Introduction to special section on Cenozoic Paleoceanography of the Central Arctic Ocean
2008; American Geophysical Union; Volume: 23; Issue: 1 Linguagem: Inglês
10.1029/2007pa001516
ISSN1944-9186
Autores Tópico(s)Geology and Paleoclimatology Research
ResumoIn late summer 2004, the Integrated Ocean Drilling Program (IODP) conducted one of the most transformational missions in the almost 40 year history of scientific ocean drilling: the Arctic Coring Expedition (ACEX) [Backman et al., 2006; Moran et al., 2006a]. This technically challenging expedition [Moran et al., 2006b] to the Lomonosov Ridge near 88°N recovered the first long-term Cenozoic sediment record from the Arctic Ocean–extending previous records from ∼1.5 Ma to an unprecedented ∼56 Ma. Glimpses of the breadth of this transformation were seen during the ACEX cruise when the massulae from freshwater ferns (Azolla) were found in 49 Ma old sediments [Brinkhuis et al., 2006] and the presence of Apectodinium augustum confirmed that the Paleocene-Eocene thermal maximum (PETM) (55 Ma) was unexpectedly recovered [Sluijs et al., 2006]. Soon after the expedition, when the cores were opened and analyzed, ice-rafted debris was found in surprisingly old (46 Ma) sediments [Moran et al., 2006a]. These middle Eocene sediments are characterized also by their high organic carbon content [Stein et al., 2006]. The exciting early results attracted other investigators that expanded the scientific investigating team to more than 40 people. Analyses were extended to include studies that revealed surprisingly high Arctic Ocean surface water temperatures [Sluijs et al., 2006] and an amplified hydrological system [Pagani et al., 2006] during the PETM. Initial analyses revealed an extensive hiatus encompassing about 26 million years that occurred below a short interval showing starkly alternating black and white layers that is now dubbed the "zebra" interval; thus the time interval from the late early Miocene (18 Ma) to the middle middle Eocene (44 Ma) is missing. Although the hiatus is a lost window in time for the Arctic paleoclimate record, it spawned other studies that integrated the regional tectonic history with ACEX results revealing a major oceanographic reorganization at 17.5 Ma–ventilation of the Arctic Ocean to the North Atlantic through the Fram Strait [Jakobsson et al., 2007]. In terms of ventilation and water mass history, Neogene neodymium isotope data reveal that Arctic Intermediate Waters originate from brine formation on Eurasian shelf areas between 15 and 2 Ma, and during glacial periods throughout Pleistocene times [Haley et al., 2008a]. They conclude that this mode of Arctic circulation differs markedly from today's situation, showing relatively high contributions of North Atlantic Intermediate Water and minor input from brine formation. In this special section, the transformation continues. Results from the large ACEX scientific "family" include: a robust age model; detailed analyses of the middle Eocene that document a unique brackish water environment; sea ice and iceberg history reconstructions as well as the provenance of ice-rafted debris from the Eocene to present; evolution of depositional environments; times of isolation and connection to the global ocean; geochemical analyses of the organic carbon-rich sediments; and unique applications of high-resolution proxies and cyclostratigraphy. Highlights from this special section are presented here. St. John [2008] presents a detailed study of terrigenous sand content, revealing an ice-rafted record that extends back to the middle Eocene. These landmark results confirm that seasonal sea ice existed at circa 46 Ma in the central Arctic Ocean. The timing of this cooling slightly precedes the earliest signs of glaciation reported for Greenland (38 Ma) [Eldrett et al., 2007] or Antarctica (40 Ma) [Lear et al., 2000]. St. John concludes, however, that the onset of ice-rafted sedimentation in the central Arctic compares in timing with a major decrease in atmospheric concentrations of CO2 and the onset of the early mid-Eocene cooling [Pearson and Palmer, 2000; Lowenstein and Demicco, 2006]. St. John also summarizes the flux and provenance of the ice-rafted debris, indicating an average Eocene flux of 0.13 g cm−2 ka−1, a rate that was tripled in the Neogene sediments, and points to a fairly consistent source area (Russian margin). Krylov et al. [2008] produce a detailed analysis of IRD provenance from the middle Eocene and throughout the Neogene. Among the minerals studied, they identify and use two (clinopyroxene and hornblende) minerals as key indicators of a major change in the source of ice-rafted debris that occurred between 14 and 13 Ma. Since the source areas for hornblende are distal (eastern Arctic Ocean from the Lena River to eastern Siberia) from the ACEX drill site, the authors argue that the increased occurrence of this mineral at 13 Ma indicates the initiation of a permanent ice cover in the Arctic Ocean. The timing of this critical Arctic sea ice event is notable because it is concomitant with the