Variations in the radiolarian assemblages in the Bering Sea since Pliocene and their implications for paleoceanography

Palaeogeography, Palaeoclimatology, Palaeoecology 410 (2014) 337–350

Contents lists available at ScienceDirect

Palaeogeography, Palaeoclimatology, Palaeoecology journal homepage: www.elsevier.com/locate/palaeo

Variations in the radiolarian assemblages in the Bering Sea since Pliocene and their implications for paleoceanography Qiang Zhang a,b,⁎, Muhong Chen a, Lanlan Zhang a, Weifen Hu a,b, Rong Xiang a a b

Key Laboratory of Marginal Sea Geology, South China Sea Institute of Oceanology, Chinese Academy of Sciences, Guangzhou 510301, China University of Chinese Academy of Sciences, Beijing 100049, China

a r t i c l e

i n f o

Article history: Received 28 November 2013 Received in revised form 30 April 2014 Accepted 30 May 2014 Available online 11 June 2014 Keywords: The Bering Sea Site U1340 Pliocene Radiolarian assemblages Paleoenvironmental conditions

a b s t r a c t Radiolarian assemblages were analyzed using samples from the IODP Site U1340 to reconstruct the paleoceanographic conditions in the Bering Sea. Based on the characteristic faunal changes, the radiolarian evolution at Site U1340 was divided into four major intervals from Stages I to IV, with Stage IV divided into Substages IVa and IVb. The radiolarians in each stage recorded significant paleoenvironmental conditions. In general, the Bering Sea was governed by an ocean ecological environment with stable warm and saline surface water during Stage I (4.15 Myr to 3.91 Myr). The environment in the Bering Sea fluctuated strongly during Stage II (3.91 Myr to 2.75 Myr) and was controlled by the cold-water masses and sea ices during Stage III (2.75 Myr to 1.07 Myr) with the gradual development of cold and well-ventilated intermediate water. Stage IVa (1.07 Myr to 0.47 Myr) was a transitional period characterized by the enhanced formation of cold subsurface and intermediate water as well as of the oxygen-rich deep water. During Stage IVb (after 0.47 Myr), the Bering Sea was mainly characterized by enhanced warmth during interglacial episodes and welldeveloped water layers that were generally comparable to those of the modern Bering Sea. These conditions indicated that the vertical water-mass structure of the modern Bering Sea began to form since 0.47 Ma. Every Stage boundary in the studied core was marked by notable changes in the radiolarian assemblages. These changes corresponded to the climatic cooling event at ~ 3.91 Ma, the intensification of Northern Hemisphere glaciation at ~2.75 Ma, the beginning of the middle Pleistocene transition at ~1.07 Ma, and the low-latitude radiolarian ecology event at ~0.47 Ma. In addition, the relative abundance pattern of Cycladophora davisiana indicates that the Bering Sea was the main source of the past North Pacific Intermediate Water at ca. 0.85 Ma (MIS22), ca. 0.63 Ma (MIS16), and ca. 0.18 Ma (MIS6), just as it was during the last glacial maximum. © 2014 Elsevier B.V. All rights reserved.

1. Introduction The Bering Sea, which is a marginal sea with high primary production, plays a significant role in modern and past global climatic systems (Takahashi, 1998, 1999b). As an outlet to both the North Pacific and the Arctic Oceans, the Bering Sea serves important functions in water mass exchange and heat equilibriums as well as in the chemical properties of different water types (Takahashi, 1998, 1999a). Given its pivotal role and unique location, the Bering Sea is a critical place for the reconstruction of subarctic climate variability on various time scales (Takahashi, 1999b, 2005). Therefore, a comprehensive knowledge of the paleoceanographic features of the Bering Sea can improve our

⁎ Corresponding author at: South China Sea Institute of Oceanology, Chinese Academy of Sciences, Guangzhou 510301, China. E-mail addresses: [email protected] (Q. Zhang), [email protected] (M. Chen), [email protected] (L. Zhang), [email protected] (W. Hu), [email protected] (R. Xiang).

http://dx.doi.org/10.1016/j.palaeo.2014.05.048 0031-0182/© 2014 Elsevier B.V. All rights reserved.

understanding of the evolutionary history of the paleoenvironments related to the climate changes in the subarctic (Takahashi, 1998). The Bering Sea's sedimentary sequence is relatively rich in siliceous plankton but has minimal calcareous microfossil content because of its high latitude location and special hydrological conditions. Thus, radiolarians with siliceous skeletons and various depth habitats are potentially effective proxies for paleoceanographic study in the Bering Sea. Several previous studies on paleoceanography have been conducted in the Bering Sea with radiolarians serving as proxy. Blueford (1983) and Wang et al. (2006) reported the depth distribution of the late quaternary radiolarians in the western and eastern parts of the Bering Sea, respectively. Morley and Robinson (1986) explain the stratigraphic significance of the radiolarian species Cycladophora davisiana (Ehrenberg) in the comparison of late Pleistocene/Holocene sequences with unusually high sedimentation rate in the Bering Sea. Wang and Chen (2005) developed the application of C. davisiana in stratigraphic correlations, and Itaki et al. (2009) discussed the change in intermediate water conditions based on the abundance profile of C. davisiana. Employing the characteristic radiolarian species, including C. davisiana,

338

Q. Zhang et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 410 (2014) 337–350

Fig. 1. The location of Site U1340, and surface water circulation in the Bering Sea (Revised after Takahashi et al., 2011). KC, Kamchatka Current; BSC, Bering Slope Current; ANSC, Aleutian North Slope Current.

Ceratospyris borealis (Bailey), and Stylochlamydium venustum (Bailey), Tanaka and Takahashi (2005) revealed the paleoenvironmental changes in the water-mass structures in the Bering Sea and determined the source regions of the past North Pacific Intermediate Water (NPIW) during the last 100 kyr. Itaki et al. (2012) reconstructed the millennial-scale variations in the vertical water column in the Bering Sea during the late Pleistocene with the combined data on radiolarians and geochemical components. In addition, various paleoceanographic studies have been conducted in the Bering Sea based on stable isotopes as well as on geochemical, lithological, and micropaleontological analyses (e.g., Sancetta, 1983; Sancetta and Robinson, 1983; Nakatsuka et al., 1995; Gorbarenko, 1996; Gorbarenko et al., 2005; Khusid et al., 2006; Gorbarenko et al., 2010; Zou et al., 2012). However, these studies only focused on the paleoenvironmental reconstruction of the late Pleistocene because of the limited length of the available sedimentary sequence. DSDP Leg 19 first drilled the long sediment cores covering the quaternary of the Bering Sea, but the recovery and quality of the core material was too poor to be used in paleoceanographic studies (Scholl and Creager, 1973). Thus, detailed paleoceanographic information on the Bering Sea that covers long time scales remained lacking until today. In 2009, the Integrated Ocean Drilling Project (IODP) Expedition 323 retrieved high-quality sediment cores to investigate the Pliocene– Pleistocene paleoenvironmental conditions of the Bering Sea. Using the samples from Site U1340, IODP 323, Zhang et al. (2014) defined the Pliocene–Pleistocene radiolarian zones and updated the shipboard preliminary age model of Site 1340 (Takahashi et al., 2011). This model provided basic data for the paleoceanographic study in the Bering Sea.

In the present work, we examine the variations in the radiolarian assemblages at Site U1340 to further reveal the paleoceanographic characteristics of the Pliocene–Pleistocene Bering Sea. This study aims to reconstruct radiolarian evolutions, paleoceanographic changes (including the oceanic conditions through the water column) and global climate changes. 2. Materials and methods The samples used in the present study were obtained from the sedimentary sequence of Site U1340, IODP 323, which is located on the Bowers Ridge at 53°24.0008’ N and 179°31.2973’ W at a water depth of 1295 m (Fig. 1). The studied core with a maximum composite depth of 604 m was obtained by establishing the stratigraphic correlations among the three holes cored at Site U1340. The sediments in the studied core consisted mostly of dark green–gray diatom ooze or diatom silt, except for a few volcanic ash layers and siliciclastic sediments (Takahashi et al., 2011). Radiolarian preservation ranged from moderate to good. A total of 243 samples were taken with a sampling interval of 40 cm from the top 15 m and 3 m for the rest of the core. During the sample processing, the samples were not washed to avoid the breakage and loss of radiolarian specimens. The detailed procedure for this process is as follows. (1) Approximately 1 g of the dried sample was weighed into a 50 ml beaker, which was then added with 25 ml of 10% H2O2 solution. (2) After approximately 30 min, the beaker was placed in a sonic oscillator for approximately 2 min to separate the clays that adhered to the radiolarian tests. The beaker

