Drake Passage and Cenozoic climate: An open ... - Wiley Online Library

27 downloads 0 Views 1MB Size Report
Jan 17, 2007 - Alfred Wegener Institute for Polar and Marine Research, Postfach 120161, ... [1] Drake Passage opening has often been viewed as a single, ...
Geochemistry Geophysics Geosystems

3

Article

G

Volume 8, Number 1 17 January 2007 Q01005, doi:10.1029/2005GC001224

AN ELECTRONIC JOURNAL OF THE EARTH SCIENCES Published by AGU and the Geochemical Society

ISSN: 1525-2027

Click Here for

Full Article

Drake Passage and Cenozoic climate: An open and shut case? Roy Livermore British Antarctic Survey, High Cross, Madingley Road, Cambridge CB3 0ET, UK Now at Department of Earth Sciences, University of Cambridge, Downing Street, Cambridge CB2 3EQ, UK. ([email protected])

Claus-Dieter Hillenbrand and Mike Meredith British Antarctic Survey, High Cross, Madingley Road, Cambridge CB3 0ET, UK

Graeme Eagles Alfred Wegener Institute for Polar and Marine Research, Postfach 120161, D-27515, Bremerhaven, Germany

[1] Drake Passage opening has often been viewed as a single, discrete event, possibly associated with abrupt changes in global circulation and climate at or near the Eocene-Oligocene boundary. A new plate tectonic model, based on recent reinterpretations of the opening history of basins in the Scotia Sea, suggests that an effective ocean gateway may have developed even earlier, during the middle Eocene. This is consistent with a growing body of evidence from sediment core proxy data for Eocene changes in Southern Ocean circulation and biological productivity. The period between earliest opening after 50 Ma and the latest Eocene was characterized by the evolution of various current pathways across the subsiding continental shelves and intervening deep basins. This shallow opening may have caused important changes in Southern Ocean circulation, contributing to Eocene cooling and the growth of Antarctic ice sheets. Components: 7146 words, 3 figures. Keywords: gateways; Southern Ocean; Antarctic Circumpolar Current. Index Terms: 3045 Marine Geology and Geophysics: Seafloor morphology, geology, and geophysics; 4901 Paleoceanography: Abrupt/rapid climate change (1605). Received 20 December 2005; Revised 22 June 2006; Accepted 18 July 2006; Published 17 January 2007. Livermore, R., C.-D. Hillenbrand, M. Meredith, and G. Eagles (2007), Drake Passage and Cenozoic climate: An open and shut case?, Geochem. Geophys. Geosyst., 8, Q01005, doi:10.1029/2005GC001224.

1. Introduction [2] There has been much debate about the timing of Drake Passage opening to deep currents like the present Antarctic Circumpolar Current (ACC), and the possibility that this could have been a primary cause of changes in global circulation and climate, and the rapid expansion of Antarctic ice sheets [e.g., Kennett, 1977; Lawver and Gahagan, 1998; Copyright 2007 by the American Geophysical Union

DeConto and Pollard, 2003a; Barker and Thomas, 2004; Sijp and England, 2004; Kennett and Exon, 2004]. Some workers have suggested that a deepwater connection between the South Pacific and South Atlantic was established by the earliest Oligocene, on the basis of estimates of the positions of the South American and Antarctic plates [Lawver and Gahagan, 1998, 2003], while others have argued that an effective pathway was estab-

1 of 11

Geochemistry Geophysics Geosystems

3 livermore et al.: drake passage and cenozoic climate 10.1029/2005GC001224

G

lished only in the Miocene, after various blockages had cleared [Barker and Burrell, 1977; Barker, 2001]. [3] Recent reinterpretations of basin opening in the western, central and southern Scotia Sea [Eagles et al., 2005, 2006; R. Livermore et al., The role of the central Scotia Sea in Drake Passage gateway opening, submitted to Marine Geology, 2006 (hereinafter referred to as Livermore et al., submitted manuscript, 2006)], although controversial in some cases, provide the basis for a comprehensive and self-consistent model for the movement of plates and plate fragments in the Drake Passage region, commencing in the middle Eocene. Placing this model in the context of new estimates for the motion of the South American and Antarctic plates [Livermore et al., 2005], we present a series of reconstructions, which we have used to estimate when and where pathways for shallow and deepwater flow may have developed during the early evolution of the Scotia Sea.

