Mid-Pliocene Asian Monsoon Intensification and the Onset of Northern Hemisphere Glaciation

Mid-Pliocene Asian monsoon intensiﬁ cation and the onset of Northern Hemisphere glaciation

Yi Ge Zhang1, 2*, Junfeng Ji1, William Balsam3, Lianwen Liu1, and Jun Chen1 1State Laboratory for Mineral Deposits Research, Institute of Surﬁ cial Geochemistry, Department of Earth Sciences, Nanjing University, Nanjing 210093, China 2Department of Marine Sciences, University of Georgia, Athens, Georgia 30602, USA 3Department of Earth and Environmental Sciences, University of Texas, Arlington, Texas 76019, USA

ABSTRACT Here we present a high-resolution hematite to goethite ratio (Hm/ The late Pliocene onset of major Northern Hemisphere glacia- Gt), an Asian monsoon precipitation proxy (Zhang et al., 2007), for the tion (NHG) is one of the most important steps in the Cenozoic global last fi ve million years, reconstructed from southern South China Sea cooling. Although most attempts have been focused on high-latitude sediments. This record enabled us to reveal the details of the variability climate feedbacks, no consensus has been reached in explaining the of key low-latitude climate systems over critical climate change inter- forcing mechanism of this dramatic climate change. Here we present vals, and to explore its possible interplay with the NHG initiation. a key low-latitude climate record, the high-resolution Asian monsoon precipitation variability for the past fi ve million years, reconstructed MATERIALS AND METHODS from South China Sea sediments. Our results, with supporting evi- Samples were recovered at Ocean Drilling Program (ODP) Site dence from other records, indicate signifi cant mid-Pliocene Asian 1143 (9°21.72′N, 113°17.11′E; 2777 m water depth), the same core we monsoon intensifi cation, preceding the initiation of NHG at ca. 2.7 used in our previous study in which a transfer function was established Ma ago. This 1.4-million-year-long monsoon intensifi cation probably (Zhang et al., 2007), enabling us to calculate hematite and goethite con- enhanced monsoon-induced Asian continental erosion and chemical centration in sediment samples from diffuse refl ectance spectroscopy weathering and in the process left fi ngerprints in marine calcium iso- (DRS). We applied this technique to ODP 1143 samples covering the topes. Furthermore, increased rock weathering and/or organic car- past 600 ka. Here we report on an additional 1807 samples measured bon burial probably lowered the contemporary atmospheric CO2 and using a Perkin Elmer Lambda 900 diffuse refl ectance spectrophotom- may have triggered the NHG onset. eter, extending the Hm/Gt record back to 5 Ma ago. Sample preparation, analysis, and data processing were exactly the same as previously INTRODUCTION described (Zhang et al., 2007). DRS results and hematite and goethite Although ice sheet buildup in the Northern Hemisphere began ~11 concentration are listed in Table DR1 in the GSA Data Repository.1 million years ago, sustained major Northern Hemisphere glaciation (NHG) The benthic δ18O of samples from this site has been tuned to Earth’s did not occur until ca. 2.7 Ma ago (Raymo, 1994; Shackleton et al., 1984). orbit, and together with magnetostratigraphic and biostratigraphic age This was the fi nal step in the Cenozoic global cooling trend, switching controls, yields an astronomically calibrated, high-resolution chronol- Earth from a greenhouse world to an icehouse world with periodical wan- ogy for ODP 1143 (Tian et al., 2002). According to this chronology, the ing and waxing of ice sheets. To understand what drives this important cli- 2122 samples used in this study (including 315 from Zhang et al., 2007) mate change, a number of hypotheses have been proposed, most of which encompass the past fi ve million years with an average resolution of 2.4 focused on the climatic infl uence of high latitudes, because high latitudes ka between each sample. are the places where ice sheets occur. For example, enhanced North Atlan- tic Deep Water production driven by tectonic process (“Panama Hypoth- PLIO-PLEISTOCENE MONSOON PRECIPITATION esis”; Haug and Tiedemann, 1998; Keigwin, 1982) and stratifi cation in VARIATIONS the subarctic Pacifi c (Haug et al., 2005) have been interpreted to increase Hm/Gt is interpreted as a proxy for the variability of Asian monsoonal precipitation, favoring ice sheet formation and accumulation in Northern precipitation in SE Asia because during the soil formation process in the Hemisphere high latitudes; favorable Milankovitch orbital parameters Mekong Basin, hematite and goethite formation are considered competitive; were also thought to facilitate the continental glacier buildup (Maslin et i.e., dry conditions are favorable for hematite formation, whereas humid al., 1998). However, these scenarios are still being debated, and the ulti- conditions are favorable for goethite formation (Ji et al., 2004; Kampf and mate driving force of NHG initiation is by no means resolved (Bartoli et Schwertmann, 1983; Schwertmann, 1988; Schwertmann and Murad, 1983), al., 2005; Klocker et al., 2005; Lunt et al., 2008a, 2008b; Molnar, 2008). and their ratio preserved in ODP 1143 via fl uvial and marine transportation Changes in low-latitude climate systems such as the Asian mon- would be precipitation-dependent (Zhang et al., 2007). Our previous study soon were generally thought unimportant when studying NHG onset. reveals that in ODP 1143 sediments, Hm/Gt provides excellent recording However, recent studies highlighted the key role of low-latitude climate of monsoonal precipitation variability because (1) iron oxide minerals in dynamics in regulating global climate changes. For example, the Asian ODP 1143 are highly climate-sensitive, (2) monsoonal precipitation, rather monsoon appears to be linked to El Niño Southern Oscillations in the than temperature, predominantly controls the Hm/Gt ratio, and (3) the Fe equatorial Pacifi c and has a profound infl uence on extratropical climate oxide signal is well preserved with little reduction/reoxidation after deposi- (Zhang et al., 2007). The monsoon was also thought to be capable of tion (Zhang et al., 2007). And our Hm/Gt record is comparable to stalagmite affecting global biogeochemical cycling of carbon on geological tim- δ18O records from South China (see Zhang et al., 2007, their fi gure 4). escales (Wang et al., 2003, 2004). Uncovering the history of the Asian monsoon and the role it played in the shift to an icehouse climate could 1GSA Data Repository item 2009140, diffuse refl ectance spectroscopy be crucial to understanding global climate in general. results for ODP 1143 samples, is available online at www.geosociety.org/pubs/ ft2009.htm, or on request from [email protected] or Documents Secretary, *E-mail: [email protected]. GSA, P.O. Box 9140, Boulder, CO 80301, USA.

