Previous   |   Next

doi:10.2204/iodp.proc.320321.101.2010

Results and highlights

During Expedition 320, 16 holes at 6 sites (Holes U1331A–U1331C, U1332A–U1332C, U1333A–U1333C, U1334A–U1334C, U1335A and U1335B, and U1336A and U1336B) (Table T1) were cored as part of the PEAT program. During Expedition 321, two major Neogene equatorial Pacific sediment sections were recovered by drilling seven holes at two sites (Holes U1337A–U1337D and Holes U1338A–U1338C) (Table T2). One additional hole (U1338D) was drilled at Site U1338 to provide sediment cores for laboratory training conducted during IODP Expedition 323. By drilling a series of sites that follow the position of the paleoequator and a limited latitudinal and depth transect, as outlined below, we recovered cores that allow us to address the combined PEAT objectives (see "Scientific objectives").

Lithologies recovered and stratigraphic summary

Figure F12 summarizes the lithostratigraphy of the northwest–southeast transect of sites drilled during Expedition 320/321 together with the sedimentary sequence from Site 1218, which is included in the PEAT flow-line strategy (see "Scientific objectives, introduction, and background"). In this figure, the Eocene sequence (green shading) thins from north to south, pinching out before Site U1336 where the basement is of early Oligocene age. The Miocene sequence (yellow shading) thins substantially from south to north, whereas the Oligocene sequence (blue shading) is thickest in the middle of the transect (Sites U1334 and U1335) and thins both north and south. Sediments of early Eocene age are only present at Site U1331, and sediments of Pliocene–Pleistocene age are present at Sites U1331, U1334, U1335, U1337, and U1338. The thickness of sediments from different parts of the age transect is compatible with that expected from our drilling strategy (see "Site selection strategy and site targets").

Six main lithologies are present in these sites:

• Surface clays,

• Nannofossil oozes and chalks,

• Radiolarian ooze and radiolarite,

• Diatom ooze and diatom nannofossil ooze,

• Porcellanite and chert, and

• Basal chalks, limestones, and clay.

Surface clays are present at Sites U1331–U1335. These sediments are affected by intense dissolution, which has removed most of the siliceous and calcareous biogenic components, and only the clay components remain. This pattern reflects the northward motion of the Pacific plate, which transports the sites out of the equatorial upwelling zone of high production as the underlying crust subsides with age.

Nannofossil oozes and chalks are the main lithology of the Oligocene in the PEAT transect. To the south, at the sites near the paleoequator, nannofossil ooze is also the main lithology in strata of early and middle Miocene age. At the older sites nannofossil ooze is also present but in decreasing importance moving northwestward as basement age becomes progressively older and water depth increases. At the northernmost end of the transect (Site U1331), nannofossil ooze is primarily restricted to an interval of early and middle Eocene and Oligocene age. The middle Eocene interval is broadly correlative to carbonate accumulation event (CAE) 3 (Lyle et al., 2005).

Radiolarian ooze is present at Sites U1331–U1336 (Fig. F12), but it is a far more abundant lithology in the north (Sites U1331 and U1332), where it is the main lithology of the Eocene, than in the south (Sites U1334–U1336).

Porcellanite is a major lithology in sediments of early and early middle Eocene age toward the northern end of the Expedition 320/321 transect (especially at Sites U1331 and U1332), where it is associated with thin clay horizons that are interbedded with radiolarian ooze. Chert is a major lithology in sediments of early and early late Oligocene age at Site U1336, where it proved a major impediment to recovery of a complete section. One stratigraphically equivalent chert/porcellanite layer was found in middle Miocene sediments at Sites U1337 and U1338.

At nearly all sites drilled on the PEAT transect, the recovered basal sediments overlying basalt are calcareous in lithology, indicating that, at the onset of sediment accumulation at these sites, the seafloor lay above the local CCD. This result is in keeping with the Expedition 320/321 rationale of drilling a flow-line of sites on crust of increasing age (southeast to northwest) to recover stratigraphic "windows" of calcareous sediments overlying contemporaneously young crust prior to its subsidence below the CCD. The only site on the PEAT transect where calcareous sediments were not found overlying basement is Site U1332. The basal sediment unit is primarily zeolite clay of middle Eocene age, although calcareous nannofossils were common in several samples. This result indicates that the crust at Site U1332 lay below the CCD, even at the point of its formation (~50 Ma), pointing to a very shallow CCD (<2700 m) at this time (see "Cenozoic CCD in the equatorial Pacific").

The combined results of Leg 199 and the PEAT program (Figs. F13, F14, F15) will allow us to decipher paleoceanographic and paleoclimatic changes within a latitudinal and depth transect in the equatorial Pacific Ocean. Intervals of interest include the EECO (Zachos et al., 2001a; Lyle, Wilson, Janecek, et al., 2002), the MECO (Bohaty and Zachos, 2003; Bohaty et al., 2009), the middle through late Eocene CAE events (Lyle et al., 2005), the Eocene–Oligocene transition (Coxall et al., 2005), late Oligocene warming (see supplementary material in Pälike et al., 2006b), the Oligocene–Miocene transition (Zachos et al., 2001b; Pälike et al., 2006a), the middle Miocene glaciation intensification event (Holbourn et al., 2005), and late Miocene–Pleistocene cooling.

Sediments of early Eocene age in the Pacific equatorial age transect are characterized by clay, cherts, and porcellanite. In general, sediments of middle Eocene age from the equatorial Pacific transect are dominated by radiolarian ooze, radiolarite, and clay, but carbonate-rich intervals also occur and appear correlative between Expedition 320 sites and Sites 1218 and 1219, where the CAE events were originally defined. Sediments of middle Eocene age at Site U1333 are carbonate rich compared to those at Site 1218, an observation that cannot be explained by their assumed relative paleodepths during this interval. The Eocene–Oligocene transition is characterized by a major lithologic change from Eocene radiolarian oozes to Oligocene nannofossil oozes at ODP Sites 1218–1221 and Sites U1331–U1333. At Site U1334, sediments of latest Eocene age are more carbonate rich than in any of the other Eocene intervals.

The Oligocene–Miocene transition occurs within a succession of pale and dark nannofossil ooze cycles at all sites where it was recovered except for Site 1220, where it is characterized by a transition from Oligocene radiolarian ooze to Miocene clay. Clay- and diatom-rich sediments characterize the middle to late Miocene interval at Sites U1331–U1335.

Within the PEAT program, only Sites U1337 and U1338 recovered sedimentary sections of late Miocene–Holocene age with relatively high sedimentation rates and preserved carbonate (~15 m/m.y.) (Fig. F14). Site U1335 recovered this interval, but sedimentation rates are ~6 m/m.y. Sites further west and north on older crust (Sites U1331–U1336) suffer from low sedimentations rates or hiatuses in the younger section because they are far from the modern equatorial productivity zone and on deeper (older) ocean crust where depth-dependent carbonate dissolution has become strong. Of the older sites, U1334–U1336 had the highest sedimentation rates in the interval from middle Miocene through late Oligocene.

Both Sites U1337 and U1338 recovered continuous, essentially complete Neogene sedimentary sections, Site U1338 from 0 to >17 Ma and Site U1337 from 0 to >23 Ma, just beyond the Oligocene/Miocene boundary. These represent the only complete Neogene sections in the equatorial Pacific, possibly for all the tropics, that have high enough sedimentation rates to resolve orbitally forced sediment cycles. Coring during Legs 8, 9, and 16 was accomplished with the rotary core barrel, and the sediments were highly disturbed. During Leg 85 at Site 574, the middle Miocene to present was double APC cored, but sedimentation rates were slow from 0 to 10 Ma (~5 m/m.y.) and the middle Miocene section was never spliced (Mayer et al., 1985). Below the middle Miocene, only one hole was drilled. Similarly, Site 572 had only a single hole drilled through the upper and middle Miocene interval (Mayer et al., 1985). Most Leg 138 sites were drilled on 10 Ma crust and did not contain significant middle and lower Miocene sediments (Mayer, Pisias, Janecek, et al., 1992). Site 846 was drilled on 17 Ma crust on the Nazca plate but experienced poor recovery below 12 Ma. In addition, Site 845 in the Guatemala Basin has a spliced sediment section to only ~12 Ma (Mayer, Pisias, Janecek, et al., 1992), even though it was located on 17 Ma crust. Even in the tropical Atlantic on the Ceara Rise, the middle and early Miocene were discontinuously cored and longer records were spliced between different sites (Shipboard Scientific Party, 1995; Raffi et al., 2006).

Cenozoic CCD in the equatorial Pacific

One of the primary objectives of the PEAT program was to detail the nature and changes of the CCD throughout the Cenozoic in the paleoequatorial Pacific (Objective 1 in "Scientific objectives"), with potential links to organic matter deposition (Olivarez Lyle and Lyle, 2005). The choice of drilling locations, targeting positions on the paleoequator to track carbonate preservation during crustal subsidence throughout time (Figs. F13, F16), followed the initial work on DSDP sites by van Andel (1975). This first reconstruction of the Cenozoic CCD was augmented by additional results from Leg 199 (Lyle, Wilson, Janecek, et al., 2002; Rea and Lyle, 2005). One of the significant contributions of Leg 199 drilling was the latitudinal mapping of CCD variations with time. During the Eocene, the CCD appeared to be deeper outside the zone ±4° from the Equator, opposite to the pattern established during the Neogene (Lyle, 2003). The PEAT cores will allow us to refine our knowledge of temporal and spatial variation in sediment accumulation rates resulting from plate movement, varying biologic productivity at the equatorial divergence, and carbonate preservation (Fig. F13). The shipboard sampling program allowed for >1000 determinations of CaCO₃ concentrations, approximately one every section from one hole of each site. Shipboard results reveal the carbonate accumulation events of Lyle et al. (2005) as sharp carbonate concentration fluctuations at ~44, 41, 39, and 36 Ma across Sites U1331–U1334 and 1218, followed by a sharp transition into much higher carbonate accumulation rates from the Eocene into the Oligocene. Results from the PEAT expeditions reveal a complex latitudinal pattern where Sites U1331, U1332, and U1334 track the equatorial CCD that matches well with the signal observed from Site 1218. However, Site U1333, which is slightly north of the equatorial zone during the Eocene–Oligocene transition, shows significantly more carbonate accumulation than expected at this latitude and paleodepth.

