MATERIALS AND METHODS

Modern Database. The Modern Analog Technique requires access to a large modern database that captures much of the variability present in modern faunas. Our modern Pacific database (Appendix) is a component of the global planktonic coretop database described by Prell (1985) and used by many workers. The Pacific database (Fig. 3) has 499 coretop faunal analyses from the work of Thompson (1976, 1981), Parker and Berger (1971), Coulborn et al. (1980), and some unpublished data. The samples are distributed from 47° N to 64° S and represent winter and summer temperatures between 0°C and 30.3°C. For this study we regrouped the modern faunal data into the same species groups as Dowsett and Poore (1990) and Dowsett (1991) with the following changes. Species with no more than one individual in a sample were deleted from the dataset at the outset. Dowsett and Poore (1990) combined sinistral and dextral coiling varieties of Neogloboquadrina pachyderma with the artificial N. pachyderma - N. dutertrei intergrade category of Kipp (1976) to create a "cold" end member Neogloboquadrina category for the North Atlantic. The modern "warm" Neogloboquadrina end member consists of N. dutertrei which is important in low-latitude and gyre-margin upwelling regions (Dowsett 1991). In this study we chose to leave the N. pachyderma categories and N. dutertrei - N. pachyderma category as distinct elements of the modern fauna. These conventions resulted in 21 counting categories (Table 1).

Sample Preparation, Counting Techniques and Taxonomy. To test the MAT, three Pliocene sequences were selected from the western North Pacific (Fig. 3). These deep-sea cores represent a range of oceanographic conditions and thus provide a good test of the MAT technique.

The samples used in this study were washed using low temperature (isotope) procedures. Sediment samples were dried in an oven at less than or equal to 50°C and weighed. The dried bulk sample was disaggregated in a beaker with warm tap water and about 2 ml of dilute Calgon ™ solution (5 gm Calgon ™ to 1 liter water). The beaker was agitated on a vibrating hot plate without heating. Samples were then washed through a 63 micron sieve using a fine spray hose and dried in an oven at less than or equal to 50° C. Ocean Drilling Program Hole 769B samples required an additional treatment with NaCO₃ added to the wash in order to obtain clean specimens. Weights were then obtained for the fine and coarse fractions of each sample.

A split of 300 to 350 planktonic foraminifer specimens was obtained from the >149 micron size fraction of each sample using a Carpco ™ sample splitter. Specimens were identified, sorted, and fixed to a standard 60-square micropaleontological slide. The taxonomic names used in Tables 2, 3, and 4 are summarized in Polanco and Dowsett (1993) and Dowsett and West (1993). In general, our taxonomic concepts follow Parker (1962, 1967) and Blow (1969). Pliocene census data were retabulated and converted to percent using the same counting categories used in the modern database. Deep Sea Drilling Project and ODP sample designations are abbreviated as core-section, depth within section in centimeters (e.g., 10-5, 34 = core 10, section 5, 34 cm below top of section 5). The depth column lists depth of sample below sea floor in meters. Wherever ages are provided they refer to the Berggren et al. (1985) time scale.

DSDP Site 445, Hole 445. Deep Sea Drilling Project Hole 445 was drilled on the Daito Ridge in the Northern Philippine Sea (25.52°N, 133.20°E) in 3377 meters of water. The upper 150 m of sediment recovered at Hole 445 consists of nannofossil ooze with interbedded foraminifer-nannofossil oozes (Klein et al. 1980). Calcareous microfossils are, over certain intervals, abundant and well preserved at this locality (Echols 1980, Okada 1980). From selected microfossil datums (Echols 1980, Okada 1980), we determined that cores 10 through 12 represent the middle Pliocene. Twenty-three samples between the bottom of Core 8 and the top of Core 13 were chosen for this study. Graphic correlation was used (Fig. 4) to determine the relationship between depth at Hole 445 and the composite standard reference section (CSRS) of Dowsett (1989a,b). Depth in Hole 445 can be converted to composite units (cu) using the equation:

y = 8.48 + 0.874 x, for x > 20 and x < 120 (2)

where x is depth in Hole 445 in meters and y is composite unit position of the CSRS of Dowsett (1989a, b). This line of correlation is well constrained down through Core 10 by the last occurrences of Helicosphaera selli, Calcidiscus macintyrei, Discoaster pentaradiatus, Dentoglobigerina altispira, and Sphaeroidinellopsis spp., and the first occurrence of Globorotalia truncatulinoides. Equation (2) is used between 20 and 120 m sub-bottom. Because the CSRS of Dowsett (1989a, b) exhibits a nearly linear fit to absolute age, CSRS position can be converted to absolute age using the equation:

y = 0.354 + 0.034 x (r² = 0.99) (3)

where y is age in Ma (Berggren et al. 1985) and x is composite unit position of the CSRS. Applying equations (2) and (3) above to the samples chosen for this study indicates an age range of 3.82 Ma to 2.56 Ma.