© 2009 Geological Society of America. For permission to copy, contact Copyright Permissions, GSA, or [email protected]. GEOLOGY,Geology, July July 2009; 2009 v. 37; no. 7; p. 599–602; doi: 10.1130/G25670A.1; 3 fi gures; Data Repository item 2009140. 599 δ18 Δδ18 As we extend the Hm/Gt record from late Pleistocene to early Plio- ifera O ( Ob-p) from the same core (Fig. 1B). The signifi cant increase Δδ18 cene, one may argue that the Mekong River northern headwaters extend up in Ob-p in ODP 1143 from 4.2 to 2.7 Ma ago was interpreted to result all the way into the tectonically active areas of SW China. Late Cenozoic from an enhanced monsoon precipitation that diluted the surface seawa- uplift in this area was recorded, and the mountain building was explained ter with relatively 18O-depleted rainwater (Fig. 1B) (Tian et al., 2004). by the lithospheric intrusion of the Tibetan Plateau (Burchfi el, 2004; Roy- Furthermore, this monsoon increase was indicated by other records from den et al., 2008). The very recent tectonic events in the Mekong drainage a variety of locations. In Figure 1 the Hm/Gt record from ODP 1143 area may alter the chemical runoff down the Mekong River by the balance (Fig. 1A) is compared to magnetic susceptibility of the Chinese Loess– between material eroded in the cooler/drier uplifted headwater regions ver- Red Clay sequences from central China (Fig. 1C) (Sun et al., 2006), and sus the warmer/wetter more stable downstream regions in the Indochina Indian Ocean mean sediment fl ux (Fig. 1D) (Rea, 1992). peninsula, challenging the idea that the Hm/Gt record in the South China The Chinese Loess Plateau located in central China is covered Sea is a monsoon proxy. However, we argue that the tectonic infl uence on by thick wind-deposited Chinese Loess–Red Clay sequences. These Hm/Gt is limited because the lower reaches of the Mekong—Cambodia and sequences provide a continuous continental climate record and have been Vietnam—that were relatively tectonically stable contribute most of the Fe used extensively for Asian monsoon reconstruction. After dust deposition, oxides in ODP 1143. In addition, hematite and goethite, both of which are the monsoonal precipitation-induced pedogenic processes that enhanced secondary minerals, are low in concentration in the unweathered rocks or magnetic mineral (e.g., magnetite, maghemite) formation resulted in an weakly weathered soils that characterized the upper reaches of the Mekong increase in magnetic susceptibility (Zhou et al., 1990). Although soil mag- (Liu et al., 2005). Instead, high annual temperature (~27 °C), precipitation netic susceptibility can be affected by a variety of factors, magnetic sus- (~3000 mm/a), and chemical weathering rate in the downstream Mekong ceptibilities in Chinese Loess–Red Clay sequences (Fig. 1C) are generally suggest it very likely will produce the majority of Fe oxide minerals. interpreted as a summer monsoon proxy (Sun et al., 2006). Hm/Gt variations indicate that Asian monsoon evolution for the past 5 Indian Ocean mean sediment fl ux (Rea, 1992) is employed to moni- Ma can be divided into three periods (Fig. 1): (1) early Pliocene (5.0–4.2 Ma tor the erosion rate in South Asia and the transportation of terrigenous ago) monsoon decline, (2) signifi cant monsoon intensifi cation during the materials to the Bengal Fan, thought to be infl uenced by both the Asian mid-Pliocene (4.2–2.7 Ma ago), and (3) slow decline with higher variability monsoon and the tectonics of the Himalayan-Tibetan Plateau (Rea, 1992). after the onset of NHG (2.7 Ma ago to present). These observations are gen- Although the four different records from diverse locations vary in erally consistent with previous studies based on Chinese Loess (An et al., details as we expected, monsoon intensifi cation during the mid-Pliocene 2001). Among these three periods, the mid-Pliocene monsoon enhancement (ca. 4.2–2.7 Ma ago) is clearly exhibited in each one of them (Fig. 1, is highlighted here because it may have an impact on NHG initiation. highlighted with a shaded bar). These records confi rm that mid-Pliocene The Hm/Gt-revealed mid-Pliocene monsoon rainfall rise in SE Asia monsoon intensifi cation occurred in all major Asian monsoon areas, and is supported by the differences between benthic and planktonic foramin- consequently enhanced erosion of the Asian continent and increased terrigenous material production/transportation.