timing of the initiation of a permanent East Antarctica ice sheet [Zachos et al., 2001]. Haley et al. [2008b] report their results from pore fluid and bulk sediment isotopic analyses and use them to reconstruct a 15 Ma record of sea ice and ocean circulation patterns. They interpret the source of detrital ACEX sediment to be the Eurasian margin. This interpretation is consistent with St. John and Krylov et al. who used other independent techniques. Haley et al. conclude, further, that the Transpolar Drift has remained the dominant transport vehicle since 15 Ma. The authors use Sr isotope data to investigate the exchange history between the Arctic Ocean and the North Atlantic. They agree with Jakobsson et al. [2007], that Arctic seawater evolved from an initial condition of unidirectional surface outflow from the Arctic to the North Atlantic prior to 17 Ma. Haley et al. argue for a limited and pulsed exchange with the North Atlantic until about 13 Ma, when Arctic Sr isotopes became identical with the global ocean Sr isotope curve. The other new provenance study, by Darby [2008], spans a slightly shorter time period than that by Krylov et al. and presents results using a different method, the Fe grain "fingerprinting" approach pioneered by Bischof and Darby [1997]. The results are consistent with Krylov et al. in one aspect but in stark contrast in another aspect. Both studies interpret the Arctic Ocean to have been perennially covered by sea ice since 13–14 Ma. The contrast lies in the interpreted source area of the ice and hence, the ice-rafted debris. Darby infers a predominantly Canadian Archipelago source for the IRD over the past 14 Ma at the ACEX site. This implies that sources from the Eurasian margin were shut down and that sea ice from the Canadian Archipelago was the primary source, a radically different interpretation of circulation when compared with Haley et al., Krylov et al., and St. John. Spofforth et al. [2008] set the stage for two other middle Eocene studies of Pälike et al. [2008] and Sangiorgi et al. [2008a] by presenting their results and interpretation of high-resolution X-ray fluorescence (XRF). Spofforth et al. (submitted manuscript, 2007) use these first detailed Eocene major element concentration data from the Arctic Ocean to describe an environment dominated by euxinic conditions. They further detail that over this time interval, the euxinic environment changed from one with limited terrestrial input to increasing terrigenous deposition, probably as a response to expanding sea and/or glacial ice. Pälike et al. apply cyclostratigraphic techniques to extract sedimentation rates and to interpret the relative importance of obliquity and precession to insolation of the central Arctic Ocean. To accomplish this, they took "snapshots" from the discontinuously recovered ACEX record and applied a frequency ratio technique to capture the dominant astronomically driven frequency. Pälike et al. borrowed methods from meteorology to better utilize the multiple high-resolution proxy data sets available (XRF, multisensor core logger, color reflectance), to identify the most robust cycles among all proxies. Sangiorgi et al. [2008a] focus on a middle Eocene interval deposited circa 46 Ma by detailing the cyclostratigraphy of one core. Their analyses of the physical, biological and sedimentological data show that obliquity and precessional orbital forcing dominate. They further interpret each of these proxy types and conclude that the biological components of the climate system are primarily responding to changes in the growing season, while the physical/sedimentological data that represent input from the continents are driven by insolation and the production of sea ice and/or glacial ice. An unexpected approximately 26 million years long hiatus, from the middle Eocene to the early Miocene, occurs in the ACEX record. Sangiorgi et al. [2008b] address possible cause(s) of this hiatus. Their multiproxy approach, using palynology, sediment geochemistry, and TEX86 as well as BIT data for paleotemperature estimates, suggests a shoaling depositional environment, located close to or at sea level, shortly prior to the hiatus. Furthermore, they suggest that sediments deposited shortly after the hiatus also represent a shallow marine environment, influenced by both fresh and marine waters with alternating oxic and anoxic conditions. They conclude that the major hiatus was caused by erosion and/or nondeposition resulting from the progressive shoaling of the ridge, and that the drill site remained in a such a shallow water setting throughout the course of this long mid-Cenozoic hiatus. But why was there a shallow water environment on the ridge crest around the hiatus, more than 12 million years after that the ridge was rifted from the Eurasian margin? This contrasts sharply to predictions made by standard tectonic subsidence models. O'Regan et al. [2008a] address this problem. Their review of the geodynamic evolution of the central Arctic Ocean, together with a reanalysis of published seismic reflection and gravity