METHODS AND MATERIALS

Three ODP sites drilled in the Caribbean during Leg 165 are used in this study. More than 350 m of core from the three sites was analyzed. Exact locations and water depths are listed for each site in Table 1 and illustrated in Figure 3 and Figure 4. Samples measuring 10 cm³ were collected approximately every 50 cm. Care was taken to avoid sampling within turbidites and ash layers. However, samples collected between ash layers can still consist of 5%-10% dispersed ash relative to the bulk sediment (Sigurdsson, Leckie, Acton, et al., 1997).

Site 998, the northernmost site analyzed for this study, is located in a water depth of 3101 m in the Yucatan Basin on the northern flank of the Cayman Rise (Fig. 3). The sections sampled in interval 165-998A-15H-1, 21-23 cm, to 22X-6, 55-57 cm, (132.51-203.15 meters below seafloor [mbsf]) consist of nannofossil ooze with foraminifers and clays, clayey nannofossil mixed sediment interbedded with turbidites and ash layers, clays with nannofossils, nannofossil chalk with clays, and foraminifer chalk with clay (Sigurdsson, Leckie, Acton, et al., 1997). This 71-m-long section from Hole 998A corresponds to an early middle Miocene to late Miocene interval and contains ~32 turbidites adding up to 6.62 m of sediment (~9.5% of the total sediment) and 2.11 m of ash within 17 distinct layers (~3% of the total sediment).

Site 999 is located on the Kogi Rise within the Colombian Basin (Fig. 3). The water depth of this site (2839 m) is almost a kilometer above the surrounding Colombian abyssal plain. It is the most proximal site to the Isthmus of Panama and also the closest to the mouth of the Magdalena River, an important source of fine terrigenous sediment. The 123-m-thick section of early middle to late Miocene-age sediment in Hole 999A (interval 165-999A-28X-1, 18-20 cm, to 38X-CC, 9-11 cm; 231.91-354.17 mbsf), consists of clayey nannofossil sediment with siliceous components, interbedded minor volcanic ash layers, clay with nannofossils, siliceous clayey mixed sediment, and clayey calcareous chalk with foraminifers and nannofossils (Sigurdsson, Leckie, Acton, et al., 1997). Samples below 300 mbsf are extremely indurated. In contrast to Site 998, no turbidites were observed in the middle to late Miocene core interval. Approximately 3.21 m of sediment (~3% of the total core) corresponding to 57 distinct ash layers was observed within this cored interval in Hole 999A (Sigurdsson, Leckie, Acton, et al., 1997).

The water depth of Site 1000, at 927 m, is the shallowest of the three sites included in this study (Fig. 3, Fig. 4). Located within the Pedro Channel, a seaway across the northern Nicaraguan Rise, Site 1000 is adjacent to active carbonate banks that export neritic, bank-derived carbonate sediment into the channel (Glaser and Droxler, 1993; Schwartz, 1996). The pelagic sediment is mixed with lateral influx of neritic carbonate and terrigenous sediments. The 172-m middle to late Miocene core interval of Hole 1000A consists of micritic nannofossil ooze with foraminifers, foraminiferal micritic ooze with nannofossils, and micritic nannofossil chalk with clay and foraminifers (Sigurdsson, Leckie, Acton, et al., 1997). The 0.99 m of combined ash layers and 1.21 m of turbidites (~0.5% and ~0.7% of the core interval, respectively) represent minor lithologies in this interval.

Age/depth models in this study are based on coccolith datums identified by Kameo and Bralower (Chap. 1, this volume). These models are based upon the biostratigraphy of Raffi and Flores (1995) (Fig. 7; Table 2). Figure 8B compares the planktonic foraminifer and nannofossil datums of the three sites. Hole 998A also includes the magnetostratigraphy of Shackleton et al. (1995), where the time scale is based upon astronomical tuning as well as gamma-ray attenuation porosity evaluator density correlation to further refine the work of Cande and Kent (1992). Good agreement between the nannofossil and foraminifer datums is observed. The nannofossil datums, however, suggest more uniform sedimentation rates. Sedimentation rate changes are calculated by linear interpolation between these datums (Fig. 8B).