Plate I. All scale bars equal 100 μm. The meanings of the Sample ID are as follows, taking “U1340A-2H-3W, 52–54 cm; 7.42 m” for example, U1340: the number of the site providing the sample; A: the number of the Hole at Site U1340; 2: the second sediment core of the studied core in downward sequence; H: ‘Hard Core’, the type of the core; 3: the third section of the second sediment core; W: for working, the purpose of the sample; 52–54 cm: the distance from the sampling position to the top of the third section; 7.42 m: the depth of the sample. Figs. 1–2, Spongotrochus glacialis Popofsky: 1. (U1340B-1H-1W, 125–127 cm; 1.25 m), 2. (U1340A-70X-5 W, 2–4 cm; 591.42 m). Figs. 3–4, Stylochlamydium venustum (Bailey): 3. (U1340B4H-4W, 2–4 cm; 29.92 m), 4. (U1340A-68X-1 W, 2–4 cm; 566.02 m). Figs. 5–6, Spongopyle osculosa Dreyer: 5. (U1340A-6H-4W, 2-4 cm; 46.44 m), 6. (U1340A-50X-1 W, 2-4 cm; 419.32 m). Figs. 7–8, Stylodictya vilidispina Jørgensen: 7. (U1340B-4H-4W, 2–4 cm; 29.92 m), 8. (U1340A-39H-1W, 2-4 cm; 333.72 m). Figs. 9–10, Actinomma boreale Cleve: 9. (U1340A-9H-1W, 2–4 cm; 70.42 m), 10. (U1340A-11H-5W, 2–4 cm; 95.42 m). Figs. 11–12, Actinomma leptoderma (Jørgensen): 11. (U1340B-1H-1W, 15–17 cm; 0.15 m), 12. (U1340A-9H-1 W, 2–4 cm; 70.42 m). Figs. 13–14, Larcopyle buetschlii Dreyer: 13. (U1340A-3H-6W, 72–74 cm; 21.62 m), 14. (U1340A-16H-7 W, 2–4 cm; 145.92 m). Figs. 15–16, Lithomellisa setosa Jørgensen: 15. (U1340A-11H7 W, 2–4 cm; 98.42 m), 16. (U1340A-27H-7 W, 2–4 cm; 233.72 m). Figs. 17–18, Cyrtopera laguncula Haeckel: 17. (U1340A-2H-1 W, 14–16 cm; 4.04 m), 18. (U1340A-56X-7 W, 2–4 cm; 485.62 m). Figs. 19–20, Antarctissa (?) sp. 1: 19. (U1340B-1H-1 W, 15–17 cm; 0.15 m), 20. (U1340A-7H-2W, 2–4 cm; 52.92 m). Figs. 21–22, Antarctissa (?) sp. 2: 21. (U1340A-7H-2W, 2–4 cm; 52.92 m), 22. (U1340B-1H-1 W, 15–17 cm; 0.15 m).

Q. Zhang et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 410 (2014) 337–350

339

340

Q. Zhang et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 410 (2014) 337–350

was then moved to the lab table. (3) After letting the beaker stand for 2 min, the clay suspension was removed, and deionized water was added into the beaker. (4) Step (3) was repeated until the water in the beaker was clear. (5) The clear deionized water was removed, and the residue was transferred to the centrifuge tube. (6) After approximately 3 min, the residual water was removed, and all the residual substances were strewn on the microslides. Several drops of Canada balsam were placed on the slide, and a cover glass was mounted on the slide. Radiolarians were observed and identified under an optical microscope at 100× or 200× magnification, and the minimum size of extracted specimens is about 40 μm in length (see Plate I). More than 500 specimens were identified on the slides for most samples, while only 150 to 250 specimens were examined on few of the samples because of the scarce radiolarian fossils. The species used in this paper are illustrated in Plate I and Plate II. This study adopted four radiolarian parameters (FRPs), namely, abundance, radiolarian accumulation rate (RAR), simple diversity (S), and Shannon–Wiener diversity (H), to describe the characteristics of the radiolarian assemblages at Site U1340. Abundance represents the individuals per gram (n.g-1), and S indicates the number of species (species) in each sample. The RAR that removed the effect of dilution by terrigenous input and H, which represents the stability of the radiolarian community structure, were respectively calculated as follows (Shannon and Wiener, 1949; Tanaka and Takahashi, 2005):  h h i i −2 −1 −1 −1 RAR No: shells cm kyr ¼ No: shells g dry sediment No:g h i −3  ðdry bulk densityÞ g cm h i −2 −1  ðsedimentation rateÞ cm kyr

H = −∑ni = 1PilnPi, where Pi is the proportion of each species. The age–depth plot used in this study is an update of the preliminary age model of Takahashi et al. (2011), which was re-established using the synthetical datum of the effective biostratigraphic and magnetostratigraphic events (Zhang et al., 2014). Age estimation of various depths was performed by linear interpolation between two control points. The base age of the studied core was calculated at approximately 4.15 Ma (i.e., the late Early Pliocene). 3. Results The Site U1340 FRPs exhibit similar fluctuation patterns (Fig. 2). They present high values in the upper section of the core above 66.92 m and in the lower section below the 547.42 m, relatively low values at the intervals 297.22 m to 547.42 m and 66.92 m to 155.92 m, and extremely low values at the interval 155.92 m to 297.22 m (Fig. 2). This result indicates that the radiolarian assemblages experienced four remarkable changes in the Bering Sea since the late Early Pliocene. Based on the age–depth plot from Zhang et al. (2014), the numerical ages of the depths 66.92, 155.92, 297.22, and 547.42 m were estimated by linear interpolation to be 0.47, 1.07, 2.75, and 3.91 Ma, respectively. Based on the changes in the FRPs profiles, the radiolarian assemblages were divided into four intervals from Stages I to IV, with Stage IV divided into Substages IVa and IVb. Stage I (4.15 Myr to 3.91 Myr, the late Early Pliocene): This stage is characterized by high abundance, S, RAR, and H with average values of 10,204 n.g-1, 37 species, 170,599 n.cm-2.kyr-1, and 1.92, respectively (Table 1). The species Stylochlamydium venustum, Spongotrochus glacialis (Popofsky), and Siphocampe arachnea (Ehrenberg), which occur abundantly in the surface sediments of the modern Bering Sea

(Tanaka and Takahashi, 2005; Wang et al., 2005, 2006; Itaki et al., 2012), predominated a mean total relative abundance of over 65%. Ceratospyris borealis and Cycladophora davisiana, which are the common species in the modern subarctic, were rare or absent in this stage (Fig. 3). Stage II (3.91 Myr to 2.75 Myr, late Early Pliocene to late Late Pliocene): The abundance, S, RAR, and H parameters of the radiolarian assemblage presented relatively low average values of 3748 n.g-1, 24 species, 47,334 n.cm-2.kyr-1, and 1.68, respectively (Table 1). The abundance of Stylochlamydium venustum, Spongotrochus glacialis, and Siphocampe arachnea sharply decreased at the boundary Stage I/II. Then, the abundance of S. venustum and S. glacialis continued to decline, while the abundance of S. arachnea presented an increasing trend. Ceratospyris borealis and Dictyophimus hirundo (Haeckel)/Dictyophimus crisiae Ehrenberg groups were common, and Cycladophora davisiana and S. arachnea became the dominant species in the late period of this stage (Fig. 3). Stage III (2.75 Myr to 1.07 Myr, late Late Pliocene to middle Pleistocene): The abundance, S, RAR, and H parameters exhibit extremely low average values of 2853 n.g-1, 21 species, 16,592 n.cm-2.kyr-1, and 1.61, respectively (Table 1). Cycladophora davisiana, Ceratospyris borealis, and Siphocampe arachnea were the common species, with the first two slightly increased at the expense of S. arachnea. In addition, the species Lithomelissa setosa Jørgensen abruptly increased between 2.5 and 1.7 Myr. Stage IV (after 1.07 Myr, middle Pleistocene to present): The FRPs increased rapidly at the early period of this stage. Then, S and H show high values throughout except at the IVa/IVb transition, while abundance and RAR fluctuate frequently from very high to very low. Based on the ranges of the radiolarian parameters, we divided Stage IV into Substages IVa and IVb. Stage IVa was defined as the period from 1.07 Ma to 0.47 Ma and is characterized by relatively high average values of abundance, S, RAR, and H (10,521 n.g-1, 38 species, 113,668 n.cm-2.kyr-1,and 1.79, respectively) with peaks of 21,110 n.g-1, 59 species, 249,088 n.cm-2.kyr-1, and 2.26, respectively (Table 1). Siphocampe arachnea, Cycladophora davisiana, Dictyophimus hirundo/crisiae group, and Cyrtopera laguncula Heackel were the common species at this substage. Cycladophora davisiana exhibited high abundance early in the substage and decreased later, while S. arachnea, D. hirundo/crisiae group, and C. laguncula exhibited the opposite trend (Fig. 3). Additionally, the abundance of Actinomma leptoderma (Jørgensen)/Actinomma boreale Cleve group shows distinct peak at around 0.65 Ma and 0.49 Ma, respectively. Stage IVb covers 0.47 Ma to the present, and abundance, S, RAR, and H parameters exhibit high average values of 15,870 n.g-1, 52 species, 147,805 n.cm-2.kyr-1, and 2.19, respectively, with peaks of 55,892 n.g-1, 72 species, 542,048 n.cm-2.kyr-1, and 2.52, respectively (Table 1). Siphocampe arachnea and C. davisiana were the dominant species. Dictyophimus hirundo/crisiae group, C. laguncula and Ceratospyris borealis were common. Spongotrochus glacialis and S. venustum, which sporadically occurred in Stage III, gradually increased and became common species (Fig. 3). 4. Discussion 4.1. Radiolarian assemblage variations and paleoenvironmental changes Sediment trap experiments and piston core studies show that radiolarian assemblages have distinct variations in their responses to changes in oceanic environmental conditions, such as water temperature, salinity, and food supply (e.g., Anderson, 1983; Takahashi, 1997; Okazaki et al., 2003a; Abelmann and Nimmergut, 2005; Motoyama and

