The base of the recovered sedimentary section in Hole 985A (Sample 162-985A-62X-CC) contains several stratigraphically important forms. Spirosigmoilinella compressa (Matsunaga) was originally described from the Miocene of Japan. Its FO in the North Atlantic region was calibrated to the standard nannofossil zonation at Site 647 in the Labrador Sea (Kaminski et al., 1989). In this area, its FO lies within the upper Eocene (nannofossil Zone NP17). In the North Sea exploration wells, its FO usually marks the base of the Oligocene sedimentary sequences. In the North Sea area, however, the Eocene/Oligocene boundary is represented by a hiatus, and much of the upper Eocene is generally missing. At Site 643 on the Vøring Slope, its FO was observed in Sample 162-643A-49X-5, 79-84 cm. Based on this correlation, we can assume that the base of the recovered sediments at Site 985 is no older than early Oligocene. S. compressa was reported to range into the upper Miocene in Hole 909C in the Fram Strait (Osterman and Spiegler, 1996). However, in their study of the Leg 151 material, Osterman and Spiegler did not separate this species from Psamminopelta gradsteini.
At Labrador Sea Site 647, the LO of Reticulophragmium amplectens was observed at the Eocene/Oligocene boundary as determined by nannofossil and magnetostratigraphy. This species, however, ranges into the Oligocene in the Norwegian Sea at Site 643 (Kaminski et al., 1990) and in the Canadian Arctic (Schröder-Adams and McNeil, 1994). Our observations at Site 985 confirm its overlap with S. compressa and its continued range into younger sediments north of the Greenland-Scotland Ridge. In the northern part of the Norwegian Sea at Site 909, it was reported to range into the middle Miocene (Osterman and Spiegler, 1996).
In Core 162-985A-59X, an assemblage with common specimens of Reticulophragmium rotundidorsatum is observed. This species was originally described from the middle Oligocene Kiscell Marls of Hungary (Hantken, 1875). However, the specimens at this level are much smaller than the types preserved at the Natural History Museum in Vienna. They are more similar, both in size and shape, to late Eocene specimens from the Carpathian Flysch deposits where an upper Eocene R. rotundidorsatum Zone has been recognized (Geroch and Nowak, 1984). This would imply that the base of the recovered section is older than the mid-Oligocene. Larger specimens of R. rotundidorsatum are observed higher in the section, in Sample 162-985A-46X-7. The occurrence of these ancestral forms at the base of Hole 985A implies that the sediments are older than the "Kiscellian" of Hungary. R. rotundidorsatum is a cosmopolitan form that ranges into the middle to upper Miocene, occurring also in the Oligocene to upper Miocene in the Beaufort-MacKenzie Basin (Schröder-Adams and McNeil, 1994) and in the Miocene "Agua Salada Fauna" in the Gulf of Mexico (M.A. Kaminski, unpubl. data). In the northern Norwegian Sea, Osterman and Spiegler (1996) reported an R. rotundidorsatum assemblage of late Miocene age in Hole 909C.
Karreriella siegliei, observed at the base of Hole 985A, is the nominate taxon of the early Oligocene Dorothia siegliei Zone of Gradstein et al. (1994) in the North Sea. At Site 643, its FO was observed in Core 162-643A-47X, which was assigned an Oligocene age (Kaminski et al., 1990). Its LO was reported within the upper Miocene R. rotundidorsatum assemblage at Site 909 (Osterman and Spiegler, 1996).
At DSDP Sites 338, 346, 349, and 350, Verdenius and van Hinte (1983) found the distinctive species Spiroplectammina spectabilis in sediments of Eocene age. At the Labrador Sea Site 647, the LO of this species coincided with the Eocene/Oligocene boundary (Kaminski et al., 1989). Based on the occurrence of the above-mentioned species and the lack of S. spectabilis in our samples, we conclude that the base of the recovered sediment column at Site 985 is no older than early Oligocene.
To assess the utility of DWAF for stratigraphic correlation in the area of the southern Norwegian Sea, we compiled the LOs of common species in the Eocene-Oligocene at three localities (Fig. 3). M.A. Kaminski has studied all three localities, ensuring consistent taxonomic concepts. Several stratigraphically important species are common to all three locations, including S. compressa, R. elongatus, H. walteri, Cyclammina placenta, R. amplectens, and D. seigliei. This comparison of the LOs also reveals information about the paleoecological preferences of the species.
