Environmental changes o¡ North Iceland duringthe deglaciation and the Holocene: foraminifera,

diatoms and stable isotopes

K.L. Knudsen a;�, H. Jiang a;b, E. Jansen c, J. Eir|¤ksson d, J. Heinemeier e,M.-S. Seidenkrantz a

a Department of Earth Sciences, University of Aarhus, DK-8000 Aarhus C, Denmarkb Laboratory of Geographic Information Science, East China Normal University,Shanghai, 200062, PR China

c Bjerknes Centre for Climate Research and Department of Geology, Alle¤gt. 55, N-5007 Bergen, Norwayd Science Institute, University of Iceland, IS-101, Reykjavik, Iceland

e The AMS 14C Dating Laboratory, Institute of Physics and Astronomy, University of Aarhus, DK-8000 Aarhus C, Denmark

Received 13 November 2002; received in revised form 10 June 2003; accepted 26 June 2003

Abstract

A combined study of foraminifera, diatoms and stable isotopes in marine sediments off North Iceland recordsmajor changes in sea surface conditions since about 15 800 cal years (yr) BP. Results are presented from two gravitycores obtained at about 400 m water depth from two separate sedimentary basins on each side of the submarineKolbeinsey Ridge. The chronology of the sedimentary record is based partly on AMS 14C dates, partly on the Veddeand the Saksunarvatn tephra markers, as well as the historical Hekla AD 1104 tephra. During the regionaldeglaciation, the planktonic foraminiferal assemblages are characterised by consistently high percentages of sinistrallycoiled Neogloboquadrina pachyderma. However, major environmental variability is reflected by changes in stableisotope values and diatom assemblages. Low N

18O values indicate a strong freshwater peak as well as possible brineformation by sea-ice freezing during a pre-B=lling interval (Greenland Stadial 2), corresponding to the Heinrich 1event. The foraminifera suggest a strong concurrent influence of relatively warm and saline Atlantic water, and boththe foraminifera and the diatoms suggest mixing of cold and warm water masses. Similar but weaker environmentalsignals are observed during the Younger Dryas (Greenland Stadial 1) around the level of the Vedde Ash. Eachfreshwater peak is succeeded by an interval of severe cooling both at the beginning of the B=lling^Aller=d InterstadialComplex (Greenland Interstadial 1) and during the Preboreal, presumably associated with the onset of intense deepwater formatiom in the Nordic Seas. The Holocene thermal optimum, between 10 200 and about 7000 cal years (yr)BP, is interrupted by a marked cooling of the surface waters around 8200 cal yr BP. This cold event is clearlyexpressed by a pronounced increase in the percentages of sinistrally coiled N. pachyderma, corresponding to atemperature decrease of about 3‡C. A general cooling in the area is indicated after 7000^6000 cal yr BP, both by thediatom data and by the planktonic foraminiferal data. After a severe cooling around 6000 cal yr BP, the planktonicforaminiferal assemblages suggest a warmer interval between 5500 and 4500 cal yr BP. Minor temperaturefluctuations are reflected both in the foraminiferal and in the diatom data in the upper part of the record, but the time

0377-8398 / 03 / $ ^ see front matter G 2003 Elsevier B.V. All rights reserved.doi:10.1016/S0377-8398(03)00075-6

* Corresponding author. Tel. : +45-8942-3557; Fax: +45-8618-3936.E-mail address: karenluise.knudsen@geo.au.dk (K.L. Knudsen).

MARMIC 953 2-2-04

Marine Micropaleontology 50 (2004) 273^305

R

Available online at www.sciencedirect.com

www.elsevier.com/locate/marmicro

resolution of the present data is not high enough to pick up details in environmental changes through the lateHolocene.G 2003 Elsevier B.V. All rights reserved.

Keywords: palaeoceanography; diatoms; foraminifera; stable isotopes; sediments; last 16 000 years; North Atlantic

1. Introduction

Strong climatic gradients between the Arcticand the North Atlantic realms traverse the NorthIcelandic shelf, enhancing the signi¢cance of thiskey area for the study of oceanographic variabil-ity through the deglaciation and the Holocene.The shelf is located in a sensitive boundary region

between Atlantic water, which is brought to thearea by the relatively warm, high-salinity IrmingerCurrent, and cold, low-salinity surface water ofthe East Icelandic Current (Fig. 1). The Arcticsurface water of the East Icelandic Current ispartly derived from the East Greenland Current(Polar water) and partly from westerly eddies ofthe Norwegian Atlantic Current (Atlantic water)

Fig. 1. The regional modern surface circulation around Iceland and the position of the marine Polar Front. Depth contour inter-vals 1000 m (modi¢ed after Knudsen and Eir|¤ksson, 2002 and Hurdle, 1986, Map A8). The extent of Fig. 2 is indicated by abox.

MARMIC 953 2-2-04

K.L. Knudsen et al. /Marine Micropaleontology 50 (2004) 273^305274

(e.g. Stefansson, 1962; Johannesen, 1986; Swift,1986; Malmberg and Jonsson, 1997; Hansen andMsterhus, 2000).

The strength of the North Atlantic Current isgenerally related to deep water formation in theNordic Seas, which is again associated with south-ward over£ow across the Greenland^Iceland^Fae-roe^Scotland Ridge. The North Atlantic Currentis expected to be strong during active deep waterformation, and weak during periods of fresheningof the surface waters north of Iceland, i.e. duringperiods with strong input of Polar water from theEast Greenland Current and the East IcelandicCurrent. At present the Norwegian Sea DeepWater (NSDW) replaces the mixed surface watermasses at about 3^400 m depth o¡ North Iceland(Figs. 2 and 3a,b). During periods of strong over-£ow in the Denmark Strait and across the Ice-

land^Faeroe Ridge, the cold deep water masses(NSDW) may be expected to in£uence the topo-graphic basins north of Iceland (Eir|¤ksson et al.,2000a).

The modern summer sea surface temperature inthe study area is relatively constant, generallyaround 6^7‡C, while the winter sea surface tem-perature is around 1^3‡C (Jiang et al., 2002).These temperatures were lower during past coldperiods when the East Icelandic Current had anincreased East Greenland Current component,while warmer and more saline waters werebrought into the area during periods of relativelystronger in£uence of the Irminger Current (e.g.Eir|¤ksson et al., 2000a; Andrews et al., 2001a;Andrews and Giraudeau, 2002). Consequently,past changes in the position of the oceanic PolarFront across the North Icelandic shelf can be reg-

Fig. 2. Location of the core sites north of Iceland (contour intervals 100 m). The arrow indicates the location of the oceano-graphic section on Fig. 3. Abbreviation: TFZ, Tjo«rnes Fracture Zone.

MARMIC 953 2-2-04

K.L. Knudsen et al. /Marine Micropaleontology 50 (2004) 273^305 275

istered in the sedimentary record of the region.This has previously been demonstrated both forthe areas around Iceland (e.g. KocS et al., 1993;Sarnthein et al., 1995; Voelker et al., 1998) andfrom the shelf west of Iceland (Jennings et al.,2000) as well as north of Iceland (Andrews etal., 2000; Eir|¤ksson et al., 2000a,b; Andrews etal., 2001a). Consistency between marine late Ho-locene proxies from the North Icelandic shelf andatmospheric and terrestrial data from the Green-land ice sheet and from Iceland glacier advancesshows that there is a close relationship betweenthe oceanic system and the atmospheric circula-tion in the region (Eir|¤ksson et al., 2000b; An-drews et al., 2001b,c; Andrews and Giraudeau,2002; Jiang et al., 2002).

The sea£oor of the North Icelandic shelf ischaracterised by the Tjo«rnes Fracture Zone fea-

turing numerous active basins in a mud dominat-ed environment (Fig. 2). The sedimentary recordof the two gravity cores HM107-05 and HM107-04 (Fig. 2), extending back to almost 15 800 cal yrBP, represents part of the deglaciation as well asthe entire Holocene. Local and temporal variabil-ity in sedimentation rates are observed on eachside of the ridge. A high sedimentation rate isrecorded during the Lateglacial and the Preborealin core HM107-05 east of the ridge (mean = 48cm/1000 yr), probably re£ecting proximity of gla-cial and glaciomarine processes. This was fol-lowed by reduced Holocene rates (mean = 12.5cm/1000 yr) east of the ridge, while the sedimen-tation rate was higher in core HM107-04 west ofthe ridge (mean = 33 cm/1000 yr). As the eastward£ow of the Irminger Current along the outer shelfis parallel to the depth contours, the Kolbeinsey

Fig. 3. Oceanographic section from the north coast of Iceland towards north across the shelf to the slope (the Siglunes pro¢le).Summer salinity (a) and temperature (b) data (August 1998) are presented as an example of the distribution of water masses.The oceanography is based on data from the Iceland Marine Institute (Hedinn Valdimarsson, pers. commun.; see OceanographicGroup Homepage, http://www.hafro.is/hafro/Sjora/index.htm). Abbreviations: db, decibarW1 m; nm, nautical mile. (After Knud-sen and Eir|¤ksson, 2002.)