0.0 GEOCHEMICAL IMPLICATIONS OF THE RISING MONSOON ODP 1143 A The Asian monsoon dominates in vast areas of East, South, and 0.1 Southeast Asia. During the late Cenozoic, these regions were characterized by the highest suspended and dissolved load in river discharges to

Hm/Gt 0.2 the ocean on Earth (Raymo and Ruddiman, 1992). Those suspended and

0.3 dissolved loads were the product of rock erosion and chemical weathering. Thus, the evolution of the Asian monsoon would be critical in regulat- B 5.5 ing both regional and global weathering and the geochemical cycling of South China Sea

ODP 1143 (‰, PDB) important elements. For example, nine major rivers in the Asian monsoon b-p

O area contribute more than one-third of the global riverine dissolved Ca 300 18 Zhaojiachuan Δδ ﬂ ux to the ocean (data from Gaillardet et al., 1999) (Fig. 2). The Ca cycle 250 5.0 Loess-Red Clay is also an important player in global weathering, a player that is most sen- 200 ) -1 sitive to atmospheric CO because Ca-Mg silicate weathering is the largest 150 C 2 kg 3

m sink for CO on geological time scales (Berner and Berner, 1997). -8 100 2

(10 The global Ca cycle can be better constrained by Ca isotopes. For 50 Ca the most commonly used isotope pairs are 44Ca and 40Ca, which Magnetic Susceptibility 0 0.8 are the stable isotopes with highest natural abundance, with the delta Chinese Loess Plateau HG notation expressed as δ44Ca. Some workers also use 44Ca and 42Ca pairs 0.6 (δ44/42Ca) because of the interference at mass 40 from argon in multicol- /ka) 2 lector inductively coupled plasma–mass spectrometry (MC-ICP-MS)