The early middle Eocene equatorial CCD is much shallower than previously thought, as shown by results from Site U1332, where we recovered little or no carbonate in the basal sediment section above basement, in contrast to Site U1331, which is only ~2 m.y. older. Our estimated CCD at ~49 Ma is only ~3000 m deep. Surprisingly, Expedition 320 results also show a shallower CCD than previously known during the late Oligocene, perhaps 300 m shallower in the time interval between 23 and 27 Ma. This shallower CCD, at a paleodepth of ~4.5 km, and associated reduced carbonate fluxes to the seafloor could be linked to the gradual late Oligocene cooling, first fully documented at Site 1218 (supplementary fig. 3 in Pälike et al., 2006b). The design of our drilling locations in combination with existing data will allow us to generate a 3-D view of CCD evolution during the Cenozoic during postcruise research.

Late Miocene carbonate crash and middle Miocene glaciation

Sites U1337 and U1338 recovered a full suite of Pleistocene to earliest Miocene calcareous and siliceous microfossils. All groups exhibit prominent downcore variations in preservation and abundance that appear related to major changes in primary production, export flux, and water-column and seafloor dissolution, thus reflecting fundamental changes in global climate. The carbonate crash, an extended period of low carbonate deposition widely recorded throughout the eastern Pacific Ocean at ~9–11 Ma (Lyle et al., 1995; Farrell et al., 1995), is strikingly expressed at Sites U1337 and U1338 by sharp decreases or the disappearance of planktonic foraminifers, high benthic to planktonic foraminifer ratios, and generally impoverished benthic foraminifer assemblages (Fig. F17). Calcareous and siliceous microfossils indicate that this prominent dissolution event was less intense and of shorter duration at Site U1338 than at Site U1337 (Fig. F16, F17), probably because it is located on younger oceanic crust and in closer proximity to the Equator during this time window. Site U1338 was ~200 m shallower than Site U1337 and essentially at the Equator during the carbonate crash interval.

During episodic expansion of the Antarctic ice sheet at ~14–15 Ma, calcareous microfossils exhibit overall good preservation and relatively high diversity, suggesting a vigorous Pacific Ocean circulation and deep CCD. In contrast, the preceding prolonged period of global warmth, often referred to as the Middle Miocene Climate Optimum (MMCO), appears marked by transient changes in CCD. Postcruise studies and integration of paleontological data will provide an opportunity to further investigate temporal and spatial variations in microfossil distribution across the equatorial Pacific Ocean and to unravel links with global climatic and oceanographic events during the Neogene (Objectives 1 and 9 in "Scientific objectives").

Biostratigraphy and preservation

Biostratigraphic integration

A virtually complete composite section with biogenic sediments spanning 52 m.y. from the upper Pleistocene to the lower Eocene was recovered during Expedition 320/321 (Objectives 2, 3, 4, and 7 in "Scientific objectives"). The youngest record of the last 12 m.y. (late middle Miocene to Holocene) was well preserved at Sites U1335, U1337, and U1338 but elsewhere is only present as a thin (5–10 m) section of noncalcareous brown clay. Biostratigraphic records spanning the middle Miocene through lower Eocene are composed of nannofossil and radiolarian oozes as two major biogenic components. At Sites U1331 and U1335 turbidite beds containing reworked microfossils were present, with mixing most obvious at Site U1331. At the shipboard biostratigraphic resolution, all drilled sites contribute apparently continuous successions to this composite section and stratigraphic highlights include multiple recoveries of Eocene–Oligocene transitions at Sites U1331–U1334 and Oligocene/Miocene boundaries at Sites U1332–U1337. These sections provide excellent records of biotic response to rapid environmental change in the principal phyto- and zooplankton groups as well as benthic foraminifers.

For Expedition 321, one of the highlights was the recovery of well-preserved and diverse assemblages of siliceous radiolarians and diatoms and calcareous planktonic foraminifers and nannofossils for the Neogene. Although the zonation and calibrations of each of these groups have been established and refined over the last 50 y, there are very few locations with both well-preserved and abundant siliceous and calcareous microfossils. Leg 199 established a well-defined radiolarian stratigraphy and calibration for the equatorial Pacific Ocean (Nigrini et al., 2006); unfortunately only for the Oligocene were the planktonic foraminifers and calcareous nannofossils sufficiently preserved to provide robust calibrations (Pälike et al., 2006b; Wade et al., 2007).

In spite of a well-defined Neogene diatom stratigraphy and calibration for the low-latitude Pacific (Baldauf and Iwai, 1995; Barron, 1985), it is clear from the initial results that the presence of some of the diatoms used in the diatom stratigraphy is discontinuous in the expanded Miocene section. This has been noticed in sediments recovered during Leg 138 (Baldauf and Iwai, 1995). The discontinuous occurrences may reflect large variations in the abundances of these species with time and changing ecologic conditions (Objective 7 in "Scientific objectives").

The diatom, radiolarian, calcareous nannofossil, and planktonic foraminifer stratigraphies at all sites generally agree, with some minor discrepancies. The fully integrated biostratigraphies coupled with the cyclostratigraphy and paleomagnetostratigraphy will lead to an integrated magneto-, astro-, and biochronology of all four microfossil groups for the equatorial Pacific Ocean (Objective 3 in "Scientific objectives").

Dissolution and microfossil preservation

The preservation of carbonate microfossils varies dramatically throughout the succession as a result of biotic production and export rates, water-column and seafloor dissolution, and other processes. The strength of dissolution reflects the depths of the drilling sites, which are all presently bathed at >4.3 km water depth, whereas the amount of dissolution strongly depends on the paleodepth (subsidence) history at each site and fluctuations of the CCD on a regional and basin-wide scale. The dissolution effect is greatest in the oldest successions at Sites U1331–U1333 (Figs. F13, F18). The Eocene equatorial CCD has been estimated at a depth shallower than ~3.5 km (Lyle et al., 2005), with short-term CCD fluctuations occurring during the middle to late Eocene based on preservation of calcareous microfossils and calcium carbonate records at Sites U1331–U1334. The most striking CCD change has been recorded close to the Eocene–Oligocene transition where the sediments change from radiolarian-dominated Eocene sediments to Oligocene nannofossil oozes. The depth transect of these sites indicates a deepening of at least 1 km over this short time interval. The recovery of carbonate-rich Oligocene successions at all sites that penetrated an Oligocene sediment section is evidence for a considerably deeper CCD (>4.5 km water depth) throughout this interval (see Fig. F13).

A compilation of semiquantitative estimates of preservation and abundance of calcareous microfossils reveals a strong coupling of the fossil records with paleodepth history and the CCD at these drilling sites. The carbonate dissolution effect is strikingly different between microfossil groups (Fig. F18). Planktonic foraminifers are the most sensitive to dissolution, and well-preserved specimens were found in sediments if carbonate contents exceeded at least 60–70 wt%. At the deepest Site U1331, planktonic foraminifers are only present during carbonate maxima in the Oligocene, middle Eocene, and early Eocene sediments. At Site U1332 they were present only in the carbonate-rich Oligocene nannofossil oozes (Fig. F18). At Sites U1334 and U1335 high-carbonate sediments (80–90 wt%) contain abundant and well-preserved planktonic foraminifers.

Calcareous nannofossils and benthic foraminifers are less susceptible to dissolution than planktonic foraminifers and closely track the presence or absence of carbonates in the sediments. The preservation of both groups varies with carbonate content, but the preservation of calcareous nannofossils varies even in sediments with low carbonate content that are barren of planktonic foraminifers. Poor preservation of specimens is observed in sediments of 5–25 wt% carbonate contents, moderate preservation in sediments of 30–70 wt%, and good benthic foraminifer and moderate to good nannofossil preservation in sediments with >75 wt% CaCO₃.

Within the nannofossil assemblages, however, certain taxa are never present in these sediments, such as holococcoliths. The relatively robust holococcolith Zygrhablithus bijugatus was only recorded in one or two samples from Sites U1335 and U1336, and other taxa show distributions that are more similar to those of planktonic foraminifers, such as the long-ranging heterococcolith genus Helicosphaera (Fig. F18).

Biotic response to climatic change, Neogene Sites U1337 and U1338

Phytoplankton, zooplankton, and benthic foraminifers provide an excellent opportunity to document the biotic response to intervals of rapid climatic change in both the surface and deep ocean, addressing PEAT Objectives 1, 2, 4, and 7 (see "Scientific objectives").

