The depth distribution of Ostracoda from the Greenland Sea

Sixteen box core samples of Recent sediment collected in three transects across the East Greenland shelf and slope were analysed for Ostracoda. The samples which range in depth from 274 m to 3355 m, yielded a total of 52 species belonging to 25 genera. No less than 26 of the species also occur in the adjacent Scoresby Sund fjord complex. The fauna represents an interesting mixture of high latitude shallow water Arctic species and others known from bathyal and abyssal depths in the North Atlantic, including some pandemic deep sea species. Many species occur in shallower water here than in the North Atlantic due to the colder water. The fauna comprised of three associations (Shelf/Upper Slope; Slope: Lower Slope/Abyss). A marked faunal turnover occurs at the Upper Slope. The study raises questions concerning the constancy and universality of the relationship between certain ostracod species and water masses.


INTRODUCTION
The principal aims of this study are to investigate the bathymetrical distribution of benthonic Ostracoda in three transects from the outer shelf to the abyss and spanning depths from 265m to 3555m off the East Coast of Greenland. Other samples, from the Scoresby Sund Fjord system, ranging in depth from 270m to 1280m are considered elsewhere (Whatley et al., in press).
The samples were collected during September 1990 by the German research vessel R.V. Polarstern on cruise ARK VII3b, as part of the PONAM research project. This project (Late Cainozoic evolution of the P o l a r North Atlantic Margins) seeks to investigate the mechanisms involved in glacial/interglacial exchange with the aim of increasing our understanding of long term climatic change, land/shelf/ocean sediment budgets and ice/ocean mass balance. Figure 1 shows the general area of the Greenland Sea and the east coast of Greenland, and gives the present day maximum extent of ice cover. Figure 2 gives the position of the sampling stations and also the bathymetry below 500 m offshore.

METHODS
The samples were collected between 31st August and 27th September 1990, by a 0.5x0.5x0.5m box-corer, recovering an average of 30-40 cm of undisturbed sediment from the surface of the sea bed (Solheim et al., 1991). From each box core, 60 cm3 of sediment was taken, preserved in buffered formalin and, in the laboratory, washed over a 200 mesh/inch (75 p ) sieve. The dry residue was passed through a nest of sieves, and the 30, 60 and 100 meshiinch (500, 250, l50p) fractions were examined and all the Ostracoda picked out. Location, depth and ecological data for each sample, are given in the Appendix.