OnsetofN D 0.4 δ44 Indian Ocean (g/cm (e.g., Sime et al., 2007). Ca records from both bulk marine calcite Integrated (e.g., De La Rocha and DePaolo, 2000; Fantle and DePaolo, 2005, 2007) Indian Ocean 0.2 Mean sediment flux and foraminiferal shells (e.g., Heuser et al., 2005; Sime et al., 2007) have 0 1000 2000 3000 4000 5000 been reconstructed for the late Cenozoic. Since the residence time of Age (ka) Ca in the ocean is ~106 years and the mixing time of the oceans is ~103 ΔΔδδ18 years, the oceans are likely to be homogeneous with respect to Ca, and Figure 1. Comparison of Hm/Gt (A) and Ob-p (B) from ODP 1143, South China Sea (Tian et al., 2004); magnetic susceptibility (C) from the isotopic record at a particular site may be considered a global sig- Zhaojiachuan Loess–Red Clay sequences, Chinese Loess Plateau nal (Fantle and DePaolo, 2005). However, an interesting observation is (Sun et al., 2006); Indian Ocean integrated mean sediment ﬂ ux (D) (Rea, δ44 1992). The mid-Pliocene (4.2–2.7 Ma) monsoon intensiﬁ cation is high- that the Ca records from bulk carbonates and foraminifera are appar- lighted with a shaded bar; the dashed line indicates the onset of NHG ently inconsistent. Although this discrepancy has not been successfully at ca. 2.7 Ma. ODP—Ocean Drilling Program; PDB—Peedee belemnite. explained (Sime et al., 2007), existing carbonate δ44Ca records from dif-