2. Gateway Opening Models [4] Drake Passage opening remains controversial. Marine geophysical data have been used to suggest that opening occurred prior to chron C8 (26 Ma), and, probably, chron C10 (29 Ma), on the basis of identification of magnetic anomalies adjacent to the South American margin and in the southwest Scotia Sea [Barker and Burrell, 1977; Lodolo et al., 1997; Livermore et al., 2005]. However, precise dating of the earliest formation of a deep-water pathway is complicated by the existence of plate fragments, which may have blocked Drake Passage even after seafloor spreading began. For example, it has been suggested that the effect of Drake Passage opening on circulation and climate was ameliorated initially by the existence of narrow, overlapping slivers of continental crust attached to the retreating South American and Antarctic continents, forming a barrier which only cleared during the early Miocene (22 Ma [Barker and Burrell, 1977; Barker, 2001]). [5] Lawver and coworkers [Lawver et al., 1992; Lawver and Gahagan, 1998, 2003] showed plate tectonic reconstructions of the Southern Ocean and circum-Antarctic continents, which they used to deduce the age of the earliest circumpolar deepwater pathway. The South America – Antarctica rotations upon which these reconstructions were based were derived by addition of published rotations around a plate circuit through Africa, and

show Drake Passage open in all maps from 50 Ma onward. Despite this, Lawver and Gahagan argued for a date of 31 ± 2 Ma for the first deep-water connection. These authors did not consider uncertainties in their reconstructions, which arise from compounding errors in the South America–Africa and Africa-Antarctica rotations they used [Kirkwood et al., 1999], nor from possible non-closure of rotations about the Bouvet triple junction, nor Cenozoic motion between the Antarctic Peninsula and East Antarctica [Cande et al., 2000]. Moreover, the motions of the major plates have been decoupled from those of the plates within Drake Passage for much of the Cenozoic by additional plate boundaries corresponding to the present North Scotia Ridge and South Scotia Ridge [Eagles et al., 2005]. As a result, reconstructions using the present outlines of South America and Antarctica overestimate the amount of separation between Tierra del Fuego and the Antarctic Peninsula by over 100 km. In addition, considerable deformation has taken place in southernmost South America during the Late Cretaceous and Cenozoic, including the formation of the Patagonian orocline [Cunningham et al., 1991], movement on numerous strike-slip faults and thrusts [Kraemer, 2003], and crustal extension, so that continental outlines during the critical Eocene-Oligocene interval would have been different to those of today. Likewise, the conjugate margin at the South Scotia Ridge has been deformed by more recent strike-slip tectonics, and is unrecognizable as a passive margin [Lodolo et al., 1997]. For these reasons, major plate reconstructions alone cannot be used to determine precisely when Drake Passage opening occurred. [6] We have therefore made detailed reconstructions based on a combination of major and minor plate motions, derived directly from marine geophysical studies in the Scotia Sea, South Atlantic and Weddell Sea. The boundary conditions for our reconstructions are the Cenozoic motions of the South American and Antarctic plates, which enclose the Drake Passage and Scotia Sea. The latter have been rigorously evaluated by Livermore et al. [2005], using joint inversion of magnetic isochrons and gravity flow lines from the Weddell Sea, and show an increase in separation rates in the proto-Drake Passage region at 50 Ma, from just a few mm/yr in a N-S direction, to 22–24 mm/yr in a WNW-ESE direction. The initial response to this change was most likely a period of crustal extension and subsidence during the middle Eocene, and 2 of 11

Geochemistry Geophysics Geosystems

3 livermore et al.: drake passage and cenozoic climate 10.1029/2005GC001224

G

Figure 1. Drake Passage and the Scotia Sea. Bathymetric contours [from Smith and Sandwell, 1997] are shown at 1000 (filled light gray), 2000, and 3000 m (filled dark gray). Approximate positions of oceanographic fronts shown in blue. The position of the Polar Front is based on SST data and is taken from Moore et al. [1999]; other fronts are from Orsi et al. [1995], as follows: SAF, Subantarctic Front; SACCF, Southern ACC Front. Labeled fragments and basins are discussed in text.