OCEANOGRAPHY
The Greenland and Norwegian seas are characterized by a complex exchange and mixing of Polar and Subtropical water masses and by the formation of deep, dense water. The region is controlled by submarine ridges, bounded to the North by the Nansen Ridge and to the South by the relatively shallow (600-800 m) Greenland-Iceland Ridge and Iceland-Shetland Ridge and these ridges are important constraints upon circulation.
A reduced flow of water passes through the Northern Gateway of the Fram Strait, the Southern Gateway of the Denmark Strait and the region between Iceland and Norway. The exiting water which cascades over the Southern Gateway forms a very large proportion of the North Atlantic Deep Water (NADW). Dingle & Lord (1990) demonstrated the importance of this water mass to the distribution of benthonic Ostracoda in the Atlantic.
The Greenland Basin has depths greater than 4000m. The Continental Shelf off East Greenland is deep with the slope break occurring at some 500m; it is widest to the North and decreases in width towards the South.
The Arctic seas are regarded as mediterranean type basins because of their enclosed nature and from the input of light surface waters which balances the southward outflow of deep, dense water. This results in the density of the water of the Greenland Sea being among the densest in the world. Worthington (1970) attributes the high density of the water as being due to its low temperature which is brought about by high latitude and heat loss to the atmosphere. The salinity is generally low, being between 34%0 and 35%0.
There has been much confusion and imperfect understanding of the physical mechanisms at work in the Greenland-Norwegian Sea, which mainly stems from the problem of renewal rates of deep waters and where convective mixing and the amount of 'new' water being added to the deeper layers cannot be easily determined (Peterson & Rooth, 1976). Aagaard (1 982) compared the major process of the Greenland/Norwegian Sea to a 'giant snorkel', which draws down the upper layers of water into the abyss and through into the Arctic, forming the deep polar waters. Details of the circulation and water mass budgets are given by Aagaard et al. (1985) who also discuss the water mass structure of the Arctic seas. They divided the water column into four water masses, largely on the basis of increased density with depth, and show that the boundaries between the various water masses occur at different depths throughout the Arctic basins as a consequence of them being greatly influenced by submarine topographic features: (1) The North Atlantic Drift (NAD) (surface waters). The surface waters of the Greenland Sea lie above the density level of 27.9. These waters are warmest in the Norwegian Sea ( > 1°C) and relatively highly saline (34.5%0). They originate in the Atlantic as part of the NAD and, while much thicker in the Norwegian Sea ( < 200m) and Iceland Sea ( < 150 m), here the surface water is usually only some 50 m thick and characterized by relatively high salinity ( > 35%0) and temperature ( > 1°C).
(2) Arctic Intermediate Water (AIW). This originates from the Iceland and Greenland gyres, occurs close to the surface, is well oxygenated and cold. In the Iceland and Norwegian seas, this is the primary water body to flow over the Faeroes Bank; in the latter sea, these intermediate waters are contaminated by deeper waters by upwelling. Peterson and Rooth (1976) interpreted the hydrogen isotope data, indicating that these waters have a short residence time in the Greenland Norwegian Sea Area (GNSA). They concluded that the overflow from the Norwegian Sea into the Atlantic (forming the NADW), through the Southern Gateways originates above the permanent pycnocline and, therefore, these dense overflow waters which cascade over the Greenland-Scotland Ridge are separate from the Norwegian and Greenland sea Deep Waters.
(3) The Greenland Sea Deep Water (GSDW). The N A D enters through the Southern Gateway but in the Norwegian Sea, the surface waters are too warm to sink, although the salinities are great enough to allow this. It is not until it reaches the Greenland Sea that the cooling of the surface waters by the atmosphere and ice cover, reduces its temperature enough to allow it to sink into the basins. Peterson & Rooth (1976) estimated from the hydrogen isotope decay rate that some 30years was required for the deep convective mixing of the N A D water in the Greenland Sea. Since some of the basins are 3000m in depth, this suggests that some 1OOm of the surface waters is mixed down annually. Aagaard et al. (1985) divided the water column of the region by the increase in density, in this very dense water. In the Greenland Sea, the boundary between intermediate and deep water, the 32.785 density level, occurs at c. 200m , while in the Iceland Sea it occurs at c. 650m. The GSDW extends down to some 900m where, below the 37.445 density level, true Arctic deep water occurs. The GSDW is characterized by temperatures of < 1°C and salinities of 34.88 to 34.90%0, with the Norwegian Sea Deep Water (NSDW) being slightly warmer and more saline. The Eurasian Basin Deep Water (EBDW) of the Arctic Ocean is even warmer and saltier, and, with the Lomonosov Ridge restricting the mixing of the EBDW and the Canadian Basin Deep Water (CBDW), this results in the latter water mass being the most saline and the warmest of the Arctic deep waters (Aargaard et al., 1985). this water into the Greenland Sea was discovered by Aargaard et al. (1985) when they recorded the salinities of the water column in a transect away from the coast of East Greenland This is water from the floor of the Arctic Ocean which spills over the Fram Straits into the Greenland Sea. It has a temperature of 0.95OC and a salinity of 34.93%0, and is characteristically high in silica.
The dissolved oxygen content of the Norwegian and Greenland seas is very high throughout the water column and is among the most highly oxygenated water in the world with values in the Greenland Sea between 7.82 mil-' and 7.19 11-I. These high levels are the result of low productivity, relatively low nutrients and the sinking of the NAD in the Greenland Sea. The oxygen minimum zone occurs here between 800m and 1200m, with oxygen levels of 7.2mll-'; oxygen levels of 7.4 ml1-' obtain below 2000m.
Useful diagrams, showing the water mass structure of the Greenland Sea and the physico-chemistry of its components, are to be found in Aagaard (1981), Aagaard et al. (1985) and Belanger & Streeter (1980). Some additional details can be found in Eynon (unpublished MSc Thesis, University of Wales, Aberystwyth).

Nutrients
The Greenland Sea is one of rather low productivity because of the low temperature and ice cover which restrict solar energy input into the system. Nutrient levels, through the water column, are generally characterized by low concentrations in the surface layers (consumed by plants) and higher values at lower depths (concentration due to the death and decay of organisms). In the Greenland Sea, nutrients are fairly constant throughout the water column. Belanger & Streeter (1980), show that, through the column, phosphates are constant, with a slight peak below the Oxygen Minimum Zone (at 1200m), silicates tend to increase to c. 1OOOm before becoming relatively constant and nitrates have their maximum between 500m and 1500m, before decreasing gradually with greater depth.