MARMIC 953 2-2-04

K.L. Knudsen et al. /Marine Micropaleontology 50 (2004) 273^305276

Ridge forms a barrier favouring deposition of bedload silts on its westward side where the currentsare slowed down and de£ected across the ridge.

The purpose of this paper is to present highresolution planktonic foraminifera, stable isotopeand diatom evidence of variable sea surface con-ditions o¡ North Iceland through the Lateg1acialand across the Pleistocene^Holocene transition,and a lower resolution record for the Holocene.Benthic foraminiferal species and sedimentarydata are included when important for the environ-mental interpretation. An overview of benthic fo-raminiferal distribution and sedimentological datain a lower time resolution study of the same twocores was previously published by Eir|¤ksson et al.(2000a).

The chronostratigraphy in this study is basedon a combination of well-known tephra markersand a series of AMS 14C dates. The record isdiscussed in terms of the Greenland ice-coreevents (Bjo«rck et al., 1998; Walker et al., 1999;Lowe et al., 2001), but for reason of clarity, theclassic chronostratigraphical terminology of Man-gerud et al. (1974) is indicated as well.

2. Materials and methods

Two gravity cores, HM107-05 (66‡54P08QN,17‡54P19QW; water depth 396 m, core length394.3 cm) and HM107-04 (67‡13P38QN,19‡03P00QW; water depth 458 m, core length392.7 cm) from the North Icelandic shelf (Fig.2), were obtained in 1995 on the HM107 BIOICEcruise with RV Haakon Mosby, 90 and 120 kmo¡shore (see also Eir|¤ksson et al., 2000a). Thecores are located on each side of the KolbeinseyRidge.

2.1. Foraminifera

The foraminiferal samples generally represent1-cm sediment slices, and analyses are carriedout at intervals varying between 1 and 10 cm(the time resolution varies between 15 and a max-imum of 1200 years). The samples were washedthrough 1000-, 125- and 63-Wm sieves, accordingto the methods described by Feyling-Hanssen et

al. (1971) and Knudsen (1998). The foraminiferawere concentrated from the 125^1000-Wm dryfraction of sediment by means of the heavy liquidC2Cl4 (speci¢c gravity 1.6 g cm33). At least 300specimens of planktonic as well as benthic speci-mens were analysed in the 125-Wm fraction, whenpossible. In some intervals the number of plank-tonic specimens was relatively low, and samplescontaining less than 30 specimens were excludedfrom the statistics. The Preboreal sediments weregenerally very poor in foraminifera, especiallyfrom the planktonic group. Some of the samplesfrom this interval were, therefore, supplementedwith neighbouring 1-cm sediment slices up to anaggregate thickness of 3 cm. For taxonomic notes,see Appendix A.

2.2. Diatoms

Each diatom sample represents a 1-cm sedimentslice, and analyses were carried out at intervals of5 or 10 cm. All the samples were treated with 10%HCl to remove the calcareous matter, washedwith distilled water, and treated with 30% H202

(1^2 h in a water bath at 60‡C) in order to destroythe organic material. Diatom slides were made byusing Naphrax (dn = 1.73). In all 102 sampleswere analysed (51 samples for each core). Morethan 300 diatoms valves were counted in eachsample (excluding Chaetoceros resting spores),although at some levels the limited amount ofdiatoms in the sediments prevented reaching this¢gure. In general, the diatoms were well pre-served. Diatom assemblage zones were identi¢edusing the cluster analysis technique of Grimm(1987). For taxonomic notes, see Appendix B.

2.3. Stable isotopes

Stable isotopes were measured on the plankton-ic sinistrally coiled Neogloboquadrina pachydermaas well as on the benthic species Islandiella nor-crossi from the 125^1000-Wm fraction of each fo-raminiferal sample, when possible. For intervalswith sparse contents of I. norcrossi the benthicspecies Cassidulina neoteretis was measured in-stead. The data set was supplemented by a seriesof overlapping samples with measurements of

MARMIC 953 2-2-04

K.L. Knudsen et al. /Marine Micropaleontology 50 (2004) 273^305 277

both benthic species. This made it possible to nor-malise the N

18O values of C. neoteretis to I. nor-crossi (correction factor +0.69 V 0.29). All theN

18O values presented in the diagrams are cor-rected for the ice volume e¡ect, using the Fair-banks (1989) sea-level curve as dated by Bard

(1990) with a correction of 0.11x N18O per 10

m sea-level change (subtracted from the measuredN

18O values). The stable isotopes were measuredon a Finnigan MAT 251 mass spectrometer at theStable Isotope Laboratory, University of Bergen.Results are given with respect to PeeDee Belem-nite, after calibration to the National Institute ofStandard and Technology NBS 19 standard.

Fig. 4. Lithology of core HM107-05 based on X-ray photo-graphs, and mineralogical analyses. The whole-core X-raylogs show laminations and structures drawn after the X-raynegatives. Pebbles s 2 mm in diameter are shown as fre-quency per cm core length. For de¢nition and discussion ofIRD, see text. Benthic foraminiferal biozones (from Eir|¤kssonet al., 2000a) and diatom biozones (this paper) are indicated,as well as the level of the Vedde and Saksunarvatn tephras(stippled lines).

Fig. 5. Lithology of core HM107-04 based on X-ray photo-graphs, and mineralogical analyses. The level of the Veddeand Saksunarvatn tephras and the historical Hekla AD 1104tephra marker are indicated (stippled lines). See also captionof Fig. 4.

MARMIC 953 2-2-04

K.L. Knudsen et al. /Marine Micropaleontology 50 (2004) 273^305278

2.4. Sediments

The basic sedimentological properties of coresHM107-05 and HM107-04 were published byEir|¤ksson et al. (2000a), including grain size, car-bonate content, water content, shear strength,mineralogical composition and tephra geochemis-try. The present paper, therefore, only presents anoverview of the lithology, shown on a depth scalein Figs. 4 and 5 together with important tephramarkers as well as ice rafted debris (IRD) con-tents. We also include some additional grain sizeparameters, pebble content and higher resolutionmineralogical data.

Sorting (standard deviation) was calculated us-ing moment statistics (McBride, 1971), and thesortable silt percentage and mean were calculatedas the percentage of the 10^63-Wm fraction out ofthe total ¢nes and the mean grain size of thatfraction (McCave et al., 1995).

The pebble content and speci¢c mineralogicalcomponents are used as proxies for sedimentationfrom melting ice in the waters north of Iceland.The number of clasts over 2 mm in diameter wascounted on whole-core X-ray negatives, and thenumber of pebbles per 1-cm-sediment thickness inthe core was calculated. This method may under-estimate the pebble frequency when two or morepebbles coincide in core depth and exposure di-rection, but the generally very low concentrationof pebbles in both cores is considered to reducethis potential error.

For the study of mineralogical provenance, atleast 300 grains were identi¢ed and counted in thes 125-Wm sand fraction of the sieved sample. Thisgrain size interval has been used in previous stud-ies of IRD content in the Nordic Seas (e.g. Hen-rich et al., 1995). Most of the allochtonous, non-volcanic sand grains fall into three groups: crys-tals (mainly quartz and feldspar), rock fragments(plutonic and sedimentary), and altered orabraded volcanic fragments. A substantial portionof the ¢rst two groups must originate outside Ice-land, the reworked volcanics may have Icelandicor Greenlandic provenance. It is considered mostlikely that the sand grains with provenance out-side Iceland were transported and released to theNorth Icelandic shelf by sea-ice and icebergs from

Greenland and the Arctic Ocean. Some of thepotentially local sand grains may have becomeincorporated in sea-ice through eolian, coastalor shelf processes. Statistically, however, the £uxof the three mineralogical groups is positively(r ranging from 0.35 to 0.75) and signi¢cantly(P6 0.0005 in all cases) correlated in both cores,and each group can be used to predict the othertwo. This supports the interpretation that all threegroups share a common origin, and can be con-sidered to represent IRD. Substantial coastal sea-ice cover may enhance eolian transport of sand tothe middle and outer shelf. The IRD £ux wascalculated as the product of the mass accumula-tion rate (as de¢ned by van Andel et al., 1975)and the number of IRD grains per gram sediment.

