Intermediate water ventilation change in the subarctic northwest Pacific during the last deglaciation



[1] The laminated sediments suggest significant depletion in dissolved oxygen of intermediate water in the North Pacific during the last deglaciation. However, the cause of depleted oxygen during deglaciation is still a matter of debate. Here, we present foraminiferal Mg/Ca-based intermediate and surface temperatures and δ18O of seawater (δ18OW; a proxy of salinity) in the Oyashio region in order to investigate intermediate water circulation through the last deglaciation. Results indicate that the intermediate water temperature and δ18OW increased in association with ventilation age at the Bølling-Allerød period, suggesting that the poor ventilation at that time. The surface record shows that the decrease of salinity preceded the slowdown of intermediate ventilation. The coupled variation of intermediate and surface water suggests that the supply of cold, fresh intermediate water diminished because of poor ventilation at the high latitude due to surface freshening.

1. Introduction

[2] The intermediate depths of the North Pacific experienced drastic environmental changes during the last deglaciation. Laminated sediments were deposited at the intermediate depths of the North Pacific margin [Behl and Kennett, 1996; Ikehara et al., 2006; Zheng et al., 2000] and the Bering Sea [Cook et al., 2005] during the Bølling-Allerød (B-A) period. The lamina preservation suggests that the oxygen concentration at those depths was so low as to prohibit benthos from surviving. The dissolved oxygen concentration at intermediate depths is determined by the balance between supply and consumption, and two mechanisms that might explain the enhancement of the oxygen minimum zone (OMZ) have been proposed: a reduction in ventilation [Ahagon et al., 2003; Kennett and Ingram, 1995; Zheng et al., 2000], and enhanced productivity [Crusius et al., 2004]. Under modern conditions, the Okhotsk Sea is the source region of the North Pacific Intermediate Water (NPIW) [Yasuda, 1997]. This region provides fresh water, which is characterized by low temperature, low salinity, and high dissolved oxygen, to the intermediate depth of the Oyashio region. Therefore, the records of intermediate temperature and salinity might help to elucidate paleo-ventilation. To investigate the variation of intermediate ventilation and its relationship to surface climate change during the last deglaciation, we analyzed the changes during the last deglaciation period in the intermediate and surface Mg/Ca-based temperatures and δ18OW in the Oyashio region off the eastern coast of northern Japan.

2. Materials and Methods

[3] A piston core (GH02-1030; 42°13.77′N, 144°12.53′E; water depth, 1212 m) was extracted from the slope off Tokachi (Figure 1). The core site was located in the main path of the southwestward Oyashio Current [Yasuda et al., 2001]. An age model was established with Ikehara et al. [2006] data (see auxiliary materials for a detailed description). Very weak lamination was found within two horizons in the middle part of the core. The upper laminated horizon was found at 11.4–12.2 cal.kyr BP (Lamina A), and the lower horizon was at 14.1–14.7 cal.kyr BP (Lamina B). We calculated the ventilation ages using the age projection method [Adkins and Boyle, 1997] (auxiliary material). The INTCAL04.14c atmospheric 14C data served as the reference, and published ventilation data from Adkins and Boyle [1997] and Ahagon et al. [2003] were recalculated.

Figure 1.

Locations of core GH02-1030 and other sites. Arrows indicate surface water flows in the North Pacific.

[4] The planktonic foraminifera Globigerina bulloides and benthic foraminifera Uvigerina akitaensis were picked from sieve fractions of 180–250 μm and 250–355 μm, respectively. For isotope analyses, the shell fragments were cleaned by ultrasonication in Milli-Q water and methanol. The oxygen isotope was analyzed with a precision of ±0.05‰. Isotope results were expressed as ‰ VPDB calibrated with a NBS-19 standard. Additional cleaning steps, including reductive, oxidative, and weak acid-leaching steps, were adopted for elemental analyses. Cleaned samples in acid-cleaned microfuge tubes were dissolved in 0.75 M HNO3 and then diluted to approximately equal Ca concentrations [de Villiers et al., 2002]. The element ratio was determined with precision of ±1% (1σ), using ICP-AES. Mn/Ca was used to examine Mn-oxide contamination, with a threshold of<50 μmol/mol [Boyle, 1983].

[5] The Mg/Ca values of G. bulloides were converted to calcification temperature using the calibration equations, which is based on core top sediments from the sub-Antarctic and laboratory-cultured samples, proposed by Mashiotta et al. [1999]. The validity of this temperature calibration for the northwest Pacific has been confirmed by Sagawa et al. [2006]. For the benthic Mg/Ca temperature, we used a linear calibration based on Uvigerina spp. from Arabian Sea surface sediments [Elderfield et al., 2006]. To reconstruct the surface and intermediate salinity changes during the deglaciation, we calculated the oxygen isotope of seawater (δ18OW) with an oxygen isotope temperature scale [Shackleton, 1974]. The global sea level curve [Waelbroeck et al., 2002] was subtracted from δ18OW to assess the local salinity change (Δδ18OW).