data, suggests that the subsidence began at the end of a prolonged phase (early Eocene through early Miocene) of basin wide compression. Sluijs et al. [2008] have expanded their initial palynological and geochemical data sets across the PETM [Sluijs et al., 2006] to cover the entire early Eocene. New XRF scanning data are presented that shows how cyclicity in Fe is related to precession. The Eocene thermal maximum 2 (ETM2) (53 Ma), also dubbed the Elmo event, is confidently identified. TEX86 data indicate temperatures of about 15°–18°C throughout the early Eocene, which is characterized also by brackish and productive surface waters interpreted from palynological data. Surface water salinities are addressed by Waddell and Moore [2008], using oxygen isotopes from fish bone carbonate. Their approach yields an average salinity of about 21–25‰ throughout the preserved Eocene section, with three negative excursions to lower salinity values (16–18‰) at the PETM (55 Ma), the freshwater Azolla event (49 Ma), and at 47.6 Ma. These reduced surface water salinities are consistent with the virtual absence of radiolarians in the biosiliceous rich interval. The Azolla event is characterized by a distinct positive δ13C excursion in the fish bone carbonate, reflecting a sharp increase in surface water productivity. Waddell and Moore suggest a salinity of 34.1‰ for a sample from the "zebra" interval at circa 18 Ma. The composition of the biosilica groups in the middle Eocene sediment, with largely endemic assemblages of marine diatoms, ebridians, and chrysophyte cysts (resting stage of freshwater algae), and lack of radiolarians, is consistent with brackish surface water conditions, as argued by Stickley et al. [2008]. The variation in group dominance over time, however, points to variable surface water conditions with respect to salinity, water column stratification and productivity. Of particular note is that the laminated middle Eocene Arctic Ocean sediments hold the most diverse, abundant and sustained levels of chrysophyte cysts ever observed in a Paleogene setting. Furthermore, pale laminae are nearly pure biosilica, containing virtually monospecific diatom assemblages, and are considered to reflect the beginning of the growth season when daylight returns to the high Arctic. Dark laminae contain mixed biosiliceous taxa plus organic material and clay, reflecting deposition later in the growth season. The middle Eocene laminated intervals thus may represent seasonal layers. Onodera et al. [2008] address the question of connections between the Arctic Ocean and other oceans. The endemic composition of the silicoflagellate and ebridian assemblages suggests that the Arctic Ocean was by and large isolated from the Pacific and Atlantic during the early part of the middle Eocene (circa 50–45 Ma). This isolation ended at about 45 Ma, indicated by high abundances of the silicoflagelle genus Corbisema, presumably when shallow connection(s) to the Atlantic (Nordic Seas) became established. The present lack of independent age information implies that the rapid evolution, and biostratigraphy, of these endemic early middle Eocene taxa still have limited value in terms of age control. Biogenic carbonate is sporadically preserved in a few horizons over the upper 20 m of the sediment column. Cronin et al. [2008] investigate both benthic and planktic foraminifers, and the transition from calcareous (above) to agglutinated (below) benthic assemblages at about MIS 9 time (circa 300 ka). They speculate that abundance peaks in agglutinated benthic foraminifers may reflect seasonally ice-free conditions during interglacials in the central Arctic Ocean. Thus judging from this data set, a permanent Arctic sea ice cover was perhaps not developed until well into the later half of the Pleistocene. This interpretation differs from the views independently presented, using various lithologic data sets, by St. John, Darby, and Krylov et al., the radiogenic isotope data of Haley et al., and the cosmogenic isotope data of Frank et al. [2008], who all suggest that a continuous sea ice cover has existed since the middle Miocene. Establishing accurate age models using short sediment cores has been a well-documented problem for the Arctic Ocean [e.g., Backman et al., 2004]. The problem arises because of the discontinuous and rare occurrences of calibrated biostratigraphic datums, the problematic record of paleomagnetic polarity patterns, and the general lack of other chronological tools in these sediments. With the advent of ACEX, penetrating much deeper into the sediment column than ever before, it was anticipated that the more deeply buried sediments, beyond the piston core range, would contain some biostratigraphic information that could provide a reliable indication on whether or not the sedimentation rates were slow (millimeter per kiloannum scale) or fast (centimeter per kiloannum scale). An unequivocal answer to this long-standing question emerged during ACEX. At about 100 mbsf, a few dinoflagellate biohorizons were observed, which all indicated a late