Late Neogene-Quaternary radiolarian biostratigraphy: a brief review

Since the 1950s, it has become apparent that Radiolaria have significant biostratigraphical potential throughout Phanerozoic time, including the Late Neogene and Quaternary. Radiolarian biozonation schemes for this period have been developed, including a Standard Tropical Zonation, which illustrates the pan-oceanic application of radiolarian biostratigraphy to Pliocene–Quaternary sediments. The biostratigraphical resolution obtainable using Radiolaria is equivalent to other microfossil groups, such as planktonic foraminifera. The recognition of abundance events of Cycladophora davisiana, and of some other species, are an alternative radiolarian dating technique for the Pliocene–Quaternary, akin to dating sediment using oxygen stable isotope (δ18O) records and with similar resolution. A number of studies have used astronomical timescales, derived from orbitally tuning δ18O and gamma ray attenuation porosity evaluator (GRAPE) records, to provide ages for radiolarian biodatums. This approach should be adopted as a more accurate alternative to palaeomagnetic chronologies with their inherent flaws. This commentary concludes that Radiolaria are important microfossils and, as a group, continue to offer significant potential as a biostratigraphical tool in future studies of the marine Pliocene–Quaternary.


INTRODUCTION
work provided a means of relatively dating tropical Quaternary sediments using Radiolaria. Absolute ages for these biodatums were later provided in Johnson & Knoll (1975). The stratigraphic resolution obtainable using Nigrini's (1971) scheme, with only four biozones spanning the entire Quaternary, is inevitably coarse. Yet, despite this, the scheme has been widely used, mainly because it is easy to apply, and has been incorporated into a 'standard' Cenozoic low-latitude radiolarian biozonation (e.g. Riedel & Sanfilippo, 1978;Sanfilippo et al., 1985). This is not for the lack of alternative schemes, as Goll (1980), for example, erected a more refined biozonation for the tropical Pacific. Also, other schemes have been erected for other geographical regions, such as the North Pacific (Hays, 1970;Kling, 1973;Foreman, 1975), North Atlantic (see review by Haslett, 1995), northeast Atlantic (Björklund, 1976;Goll & Björklund, 1980) and Antarctica (Hays, 1965(Hays, , 1967Hays & Opdyke, 1967;Chen, 1975).
In the late 1980s a concerted attempt was made to provide a new higher resolution Pliocene-Pleistocene radiolarian biozonation. This was initiated by an extensive appraisal of the timing of known Neogene radiolarian biodatums in palaeomagnetically and biostratigraphically dated cores from the equatorial Indian and Pacific Oceans (Johnson & Nigrini, 1985). This study indicated that a number of biodatums previously used as biozonal boundary markers were diachronous and unsuitable for inter-oceanic biostratigraphy (see also Spencer-Cervato et al., 1993). These results contributed further to the lack of suitable radiolarian species available for Pliocene-Pleistocene biostratigraphy. Perhaps, in response to this situation, taxo-nomic studies were undertaken on two potentially useful genera of Radiolaria, Pterocorys  and Anthocyrtidium , which provided a number of new biodatums based on Indian and Pacific Ocean material, again from palaeomagnetically and biostratigraphically dated cores. Subsequentally, Johnson et al. (1989) erected a new palaeomagnetically calibrated Pliocene-Pleistocene radiolarian biozonation for the tropical Indian Ocean, later refined by Caulet et al. (1993) (Fig. 1; Table 1). Johnson et al. (1989) recognized 33 radiolarian biodatums and erected 11 biozones for the Pliocene-Pleistocene, nearly doubling the number of biozones reported earlier for the same interval by Sanfilippo et al. (1985). This increase in biozones put Radiolaria on a par with other microfossil biozonations, such as planktonic Foraminifera (Bolli & Saunders, 1985) and calcareous nannofossils (Martini, 1971), but exceeding the resolution obtainable through diatoms (Barron, 1985).
All of this work referred to the Indian and Pacific oceans, and there was little information concerning the biostratigraphic application of these species in the Atlantic Ocean. Haslett (1994) published an account of Pliocene-Pleistocene radiolarian biostratigraphy at DSDP Site 609 in the North Atlantic ( Fig. 1) and demonstrated that elements of Johnson et al.'s (1989) biozonation could be applied in the Atlantic Ocean. This suggests that this scheme offered the potential to become a standard Pliocene-Pleistocene biozonation, enabling correlation throughout the world's tropical to mid-latitude oceans. This is hampered somewhat, however, by the apparent diachroneity of some biodatums between the Indian and Atlantic oceans ( Table 2).  Tables 1 and 3 for explanation of zone abbreviations). References: 1 Shackleton et al. ( , 1995; 2 Nigrini (1971), Sanfilippo et al. (1985), Moore et al. (1993); 3 Johnson et al. (1989); 4 Haslett (1994); 5 Sanfilippo & Nigrini (1998). Dashed line indicates tentative boundary position.