600 GEOLOGY, July 2009 ferent sediment cores around the world are generally consistent (De La differ signifi cantly in time resolution. Particularly, the most remarkable Rocha and DePaolo, 2000; Fantle and DePaolo, 2005, 2007), corrobo- increase of monsoon precipitation and δ44Ca both occur during the mid- rating the idea that bulk carbonate δ44Ca might be a reasonable recorder Pliocene, from ca. 4.2 to 2.7 Ma ago (Fig. 3). of the Ca isotope signature of global seawater. To understand the general correlation between Asian monsoon pre- The bulk carbonate δ44Ca record from Deep Sea Drilling Project cipitation and marine δ44Ca over the last fi ve million years, identifying (DSDP) Site 590 for the past 5 Ma (Fantle and DePaolo, 2005) is pre- the controlling factors of seawater Ca isotope composition variations sented in Figure 3, and this record is compared to our monsoon precipita- would be helpful. Although the interpretation of marine Ca isotope tion proxy Hm/Gt (Fig. 3). Most strikingly, Asian monsoon precipitation records is an unsettled issue, and less well constrained hydrothermal Ca recorded in South China Sea Hm/Gt records and marine carbonate δ44Ca inputs complicate the problem even more, a recent model indicated that variations for the last 5 Ma seem to be correlated, although the two records small changes in mean Ca isotope values of riverine Ca input could read- ily explain the observed δ44Ca variations in the Neogene ocean (Sime et al., 2007). This study highlights the role of river discharge and hence the δ44Ca role of precipitation in the Ca cycle. 1.0 Rivers in the Asian monsoon region are characterized by high Ca2+ -0.29 content, with nine of the major rivers in the area contributing more than -0.22 one-third of the dissolved Ca2+ discharge of the 63 major global rivers to the 0.8 -0.15 ocean (2.56 Tmol Ca a–1 divided by 7.64 Tmol Ca a–1). The Yangtze River -0.08 2+ -0.01 in China represents the largest worldwide annual riverine Ca fl ux (0.903 0.6 0.06 Tmol a–1; all data from Gaillardet et al., 1999). This could possibly be 0.13 explained by the Cenozoic uplift of the Tibetan Plateau and the associated 0.20 Asian monsoon development in this area (Raymo and Ruddiman, 1992). 0.4 0.27 Rivers in the Asian monsoon region also have distinct δ44Ca signa- tures. Although a comprehensive Ca isotope study of all Earth’s major Annual flux (Tmol/a) 44 0.2 rivers has not yet been undertaken and riverine δ Ca may vary in the geological past, existing data suggest that rivers in East and South Asia such as the Yangtze, Ganges, and Huanghe are generally enriched in 44Ca, with 0.0 the most positive δ44Ca occurring in the Yangtze River (δ44Ca = 0.27‰; Schmitt et al., 2003). On the other hand, rivers from the rest of the world, Amazon Yangtze Kolyma Huanghe Lena such as the Amazon, which has the second largest riverine Ca fl ux, seem Orinoco Ganges Indus Yana to be 44Ca-depleted (Fig. 2). Thus, if the Asian monsoon precipitation Figure 2. Dissolved Ca fl ux (Gaillardet et al., 1999) and Ca isotopes increases, river discharge in East Asia increases, and presumably the (Schmitt et al., 2003, and references therein) at nine river mouths. mean δ44Ca of global riverine Ca increases. If changes in the riverine δ44Ca Nine-tenths (0.9) has been added to the data reported by Schmitt lead δ44Ca variations in seawater (Sime et al., 2007), this then provides a et al. (2003) to make it in accordance with the data reported on the possible explanation to the general correlation between Hm/Gt in ODP “bulk Earth” scale used by Fantle and Depaolo (2005, 2007). Note δ44 that rivers in Asian monsoon regions such as the Yangtze, Ganges, 1143 and Ca in DSDP 590 as we observed. In other words, the Asian 44 44 and Huanghe are generally enriched in 44Ca compared to the rest monsoon infl uences marine δ Ca via the Asian fl uvial discharge of Ca- such as the Amazon River. enriched Ca into the ocean. Elevated riverine Ca discharge in Asian monsoon regions is also a likely indicator of enhanced chemical weathering, which is the largest sink 0.0 100 of atmospheric CO over geological time scales. In addition, increased Onset of Northern 2 Hemisphere glaciation detrital matter load and input to the ocean induced by monsoon intensifi - 0.0 150 cation would facilitate organic carbon burial, which is the second largest

2 CO2 sink (Berner and Berner, 1997). As discussed above, we infer that the 200 CO ppmv p Asian monsoon and the corresponding chemical weathering and terrig- 0.2 enous material production/transportation in monsoon regions would have -0.5 250 intensiﬁ ed during the mid-Pliocene, all of which will lead to a pCO2 drop. Hm/Gt Actually, a pronounced decline of pCO2 has been observed from ca. 4 to

ODP Site 1143 300 3 Ma ago, with pCO2 decreasing from ~250 ppm to 180 ppm (Demicco

0.4 et al., 2003; Pearson and Palmer, 2000) (Fig. 3). This 70 ppm CO2 drop -1.0 recently has been invoked to explain the late Pliocene NHG initiation in Late-Pliocene-Pleistocene Mid-Pliocene Early-Pliocene monsoon monsoon Ca (‰) monsoon decline 44 Greenland (Lunt et al., 2008a). intensification decline δ DSDP Site 590 DSDP Site 0 1000 2000 3000 4000 5000 Age (ka) CONCLUSIONS In this study we utilized an Asian monsoon precipitation variability δ44 Figure 3. Comparison of Hm/Gt from ODP Site 1143 and Ca from proxy, the Hm/Gt ratio, from ODP 1143, southern South China Sea, to DSDP Site 590 (Fantle and DePaolo, 2005) for the past 5 Ma. Note the time resolution difference between the Hm/Gt record (2.4 ka per sample) reveal low-latitude climate dynamics since the Pliocene. By comparison and the Ca isotope record (250 ka per sample). Dashed lines separate with other monsoon records, we confi rmed that the mid-Pliocene (4.2–2.7 different phases of the evolution of the Asian monsoon: (1) 5–4.2 Ma, Ma ago) was characterized by a signifi cant monsoon intensifi cation. This early Pliocene monsoon decline; (2) 4.2–2.7 Ma, mid-Pliocene monsoon intensifying monsoon probably left fi ngerprints in marine Ca isotopes. And intensifi cation; (3) 2.7 Ma–present, late Pliocene–Pleistocene monsoon monsoon-induced chemical weathering, continental erosion, and terrige- decline. Available pCO2 before and after the onset of NHG is also indicated (Demicco et al., 2003; Pearson and Palmer, 2000). DSDP—Deep nous inputs likely decreased atmospheric CO2. Subsequently, this CO2 drop Sea Drilling Project; ppmv—parts per million in volume. could have triggered the onset of NHG as suggested by Lunt et al. (2008a).