3. Results and Discussion

[6] The benthic foraminiferal Mg/Ca-derived intermediate temperature increased in the B-A and PB and decreased in the Younger Dryas (YD) (Figure 2c). U. akitaensis was insufficient for Mg/Ca analysis at the earliest B-A, because of dysoxic condition [Shibahara et al., 2007]. It is known that the post-depositional dissolution affects foraminiferal Mg/Ca [e.g., Brown and Elderfield, 1996]. The carbonate content of the GH02-1030 core during the deglacial period was relatively high in comparison with that of other periods (B. K. Khim et al., personal communication, 2007), implying that carbonate dissolution was negligible or very weak. The Δδ18O of intermediate water, which is a proxy of salinity, showed changes synchronous with temperature (Figure 2d). The similarity of temperature and salinity variations may imply that the amplitude of temperature change is overestimated due to application of Mg/Ca-T calibration, which derived from the Arabian Sea sediment, to this region. However, the calculated Δδ18OW shows that the saline period corresponds to the warm period with any Mg/Ca-T calibration of Uvigerina spp. [Elderfield et al., 2006; Lear et al., 2002], and the Δδ18OW amplitude difference is less than 0.08‰ among calibrations. As a result, variation in water mass properties involved the alternation of warm-saline and cold-fresh components. The general trend of the reconstructed ventilation age from GH02-1030 is consistent with that of neighboring cores from shallow (CH84-14; 978 m) [Adkins and Boyle, 1997] and deep (PC4/5; 1366 m) [Ahagon et al., 2003], indicating that the ventilation age was great in the B-A when lamina preserved (Figure 2e). The small ventilation age at 14.0 cal.kyr BP coincided with the disappearance of lamination in GH02-1030. This age agrees with the horizons between lamina 2 and 3 in MD01-2409 [Ikehara et al., 2006]. Large ventilation ages at 13–13.5 cal.kyr BP were observed in other core, corresponding to lamina 2 in MD01-2409, but there is no radiocarbon data in GH02-1030. The ventilation age variation was associated with changes in intermediate water properties, although the transition from YD to PB is not clear.

Figure 2.

Down-core results of the GH02-1030 core proxy record for the last deglaciation. (a) δ18O record of the Greenland ice sheet GISP2 [Grootes et al., 1993]. (b) Benthic δ18O and (c) Mg/Ca-temperature. AMS 14C calendar ages from planktonic and benthic foraminifera are shown as open and solid triangles, respectively, on the top axis of Figure 1b. (d) Calculated intermediate water Δδ18OW change. (e) Ventilation age of intermediate water from GH02-1030 (blue), from Adkins and Boyle [1997] (red), and from Ahagon et al. [2003] (green). (f) Planktonic δ18O, (g) Mg/Ca-temperature, and (h) surface Δδ18OW.

[7] According to the modern observation, the dissolved oxygen concentration at northwest Pacific intermediate depth has decreased for the last several decades [Nakanowatari et al., 2007; Ono et al., 2001; Watanabe et al., 2001]. The decrease of dissolved oxygen coincides with increase of salinity and phosphate [Ono et al., 2001], CFC tracer ages [Watanabe et al., 2001], and temperature [Nakanowatari et al., 2007]. These results clearly show that the weakened intermediate circulation results in increasing temperature and salinity because the supply of cold and low-salinity water to intermediate depth has decreased. In our sedimentary records lamina preservation and the increase in the intermediate temperature, salinity, and ventilation age in the B-A suggested that the intermediate ventilation was significantly reduced.

[8] The radiolarian and benthic foraminiferal assemblages suggest that the intermediate water was cold, well-oxygenated, and ventilated during the LGM [Ohkushi et al., 2003; Shibahara et al., 2007]. This is consistent with geochemical evidence indicating that the intermediate depth ventilation (<2 km) was similar or more active than modern conditions at the LGM and that deep waters (>2 km) were nutrient rich [Galbraith et al., 2007; Keigwin, 1998; Matsumoto et al., 2002]. At the onset of the B-A, the dominance of oxic to suboxic taxa changed abruptly to dysoxic taxa [Shibahara et al., 2007], suggesting that the supply of glacial cold and low-salinity intermediate water to the study site abruptly weakened. The significant decrease of C. davisiana relative abundance in the Bering Sea sediment suggests that intermediate water formation was reduced at the end of the glacial period [Tanaka and Takahashi, 2005]. The high resolution record of benthic assemblage from MD01-2409 (976 m) shows a brief oxic period during the B-A [Shibahara et al., 2007], which agrees with the small ventilation age at 14.0 cal.kyr BP in GH02-1030 (Figure 2e). After this brief event, dissolved oxygen returned to suboxic to dysoxic level as indicated by benthic assemblage, and the lamination was observed in MD01-2409 but in GH02-1030. Therefore, the enhancement of OMZ at the late B-A might be centered at shallower depth than the previous one. Intermediate ventilation was enhanced again at the YD, as suggested by the return of suboxic taxa.

