In recent years, dinoflagellate cysts have been used to interpret Neogene coastal, shelf, and deep-sea depositional environments of the Atlantic Ocean and adjacent seas (Harland, 1988; Mudie, 1989; de Vernal et al., 1992; McCarthy and Mudie, 1996; Matthiessen and Brenner, 1996; Versteegh, 1997; Zonneveld and Bossenkool, 1996) and to estimate shifts in paleotemperatures and ocean circulation patterns (Edwards et al. 1991; Mudie, 1992). Although the ecological preferences of most dinoflagellate cysts are not known in detail, the identified distribution pattern of many extant species found in Atlantic Ocean surface sediments (Mudie, 1992; Matthiessen, 1995; Dale, 1996) can be used to interpret the paleoenvironments of the upper Pliocene-Pleistocene succession at Site 986. There is an intimate relationship between dinoflagellate cyst assemblages and environmental conditions; further, regional ecostratigraphies reflect climatic changes with accuracy. It is also established that interpretations of Quaternary dinoflagellate cyst assemblages, based on the ecology of modern cysts, give results that are in good agreement with interpretations derived independently from oxygen isotope data, foraminifers, and coccolith microfloras (Mudie and Harland, 1996). As Matthiessen and Brenner (1996) pointed out, however, it must be kept in mind that dinoflagellate cysts may be transported with surface and bottom currents over considerable distances, and that species preferring warmer water masses can be advected from the Norwegian Sea with the West Spitsbergen Current into the Arctic Ocean. Therefore, the dinoflagellate cyst record at Site 986 reflects not only local surface-water mass conditions but also regional changes in the West Spitsbergen Current and the North Atlantic Current. Thus, as in the approach used by Matthiessen and Brenner (1996), only changes in the abundance of taxa that are common to abundant are used to interpret changes in paleoenvironments through the upper Pliocene and Pleistocene succession.
In addition to determining fluctuations in the abundance of selected taxa, the gonyaulacoid to peridinioid value (G:P ratio) is used to detect changes in paleoenvironmental conditions in the upper Pliocene of Hole 986D. This ratio provides a rough index of surface-water temperature and sea-ice conditions (Edwards et al., 1991; Mudie, 1992; Mudie and Harland, 1996). Edwards et al. (1991) correlated five groups of G:P ratios with modern environments as follows: values <0.2 mark perennial pack-ice and arctic estuaries, ratios of 0.2-0.6 mark the seasonal ice zone, ratios of 0.8-2 characterize subarctic and transitional waters, and ratios greater than 4 indicate subtropical to tropical waters. The G:P ratios for Hole 986D are shown in Figure 7. At some levels with low counts of dinoflagellate cysts, the ratio has not been calculated. A major limitation of the G:P method is that the ratio often reflects other ecological factors. Low values may also indicate upwelling or deltaic conditions as well as the influx of large amounts of land-derived nutrients. According to Edwards et al. (1991), the method yields a very broad summer sea-surface temperature of 14°-26ºC only when protoperidinioids are rare or absent. Consequently, the method may be imprecise for larger parts of the upper Pliocene-Pleistocene succession at Site 986.
Recent data from ODP Legs 151 in the Norwegian-Greenland Sea (Thiede and Myhre, 1996) and 152 off southeast Greenland (Larsen et al., 1994) have provided evidence that the middle Miocene was fairly mild, that cooling most likely started shortly after 10 Ma within the early late Miocene, and that full glacial conditions in southeast Greenland were not established until the middle part of the late Miocene. According to Thiede and Myhre (1996), the influence of sea ice and of the influx of icebergs, first from Greenland and later from northwestern Europe, increased during the Pliocene until it reached an important threshold ~3-4 Ma off southern Greenland and 2.6 Ma in the Fram Strait and off northwestern Europe.