Miocene age, hence providing an average sedimentation rate of about 1.5 cm ka−1 for the upper 100 m of the sediment column on the ridge crest [Backman et al., 2006; Moran et al., 2006a]. It turned out, however, that only a few biostratigraphic datums, chiefly provided by dinoflagellate cysts, are preserved in the Neogene sediments. In order to improve the chronological resolution across the upper 150 m of the sediment column, Frank et al. examined Be isotope ratios. Apart from providing age control well back into middle Miocene time and revealing a 2.2 million years long hiatus near the middle/late Miocene boundary, this study also observed that the flux of 10Be to the ACEX site is about a factor of two lower than the global average production rate. This is interpreted to reflect the existence of a continuous sea ice cover throughout the entire range of productive 10Be, that is, since 12.3 Ma. An approximately 45 m thick interval, from the deepest productive 10Be sample at 151.3 mcd to shortly above the major mid-Cenozoic hiatus at 198.7 mcd, essentially lacks age control. Uncertainties and possible sedimentation rates in this interval are discussed by Jakobsson et al. [2007]. A stratigraphic and chronologic synthesis of the entire ACEX sequence is presented by Backman et al. [2008]. Here all biostratigraphic and cosmogenic isotope data are merged with cyclostratigraphic data and a single magnetostratigraphic reversal boundary, into a coherent age model that is placed on a clearly defined, amalgamated timescale. O'Regan et al. [2008b] focused in on the ACEX Pleistocene records that help to illuminate the evolution of glacial cycles in the Arctic Ocean. They investigate the upper 27 m continuous sedimentary interval using biostratigraphic datums and magnetic inclination data to show that bulk and mineral magnetic properties covary with Milankovitch frequencies and delineate glacial and interglacial modes of sediment deposition. Articles by Revkin [2004] and Traufetter [2006] were two of many news articles that highlighted the fact that ACEX scientists discovered elevated concentrations (1–14%) of organic carbon in the early Miocene through Eocene sediments near the North Pole. Concentrations of organic carbon, nitrogen and sulfur in the approximately 200 m thick Eocene and early Miocene section were presented by Stein et al. [2006]. Low C/S ratios and presence of pyrite in laminated sediments are suggested to reflect euxinic conditions in the early middle Eocene section, probably caused by widespread salinity stratification in an isolated Arctic Ocean. The euxinic conditions are suggested to become more intense shortly prior to the hiatus as judged from sulfur excess values. A plot of TOC versus pyritic sulfur values from the "zebra" interval shows that its alternating dark and light bands represent anoxic and oxic condition, respectively. The study of Stein et al. [2006] contains the first indication, based on δ13C data, that the "Elmo" event (EMT2) was partially recovered in the ACEX sediments. Weller and Stein [2008] took a fresh look at these organic carbon-rich sediments, this time with a focus on source-related biomarkers. They conclude that the organic material has a terrestrial signature during the late Paleocene and earliest Eocene, whereas a marine influence becomes more important in the middle Eocene biosilica rich sediments. In discussing the source rock potential of the ACEX sediments, Stein [2007, p. 72] remarks on the low level of maturity throughout the study interval, hence not so much potential for "… oil bonanzas beneath Arctic Ocean" [Revkin, 2006] after all in the central Arctic, judging from these ACEX drill data as well as from the sparse reflection seismic data available from the deep Amerasian and Eurasian basins [Grantz et al., 1990; Jokat et al., 1995]. A study of bulk nitrogen and phosphorus contents, and nitrogen isotopes by Knies et al. [2008] reinforces the interpretations about high Paleogene and low Neogene productivity. They present a model showing that, during the Eocene, the waters that bathed the ridge crest (drill site) were depleted in oxygen because of a well developed pycnocline in the uppermost water column. Studies of the ACEX sediments have transformed our understanding of the Cenozoic paleoenvironmental evolution in the central Arctic Ocean. The key questions formulated in the underlying drilling proposal are addressed in these ACEX-related Paleoceanography papers and in six published Nature and Nature Geoscience papers, among others. In addition, a series of papers describing new species and genera of several microfossil groups will be published in Micropaleontology. Among the rich results, it appears pertinent to highlight the new information obtained with respect to the time-dependent distribution of sea ice in the Arctic Ocean, which was a centerpiece in the drilling proposal. These studies reveal that a cover of continuous sea ice has existed since just after the middle Miocene climatic optimum. This is many million years further back in time, compared to previous