Planktonic foraminifer assemblages are generally well preserved at Sites U1337 and U1338. As with other microfossil groups, distinct changes in assemblage composition reflect preservational and evolutionary changes, fluctuations in the water-column structure, and the position of each site relative to the paleoequator (Fig. F19). Globoquadrinids and dentoglobigerinids dominate the assemblages from the Oligocene to middle–late Miocene. This interval is also associated with extremely abundant and diverse paragloborotaliids. These taxa are considered to occupy the thermocline and prefer nutrient-rich environments (Wade et al., 2007). The decrease in globoquadrinids and dentoglobigerinids is associated with an increase in Globigerinoides, which increases in abundance from the middle Miocene. Keeled globorotaliids dominate the Pleistocene assemblages. Peak abundances of Globigerinoides spp. occur within the late Miocene.

There appears to be a relationship between the diversity in the planktonic foraminifers and climatic change. At Site U1338, a high diversity of planktonic foraminifers is recorded in Zone M5, with many microperforate taxa (Globigerinita, Globigerinatella, Mutabella, and Tenuitella). The diversity of this group is particularly high through the early middle Miocene. These taxa occupied the warmer mixed layer (Pearson et al., 2001b; Majewski, 2003), and the high diversity may be related to higher SSTs and increased stratification associated with the MMCO.

Planktonic foraminifers, because they dissolve more easily than benthic foraminifers or nannofossils, are useful to evaluate the degree of dissolution in calcareous biogenic sediments (Objective 1 in "Scientific objectives") (Fig. F17). Intervals of enhanced dissolution have been detected at Sites U1337 and U1338 and are associated with laminated diatom oozes between 11.5 and 9 Ma, the "carbonate crash" interval. Another short interval is barren of planktonic foraminifers and marked by a high benthic/planktonic foraminifer ratio at ~17 Ma at Site U1337.

Much like the planktonic foraminifers, calcareous nannofossil assemblages reflect global and regional oceanographic changes. Neogene placolith diversity peaks in the early to middle Miocene, during the MMCO. At ~14 Ma, nannofossil assemblages begin to fluctuate, indicating changes in productivity. At Site U1338 there are several instances when assemblages are dominated by very small (<3 µm) Dictyococcites species. These events are also less frequently present at Site U1337. These dominance intervals may reflect greater surface water productivity and perhaps suggest a more intense upwelling regime at the MMCO. Another indication that Site U1338 records a higher productivity signal is the exclusion of certain taxa, such as the marker species Helicosphaera ampliaperta and Catinaster coalitus, both of which are present but rare at Site U1337.

Lower bathyal to abyssal benthic foraminifers occur throughout the early Miocene to Pleistocene at Sites U1337 and U1338 and show relatively good preservation. Marked variations in downcore abundance and assemblage composition may relate to fundamental changes in global climate, major ice-volume fluctuations, and reorganizations in oceanic circulation during the Neogene. Organic flux–sensitive taxa track substantial changes in equatorial Pacific Ocean surface productivity throughout the latest Miocene to Pleistocene. Impoverished assemblages and high benthic/planktonic ratios prominently mark the late Miocene carbonate crash (Fig. F17), whereas diversified assemblages, including high numbers of epifaunal or near surface dwellers such as Cibicidoides, signal a marked improvement in deep ocean ventilation following middle Miocene high-latitude ice sheet expansion. Early to early middle Miocene assemblages exhibit significant fluctuations in abundance and diversity, hinting at major changes in upper ocean structure and deepwater ventilation at the onset of and during the MMCO.

Basal carbonates

Eocene basal carbonate sediments were confirmed on top of basaltic basement at all PEAT sites apart from U1332 and U1336 (Fig. F20). At Site U1332, several samples immediately on top of basement basalt contained relatively common calcareous nannofossils. The existence of carbonates suggests that paleodepths were above the shallow Eocene CCD during the early to middle Eocene. At Sites U1331 and U1333–U1335, these carbonate intervals are thin and their lower parts are lithified to limestones, probably because of the combined influence of hydrothermal fluids and overburden. These sediments contain slightly diagenetically modified calcareous microfossils and are barren of siliceous microfossils. The sites where Eocene sediments were recovered were located within 2° latitude of the paleoequator at the time of first sediment accumulation, so evidence of equatorial upwelling might be expected to have influenced the assemblages (see below). The absence of siliceous microfossils is likely a result of dissolution associated with hydrothermal flow of the crust (Moore, 2008a).

Productivity indicators

Shipboard analyses of quantitative microfossil faunal assemblages allow only preliminary speculation on potential microfossil-based productivity indicators; quantitative work will be required to follow up on these initial observations and fully address Objective 2 of the PEAT program (see "Scientific objectives"). Calcareous nannofossil assemblages at these lowest stratigraphic levels are not strikingly different from younger examples. At several sites, however, the common presence of taxa that are considered to be higher productivity (or cooler water) indicators is suggestive of an upwelling signal (e.g., common Chiasmolithus and small reticulofenestrids [Reticulofenestra minuta] at Sites U1333 and U1334). At all sites sphenoliths are also common at these levels, and although some representatives of this group are considered to be oligotrophs (e.g., Gibbs et al., 2004), certain species clearly display more opportunistic behavior, which may explain their abundant presence here (e.g., Wade and Bown, 2006; Dunkley Jones et al., 2008).

The absence or relative rarity of the warm-water oligotrophic discoasters in the lowest parts of Sites U1331, U1333, and U1334 is also suggestive of higher productivity surface waters, at least for the sediments immediately overlying basement. Discoasters are actually common in the lowest sediments at Site U1331, and this either represents selective concentration due to the dissolution of less robust taxa or indicates that these species (Discoaster deflandrei and Discoaster lodoensis) were adapted to more eutrophic paleoenvironments. Planktonic foraminifer assemblages in the basal carbonates at Sites U1333 and U1334 are dominated by relatively robust taxa: subbotinids, parasubbotinids, and paragloborotalids. These genera are thought to occupy a (sub)thermocline habitat (Wade et al., 2007; Sexton et al., 2006) and are often associated with high-productivity environments (Wade et al., 2007), an association consistent with both sites being situated in the equatorial upwelling region. However, planktonic foraminifer assemblages at Sites U1333 and U1334 may also be biased toward these more robust taxa by the effects of dissolution. Detailed assemblage study is required to elucidate the relative contributions of calcium carbonate dissolution and the true paleocological signal.

Radiolarian stratigraphy

The radiolarian stratigraphy in sediments recovered during Expedition 320 (Table T3) spans Zones RN14 (lower Pleistocene) to RP10 (lower middle Eocene) and provides the highest shipboard biostratigraphic resolution for most sections within the Eocene. The preservation of assemblages is generally good with only a few scattered intervals of moderate to poor preservation. Nigrini et al. (2006) took a comprehensive approach toward establishing ages for all radiolarian datums recovered at Leg 199 sites. In so doing, they produced age estimates for >300 radiolarian datums, greatly enhancing our ability to date the Cenozoic section in the tropical Pacific. However, several of these calibrations need to be checked and/or refined. In addition, some of the datums appear to be more reliable than others, and this needs to be further evaluated. Results from Sites U1331–U1334 will allow recalibration of radiolarian datums near the Eocene/Oligocene boundary.

In some cases, variation in the levels of first and last appearances of species may be caused by variation in taxonomic interpretation, but more often these variations are due either to real differences in the ranges of species at different locations in the tropical Pacific or to the extremely low abundance of certain species in samples from a given site. Thus, taxonomic difficulty, abundance, and preservation all figure into how well a species serves as a stratigraphic marker.

A few levels were not adequately covered by Leg 199 sites. For the upper Miocene we have relied on the radiolarian studies of Leg 138 in the far eastern tropical Pacific (Moore, 1995). In their studies of Leg 199 sites, Kamikuri et al. (2005) and Funakawa et al. (2006) added substantially to our understanding of the radiolarian assemblage transitions at the Oligocene/Miocene and Eocene/Oligocene boundaries. However, these studies focused on the statistical changes in the faunal assemblages as a whole and made no wholesale attempt to recalibrate first and last appearance datums. During Expedition 320 we were able to add to the stratigraphic control in the lower part of the middle Eocene collected at Site U1331 and test the usefulness of individual datums at the sites drilled. The full integrated biostratigraphies will address Objective 3 (see "Scientific objectives").

Diatoms

The diatom stratigraphy in sediments recovered during Expedition 320/321 represent the Pleistocene Fragilariopsis doliolus through the lower Oligocene Coscinodiscus excavatus Zones. Not all zones are represented at all sites given variability in the state of diatom preservation and the age of the sedimentary sections recovered.

The species observed are typical of the low-latitude eastern equatorial Pacific Ocean (Barron, 1985; Baldauf and Iwai, 1995). Primary biostratigraphic markers are typically applicable. Noted exceptions include the often inconsistent occurrences of Coscinodiscus gigas var. diorama, Actinocyclus moronenesis, and Cestodiscus peplum. When necessary, secondary stratigraphic indicators were used for zonal assignments.

Diatoms are most common in the Holocene through upper lower Miocene sediments (0–18 Ma) (Fig. F21). This interval is best represented at Sites U1335, U1337, and U1338. Distinct within this interval is the occurrence of thick sequences of laminated diatom oozes (diatom mats) at Site U1337 (Fig. F22). These diatom intervals are mainly composed of the needle-shaped diatom Thalassiothrix and are deposits that record unusual oceanographic conditions that repeated several times in the eastern equatorial Pacific region during the middle to late Miocene and early Pliocene (Fig. F23) (Kemp and Baldauf, 1993). One of the unresolved questions of these expeditions is why the diatom mats are prominent at Site U1337, whereas at Site U1338 the same time intervals contain diatom rich sediment, but only rarely laminated diatom intervals.