The upper Pliocene sediment recovered below Reflector R7 at Site 986 (Unit SV-VIII) comprises silty clay, nearly barren of dropstones. High proportions of in situ dinoflagellates compared to reworked forms suggest a relatively low influence of redepositional processes compared to the overlying units. The relatively high diversity of in situ taxa and the high content of amorphous organic matter in the palynological preparations further support a hemipelagic depositional regime for these oldest sediments (Forsberg et al., Chap. 17, this volume). G:P ratios of dinoflagellate cysts are between 2 and 3, suggesting temperate to relatively warm sea-surface conditions. The sample at 946.04 mbsf, however, records a distinct cooling event, whereas one sample at 898.04 mbsf recorded a G:P ratio typical of subtropical to tropical waters. The most prominent gonyaulacoid taxa in Unit SV-VIII are Operculodinium centrocarpum, O. israelianum, Spiniferites spp., and Reticulatosphaera actinocoronata. A distinct acme of Impaginidium patulum is observed at 928.34 mbsf. O. centrocarpum is a ubiquitous species, tolerating a wide range of temperature and salinity conditions. This species is dominant in recent sediments from the Norwegian Sea (Matthiessen, 1995). In contrast, O. israelianum is a typical warm (temperate to tropical) species. The occurrence of this species supports the G:P ratio evidence of periods with warm water during the late Pliocene. The acme of I. patulum can be linked to the influx of oceanic water masses and suggests a strong influence of the North Atlantic Current during deposition of the sediments at 928.34 mbsf.
Unit SV-VIII also contains moderate counts of the protoperidinioids Brigantedinium spp. and Selenopemphix brevispinosa. Low-diversity dinoflagellate cyst assemblages with high quantities of Brigantedinium spp. are often a characteristic of polar to subpolar environments. However, assemblages dominated by Brigantedinium spp. are also common in temperate regions, especially in coastal upwelling areas (Wall et al., 1977). These assemblages, having a high diversity (especially of protoperidinioid cysts [Mudie, 1992; Mathiessen and Brenner, 1996]), differ from polar to subpolar ones. Thus, the presence of fairly common protoperidinioids in Unit SV-III does not contradict the overall picture of the influence of relatively warm water masses during deposition of the upper Pliocene clays.
Reflector R7 is interpreted to mark the onset of a major glacial wedge along the Svalbard-Barents Shelf margin. There are, however, no distinct changes in dinoflagellate assemblages and no marked lithologic changes across this seismic boundary. As in the underlying unit, O. centrocarpum and Spiniferites spp. are the most common taxa. O. israelianum is present throughout Unit SV-II, mostly in low numbers but with distinct abundance peaks at 830.64 mbsf and between 637.99 and 564.44 mbsf. Most notable is the absence of Brigantedinium spp. in the upper unit SV-III and in lower unit SV-II (up to 795.04 m).
The acmes of O. israelianum call for particular comment since these cysts suggest a distinct inflow of warm water into what is interpreted as a glacially dominated depositional area. These influxes of O. israelianum coincide with contemporaneous high abundances of this species at Sites 898 and 900, Leg 149, off Spain, in the late Pliocene to early Pleistocene (McCarthy and Mudie, 1996). The species is, however, only recorded in moderate to low numbers in the late Pliocene in southwestern and eastern England (Head, 1993, 1996, respectively) and in the late Pliocene to early Pleistocene of the Norwegian Sea (Mudie, 1989). Mathiessen and Brenner (1996) recorded low numbers of O. israelianum (as O. crassum) at two restricted intervals in the upper Pliocene and lower Pleistocene deposits at Hole 911A on the Yermak Plateau. These probably correlate with some of the warm-water influxes, which also are recorded at Site 986. Both the marked acme at 584.74 mbsf in the upper Pliocene at Site 986 and the minor abundance peak at 526.04 mbsf in the lower Pleistocene are accompanied by relatively high G:P ratios. The event at 526.04 mbsf corresponds to a similar and approximately contemporaneous event noted from 268.39 to 249.56 mbsf in Hole 911A on the Yermak Plateau (Fig. 7).