3. Chronology

The chronology for this work is based on com-bined tephrochonology and AMS 14C datings ofeither molluscs or benthic or planktonic forami-nifera. Most of the AMS 14C dates have previ-ously been published by Eir|¤ksson et al. (2000a),who also presented an age^depth model based on14C ages (uncalibrated) for each of the two cores.A standard marine reservoir correction of 400years was used throughout the entire record byEir|¤ksson et al. (2000a). For the age model ofthis paper, all the 14C dates have been calibratedwith CALIB4 (Stuiver et al., 1998a). The marinemodel calibration curve with vR= 0, correspond-ing to a reservoir correction of approximately 400years (Stuiver et al., 1998b; see also Andersen etal., 1989), has been used for the Holocene record,while a reservoir correction of 800 years(vR= 400) has been applied for older samples(Tables 1^4). The present-day reservoir age ofthe coastal waters of Iceland is consistent withthe vR = 0 value, and Eir|¤ksson et al. (2000a)did show that a vR= 0 value is also applicableat the level of the Saksunarvatn ash (9000 14Cyr BP), whereas a reservoir correction of 750^800 years was necessary at the Vedde Ash level(10 300 14C yr BP). Clearly, a change of reservoirage took place in the waters north of Iceland be-tween the deposition of the Vedde and the Saksu-

MARMIC 953 2-2-04

K.L. Knudsen et al. /Marine Micropaleontology 50 (2004) 273^305 279

Table 1Radiocarbon datings and tephra markers in core HMI07-05, calibrated with CALIB4 (Stuiver et al., 1998a) using the marine model calibration curves (Stuiver etal., 1998b)

HM107-05, depth Laboratory No. Material/dated tephra layer 14C age Cal age(s) BP Cal V 1 s Cal age(s) BP Cal V 1s N13C

(cm) (BP) V 1 s (R =400) (BP) (R =800) (BP)

18^19 AAR-3381 Foraminifera, total benthic fauna 2 015 V 45 1 560 1 620^1 520 31.149^51 AAR-4117 Foraminifera, total benthic fauna 5 050 V 45 5 430 5 460^5320 30.9122^125 AAR-4419 Foraminifera, total benthic fauna 9 190 V 80 9 820 9 930^9 620 31.7125^128 AAR-4420 Foraminifera, total benthic fauna 9 250 V 70 9 840 10 260^9 730 31.4127 Saksunarvatn ash 10 200137^141 AAR-3380 Foraminifera, total benthic fauna 9 730 V 60 10 330 10 600^10 300 31.1212^213 AAR-3382 Foraminifera, total benthic fauna 10 900 V 100 12 570^12 340 12 800^11 770 11 620^11 400 11 910^11 160 31.1222 Vedde Ash 12 000224^227 AAR-4118 Foraminifera, Neoglob. pach. sin. 10 970 V 60 12 610^12 370 12 810^12 150 11 660 11 910^11 370 +0.3224^228 AAR-4119 Foraminifera, total benthic fauna 11 090 V 80 12 800^12650 12 880^12 370 11 920^11 750 12 280^11 590 30.5228^231 AAR-3379 Foraminifera, total benthic fauna 11 440 V 90 12 980 13 130^12 870 12 780^12 630 12 870^12 340 30.8331^334 AAR-4116 Foraminifera, total benthic fauna 12 920 V 80 14 350 15 290^14 180 14 070 14 270^13 710 31.3364^366 AAR-4115 Foraminifera, total benthic fauna 13 560 V 90 15 700 15 950^15 480 15 300 15 510^14 380 31.2379.5^381 AAR-3377 Foraminifera, Neoglob. pach. sin. 13 790 V 120 15 970 16 240^15 710 15 520 15 760^15 280 31.1380^381 AAR-3378 Foraminifera, total benthic fauna 14 010 V 120 16 230 16 490^15 970 15 760 16 020^15 520 30.9393^394.3 AAR-3375 Foraminifera, Neoglob. pach. sin. 14 100 V 140 16 330 16 610^16 060 15 870 16 140^15 600 31.3393^394.3 AAR-3376 Foraminifera, total benthic fauna 13 690 V 100 15 860 16 110^15 610 15 430 15 630^14 480 31.2393^394.3 AAR-3383 Molluscs, Yoldiella lenticula, Thyasira gouldi 13 980 V 90 16 190 16 440^15 950 15 730 15 970^15 500 31.3

A standard reservoir correction of ca. 400 years (vR= 0) is built into this model for samples younger than 10 000 14C yr, while ca. 800 years (vR=400) has beenused for older samples. The datings were carried out at the AMS 14C Dating Laboratory, University of Aarhus, Denmark.

MA

RM

IC953

2-2-04

K.L.Knudsen

etal./M

arineMicropaleontology

50(2004)

273^305280

narvatn tephras, close to the Pleistocene^Holo-cene boundary.

Several studies from the northern North Atlan-tic have previously shown higher reservoir agevalues for the Lateglacial than observed at present(e.g. Austin et al., 1994; Bard et al., 1994; Ha£i-dason et al., 1995; Bondevik et al., 2001). Wael-broeck et al. (2001) have presented data showingthat the reservoir age was not constant during theLateglacial, amounting to 930 V 250 yr during theYounger Dryas, but as much as 1880 V 750 yr atthe end of the Heinrich 1 event (14 500 cal yr BP).Both Bondevik et al. (2001) and Waelbroeck et al.(2001) found that the reservoir ages were compa-rable to the present-day values during the warmB=lling^Aller=d period. Taking these variationsinto account is not straightforward, however, asadditional complications may well have been in-troduced by anti-phase relationships between thewater masses of the North Icelandic shelf regionand those of the eastern North Atlantic duringthe Lateglacial. We have therefore chosen toadopt a vR= 400 value for age calibration of allsamples older than 10 000 14C yr BP.

The age^depth models of HM107-05 andHMI07-04 are shown in Figs. 6 and 7 (Tables 1and 3). The age models were constructed bymeans of age zones that are delimited by tephramarkers, 14C dates and biozone boundaries. Sed-imentation rate for each age zone has been ob-tained with linear interpolation (Tables 2 and 4).

Ice-core ages are used for the Vedde (12 000 cal yrBP) and the Saksunarvatn (10 200 cal yr BP) teph-ras (Gro«nvold et al., 1994; Zielinski et al., 1997;Gulliksen et al., 1998). The historical Hekla AD1104 has simply been converted to cal yr BP. Inthe absence of constraining dates, we have chosento let sedimentation rates change at benthic fora-miniferal biozone boundaries. This is based on theassumption that major faunal changes are consid-ered to coincide with changes in sedimentationrates because they re£ect palaeoceanographicchanges in the region (see also Eir|¤ksson et al.,2000a). In the age models, zero age is assumedfor the core tops, based on correlation with cor-responding box cores with intact core tops. Incore HMl07-04, AMS 14C dates from the baseof biozone F4B and the top of F4A indicate anhiatus of almost 700 years between those two bio-zones, coinciding with a level of major faunalchanges (see Eir|¤ksson et al., 2000a).

4. Lithology

The origin of sand-sized tephra particles incores HMl07-05 and HMl07-04 is primarily at-tributed to volcanic events in Iceland (Eir|¤kssonet al., 2000a). The sand fraction constitutes up to40%, but is generally less than 20% of the sedi-ment. The dispersal of volcanic material from Ice-land is considered to be largely wind-driven with

Table 2Age^depth model boundaries for core HM107-05 and sedimentation rates for each linear segment of the model

Top Base Base Sedimentation rate Base, boundary type(cm) (cm) (cal BP) (cm/kyr)

0 18.5 1 560 12 Age zone boundary (AMS date)18.5 50 5 430 8 Age zone boundary (AMS date)50 67 6 430 16 Zone F5F/F5E boundary67 127 10 200 16 Age zone boundary, Saksunarvatn ash127 128 10 220 50 Age zone boundary, Zone F5D/F5E boundary128 212 11 470 336 Age zone boundary, Zone F5C/F5D boundary212 222 12 000 19 Age zone boundary, Vedde Ash222 229.5 12 400 19 Age zone boundary229.5 334 14 130 60 Age zone boundary, Zone F5B/F5C boundary334 366 15 320 27 Age zone boundary, Zone F5A/F5B boundary366 394.3 15 740 64 Core base

The ages of the benthic foraminiferal biozone boundaries (major faunal changes) are indicated. Ages of diatom zone boundariesare entered in the text (see also Fig. 6).