GEOLOGY, July 2009 601 ACKNOWLEDGMENTS Maslin, M.A., Li, X.S., Loutre, M.F., and Berger, A., 1998, The contribution of We thank Fengmei Li for the help with sample preparation and DRS mea- orbital forcing to the progressive intensifi cation of Northern Hemisphere surement, Gaojun Li for stimulating discussion, Mian Liu for proofreading, and Li glaciation: Quaternary Science Reviews, v. 17, p. 411–426, doi: 10.1016/ Li for sampling. William Ruddiman and Matthew Fantle offered constructive and S0277-3791(97)00047-4. thoughtful review. Samples used in this study were provided by the Ocean Drilling Molnar, P., 2008, Closing of the Central American Seaway and the Ice Age: A criti- Program. This work was supported by the National Natural Science Foundation of cal review: Paleoceanography, v. 23, PA2201, doi: 10.1029/2007PA001574. China through grants 40573054, 40773056, and 40625012. Pearson, P.N., and Palmer, M.R., 2000, Atmospheric carbon dioxide concen- trations over the past 60 million years: Nature, v. 406, p. 695–699, doi: REFERENCES CITED 10.1038/35021000. An, Z.S., Kutzbach, J.E., Prell, W.L., and Porter, S.C., 2001, Evolution of Asian Raymo, M.E., 1994, The initiation of Northern Hemisphere glaciation: Annual monsoons and phased uplift of the Himalaya-Tibetan plateau since Late Review of Earth and Planetary Sciences, v. 22, p. 353–383, doi: 10.1146/ Miocene times: Nature, v. 411, p. 62–66, doi: 10.1038/35075035. annurev.ea.22.050194.002033. Bartoli, G., Sarnthein, M., Weinelt, M., Erlenkeuser, H., Garbe-Schonberg, D., Raymo, M.E., and Ruddiman, W.F., 1992, Tectonic forcing of late Cenozoic cli- and Lea, D.W., 2005, Final closure of Panama and the onset of northern mate: Nature, v. 359, p. 117–122, doi: 10.1038/359117a0. hemisphere glaciation: Earth and Planetary Science Letters, v. 237, p. 33– Rea, D.K., 1992, Delivery of Himalayan sediment to the northern Indian Ocean 44, doi: 10.1016/j.epsl.2005.06.020. and its relation to global climate, sea level, uplift, and seawater strontium, Berner, R.A., and Berner, E.K., 1997, Silicate weathering and climate, in Rud- in Duncan, R.A., et al., eds., Synthesis of results from scientifi c drilling in diman, W.F., ed., Tectonic uplift and climate change: New York, Plenum the Indian Ocean: American Geophysical Union Geophysical Monograph Press, p. 353–365. 70, p. 387–402. Burchfi el, B.C., 2004, New technology, new geological challenges: GSA Today, Royden, L.H., Burchfi el, B.C., and van der Hilst, R.D., 2008, The geological evo- v. 14, no. 2, p. 4–10, doi: 10.1130/1052-5173(2004)014<0004:NTNGC> lution of the Tibetan Plateau: Science, v. 321, p. 1054–1058, doi: 10.1126/ 2.0.CO:2. science.1155371. De La Rocha, C.L., and Depaolo, D.J., 2000, Isotopic evidence for variations in Schmitt, A.D., Chabaux, F., and Stille, P., 2003, The calcium riverine and hy- the marine calcium cycle over the Cenozoic: Science, v. 289, p. 1176–1178, drothermal isotopic fl uxes and the oceanic calcium mass balance: Earth doi: 10.1126/science.289.5482.1176. and Planetary Science Letters, v. 213, p. 503–518, doi: 10.1016/S0012- 821X(03)00341-8. Demicco, R.V., Lowenstein, T.K., and Hardie, L.A., 2003, Atmospheric