Recent Sediments
Sands, clays and muds prevail on the shelf being increasingly contaminated by ice rafted pebbles and boulders towards the coast (Vogt et al., 1981). In some areas substantial gouging by icebergs has taken place.

THE OSTRACODA Previous studies on benthonic Ostracoda from Greenland and the Arctic
Early taxonomic studies were by Brady (1866)  All existing published works on the Ostracoda of East Greenland are based on samples collected from depths of less than 250m. Sars (1909) recorded the total Crustacea collected by the 2nd Norwegian Arctic Expedition (1898)(1899)(1900)(1901)(1902) in the Fram Strait. Stephensen (1913) and Skogsberg (1920) recorded the shallow marine ostracods from between 68"N and 76"N off the East Greenland coast and, in 1939 Stephensen listed the ostracods encountered in Icelandic waters. Hazel (1967Hazel ( , 1970 studied the ostracods between Cape Stosch, (Latitude 74"04'N; Longitude 21"45'W) from a depth of 7fm and Clavering Island (Latitude 74"15'N; Longitude 21"OO'W) at 50fm as part of his study of the distribution of ostracods along the North American seaboard. Neale and Howe (1975) recorded the ostracods from Russian Harbour, Novaya Zemlya and also from Shannon Island (Latitude 75' 20"; Longitude 19'00'E) at 11 fm and the H.M.S. Vidal Station 46 (Latitude 75"11,2'N; Longitude 22"14'E) at 7 fm.
More recently, Hawley (1980, unpublished MSc Thesis, Aberystwyth) studied the shallow water fauna of Lysefjord in Southwest Greenland, Whatley (1982) listed the littoral and immediate sublittoral ostracods from Sisimiut in Western Greenland. and Penney (1989) reported on the fauna from Ikerssuak, Southwest Greenland.  describe three new species of Cytheropteron and one of Eucytherura from East Greenland and  discuss the fauna from the Scoresby Sund Fjord complex, East Greenland. Hartmann ( , 1993Hartmann ( , 1994 has described the Recent and subfossil Ostracoda of the Liefdejords in Spitzbergen. A number of previous studies on late Cainozoic cold water Ostracoda from Europe, North America and Greenland are relevant to this study. In Europe, Brady et al. (1874) described many species from the British Quaternary which now live in the Arctic. Lord (1980) described an important fauna from the Sandness Clay of Norway and Penney (1990) from the North Sea.
In North America, Benson et al. (1983) demonstrated the distribution and biofacies of Ostracoda from the Newfoundland slope and rise, and Cronin (1979Cronin ( , 1980 described the Pleistocene ostracod faunas from the southeastern Atlantic Coastal Plain and the St Lawrence Lowlands, respectively, and in the Greenland Sea area, Malz (pers. comm.) described the very poor fauna recovered from the Cainozoic of ODP Leg 104 on the Voting Platform off Norway, and Whatley (pers. obs.) recovered very few ostracods from the late Cainozoic of ODP Leg 151 in the region of the Fram Strait. Brouwers et al. (1991) documented the fauna of the Pliocene Kab Kobenhavn Formation, northern Greenland. Cronin (199 1) described an essentially warm water Pliocene fauna from Tjornes, Iceland, and Cronin et al. (1994) discussed Quaternary palaeoceanography based on the Ostracoda of the deep Arctic Ocean Modern studies on the deep water faunas of the North Atlantic include those by Whatley (1993), Coles (1987, 1990), Coles (1990) and Coles et al. (1990).