MARMIC 953 2-2-04

K.L. Knudsen et al. /Marine Micropaleontology 50 (2004) 273^305 281

sand-sized and ¢ner grained air-fall tephras set-tling conformably on the shelf sea£oor. Localeruptions may have formed submarine or subae-rial eruption plumes and caused tephra depositionfrom suspension or submarine turbidity currentsacross local basins (Lackschewitz et al., 1994).Mineralogical analyses have revealed a relativelyhigh background of sand-sized reworked tephra.This was attributed to occasional storm rework-ing of local shallow water deposits on the Kol-beinsey Ridge by Eir|¤ksson et al. (2000a). Rockfragments and minerals of ‘crustal’ origin, as wellas rounded and altered volcanic fragments wereinterpreted as IRD. The provenance of the mudfraction has not been examined. Both ice raftingand volcanic processes have undoubtedly contrib-uted to the ¢nes, in addition to benthic organicactivity, redeposition by bottom currents and sus-pension deposition of material carried to the shelfenvironment by eolian and £uvial processes, aswell as organic processes in the water column.

There is a distinct relationship between meangrain size (Eir|¤ksson et al., 2000a) and sortablesilt content, indicating that the bulk of the sedi-ment volume has been deposited by deceleratingbottom currents. During periods of weak bottomcurrents, the sedimentation was dominated by de-position of ¢ne silt and clay from suspension, butduring periods of stronger currents carrying sort-able silt, the mean grain size was pushed up. Onthe freshly split core surfaces and X-ray images ofboth cores, the predominantly silty sediments ap-pear faintly streaky due to grain size variationswithin the silt range. This is consistent with depo-sition from suspension and concomitant weakbottom currents. Distinct bedding is con¢ned totephra layers and rare coarse sandy or pebblyunits. Bioturbation was only observed sporadical-ly as isolated burrow marks.

The pebble content (Figs. 4 and 5) at both sitesdisplays roughly three intensity levels. Below theVedde Ash, the average pebble content amountsto 5 pebbles/cm core, £uctuating between 0^30.The average value decreases to 2, £uctuating be-tween 0^10, in the interval between the Vedde andthe Saksunarvatn tephras, and the average post-Saksunarvatn value is 1 pebble/cm core. However,a marked increase is observed in the uppermost 20T

able

3R

adio

carb

onda

ting

san

dte

phra

mar

kers

inco

reH

MI0

7-04

HM

I07-

04,de

pth

Lab

orat

ory

No.

Mat

eria

l/dat

edte

phra

laye

r14

Cag

eC

alag

e(s)

BP

Cal

V1

sC

alag

e(s)

BP

Cal

V1

sN

13C

(cm

)(B

P)

V1

s(R

=40

0)(B

P)

(R=

800)

(BP)

7Hekla

1104

850

87^8

8A

AR

-420

2M

ollu

sc,Siphonodentalium

lobatum

534

5V

55570

0575

0^5

630

+1.

714

2^14

3A

AR

-504

0M

ollu

sc,Thyasiraequalis

634

0V

70678

0688

0^6

720

0‡16

5^16

6A

AR

-293

2M

ollu

scs,

Thyasiraequalis,Yoldiella

sp.

683

0V

60736

0741

0^7

280

32.

020

7^20

8A

AR

-420

3-2

Mol

lusc

,Siphonodentalium

sp.

770

0V

100

816

0828

0^8

020

+1.

321

1^21

2A

AR

-574

3M

ollu

scs,

Thyasiragouldi,Yoldiella

lenticula

735

0V

90779

0791

0^7

700

0‡26

9^27

0A

AR

-574

5M

ollu

sc,Siphonodentalium

lobatum

879

0V

45940

0^9

110

960

0^9

070

+1.

229

2^29

3A

AR

-338

4M

ollu

sc,cf

.Siphonodentalium

lobatum

907

5V

60979

0984

0^9

550

+1.

030

8^31

3A

AR

-510

3For

amin

ifer

a,to

talbe

nthi

cfa

una

941

5V

6510

290^

1014

010

250^

986

03

1.7

317

Sak

sunarvatnash

10200

362^

366

AA

R-5

194

For

amin

ifer

a,to

talbe

nthi

cfa

una

1054

0V

4511

650

1191

0^11

360

1111

0^10

850

1129

0^10

840

31.

136

9^37

1A

AR

-442

1For

amin

ifer

a,to

talbe

nthi

cfa

una

1110

0V

8012

800^

1265

012

890^

1237

011

930^

1175

012

280^

1167

03

0.3

369^

371

AA

R-4

422

For

amin

ifer

a,Neoglob.pach

.sin.

1098

0V

9012

620^

1238

012

810^

1215

011

670

1193

0^11

360

+0.

1370

Vedde

Ash

12000

371^

373

AA

R-4

424

For

amin

ifer

a,Neoglob.pach

.sin.

1116

0V

9012

820^

1267

012

900^

1263

012

080^

1196

012

340^

1170

0+0.

137

1^37

3A

AR

-442

3For

amin

ifer

a,to

talbe

nthi

cfa

una

1129

0V

110

1289

012

980^

1266

012

330

1280

0^11

760

30.

739

1^39

2.7

AA

R-4

418

For

amin

ifer

a,to

talbe

nthi

cfa

una

1204

0V

8013

470

1380

0^13

400

1314

013

730^

1301

03

0.6

0‡=

not

mea

sure

d(s

tand

ard

valu

eas

sum

ed).

See

also

capt

ion

ofTab

le1.

MARMIC 953 2-2-04

K.L. Knudsen et al. /Marine Micropaleontology 50 (2004) 273^305282

cm of core HM107-04, indicating increased icerafting north of Iceland in the last 2000 years ofthe Holocene.

5. Planktonic foraminifera and stable isotopes

The percentage distributions of selected taxa ofplanktonic foraminifera in cores HM107-05 andHM107-04 are shown on Figs. 8 and 9 togetherwith the stable isotopic records for the planktonicand the benthic foraminifera, as well as the IRD£uxes.

5.1. Core HM107-05

5.1.1. The Lateglacial interval (15 740^11 500 calyr BP)

There is a very high dominance of the arcticsinistrally coiled Neogloboquadrina pachyderma(mostly 95^100% of the total planktonic assem-blage) throughout the Lateglacial part of coreHM107-05 (Fig. 8). Other planktonic species arerare but include some specimens of dextrallycoiled N. pachyderma, speci¢cally in the lower-most part of the core (before about 15 300 calyr BP, i.e. in pre-B=lling time) and on both sidesof the Vedde tephra horizon. This corresponds tointervals of increased planktonic foraminiferal£uxes, which would also indicate temporalchanges in the water mass distribution in thearea. An increased input of Atlantic water fromthe Irminger Current during these intervals has

been assumed to be the explanation for thesechanges (see also Eir|¤ksson et al., 2000a).

A very light planktonic oxygen isotope signal inthe bottom part of the core presumably indicatesa marked freshwater input to the surface waters inthe area, which appears to continue into theB=lling^Aller=d, almost until 14 000 cal yr BP.Similar low isotopic values in the benthic curvesuggest a mixing of surface and bottom waters.These coupled signals are most likely due to sea-ice freezing processes (Veum et al., 1992; Vidal etal., 1998; Dokken and Jansen, 1999). Therefore, itis concluded on the basis of the oxygen isotopevalues that bottom water produced by brine for-mation probably in£uenced the core site, speci¢-cally before 15 300 cal yr BP, but presumably alsoduring the early part of the B=lling^Aller=d com-plex and during the Younger Dryas. Similaritiesbetween the planktonic and benthic carbon iso-tope records also indicate active exchange be-tween surface and bottom waters during most ofthe time span covered by the cores.

A strong in£uence of the benthic species Cassi-dulina neoteretis, combined with a high percentageof the high-salinity group Miliolida in pre-B=llingdeposits on the North Icelandic shelf, has previ-ously been related entirely to a temporary strongin£uence of Atlantic waters brought to the areaby the Irminger Current (Eir|¤ksson et al., 2000a).A similarly high in£uence of C. neoteretis has,however, also been registered during Heinrichevents (and related to brine formation) in theNorth Atlantic by Lassen et al. (in press). The

Table 4Age^depth model boundaries for core HMI07-04. See caption of Table 2

Top Base Base Sedimentation rate Base, boundary type(cm) (cm) (cal BP) (cm/kyr)

0 6.5 850 8 Age zone boundary, Hekla 1104 tephra6.5 87.5 5700 17 Age zone boundary (AMS date)87.5 117 6300 49 Zone F4C/F4D boundary117 207.5 8160 49 Age zone boundary (AMS date)207.5 317 10200 54 Age zone boundary, Saksunarvatn ash317 322 10290 56 Zone F4B/F4C boundary322 366 11110 54 Top of hiatus, base Zone F4B366 366 11800 Base of hiatus, top Zone F4A366 370 12000 20 Age zone boundary, Vedde Ash370 392.7 13140 20 Core base

See also Fig. 7.