Kemp and Baldauf (1993) proposed that the diatom mats record high open-ocean sedimentation rates and may represent vast sinks of silica and carbon. However, a change in sedimentation rates across these intervals was not observed based on the coarse shipboard biostratigraphic resolution. The alternation between Thalassiothrix and mixed sediment laminae could reflect the periodic movement of the frontal zone in the eastern equatorial Pacific Ocean. An improved understanding of these diatom mats will be an important prerequisite for enhancing the understanding of the marine biogeochemical cycling and for assessing the impacts of rapid climate change on export production in the low-latitude eastern Pacific Ocean) Objectives 1, 2, 4, and 7 in "Scientific objectives").

Diatoms are less common in the lowermost Miocene through uppermost Oligocene sediments than those from the upper Neogene. The lower diatom numbers in part reflect a deterioration, but variability, in diatom preservation. Diatom abundance and preservation improves for the lowermost upper Oligocene and lower Oligocene and typically allow identification of the Bogorrovia veniamini through Coscinodiscus excavatus Zones. These zones are best represented at Sites U1333 and U1334. The base of the C. excavatus Zone approximates the Oligocene/Eocene boundary. Diatom preservation and therefore abundance deteriorates at or immediately below this boundary, with diatoms typically not observed in the majority of samples examined from the Eocene.

Paleomagnetism

Paleomagnetism and magnetostratigraphic studies are important observations needed to fulfill the expedition objectives of obtaining a well-intercalibrated Cenozoic megasplice and constraining Pacific plate tectonic motion (Objectives 3 and 5 in "Scientific objectives"). Results obtained so far indicate that the sediments recovered will provide the needed archive to address these objectives in a comprehensive fashion. For example, during Expedition 320, shipboard paleomagnetic results were obtained from 56,222 intervals measured along ~2000 split-core sections and from detailed progressive alternating-field (AF) and thermal demagnetization of 411 small discrete samples (Figs. F24, F25) These data indicate that a useful magnetic signal (characteristic remanent magnetization [ChRM]) is preserved in APC cores after removal of the drilling-induced overprint by partial AF demagnetization at 20 mT except for intervals affected by reduction diagenesis. Cleaned paleomagnetic data were characterized by shallow inclinations, consistent with the sites being near the paleoequator, and by 180° alternations in declinations downhole, reflecting magnetic polarity zones (magnetozones). These qualities, along with demagnetization results from discrete samples, indicate that the ChRM is the primary depositional remanent magnetization. Magnetostratigraphy at each site is constructed by correlating the distinct declination alternations with the geomagnetic polarity timescale (GPTS) (Fig. F24).

The magnetic polarity stratigraphy of the entire middle Eocene to Quaternary interval was resolved at the eight sites occupied during Expedition 320/321. The middle Eocene and Oligocene intervals were well documented at Sites U1332 and U1333. At Site U1332, mean sedimentation rates in this interval range from 3 to 8 m/m.y., and at Site U1333 the mean sedimentation rates reach ~12 m/m.y. in the lower Oligocene. At Site U1334 the upper Oligocene and lower Miocene magnetic stratigraphy was well resolved with mean sedimentation rates of 12 m/m.y. in the lower part of the interval and 4 m/m.y. in the upper part. At Site U1335, the upper Miocene and Quaternary magnetic stratigraphy was well resolved at sedimentation rates of ~6 m/m.y., and below these sediments lies an interval of upper middle Miocene where the magnetic stratigraphy is not resolved because of weak magnetization intensities. This ~20 m long unresolved interval is underlain by sediments that carry a magnetic stratigraphy of part of the middle and lower Miocene, although the interval is interspersed with turbidites and evidence of reworking. At Site U1336, the magnetic stratigraphy of part of the middle and lower Miocene is resolved in the 0–80 m CSF interval at mean sedimentation rates of ~10 m/m.y.; however, below ~80 m CSF magnetization intensities are too low for resolvable magnetization directions. At Site U1337, although the magnetic polarity stratigraphy is interpretable for the late Miocene to Quaternary, data quality is not high for the upper Miocene. At Site U1338, the magnetic stratigraphy is resolvable in three intervals: part of the Pliocene–Pleistocene (0–4 Ma), part of the late Miocene (8.7–11 Ma), and part of the middle Miocene (12.7–16 Ma). These three intervals at Site U1338 are separated by intervals of low magnetization intensities where the magnetostratigraphy was not resolvable.

In total, the magnetostratigraphies from Expedition 320/321 yield 868 dates ranging from 51.743 Ma (base of Chron 23n.2n at Site U1331) to the present. In addition, 83 short polarity intervals were observed that might correspond to cryptochrons or geomagnetic excursions. At these short events and at the geomagnetic reversals, magnetization intensities are low, as would be expected if the sediments are accurately recording the past paleomagnetic field intensity. Analysis of paleomagnetic directions over stable polarity intervals (full chrons) indicates the long-term record provides paleolatitude information that will aid in refining the Pacific apparent polar wander path and directional dispersion information for studying geomagnetic secular variation (Objective 5 in "Scientific objectives"). Thus, besides providing ages for the eight sites, the high-quality paleomagnetic records have the potential to resolve long- and short-term geomagnetic field variability and provide important plate kinematic constraints.

Cenozoic megasplice and stratigraphic correlation

One of the major shipboard scientific efforts for every paleoceanographic drilling cruise is to place the depth column into a time framework. For the PEAT program a prime objective was a detailed intercalibration of bio-, magneto-, and chemostratigraphic records for the Cenozoic from the early Eocene to the present within an astronomically age-calibrated framework (Objective 3 in "Scientific objectives"). During Expedition 320/321 we depended on shipboard biostratigraphy and magnetostratigraphy for the basic time framework. Shipboard biochronostratigraphy is based upon the PEAT timescale (see "Reconstruction of geologic age during Expedition 320/321" in the "Methods" chapter) developed before the PEAT expeditions and to be refined postcruise. Depth positions of recognized biostratigraphic zones and microfossil datum levels will be listed in each of the site summaries.

The PEAT program was designed to incorporate results from Leg 138 for the younger Neogene part and Leg 199 for time intervals in the Eocene (Pälike et al., 2008) (Fig. F2). Expedition 320/321 shipboard results indicate that we can achieve this objective, based on the observation that even decimeter-scale features in the sedimentary record from the drilled sites can be correlated over large distances across the Pacific seafloor (Fig. F9) (Pälike et al., 2005). The PEAT program will leave a long-lasting legacy for the detailed intercalibration of all major fossil groups, a detailed magnetostratigraphy with 870 identified dated reversals, and sedimentary cycles that can be calibrated across large distances in the Pacific Ocean. Figure F26 demonstrates that a Cenozoic megasplice can be constructed from the material recovered and spliced onto previous Leg 138 and 199 sites. Physical property data that proxy calcium carbonate oscillations at Sites U1331 and U1332 show a remarkable match with those from Site 1220, which also has an excellent magnetostratigraphy. Similarly, Sites U1333 and U1334 can be spliced to Site 1218, providing a coherent and integrated record of large-scale Pacific sedimentation patterns of biogenic material from the Eocene through the Miocene and younger (Fig. F27, F28). Such stratigraphic correlation makes possible the study of sedimentation patterns and mass accumulation rates at orbital resolution. The material recovered will also allow us to verify existing calibrations (e.g., Pälike et al., 2006b) and further extend these up into the Miocene and down into the Eocene.

We can exemplify the approach toward the Cenozoic by constructing a preliminary multisite splice of gamma ray attenuation (GRA) bulk density (Fig. F29). This record of GRA bulk density from tropical Pacific sediments encompasses a major part of the Cenozoic, stretching from the upper Eocene to the recent, using data from Legs 138 and 199 and the PEAT program.

Sedimentation rates

A study of paleoceanographic processes and variations of mass accumulation rates across the PEAT latitudinal transect and its evolution over time depends on a detailed knowledge of sedimentation rates (Fig. F14). The integrated bio- and magnetostratigraphies obtained for all Expedition 320/321 sites will allow us to fully exploit and understand the complex interplay of productivity, dissolution, and spatial biogenic sedimentation patterns, which leave their imprint in the sedimentation rates recorded at different drill sites. Depending on the crustal subsidence and age for each site, sedimentation rates vary from site to site over time (Figs. F14, F30).

Our results reveal the change in linear sedimentation rates (LSRs) in both the latitudinal and age transect components of the PEAT program. LSRs of the middle Eocene are extremely high, frequently >10 m/m.y., with a maximum of 18 m/m.y. at Site U1331. Rates at Sites U1332 and U1333 are similar (8–6 m/m.y.). LSRs of the late Eocene decrease to 3.5–6 m/m.y. at Sites U1331–U1333. The highest LSR peaks (>20 m/m.y.) exist in the early to late Oligocene section at Sites U1333 and 1334 and in the early Miocene for Sites U1336 and U1335. Middle to late Miocene sedimentation rates are in the range of 20–30 m/m.y. at Sites U1337 and U1338. LSRs also increase from the west (Site U1331) to the east (Site U1336), reflecting the relative age, depth, and latitudinal position of the sites. LSRs frequently show high rates >20 m/m.y in the east (Sites U1334–U1338) but rarely exceed 15 m/m.y. in the western sites. The LSRs of Sites 1218 and 1219 are also reflected at Site U1334 but show slightly lower values during the early Miocene (15–20 m.y.). By combining the available data from Legs 138 and 199 and Expeditions 320 and 321, we will obtain a continuous history of sedimentation rates in the equatorial Pacific region for the past 56 m.y.