pCO2 since 60 Ma from records of seawater pH, calcium, and primary carbonate Schwertmann, U., 1988, Occurrence and formation of iron oxides in various pe- mineralogy: Geology, v. 31, p. 793–796, doi: 10.1130/G19727.1. doenvironments, in Stucki, J.W., ed., Iron in soils and clay minerals: Nor- Fantle, M.S., and DePaolo, D.J., 2005, Variations in the marine Ca cycle over the well, D. Reidel, p. 267–308. past 20 million years: Earth and Planetary Science Letters, v. 237, p. 102– Schwertmann, U., and Murad, E., 1983, Effect of pH on the formation of goethite 117, doi: 10.1016/j.epsl.2005.06.024. and hematite from ferrihydrite: Clays and Clay Minerals, v. 31, p. 277–284, Fantle, M.S., and DePaolo, D.J., 2007, Ca isotopes in carbonate sediment and doi: 10.1346/CCMN.1983.0310405. pore fl uid from ODP Site 807A: The Ca2+(aq)-calcite equilibrium fraction- Shackleton, N.J., Backman, J., Zimmerman, H., Kent, D.V., Hall, M.A., Rob- ation factor and calcite recrystallization rates in Pleistocene sediments: erts, D.G., Schnitker, D., Baldauf, J.G., Desprairies, A., Homrighausen, R., Geochimica et Cosmochimica Acta, v. 71, p. 2524–2546, doi: 10.1016/j. Huddlestun, P., Keene, J.B., Kaltenback, A.J., Krumsiek, K.A.O., Morton, gca.2007.03.006. A.C., Murray, J.W., and Westbergsmith, J., 1984, Oxygen isotope calibra- Gaillardet, J., Dupre, B., Louvat, P., and Allegre, C.J., 1999, Global silicate weath- tion of the onset of ice-rafting and history of glaciation in the North Atlantic region: Nature, v. 307, p. 620–623, doi: 10.1038/307620a0. ering and CO2 consumption rates deduced from the chemistry of large rivers: Chemical Geology, v. 159, p. 3–30, doi: 10.1016/S0009-2541(99)00031-5. Sime, N.G., De La Rocha, C.L., Tipper, E.T., Tripati, A., Galy, A., and Bickle, Haug, G.H., and Tiedemann, R., 1998, Effect of the formation of the Isthmus of M.J., 2007, Interpreting the Ca isotope record of marine biogenic car- Panama on Atlantic Ocean thermohaline circulation: Nature, v. 393, p. 673– bonates: Geochimica et Cosmochimica Acta, v. 71, p. 3979–3989, doi: 676, doi: 10.1038/31447. 10.1016/j.gca.2007.06.009. Haug, G.H., Ganopolski, A., Sigman, D.M., Rosell-Mele, A., Swann, G.E.A., Sun, Y., Lu, H.Y., and An, Z.S., 2006, Grain size of loess, palaeosol and Red Tiedemann, R., Jaccard, S.L., Bollmann, J., Maslin, M.A., Leng, M.J., and Clay deposits on the Chinese Loess Plateau: Signifi cance for understand- Eglinton, G., 2005, North Pacifi c seasonality and the glaciation of North ing pedogenic alteration and palaeomonsoon evolution: Palaeogeogra- America 2.7 million years ago: Nature, v. 433, p. 821–825, doi: 10.1038/ phy, Palaeoclimatology, Palaeoecology, v. 241, p. 129–138, doi: 10.1016/ nature03332. j.palaeo.2006.06.018. Heuser, A., Eisenhauer, A., Bohm, F., Wallmann, K., Gussone, N., Pearson, P.N., Tian, J., Wang, P.X., Cheng, X.R., and Li, Q.Y., 2002, Astronomically tuned Plio- Nagler, T.F., and Dullo, W.C., 2005, Calcium isotope (δ44/40Ca) variations of Pleistocene benthic δ18O record from South China Sea and Atlantic-Pacifi c Neogene planktonic foraminifera: Paleoceanography, v. 20, PA2013, doi: comparison: Earth and Planetary Science Letters, v. 203, p. 1015–1029, doi: 10.1029/2004PA001048. 