). Our results contribute extra-regional significance to this mechanism and point to the low latitudes as a source of available heat (and moisture) during this season. As a sea surface temperature increase may enhance evaporation, a concomitant increase in surface-water salinity may have dampened the δ 18O change, masking the full temperature range resolved by the planktic foraminiferal δ 18O values. Alternatively, that the entire 0.3‰ decrease in the Δδ 18O record reflects a surface-water freshening is possible considering that dramatic freshening and decreased seawater δ 18O values (by 2–6‰) occurred in the open western subarctic North Pacific (Site 882), ostensibly associated with meltwater during glacial intervals beginning at 2.7 Ma (Swann, 2010). It is conceivable that through the Oyashio Current, the influence of 18O-depleted meltwater might have even reached the KCE (e.g., Fig. 1). However at 2.7 Ma, ice volume increased; thus, the Δδ 18O decrease recorded at this time is not likely to reflect the influence of meltwater. There is no marked change in the Δδ 18O record at 2.2 Ma that would show a response of the KCE sea surface to hydrographic

a

b

c

Fig. 5. Cross-spectral analysis between Δδ18O, peak summer (June 21–July 20) insolation at 36°N (Laskar et al., 1993), and the section's benthic foraminiferal δ18O record: power spectra (a), coherence (b), and phase (c).

112

N.L. Venti, K. Billups / Marine Micropaleontology 101 (2013) 106–114

a

b

c

Fig. 6. Cross-spectral analysis between late-summer (August 21–September 20) insolation at 36°N (Laskar et al., 1993), Globigerinoides ruber δ18O, and Δδ18O: power spectra (a), coherence (b), and phase (c).

changes in the eastern equatorial Pacific Ocean. At 2.2 Ma, SSTs in the eastern equatorial Pacific show the first pronounced cooling events, and by 2 Ma an east–west temperature gradient is clearly established, signifying the end of perennial El Niño-like conditions at the equator (Lawrence et al., 2006; Ravelo et al., 2007; Etourneau et al., 2010). This event is also associated with an increase in the ratio of warmwater foraminifer species to cool-water species at Site 1208 (Venti, 2006), but no corresponding increase in summer/fall SSTs is implied by the Gs. ruber Δδ 18O record (Section 4.1). This lack of a summer/ fall hydrographic response might suggest that summer KCE temperatures do not necessarily correspond to KCE flow intensity. The end of perennial El Niño-like conditions may have been accompanied by strengthened Walker circulation. To the extent that this enhanced wind stress curl in the subtropics, KCE flow would have strengthened as discussed above (Section 2) to more consistently deliver warmwater planktonic taxa without necessarily increasing summer/fall temperatures. 5.2. The summer/fall KCE on orbital timescales Of the three variables impacting δ 18O (ice volume, temperature, and salinity), temperature changes are most likely to account for the precessional-scale variability in the Gs. ruber Δδ 18O record. It is unlikely that orbital-scale Δδ 18O cycles can be explained by the effect of ice volume on seawater δ 18O values. Even if the hydrography approximation does not effectively account for ice-volume-derived seawater δ 18O changes, precessional-scale δ 18O variability available to represent changes in seawater δ 18O (ice volume) is only 0.1‰ in

amplitude, less than that in the Δδ 18O record, which has an amplitude of up to 0.5‰. Moreover, precession-scale Δδ 18O variability is not coherent with benthic foraminifer δ 18O variability (Fig. 5b). It is similarly unlikely that salinity changes dominate the orbital-scale Δδ 18O variability at Site 1208's open-ocean setting: a 1.0‰ change in Δδ 18O is equivalent to the difference between the center of the subtropical gyre and the Gulf of Alaska (e.g., Bigg and Rohling, 2000), opposite ends of the hydrologic (evaporation–precipitation) spectrum. Summer insolation at 36°N, and at low and middle latitudes is dominated by precessional forcing (e.g., Figs. 4a, 5a, and 6a; Laskar et al., 1993). The Gs. ruber Δδ18O record reflects this insolation signal; it contains significant concentration of power at the 19-kyr precessional period. Thus, the KCE hydrographic reconstruction is consistent with the idea that subtropical climate is precession-dominated (e.g., Kutzbach, 1981; Clemens and Prell, 1990; Ruddiman, 2006, and references therein; Clemens and Prell, 2007, and references therein; Ao et al., 2011). Aside from the climatic significance, this observation validates the assumption of constant sedimentation rates between obliquity-scale marine isotope stage transitions that provide the age model. Summer/ fall KCE temperature variability corresponding to a precessionally influenced low-latitude climate is consistent with the modern ocean wherein subtropical warmth extends to mid latitudes during summer (Section 2). The in-phase response to late-summer subtropical insolation likely reflects particularly warm and favorable local conditions for Gs. ruber resulting directly from increased overhead solar radiation during the species' growth season. Regional impacts, such as variations in the intensity of the subtropical gyral circulation, determined primarily by wind-stress curl (e.g., Deser et al., 1999; Qiu, 2002) and sea surface