Middle Eocene Climatic Optimum

A complete downhole transition of the MECO event was recovered at Sites U1331–U1333 (Bohaty and Zachos, 2003; Bohaty et al., 2009) (Fig. F31). Based on bio- and magnetostratigraphic datums, the MECO event (40–41 Ma) occurs between magnetic Chrons C18n.1n and C18r and falls into the radiolarian Zone RP15 to lowermost RP16. Bohaty et al. (2009) revised the position of peak middle Eocene warming to 40.0 Ma in Chron C18n.2n. At Site U1333 the lithostratigraphy of the MECO event is characterized by an alternating sequence of nannofossil ooze and radiolarian nannofossil ooze interrupted by an interval of radiolarian clay as thick as 4.2 m (Fig. F31). The MECO event at Site U1332 is marked by an alternating sequence of nannofossil ooze, radiolarian nannofossil ooze, radiolarian ooze, and clayey radiolarian ooze. Site U1331 lithostratigraphy also shows an alternating sequence of nannofossil radiolarian ooze, radiolarian nannofossil ooze, and nannofossil ooze. However, at Site U1331 the interval is interrupted by coarse-grained gravity flows. These lithostratigraphic results for the MECO event are similar to those obtained from ODP Sites 1218–1221; at Site 1222 the interval is dominated by clay (Lyle, Wilson, Janecek, et al., 2002).

Eocene–Oligocene transition

Lithology

An Eocene–Oligocene transition was recovered at four sites drilled during Expedition 320 (Sites U1331–U1334). The Eocene/Oligocene boundary as formally defined cannot be identified at these sites because of the absence of the planktonic foraminifer biostratigraphic marker Hantkenina. Magnetostratigraphy from APC-cored intervals and biostratigraphy (radiolarians and nannofossils) provide excellent age control, however, with the Eocene/Oligocene boundary falling just below the Magnetochron C13n/C13r reversal, near the middle of nannofossil Biozone NP21, and just above the radiolarian Biozone RP20/RP19 boundary. (Fig. F32).

Sites U1331–U1334 capture the lithostratigraphy of the Eocene–Oligocene transition in the equatorial Pacific Ocean in a depth transect from ~3600 to 4300 m paleowater depth (~34 Ma) (Fig. F32). At each site, a downhole transition takes place from white to pale brown nannofossil ooze of earliest Oligocene age to much darker brown sediments of Eocene age. At the deep end of the depth transect where the Eocene–Oligocene transition is thinnest (Site U1331), the transition is sharp (over a ~5 cm thick interval) from carbonates into homogeneous dark brown clayey radiolarian ooze. At the shallow end of the transect (Site U1334), where the correlative section is much more expanded, it is less sharp and takes place through dark clayey nannofossil chalk to alternations of dark nannofossil chalk and even darker clayey nannofossil chalk. At Site U1332 the lithologic transition is through radiolarian nannofossil ooze to radiolarian ooze with clay; at Site U1333 the transition is through radiolarian ooze to alternations of radiolarian nannofossil ooze with clay and clayey radiolarian ooze. At Sites U1332–U1334 prominent ~50–100 cm thick beds of particularly dark clays or radiolarian clays are noticeable (Fig. F32). At all sites, the line-scan core images reveal stepwise downhole transitions in sediment color (Fig. F32). Associated pronounced downhole stepwise increases occur in magnetic susceptibility, a*, and b*, together with pronounced downhole deceases in GRA bulk density, L*, and CaCO₃ content (see "Lithostratigraphy" in the "Site U1331" chapter, "Lithostratigraphy" in the "Site U1332" chapter, "Lithostratigraphy" in the "Site U1333" chapter, and "Lithostratigraphy" in the "Site U1334" chapter).

The lithostratigraphy of the Eocene–Oligocene transition from Expedition 320 sites is remarkably consistent with both the expedition rationale for drilling these sites and Leg 199 results and will allow the study of the early history of Cenozoic glaciation and CCD behavior (see Coxall et al., 2005; Pälike et al., 2006b) across a depth transect. Two major lithostratigraphic results from the Eocene–Oligocene transition from Expedition 320 are unexpected and also demand evaluation:

1. The discovery that the section at Site U1333 (paleowater depth = ~34 Ma; ~4000 m) is more carbonate rich than Site 1218 (paleowater depth = ~34 Ma; ~3850 m), which perhaps reflects a paleolatitudinal signal given that this site is at the northern edge of the latitudinal transect of PEAT sites during the Eocene–Oligocene; and

2. Recovery in the most expanded section (Site U1334) of darker and more clay rich and yet more calcareous lithologies than at either Site U1333 or Site 1218 (Figs. F32, F33).

The datums throughout a major part of the Cenozoic section provided by the Nigrini et al. (2006) radiolarian calibration offsets any taxonomic, abundance, or preservation problems that were encountered with a few of the species. One interval, however, has proven to be particularly troublesome: the Eocene/Oligocene boundary. This interval not only shows substantial turnover in the radiolarian fauna (Fig. F34) (Funakawa et al., 2006), it is often represented by a hiatus and is associated with reworked older Eocene radiolarians being deposited in uppermost Eocene and lower Oligocene sediments. Thus, the upper appearance limit of many Eocene species is problematic. A further complication in establishing the true age of first and last appearances of these lower Oligocene and upper Eocene species arises from the impact of missing sections on establishing a paleomagnetic stratigraphy. With part of the section missing, chron boundaries can be truncated, giving an inaccurate estimate of the age of the sediment marking that boundary. Finally, the sharp change in the CCD across the Eocene/Oligocene boundary often makes it difficult to correlate the section recovered at one site to that of another site in the same region (Fig. F35). Of all the means of making regional correlations across the Eocene/Oligocene boundary, paleomagnetic chron boundaries appear to be the most reliable (Fig. F35).

Through the many studies on the material collected during Leg 199, we have come to appreciate more fully the true stratigraphic nature of the Eocene/Oligocene boundary. From the study and comparison of these sections we have been able to identify those sites that appear to have the most complete record across this boundary. From Leg 199, Site 1218 appears to provide the most complete stratigraphic record. From this expedition, a complete boundary section appears to have been recovered at Sites U1333 and U1334; however, only Site U1333 has paleomagnetic control.

In an effort to make a detailed correlation of these three sites, we compared their magnetic susceptibility records and, where available, the paleomagnetic stratigraphy (Fig. F35). This comparison is revealing in many ways. Although Site 1218 and U1334 magnetic susceptibility records look quite similar superficially, when compared in detail they are substantially different. The two-step change in magnetic susceptibility (as well as in other geochemical variables) at the base of Chron C13n is common to Sites 1218 and U1334, as well as to Site U1333. This is one of the first indications that the Eocene/Oligocene boundary section at a site is relatively complete (e.g., Coxall et al., 2005). By simple "peak counting," the maximum in magnetic susceptibility at the top of Chron 15n is also fairly easily identifiable at Site U1334 (as well as at Site U1333). However, the broad, major peak in magnetic susceptibility seen at Site 1218 is only partially represented at Site U1334, with only the younger part of this broad maximum seen at Site U1334. It is not seen at all at Site U1333. A sharp minimum in magnetic susceptibility just below the top of Chron C17n.1n at Site 1218 can also be seen, along with similar minima at Sites U1334 and U1333. Finally, the minimum in magnetic susceptibility just below the broad maximum near the top of Chron 18n.1n is associated with the first appearance of Calocyclas turris at Site 1218, as well as at Sites U1334 and U1333.

The magnetic susceptibility record from Site U1333 looks very dissimilar to those of the two sites drilled on younger basement (Fig. F35). It looks more similar to those records from Sites 1219 and 1220 that were drilled on 56 Ma crust (Fig. F36). These differences and similarities exist even though Site U1334 has fairly good preservation of calcium carbonate in the upper Eocene section, whereas at Site 1218 carbonate preservation is relatively poor, and at Sites 1219, 1220, and U1333 carbonate is only occasionally present. Thus, it appears that the high-amplitude excursions in magnetic susceptibility records near the top of Chron 16n.1n are found only in sections deposited on crust that is only a few million years older than this chron (~35.5 Ma).

One would hope that the biostratigraphic record of the Eocene/Oligocene boundary region at the more complete sites would reveal a consistent record of faunal turnover. This may be true for the first appearances of species in the lower Oligocene; however, the position of the last occurrences of Eocene species within the uppermost Eocene seem to indicate that the reworking of older fossils into younger sections is common, even if there are no apparent breaks in the sections. For example, the position of the last occurrence of the radiolarian Cryptocarpium azyx at Site 1218 is considerably higher in the section than at Sites U1333 and U1334 (Fig. F35). The first appearance of species in these sections is somewhat more reliable (Fig. F35), but it is the very nature of such large faunal turnovers as seen near the Eocene/Oligocene boundary that most of the biostratigraphic datums involve extinctions. Only a very detailed sampling of the section and a more quantitative analysis of the fauna can reveal when the presence of a species in a sample is likely to indicate reworking of older sediments.

