We test the hypothesis that changes in radiolarian faunal indices correspond to climatic change across the Eocene/Oligocene boundary interval and within the lower Oligocene, the null hypothesis being that radiolarian faunal indices develop independently from reconstructed changes in Earth-system parameters. In contrast to low latitudes, the sedimentary record of Eocene siliceous microplankton is generally patchy in southern high latitudes. This applies to diatoms as well as radiolarians (Baldauf, 1992; Baldauf and Barron, 1990). The observation that radiolarian preservation and accumulation rates increase gradually from the Eocene to the Oligocene has been made at many Southern Ocean Deep Sea Drilling Project and ODP sites (Table T4) including Antarctica. In nearly all regions where both Eocene and Oligocene sediments were studied, the Oligocene yields better-preserved faunas than the Eocene. This pattern is quite opposite to that in the tropics where the average preservation is much better in the late Eocene (Nigrini and Sanfilippo, 2000), but it is coherent with findings from northern high latitudes (Lazarus and Pallant, 1989). Although well-preserved faunas were recovered from the Eocene at several localities including the Kerguelen Plateau (Caulet, 1991), the mean preservation of radiolarians is significantly (p < 0.001, based on t-test) different between Eocene and Oligocene samples. One of the few exceptions is reported from the Falkland Plateau (Weaver, 1983), where well-preserved faunas are apparently present throughout the Eocene and Oligocene, although preservation varies between holes.
The general paucity of silica accumulation/preservation in many Eocene age southern high-latitude sites has been attributed to low productivity. Enhanced Oligocene silica accumulation and productivity are usually explained by high-latitude cooling, which increased latitudinal temperature gradients and led to stronger oceanic turnover and thus higher nutrient supply (Kennett, 1977). Tectonic uplift and enhanced weathering may also have contributed to increasing nutrient concentrations (Zachos et al., 1999). The development of the Antarctic Bottom Water formation may additionally have aided radiolarian preservation and declination of planktonic foraminiferal preservation (Diester-Haass, 1995). Although authors have often argued for a continuous and gradual climatic deterioration (Keller et al., 1992), it is now clear that cooling was punctuated in the earliest Oligocene, at least in the Southern Ocean (Wei, 1991; Zachos et al., 1999; Zachos et al., 2001).
Considering the arguments above, we are tempted to invoke climatic cooling and associated productivity fluctuations as the prime control of radiolarian preservation, abundance, and diversity. Although the relationship may not be as straightforward (Diester-Haass, 1996), the lower Oligocene opal maximum is usually associated with high values of other productivity proxies such as benthic foraminiferal accumulation rates and carbonate dissolution (Diester-Haass, 1996; Diester-Haass and Zahn, 1996). Carbonate dissolution is invoked as a paleoproductivity proxy owing to the increase of calcite dissolution with increasing organic carbon supply in a well-oxygenated environment (Diester-Haass, 1995). In our material, a temporal decoupling of the productivity proxies is evident. Carbonate dissolution appears to be most substantial in the upper Eocene and earliest Oligocene, judging from the incomplete stratigraphic record in Core 183-1138A-36R and the pronounced hiatus at the top of Core 183-1138A-37R (Fig. F2). Radiolarian preservation, in contrast, is not significantly enhanced before the uppermost lower Oligocene (Section 183-1138A-35R-2; Chron C10r or possibly Subchron C11n.1), immediately after the end of the condensed sequence of Core 183-1138A-36R. This pattern suggests regional differences on the Kerguelen Plateau. The sediments from Southern Kerguelen Plateau show a profound synchronous increase in productivity in the earliest Oligocene, Chron C13n (Sites 738 and 744; Diester-Haass, 1995, 1996; Zachos et al., 1999; Site 748; Wise et al., 1992), whereas the central and Northern Kerguelen Plateau (Sites 1138, 1139, and 1140) sediments record an increase in radiolarian and diatom preservation 3 to 4 m.y. later. This interpretation is consistent with the observation of Lazarus and Caulet (1993) that endemic radiolarian faunas, indicative of a distinct surface water mass, developed first during the late Eocene only close to the Antarctic continent on the Southern Kerguelen Plateau and spread throughout the Antarctic region only later, during the Oligocene.