N.L. Venti, K. Billups / Marine Micropaleontology 101 (2013) 106–114

height in the subtropics (e.g., Imawaki et al., 2001), might have also enhanced precession-scale summer/fall warming of the KCE. We cannot rule out that salinity changes contributed to the observed precessional-scale hydrographic variability. Although decreases in salinity in the subarctic North Pacific are associated with meltwater during glacial intervals (Swann, 2010), it is unlikely that this particular mechanism contributed to precessional-scale hydrographic variability at Site 1208. This is because the salinity variability in the subarctic North Pacific closely follows the marine δ18O record, and so like NHG, is presumably dominated by the obliquity pacing (Swann, 2010). Moreover, the Site 1208 Δδ18O and benthic foraminifer δ18O records are not coherent at either precessional band. More likely, orbital-scale salinity changes in the Kuroshio Extension might reflect shifts in the hydrologic cycle of Southeast Asia (e.g., Nie et al., 2008), though unfortunately the lack of coherence between Site 1208 benthic δ18O and Δδ18O limits comparison to hydrologic reconstructions from the South China Sea that tie into the marine δ 18O chronology. Nonetheless, hematite/ goethite ratios from this region suggest precession-scale precipitation variability (Ao et al., 2011). The absence of a 41-kyr cycle, the dominant high-latitude climate pacing, in the summer/fall hydrography reconstruction is consistent with the way orbital parameters affect subtropical sunlight distribution (i.e., relatively little power in the obliquity band). It emphasizes the insensitivity of the KCE summer/fall sea surface to high-latitude climate change, where obliquity is a more-prominent forcing agent. These results suggest minimal far-field impacts of the growth and decay of glaciers during the Plio/Pleistocene on the summer/fall subtropical sea surface. Specifically, the lack of a 41-kyr signal indicates that there were no large-scale meridional changes in the surface water masses or the position of the gyre on these time scales. These results are consistent with SST reconstructions from foraminiferal faunal assemblages in the western tropical and subtropical Pacific that show relatively stable summer SSTs during the Pliocene/Pleistocene (Wang, 1994; Sato et al., 2008). Insensitivity of the gyre boundary to glacial–interglacial climate shifts has been proposed based on small differences in pollen transport (Kawahata and Ohshima, 2002) and planktic foraminiferal stable isotopes (Yamane, 2003) and assemblages (Ujiié, 2003) across marine isotope stages during the late Pleistocene. 6. Conclusions The Site 1208 Gs. ruber δ 18O record provides the first continuous, orbital-scale coverage of the Pliocene–Pleistocene climate transition from the subtropical northwest Pacific. By subtracting the benthic foraminiferal δ 18O record, we obtained a measure of sea surface hydrographic variability (Δδ 18O), which most likely reflects SSTs. Gs. ruber is a (sub)tropical species that inhabits the KCE mixed layer predominantly during summer/fall; thus Δδ 18O changes mostly reflect summer/fall SSTs. The hydrographic reconstruction (Δδ18O) of the summer/fall KCE shows a significant change with the expansion of NHG to the mid latitudes at 2.7 Ma. At this time, the summer/fall KCE warmed by as much as 1.5 °C, supporting the idea that the North Pacific was an important moisture source for the expanding NH glaciers. The Gs. ruber δ18O record varies at all the major orbital frequencies: precession, obliquity and eccentricity. Removing marine reservoir changes in seawater δ18O, which vary predominantly on the obliquity scale with glacial cycles, reveals that summer/fall KCE surface hydrography varied primarily at precessional and eccentricity frequencies. The lack of 41-kyr variability indicates that there were no large-scale meridional changes in the surface water masses or the position of the gyre on these time scales, reflecting a relatively stable hydrographic regime in the face of evolving NH boundary conditions. High-amplitude precessional-scale hydrography (presumably temperature) cycles, correspond to and are phased with summer/fall insolation, consistent with this species' calcification season in the northwestern subtropical Pacific Ocean. Thus, the