Plate II. Figs. 1–4, Dictyophimus hirundo (Haeckel)/crisiae Ehrenberg group: 1, 3. (U1340A-35H-7W, 2–4 cm; 309.72 m), 2, 4. (U1340A-13H-7W, 2–4 cm; 117.42 m). Figs. 5–6, Pterocanium korotnevi (Dogiel): 5. (U1340B-1H-1 W, 125–127 cm; 1.25 m), 6. (U1340A-7H-2W, 2–4 cm; 52.92 m). Figs. 7–8. Amphimelissa setosa (Cleve): 7, 8. (U1340A-3H-6W, 72–74 cm; 21.62 m). Figs. 9–10. Ceratospyris borealis Bailey: 9. (U1340A-57X-3 W, 2–4 cm; 489.22 m), 10. (U1340B-1H-1 W, 15–17 cm; 0.15 m). Figs. 11–12, Zygocircus productus (Herting) group: 11. (U1340A-31H-1 W, 2–4 cm; 262.72 m), 12. (U1340A-2H-3W, 70–72 cm; 7.6 m). Figs. 13–14. Siphocampe arachnea (Ehrenberg): 13. (U1340A-10H-3W, 2–4 cm; 82.92 m), 14. (U1340A-24H-1 W, 2–4 cm; 196.22 m). Figs. 15–16, Cycladophora cornutoides (Petrushevskaya): 15. (U1340A-2H-5W, 64–66 cm; 10.55 m), 16. (U1340A-9H-1 W, 2–4 cm; 70.42 m). Figs. 17–18. Cycladophora davisiana Ehrenberg: 17. (U1340B-1H-2W, 45–47 cm; 1.95 m), 18. (U1340A-56X-1 W, 2–4 cm; 476.62 m). Fig. 19–20, Cycladophora sakaii Motoyama: 19. (U1340A-68X-1 W, 2–4 cm; 566.02 m), 20. (U1340A-56X-1 W, 2–4 cm; 476.62 m).

Q. Zhang et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 410 (2014) 337–350

341

342

Q. Zhang et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 410 (2014) 337–350

Fig. 2. Fluctuation patterns of the FRPs for radiolarian assemblages at Site U1340. RAR, radiolarian accumulation rate; S, simple diversity; H, Shannon–Wiener diversity.

Nishimura, 2005; Itaki et al., 2008, 2012). In general, warm and stable oceanic conditions with high productivity are beneficial to radiolarian production. Conversely, radiolarian populations, particularly the species dwelling near the surface, such as Stylochlamydium venustum, significantly decline when the marine environment becomes severe, for example with the development of sea-ice (Abelmann, 1992; Okazaki et al., 2003a; Abelmann and Nimmergut, 2005; Tanaka and Takahashi, 2005). Therefore, we are able to deduce the significant events of paleoenvironment and the climate changes in the Bering Sea during geological time based on the variations in radiolarian assemblages at Site U1340. During Stage I, the high FRPs values (Fig. 4) indicate an optimum ocean ecological environment for high radiolarian production, probably related to the Climatic Optimum known as the late Early Pliocene Warmth (Ciesielski and Grinstead, 1986; Hodell and Kennett, 1986; Abelmann et al., 1990; Barron, 1992; Dowsett et al., 1992). In addition, the production of shallow water species Stylochlamydium venustum is

very high, which indicates the near absence of sea-ice coverage in the region of the studied core. This finding can be well confirmed by the extremely low percentage of sea-ice diatoms during this period (Takahashi et al., 2011). At the Stage I/II boundary (3.91 Ma), the FRPs values and the abundance of S. venustum declined abruptly with a slight increase in sea-ice diatoms (Takahashi et al., 2011) and an abrupt increase in cold-water radiolarians (Figs. 3, 4). This finding shows a slightly enhanced formation of the sea ice and a climatic transition from a warm period to relatively cold conditions in the Bering Sea at 3.91 Ma. Dowsett et al. (1996) reported the global warming of the ocean surface during the late Early Pliocene, but some studies show several cooling events against the warm background for the period. Heusser and Morley (1996) documented late Early Pliocene cooling at ~3.7 Ma in the western Pacific Ocean in the vicinity of Japan based on combined pollen and microfaunal studies. Anderson (1997) reported an extended cool period between 3.35 and 3.05 Ma at ODP Site 806 in the western equatorial Atlantic Ocean. Marlow et al. (2000) used alkenone-

Table 1 The FRP values of radiolarian assemblages in each stage during the last 4.15 Myr. Stages

IVb IVa III II I

Time Interval (Myr) 0.47–0 1.07–0.47 2.75–1.07 3.91–2.75 4.15–3.91

Abundance (n.g-1)

RAR (n.cm-2.kyr-1)

S (species)

H

Min.

Max.

Avg.

Min.

Max.

Avg.

Min.

Max.

Avg.

Min.

Max.

Avg.

1922 662 107 214 1584

55,892 21,110 14,823 24,559 26,833

15,870 10,521 2853 3748 10,204

23 10 9 7 15

72 59 36 43 52

52 38 21 24 37

8826 8951 643 2502 30,256

542,048 249,088 58,082 23,7710 46,8757

147,805 113,668 16,592 47,334 17,0599

1.79 1.23 1.03 1.38 1.60

2.52 2.26 2.03 2.54 2.28

2.19 1.79 1.61 1.68 1.92

Notes: RAR, the radiolarian accumulation rate; S, the simple diversity; H, the Shannon–Wiener diversity. Min. = Minimum; Max. = Maximum; Avg. = Average.

Q. Zhang et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 410 (2014) 337–350

343

Fig. 3. Relative and absolute abundance changes for the major radiolarian taxa living in various depths and the pattern of total sulphur (from Takahashi et al., 2011) at Site U1340. FO = first occurrence; FCO = first common occurrence.

344

Q. Zhang et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 410 (2014) 337–350

Fig. 4. The FRPs and relative abundance profiles of cold-water radiolarian species (Actinomma boreale, Actinomma leptoderma, Amphimelissa setosa (Cleve), Cycladophora cornutoides (Petrushevskaya), Cycladophora davisiana, Larcopyle buetschlli Dreyer, Petrocanium korotnevi (Dogiel), Siphocampe arachnea, Spongopyle osculosa Dreyer) (Bjørklund and Kruglikova, 2003; kamikuri et al., 2007) and sea-ice diatoms at Site U1340.

calculated sea surface temperatures and observed the presence of marked Pliocene cooling at 3.7 Ma at ODP Site 1084 off Namibia. Carter (2005) reported a sharp mid-latitude atmospheric cooling from 3.68 Myr to 3.63 Myr during the late Early Pliocene in New Zealand based on the natural gamma ray at ODP Site 1119. The pulsatory cooling episodes mentioned above indicated that the cooling episodes of the late Early Pliocene to the late Pliocene were not only regional phenomena. The variations in the radiolarian assemblages in the Bering Sea provided new evidence for the climatic cooling during the late Early Pliocene, which was probably a global event. Interestingly, the cooling event at 3.91 Ma in the Bering Sea seemed to have occurred earlier than those in the low and middle latitudes. During Stage II (3.91 Myr to 2.75 Myr), the radiolarian assemblages exhibit relatively low average FRPs values as a whole (Fig. 4), thus indicating that the radiolarian production was distinctly affected by the climate cooling and the development of sea ice at the end of Stage I (~3.91 Ma). At the boundary of Stage II/III (2.75 Ma), the further decrease in the FRPs values and the abundance of Stylochlamydium venustum, which is associated with the increase in percentages of cold-water radiolarian species and sea-ice diatoms (Takahashi et al., 2011), indicate the enhancement of the sea-ice cover and another significant climatic cooling at 2.75 Ma; this finding is in agreement with the intensification of the Northern Hemisphere Glaciation (NHG) at around 2.7 Ma (Raymo, 1994; Maslin et al., 1996). Then, the evolutionary history of the radiolarian assemblages stepped into Stage III, which was characterized by cold climatic conditions.

During Stage III (2.75 Myr to 1.07 Myr), the low values of the FRPs (Fig. 4) and the high abundance of sea-ice diatoms (Takahashi et al., 2011) indicate that the region of the studied core was controlled by the cold-water masses and sea ice that seriously inhibited the propagation of radiolarians. This relatively stable and cold climate continued until the end of Stage III. At 1.07 Ma, the proportion of cold-water radiolarians decreased from 40% to 24% even though the sea-ice diatoms continuously increased (Takahashi et al., 2011). The values of the FRPs increased distinctively, and the production of Stylochlamydium venustum increased slightly when the total radiolarian abundance increased from approximately 2500 n.g-1 to 13,000 n.g-1 and when the diversity increased from 23 species to 47 species (Fig. 3, 4). These changes imply the reduction of the sea ice and the occurrence of a significant climatic transition in the Bering Sea at around 1.07 Ma, approximately coincident with the beginning of the Middle Pleistocene Climatic Transition (MPT) at around 1.2 Ma (e.g., Pisias and Moore, 1981; Mudelsee and Schulz, 1997; Clark et al., 2006). During Stage IV (after 1.07 Ma), the FRPs fluctuate frequently but show a gradual increase overall (Fig. 4) and the sea-ice diatom percentages mirror these fluctuations, decreasing overall (Takahashi et al., 2011) which indicate that the climate was gradually getting warmer but with frequent large deviations from the trend. This result is consistent with the result revealed by the benthic foraminifer δ18O values from the equatorial Pacific Ocean (Shackleton and Opdyke, 1973). In addition, both radiolarian abundance and RAR were genenally high during the interglacials and low during the glacials based on the relationship

Q. Zhang et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 410 (2014) 337–350

Fig. 5. Changes in the radiolarian abundance and radiolarian accumulation rate during the last 1.07 Myr at Site U1340 and benthic foraminifer oxygen stable isotope at Site U1143 (from Tian et al., 2006).