There is relatively good agreement between the sequence of LOs in the two ODP records from the deep Norwegian Sea (Fig. 3). In both ODP holes, the "coarse agglutinated species" (i.e., Recurvoides, Cribrostomoides, and Rhabdammina) and Psamminopelta gradsteini are among the last of the Oligocene forms to disappear. Within the Oligocene sediments, the species Ammosphaeroidina pseudopauciloculata, C. placenta, H. walteri, Haplophragmoides kirki, and H. rugosa display a consistent sequence of last occurrences. This relatively good correlation between the FOs in Holes 985A and 643A suggests that similar ecological controls on DWAF may have been present at both localities. Deep-water masses in the Norwegian Sea were undersaturated with respect to carbonate, and DWAF persisted well into the Oligocene in the deeper parts of the basin. At Site 909 situated in the Fram Strait, a diverse DWAF assemblage was recovered in sediments of middle to late Miocene age (Osterman and Spiegler, 1996), indicating that favorable ecological conditions persisted much later at the high latitudes. There are some inexplicable differences between the foraminiferal records at Sites 985 and 643. For example, R. rotundidorsatum was not found at Site 643, and C. voeringensis ranges into the Oligocene at Site 985. It was found only in sediments of the Eocene age at Site 643.
However, along the shallow margins of the Norwegian Sea and in the central North Sea situated toward the south, a change from predominantly agglutinated assemblages to calcareous assemblages is generally observed within the lower Oligocene (Gradstein et al., 1994; Gradstein and Bäckström, 1996). At bathymetrically shallow sites on the Vöring Plateau, the Oligocene sediments contain a benthic assemblage consisting of both calcareous and agglutinated species with Turrilina alsatica and Rotaliatina buliminoides. Some species known from the Haltenbanken area are absent at Sites 985 and 643 (e.g., Annectina biedai Gradstein and Kaminski and Aschemocella grandis [Grzybowski]), whereas others have markedly different stratigraphic ranges. A number of DWAF species that persisted into the Oligocene at the deeper sites display LOs within the Eocene sediments in the quantitative zonation of Gradstein and Bäckström (1996), including R. amplectens, A. pseudopauciloculata, and H. kirki (Fig. 3). Dorothia siegliei does not range above the Eocene in wells from the Haltenbanken area (Gradstein and Bäckström, 1996), but it does range into the Oligocene sequence in the northern North Sea and in our ODP holes.
Some species display stark differences in their stratigraphic ranges in different areas of the Norwegian Sea. For example, Hyperammina rugosa was first described from the Oligocene of the DSDP Leg 38 sites (Verdenius and van Hinte, 1983). Its LO in the North Sea and western Barents Sea consistently lies within the upper Paleocene. At Site 643, this species was also found in the lower Eocene. Osterman and Spiegler (1996) recorded it from middle to upper Miocene sediments at Site 909.
The diachronous LOs of deep-water species such as Haplophragmoides walteri, Ammosphaeroidina pseudopauciloculata, Reticulophragmium amplectens, and the coarsely agglutinated forms are consistent with a bathymetric shallowing of the Vøring Plateau (owing to infill of the basin and lower Oligocene sea levels). The early Oligocene faunal turnover also reflects changes in surface-water masses in the Norwegian Sea, which allowed the establishment and preservation of calcareous microfossils.
The role of the Greenland-Scotland Ridge as an effective barrier to the exchange of deep water between the Atlantic and the Arctic throughout the Cenozoic is well known (Thiede and Eldholm, 1983). Today, the Norwegian Sea provides the largest volume of Northern Hemisphere deep water (Worthington, 1976). The Greenland-Scotland Ridge acts as a bathymetric barrier, restricting the flow of Norwegian Sea deep water into the North Atlantic. Similarly, this barrier prevents the northward flow of southern-source deep water into the Norwegian Sea. Berggren and Schnitker (1983) suggested that the Faeroe-Shetland Channel may have been opened by subsidence associated with the Hebridean volcanism and that there may have been an exchange of water across the Greenland-Scotland Ridge during the early Eocene. In the geologic past, this blocking effect of the Greenland-Scotland Ridge must have been even greater, as portions of the ridge have been subsiding since the Eocene. These findings are consistent with our understanding of the paleobiogeography of DWAF, which display some interesting differences across the Greenland-Scotland Ridge.
By comparing the faunal composition of assemblages at Site 985 with the assemblages from the Paleogene of the Labrador Sea farther south, we can speculate about the extent of faunal connections across the Greenland-Scotland Ridge. These faunal records are important in understanding the history of deep-water connections across this ridge.
If we contrast the Oligocene DWAF species occurring at either side of the Greenland-Scotland Ridge, we notice some paleobiogeographical differences. We can divide the species occurring at Sites 647 (Labrador Sea), 643, and 985 into three categories (Table 2).
The first species group constitutes forms that are known from the deep Labrador Sea in the late Eocene to early Oligocene but have not been found in approximately coeval sediments in the southern Norwegian Sea. This group includes a number of species first described from deep-water sediments occurring in the Caribbean and in the Polish Carpathians. Several typically Atlantic abyssal forms such as Spiroplectammina cubensis, Trochamminoides spp., and Paratrochamminoides spp. have never been observed in the Norwegian Sea region. These forms are relatively common in the Oligocene Cipero Formation of Trinidad, for instance. Another example is the widely distributed deep-water species Glomospira charoides (Jones and Parker), which is common in the lower Oligocene of Hole 647A in the Labrador Sea. It disappeared from the Vøring Slope area during the Eocene. This species is occasionally observed in older (Paleocene to lower Eocene) strata in the Norwegian Sea and North Sea regions. Its absence north of the Greenland-Scotland Ridge (Holes 985A, 643A, and 913B) during the Oligocene again points to the faunal isolation of the deep Norwegian Sea Basin.