10.1016/S0012-821X(02)00923-8. Ji, J., Chen, J., Balsam, W., Lu, H., Sun, Y., and Xu, H., 2004, High resolution Tian, J., Wang, P.X., and Cheng, X.R., 2004, Development of the East Asian mon- hematite/goethite records from Chinese loess sequences for the last gla- soon and Northern Hemisphere glaciation: Oxygen isotope records from the cial-interglacial cycle: Rapid climatic response of the East Asian Monsoon South China Sea: Quaternary Science Reviews, v. 23, p. 2007–2016, doi: to the tropical Pacifi c: Geophysical Research Letters, v. 31, L03207, doi: 10.1016/j.quascirev.2004.02.013. 10.1029/2003GL018975. Wang, P.X., Tian, J., Cheng, X.R., Liu, C.L., and Xu, J., 2003, Carbon reservoir Kampf, N., and Schwertmann, U., 1983, Goethite and hematite in a climose- changes preceded major ice-sheet expansion at the mid-Brunhes event: Ge- quence in southern Brazil and their application in classifi cation of kaolinitic ology, v. 31, p. 239–242, doi: 10.1130/0091-7613(2003)031<0239:CRCP soils: Geoderma, v. 29, p. 27–39, doi: 10.1016/0016-7061(83)90028-9. MI>2.0.CO;2. Keigwin, L., 1982, Isotopic paleoceanography of the Caribbean and East Pa- Wang, P.X., Tian, J., Cheng, X.R., Liu, C.L., and Xu, J., 2004, Major Pleistocene cifi c—Role of Panama uplift in late Neogene time: Science, v. 217, p. 350– stages in a carbon perspective: The South China Sea record and its global 352, doi: 10.1126/science.217.4557.350. comparison: Paleoceanography, v. 19, PA4005, doi: 10.1029/2003PA000991. Klocker, A., Prange, M., and Schulz, M., 2005, Testing the infl uence of the Cen- Zhang, Y.G., Ji, J., Balsam, W.L., Liu, L., and Chen, J., 2007, High resolution tral American Seaway on orbitally forced Northern Hemisphere glaciation: hematite and goethite records from ODP 1143, South China Sea: Co-evo- Geophysical Research Letters, v. 32, L03703, doi: 10.1029/2004GL021564. lution of monsoonal precipitation and El Niño over the past 600,000 years: Liu, Z.F., Colin, C., Trentesaux, A., Siana, G., Frank, N., Blamart, D., and Farid, Earth and Planetary Science Letters, v. 264, p. 136–150, doi: 10.1016/j. S., 2005, Late Quaternary climatic control on erosion and weathering in the epsl.2007.09.022. eastern Tibetan Plateau and the Mekong Basin: Quaternary Research, v. 63, Zhou, L.P., Oldfi eld, F., Wintle, A.G., Robinson, S.G., and Wang, J.T., 1990, p. 316–328, doi: 10.1016/j.yqres.2005.02.005. Partly pedogenic origin of magnetic variations in Chinese loess: Nature, Lunt, D.J., Foster, G.L., Haywood, A., and Stone, E.J., 2008a, Late Pliocene v. 346, p. 737–739, doi: 10.1038/346737a0.

Greenland glaciation controlled by a decline in atmospheric CO2 levels: Nature, v. 454, p. 1102–1105, doi: 10.1038/nature07223. Manuscript received 23 November 2008 Lunt, D.J., Valdes, P.J., Haywood, A., and Rutt, I.C., 2008b, Closure of the Pan- Revised manuscript received 13 February 2009 ama Seaway during the Pliocene: Implications for climate and Northern Manuscript accepted 16 February 2009 Hemisphere glaciation: Climate Dynamics, v. 30, p. 1–18, doi: 10.1007/ s00382-007-0265-6. Printed in USA

602 GEOLOGY, July 2009