The Greenland Sea fauna
The species encountered in the present study are listed below. Those marked with an asterisk also occur in the adjacent Scoresby Sund Fjord complex . All the species are illustrated in Plates 1-3, and the catalogue numbers refer to the Eynon Collection in the Micropalaeontology Museum at Aberystwyth, where the specimens are housed. Acetabulastoma hyperborea Schornikov, 1970 *Argilloecia conoidea Sars, 1923 *Argilloecia cylindrica Sars, 1923 Bafinicythere howei Hazel, 1967 Bythocythere scaberrima (Brady, 1886 Sars, 1869 Explanation of Plate 1 All external lateral views unless otherwise stated. fig. 1. Argilloecia conoidea Sars, 1923. Female RV, ME/l/P3, x62.5. figs 2, 3. Argillaecia cylindrica Sars, 1923; fig. 2. RV, ME/2/T, ~7 6 ; fig. 3, LV, ME/2 /PI, ~7 6 .   (Sars, 1866). Henryhowella dasyderma (Brady, 1880 (Sars, 1865). Nannocythere sp. *Swainocythere nanseni (Joy & Clark, 1977 (Brady, 1868). Haplocytheridea bradii (Norman, 1865) Semicytherura afinis (Sars, 1865) * Thaerocythere crenulata (Sars, 1865) This fauna represents an interesting mixture of high latitude shallow water Arctic species and those known from bathyal and abyssal depths in the North Atlantic, which include some species which are virtually world wide (the major exception being the Arctic Ocean) at these depths. Since it is temperature, a secondary function of depth which, except in most unusual circumstances, is the primary controlling feature on the bathymetrical distribution of all marine benthos, at such high latitudes with such low temperatures, it is not surprising to encounter some considerable difference in the depth distribution of ostracod species between the Greenland Sea and further south in the North Atlantic, although this has not been previously documented.
For example, six species whose depth range extends below 1000 m and which are common to the North Atlantic (data from the compilation of Dingle Lord, 1990) and the area of the present study in the Greenland Sea, are seen to occur in shallower water in the latter area. Two of the species (Cytheropteron porterae and C. testudo actually occur within the adjacent Scoresby Sund Fjord system on the East coast of Greenland at depths between 277 and 1262m . The implications for the recognition of palaeo-watermass using ostracods is obvious. The comparative depth distribution (in m) of the six species in the two areas is given in Table 1. The relationship of the fauna to depth is shown in Tables 2 and 3 which plot the species in depth order by first appearance of upper and lower limit, respectively. These two tables also show at what levels various species are represented by live individuals. Figures 3 and 4 depict graphically the appearance and disappearance of species with increasing depth, respectively. The depth ranges of the ostracods plotted in Tables 2 and 3 show that certain taxa are restricted to the shelf/upper slope, the slope and the lower slope/Greenland Basin abyss. Some species, however, range across the entire bathymetry of the East Greenland continental margin. The various depth categories of these associations are given below: The shelf/upper slope association

Faunal turnover
The Greenland Sea is characterized by a faunal turnover which occurs in the depth range of the upper slope. This is shown in Tables 2 and 3 of species distribution by depth, arranged by upper and lower limit, respectively, and by Figs 3 and 4 showing the appearances and disappearances of species against depth.
They show that there is an increase in both the number of species disappearing and appearing between the depths of 600 m and 1100 m, which is exhibited in two major peaks at 650 and 1090m respectively. The majority of taxa which inhabit the continental slope first appear on the upper slope, between c. 600m and 1100m. Only seven species which occur on the shelf, range through to the middle and lower slope. This is probably to some extent a function of the fact that the thermocline, which is such a formidable ecological barrier in lower latitudes, is much less important in the Greenland Sea. A further peak in the number of disappearing taxa occurs at 2687 (Figs 3 and 4) but here, unlike the upper slope peak, the fauna is not replaced by newly appearing taxa. The ostracod counts for each sample when converted to percentage data are shown in Fig. 5. This illustrates the upper slope faunal turnover and also shows that possibly the most important correlative is the oxygen minimum zone. Figure 6. represents, in depth order of the samples, the abundance, the species diversity, the number of live specimens recovered, the percentage of live specimens and generic diversity across the East Greenland continental margin. Table 4 gives most of this data numerically. The greatest number of specimens was recovered from the continental slope, although the lowest number of specimens was also recorded from this depth zone. The greatest number and percentage of live specimens were from the shelf and upper slope samples, above 65 1 m, while a localized peak of live specimens occurs on the lower slope at 1735m.