thought [e.g., Driscoll and Haug, 1998]. ACEX data also demonstrate that seasonal sea ice began to occur in the middle Eocene (46 Ma), preceding the first signs of glaciation on Greenland later in the middle Eocene (38 Ma) [Eldrett et al., 2007]. Glaciation on Antarctica is also considered to have begun in the middle Eocene, around 40 Ma [Lear et al., 2000; Pearson and Palmer, 2000], although the size, exact timing, location, and impact of these early Cenozoic glaciations are still debated [Tripati et al., 2005; Edgar et al., 2007]. The details in this middle Eocene paleoclimate evolution are thus still unclear. Yet it appears reasonable to suggest that the formation of seasonal sea ice in the central Arctic Ocean indeed contributed to the onset of the Cenozoic cooling trend. This is a major change in the way we think about how the planet has changed from a greenhouse to icehouse world because the initial deep freeze of the two poles appears to have occurred in concert. Moreover, this eliminates the problem of having one pole cold (Antarctica) and the other warm(er) (Arctic) in the late middle Eocene through middle Miocene interval [Zachos et al., 2001]. In this context, one must remember that the age control in the middle Eocene ACEX sediments is comparatively uncertain. Perhaps the present differences in timing of the onset of cooling between the Arctic (ACEX) and lower-latitude records will disappear and merge when complete Paleogene paleoceanographic sections are recovered from the Arctic Ocean sometime in the future. It came as a complete surprise to us after over 6 years of planning (between submission of the first preproposal to the Ocean Drilling Program in early 1998 to the execution of ACEX by the IODP in mid-2004) that we encountered a 26 million years long hiatus in what was originally interpreted to be a continuous middle Eocene through Holocene layer cake stratigraphy in the seismic reflection record [Jokat et al., 1992, 1995]. The Arctic Ocean is still holding on to its secrets, from the middle middle Eocene to the late early Miocene. Filling that gap, and other poorly recovered intervals in the ACEX record, is an obvious and worthy target for other, future scientific ocean drilling expedition(s) into the central Arctic Ocean. Any such attempt should plan from lessons learned during ACEX, both in terms of logistics, how to keep station in the Arctic pack ice during drilling operations which is a separate story and not dealt with here, and science strategies. Regarding the latter, a few critical points are worth considering: 1. Locations of future drill sites should continue to focus on ridges/topographic highs because the abyssal plains are filled with 2–3 km thick sections of turbiditic sediments in the Eurasian Basin. The Amerasian Basin abyssal plains show sediment thicknesses on the order of 6–10 km [Jackson and Oakey, 1990], of presumably turbiditic origin and hence less suitable for paleoceanographic reconstruction. 2. Future drilling sites must be placed so as to avoid the long mid-Cenozoic hiatus encountered in the ACEX section, which requires detailed and highly resolved reflection seismic characterization of the sediments. This hiatus is most likely related to the tectonic history of the Lomonosov Ridge [O'Regan et al., 2008a]. 3. ACEX sediments were cored at a present water depth of nearly 1.3 km. The Paleogene through early Miocene sediments, however, appear to have been deposited at considerably shallower depths, probably corresponding to shelf or neritic depths [Backman et al., 2006; Moran et al., 2006a; Sangiorgi et al., 2008b]. The ACEX record of Paleogene through early Neogene sediments therefore may represent only a partial, uppermost water column, view of the paleoceanographic history of the central Arctic Ocean. For example, were the euxinic conditions encountered in the Paleogene and early Miocene sediments typical for the entire water column? Or were they restricted to the upper water column? Drilling a depth transect will help answer this fundamental question. Drilling other ridges, that have had a different tectonic history than the Lomonosov Ridge, may also contribute to answer this question about Arctic's crucial ventilation history. 4. Ice-rafted dropstones were never a problem during the ACEX operations. In terms of lithologies, fairly standard paleoceanographic drilling conditions were encountered, mostly silty muds intersected by thin sand layers which became progressively stiffer with increasing penetration depth, until softer Eocene biosiliceous oozes were encountered. This holds great promise with respect to future core recoveries, which were modest during ACEX (68%). 5. Finally, future paleoceanographic expeditions to the Arctic Ocean, and elsewhere too, should adopt the use of the Rhizon pore water sampling technique that was systematically and successfully employed for the first time during ACEX [Dickens et al., 2007]. This editorial summary benefited from the careful reviews provided by Paleoceanography Editors Jerry Dickens and Eelco Rohling.
Referência(s)