[9] What determined the strength of the intermediate water circulation at the last deglaciation? On the basis of modern observations, intermediate water properties are associated with surface water variation [e.g., Ono et al., 2001]. Surface water properties at high latitude have an impact on surface and vertical oceanic conditions [Sigman et al., 2004]. The planktonic δ18O (δ18OPF) variations during the last deglaciation are shown in Figure 2f. The δ18OPF value shows a rapid decrease of 0.8‰, from 15.2 to 14.6 cal.kyr BP, comparable to the onset of B-A warm period in the North Atlantic. The decrease of δ18OPF in this period was also observed in the subarctic region, including the eastern coast of Japan [Hoshiba et al., 2006; Kallel et al., 1988], the western subarctic gyre [Gorbarenko, 1996; Keigwin et al., 1992; Keigwin, 1998], the Alaskan gyre [Galbraith et al., 2007; Zahn et al., 1991], the Okhotsk Sea [Gorbarenko, 1996], and the Bering Sea [Cook et al., 2005; Gorbarenko, 1996]; note that some of these did not have sufficient age control. Surface water temperature and Δδ18OW were reconstructed using δ18OPF combined with Mg/Ca-temperature (Figures 2g and 2h). Because of low sedimentation rate, the data were sparse during 13.5–12.5 cal.kyr BP. The Mg/Ca-temperature increased approximately 2° and 1°C in the B-A and PB, respectively. Calculated surface Δδ18OW at the late glacial was 0.77‰, which was higher than LGM value of 0.3‰ (not shown), and decreased to 0.34‰ at 14.6 cal.kyr BP. The salinity minima were recorded at 14.6 and 13.6 cal.kyr BP, coincided with the bottom of lamina B in GH02-1030 and lamina 2 in MD01-2409, respectively. The lamina B disappeared after the fluctuation of temperature and salinity at 14.0 cal.kyr BP. The decrease of surface salinity preceded the slowdown of intermediate ventilation at the B-A. Surface freshening in the B-A has been reported in several studies from the subarctic North Pacific [Zheng et al., 2000] and proximal marginal seas [Gorbarenko et al., 2004, 2005; Seki et al., 2004]. The nearly synchronous phenomena at the surface and intermediate depth imply that a strong link existed between them. Given that surface salinity decreases cause seawater to gain buoyancy, surface conditions at the onset of B-A would have yielded a reduction of ventilation.

[10] Another possible explanation of the enhanced OMZ is higher surface productivity. Crusius et al. [2004] suggested that increased productivity at the B-A contributed to oxygen depletion at the intermediate depth. However, no lamination was found in the Holocene section, despite high productivity in the late Holocene compared with the deglacial period [Ikehara et al., 2006; Narita et al., 2002]. Moreover, the benthic foraminiferal assemblage was dominated by suboxic taxa throughout the Holocene, in contrast to the dysoxic taxa dominating the B-A [Shibahara et al., 2007]. This is because the modern intermediate ventilation is more active than the B-A. Therefore, the enhanced OMZ during deglaciation cannot be explained by high surface productivity alone. Another mechanism is needed to explain the preservation of lamina.

[11] The timing of the surface freshening shown in this study was coincident with NADW regeneration [McManus et al., 2004]. The modeling studies showed the Atlantic meridional overturning circulation (MOC) changes had significant impacts on the Pacific climate. Schmittner et al. [2007] demonstrated that the MOC variation had a significant effect in the NPIW formation. Furthermore, the weakening MOC reduced summer precipitation over the East Asian monsoon region and the tropical western Pacific [Zhang and Delworth, 2005]. The relationships between the MOC change and Pacific response illustrated in modeling simulations were similar to paleo-records from this and previous studies. The activity of the East Asian summer monsoon was abruptly enhanced during the B-A warming interval [Wang et al., 2001]. Increased precipitation from the tropical Pacific was also reported during this interval [Rosenthal et al., 2003; Stott et al., 2002]. The similar variation in the entire Pacific might reflect inter-basin moisture redistribution during the last deglaciation.

[12] Our data suggest that rapid freshening of the surface water could have reduced intermediate ventilation during the B-A and caused depletion of dissolved oxygen in the OMZ, possibly combined with high production. These results show strong coupling of atmosphere-ocean subsystems during climate change and indicate that intermediate water circulation in the North Pacific greatly influenced biogeochemical cycles. Further study is needed to understand why the glacial intermediate circulation pattern switched to the modern pattern.


[13] We thank M. Murayama, K. Horikawa, and the other members of KCC, as well as M. Yamamoto, T. Irino, Y. Igarashi, A. Shibahara, B. K. Khim, H. Katayama, and A. Noda for their help and valuable discussions. We also thank the captain, onboard scientists, and crew of the GH02 cruise of the R/V Hakurei-Maru 2 for their help with the fieldwork. J. R. Toggweiler and two anonymous reviewers made valuable contributions to the paper during the review process. This research was supported in part by a Sasagawa Scientific Research Grant from the Japan Science Society and by Grant-in-aid for Young Scientists (B) 20710007.