18O from 0.97
(232.10 mbsf) to 1.66 (233.88 mbsf) with the 26-Ma warming event and our downhole decrease from 1.99 (175.03 mbsf) to 1.22 (184.04 mbsf) with the 23-Ma cooling event.
Mixed terrigenous-pelagic sediments from Hole 1139A between 150 and 390 meters below seafloor (mbsf) constitute an approximately two-component system. To determine the cause of fluctuations in carbonate concentration, noncarbonate and carbonate fluxes must be calculated independently. The fluxes, or mass accumulation rates (MARs), are determined by multiplying the weight percent of each component, dry bulk density, and sedimentation rate. For example:
where,
MAR = g/cm2/m.y.,
wt% CaCO3 = g CaCO3/g sediment,
bulk density = g/cm3, and
sedimentation rate = cm/m.y.
Thus, the resolution of the depth-age curve (Fig. F4) strongly influences the flux estimates. Also, the low age resolution prohibits making flux estimates at the shorter timescales that correspond to the second- and third-order carbonate variations.
Carbonate accumulation rate varied only slightly at Site 1139 (Fig. F5A). It reached >2 kg/cm2/m.y. in the late early Oligocene (~30 Ma) and gradually fell to <1 kg/cm2/m.y. in the Miocene.
The two prominent minima in carbonate concentration at 100-160 and 250-290 mbsf (Fig. F3) correspond to high sedimentation rates and peaks in the noncarbonate flux. Specifically, noncarbonate accumulation peaked at >3 kg/cm2/m.y. at ~29 Ma, decreased to <1 kg/cm2/m.y. between 26 and 23 Ma, peaked at >4 kg/cm2/m.y. at ~22 Ma, and became minimal by 21 Ma (Fig. F5B). Hence, variable terrigenous input is largely responsible for the long-term variation of carbonate concentration in Hole 1139A cores.
The flux of clay minerals from Skiff Bank to the Site 1139 basin is judged to have been limited because the terrigenous fraction is mostly silt sized (Fig. F2B). The magnetic susceptibility record and magnetic grains observed during sample preparation also indicate that the terrigenous component is relatively coarse (>2 µm). Although the shipboard X-ray diffraction results list clay minerals, clay minerals were not prominent in XRD analyses of decarbonated samples of this study (Table T3). Whole-rock elemental analyses also suggest a predominance of igneous (i.e., unweathered) material in the terrigenous component (Table T8).
The flux of organic carbon at Site 1139 was ~5 g/cm2/m.y. (Fig. F5C). This falls at the low end of the range of carbon fluxes from different settings around the world including surficial pelagic sediments from the open oceans (Berner, 1982).
Lithology was most likely constant through the history of weathering and erosion of Skiff Bank. Ideally for the purposes of this study, the provenance of Skiff Bank sediments would be uniform flood basalts. Skiff Bank basement, however, comprises evolved alkaline materials, including trachytic basalts, trachytes, and even rhyolites. Volcanic and hydrothermal features indicate that erosion did not progress to a deep level in the basement at Site 1139 (Shipboard Scientific Party, 2000). There is no indication of a change in provenance through this interval as indicated by the gross mineralogy of the sediments (Table T3).
Despite small numbers of specimens compared to Leg 120 studies, the quality of our benthic foraminifer data is as good. The lack of significant variation indicates that any changes in relative sea level were not large enough to be recognized by this technique, which places Site 1139 in the lower bathyal zone (1000-2000 m) throughout the Oligocene and Miocene Epochs.
The nannofossil Braarudosphaera bigelowii (Sample 183-1139A-30R-CC) is normally a neritic taxon (Shipboard Scientific Party, 2000). Its occurrence, which coincides with the peak of terrigenous sedimentation rate at ~29 Ma, affords weak evidence for a minimum in water depth at this time. Possibly, the erosional pulse correlates with the mid-Oligocene glacioeustatic sea level fall TB1.1 and subsequent lowstand (Haq et al., 1987) and oxygen isotopic zone Oi2 (e.g., Pekar and Miller, 1996).
The subsidence and uplift history of Skiff Bank is constrained by benthic foraminifers (Fig. F6). An unexpectedly old basement age of 68-69 Ma (Duncan, 2002) suggests that the majority of subsidence occurred in the early Paleogene. The lack of variation in benthic foraminifers is consistent with Skiff Bank lithosphere having stabilized by the Oligocene (~35 m.y. later).
Skiff Bank lies ~350 km from the Kerguelen archipelago (Fig. F1). Flood basalt activity on the archipelago peaked at ~25 Ma (Nicolaysen et al., 2000) (Fig. F6) when the terrigenous sedimentation rate at Skiff Bank reached a minimum (Fig. F5B). This broad temporal coincidence suggests a possible connection through depression of the lithosphere around the thick pile of flood basalts that accumulated to form the archipelago. This connection, however, is considered dubious because the peripheral forebulge around Hawaii has a similar radius but Hawaii is built on relatively old, thick lithosphere; most likely, Skiff Bank lies too far from the Kerguelen archipelago to have been affected by plume activity during the Oligocene.