A second group of species is widely distributed along the Labrador margin and North Sea regions (Gradstein et al., 1994) as well as in the deep ODP holes. This group includes several index species that may be useful for correlation between the Tethys, Atlantic, and North Sea regions such as R. rotundidorsatum, R. amplectens, A. latus, R. elongatus, S. compressa, and P. gradsteini. These species apparently possessed a very wide bathymetric distribution during the Oligocene. We believe that the Greenland-Scotland Ridge apparently did not constitute a barrier to their migration into the Norwegian Sea and North Sea regions. It is likely that these species migrated into the Norwegian Basin with the incoming surface-water masses. Latest Eocene to Oligocene diatom assemblages from Holes 908 and 913 in the northern Norwegian Sea contain many low-latitude species, suggesting that warm waters were present (Scherer and Koç, 1996). This supports the premise that surface waters entered the Norwegian Sea from the Atlantic (or also via the North Sea), bringing temperate floras and faunas as well as neritic benthic foraminifers (see also discussion by Thiede and Myhre, 1996). The species R. rotundidorsatum, R. amplectens, S. compressa, and A. pseudopauciloculata have also been reported from the Oligocene to lower Miocene of the Beaufort-Mackenzie Basin (Schröder-Adams and McNeil, 1994), indicating further faunal connections with the Arctic province, probably via the Fram Strait.
The third group of species constitutes "boreal" forms such as Karreriella siegliei, Adercotryma agterbergi, Conotrochammina voeringensis, and Annectina biedai. These species, so distinctive for the Oligocene of the North Sea and Norwegian Sea regions, have so far not been observed south of the Greenland-Scotland Ridge. These species are widely distributed in Paleogene sediments throughout the North Sea region (Charnock and Jones, 1990; Gradstein et al., 1994; Gradstein and Kaminski, 1997). Because their distribution in the Norwegian Sea is also isobathyal, their absence south of the ridge is more difficult to explain.
The absence of deep Atlantic species from the boreal seas (and the occurrence of some "endemic" species) is consistent with the theory that the Greenland-Scotland Ridge was a barrier to southern-source deep water and contributed to the isolation of Norwegian Sea deep water. During the Paleogene, marine connections across the Greenland-Scotland Ridge would have had shallower sill depths than at present, causing greater isolation of the deep-water mass than in the modern ocean. Instead, it is likely that the Oligocene deep waters of the Norwegian Sea had a greater affinity to Arctic deep waters because of the opening of the Fram Strait during Anomaly 13 (Berggren and Schnitker, 1983). At present, deep-ocean convection in the southern Norwegian sea occurs to depths of 3000 m, resulting in a deep lysocline and well-oxygenated deep water (Jansen and Raymo, 1996). During the Oligocene, however, both sedimentological and faunal evidence point to a very different oceanographic setting. The lack of carbonate at the deep Norwegian Sea sites indicates a bottom-water mass that is both vertically stratified and corrosive. Although the presence of agglutinated foraminifers in the Oligocene sediments indicates that the bottom waters were probably never anoxic, the fact that calcareous benthic foraminifers have only been found at shallow sites on the Vöring Plateau suggests that convection, if present, involved only the thermocline waters.
In the absence of deep convection and open connections to the North Atlantic, it is reasonable to assume that the Norwegian Sea deep waters were poorly oxygenated during the Oligocene. The foraminiferal data at Site 985 and at other DSDP sites support this idea. The change from an early Oligocene coarsely agglutinated assemblage at the base of Hole 985A to a sparse assemblage with Psamminopelta gradsteini and pyritized diatom/radiolarian steinkerns in the upper Oligocene-Miocene sequence can be interpreted as reflecting increasingly dysaerobic conditions at the seafloor. In a study of Holocene DWAF from the California Borderlands basins, the change from a relatively diverse assemblage with coarsely agglutinated tubular forms to an assemblage displaying high dominance of small, thin-walled forms was observed as the bottom-water oxygen content falls below 0.5 mL/L (Kaminski et al., 1995). The Oligocene-Miocene Psamminopelta gradsteini assemblage found at Site 985 and at other DSDP sites in the central Norwegian Sea (Verdenius and van Hinte, 1983) probably indicates a deep-sea environment that is strongly dys-aerobic. By contrast, the deep-sea environment at Site 647 south of the Greenland-Scotland Ridge was well ventilated during the Oligocene, and assemblages contain diverse calcareous benthic foraminifers.
Our comparisons of Oligocene benthic foraminifers from various localities in the northern North Atlantic lead us to conclude that the Greenland-Scotland Ridge not only constituted a physical barrier (perhaps preventing some truly abyssal species from migrating into the boreal seas), but by isolating the Norwegian Sea deep water it caused extreme environmental conditions that excluded many of the isobathyal cosmopolitan species.