113

precession-scale variability in the summer/fall hydrographic reconstruction seemingly illustrates a direct response to subtropical insolation.

Acknowledgments Comments of Editor Richard Jordan, Tadamichi Oba, and two anonymous reviewers have greatly improved this manuscript. We thank Jessica Masterman, Livia Gong, Nicole Voutsina, Kevin Gielerowski, Elisa Sarantchin, Nathan Rabideaux, Christine Thomas, and Gabrielle Munn for their careful laboratory assistance. This research used samples provided by the Ocean Drilling Program (ODP). ODP is sponsored by the U.S. National Science Foundation (NSF) and participating countries under the management of the Joint Oceanographic Institutions, Inc. Funding was provided by NSF grant 0902729 to K.B.

References Antonov, J.I., Seidov, D., Boyer, T.P., Locarnini, R.A., Mishonov, A.V., Garcia, H.E., Baranova, O.K., Zweng, M.M., Johnson, D.R., 2010. World Ocean Atlas 2009, Volume 2: Salinity. In: Levitus, S. (Ed.), NOAA Atlas NESDIS 69. U.S. Government Printing Office, Washington, D.C. (184 pp.). Ao, H., Dekkers, M., Qin, L., Xiao, G., 2011. An updated astronomical timescale for the Plio-Pleistocene deposits from South China Sea and new insights into Asian monsoon evolution. Quaternary Science Reviews 30, 1560–1575. Balco, G., Rovey, C.W.I.I., Stone, J.O.H., 2005. The first glacial maximum in North America. Science 307, 222. Bemis, B.E., Spero, H.J., Bijma, J., Lea, D., 1998. Reevaluation of the oxygen isotopic composition of planktonic foraminifera: experimental results and revised paleotemperature equations. Paleoceanography 13, 150–160. Bigg, G.R., Rohling, E.J., 2000. An oxygen isotope data set for marine waters. Journal of Geophysical Research 105, 8527–8535. Bradshaw, J.S., 1959. Ecology of living planktonic foraminifera in the north and equatorial Pacific Ocean. Contributions from the Cushman Foundation for Foraminiferal Research 10, 25–64. Cannariato, K.G., Ravelo, A.C., 1997. Pliocene–Pleistocene evolution of eastern tropical Pacific surface water circulation and thermocline depth. Paleoceanography 12, 805–820. Chaisson, W.P., Ravelo, A.C., 2000. Pliocene development of the east–west hydrographic gradient in the equatorial Pacific. Paleoceanography 15, 497–505. Clemens, S.C., Prell, W.L., 1990. Late Pleistocene variability of Arabian Sea summer monsoon winds and continental aridity: eolian records from the lithogenic component of deep-sea sediments. Paleoceanography 5, 109–145. Clemens, S.C., Prell, W.L., 2007. The timing