MARMIC 953 2-2-04

K.L. Knudsen et al. /Marine Micropaleontology 50 (2004) 273^305 283

present isotopic data, combined with the faunalindication and a high biological production, re-£ected both in the planktonic and the benthic fo-raminiferal £uxes, suggest that the area was prob-ably in£uenced both by freezing processes and ahigh input of Irminger Current water in the sub-

Fig. 6. Age^depth diagram for core HM107-05. Tephramarkers and AMS 14C datings, based on molluscs and fora-minifera (Table 1), are used as a basis for the reconstructionof linear segments in the age^depth model (see also text andTable 2). The age calibration is based on Stuiver et al.(1998a,b). A standard reservoir correction of about 400 years(vR= 0) is built into this model for the marine samplesyounger than 10 000 14C years, while a reservoir correctionof about 800 years (vR=400) has been used for older sam-ples ( V one standard deviation is indicated by error bars).An age of 12 000 cal yr BP is applied for the Vedde Ash and10 200 cal yr BP for the Saksunarvatn ash. The ages ofbenthic foraminiferal zone boundaries are given in Table 2.

Fig. 7. Age^depth diagram for core HM107-04. Tephramarkers, including the historical Hekla AD 1104, and AMS14C datings, based on molluscs and foraminifera (Table 3),are used as a basis for the reconstruction of linear segmentsin the age^depth model (see also text and Table 4). An hia-tus between 11 800 and 11 110 cal yr BP is shaded (see alsotext). The ages of benthic foraminiferal zone boundaries aregiven in Table 4. See also caption of Fig. 6.

MARMIC 953 2-2-04

K.L. Knudsen et al. /Marine Micropaleontology 50 (2004) 273^305284

Fig. 8. Percentage distribution (calculated from total planktonic contents) of selected species and species groups of planktonic foraminifera, as well as the plankton-ic foraminiferal £ux (number/cm2/yr) in core HM107-05, shown on a calibrated age scale. Other subarctic planktonic species include the following species: Turboro-talita quinqueloba, Globigerina bulloides and Globigerinita glutinata. The N

18O (corrected for ice volume e¡ect) and N13C curves are shown for the planktonic Neoglo-

boquadrina pachyderma sinistral (NPS) and for the benthic Islandiella norcrossi (I.n.) and Cassidulina neoteretis (C.n.). C. neoteretis has been normalised toI. norcrossi values for the benthic N

18O curve. The di¡erence between benthic and planktonic N18O values is calculated by interpolation of ¢xed 100 years distances

between data points. The IRD £ux (sand-sized) is shown as number/cm2/yr. The ages of the tephra markers (Vedde and Saksunarvatn) are shown with stippledlines. For comparison, the benthic foraminiferal zonation and the diatom zonation are also shown. For stratigraphical correlation, see Fig. 15.

MA

RM

IC953

2-2-04

K.L.Knudsen

etal./M

arineMicropaleontology

50(2004)

273^305285

Fig. 9. Environmental data for core HM107-04. The ages of the tephra markers (Vedde, Saksunarvatn and Hekla 1104) are shown with stippled lines. The hiatusbetween 11 800 and 11 110 cal yr BP is shaded. For stratigraphical correlation, see Fig. 15. See also caption of Fig. 8.

MA

RM

IC953

2-2-04

K.L.Knudsen

etal./M

arineMicropaleontology

50(2004)

273^305286

surface layers. The marked di¡erence in faunalindication of the surface water masses comparedto the bottom waters strongly suggests that theIrminger water in£uenced the subsurface and bot-tom waters much more than the surface waters.

It is interesting to notice that a similar, but lesspronounced decrease in oxygen isotope valuesalso occurs just before the Vedde tephra horizonand continues to the end of the Younger Dryas(11 500 cal yr BP). This coincides with anothermarked maximum of Cassidulina neoteretis, last-ing for a maximum of 800 years according to thepresent age model, a value which is presumablytoo high due to increased tephra input, in£uenc-ing the age model during this time period (Fig.12). This interval presumably represents anothersea surface freshwater peak, combined with brineformation as well as increased Atlantic water in-£ow to the subsurface and bottom waters.

5.1.2. The Holocene interval (11 500 cal yr BP tothe present)

Only very few planktonic foraminifera werefound in the Preboreal interval, but a low densityfauna is present from about 10 000 cal yr BP tothe present time. Since rather few samples wereanalysed in this interval, the results only give anoverview of the general trend in climate change.The percentage of sinistrally coiled Neogloboqua-drina pachyderma is high, generally above 90% ofthe fauna, but there is a clear indication of warm-er surface water conditions just after 10 000 cal yrBP. After 7000^6000 cal yr BP, the percentages ofsinistrally coiled N. pachyderma increase to morethan 95%. This shows that Arctic or Polar watermasses become predominant again (see Johanne-sen et al., 1994).

The planktonic N18O values £uctuate around

3x through the entire interval from 10 000 calyr BP to the present, while there is a clear increas-ing trend in the benthic values from 4 to about4.4x between 10 000 and around 5000 cal yr BP,and the benthic values £uctuate around that levelthroughout the late Holocene. The increasing dif-ference between benthic and planktonic values inthe late Holocene indicates an increasing strati¢-cation of the water column.

There is also a Holocene trend in the carbon

isotope values of core HM107-05, with low valuesin the early Holocene and more positive values inthe mid- to late Holocene intervals. This may bethe general interglacial trend due to changes in thecarbon cycle, which occurs in many cores in theNordic Seas both in the Holocene and in MarineIsotope Substage 5e (e.g. Vogelsang, 1990; Fron-val and Jansen, 1997), but the resemblance withthe trend towards more cold water dominance inthe planktonic foraminiferal fauna (Fig. 8) wouldindicate that it may be due to a diminishing in£u-ence of Atlantic water and stronger dominance ofthe Arctic water domain with its higher N

13C val-ues (Johannesen et al., 1994). The somewhat re-duced planktonic values towards the top of thecores may be due to a further in£uence of Polarwaters, which have lower N13C values. This is cor-roborated by the concurrent highest values of thepolar foraminiferal taxon Neogloboquadrina pa-chyderma sinistral and indication of increasedIRD £ux.

5.2. Core HM107-04

5.2.1. The Lateglacial interval (13 140^11 800 calyr BP, i.e. base hiatus)

The Lateglacial interval in core HM107-04 isrepresented by relatively few samples, coveringless than 1400 years. The percentage of Neoglobo-quadrina pachyderma sinistral is very high (95^100% of the total planktonic assemblage; Fig.9). There is an increase both in planktonic andbenthic foraminiferal £uxes just below the VeddeAsh (see also Eir|¤ksson et al., 2000a), and thevalues continue at a relatively high level afterthat marker horizon. A decrease in both plank-tonic and benthic oxygen isotope values is alsoseen around the level of the Vedde Ash.

Even though the data from this interval aresparse, they strongly support the results fromcore HM107-05. The benthic faunal compositionshows a pattern similar to the one described forcore HM107-05, with an interval of high percen-tages of Cassidulina neoteretis on both sides of theVedde Ash. In core HM107-04 this is associatedwith a maximum in the high-salinity group Mil-iolida. The isotope records and the faunal indica-tions suggest that the surface waters at this site

MARMIC 953 2-2-04

K.L. Knudsen et al. /Marine Micropaleontology 50 (2004) 273^305 287

were probably also in£uenced by freezing process-es during part of the Younger Dryas, and thatthere was an increased Atlantic water in£ux tothe subsurface and bottom waters.

5.2.2. The Holocene interval (11 110 cal yr BP,i.e. top hiatus to the present)

Only very few samples from the Preboreal (ataround 11 000 cal yr BP) contained enough plank-tonic foraminifera for interpretation. However,the high percentages of sinistrally coiled Neoglo-boquadrina pachyderma show that Arctic or Polarsurface waters did predominate. The planktonicoxygen isotope values decrease from 3 to 2.6xwithin this short interval. The benthic oxygen iso-tope values £uctuate through the Preboreal, butwith a general increase from ca. 3.5x at 11 000cal yr BP to ca. 4.0x around the level of theSaksunarvatn ash (10 200 cal yr BP). The benthicforaminiferal assemblages, however, indicate in-creasing temperatures through the Preboreal in-terval (Eir|¤ksson et al., 2000a).