A bloom of the nannofossil Braarudosphaera bigelowii (Sample 183-1139A-30R-CC) suggests biologic forcing in the mid-Oligocene (Shipboard Scientific Party, 2000) and possibly contributed to the slightly higher carbonate accumulation rates between ~31 and ~27 Ma (Fig. F5A). This event, however, cannot account for the subsequent relatively high carbonate concentrations between ~27 and ~23 Ma (Fig. F7A), a time when the carbonate accumulation rate was relatively slow and constant (<2 kg/cm2/m.y.) (Fig. F5A). Hence, high biological productivity is not a likely cause of the high carbonate concentrations in the latest Oligocene.
We have few direct constraints on the local climate at Site 1139. Surficial conditions, however, may be inferred from our oxygen isotopic analyses on benthic foraminifers from Hole 1139A (Fig. F7B) and by correlating with more detailed records from other sites. In particular, the detailed records indicate a prominent late Oligocene warming at around 26 Ma and a glaciation at the beginning of the Miocene (just before 23.0 Ma) (Zachos et al., 2001). A future study should explore an age model that correlates our downhole increase in 18O from 0.97 (232.10 mbsf) to 1.66 (233.88 mbsf) with the 26-Ma warming event and our downhole decrease from 1.99 (175.03 mbsf) to 1.22 (184.04 mbsf) with the 23-Ma cooling event.
An important aspect of climate often forgotten is wind strength. Considering that Skiff Bank may have been a small island in a large ocean, variable wind speed and wave intensity through climate cycles may have been very significant in determining the erosional history of the coastline.
Carbonate productivity and tectonic factors are less likely to have controlled the variations of noncarbonate concentration at Site 1139 than glacioeustatic sea level and climate changes. First, on a multimillion-year timescale, the coincidence of high noncarbonate concentration and fast sedimentation rate indicates that carbonate was diluted (Figs. F3, F4); this relationship might extend to shorter timescales. Second, grain sizes increase with increasing noncarbonate (Fig. F2B), suggesting that the gradational boundary between terrigenous and pelagic facies shifted basinward during lowstands of sea level. Third, opal concentration increases with increasing noncarbonate (Fig. F2B), which is consistent with latitudinal shifting of the gradational facies boundary between cold-water silica-rich sediments and warm-water carbonate-rich sediments. (Alternatively, opal may partially dissolve in carbonate-rich sediment due to higher pH.) Fourth, the carbonate pattern resembles the glacioeustatic record (Zachos et al., 2001). Fifth, the setting of Site 1139 resembles that of other sites of mixed terrigenous-pelagic sedimentation (e.g., Site 594 and Ceara Rise), where noncarbonate tracks the benthic foraminifer oxygen isotope record (low carbonate during glacials is attributed to dilution by terrigenous sediments).
In the carbonate-rich intervals, intermediate (>1 m) cycles may correlate with million-year to 400-k.y. cycles identified in oxygen isotope and continental margin studies (Oi and Mi events of Miller et al., 1991). The highest frequency cycles may be 40 k.y. in duration, which are conspicuous in Ceara Rise sediments that accumulated at a similar rate (Zachos et al., 2001).
How did silicate weathering vary and respond to environmental changes at Skiff Bank? The paucity of clay minerals and clay-sized sediments indicates that physical weathering and erosion were dominant (i.e., weathering-limited regime of Stallard and Edmond, 1983) and possibly that consumption of atmospheric carbon dioxide by silicate weathering was minimal (although this flux is high in Iceland, where high rainfall and mechanical breaking of rocks compensate for the low temperatures) (Gislason et al., 1996).
The organic carbon concentrations are low and relatively constant (Fig. F5C). The likely causes include minimal terrigenous OC production in a cold, hostile climate and poor preservation in coarse sediments with low mineral surface area that were extensively burrowed at a depth below the oxygen minimum layer.
A crude comparison between Skiff Bank and Himalayan carbon fluxes can now be made. Carbon dioxide consumption related to silicate weathering is unlikely to have been excessive at either location, at Skiff Bank because of limited weathering and in the Himalayan system because of inappropriate materials (paucity of soluble Ca- and Mg-silicates) (France-Lanord and Derry, 1997). The organic carbon flux at Skiff Bank (~5 g/cm2/m.y.) is minor compared to the Bengal Fan, where long-term sedimentation rates are triple and OC concentrations are up to an order of magnitude greater (France-Lanord and Derry, 1994).