of orbital-scale Indian monsoon changes. Quaternary Science Reviews 26, 275–278. Da Silva, A., Young-Molling, A.C., Levitus, S., 1994. Atlas of Surface Marine Data 1994, vol. 6, NOAA Atlas NESDIS, vol. 15National Oceanic and Atmospheric Administration, Silver Spring, MD, USA. Deser, C., Alexander, M.A., Timlin, M.S., 1999. Evidence for a wind-driven intensification of the Kuroshio Current Extension from the 1970s to the 1980s. Journal of Climate 12, 1697–1706. Eguchi, N.O., Ujiié, H., Kawahata, H., Taira, A., 2003. Seasonal variations in planktonic foraminifera at three sediment traps in the Subarctic. Transition and Subtropical zones of the central North Pacific Ocean Marine Micropaleontology, 48, pp. 149–163. Etourneau, J., Schneider, R., Blanz, T., Martinez, P., 2010. Intensification of the Walker and Hadley atmospheric circulations during the Pliocene–Pleistocene climate transition. Earth and Planetary Science Letters 297, 103–110. Hagen, S., Keigwin, L.D., 2002. Sea-surface temperature variability and deep water reorganization in the subtropical North Atlantic during Isotope Stage 2–4. Marine Geology 189, 145–162. Harris, S.A., 2005. Thermal history of the Arctic Ocean environs adjacent to North America during the last 3.5 Ma and a possible mechanism for the cause of the cold events (major glaciations and permafrost events). Progress in Physical Geography 29. http://dx.doi.org/10.1191/0309133305pp444ra. Haug, G.H., Maslin, M.A., Sarnthein, M., Stax, R., Teidemann, R., 1995. Evolution of northwest Pacific sedimentation patterns since 6 Ma (Site 882). Proceedings of the Ocean Drilling Program: Scientific Results 145, 293–314. Haug, G.H., Ganopolski, A., Sigman, D.M., Rosell-Mele, A., Swann, G.E., Tiedemann, R., Jaccard, S.L., Bollmann, J., Maslin, M.A., Leng, M.J., Eglinton, G., 2005. North Pacific seasonality and the glaciation of North America 2.7 million years ago. Nature 433, 821–825. Howell, P., Pisias, N., Ballance, J., Baughman, J., Ochs, L., 2006. ARAND Time-Series Analysis Software. Brown University, Providence, RI. Imawaki, S., Uchida, H., Ichikawa, H., Fukasawa, M., Umatani, S., the ASUKA group, 2001. Satellite altimeter monitoring of the Kuroshio transport south of Japan. Geophysical Research Letters 28, 17–20. Jansen, E., Sjøholm, J., 1991. Reconstruction of glaciation over the past 6 Myr from ice-borne deposits in the Norwegian Sea. Nature 349, 600–603. Jenkins, G.M., Watts, D.G., 1968. Spectral Analysis and its Applications. Holden-Day, Merrifield, VA.