The planktonic foraminiferal assemblages showthat the highest surface water temperatures werereached between 10 200 and 7000 cal yr BP at thiscore site, but with a generally decreasing trend.The percentages of sinistrally coiled Neogloboqua-drina pachyderma increased from less than 50 atthe beginning to 80^90% towards the end of theHolocene climatic optimum. The relatively warmperiod was, however, interrupted by a markedcooling between ca. 8600 and 8000 ca1 yr BP,when percentages of sinistrally coiled N. pachyder-ma increased to a maximum of 95%. This corre-sponds to a temperature decrease of about 3‡C(see Johannesen et al., 1994). After a short periodwith relatively high in£uence of subpolar speciesbetween about 5500 and 4500 cal yr BP, theamount of sinistrally coiled N. pachyderma re-mains constantly high through the late Holocene,indicating the predominance of Arctic or Polarsurface waters in the area.

The planktonic oxygen isotope values £uctuate,but display an increasing trend through the Ho-locene climatic optimum. Values as low as 2.1^2.2x are reached during the cooling event be-tween 8600 and 8000 cal yr BP. These low valuespresumably indicate the in£uence of low-salinity

surface waters in the area. There is a clear de-crease in planktonic oxygen isotope valuesthrough the last 5000 cal years.

Benthic oxygen isotope values display a slightgeneral increase through the Holocene, indicatinga minor cooling trend of about 1‡C. There is aclear increase from the Preboreal to the remainingHolocene, which would indicate that brine waterin£uences the oxygen isotope signal until 10 000cal yr BP. The increase in the benthic to plank-tonic di¡erence at the same time would tend tounderscore this conclusion.

As observed for core HMl07-05, reduced plank-tonic N

13C values towards the top of this core maybe due to in£uence of Polar waters with its lowerN

13C values, an assumption which is supported bythe high values of the polar sinistrally coiled formof Neogloboquadrina pachyderma.

6. Diatoms

A recent study of the modern distribution ofdiatoms in surface sediments around Iceland dem-onstrated a close relationship between diatom as-semblages and the modern distribution of watermasses (Jiang et al., 2001). These results, coupledwith previous diatom research in the North At-lantic form the basis of the environmental inter-pretation of the diatom results presented in Figs.10 and 11, which show the percentage distributionof the most common diatom taxa in coresHMl07-05 and HMl07-04. A brief summary ofthe most important environmental indicators isgiven below as it forms the background for theinterpretation of the diatom results.

A strong in£uence of the Irminger Current onthe North Icelandic shelf is indicated by Thalas-siosira oestrupii, which is the main diatom speciesof the Atlantic assemblage in the modern warmNorth Atlantic water mass with salinities higherthan 35.3 and sea surface temperatures above 3‡C(KocS Karpuz and Schrader, 1990). In modern as-semblages around Iceland, it is mainly found inthe surface sediments south and west of Iceland,where the warm Irminger Current has a strongimpact today (Jiang et al., 2001). The presenceof the Irminger Current is also indicated by Tha-

MARMIC 953 2-2-04

K.L. Knudsen et al. /Marine Micropaleontology 50 (2004) 273^305288

lassionema nitzschioides and Paralia sulcata, whichconstitute the main species of the Norwegian^At-lantic Current assemblage in the Greenland^Ice-land^Norwegian seas (KocS Karpuz and Schrader,1990). Both species are associated with the Atlan-tic waters £owing into the Skagerrak^Kattegat inthe eastern North Atlantic (Jiang, 1996). They aremainly found in the modern surface sedimentssouth and west of Iceland, where the IrmingerCurrent has a strong impact (Jiang et al., 2001).

One of the most important taxa in the cold-water diatom assemblage associated with thecold, low-salinity East Icelandic Current is Tha-lassiosira antarctica resting spores (Jiang et al.,2001). It is one of the main taxa of the Arctic

water assemblage, which at present coincides geo-graphically with the seasonally sea-ice coveredIceland Plateau with salinity between 34.7 and34.9 and temperatures ranging from freezing to8‡C (KocS Karpuz and Jansen, 1992).

Sea-ice conditions are indicated by Fragilariop-sis cylindrus, which is bipolar in distribution, andF. oceanica, which is an arctic and/or sea-ice spe-cies (Hasle and Syvertsen, 1997). Both species arethe main components of the sea-ice diatom assem-blage in surface sediments from the sea£ooraround Iceland (Jiang et al., 2001). An assem-blage containing F. cylindrus, F. oceanica andThalassiosira nordenskioeldii was considered tobe indicator of the spring bloom associated with

Fig. 10. Diatom distribution in core HM107-05 on a calibrated time scale. Frequencies are shown as percentages of the total as-semblage (excluding Chaetoceros resting spores). Ages of zone boundaries are given in the text. For comparison, the benthic fora-miniferal zonation is shown as well (ages of zone boundaries in Table 2). For stratigraphical correlation, see Fig. 15. The ages ofthe tephra markers (Vedde and Saksunarvatn) are shown with stippled lines.

MARMIC 953 2-2-04

K.L. Knudsen et al. /Marine Micropaleontology 50 (2004) 273^305 289

melting ice in modern Bering Sea assemblages(Sancetta, 1981).

Mixed water masses are indicated by Thalassio-sira nordenskioeldii and Thalassiothrix longissima.Thalassiosira nordenskioeldii is characteristic ofmodern diatom assemblages in areas in£uencedby a mixing of the warm Irminger Current andthe cold waters of the East Icelandic Currentnorth of Iceland. Thalassiothrix longissima isabundant (as the main species of the taxon Tha-lassiothrix spp.) in the modern surface sedimentswest of Iceland (Jiang et al., 2001). It is also as-sociated with the West Greenland Current (DeSe¤ve, 1999), and it is very common in the NorthAtlantic, extending to the northern Arctic (Hus-tedt, 1930^1960; Hendey, 1964). It is the maincomponent of the Arctic^Norwegian waters mix-ing assemblage, which coincides with areas ofmixture between the Arctic and the Atlanticwaters (KocS Karpuz and Schrader, 1990).

Strati¢ed water masses are inferred by abun-dant Thalassiosira sp. resting spores. The taxono-my of Thalassiosira sp. resting spores is unclear(Hasle, pers. commun.). It was only found insmall amounts in a few modern surface samplesnorth of Iceland (Jiang et al., 2001). O¡ NorthIceland it only occurs in the interval from about11 500^9500 cal yr BP, where it is usually one ofthe dominant taxa. It was also recorded (as Tha-lassiosira cf. scotia resting spore, which is mor-phologically close to the present one) from Youn-ger Dryas sediments in the Skagerrak^Kattegat inthe eastern North Atlantic, where the watermasses were highly strati¢ed with high-salinitybottom waters and more brackish surface waters(Jiang et al., 1997). Therefore, we suggest thatabundance of Thalassiosira sp. resting sporesnorth of Iceland may re£ect a local environmentof highly strati¢ed water masses rather than lowsea surface temperature.

6.1. The Lateglacial and the Pleistocene^Holocenetransition

6.1.1. Zone D5A, HM107-05 (15 740^15 000 calyr BP)

This zone is dominated by Fragilariopsis oceani-ca and Thalassiothrix longissima together with

Fragilariopsis cylindrus, Thalassiosira antarcticaresting spores and Actinocyclus curvatulus (Fig.10) indicating periodic seasonal ice cover in theoceanographic mixing zone between the cold EastGreenland Current and the Irminger Current. Amixing of the two water masses in the area isshown by the co-occurrence of sea-ice speciesand the relatively warm species T. longissima.

6.1.2. Zone D5B, HM107-05 (15 000^13 800 calyr BP)

There is a rapid decrease in Fragilariopsis cy-lindrus and Thalassiothrix longissima in this zone,and Thalassiosira antarctica resting spores becomethe most important component (Fig. 10). Thischange in assemblage suggests that a majorchange in the current pattern occurred at around15 000 cal yr BP. The East Icelandic Current hada major in£uence on the area during the timeinterval represented by zone D5B.

6.1.3. Zone D5C, HM107-05 (13 800^11 150 calyr BP); Zone D4A, HM107-04 (13 140^11 800 calyr BP, i.e. base hiatus)

These zones are characterised by abundant Fra-gilariopsis cylindrus and Thalassiosira norden-skioeldii with a rapid decline of Thalassiosira ant-arctica resting spores (Figs. 10 and 11), but theyalso contain a few relatively warm taxa such asThalassiothrix longissima, Odontella aurita andThalassionema nitzschioides. The composition ofthe diatom assemblages suggests that watermasses from the cold East Greenland and EastIcelandic currents still had a strong in£uence onthe North Icelandic shelf. The assemblages, how-ever, also reveal an in£ux of the warm IrmingerCurrent and mixing of the cold and warm watermasses in the region. The more westerly core siteHM107-04 was more in£uenced by the warm cur-rents than HMI07-05 during the overlapping pe-riod.