Neogene-Quaternary radiolarian biostratigraphy
Recognition of five Pleistocene radiolarian zones (NR1-NR5) has been made in the South China Sea (Wang & Abelmann, 1999), further extending the wide application of this radiolarian biozonation scheme. The absolute ages for these biodatums were calculated in all cases cited above through interpolation between dated palaeomagnetic reversals using the time-scale of Berggren et al. (1985a, b). Johnson et al. (1989) suggest that the precision of some biodatums is less than 100 ka; however, little consideration is given to possible variations in sedimentation between reversals. Ruddiman et al. (1986) and Shackleton et al. ( , 1995 provide an alternative method of dating sedimentary cores by tuning 18 O and gamma ray attenuation porosity evaluator (GRAPE) density records to the periodicity of astronomical cycles: 400 ka and 100 ka eccentricity cycles, 41 ka obliquity cycles and 23-19 ka precessional cycles. This method commonly allows for samples to be dated to within c. 20 ka (i.e. interval between adjacent precessional excursions) and enables very high precision ages to be applied to biodatums. Moore et al. (1993) and Moore (1995) exploit astronomical time-scales constructed for ODP Leg 138 cores from the eastern equatorial Pacific to derive ages for the assessment of diachroneity of Neogene-Quaternary radiolarian biodatums in the region (Table 2). They categorized all events occurring throughout the region within 150 ka as synchronous, and six of the eleven Quaternary (c. 0-2 Ma) biodatums in the eastern equatorial Pacific met this criteria. Of the diachronous events, only one was significantly greater than 300 ka. This study is important for the accuracy of the astronomically tuned ages applied to the radiolarian biodatums, which does not involve the same degree of interpolation as palaeomagnetic time-scales.
Extending this approach into the Indian Ocean, Haslett et al. (1995) investigated a Pliocene-Early Pleistocene section from ODP Site 709 and identified ten radiolarian biodatums. Oxygen isotope stratigraphy is available for part of the section (Shackleton & Hall, 1990), which was astronomically tuned by correlation with ODP Site 677 ; see also Funnell et al., 1996), allowing dates to be assigned to six biodatums (Table 2). Only one of the biodatums identified in ODP Site 709 (LAD Anthocyrtidium jenghisi) appears to be synchronous with the eastern equatorial Pacific, as defined by Moore et al. (1993). To date, no attempt has been made to extend this approach of dating radiolarian biodatums through the entire Quaternary of the Indian Ocean, or at all in the Atlantic Ocean. These are pressing objectives that are required for a fully integrated low-latitude radiolarian biozonation of the world's oceans.
Recently, Sanfilippo & Nigrini (1998) have offered a revised biozonal nomenclature for the Cenozoic, which provides a Standard Tropical Zonation based on Radiolaria with the introduction of new code numbers ( Fig. 1 and Table 3). In terms of the Pliocene-Pleistocene part of this zonation the main departure from Johnson et al. (1989) occurs in the Pliocene where their zones NR6 and NR7 are replaced with a single zone Table 2. Radiolarian biodatum ages in the Atlantic, Indian and Pacific oceans. Ages for Atlantic and Indian Ocean biodatums are recalculated here using astronomically derived ages for palaeomagnetic reversals (Shackleton et al., , 1995 Haslett (1994) and Westberg-Smith et al. (1986) b after Johnson et al. (1989) and Haslett et al. (1995) c Moore et al. (1993) and Moore (1995) Sanfilippo & Nigrini (1998) b where more than one date is given by Sanfilippo & Nigrini (1998), the mean age is used Neogene-Quaternary radiolarian biostratigraphy (RN12), subdivided into two subzones, and NR8 and NR9 are replaced with a single zone (RN11), again with two subzonations. The varied use of zonal codes, where one scheme descends with time and the other ascends, is clearly unsatisfactory at present and may lead to confusion. It is suggested that efforts to harmonise the schemes should be made, but the present review is not a suitable vehicle for this task. All ages given by Sanfilippo & Nigrini (1998, fig. 2) for their zonation are palaeomagnetically derived, incorporated within the Geomagnetic Polarity Time Scale (GPTS) and, whilst this is justified in older Cenozoic sediments where orbitally tuned time-scales are lacking, ages for Late Neogene and Quaternary biodatums should now be (where possible) astronomically derived. They do provide tables for the conversion of ages from palaeomagnetic time-scales (e.g. Berggren et al., 1985aBerggren et al., , b, 1995Cande & Kent, 1992 to an orbitally tuned time-scale (Shackleton et al., 1995) but, as mentioned above, interpolation of ages between magnetic reversals cannot incorporate sedimentation rate fluctuations, which may lead to considerable ageassignment errors. However, orbitally tuned time-scales are erected independently, unaffected by sediment accumulation rates and may establish age control points at c.25 ka intervals. Therefore, a biodatum inaccurately aged by palaeomagnetic interpolation will convert to an erroneous astronomically derived age.

