RESULTS

The Modern Analog Technique (MAT) was applied to the planktonic foraminiferal assemblages from Sites 445, 463, and 769 to estimate February and August SST. Tables 5 (5a samples 1-100, 5b samples 101-199, 5c samples 200-299, 5d samples 300-399, 5e sample 400-499), 6 (6a samples 1-100, 6b samples 101-199, 6c samples 200-299, 6d samples 300-399, 6e samples 400-499), and 7 7a samples 1-100, 7b samples 101-199, 7c samples 200-299, 7d samples 300-399, 7e samples 400-499)show the squared chord distance (SCD) between each sample from the Pliocene sequences and the 499 modern coretop samples. Table 8 summarizes the analyses by showing the minimum, maximum, mean and standard deviation of the lowest SCD values for each site. Distance to the closest analogs of Site 445 samples range between 0.19 and 0.51 with a mean of 0.27. The SCD values from sites 463 and 769 have a wider range (0.07 to ~0.60), the mean is slightly higher (~0.30), and the standard deviation is twice that of Site 445 samples. Table 9 gives SCD between each sample from the three sites and the closest modern analog, the weighted average February and August SST values, and the standard deviations of those values. To generate SST we used the weighted mean of winter and summer SST associated with the five closest analogs to each Pliocene sample. An SCD value of 0.26 was used as a cutoff for analogous samples. Inspection of faunal spectra at a range of distances indicates that samples with distances greater than 0.26 are distinctly dissimilar. On Table 9 and Figure 7, gray bands are used to highlight samples and SST estimates with SCD values = 0.26. These are referred to as "no analog" samples.

Below sample 11-2, 45 (3.23 Ma), Site 445, SCD values are generally higher, indicating less similarity between Pliocene faunal assemblages and modern Pacific assemblages. Based upon the numbers of planktonic fragments recovered in each sample, carbonate dissolution increases between samples 11-2, 74 (3.24 Ma) and 13-2, 108 (3.82 Ma) relative to Core 10 samples. The only detectable faunal change is a down-core increase in abundance of Globorotalia crassaformis to levels greater than that in any modern sample. We conclude the increased dissolution and resultant no analog levels of Globorotalia crassaformis explain the higher SCD values. Alternatively the non-analog abundance of Globorotalia crassaformis could be due to a Pliocene ecological strategy different from that associated with this species today.

Oceanographic characteristics of the nearest analogs to Hole 445 samples indicate Pliocene SST that are generally the same as those of today with two instances of cooling (Fig. 7). The nearest analogs for these cool samples come from a region in the northern part of the Philippine Sea, close to the present location of DSDP Site 445. For the PRISM time slab (3.15-2.85 Ma), the peak averaging technique (Dowsett and Poore 1991) was used to determine an average deviation from modern winter and summer of +1.9°C and +0.8°C respectively.

At Hole 463, samples below 2-6, 31 (4.35 Ma) have high SCD values indicating a "no analog" situation. Samples below this level have greater numbers of benthic foraminifers, but unlike Hole 445, there is no correlation between high SCD values and high numbers of test fragments. The "no analog" samples are marked by a sharp downcore increase in the percentages of Globigerina nepenthes and Orbulina universa and a decrease in the percentages of Globigerinoides. This faunal change (increase in cooler and concomitant decrease in warmer taxa downcore) explains the step-like shift toward warmer temperatures that takes place at 4.4 Ma.

At Hole 769 only Section 6 of Core 13 and the uppermost two sections of Core 14 (3.26 Ma - 3.01 Ma) proved suitable for MAT. Generally, the upper samples gave very high SCD values. Those samples with high SCD values have low numbers of planktonic foraminifers and/or large numbers of dissolution resistant forms (Tables 3 and 9). Accordingly, we interpret these high SCD values to be the result of dissolution. The samples in cores 14 and 13 that have acceptable SCD values are at or near modern conditions for winter and summer, but show a marked departure toward cooler conditions at 3.0 Ma. The nearest modern analogs of these samples come from near 27°N, 126°E in the southern part of the South China Sea.