Just as the rapid extinction of many Eocene species is a dramatic illustration of the impact of climate change on the planktonic fauna, the recovery from this event is also of interest. A few radiolarian species survived the transition from the warm Eocene to the cooler Oligocene (Funakawa et al., 2006), but the rapid appearance of new species did not occur until several hundred thousand years after the Eocene/Oligocene boundary at 33.7 Ma. Two lower Oligocene radiolarian marker species (Lithocyclia crux and Theocyrtis tuberosa) first appear near 33.4 Ma (recalibrated age), followed by Dorcadospyris pseudopapilio, Dorcadospyris quadripes, and Centroboytris gravida near 33.0 Ma (recalibrated age) (Fig. F34). Abundant diatoms are found in the coarse fraction slides alongside the radiolarians starting in the uppermost Eocene at the same level as the first (older) step in the magnetic susceptibility record (Fig. F34), and the species makeup of this flora changes rapidly upsection. Thus, it appears that there is an increase in the productivity of larger diatoms associated with the Eocene–Oligocene transition (Fig. F21) and that the changeover of the species makeup of the flora occurs quickly in the diatoms relative to the radiolarians. Radiolarians allow the correlation of Eocene/Oligocene boundary sections between different sites without relying directly on the lithostratigraphy (Fig. F37).

Oligocene–Miocene transition

At the end of the Oligocene a significant multimillion year–long rise in the oxygen isotope record (Lear et al., 2004) is closely followed by a relatively short, sharp increase in oxygen isotope values that has been interpreted as a major glacial episode (Mi-1) (Zachos et al., 1997, 2001a, 2001b; Pälike et al., 2006a, 2006b) and correlated to a pronounced drop in sea level (Miller et al., 1991). This event is very close to the Oligocene/Miocene boundary and has now been astronomically age calibrated in several ocean basins (Shackleton et al., 2000; Billups et al., 2004; Pälike et al., 2006a). Although there are clear periodic isotopic signals indicating major changes in ice volume, ocean temperatures, and/or ocean structure, this biostratigraphic boundary has always been somewhat of an enigma. Unlike the major changes in the isotopic stratigraphy, the biostratigraphies of the planktonic microfossils show very little change across this boundary. In fact, it is one of the most difficult epoch boundaries to pick using solely microfossil biostratigraphies.

At Sites 1218 and 1219 this interval was well recovered; however, carbonate preservation still presented a problem for foraminifer stratigraphy. Both sites were deep and well within the lysocline, making the application of temperature proxies such as Mg/Ca ratios in foraminifer tests more difficult (Lear et al., 2008). At the time Miocene–Oligocene sediments were deposited, Site 1218 already resided on 18 Ma old crust and was ~4.1 km deep. Site 1219 was on ~34 Ma crust and was ~4.5 km deep (Lyle, Wilson, Janecek et al., 2002). A relative increase in large diatoms near this boundary in the siliceous coarse fraction suggests increased productivity; however, detailed, high-resolution flux rates across this interval have yet to be determined.

Complete sequences to the biozone and magnetochron level of the Oligocene–Miocene transition were recovered at Sites U1332–U1336 and at the very base of Site U1337 (Fig. F38), providing an excellent integrated stratigraphy. Sites U1332–U1334 display unambiguous magnetostratigraphy coherent with biostratigraphy and a distinct record of alternations in sediment constituents and physical properties. Because of Fe reduction, late Oligocene and early Miocene sediments from Sites U1335 and U1336 do not retain a sufficiently strong remanent intensity to allow retrieval of a reliable shipboard magnetostratigraphy, but good biostratigraphic control is available. The northwesternmost Site U1331 (Fig. F1) does not record the Oligocene–Miocene transition. The Oligocene/Miocene boundary is approximated by the first occurrence of the planktonic foraminifer Paragloborotalia kugleri (23.0 Ma) and the total range of the short-lived (~100 k.y.) calcareous nannofossil Sphenolithus delphix (23.1–23.2 Ma) just below Chron C6Cn.2n. The Oligocene–Miocene transition in Expedition 320 sediments is characterized by alternations of nannofossil- and radiolarian/clay-dominated intervals upsection of the Oligocene/Miocene boundary at Sites U1332 and U1333 and by subtle light–dark color alternations of nannofossil ooze at Sites U1334, U1336 (Fig. F38), and U1337.

Miocene

Carbonate-bearing Miocene sediments were recovered primarily at Sites U1335–U1338. At Site U1335, U1337, and U1338, the early and middle Miocene sections are expanded and sedimentation rates are as high as 20–30 m/m.y., allowing us to achieve our Miocene specific objectives (see "Latest Oligocene–earliest Miocene [Site U1335; 26 Ma crust]," "Miocene [Site U1337; 24 Ma crust]," and "Early and middle Miocene [Site U1338; ~18 Ma crust]"). The Miocene and younger periods form the prime focus for Expedition 321.

Neogene biogenic events and cycles

The Neogene saw significant changes in temperature, glaciation, plankton community, and carbon cycle, including a significant warm interval at ~15–16 Ma, and multiple glaciations first in Antarctica and then in the Northern Hemisphere (Zachos et al., 2001a). There were also important changes in the CCD. Changes in isotope composition and SST proxies can only be monitored by shore-based geochemical studies. However, shipboard monitoring of changes in plankton composition and relative abundance indicate that dissolution has taken place in the Neogene equatorial Pacific (Figs. F17, F19), achieving parts of Objectives 2, 3, 4, and 7 listed in "Scientific objectives."

The eastern equatorial Pacific has times of known change in CCD at ~18 and ~10 Ma (Lyle et al., 1995; Lyle, 2003). There are also important intervals where biogenic fluxes have changed significantly (Farrell et al., 1995) and important time intervals of high diatom flux (Kemp and Baldauf, 1993). We were able to identify these intervals at Sites U1337 and U1338 (Fig. F23) and expect that further work will identify new intervals, thereby achieving Objective 1 listed in "Scientific objectives."

Equatorial productivity and redox-related color changes

White nannofossil oozes are the backdrop for the vivid color changes observed midsection at Sites U1334–U1336 (Oligocene–Miocene but independent of section age). Sediment color shifts from brown/very pale brown to light greenish gray and white. At each of the three sites, the shift in color is illustrated by a steplike drop in b* reflectance (yellow-blue) and a near complete loss of magnetic susceptibility (Fig. F39). Dissolved Fe concentrations in pore fluids increase at least sixfold in the zone of greenish gray sediments. Dissolved Mn concentrations increase at least fivefold in peaks just shallower than the dissolved Fe peaks, consistent with remobilization of Mn under less reducing suboxic diagenesis. The link between sediment color and suboxic diagenesis is clearest at Site U1335, where the light greenish gray color is interrupted by a small interval of very pale brown before returning to greenish gray again (~180 m CSF; lower Miocene). That very pale brown interval corresponds to a pronounced dip in dissolved Fe and a small increase in dissolved Mn. Together with the reduced remanent intensity and magnetic susceptibility, increases in dissolved Fe concentrations and changes in sediment color indicate intensified microbial Fe reduction, perhaps fueled by higher organic carbon accumulation rates across this interval. The intensification of suboxia at depth is largely controlled by site location with respect to the core of the equatorial upwelling system (Figs. F6, F39B). The greenish gray coloration is restricted to the time interval when each site was located south of 3°N. This pattern of geographic control on organic matter deposition and sediment diagenesis is supported by sediment color change observations in three additional DSDP sites (78, 79, and 574) (Fig. F39B). Further postcruise research will establish to what extent geochemical proxies of organic matter productivity, burial, and degradation contribute to the observed patterns, following Olivarez Lyle and Lyle (2005).

While Expedition 320 scientific party noted a relationship among magnetic susceptibility, brown-green color change, and Mn and Fe contents in interstitial waters at Sites U1334–U1336, the results are not as clear-cut at Sites U1337 and U1338 (Fig. F39). All sites have significant dissolved manganese peaks in the upper sediment column. The low-latitude younger sites (U1335, U1337, and U1338) have the Mn peak very near the sediment/water interface, whereas the dissolved manganese peak in the older, more northerly sites is found tens of meters into the sediment column. The dissolved Mn peaks in Site U1337 and U1338 sediments are about a factor of 30 higher than those found in the other sites, however, perhaps because of higher water-column scavenging of Mn to the sediments. The dissolved Fe peak does not clearly match the green intervals, especially at Site U1338. Although not plotted on the paleolatitude graph, green sediments at Sites U1337 and U1338 are associated with intervals when they were near or within the equatorial zone, showing an apparent sedimentary "memory" of the equatorial productivity zone.

Geochemistry

Bulk sediment geochemistry: sedimentary organic carbon and carbonate burial

A high-priority objective of the PEAT program is to reveal the history of CCD fluctuations during the Cenozoic. The findings resulting from >1000 coulometric carbonate measurements (Figs. F16, F40) are described in "Cenozoic CCD in the equatorial Pacific" (Fig. F13). Calcium carbonate, inorganic carbon (IC), and total carbon (TC) concentrations were determined on sediment samples from one hole at every site.