The record of warming episodes in the upper Pliocene-lowermost Pleistocene of Site 986 supports earlier observations from the Kap København Formation on Perry Land in North Greenland. Here Funder et al. (1985) recovered frequently well-preserved remains of terrestrial vegetation and invertebrate faunas indicative of a forested tundra environment intercalated within the glacial sequences. More recently, Willard (1996) recovered late Pliocene terrestrial microfloras at Sites 910 and 911, Leg 151, that are suggestive of open boreal vegetation with relatively temperate deciduous elements. These microfloras indicate warmer late Pliocene conditions in the source area adjacent to the Yermak Plateau. Similarly Spiegler (1996), Cronin and Whatley (1996), and Osterman (1996) recorded microfaunas typical of warm surface waters in upper Pliocene marine deposits at Site 910 on the Yermak Plateau. Based on these findings, Spiegler (1996) suggested that warm and subtropical surface-water masses invaded the generally cold Norwegian-Greenland Sea and Arctic Ocean in the late Pliocene during short-lived episodic events. This interpretation is supported by the results from Site 986, both from the dinoflagellate cyst record and from observation of the foraminifer fauna by Eidvin and Nagy (Chap. 1, this volume). The present data further suggest a relatively strong influence of the North Atlantic Current during periods of the late Pliocene to early Pleistocene and that warm surface-water masses were transported along the West Spitsbergen Current into the Fram Strait. These periods of warm-water inflow must have alternated with periods of continuous sea-ice cover in the Arctic Ocean (Matthiessen and Brenner, 1996).
The Pleistocene sediments recovered at Site 986 are characterized by marked fluctuations in abundances of the dominant dinoflagellates Brigantedinium spp., Bitectatodinium tepikiense, Spiniferites spp., and Operculodinium centrocarpum. This compares well with the distribution pattern seen through the contemporaneous deposits at Hole 911A on the Yermak Plateau (Matthiessen and Brenner, 1996).
The sediments directly above seismic Reflector R6 (i.e., up to 526.04 mbsf) contain relatively high abundances of Brigantedinium spp., with the maximum at 545.24 mbsf. High abundances of these forms are often associated with polar to subpolar conditions. The occurrence of the distinct arctic species Algidaspheridium? minutum at 564.44 mbsf supports the influence of cold-water masses during deposition of these sediments. This species is only a minor constituent of the Pleistocene dinoflagellate assemblages at Site 986, but when present, it is generally associated with common Brigantedinium spp.
Between 535.64 and 526.04 mbsf, Brigantedinium is replaced by common to abundant Bitectatodinium tepikiense. This species has broad thermal tolerance but is absent or rare at present in the Arctic Ocean (Mudie, 1992; Matthiessen and Brenner, 1996). In the northwestern Atlantic, this species appears to prefer environments with cold winter and warm summer sea-surface temperatures (de Vernal et al., 1992). According to Matthiessen and Brenner (1996), high percentage abundances of B. tepikiense are attributed to offshore mixing of polar-influenced oceanic waters with cold, brackish meltwaters from ice. They are associated with salinities around 30% and a relatively short seasonal sea-ice cover or even ice-free conditions.
Between 516.44 and 455.84 mbsf, there is a marked drop in the abundance of in situ dinoflagellates. The abundance of dinoflagellate cysts declines strongly with decreasing surface temperatures in polar regions; areas of permanent sea ice have <100 dinoflagellate cysts per gram, and most sediments under sea ice are barren (Mudie, 1992; Mudie and Harland, 1996). Quaternary glacial-stage sediments also have barren intervals, and the distinct drop in abundance between 516.44 and 455.84 mbsf can probably be related to a change to glacially dominated depositional conditions. Minor acmes in the distribution of Brigantedinium spp. and Algidasphaeridium? minutum at 484.18 mbsf appear to support this interpretation.
A new inflow of warmer surface-water masses is suggested by new high abundances of Bitectatodinium tepikiense, Spiniferites spp., and Operculodinium centrocarpum, and by the presence of Operculodinium israelianum at 436.64 mbsf. There also is an increase in the relative abundance of Brigantedinium spp. at this level, but the absence of A.? minutum supports the interpretation of a change to warmer conditions. The increased abundance of Brigantedinium spp. may here be attributed to increased nutrient levels, perhaps supplied from melting sea ice.