6.2. The Holocene

6.2.1. Zone D5D, HM107-05 (11 150^10 100 calyr BP); Zone D4B, HM107-04 (11 110, i.e. tophiatus^9600 cal yr BP)

These early Holocene zones are characterised

MARMIC 953 2-2-04

K.L. Knudsen et al. /Marine Micropaleontology 50 (2004) 273^305290

by abundant Thalassiosira sp. resting spores (Figs.10 and 11), which may re£ect an environment ofhighly strati¢ed water masses. The lower part ofzone D5D contains relatively abundant sea-icespecies, while the assemblages in the upper partindicate higher water temperatures. A rapid de-crease in the sea-ice species Fragilariopsis cylin-drus through the zone suggests a climatic amelio-ration since around 11 100 cal yr BP, althoughthere were still abundant cold water diatomssuch as Thalassiosira antarctica and Thalassiosirasp. resting spores. Fragilariopsis oceanica andT. antarctica resting spores peak at around10 700 cal yr BP in both cores, suggesting a short

period of sea surface cooling, while the succeedingappearance of the two warm species Thalassione-ma nitzschioides and Paralia sulcata at about10 400 cal yr BP in zone D5D indicates the ¢rstmarked in£uence of the warm Irminger Current inthe area at about 200 years before the depositionof the Saksunarvatn ash. The climatic warmingset in slightly earlier at the more westerly siteHMI07-04. In general, there is a lower contentof the typical sea-ice species F. cylindrus andF. oceanica in core HMI07-04 than in coreHMI07-05 in the early Holocene, suggesting thatslightly warmer water masses prevailed at core siteHMI07-04.

Fig. 11. Diatom distribution in core HM107-04. For ages of the benthic foraminiferal zone boundaries, see Table 4. The ages ofthe tephra markers (Vedde, Saksunarvatn and Hekla 1104) are shown with stippled lines. For stratigraphical correlation, see Fig.15. See also caption of Fig. 10.

MARMIC 953 2-2-04

K.L. Knudsen et al. /Marine Micropaleontology 50 (2004) 273^305 291

6.2.2. Zone D5E, HM107-05 (10 100^5100 cal yrBP); Zone D4C, HM107-04 (9600^6000 cal yrBP)

The diatom assemblages in these zones arecomposed mainly of warm species, and cold waterdiatoms as well as sea-ice diatoms reach their low-est frequency levels (Figs. 10 and 11). These as-semblages document the strongest in£uence of theIrminger Current for the entire sedimentary rec-ord. However, there appear to be several shortcold phases, indicated by a slight increase insea-ice species and cold species, during this periodof time. A peak at around 8700 cal yr BP in coreHM107-05 is especially pronounced, where a con-temporaneous increase in Fragilariopsis cylindrus,Fragilariopsis oceanica and Thalassiosira antarcti-ca resting spores is seen.

6.2.3. Zone D5F, HM107-05 (5100 cal yr BP tothe present); Zones D4D, D4E and D4F, HM107-04 (6000^3150, 3150^1300 and 1300 cal yr BP tothe present)

These late Holocene zones are distinguishedfrom the previous interval by a decrease or analmost total disappearance of the Atlantic waterindicator Thalassiosira oestrupii (Figs. 10 and 11).A slight increase is observed in the sea-ice speciesFragilariopsis oceanica as well as the arctic Tha-lassiosira antarctica resting spores. These assem-blages show that the in£uence of the warm Ir-minger Current had decreased compared to theprevious zone, and that a climatic cooling hadoccurred.

7. Palaeoceanographic changes

The oceanographic circulation north of Icelandwas di¡erent from the present oceanic systemboth during the Lateglacial and the Preboreal(Eir|¤ksson et al., 2000a). During certain intervalsof the Lateglacial, Atlantic water, presumably re-lated to the strength of the Irminger Current, ap-pears to have had an even stronger in£uence onthe North Icelandic shelf than at any time duringthe Holocene. The present-day oceanographiccurrent system in the area appears to have beenestablished at about 10 200 cal yr BP (e.g. Eir|¤ks-

son et al., 2000a,b; Knudsen and Eir|¤ksson,2002).

7.1. The Lateglacial (15 740^11 500 cal yr BP)

In general, the data show that Polar or Arcticsurface water masses were dominant on the NorthIcelandic shelf throughout the Lateglacial (Fig.12). This is re£ected by the high dominance ofNeogloboquadrina pachyderma sinistral (s 95%of the total planktonic assemblage) and by ahigh amount of the arctic diatom Thalassiosiraantarctica resting spores as well as the diatomsea-ice species Fragilariopsis cylindrus and F. oce-anica (see also Figs. 10 and 11). In some intervals,there is a clear correlation between the amount ofdiatom sea-ice species and the IRD concentration,i.e. at the base of the record and during the Youn-ger Dryas, but in other intervals they show oppo-site trends. This is probably due to the fact thatthe sea-ice may have di¡erent origin.

A surface freshwater pulse in this shelf area isre£ected by extremely low planktonic oxygen iso-tope values (NPS) at the base of the sequence(Figs. 8 and 12). This is especially pronouncedbetween 15 740 and 15 300 cal yr BP, but contin-ues until about 14 300 cal yr BP. Low benthicoxygen isotope values show that brine formationseems to have been active during the same periodof time (Fig. 8). One could argue that the lowbenthic and planktonic N

18O values were due towarm subsurface water £owing into the regionnorth of Iceland in an anti-estuarine circulationmode. Such arguments were proposed by Sarnt-hein et al. (1995) and Rasmussen et al. (1996,1997). However, if one calculates the ice volumecorrected amplitude, the 2x benthic N

18O low-ering at 15 700 relative to 15 000 cal yr BP wouldin this scenario imply that the bottom waters ofthe site were 8‡C warmer at 15 700 than at 15 000cal yr BP, using a relationship of 0.26x per de-gree C (Shackleton, 1974). Such a warm signal isvery unlikely. The brine formation interpretationis also supported by the corresponding negativeplanktonic N

18O values occurring together with astrong dominance of polar planktonic foraminif-era. In£owing warm intermediate waters can thusbe ruled out as a sole explanation for the negative

MARMIC 953 2-2-04

K.L. Knudsen et al. /Marine Micropaleontology 50 (2004) 273^305292

oxygen isotope values, and the brine mechanism isconsidered to be the only plausible explanation.

The interval between 15 740 and 15 300 cal yrBP is, however, also characterised by a highamount of benthic Atlantic water indicatorssuch as the opportunistic species Cassidulina neo-teretis and Alabaminella weddellensis together withMiliolida, and there is a clear sedimentologicalindication of increased bottom water velocities(see also Eir|¤ksson et al., 2000a). This is in accor-dance with the observation by Hayward et al.(2002), who found that A. weddellensis is con-nected to areas of high mud content receiving alateral £ux of organic matter, i.e. regions withrelatively strong bottom currents. Thus, there isindication of a pronounced in£uence of high-sa-linity Atlantic waters from the Irminger Currentat the sea£oor. A mixing of water masses in thearea is re£ected by high £uxes of both planktonicand benthic foraminifera and by the co-occur-rence of diatom sea-ice species and the relativelywarm diatom species Thalassiothrix longissima(Figs. 10 and 12).

After a period of generally cold conditions,with dominance of the East Icelandic Currentbut with some in£uence of the Irminger Currentbetween 15 300^12 400 cal yr BP (see also Eir|¤ks-son et al., 2000a), a major oceanic change oc-curred a few hundred years before the depositionof the Vedde Ash, prevailing until about 11 500cal yr BP. A marked increase is seen in the pro-duction of planktonic foraminifera at both sites(Figs. 8, 9 and 12), and on the western side ofthe Kolbeinsey Ridge, the benthic productionalso went up (HMl07-04). The percentages ofthe benthic Atlantic water indicator Cassidulinaneoteretis increase considerably in this intervalaround the Vedde Ash (Fig. 12). In addition,there is a decrease in oxygen isotope values bothfor the planktonic (NPS) and for the benthic fo-raminifera. This presumably indicates a return toa similar environment as seen before 15 300 cal yrBP, but less pronounced, i.e. a melt water pulse,brine formation and faunal indication of mixingof the cold surface waters with the underlyingAltantic waters. Relatively strong bottom currentswere also indicated by the sortable silt percentagein this interval (see also Eir|¤ksson et al., 2000a).

The diatoms indicate continuous arctic surfacewater conditions through the Lateglacial withlong periods of seasonal sea-ice, re£ected by thedominance of sea-ice species and arctic diatoms.Similar conditions continued well into the Prebor-eal. Furthermore, abundance of Thalassiosira sp.resting spores around the Vedde Ash horizon cor-responds closely to the change in foraminiferalassemblages and the change to light oxygen iso-tope values mentioned above and continues toform a considerable part of the diatom assem-blages until about 9500 cal yr BP, i.e. well intothe Holocene, suggesting local strati¢cation of thewater masses.