Faunal census data for Hole 445 is given in Table 2 (Dowsett and West 1993). The fauna is a typical Pliocene subtropical assemblage with significant proportions of Globigerinoides obliquus, Globigerinoides ruber, Globigerinoides sacculifer, Globorotalia crassaformis, Globorotalia menardii, Globigerina woodi, Globigerina incisa, andNeogloboquadrina acostaensis. Large numbers of fragments in some samples are correlated with higher percentages of benthic foraminifers and are suggestive of increased dissolution.

DSDP Site 463, Hole 463. Hole 463 is located in the central North Pacific on the mid-Pacific Mountains at 21.35° N and 174.66°E in 2525 meters of water (Fig. 3). The upper 50 m of Hole 463 consists of highly disturbed, bioturbated, and sometimes soupy nannofossil ooze (Thiede and Vallier et al. 1981). The stratigraphic distribution of foraminifers (Vincent 1981, Polanco and Dowsett 1993) suggests a very low sediment accumulation rate with the Miocene-Pliocene boundary occurring approximately 25 m sub-bottom (top of core 4). Twenty-three samples from cores 1 through 4 were processed and analyzed for planktonic foraminifers. Graphic correlation analysis was performed on these samples using planktonic foraminiferal events from Vincent (1981) and Polanco and Dowsett (1993, see Fig. 5). The distribution of events in Figure 5 indicates a "channel" (Shaw 1964) in which the line of correlation (LOC) can be placed. The low number of events results in a fairly broad channel, and the LOC shown in Figure 5 represents only a general correlation. This LOC can be meaningfully applied between about 19 m and 5 m sub-bottom. Using the equations of the LOC:

y = -20.94 + 10.84 x for x < 14.5 (4)

y = 78.00 + 2.8 x for x > 14.5 (5)

and equation (3) above, absolute ages can be derived for most of the samples in the census data set. This age model shows our sample spacing to be inadequate to capture any of the oceanographic variability known to exist at Milankovitch periodicities during the Pliocene.

Planktonic foraminifer census data for Hole 463 are presented in Table 3 (Polanco and Dowsett 1993). These samples are dominated by Globigerinoides obliquus, Globigerinoides ruber, Globigerinoides sacculifer, Dentoglobigerina altispira, Globigerina woodi, Orbulina universa,and Sphaeroidinellopsis spp. High numbers of fragmented specimens and high percentages of benthic foraminifers in some samples suggest increased dissolution.

ODP Site 769, Hole 769B. Hole 769B (8.78°N, 121.29°E) is located on the southeastern flank of the Cagayan Ridge in the Sulu Sea in 3643 m of water (Fig. 3). The upper 18 cores recovered nearly 170 m of pelagic biogenic carbonate sediment and hemipelagic clays. Planktonic foraminifers are abundant and fair to moderately well-preserved in these samples with preservation diminishing due to increased dissolution downcore.

Rangin et al. (1990) interpreted the paleomagnetic stratigraphy to indicate the Gauss-Matuyama Chron boundary at 114.5 mbsf. The top of the Kaena subchron occurs at 117.0 mbsf and the bottom of the Cochiti occurs at 132.6 mbsf (Fig. 6). We interpolated between these boundaries using the Berggren et al. (1985) time scale to derive ages for 26 samples from cores 13, 14 (mid Gauss) and 18. This age model suggests these samples are more-or-less equally distributed between about 3.4 and 2.3 Ma with the core 18 samples clustering near 5.3 Ma. Planktonic foraminifer biostratigraphy generally supports these age assignments.

Planktonic foraminifer census data for Hole 769B are given in Table 4 (Polanco and Dowsett 1993). The assemblage is typical of tropical Pacific assemblages with high numbers of Globigerinoides, Globorotalia, and warm water Neogloboquadrina species. Fragment data was not generated for Hole 769B, but the high percentages of benthic taxa in many samples with very low numbers of planktonic specimens per 10 cc of raw material are indicative of severe dissolution.