Fluctuations in the abundance of Brigantedinium spp., Bitectatodinium tepikiense, Spiniferites spp., and O. centrocarpum continue up to 151.84 mbsf (up to ~1.0 Ma). Low abundances and diversities of dinoflagellates suggest periods with cold conditions and low productivity during deposition of the sediments at 321.56-312.44 mbsf and 254.76-245.04 mbsf. High abundances of the three latter taxa indicate a change to warmer and more productive water masses during deposition of the sediments at 392.44-382.84, 353.94, 333.54, and 302.04 mbsf and from 235.44 to 151.84 mbsf.
Maximum abundance peaks of Spiniferites spp. and O. centrocarpum are seen at 200.04 mbsf. An age of 1.2 Ma has been assigned to the sediments at this level (Channell et al., Chap. 10, this volume). This event coincides with distinct acmes in the abundance of Lingulodinium machaerophorum, Impagidinium aculetum, Spiniferites mirabilis, and Nematosphaeropsis labyrinthus. Mudie and Harland (1996) described L. machaerophorum and S. mirabilis as typical cool-temperate to tropical species. According to Dale (1996), the consistent appearance of L. machaerophorum in interglacial sequences from Norway confirms the value of this species as an indicator of temperate or warmer waters. Mudie and Harland (1996) further describe I. aculeatum as a characteristic subtropical oceanic species, whereas N. labyrinthus is typical in temperate neritic to oceanic environments. Observations from Holocene sediments in the Norwegian-Greenland Sea and the Arctic Ocean, however, have shown that N. labyrinthus is also common in the cold-temperate gyres in the Greenland and Iceland seas, the subpolar gyre, and the eastern Arctic Ocean (Mudie, 1992; Matthiessen, 1995). The combined dinoflagellate evidence, however, suggests a period with inflow of temperate to warm North Atlantic water masses into the West Spitsbergen Current around 1.2 Ma.
Above 141.89 mbsf, the assemblages show low percentages of Bitectatodinium tepikiense but are otherwise characterized by similar fluctuations in the abundances of Brigantedinium spp., Spiniferites spp., and Operculodinium centrocarpum compared with the older Pleistocene intervals below. In their study of Hole 911A, Matthiessen and Brenner (1996) found that since ~1 Ma, the dinoflagellate cyst assemblages of the Yermak Plateau reveal a similarity in composition to Holocene assemblages from oceanic surface-water environments in the Norwegian-Greenland Sea and eastern Arctic Ocean. In overall character, the contemporaneous in situ assemblages at Site 986 are comparable to those of the Yermak Plateau. The assemblages differ, however, in containing high proportions of reworked Cenozoic and Mesozoic dinoflagellate cysts. A marked abundance peak of O. centrocarpum is observed between 66.04 and 58.04 mbsf at Site 986. A similar distinct acme of O. centrocarpum is recorded at 98.95 mbsf in Hole 911A. The dominance of O. centrocarpum at this stratigraphic level on the Yermak Plateau suggests an episode with increased influence of the North Atlantic Current and West Spitsbergen Current at around 0.5 Ma.
The low-resolution sampling at Site 986 does not allow identification of any small-scale cycles through the upper Pliocene-Pleistocene. The low abundance and diversity at 30.24 mbsf is attributed to glacial conditions, whereas the increase in the relative abundance of O. centrocarpum and Spiniferites spp. between 11.24 and 3.74 mbsf may imply a new retreat of ice-dominated waters in the northern Atlantic-Arctic region. According to Lloyd et al. (1996), initial deglaciation of the Spitsbergen margin occurred at 14.1 ka. These authors, however, found no evidence of an influx of warmer subpolar waters during this time. Lloyd et al. (1996) further determined that the first influx of subpolar North Atlantic waters into the area was during the second deglacial phase of the Spitsbergen ice mass at the beginning of the Holocene. This episode, dated at ~9 ka, was as warm as if not warmer than the present day along the Spitsbergen margin. The sample at 3.74 mbsf can probably also be related to this warming event. More detailed analyses, however, are required to confirm this and to detect the high-frequency glacial-interglacial cycles through the Pleistocene succession at Site 986.