We measured TC and total organic carbon (TOC) at a similar sample resolution and determined very low TOC concentrations, as previously found during Leg 199 (Lyle, Wilson, Janecek, et al., 2002; Olivarez Lyle and Lyle, 2005). TOC concentrations were determined separately by a difference method and by an acidification method for Site U1331. However, we concluded that the TOC concentrations determined by the normal difference method were overestimates in high-weight percent CaCO₃ and very low weight percent TOC sediments because they were determined as a small difference between two numbers comparable in magnitude. Therefore, TOC analyses were performed only by the acidification method, in which TOC was determined by using carbonate-free sediments after treatment by acidification for the remaining sites. Using this acidification technique, we reduced the detection limit for TOC measurements to 0.03 wt%. TOC concentrations in sediments determined by this method are very low throughout the sediment column and near or below the detection limit for samples from Sites U1331–U1335 (Figs. F40, F41). TOC concentrations tend to be slightly higher at those depths where CaCO₃ concentrations are low at Sites U1332–U1335. The maximum TOC value determined is 0.18 wt% in surface sediments from Site U1332. Despite the very low TOC values across the PEAT sediments recovered, postcruise research will be able to measure biomarkers and alkenones from some of the more organic rich sediments, addressing Objectives 2, 4, and 9 (see "Scientific objectives"). For Sites U1337 and U1338, TOC increases from low levels at the base of the sedimentary section to levels ~0.2 wt% in the youngest sediments upcore. Postcruise research will focus upon measuring alkenones, other biomarkers, and productivity proxies to better understand the variations in TOC and productivity (Objectives 2, 4, and 9 in "Scientific objectives").

Figure F16 shows the TOC and carbonate profiles versus age at Sites U1336–U1338. None of these sites were ever below the CCD, although Site U1337 briefly approached zero carbonate levels during the 9–11 Ma carbonate crash interval. We can date other carbonate minima between 16 and 18 Ma and 3 and 4 Ma. The 16–18 Ma minimum correlates with the early Miocene carbonate flux minimum identified by Lyle (2003) but is much better defined at Site U1337. It reaches its minimum at 17.2 Ma. The 3–4 Ma carbonate minimum is not well identified in Leg 138 sediments, perhaps because these sites are on younger, shallower crust.

TOC profiles identify a trend from lower to higher TOC from the base of the cores and small intervals of high TOC that were not resolved well at the very coarse shipboard sample interval. High TOC was found in high-biosiliceous intervals at levels as high as 0.7 wt%. The distribution of high TOC deserves more study postcruise to determine how these short events have affected TOC burial fluxes. The base of each site had TOC contents of ~0.1 wt% or below, whereas younger sediments have values of 0.2 wt% or higher. Site U1336, which has a hiatus at 12 Ma, has low TOC values to the surface of the core. Earlier drilling did not find trends in TOC because the standard technique to measure TOC on ODP and DSDP legs (assigning the difference between total carbon and carbonate carbon to TOC) has very high errors in carbonate-rich sediments.

The longer TOC trend could be the result of age and long-term degradation or the strength of early diagenetic degradation, or it could result from changes in TOC particulate rain ("productivity") through time. Distinguishing between these alternatives will require postcruise study of mass accumulation rates and abundances of different biomarkers.

Interstitial water profiles: Sr and Li

Interstitial water geochemistry profiles of the different sites drilled during Expedition 320 show considerable differences in respect to dissolved Sr²⁺ concentrations (Fig. F42). Whereas Sites U1331 and U1332 show little variability with depth, Sites U1333–U1338 reveal increasing variability of Sr²⁺ concentrations. At Sites U1331 and U1332, Sr²⁺ shows mainly concentrations around seawater values; Sr²⁺ increases at Sites U1333 and U1334 up to ~110 µM, at Site U1335 up to 250 µM, Site U1336 up to 430 µM, and Site U1338 up to 400 µm. Sites U1335 and U1338 are characterized by a pronounced increase in Sr²⁺ with depth followed by a strong decrease toward basement to seawater-like concentrations. A similar pattern is revealed for Sites U1333 and U1334, but it is considerably less pronounced and developed. This pattern indicates the influence of carbonate diagenesis and recrystallization, releasing Sr²⁺ to the pore fluid at intermediate depth, and the flow of relatively unaltered seawater through basement and diffusion between end-members. Limited variability of Sr²⁺ at Sites U1331 and U1332 might be related to the relatively thin sediment thickness preventing the establishment of large gradients. At Site U1337 a rapid change in Sr²⁺ occurred at the "baby chert" layer, showing that it acts as a diffusional barrier in the sediments.

Similar to Sr²⁺, Li⁺ shows considerable differences between the different sites of Expedition 320, with the least variability at Sites U1331–U1333 and increasing depletion of Li⁺ in the pore fluid at Sites U1334 (down to 15 µM), U1336 (down to 7 µM), and U1335 and U1338 (down to 4 µM) at intermediate depth (Fig. F43). Near basement, Li⁺ increases again toward seawater-like values. The Site U1337 Li⁺ profile also shows a major offset at the baby chert layer, like Sr²⁺. Profiles of Li⁺ in the pore fluid indicate diagenetic reactions in the sediments consuming Li, possibly low-temperature clay alteration. Li⁺ concentrations similar to seawater values are compatible with observations of Sr²⁺, suggesting the flow of relatively unaltered seawater through the oceanic crust.

Seismic stratigraphy and integration, physical properties, site correlation and orbital cycles

The equatorial Pacific is a classic "binary" sediment system, with variable amounts of biogenic calcium carbonate and biosiliceous sediment components but very little clay. It is also well known that carbonate contents of equatorial Pacific sediments can be estimated from the bulk density because the carbonates have lower porosity and higher grain density than biosiliceous sediments (Mayer, 1991). Consequently, physical property records contain meter-scale cyclicity that will ultimately be useful for orbital-tuning timescales (Objective 3 in "Scientific objectives").

On longer depth scales, physical property records are useful to correlate among sites. Figure F28 shows the two high-resolution density logs at Sites U1337 and U1338 and illustrates the high degree of correlation between the two sites that are separated by ~600 km. Lines on the figure represent equivalent nannofossil biostratigraphic datum levels. The two records are aligned such that the top peaks line up. The depth scale for Site U1338 is compressed so that the bottom event also lines up. Density variations can easily be correlated between the sites even though they are separated by ~600 km.

Mayer et al. (1985) developed a seismic stratigraphy for the central Pacific at Site 574 during Leg 85. They noted that major seismic horizons were caused by density variations associated with low-carbonate intervals. They proposed that these intervals were chronostratigraphic because they were caused by paleoceanographic changes in deposition and/or dissolution of calcium carbonate.

Unfortunately, Mayer et al. (1985) did not have logs to measure in situ velocities. One of the important experiments of Expedition 321 was to use logging as well as cores to check this conclusion and to better constrain the age of the equatorial Pacific seismic horizons. For this reason, we planned a vertical seismic profile (VSP) experiment at one of the Neogene PEAT sites (Objective 6 in "Scientific objectives"). We were able to run the VSP log at both Site U1337 (Fig. F44) and U1338. Figure F44 is an initial comparison between the Site 574 seismic stratigraphy of Mayer et al. (1985) and the initial results for Site U1337. The events appear to correlate in age, as would be predicted by Mayer et al. (1985). Site 574 is at essentially the same latitude as Site U1337 but >1000 km to the west. The length scale of the correlatable seismic horizons across the Pacific helps to define the length scale of the paleoceanographic events with which they are associated. Postcruise studies will focus on better defining the seismic stratigraphy at both Sites U1337 and U1338, allowing new tie points for the seismic stratigraphic study of the equatorial Pacific sediment bulge (Mitchell et al., 2003).

Objective 6 of the PEAT program (see "Scientific objectives") is to establish the age and lithologic origin of the seismic reflections previously identified in the eastern equatorial Pacific and make use of the high level of correlation between tropical sedimentary sections and existing seismic stratigraphy to develop a more complete model of equatorial circulation and sedimentation (Lyle et al., 2002, 2006). We achieved this objective by calculating synthetic seismograms made from bulk density and sonic velocity data from core material and downhole logs for the Paleogene Site U1331 (Fig. F45).

The known depths and two-way traveltimes (TWTs) to the seafloor and basement provided an initial depth to TWT model and average sonic velocity for the sediment cover. The depth to TWT model was then adjusted to bring the reflectors in the synthetic seismogram into line with the corresponding reflectors in the seismic section. Typically, only a small number of adjustments were needed to give a good match.

Downhole logging data at two sites discriminate between radiolarian and nannofossil ooze using density, conductivity, and magnetic susceptibility, which are more useful in understanding the lithology in the poor recovery intervals composed of chert and porcellanite. In addition, a hard lithified limestone altered by hydrothermal activity occurred 10–20 m above basement at most drilling sites. Correlation between seismic profiles and drilling is important in understanding an estimation of basement age and thickness of sedimentary sequences. At Site U1335 the thickness is was originally estimated to be 361 m, but 420 m of sediment was recovered above the basement. Thus, our results contribute to the drilling strategy for future drilling in the equatorial Pacific.

MORB age transect

An age transect of sediments on top of basaltic basement in the equatorial Pacific region was recovered during Expedition 320/321. Basement ages decrease eastward from Site U1333 to Site U1338 (Fig. F20). Basement basalts are covered by hard lithified limestones spanning from early Eocene through early Miocene age except at Site U1332, where zeolitic clays were recovered above basalt. These clays yielded poorly preserved calcareous nannofossils, and the position of the clays may be due to hydrothermal alteration. At Site U1331, however, a typical calcareous ooze with high carbonate contents containing abundant planktonic foraminifers and calcareous nannofossils was recovered as observed at Site 1221. Basement basalts are highly altered with a spherulitic texture, whereas ferromagnesian minerals (mainly clinopyroxene) are replaced by chlorite. Thin section analysis indicates a sparsely phyric basalt mainly composed of phenocryst plagioclase.