Each sample was divided onto two aluminum weighing trays. The samples were then dried for at least 48 hr in an oven at 50°C. One portion was weighed, soaked in pH-balanced deionized distilled water, and sieved through a 63-µm screen. The coarse (sand sized) fraction was then dried and reweighed for relative proportions of the coarse fraction to the bulk sediment. Indurated samples were also soaked in a sodium hexametaphosphate solution to aid in disaggregation.

Stable Isotopes

The sand-sized particles were dry sieved through a 250-µm screen and picked for the benthic foraminifer Planulina wuellerstorfi, which records oxygen and carbon isotopes of the ambient waters (Shackleton and Opdyke, 1973; Belanger et al., 1981). Relative preservation of the benthic tests, such as the degree of cementation, was also recorded. The P. wuellerstorfi tests were cleaned by sonication in deionized, distilled water and analyzed in Dr. Howard Spero's stable isotope laboratory at the University of California-Davis. The foraminifers were reacted in a common phosphoric acid bath fed by an autocarbonate device and analyzed in a Fisons Optima mass spectrometer. Values for the mass ratios of ¹⁸O/¹⁶O and ¹³C/¹²C were related to the Peedee belemnite standard and reported as ¹⁸O and ¹³C, respectively, with a precision of 0.08 for ¹⁸O and 0.05 for ¹³C.

Carbonate Content

The second portion of bulk sediment was ground with a mortar and pestle. Half-gram portions of the powder were analyzed in a carbonate bomb for carbonate content (Müller and Gastner, 1971). Within the sealed cylinder, 50% concentrated HCl was reacted with the sample. The increased pressure from the generated carbon dioxide is proportional to the initial amount of carbonate (Droxler et al., 1988) when compared to a 100% pure calcium carbonate standard. The carbonate weight percent was derived using the following equation:

CaCO₃ wt% = [(sample pressure/sample weight) / (standard pressure/standard weight)] × 100.

Standards are run at the beginning and end of a batch of 10 unknowns to ensure consistency. These are the same procedures described in Glaser and Droxler (1993). Carbonate data acquired in the Rice University lab using this procedure were integrated with Leg 165 shipboard data. Inorganic geochemical analyses were conducted during ODP Leg 165 on one sample per section of core (approximately every 150 cm). Carbonate weight percent was obtained on the ship by reacting 10 mg of dried, ground, bulk sample with HCl in a Coulometrics 5011 coulometer (for methods, see the "Explanatory Notes" chapter in Sigurdsson, Leckie, Acton, et al., 1997). No discrepancy between the two methods was observed in the carbonate analysis (Fig. 9). The resolution of the carbonate records was certainly improved by merging both shipboard and postcruise carbonate data sets for Sites 998, 999, and 1000.

Accumulation rates of the bulk sediment are calculated by multiplying a linear sedimentation rate (in meters per million years) by the dry bulk density (grams of dry sediment per wet volume in cubic centimeters):

Dry bulk density × linear sedimentation rate = bulk MAR.

Carbonate MARs (CO₃ MARs) are calculated by multiplying the MAR by the calcium carbonate content:

MAR × CaCO₃ wt% × 100 = CO₃ MAR.

The dry bulk density was directly measured from each core section as routine shipboard analysis during Leg 165. The sedimentation rate was calculated using the biostratigraphic datums of Raffi and Flores (1995) (Fig. 7, Fig. 8). CO₃ MAR, as opposed to carbonate weight percent, reflects what is happening to the carbonate only, rather than how the carbonate changes relative to other constituents. Despite being derived from other measurements, the calculation of MAR can be extremely helpful in discerning changes because of dissolution of carbonates and dilution by noncarbonates.

Because of its location in relatively shallow depths and its proximity to carbonate banks, the ground bulk sediment of Hole 1000A was also used to acquire mineralogical data of two carbonate species present in the core. A Philips-Norelco model 12045 X-ray diffractometer was used to scan the 2- angles from 25.5° to 27.5° and from 28.5° to 32° at a resolution of 0.02° per step. The area above the background radiation and under the aragonite peak (d-spacing = 3.41) and calcite peak (d-spacing = 3.04) was integrated. Greater description of this procedure and its quantitative range can be found in Milliman (1974) and Droxler et al. (1988). Because the aragonite content was reported as a percentage of the total carbonate, an aragonite accumulation rate was calculated by multiplying the aragonite content by the CO₃ MAR.