114

N.L. Venti, K. Billups / Marine Micropaleontology 101 (2013) 106–114

Kawahata, H., 2005. Stable isotopic comparison of two morphotypes of Globigerinoides ruber (white) in the subtropical gyre in the North Pacific. Paleontology Research 9, 27–35. Kawahata, H., Ohshima, H., 2002. Small latitudinal shift in the Kuroshio Extension (Central Pacific) during glacial times: evidence from pollen transport. Quaternary Science Reviews 21, 1705–1717. Kuroyanagi, A., Kawahata, H., 2004. Vertical distribution of living planktonic foraminifera in the seas around Japan. Marine Micropaleontology 53, 173–196. Kutzbach, J.E., 1981. Monsoon climate of the early Holocene: climate experiment with Earth's orbital parameters for 9000 years ago. Science 214, 59–61. Laskar, J., Joutel, F., Boudin, F., 1993. Orbital, precessional, and insolation quantities from −20 Myr to +10 Myr. Astronomy and Astrophysics 270, 522–533. Latif, M., Barnett, T.P., 1996. Decadal climate variability over the North Pacific and North America: dynamics and predictability. Journal of Climate 9, 2407–2423. Lawrence, K.T., Liu, Z., Herbert, T.D., 2006. Evolution of the eastern tropical Pacific through Plio-Pleistocene Glaciation. Science 312, 79–83. Levitus, S., Boyer, T., 1994. World Ocean Atlas 1994, Vol. 4: Temperature. NOAA Atlas NESDIS 4. U.S. Gov. Printing Office, Washington, DC (117 pp.). Levitus, S., Burgett, R., Boyer, T., 1994. World Ocean Atlas 1994, Vol. 3: Salinity. NOAA Atlas NESDIS 3. U.S. Gov. Printing Office, Washington, DC (99 pp.). Lisiecki, L.E., Raymo, M.E., 2005. A Pliocene–Pleistocene stack of 57 globally distributed benthic δ18O records. Paleoceanography 20, PA1003. http://dx.doi.org/10.1029/ 2004PA001071. Locarnini, R.A., Mishonov, A.V., Antonov, J.I., Boyer, T.P., Garcia, H.E., Baranova, O.K., Zweng, M.M., Johnson, D.R., 2010. World Ocean Atlas 2009, Volume 1: Temperature. In: Levitus, S. (Ed.), NOAA Atlas NESDIS 68. U.S. Government Printing Office, Washington, DC (184 pp.). Medina-Elizalde, M., Lea, D.W., 2010. Late Pliocene equatorial Pacific. Paleoceanography 25. http://dx.doi.org/10.1029/2009/PA001780. Miller, A.J., Cayan, D.R., White, W.B., 1998. A westward-intensified decadal change in the North Pacific thermocline and gyre-scale circulation. Journal of Climate 11, 3112–3127. Mizuno, K., White, W.B., 1983. Annual and interannual variability in the Kuroshio Current system. Journal of Physical Oceanography 13, 1847–1867. Mohiuddin, M.M., Nishimura, A., Tanaka, Y., Shimamoto, A., 2004. Seasonality of biogenic particle and planktonic foraminifera fluxes: response to hydrographic variability in the Kuroshio Extension, northwestern Pacific Ocean. Deep Sea Research Part I: Oceanographic Research Papers 51, 1659–1683. Newman, M., Compo, G.P., Alexander, M.A., 2003. ENSO-forced variability of the Pacific Decadal Oscillation. Journal of Climate 16, 3853–3857. Nie, J., King, J., Liu, Z., Clemens, S., Prell, W., Fang, X., 2008. Surface-water freshening: a cause for the onset of North Pacific stratification from 2.75 Ma onward? Global and Planetary Change 64, 49–52. Oppo, D.W., Keigwin, L.D., McManus, J.F., Cullen, J.L., 2001. Persistent suborbital climate variability in Marine Isotope Stage 5 and Termination II. Paleoceanography 16, 280–292. Philander, S.G., Fedorov, A.V., 2003. Role of tropics in changing the response to Milankovich forcing some three million years ago. Paleoceanography 18. http:// dx.doi.org/10.1029/2002PA000837. Qiu, B., 2002. The Kuroshio Extension system: its large-scale variability and role in midlatitude ocean–atmosphere interaction. Journal of Oceanography 58, 57–75. Ravelo, A.C., Dekens, P.S., McCarthy, M., 2006. Evidence for El-Nino-like conditions during the Pliocene. GSA Today 16, 4–11. Ravelo, A.C., Billups, K., Dekens, P.S., Herbert, T.D., Lawrence, K.T., 2007. Onto the ice ages: proxy evidence for the onset of Northern Hemisphere glaciation. In: Williams, M., Haywood, A.M., Gregory, F.J., Schmidt, D.N. (Eds.), Deep time perspectives on climate change: marrying the signal from computer models and biological proxies. The Micropaleontological Society, Special Publications. The Geological Society, London, pp. 563–573. Raymo, M.E., Nisancioglu, K., 2003. The 41-kyr world: Milankovitch's other unsolved mystery. Paleoceanography 18. http://dx.doi.org/10.1029/2002PA000791. Reid, J.L., 1997. On the total geostrophic circulation of the Pacific Ocean: flow patterns, tracers and transports. Progress in Oceanography 39, 263–352. Ruddiman, W.F., 2006. What is the timing of orbital-scale monsoon changes? Quaternary Science Reviews 25, 657–658.