between the FRPs and the δ18O values at ODP Site 1143 (Tian et al., 2006) (Fig. 5). This finding suggests radiolarian production and consequent accumulation in the sediments of the Bering Sea were primarily regulated by the general changes in global interglacial and glacial conditions since the beginning of the MPT. During Stage IVa, the FRPs have relatively higher values (an average RAR of 113,668 n.cm-2.kyr-1) compared with those in Stage III (an average RAR of 16,592 n.cm-2.kyr-1) but lower values than those in Stage IVb (an average RAR of 147,805 n.cm-2.kyr-1) (Table 1), thus implying that the radiolarian assemblages were gradually recovering while the community structure was unstable due to the frequent fluctuations in the climate events, which are also recorded by the variations in the sea-ice diatoms (Takahashi et al., 2011). In contrast with relatively high average values of the FRPs during stage IVa, the FRPs decreased sharply both at ~0.65 Ma (MIS 16) and ~0.49 Ma (MIS 12) (Fig. 4). This change was probably caused by the severe water conditions of the long duration sea-ice coverage with the climate cooling, because the Actinomma leptoderma/ boreale group shows high abundance at around 0.65 Ma (MIS 16) and around 0.49 Ma (MIS 12) (Fig. 3), and this group is closely related to severe environments such as areas with cold sea-surface temperature, long duration or permanent sea-ice cover and extremely low productivities (Bjørklund and Kruglikova, 2003; Itaki, 2003). In addition, the percentage of sea-ice diatoms in the studied core shows the peak at ~0.65

345

Ma and ~ 0.49 Ma, respectively (Fig. 4), which indicates the enhancements of the sea-ice coverage during these two periods. Among the glacial periods since the MPT, MIS 16 was well known as a period of high global ice volume in estimates of global ice volume (Shackleton and Opdyke, 1973; Shackleton, 1987), and MIS 12 was a very cold period in agreement with the Anglian Glaciation reported in England (e.g. Rose et al., 1985; Bowen et al., 1986; Lee et al., 2004). These findings suggest that both MIS 16 and MIS 12 corresponded to the extremely cold climate which may be responsible for the severe water conditions of the long duration sea-ice coverage and low productivity. Eventually, this severe condition resulted in the low values of the FRPs in the Bering Sea at ~ 0.65 Ma and ~ 0.49 Ma. At around 0.47 Ma, accompanied by the decline in sea-ice diatoms, the values of the FRPs abruptly increased and then presented significantly higher levels during interglacial periods in Stage IVb (Figs. 4, 5). In addition, Stylochlamydium venustum also shows a higher average abundance in Stage IVb than that in stage IVa. These findings suggest an overall trend of further warming and sea-ice reducing, and this change well responded to the Mid-Brunhes Event (MBE) that occurred at ~ 0.43 Ma, which was characterized by enhanced warmth during interglacial episodes after MIS 13 (e.g. Jansen et al., 1986; Wang et al., 2003; Candy et al., 2010). During the intervals from 0.35 Ma to 0.31 Ma (MIS 9), 0.25 Ma to 0.2 Ma (MIS 7) and 0.12 to 0.08 Ma (MIS 5), the average RAR, which is associated with the abundance peaks of S. venustum, reached 212,864 n.cm-2.kyr-1, 235,311 n.cm-2.kyr-1 and 196,874 n.cm-2.kyr-1, respectively, and these values were greater than that of 170,599 n.cm-2.kyr-1 during Stage I. This finding may indicates the favorable and warm periods for high radiolarian production comparable to the late Early Pliocene Warmth. The abrupt increase in the radiolarian assemblages in the Bering Sea at 0.47 Ma is in agreement with the radiolarian ecological events reported in the Indian Ocean (Nigrini, 1991) and in the southern South China Sea (Yang et al., 2002). Because this event, which is characterized by rapid increase of radiolarian abundance and diversity after ~ 0.47 Ma, is recorded not only in low latitude but also in high latitude, changes in radiolarian assemblages at 0.47 Ma was likely a global event. 4.2. Oceanographic conditions of different water layers during the last 4.15 Myr Studies on modern radiolarian ecology show that some species inhabit particular depths through water columns and would exhibit high abundance under the optimal conditions in these water masses (e.g., Casey, 1977; Casey et al., 1979; Kling, 1979; Nimmergut and Abelmann, 2002; Okazaki et al., 2004; Abelmann and Nimmergut, 2005). Therefore, the variations in radiolarian assemblages, including in shallow and deep dwellers, can reflect changes in vertical watermass conditions. Following the lead of plankton investigations in the subarctic (Kling, 1979; Kling and Boltovskoy, 1995; Nimmergut and Abelmann, 2002; Itaki, 2003; Okazaki et al., 2003a, 2004; Tanaka and Takahashi, 2005), we obtained 14 radiolarian species dwelling at various depths; the surface layer (shallower than 50 m), subsurface layer (50 m to 200 m), intermediate layer (200 m to 500 m), and deep layer (N 500 m) (Table 2 and Fig. 3). Based on this depth zonation and the

Table 2 List of major radiolarian dwellers at various depths of the subarctic. Surface dwellers (0–50 m)

Subsurface dwellers (50–200 m)

Intermediate dwellers (200–500 m)

Deep dwellers (N500 m)

Spongotrochus glacialis Stylochlamydium venustum Lithomelissa setosa Stylodictya validispina

Actinomma leptoderma Actinomma boreale Ceratospyris borealis Antarctissa (?) sp.1 Zygocircus productus

Cycladophora davisiana

Dictyophimus hirundo Dictyophimus crisiae Siphocampe arachnea Cyrtopera laguncula

References: Kling (1979), Kling and Boltovskoy (1995), Nimmergut and Abelmann (2002), Itaki (2003), Okazaki et al. (2004), Abelmann and Nimmergut (2005), Tanaka and Takahashi (2005).

346

Q. Zhang et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 410 (2014) 337–350

corresponding radiolarian assemblages at Site 1340, we attempted to reconstruct the water conditions of each layer in the Bering Sea since the late Early Pliocene.

4.2.1. Surface water condition (upper 50 m) The assemblage of the surface dwellers was composed of Spongotrochus glacialis, Stylochlamydium venustum, Stylodictya validispina Jørgensen, and Lithomelissa setosa, with the first two species being dominant for most periods and L. setosa being dominant during Stage III (Fig. 3). Several studies discussed the ecology habits of these species at various water conditions. Tanaka and Takahashi (2005) point out that the productions of the S. glacialis and S. venustum would be strongly limited by low-temperature and low-salinity surface waters caused by the melt water or the expansion of the sea-ice in the Bering Sea; such description is well supported by the distribution patterns of these two species in the subarctic. Spongotrochus glacialis and S. venustum are abundant in the modern Bering Sea (Ling et al., 1971; Blueford, 1983; Wang et al., 2006; Itaki et al., 2012) and the western subarctic Pacific (Kruglikova, 1969) but are relatively rare in the western part of the present Okhotsk Sea (Abelmann and Nimmergut, 2005), where the surface water has low salinity due to the fresh water discharged from the Amur River. Wang et al. (2005, 2006) suggests that the high abundance of S. glacialis and S. venustum is related to the relatively high productivity based on the relationship between the radiolarian assemblages and the high biological productivity area called the “Green Belt” in the Bering Sea (Springer et al., 1996). A similar result was shown by the radiolarian data from the Okhotsk Sea, wherein the increasing number of S. venustum represents a highly stable condition of less-pronounced seasonal changes combined with an increased seasonal phytoplankton production at the sea surface in the geological past (Abelmann and Nimmergut, 2005). In addition, Robertson (1975) suggests that the low abundance north of 45° N but high abundance between 35° N and 45° N of S. venustum in the subarctic during the last glacial were caused by the decrease in the surface mixing in the Bering Sea and subarctic Pacific north of 45° N at that time (Itaki et al., 2012). This suggestion indicates that the production of S. venustum is also related to the extent of the surface mixing. As mentioned above, high S. venustum abundance seems to represent an optimal condition characterized by relatively warm and saline water with high productivity and surface mixing, whereas low abundance of S. venustum reflects severe conditions with low-temperature and low-salinity surface water. In the studied core, the abundance of S. venustum and S. glacialis exhibits high values from 4.15 Myr to 3.91 Myr and during the interglacials of the last 0.47 Myr (Fig. 3), corresponding to the warm period discussed above. Therefore, we suggest that the high production of S. venustum and S. glacialis during the geological periods was closely related to the surface condition of the relatively warm and saline water caused by the warm climate. The high abundance of Stylochlamydium venustum and Spongotrochus glacialis (Fig. 3) from 4.15 Myr to 3.91 Myr indicates that the surface water presented ideal conditions of stably warm and saline water that was beneficial to the massive propagations of S. venustum and S. glacialis; this propagation corresponds to the thriving stage of the radiolarian assemblages caused by the late Early Pliocene Warmth (Ciesielski and Grinstead, 1986; Hodell and Kennett, 1986; Abelmann et al., 1990; Barron, 1992; Dowsett et al., 1992). From 3.91 Myr to 1.07 Myr, the abundance of S. venustum and S. glacialis gradually decreased (Fig. 3), implying cold and low-salinity surface water, consistent with the climatic cooling and the significant sea-ice expansion of Stages II and III. The increase in the abundance of S. venustum and S. glacialis suggests that the conditions of surface layer gradually improved for the radiolarian ecology and were generally characterized by the relatively warm and saline water as a whole during the last 1.07 Ma.