CYCLADOPHORA DAVISIANA ABUNDANCE EVENTS
Cycladophora davisiana is generally a cold-water high-latitude species that evolved c. 2.5 Ma (Motoyama, 1997) and whose relative abundance has been shown to vary synchronously through the Late Pliocene and Quaternary of high-latitude oceans (Hays et al., 1976;Morley, 1980Morley, , 1987Morley & Hays, 1979, 1983Abelmann & Gersonde, 1988;Björklund & Ciesielski, 1994). The abundance variations of this species have been used as a means of correlating sedimentary sequences in a similar manner to 18 O stratigraphies (i.e. counting peaks and troughs in the continuous record). Indeed, some authors have dated C. davisiana abundance events using oxygen isotope stratigraphy, such as in the North Atlantic (Ciesielski & Björklund, 1995) and in the subantarctic Atlantic Ocean (Brathauer et al., 2001). Abundance events have been variously coded by authors for ease of reference when cross-correlating different cores; either by letters (e.g. Hays et al., 1976), numbers (e.g. Haslett & Funnell, 1998, Fig. 2), or alpha-numerically (e.g. Caulet, 1982Caulet, , 1986. Although stratigraphies based on C. davisiana are widely used at mid-to high-latitude sites, Haslett & Funnell (1998) demonstrated that the same method may also be applied with success to low-latitude sediments. These authors erect C. davisiana stratigraphies spanning the Olduvai subchron (Plio-Pleistocene) of ODP sites from the equatorial Indian and Pacific Ocean (sites 677, 709, 847, 850 and 851), as a means of  Haslett & Funnell, 1998). independently evaluating high-resolution time-scales developed by other methods. The results are encouraging, with a high degree of consistency in the C. davisiana stratigraphies between all sites. In particular, Haslett & Funnell (1998) recognize a major peak in C. davisiana abundance (Fig. 2) coincident with the Plio-Pleistocene boundary at 1.81 Ma (Hilgen, 1991). The percentage value of this peak varies from site to site, probably due to local palaeooceanographic factors (Haslett & Funnell, 1998), but may represent a useful new biomarker for the Tertiary-Quaternary boundary, indicating a major intrusion of cool high-latitude water into the equatorial oceans at this time, perhaps associated with 18 O stage 64. Indeed, Ciesielski & Björklund (1995) note that in the North Atlantic the initial major abundance peak of C. davisiana occurred near the stage 63/62 boundary. These results indicate that C. davisiana stratigraphy may have the potential for correlation not only in high-latitude regions, but also on intra-and inter-oceanic scales covering all latitudes. However, these promising indications require further research in the future. Furthermore, abundance variation in other species may also prove biostratigraphically useful. For example, Gupta (2002) demonstrates that temporal abundance variations in pyloniid radiolarians vary in a similar way to that of C. davisiana.

CONCLUSIONS
The range and precision of radiolarian applications in geological studies, including the Pliocene-Pleistocene, have developed since the work of William R. Riedel in the 1950s indicated their potential. This paper has focused on biostratigraphic applications. Much is now known concerning modes of evolution and extinction and specific dates when particular species appeared or became extinct. These biodatums have been dated primarily through palaeomagnetic means, interpolating time-scales between magnetic reversals. This has provided a reasonably accurate chronology, which, for most dating purposes, is adequate. However, such palaeomagnetic dating of biodatums cannot easily account for sediment accumulation rate variations between reversals. A more precise method of attributing dates is through the construction of astronomical time-scales, which involves much less interpolation, so reducing errors associated with sedimentation rates. Future refinement of radiolarian biodatums should logically employ such orbitally tuned time-scales. A number of biozonal schemes have been developed, based on radiolarian biodatums, which subdivide Pliocene-Pleistocene time and allow for sediment correlation on an inter-oceanic basis. Abundance variations of Cycladophora davisiana and some other taxa have been shown to fluctuate through time in response to environmental changes. These fluctuations have been shown to occur systematically on an inter-hemisphere and inter-oceanic scale allowing stratigraphies to be developed based on the identification of particular abundance events. This is particularly promising as it allows time-scales to be constructed in a similar manner and resolution to 18 O stratigraphies. However, this technique requires further confirmation of its potential from the analysis of deep-sea cores from widely distributed sites in the ocean. From this review, it is clear that radiolarian stratigraphy is currently not limited as suggested by Lowe & Walker (1997) and their use is hampered only by preservation.