Chert and porcellanite occurrence

Sites U1331–U1333 and U1336 recovered chert and porcellanite (Fig. F46). Cherts are characterized by their hardness and highly silicified matrix in which sediments are cemented with microcrystalline quartz. Primary pore spaces (e.g., within chambers of foraminifers and radiolarians) are sometimes filled with chalcedony. Porcellanites are silicified to a lesser extent and are richer in clay minerals than the cherts. Porcellanite is much more abundant than chert at Sites U1331–U1333, which is interbedded with radiolarian ooze and nannofossil ooze in the early middle Eocene interval at Sites U1332 and U1333 (Fig. F46). In stratigraphic intervals from the early Eocene through earliest middle Eocene at Sites U1331 and U1332, the presence and original structures of porcellanite are not clear because of poor core recovery. Site U1336 contains mostly chert. Porcellanite-bearing intervals at Sites U1331–U1333 correspond to a time interval between early Eocene and early middle Eocene (~42 Ma), roughly coincident with the chert-rich intervals at Sites 1220–1222 recovered during Leg 199 (Fig. F46). At Sites U1331 and U1332, porcellanite layers are associated with thin clay horizons interbedded with radiolarian ooze. Oligocene chert layers from Site U1336 are interbedded with nannofossil ooze and chalk and show various colors such as greenish gray, dark gray, pink, and black and contain abundant foraminifers that are occasionally replaced with microcrystalline quartz and pyrite.

At the base of the sections recovered there was always an interval in which all biogenic silica had been dissolved and which was barren of radiolarians. This "silica free zone" (SFZ) is usually not thick—only 7 to 16 m at most of the sites drilled (Table T3). This is well within the usual 1–40 m thickness of the SFZ described by Moore (2008a) for Pacific open-ocean sections recovered by scientific ocean drilling. Moore (2008a) associated this dissolution zone with the circulation of hydrothermal waters in the upper oceanic crust and related the silica removed by this process to the ultimate formation of cherts. Thus, Site U1336 stands out as very unusual. It has a SFZ almost 130 m thick from ~170 m CSF to the base of the drilled section at 298 m CSF, with abundant chert stringers. The extensive alteration suggests that hydrothermal waters may actually invade well up into the section near this site, possibly along more permeable, small-offset faults.

Within the Neogene sediments recovered, chert/porcellanite intervals were found at Sites U1336–U1338. At Site U1336, a chert interval was found in unlithified early Miocene sediments at ~130 m CSF. The section below 200 m CSF also had significant numbers of chert intervals and the carbonates had turned to limestone, suggesting significant higher temperature diagenesis.

At Sites U1337 and U1338, only one cherty interval was found—what we referred to as the baby chert interval. At both sites the cherty interval hampered core recovery around it, but it was easily penetrated by drilling. Interestingly, its position appears to be chronostratigraphic, at ~11 Ma in the late middle Miocene (~240 m CSF at Site U1337 and ~282 m CSF at Site U1338). This baby chert interval was located in a diatom mat interval at Site U1337 and a diatom-rich interval at Site U1338. The interval was well-imaged by Formation MicroScanner (FMS) logging and shows that the chert interval is 40 cm thick at Site U1337 and about one-quarter of the way up from the base of the low-resistivity diatom-rich interval (Fig. F47). At Site U1338 the chert interval was only 16 cm thick but also about one-quarter of the way through the low-resistivity interval from the base upward. The logs allowed us to estimate the amount of disturbed and missing section in the cored sediments—there is only a small loss of material (<1 m) around the chert.

At Site U1337 it was clear from interstitial water profiles that the chert interval is a barrier to diffusion. There were offsets in Li, Sr, and SO₄ in interstitial waters immediately below and above the baby chert interval (Fig. F43, F45). It is unclear why the chert formed in this interval alone at both sites because there was nothing unique about it. There were additional diatom-rich intervals below the baby chert that were unaffected. The more extensive chert layer at Site U1337 is actually formed ~30 m shallower (250 m CSF) than the interval at Site U1338 (280 m CSF). Postcruise research focused on diagenesis and fluid flow may provide insight into chert formation within these diatom-rich intervals.

Heat flow measurements

Geothermal gradient and heat flow values were determined at Sites U1331–U1338 from in situ temperature measurements made using the advanced piston corer temperature tool (APCT-3) and the sediment temperature tool (SET). At least four APCT-3 deployments were made per site to a maximum depth of 316 m. Bottom water temperature was also determined for each site, with values ranging from 1.44° to 1.65°C. Linear geothermal gradients obtained from these data ranged widely from 7.5°C/km at Site U1335 to 75.0°C/km at Site U1332.

Heat flow was calculated according to the Bullard method, to be consistent with the Leg 199 analyses and the synthesis of ODP heat flow data by Pribnow et al. (2000). Similar to the geothermal gradients, Site U1335 had the lowest heat flow, 7.0 mW/m², and Site U1332 the highest, 70.7 mW/m² (Fig. F48). These heat flows are within the range of values in the eastern equatorial Pacific heat flow data set maintained by the International Heat Flow Commission (Pollack et al., 1993).

The wide range of values emphasizes the possibility that local crustal hydrothermal circulation is strongly influencing heat flow values. Results from the PEAT expeditions will also provide heat flow information in an area of the Pacific Ocean that has only a sparse existing heat flow data set.

Gravity flow deposits

Throughout the sedimentary section drilled at Sites U1331 and U1335, sharp irregular contacts are present between lithologies (Fig. F49) (see "Lithostratigraphy" in the "Site U1331" chapter and "Lithostratigraphy" in the "Site U1335" chapter). Many are recognizable by distinct changes in color. Sharp contacts, occasionally associated with an erosional basis, are often overlain by coarser grained, more carbonate rich (including planktonic and benthic foraminifers), and/or opaque-coated sediments than those below the contact. At Site U1335 angular basalt fragments, pyritized foraminifers and/or radiolarians, and fish teeth are often present at the base of the gravity flows. The overlying sediments fine uphole and in some cases show cross or parallel laminations in the middle of the bed. These features indicate that the erosional contacts and their overlying coarse sediments at both Sites U1331 and U1335 are the product of mass flow events, typically turbidity currents. Commonly, turbidite thickness is between 2 and 25 cm, with the maximum thickness of 176 cm found at Site U1335. The total thickness of the identified turbidites at Site U1335 occupies at least 2% of the recovered sediment. The provenance of the inferred turbidites observed at Site U1331 is unknown, but their typically calcareous composition points to a source that lay above the CCD at the time that the reworked sediments were originally deposited, possibly a seamount lying a few kilometers south of the drilled location. A similar provenance of turbidites is proposed for Site U1335, where two seamounts lie ~15–20 km northeast and southeast of the drilled location.

New capabilities on the JOIDES Resolution

The shipboard laboratories have new capabilities that greatly improve the ability of shipboard scientists and later researchers to study the sediment column and that complement excellent core recovery achieved during the PEAT expeditions (Fig. F50). One important capability is a much better capacity to integrate line-scan core images into descriptions and analyses. Figure F51 shows color images from different holes at Site U1337 and how they were joined into the Site U1337 splice. The color images are particularly useful to resolve color banding and cyclic behavior in the multicolored sediment intervals. In the upper parts of all the sites, it is common to find oxidized brown sediments that become green when Fe within the sediments is reduced (Lyle, 1983). Typically, this feature has been described in the upper meter of the sediment column. Further west, where the PEAT sites are located, brown colors extend downward for tens of meters. Not all of the color information is stratigraphically useful—redox state and other diagenetic processes can overprint other sediment information. For example, the color change from green to yellow at the base of the Site U1338 sedimentary section cuts across stratigraphic correlations between different holes.

Dark–light color banding, whatever shade of color, is indicative of carbonate content in the equatorial Pacific. Darker layers usually have lower carbonate contents, higher magnetic susceptibility, higher natural gamma activity, and lower bulk density. Different scales of carbonate cycles are found in the core images. Because many of the low-carbonate events are chronostratigraphic, they can be used to correlate between cores. At the larger scale, the dark–light color banding can be used to correlate between sites.

Another new instrument on the JOIDES Resolution is the Natural Gamma Radiation Logger (NGRL). The instrument is much more sensitive than the ODP NGR and can be used as part of the normal core-flow measurements. Biogenic sediments of the equatorial Pacific are not ideal for natural gamma activity studies—they only have low levels of natural radioactivity. Nevertheless, the NGRL was able to reliably detect natural gamma ray activity variations of <1 counts per second (cps). Small natural gamma activity highs correlated with dark sediment layers. One of the objectives for the natural gamma ray measurement on cores is to produce a record that can be correlated with the wireline logging measurements. Unfortunately, the low natural gamma ray activity levels resulted in only a few peaks that could be consistently correlated between core and logs.

The seafloor was one interval that consistently had high natural gamma ray activity (Fig. F52). Whereas counts at a distance from the seafloor were <5 cps, seafloor natural radioactivity levels were as high as 160 cps within 1 m of the top of the sediments. Natural gamma ray activity dropped to the typical low count a few meters below the surface. Spectral studies using hour-long counts on the surface cores revealed that natural gamma ray activity at the surface is being produced by high uranium levels, not thorium or potassium. It is not clear yet why uranium is concentrated only near the sediment/water interface. Because the near-surface sediments have low organic carbon contents, uranium has not been concentrated by reduction and precipitation from seawater. Kunzendorf et al. (1983) instead argues that enrichments of uranium in tropical Pacific sediments result from coprecipitation of seawater uranium with Fe oxides. This solution still begs the question as to why the level of Fe oxides might be higher at the surface of the sediments.

Top of page   |   Previous   |   Next