By contrast, the abundance and percentage of Lithomelissa setosa were relatively high (Fig. 3) from 2.5 Myr to 1.7 Myr when a cold period with sea-ice expansion occurred in the Bering Sea. Lithomelissa setosa is currently a common species in the Norwegian Sea, the Okhotsk Sea, and the Japan Sea, and its high abundance is closely related with the influences of warm currents and high biological production (Bjørklund et al., 1998; Abelmann and Nimmergut, 2005; Itaki et al., 2007). In the subarctic, the Alaskan Stream is the only warm current with high production that has influence on the area of the studied core. Thus, the surface water circulation in the Bering Sea was possibly significantly influenced by the warm Alaskan Stream water mass from 2.5 Myr to 1.7 Myr. 4.2.2. Sub-surface water condition (50 m to 200 m) The subsurface water in the modern Bering Sea is characterized by the development of the cold dichothermal layer that originates from the shelf area during winter resulting from convection under the seaice (Tomczak and Godfrey, 1994). At Site U1340, the subsurface radiolarian dwellers include Actinomma leptoderma/boreale group, Ceratospyris borealis, Antarctissa (?) sp.1, Lithelius minor Jørgensen, and Zygocircus productus (Herting) group. The A. leptoderma/boreale group, which is the common species in the modern Arctic Ocean, can reproduce under the conditions of low subsurface temperature, long duration sea-ice cover, and limited primary productivity (Bjørklund and Kruglikova, 2003; Itaki, 2003). In both the Bering Sea and the Okhotsk Sea, the abundance of the A. leptoderma/boreale group significantly increased during the last glacial due to the long duration of the sea-ice season and low productivity (Itaki et al., 2008, 2012). Hence, we derive the high abundance of the A. leptoderma/boreale group indicates the severe conditions of long duration sea-ice cover and low productivity. Ceratospyris borealis, which prefers to live around the dichothermal layer in the Okhotsk Sea (Nimmergut and Abelmann, 2002), can thrive in the subsurface layer with extremely low temperatures. Moreover, Takahashi (1997) suggests that C. borealis could be used as a productivity indicator in the eastern subarctic Pacific, and this suggestion was subsequently supported by the data on the Okhotsk Sea (Nimmergut and Abelmann, 2002). Thus, we propose that the high abundance of C. borealis indicates the enhanced development of the cold dichothermal layer in the subsurface with high productivity. At Site 1340, the sporadic occurrence of Ceratospyris borealis from 4.15 Myr to 3.91 Myr (Fig. 3) indicates the relatively low productivity and the absence of the cold dichothermal layer in the subsurface at that time. This condition suggests a thicker surface layer and the absence of sea-ice in the Bering Sea resulting from the warmer climate of the late Early Pliocene. The relatively high abundance of C. borealis from 3.91 Myr to 2.75 Myr implies the enhanced formation of the cold dichothermal layer that was probably related to the development of the seasonal sea-ice for the climate cooling and relatively high productivity during this period. The subsurface cold water was weakly generated from 2.75 Myr to 1.07 Myr, as shown by the low abundance of C. borealis. It then gradually developed as shown by the increase in the abundance of C. borealis during the last 1.07 Myr. The two peaks of the abundance of the A. leptoderma/boreale group at around 0.65 Ma and 0.49 Ma indicate the occurrence of two distinct periods of long duration sea-ice cover and extremely low productivities in the Bering Sea (see Section 4.1). 4.2.3. Intermediate water condition (200 m to 500 m) Although Cycladophora davisiana inhabits in several depth zones globally and is generally considered as a deep dweller in most regions (Petrushevskaya and Bjørklund, 1974; Abelmann and Gowing, 1997; Bjørklund et al., 1998; Itaki, 2003), it mainly dwells at depths ranging from 200 m to 500 m in the subarctic Pacific and has been used as an intermediate-water indicator (Matul and Abelmann, 2001; Nimmergut and Abelmann, 2002; Hays and Morley, 2003; Okazaki et al., 2003b, 2004, 2006; Itaki and Ikehara, 2004; Tanaka and

Q. Zhang et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 410 (2014) 337–350

Takahashi, 2005; Itaki et al., 2009). Plankton investigations in the Sea of Okhotsk show that high abundance of C. davisiana is related to the intermediate water characterized by low temperature as well as high oxygen and organic matter levels (Nimmergut and Abelmann, 2002; Okazaki et al., 2004; Abelmann and Nimmergut, 2005). Itaki et al. (2012) discussed the distributional patterns of C. davisiana in various oceans, including the Southern Ocean (Abelmann and Gowing, 1997), the North Atlantic (Petrushevskaya and Bjørklund, 1974) and the Japan/East Sea (Itaki, 2003), the Okhotsk Sea (Nimmergut and Abelmann, 2002; Okazaki et al., 2004), and the Bering Sea (e.g. Morley and Hays, 1983; Tanaka and Takahashi, 2005; Itaki et al., 2009) and pointed out that ventilation depth determines the living depth of C. davisiana in highlatitude oceans and availability of organic matter determines the species' abundance at that living depth. In general, the amount of organic matter decreases with increasing water depth; intermediate waters often contain plenty of organic matter and deep waters contain little organic matter. Thus, the high abundance of C. davisiana generally corresponds to a relative shallow ventilation at intermediate depths (200 m to 500 m), whereas low abundance is related to a deep penetration of the ventilation (deeper than 1000 m, such as in the North Atlantic and the Japan/East Sea) or to oxygen-poor conditions at intermediate depth (Itaki et al., 2012). Motoyama (1997) reports that Cycladophora davisiana was evolved from Cycladophora sakaii in the Northwest Pacific during 2.8–2.4 Ma and became a common species at 2.5 Ma. At Site U1340, morphotypic specimens of C. davisiana (fig. 18 in Plate II) first appeared at ~3.6 Ma, which was associated with sporadic occurrence of C. sakaii as well as a few intermediate forms between C. sakaii and C. davisiana, and it commonly occurred at ~3.0 Ma (Fig. 3). It seems that C. davisiana appears slightly earlier in the Bering Sea than in the Northwest Pacific, and this finding may indicate that the cold and well-ventilated intermediatewater in the subarctic Pacific, which is beneficial to the propagation of C. davisiana, was first formed in the Bering Sea. Based on the abundance profile of C. davisiana, we propose that the intermediate layer was stagnant from 4.15 Myr to 3.0 Myr because of the near absence of C. davisiana (Fig. 3). By contrast, the high abundance of C. davisiana indicates that active ventilation and increased formation of the intermediate layer occurred at around 3.0 Ma, and that the intermediate layer was associated with the relatively well-ventilated water and abundant organic matter after 1.07 Myr. In addition, the abundance of C. davisiana increasing during the periods from 2.75 Myr to 1.07 Myr and decreasing during the periods from 1.07 Myr to 0.47 Myr suggests that the penetration depth of the ventilation was becoming shallow after the NHG and then becoming deeper after the beginning of the MPT. As a useful tracer of the intermediate water in the subarctic Pacific, the relative abundance of Cycladophora davisiana has been used as a proxy for tracing the source region of the North Pacific Intermediate Water (NPIW). Based on this proxy, Ohkushi et al. (2003) suggested that the cold, well-ventilated NPIW was derived from the Bering Sea and not from the Okhotsk Sea during the last glacial. Subsequently, Tanaka and Takahashi (2005) deciphered the temporal and spatial changes in the source region for the NPIW during the past 100 kyr and also pointed out that the well-oxygenated intermediate water in the North Pacific was produced largely by the Bering Sea during the last glacial maximum (LGM). At Site U1340, we reached a similar interpretation, as the relative abundance of C. davisiana at around 12 Ka was very high at 39.84% to 46.03%; We also found that the relative abundances of C. davisiana exhibited extremely high average values of 74.6%, 51%, and 47.2% at glacial periods of ca. 0.85 Ma (MIS22), ca. 0.63 Ma (MIS16), and ca. 0.18 Ma (MIS6), respectively. These results were comparable to or even exceeded the values of the LGM (Fig. 6). Therefore, we propose that the Bering Sea was the main source region of NPIW during these glacial periods. 4.2.4. Deep-water condition (N500 m) Deep dwelling radiolarians in the subarctic include the Dictyophimus hirundo/crisiae group, Cyrtopera laguncula, and Siphocampe arachnea

347

Fig. 6. Relative and absolute abundance profiles of C. davisiana at Site U1340. The black solid line and the red dashed line represent the absolute and relative abundance of C. davisiana during LGM, respectively.