Sample
Depth (m) 1923 1916 1924 1947 1948 1925 1918 1949 1951 1950 1926 1927 1919 1920 1921 1922  The fauna of station 1918Maddocks & Steineck (1987 and Steineck et al. (1990) considered the ostracod association living on experimental wood islands (between 1800 and 4000m) in various areas of the World Ocean. They discovered a unique assemblage of podocopid ostracods including the cytherurid genus Xylocythere [subsequently also encountered by Van Harten (1994) associated with vent faunas] and several species of Cytherois, Purudox-ostoma and Propontocypris all concentrated on the wood but absent from the surrounding sediment.
Station 1918, at a depth of 1090m, yielded the highest number of ostracod specimens from East Greenland, with Paradoxostoma, Cytherois and Propontocypris being well represented. Although no wood fragments were present in the sediment and Xylocythere was absent, the possibility of an adjacent wood parcel cannot be ruled out. Table 3. The depth distribution of ostracod species in the Greenland Sea by lower limit: xxxxx equals occurrence as dead specimens; LLLLL indicates Sample 1923191619241947194819251918194919511950192619271919192019211922

DISCUSSION AND CONCLUSIONS
The most significant result from this study is the pronounced faunal turnover which takes place between 600 and 1100m, where shelf species are replaced by bathyal and abyssal taxa. The Greenland Sea is characterized by fairly constant temperatures and generally little variation in nutrient levels. Several possible causes exist for the faunal turnover, such as change in temperature salinity or oxygenation at a water mass boundary. Aagaard (1981) recognised a zone of reduced salinities in the Greenland Sea between 1000 and 1600m and Belanger Streeter (1980) noted that the oxygen minimum zone occurs between 800 and 1200 m. However, although both of these phenomena occur at depths which render it possible to allow one to invoke their agency in causing the faunal turnover, the reductions in salinity and oxygen levels are very small scale. Dingle & Lord (1990) identified the water mass boundaries in the Atlantic as a cause of the depth zonation of certain deep sea ostracods. The upper limit of the Greenland Sea Deep Water occurs at 900m, which correlates with the faunal turnover as shown by the first peak in Figs 3 and 4, and probably exerts   (1978), and Barkham (1985, unpublished MSc Thesis, Aberystwyth) off the coast of Northwest Africa and Cronin (1983) off Cape Hatteras, have recorded a faunal turnover below the shelf break. The East Greenland continental margin is characterized by a very deep continental shelf, with the shelf break occurring at c. 500 m and, therefore, the change in fauna occurs in a similar position to these other studies (second peak , Figs 3 and 4).
Such Arctic shelf taxa as Thaerocythere crenulata, Muellerina abyssicola and Krithe glacialis have their lower depth limit at 1100 m. Water temperature does not decrease here significantly with depth below 200 m (see Appendix) so that temperature alone cannot be the sole barrier to downslope migration of species into the deep sea. Because of the absence of an effective thermocline at these latitudes, theoretically they are where we should expect to encounter a 'backdoor' means of entry into the deep sea and it is very likely this route has been of significance in the past and may still be operative today. It is not easy to prove that the High Latitude Doorway hypothesis has been a major route of access into the deep sea, rendered so difficult by the inception of the thermocline in the late Palaeogene (Benson, 1975;Whatley, 1996) and from that time virtually impossible in lower latitudes, but the temperature controlled stenotopic depth distribution differences between Arctic and North Atlantic assemblages of the same species (Table 1) is possible evidence in the case of certain species. Certainly, in the present study, the clear distinction between shelf and bathyal/abyssal taxa in the Greenland Sea demonstrates the significance of the ecological barriers that exist between them.
The deep sea fauna of the Arctic Ocean lacks many of the taxa which are virtually pandemic elsewhere in the world's oceans. As early as 1969, Benson after examining 'a few samples from the Arctic' was able to state that this area does not have 'a normal deep-water fauna at the present time'. We note the absence from the Greenland Sea of such genera as Bradleya Hornibrook and Poseidonamicus Benson, members of the pandemic deep sea fauna. Also, Coles (pers. comm. 1992) commented that the Krithe spp. from the Greenland Sea '. . .seem to be distinct from 'true' deep sea North Atlantic faunas further south and in deep sea water areas'. Joy & Clark (1977) in their study of the deep water ostracods of the Canadian Basin, recorded three species which are common with the fauna of the Greenland Sea. These are: Henryhowella dasyderma (1 35 1-3 193 m), Pseudocythere caudata (1531-3198 m) and Polycope punctata (1 351-2810 m). These authors also described 13 species which were then confined to the Canadian Basin and these figures indicate the importance as a barrier to migration of the Lomonosov Ridge.
More species (at least ten), are common with the fauna described from the Quaternary of the deep Arctic Basin by Cronin et al. (1994). It is not possible to be more precise since neither Krithe nor Polycope are speciated in the latter work.