Sainz-Trapaga, S.M., Goni, G.J., Sugimoto, T., 2001. Identification of the Kuroshio Extension, its bifurcation and northern branch from altimetry and hydrographic data during October 1992–August 1999: spatial and temporal variability. Geophysical Research Letters 28, 1759–1762. Sato, K., Oda, M., Chiyonobu, S., Kimoto, K., Domitsu, H., Ingle Jr., J.C., 2008. Establishment of the western Pacific warm pool during the Pliocene: evidence from planktic foraminifera, oxygen isotopes, and Mg/Ca ratios. Palaeogeography, Palaeoclimatology, Palaeoecology 265, 140–147. Schmidt, G.A., 1999. Forward modeling of carbonate proxy data from planktic foraminifera using oxygen isotope tracers in a global ocean model. Paleoceanography 14, 482–497. Schneider, N., Cornulle, B.D., 2005. The forcing of the Pacific Decadal Oscillation. Journal of Climate 18, 4355–4373. Shipboard Scientific Party, 2002. Leg 198 summary. In: Bralower, T.J., Premoli Silva, I., Malone, M.J., et al. (Eds.), Proceedings of the Ocean Drilling Program, Initial Reports, 198. Ocean Drilling Program, College Station TX, pp. 1–148. Short, D.A., Mengel, J.G., Crowley, T.J., Hyde, W.T., North, G.R., 1991. Filtering of Milankovitch cycles by Earth's geography. Quaternary Research 35, 157–173. Sosdian, S., Rosenthal, Y., 2009. Deep-sea temperature and ice volume changes across the Pliocene–Pleistocene climate transitions. Science 325, 306–310. Swann, G.E.A., 2010. Salinity changes in the North West Pacific Ocean during the late Pliocene/early Quaternary from 2.73 Ma to 2.52 Ma. Earth and Planetary Science Letters 297, 332–338. Ujiié, H., 2003. A 370-ka paleoceanographic record from the Hess Rise, central North Pacific Ocean, and an indistinct Kuroshio Extension. Marine Micropaleontology 49, 21–47. Venti, N.L., 2006. Revised late Neogene mid-latitude planktic foraminiferal biostratigraphy for the northwest Pacific (Shatsky Rise), ODP Leg 198, Master of Science thesis, Department of Geosciences. University of Massachusetts-Amherst, Amherst, MA, USA. Venti, N.L., Billups, K., 2012. Stable-isotope stratigraphy of the Pliocene–Pleistocene climate transition in the northwestern subtropical Pacific. Palaeogeography, Palaeoclimatology, Palaeoecology. http://dx.doi.org/10/1016/j.palaeo.2012.02.001. Vincent, E., Berger, W.H., 1981. Planktonic foraminifera and their use in paleoceanography. In: Emiliani, C. (Ed.), The Oceanic Lithosphere: The Sea, vol. 7. John Wiley & Sons, pp. 1025–1119. Wang, L., 1994. Sea surface temperature of the low latitude western Pacific during the last 5.3 million years. Palaeogeography, Palaeoclimatology, Palaeoecology 108, 379–436. Wang, L.J., 2000. Isotopic signals in two morphotypes of Globigerinoides ruber (white) from the South China Sea: implications for monsoon climate change during the last glacial cycle. Palaeogeography, Palaeoclimatology, Palaeoecology 161, 381–394. Wara, M.W., Ravelo, A.C., Delaney, M.L., 2005. Permanent El Niño-like conditions during the Pliocene warm period. Science 309, 758–761. Weirauch, D., Billups, K., Martin, P., 2008. Evolution of millennial-scale climate variability during the mid-Pleistocene. Paleoceanography 23. http://dx.doi.org/ 10.1029/2007PA001584. Wijffels, S.E., Hall, M.M., Joyce, T., Torres, D.J., Hacker, P., Firing, E., 1998. Multiple deep gyres of the western North Pacific: a WOCE section along 149°E. Journal of Geophysical Research 103, 12,985–13,009. Wyrtki, K., Maagard, L., Hagar, J., 1976. Eddy energy in the oceans. Journal of Geophysical Research 81, 2641–2646. Xie, P., Arkin, P.A., 1997. Global precipitation: 17-year monthly analysis based on gauge observations, satellite estimates, and numerical outputs. Bulletin of the American Meteorological Society 78, 2539–2558. Yamane, M., 2003. Late Quaternary variations in water mass in the Shatsky Rise area, northwest Pacific Ocean. Marine Micropaleontology 48, 205–223. Yasuda, I., 2003. Hydrographic structure and variability in the Kuroshio–Oyashio transition area. Journal of Oceanography 59, 389–402. Yasuda, I., Yoon, J.H., Suginohara, N., 1985. Dynamics of the Kuroshio large meanderbarotropic model. Journal of the Oceanographic Society of Japan 41, 259–273. Zhang, Y., Wallace, J.M., Battisi, D.S., 1997. ENSO-like interdecadal variability: 1900–93. Journal of Climate 10, 1004–1020. Zhong, Y., Liu, Z., 2009. On the mechanism of Pacific multidecadal climate variability in CCSM3: the role of the subpolar North Pacific Ocean. Journal of Physical Oceanography 39, 2052–2076.