(Nimmergut and Abelmann, 2002; Abelmann and Nimmergut, 2005). Although ecological habits of deep water radiolarians are not well understood, D. hirundo/crisiae group has been shown to be closely related to the conditions of high oxygen levels or the intensified ventilation in deep water as well as the enhanced nutrient supply from highly productive surface water (Abelmann and Nimmergut, 2005), and it is supposed that the deep penetration of the ventilation is responsible for the increase in the number of D. hirundo/crisiae group both in the SW Okhotsk Sea and in the Bering Sea during the last glacial (Itaki et al., 2008, 2012). In contrast, it seems that S. arachnea is not sensitive to the oxygen levels compared to D. hirundo/crisiae group, because S. arachnea is a common species and mainly lives in deep waters below depth of 500 m in modern Bering Sea (Wang et al., 2005), approximately corresponding to the oxygen minimum zone (OMZ) of 500–1500 m in this region (Conkright et al., 2002; Okazaki et al., 2005). Although the abundance of S. arachnea varies in different periods at Site U1340, it generally shows relatively high abundance as a dominate species throughout the studied core. The absence of Dictyophimus hirundo/crisiae group during the period from 4.15 Myr to 3.91 Myr indicates that the deep water of the Bering Sea was characterized by low oxygen and nutrient levels. Then, the abundance of D. hirundo/crisiae group increased during the periods

348

Q. Zhang et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 410 (2014) 337–350

from 3.91 Myr to 2.75 Myr decreased during the period 2.75 Myr to 1.07 Myr and increased again from 1.07 Myr to 0.47 Myr. These trends indicate that deep water ventilation at the site went through a single cycle from strong to weak and back again to strong in the period 3.91 Ma to 0.47 Ma. This is consistent with ventilation depth increasingly shallow after the onset of the NHG and then deeper again after the beginning of the MPT, as revealed by the abundance variations in Cycladophora davisiana (see Section 4.2.3). The relatively high abundance of D. hirundo/crisiae group during the period from 3.0 Myr to 2.8 Myr suggests that active ventilation occurred in the deep water which is probably caused by the strong brine ejection due to enhanced formation of sea-ice around the onset of the NHG. This condition is similar to the phenomenon of the ventilation penetration in the subarctic Pacific reaching the 2000 m depth during the last glacial period (Keigwin, 1998; Matsumoto et al., 2002). Additionally, high abundance of D. hirundo/ crisiae group is consistent with lower values of the TS content (3.0 Myr to 2.8 Myr and the last 1.07 Myr) in the studied core (Takahashi et al., 2011) (Fig. 3). The reduction in the TS content generally indicates the increase in oxygen content at the seafloor (Jørgensen, 1982; Sageman et al., 1991), and this condition of high oxygen level is in favor of D. hirundo/crisiae group’s propagation.

As discussed above, the oceanographic conditions of various water layers in the Bering Sea changed with time during the last 4.15 Myr, as indicated in each stage of the radiolarian evolutions. Based on the variations in the vertical water-mass conditions reconstructed above, we established five oceanographic condition stages of the water column, and their major characteristics are summarized as follows. (1) 4.15 Myr to 3.91 Myr (Fig. 7A): a thick, warm, and saline surface layer with relatively high productivity, no sea-ice, and no formation of the cold dichothermal layer in subsurface water, stagnant intermediate water and low levels of oxygen and nutrient content in deep water; (2) 3.91 Myr to 2.75 Myr (Fig. 7B): relatively low-temperature and low-salinity surface water, enhanced formation of the cold dichothermal layer because of the development of the seasonal sea-ice, stagnant intermediate water for most of the period but actively ventilated at around 3.0 Ma, oxygen level of the deep water going through one phase from extremely low to high; (3) 2.75 Myr to 1.07 Myr (Fig. 7C): cold and low-salinity surface water with expansion of the sea-ice, weak formation of the cold dichothermal layer, intermediate water gradually becoming well ventilated, oxygen in the deep water becoming poor; (4) 1.07 Myr to 0.47 Myr (Fig. 7D): relatively cold and low-salinity surface water with large fluctuations in sea-ice expansion, gradual development of the

Fig. 7. The schematic of oceanographic condition changes in each water layer of the Bering Sea during the last 4.15 Myr. The black dashed lines in D and E denote the potential position of the coastline during the glacial periods according to the global sea-level change in the LGM (Hopkins, 1973; Chappell et al., 1996). Various numbers of the symbol □ in B, C, and D represent the variations in sea-ice volume.

Q. Zhang et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 410 (2014) 337–350

cold dichothermal layer, the ventilation of the intermediate water increasing and then becoming relatively weak, the oxygen level in deep water being initially low and then becoming high; (5). 0.47 Myr to Present (Fig. 7E): relatively warm and saline surface water with seasonal sea-ice formation, well-developed cold dichothermal layer in the subsurface water, and active ventilation in both the intermediate and deep waters generally but with frequent fluctuations in ventilation depth. Each layer through the water column developed well during this period and was comparable with those of the modern Bering Sea (Fig. 7E). We therefore concluded that the structure of the modern Bering Sea's water column was initially formed in ~ 0.47 Ma. 5. Conclusions (1) Based on the fluctuation patterns of the FRPs, the evolutionary history of the radiolarian assemblages at Site U1340 was divided into Stages I, II, III, and IV, with Stage IV divided into Substages IVa and IVb. The major characteristics of the radiolarian assemblages in each stage indicate that the oceanic environments of the Bering Sea went through two phases: from warm to extremely cold (Stages I to III) and then from cold to warm (Stages III to IVb). The boundaries between Stages I to IV in the studied core correspond to the climatic cooling event at ~ 3.91 Ma, the NHG at ~2.75 Ma, the beginning of the MPT at ~1.07 Ma and to the low-latitude radiolarian ecology event at ~ 0.47 Ma, respectively. (2) The oceanographic conditions of the various water layers in the Bering Sea since the late Early Pliocene were reconstructed according to the ecological habits of radiolarian dwellers at various depths. The results show that each water layer was significantly affected by the variations in oceanic environments. The surface layer (upper 50 m) was characterized by warm and saline water during the warm periods (4.15 Myr to 3.91 Myr and after 0.47 Ma) while being governed by the severe conditions of low temperature and low salinity during the cold periods (particularly during 2.75 Myr to 1.07 Myr). A cold dichothermal layer formed in the subsurface water (50 m to 200 m) during the periods when the seasonal sea-ice was developing strongly (3.91 Myr to 2.75 Myr and the last 1.07 Myr). The well-ventilated intermediate water (200 m to 500 m) was initially formed at the end of the stage II (~ 3.0 Ma) and intensified primarily during the cold periods (2.75 Myr to 1.07 Myr and some glacial periods after 1.07 Myr). Deep water (N500 m) was rich in oxygen mainly during the relatively warm periods after 1.07 Myr, except for the period from 3.0 Myr to 2.8 Myr. All layers through the water column developed well after 0.47 Myr, and were generally comparable to those of the modern Bering Sea. We therefore conclude that the vertical water-mass structure of the modern Bering Sea initially formed in ~ 0.47 Ma. (3) The relative abundance of Cycladophora davisiana, which is an effective proxy of the intermediate water in the subarctic Pacific, was extremely high with average values of 74.6%, 51%, and 47.2% at ca. 0.85 Ma (MIS22), ca. 0.63 Ma (MIS16), and ca. 0.18 Ma (MIS6), respectively. This finding indicates that the Bering Sea was the main source region of NPIW during these glacial periods, just as during the LGM. Acknowledgments Samples used in this research were provided by the Integrated Ocean Drilling Program (IODP). We are grateful to Prof. Thierry Corrège, Dr. John Rogers and an anonymous reviewer for their constructive suggestions that substantially contributed towards improving the quality of the initial manuscript. This work was supported by National Natural Science Foundation of China (Nos. 41076026, 41276051 and

349

91228207), the National Key Basic Research Program of China (No. 2013CB956102), and IODP-China.