the formation of a shallow ocean connection between the Pacific and Atlantic oceans. [7] Although strongly influenced by these motions, seafloor spreading in Drake Passage was, as mentioned above, decoupled by the existence of additional plate boundaries, such that the rate of opening exceeded the rate of South America– Antarctica separation, with the difference being taken up by deformation in Tierra del Fuego [Cunningham et al., 1995; Kraemer, 2003; Eagles et al., 2005]. Opening of a deep ocean gateway between Tierra del Fuego and the Antarctic Peninsula was dated using magnetic anomalies as >26 Ma from the identification of anomaly C8 [Barker and Burrell, 1977]. However, this anomaly generally lies some distance from the margins of Drake Passage, and so the actual age of the earliest deep seafloor is uncertain [Eagles et al., 2005]. Using a simple extrapolation, Livermore et al. [2005] estimated this age as 32 ± 2 Ma. Even so, the time at which uninterrupted flow of significant deep currents like the ACC could have commenced depended on the position of various crustal fragments and intervening basins as the gateway developed. [8] These include South Georgia (Figure 1), which is thought to have migrated from an original

position adjacent to Tierra del Fuego [Mukasa and Dalziel, 1996], and the central Scotia Sea basin, the opening of which was probably responsible for the northward migration of South Georgia. Central Scotia Sea evolution has recently been reevaluated by R. Livermore et al. (submitted manuscript, 2006), who suggest a model based on an earlier interpretation by Hill and Barker [1980], but with a slightly later date of extinction. Other fragments believed to have been involved in gateway evolution are the Terror Rise, Pirie Bank, Bruce Bank, Discovery Bank, and South Orkney (Figure 1). Eagles et al. [2006] reviewed evidence for the opening of small basins in the southern Scotia Sea, which resulted in the rifting and isolation of the banks mentioned above, and advanced a model involving basin opening since the middle Eocene, somewhat earlier than previous suggestions [e.g., Barker and Hill, 1982]. This model is employed here, with the caveat that, in the absence of conclusive evidence, the ages of these basins must continue to be regarded as uncertain, allowing other interpretations [e.g., Galindo-Zaldivar et al., 2006]. Opening of Powell Basin and the eastward translation of South Orkney was dated by Eagles and Livermore [2002] at 40 Ma to 21.8 Ma, consistent with basement depths, although the poor quality of magnetic anomalies recorded in Powell 3 of 11

Geochemistry Geophysics Geosystems

3 livermore et al.: drake passage and cenozoic climate 10.1029/2005GC001224

G

Basin once more allows alternative interpretations [e.g., Coren et al., 1997].

3. Reconstructions [9] At 52 Ma, the time of the so-called Early Eocene Climatic Optimum, fragments were grouped into a tight cluster, forming a land bridge between South America and Antarctica (Figure 2a), allowing the dispersal into the northern Antarctic Peninsula of several groups of animals and plants [Reguero et al., 2002]. Circulation patterns in the South Pacific and South Atlantic were therefore independent, with wind-driven cyclonic gyres bringing warmer, sub-Antarctic waters southward along the Pacific margin, and cool waters northward from the Weddell Sea to influence the Atlantic margin. These circulation patterns are consistent with the results of oceanographic modeling studies conducted with a closed Drake Passage, such as performed by Sijp and England [2004], and we represent them schematically in Figure 2a. Although the details and intensity of these circulations are somewhat matters for speculation, they do appear to be the most plausible patterns for the circulation at this time, and agree with dynamical characteristics such as the topographic steering of flows. [10] We suggest that the land bridge became submerged some time after 50 Ma, as the continental crust subsided in response to an abrupt increase in the rate of separation of South America and Antarctica [Livermore et al., 2005], and the initiation of subduction of Weddell Sea crust (Figure 2b). This represented the first opening of the gateway, but was limited to water depths 2000 m)

gap between the subsiding sill and the separating blocks, which still restricted current flow to depths