References Abelmann, A., 1992. Radiolarian flux in Antarctic waters (Drake Passage, Powell Basin Bransfield Strait). Polar Biol. 12, 357–372. Abelmann, A., Gowing, M.M., 1997. Spatial distribution of living polycystine radiolarian taxa-baseline study for paleoenvironmental reconstructions in the Southern Ocean (Antarctic sector). Mar. Micropaleontol. 30, 3–28. Abelmann, A., Nimmergut, A., 2005. Radiolarians in the Sea of Okhotsk and their ecological implication for paleoenvironmental reconstructions. Deep-Sea Res. II 52, 2302–2331. Abelmann, A., Gersonde, R., Spiess, V., 1990. Pliocene–PIeistocene paleoceanography in the Weddell Sea—Siliceous microfossil evidence. In: Bleil, U., Thiede, J. (Eds.), Geological history of the polar oceans: Arctic versus antarctic. Kluwer, Dordrecht, pp. 729–759. Anderson, O.R., 1983. Radiolaria. Springer-Verlag, New York, p. 355. Anderson, C., 1997. Transfer function vs. modern analog technique for estimating Pliocene sea-surface temperatures based on planktonic foraminiferal data, western equatorial Pacific Ocean. J. Foraminifer. Res. 27, 123–132. Barron, J.A., 1992. Paleoceanographic and tectonic controls on the Pliocene diatom record of California. In: Tsuchi, R., lngle Jr., J.C. (Eds.), Pacific neogene. Univ. Tokyo Press, Tokyo, p. 2542. Bjørklund, K.R., Kruglikova, S.B., 2003. Polycystine radiolariansin surface sediments in the Arctic Ocean basins and marginal seas. Mar. Micropaleontol. 49, 231–273. Bjørklund, K.R., Cortese, G., Swanberg, N., Schrader, H.J., 1998. Radiolarian faunal provinces in surface sediments of the Green-land, Iceland and Norwegian (GIN) seas. Mar. Micropaleontol. 35, 105–140. Blueford, J.R., 1983. Distribution of Quaternary Radiolaria in the Navarin Basin geologic province, Bering Sea. Deep Sea Res. Part A 30, 763–781. Bowen, D.Q., Rose, J., McCabe, A.M., Sutherland, D.G., 1986. Quaternary Glaciations in England, Ireland, Scotland, and Wales. Quat. Sci. Rev. 5, 299–340. Candy, I., Coope, G.R., Lee, J.R., Parfitt, S.A., Preece, R.C., Rose, J., Schreve, D.C., 2010. Pronounced warmth during early Middle Pleistocene interglacials: Investigating the Mid-Brunhes Event in the British terrestrial sequence. Earth Sci. Rev. 103, 183–196. Carter, R.M., 2005. A New Zealand climatic template back to c. 3.9 Ma: ODP Site 1119, Canterbury Bight, south-west Pacific Ocean, and its relationship to onland successions. J. R. Soc. N. Z. 35, 9–42. Casey, R., 1977. Distribution of polycystine Radiolaria in the oceans in relation to physical and chemical conditions. In: Funnell, B.M., Riedel, W.R. (Eds.), The micropaleontology of the oceans. Cambridge University Press, Cambridge, pp. 151–159. Casey, R., Spaw, J., Kunze, F., Reynolds, R., Duis, T., McMillen, K., Pratt, D., Anderson, V., 1979. Radiolarian ecology and the development of the radiolarian component in Holocene sediments, Gulf of Mexico and adjacent seas with potential paleontological applications. Trans. Gulf Coast Assoc. Geol. Soc. 29, 228–237. Chappell, J., Omura, A., Esat, T., Mcculloch, M., Pandolfi, J., Ota, Y., Pillans, B., 1996. Reconciliation of late Quaternary sea levels derived from coral terraces Huon Peninsula with deep sea oxygen isotope records. Earth Planet. Sci. Lett. 141, 227–236. Ciesielski, P.F., Grinstead, G.P., 1986. Pliocene variations in the position of the Antarctic Convergence in the southwest Atlantic. Paleoceanography 1, 197–232. Clark, P.U., Archer, D., Pollard, D., Blum, J.D., Rial, J.A., Brovkin, V., Mix, A.C., Pisias, N.G., Roy, M., 2006. The middle Pleistocene transition: Characteristics, mechanisms, and implications for long-term changes in atmospheric pCO2. Quat. Sci. Rev. 25, 3150–3184. Conkright, M.E., Locarnini, R.A., Garcia, H.E., O'Brien, T.D., Boyer, T.P., Stephens, C., Antonov, J.I., 2002. World Ocean Atlas 2001: Objective analyses, data statistics, and figures: CD-ROM documentation. National Oceanographic Data Center, Silver Spring, MD, p. 17. Dowsett, H.J., Cronin, T.M., Poore, R.Z., Thompson, R.S., Whatley, R.C., Wood, A.M., 1992. Micropaleontological evidence for increased meridional heat transport in the North Atlantic Ocean during the Pliocene. Science 258, 1133–1135. Dowsett, H.J., Barron, J.A., Poore, R., 1996. Middle Pliocene sea surface temperatures: A global reconstruction. Mar. Micropaleontol. 27, 13–15. Gorbarenko, S.A., 1996. Stable isotope and lithologic evidence of late-glacial and Holocene oceanography of the northwestern Pacific and its marginal seas. Quat. Res. 46, 230–250. Gorbarenko, S.A., Basov, I.A., Chekhovskaya, M.P., Southon, J.R., 2005. Orbital and millennium scale environmental changes in the southern Bering Sea during Last Glacial–Holocene: Geochemical and paleontological evidence. Deep-Sea Res. II 52, 2174–2185. Gorbarenko, S.A., Wang, P., Wang, R., Cheng, X., 2010. Orbital and suborbital environmental changes in the southern Bering Sea during the last 50kyr. Palaeogeogr. Palaeoclimatol. Palaeoecol. 286, 97–106. Hays, J.D., Morley, J.J., 2003. The Sea of Okhotsk: A window on the ice age ocean. Deep-Sea Res. I 51 (4), 1481–1506. Heusser, L.E., Morley, J.J., 1996. Pliocene climate of Japan and environs between 4.8 and 2. 8 Ma: A joint pollen and marine faunal study. Mar. Micropaleontol. 27, 85–106. Hodell, D., Kennett, J.P., 1986. Late Miocene-Early Pliocene stratigraphy and paleoceanography of the South Atlantic and southwest Pacific Oceans: A synthesis. Paleoceanography 1, 285–311. Hopkins, D.M., 1973. Sea level history in Beringia during the past 250,000 years. Quat. Res. 3, 520–540.

350

Q. Zhang et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 410 (2014) 337–350

Itaki, T., 2003. Depth-related radiolarian assemblage in the water-column and surface sediments of the Japan Sea. Mar. Micropaleontol. 47, 253–270. Itaki, T., Ikehara, K., 2004. Middle to late Holocene changes of the Okhotsk Sea intermediate water and their relation to atmospheric circulation. Geophys. Res. Lett. 31, L24309. http://dx.doi.org/10.1029/2004GL021384. Itaki, T., Komatsu, N., Motoyama, I., 2007. Orbital- and millennial-scale changes of radiolarian assemblages during the last 220 kyrs in the Japan Sea. Palaeogeogr. Palaeoclimatol. Palaeoecol. 247, 115–130. Itaki, T., Khim, B.K., Ikehara, K., 2008. Last glacial–Holocene water structures in the southwestern Okhotsk Sea inferred from radiolarian assemblages. Mar. Micropaleontol. 67, 191–215. Itaki, T., Uchida, M., Kim, S., Shin, H.-S., Tada, R., Khim, B.-K., 2009. Late Pleistocene stratigraphy and palaeoceanographic implications in northern Bering Sea slope sediments: Evidence from the radiolarian species Cycladophora davisiana. J. Quat. Sci. 24, 856–865. Itaki, T., Kim, S., Rella, S.F., Uchida, M., Tada, R., Khim, B.-K., 2012. Millennial-scale variations of late Pleistocene radiolarian assemblages in the Bering Sea related to environments in shallow and deep waters. Deep-Sea Res. II 61, 127–144. http://dx.doi.org/10. 1016/j.dsr2.2011.03.002. Jansen, J.H.F., Kuijpers, A., Troelstra, S.R., 1986. A mid-Brunhes climatic event: Long-term changes in global atmosphere and ocean circulation. Science 232, 619–622. Jørgensen, B.B., 1982. Mineralization of organic matter in the sea bed—the role of sulphate reduction. Nature 296, 243–245. Kamikuri, S.I., Nishi, H., Motoyama, I., 2007. Effects of late Neogene climatic cooling on North Pacific radiolarian assemblages and oceanographic conditions. Palaeogeogr. Palaeoclimatol. Palaeoecol. 249, 370–392. Keigwin, L.D., 1998. Glacial-age hydrography of the far northwest Pacific Ocean. Paleoceanography 13, 323–339. Khusid, T.A., Basov, I.A., Gorbarenko, S.A., Chekhovskaya, M.P., 2006. Benthic foraminifers in Upper Quaternary sediments of the Southern Bering Sea: Distribution and paleoceanographic interpretations. Stratigr. Geol. Correl. 14, 538–548. Kling, S.A., 1979. Vertical distribution of polycystine radiolarians in the central North Pacific. Mar. Micropaleontol. 4, 295–318. http://dx.doi.org/10.1016/0377-8398(79) 90022-7. Kling, S.A., Boltovskoy, D., 1995. Radiolarian vertical distribution patterns across the southern California Current. Deep-Sea Res. I Oceanogr. Res. Pap. 42, 191–231. Kruglikova, S.B., 1969. Radiolarians in the surface layer of the sediments of the northern half of the Pacific Ocean. In: Kort, P.P. (Ed.), the Pacific Ocean, Microflora and Microfauna in the Recent of Pacific Ocean, pp. 48–72 (Nauka, Moscow, in Russian). Lee, J.R., Rose, J., Hamblin, R.J., Moorlock, B.S., 2004. Dating the earliest lowland glaciation of eastern England: A pre-MIS 12 early Middle Pleistocene Happisburgh glaciation. Quat. Sci. Rev. 23, 1551–1566. Ling, H.Y., Stadum, C.J., Welch, M.L., 1971. Polycystine radiolaria from Bering Sea surface sediments. In: Farinacci, A. (Ed.), Proceeding of the Second Planktonic Conference, Roma, Tecnoscienza, pp. 705–729. Marlow, J.R., Lange, C.B., Weger, G., Rosell-Mele, A., 2000. Upwelling intensification as part of the Pliocene-Pleistocene climate transition. Science 290, 2288–2291. Maslin, M.A., Haug, G.H., Sarnthein, M., Tiedemann, R., 1996. The progressive intensification of Northern Hemisphere glaciation as seen from the North Pacific. Geol. Rundsch. 85, 452–465. Matsumoto, K., Oba, T., Lynch-Stieglitz, J., Yamamoto, H., 2002. Interior hydrography and circulation of the glacial Pacific Ocean. Quat. Sci. Rev. 21, 1693–1704. Matul, A., Abelmann, A., 2001. Quaternary water structure of the Sea of Okhotsk based on radiolarian data. Dokl. Earth Sci. 381, 1005–1007. Morley, J.J., Hays, J.D., 1983. Oceanographic conditions associated with high abundances of the radiolarian Cycladophora davisiana. Earth Planet. Sci. Lett. 66, 63–72. http:// dx.doi.org/10.1016/0012-821X (83) 90126-7. Morley, J.J., Robinson, S.W., 1986. Improved method for correlating late Pleistocene/Holocene records from the Bering Sea: Application of a biosiliceous/geochemical stratigraphy. Deep-Sea Res. 33, 1203–1211. Motoyama, I., 1997. Origin and evolution of Cycladophora davisiana Ehrenberg (Radiolaria) in DSDP Site 192, Northwest Pacific. Mar. Micropaleontol. 30, 45–63. Motoyama, I., Nishimura, A., 2005. Distribution of radiolarians in North Pacific surface sediments along the 17 5° E meridian. Paleontol. Res. 9, 95–117. Mudelsee, M., Schulz, M., 1997. The Mid-Pleistocene climate transition: Onset of 100 ka cycle lags ice volume build-up by 280 ka. Earth Planet. Sci. Lett. 151, 117–123. Nakatsuka, T., Watanabe, K., Handa, N., Matsumoto, E., Wada, E., 1995. Glacial to interglacial surface nutrient variations to Bering deep basins recorded by δ 13C and δ 15 N of sedimentary organic matter. Paleoceanography 10, 1047–1061. Nigrini, C., 1991. Composition and biostratigraphy of radiolarian assemblages from an area of upwelling (northwestern Arabian Sea, Leg 117). Proceedings of the Ocean Drilling Program. Scientific Results, 117, pp. 89–126. Nimmergut, A., Abelmann, A., 2002. Spatial and seasonal changes of radiolarian standing stocks in the Sea of Okhotsk. Deep-Sea Res. I Oceanogr. Res. Pap. 49, 463–493. Ohkushi, K., Itaki, T., Nemoto, N., 2003. Last Glacial-Holocene change in intermediatewater ventilation in the Northwestern Pacific. Quat. Sci. Rev. 22, 1477–1484. Okazaki, Y., Takahashi, K., Yoshitani, H., Nakatsuka, T., Ikehara, M., Wakatsuchi, M., 2003a. Radiolarians under the seasonally sea-ice covered conditions in the Okhotsk Sea: Flux and their implications for paleoceanography. Mar. Micropaleontol. 49, 195–230. Okazaki, Y., Takahashi, K., Nakatsuka, T., Honda, M.C., 2003b. The production scheme of Cycladophora davisiana (Radiolaria) in the Okhotsk Sea and the northwestern North Pacific: Implication for the paleoceanographic conditions during the glacials in the high latitude oceans. Geophys. Res. Lett. 30, 1939. http://dx.doi.org/10.1029/ 2003GL018070.

Okazaki, Y., Takahashi, K., Itaki, T., Kawasaki, Y., 2004. Comparison of radiolarian vertical distributions in the Okhotsk Sea near Kuril Islands and the northwestern North Pacific off Hokkaido Island. Mar. Micropaleontol. 51, 257–284. Okazaki, Y., Takahashi, K., Asahi, H., Katsuki, K., Hori, J., Yasuda, H., Tokuyama, H., 2005. Productivity changes in the Bering Sea during the late Quaternary. Deep-Sea Res. II Top. Stud. Oceanogr. 52, 2150–2162. Okazaki, Y., Seki, O., Nakatsuka, T., Sakamoto, T., Ikehara, M., Takahashi, K., 2006. Cycladophora davisiana (Radiolaria) in the Okhotsk Sea: A key for reconstructing glacial ocean conditions. J. Oceanogr. 62, 639–648. Petrushevskaya, M.G., Bjørklund, K.R., 1974. Radiolarians in Holocene sediments ofthe Norwegian-Greenland Seas. Sarsia 57, 33–46. Pisias, N.G., Moore Jr., T.C., 1981. The evolution of Pleistocene climate: A time series approach. Earth Planet. Sci. Lett. 52, 450–458. Raymo, M.E., 1994. The initiation of Northern Hemisphere glaciation. Annu. Rev. Earth Planet. Sci. 22 (1), 353–383. Robertson, J.H., 1975. Glacial to interglacial oceanographic changes in the north-west Pacific, including a continuous record of the last 400,000 years. Ph.D. Thesis Columbia University, New York, p. 355. Rose, J., Allen, P., Kemp, R.A., Whiteman, C.A., Owen, N., 1985. The early Anglian Barham soil of eastern England. In: Boardman, J. (Ed.), Soils and quaternary landscape evolution. Wiley, Chichester, pp. 197–230. Sageman, B.B., Wignall, P.B., Kauffman, E.G., 1991. Biofacies models for oxygen-deficient facies in epicontinental seas: Tool for paleoenvironmental analysis. In: Einsele, Gerhard, Ricken, Werner, Seilacher, Adolf (Eds.), Cycles and events in stratigraphy. Springer-Verlag, Berlin, pp. 542–564. Sancetta, C., 1983. Effect of Pleistocene glaciation upon oceanographic characteristics of the North Pacific Ocean and Bering Sea. Deep-Sea Res. 30, 851–869. Sancetta, C., Robinson, S.W., 1983. Diatom evidence on Wisconsin and Holocene events in the Bering Sea. Quat. Res. 20, 232–245. Scholl, D.W., Creager, J.S., 1973. Geologic synthesis of Leg 19 (DSDP) results: Far north Pacific and Aleutian Ridge, and Bering Sea. Initial Reports of the Deep Sea Drilling Project 19. U.S. Govt. Printing Office, Washington, D.C, pp. 897–913. Shackleton, N.J., 1987. Oxygen isotopes, ice volume and sea level. Quat. Sci. Rev. 6, 183–190. Shackleton, N.J., Opdyke, N.D., 1973. Oxygen isotope and palaeomagnetic stratigraphy of the Equatorial Pacific core V28–238: Oxygen isotope temperatures and ice volumes on a 105 and 106 year scale. Quat. Res. 3, 39–55. Shannon, C.E., Wiener, W., 1949. The Mathematical Theory of Communication. University of Illinois Press, Urbana, p. 125. Springer, A.M., McROY, C.P., FLINT, M.V., 1996. The Bering Sea Green Belt: Shelf-edge processes and ecosystem production. Fish. Oceanogr. 5, 205–223. Takahashi, K., 1997. Time-series fluxes of Radiolaria in the eastern subarctic Pacific Ocean. News Osaka Micropaleontol. 10, 299–309. Takahashi, K., 1998. The Bering Sea and Okhotsk Sea: Modern and past paleoceanographhic changes and gateway impact. J. Asian Earth Sci. 16, 49–58. Takahashi, K., 1999a. The Okhotsk and Bering Seas: Critical marginal seas for the land– ocean linkage, land–sea link in Asia. In: Saito, Y., Ikehara, K., Katayama, H. (Eds.), Proceedings of an International Workshop on Sediment Transport and Storage in Coastal Sea-Ocean System. JAMSTEC and Geological Survey of Japan, Tsukuba, pp. 341–353. Takahashi, K., 1999b. Paleoceanographic changes and present environment of the Bering Sea. In: Loughlin, T.R., Ohtani, K. (Eds.), Dynamics of the Bering Sea. University of Alaska Sea Grant, Fairbanks, pp. 365–385. Takahashi, K., 2005. The Bering Sea and paleoceanography. Deep-Sea Res. II 52, 2080–2091. Takahashi, K., Ravelo, A.C., Alvarez Zarikian, C.A., the IODP Expedition 323 scientists, 2011. Expedition reports: Bering sea paleoceanography. Proceedings of the Integrated Ocean Drilling Program, Volume 323. Integrated Ocean Drilling Program Management International, Inc, for the Integrated Ocean Drilling Program. http://dx.doi. org/10.2204/iodp.proc.323.104.2011. Tanaka, S., Takahashi, K., 2005. Late Quaternary paleoceanographic changes in the Bering Sea and the western subarctic Pacific based on radiolarian assemblage. Deep-Sea Res. II 52, 2131–2149. Tian, J., Pak, D.K., Wang, P., Lea, D., Cheng, X., Zhao, Q., 2006. Late Pliocene monsoon linkage in the tropical South China Sea. Earth Planet. Sci. Lett. 252, 72–81. Tomczak, M., Godfrey, J.S., 1994. Regional oceanography: An introduction. Pergamon, Oxford, p. 422. Wang, R.J., Chen, R.H., 2005. Cycladophora davisiana (radiolarian) in the Bering Sea during the late Quaternary: A stratigraphic tool and proxy of the glacial Subarctic Pacific Intermediate Water. Sci. China Ser. D Earth Sci. 48, 1698–1707. Wang, P., Tian, J., Cheng, X., Liu, C., Xu, J., 2003. Carbon reservoir changes preceded major ice-sheet expansion at the mid-Brunhes event. Geology 31, 239–242. Wang, R.J., Chen, R.H., Xiao, W.S., 2005. Depth distribution pattern of radiolarians in surface sediments of the Bering Sea and their oceanographic implications. Acta Micropalaeontol. Sin. 22, 127–135 (in Chinese with English abstract). Wang, R.J., Xiao, W.S., Li, Q.Y., Chen, R.H., 2006. Polycystine radiolarians in surface sediments from the Bering Sea Green Belt area and their ecological implication for paleoenvironmental reconstructions. Mar. Micropaleontol. 59, 135–152. Yang, L.H., Chen, M.H., Wang, R.J., Zhen, F., 2002. Radiolarian record to paleoecological environment change events over the past 1.2 MaBP in the southern South China Sea. Chin. Sci. Bull. 47, 1478–1483. Zhang, Q., Chen, M.H., Zhang, L.L., Wang, R.J., Xiang, R., Hu, W.F., 2014. Radiolarian Biostratigraphy in the Southern Bering Sea since Pliocene. Sci. China Earth Sci. 57, 682–692. Zou, J.J., Shi, X.F., Bai, Y.Z., Zhu, A.M., Chen, Z.H., Huang, Y.H., 2012. Paleoenvironment and Paleoproductivity Variations in the Bering Sea Since the Last Deglacial. Earth Sci. J. China Univ. Geosci. 37 (Supplement), 1–10 (in Chinese with English abstract).