7.2. The Holocene (11 500 cal yr BP to thepresent)

The environmental conditions during the Pre-boreal north of Iceland appear to have beenvery di¡erent from those of the remaining Holo-cene. The amount of planktonic foraminifera wasvery restricted. Almost all specimens belong to thearctic water indicator Neogloboquadrina pachyder-ma sinistral, but the numbers were usually toosmall for quantitative analyses. There was an in-crease in benthic foraminiferal £ux, however, al-ready at the transition to the Holocene (Fig. 12),with assemblages indicating very cold bottomwater conditions, presumably in£uenced by theNSDW (see Eir|¤ksson et al., 2000a). A benthicfaunal change to more diverse assemblages, grad-ually containing more high-salinity species, re-£ects a gradual change to an increased in£uenceof Atlantic waters brought to the sea£oor with theIrminger Current. This change corresponds to agradual change to heavier benthic oxygen isotopicvalues through the Preboreal (Fig. 8).

The continued high amount of diatom sea-icespecies and the abundance of Thalassiosira sp.resting spores indicate cold freshwater in£uencedsurface water conditions and a strong strati¢ca-tion during the Preboreal. This coincides with ahigh £ux of IRD throughout the Preboreal. Slightclimatic amelioration after about 11 100 cal yr BPwas indicated by the ¢rst appearance of diatomspecies associated with Atlantic water from about10 400 cal yr BP (Figs. 10 and 11). Planktonic

MARMIC 953 2-2-04

K.L. Knudsen et al. /Marine Micropaleontology 50 (2004) 273^305 293

Fig. 12. Data from the time interval between 15 800 (bottom of core) and 11 000 cal yr BP in core HM107-05. The following parameters are shown on a calibratedtime scale: Neogloboquadrina pachyderma sinistral (calculated as percentage of total planktonic contents), planktonic foraminiferal £ux (number/cm2/yr), the N

18Oof N. pachyderma sinistral (NPS, corrected for the ice volume e¡ect), the percentages of the benthic taxa Cassidulina neoteretis and Miliolida (calculated as percen-tages of total benthic contents), the benthic foraminiferal £ux (number/cm2/yr), the percentage of sea-ice diatoms (for species, see text), IRD £ux (number/cm2/yr),and sortable silt content. For comparison, the benthic foraminiferal zonation and the diatom zonation are shown. For stratigraphical correlation, see Fig. 15.

MA

RM

IC953

2-2-04

K.L.Knudsen

etal./M

arineMicropaleontology

50(2004)

273^305294

Fig. 13. Data from the time interval between 9000 and 7500 cal yr BP in core HM107-04 (within the foraminiferal zone F4C). The following parameters are shownon a calibrated time scale: Neogloboquadrina pachyderma sinistral, N. pachyderma dextral (calculated as percentages of total planktonic contents) and other plank-tonic species, including the following subpolar species: Turborotalita quinqueloba, Globigerina bulloides and Globigerinita glutinata, as well as planktonic foraminifer-al £ux (number/cm2/yr), the N

18O records (corrected for the ice volume e¡ect) of the planktonic N. pachyderma sinistral (NPS) and the benthic Islandiella norcrossi(I.n.), the N

13C of the planktonic N. pachyderma sinistral, the percentage of sea-ice diatoms (for species, see text), the IRD £ux (number/cm2/yr) values and the sort-able silt content.

MA

RM

IC953

2-2-04

K.L.Knudsen

etal./M

arineMicropaleontology

50(2004)

273^305295

foraminifera return to the area shortly after that(at 10 200 cal yr BP; Fig. 9), for the ¢rst time witha high amount of subarctic species (6 50% Neo-globoquadrina pachyderma sinistral).

The change in foraminiferal and diatom assem-blages and a sharp decrease in IRD £ux shortlybefore the deposition of the Saksunarvatn ash at10 200 cal yr BP (Figs. 9^11; see also Eir|¤ksson etal., 2000a) indicates the establishment of the mod-ern oceanic circulation pattern in the region. Atthe same time, increased current velocities at thesea£oor are indicated by a distinct rise in sortablesilt content on both sides of the KolbeinseyRidge. Generally warm, but £uctuating sea sur-face temperatures are re£ected in the foraminifer-al and diatom data through the Holocene climaticoptimum, i.e. until about 7000 cal yr BP. Temper-atures of up to 2^3‡C higher than at present in thearea are indicated by a high amount of subarcticplanktonic foraminifera (from 6 50 up to about80% Neogloboquadrina pachyderma sinistral ; Fig.9) and by maximum numbers of the warm waterdiatom Thalassiosira oestrupii (Figs. 10 and 11) inthis interval. In addition to minor coolings, onemajor period of cooling, corresponding to the8200 cal yr BP event in the Greenland ice cores(i.e. Johnsen et al., 1992), is registered both in theplanktonic foraminiferal data and in the diatomassemblages (Figs. 9, 11 and 13). According to thepercentages of the sinistrally coiled N. pachyder-ma, this cooling of about 3‡C appears to be theculmination of a general cooling since about 9000cal yr BP. The actual cold event appears to havelasted for about 600 years (8600^8000 cal yr BP),according to the age model used here. This gen-eral climatic pattern of the Holocene optimum,including the changes during the interval 8600^8000 cal yr BP, is supported by the planktonicN

13C results, where low values are interpreted asindications of more Atlantic water in£uence,while heavier values are indicative of Arctic waterin£uence (Figs. 9 and 13). The planktonic oxygenisotope record shows generally low values (2.1^2.6x) through the cooling event (Fig. 13). Thevalues especially £uctuate at the beginning andclose to the end of the event, and the extremelylow peaks presumably indicate in£ux of low-salin-ity Polar waters in the area. The two periods of

£uctuations correspond to peaks in diatom sea-icespecies.

The 8600^8000 cal yr BP cooling event, as re-corded at core site HM107-04 west of the Kol-beinsey Ridge, coincides with a period of in-creased sortable silt content (Fig. 14), and thesortable silt mean value is also higher than imme-diately before and after the event (Fig. 15). IRD£ux values are relatively stable before and afterthe event, i.e. at around 10^15 grains/cm2/yr, butduring the event there are several peaks reachingaround 40 grains/cm2/yr. The 8600^8000 cal yr BPcooling event coincides with quite distinct peaksin the £ux values for the two benthic species Ci-bicides lobatulus and C. neoteretis (Fig. 14). TheC. lobatulus increase is probably related to a peri-od of increased bottom currents, with the C. neo-teretis peak indicating the in£uence of Atlanticwater. This suggests a strongly strati¢ed watercolumn in the area during the cooling event.

Fig. 14. The £ux (number/cm2/yr) of the benthic species Cibi-cides lobatulus and Cassidulina neoteretis plotted against agefor core HM107-04. Note prominent £ux peaks at around8200 and 6000 cal yr BP. The standard deviation (sorting inw units) of the total grain size distribution and the sortablesilt mean grain size are also shown.

MARMIC 953 2-2-04

K.L. Knudsen et al. /Marine Micropaleontology 50 (2004) 273^305296

Fig. 15. Comparison of the results from the Lateglacal and Holocene records o¡ North Iceland with the oxygen isotope record of the GRIP ice core (Johnsen etal., 1992, 1995). The Greenland ice-core events (e.g. Bjo«rck et al., 1998) are indicated as well as the chronostratigraphical units of Mangerud et al. (1974). The per-centages of sinistrally coiled Neogloboquadrina pachyderma (NPS) are shown in a combined diagram for the Lateglacial part of HM107-05 and for the Holocenepart of HM107-04, while the oxygen isotope record, the contents of diatom sea-ice species, as well as selected sedimentological parameters are shown for eachcore. The benthic foraminiferal zonation of cores HM107-04 and HM107-05 are included for correlation. Important tephra markers (the Vedde and Saksunarvatntephras) are shown. For details on the late Holocene climatic oscillations, see Eir|¤ksson et al. (2000b), Jiang et al. (2002) and Knudsen and Eir|¤ksson (2002).

MA

RM

IC953

2-2-04

K.L.Knudsen

etal./M

arineMicropaleontology

50(2004)

273^305297

The relatively poor sorting of the sediment at thislevel is probably due to the increased sand con-tent, including both the IRD material and the fo-raminiferal tests.

Both the foraminiferal and the diatom datashow a general, but £uctuating, cooling sinceabout 7000^6000 cal yr BP. West of the Kolbein-sey Ridge (HM107-04; Fig. 9), a cooling period isre£ected between 6500 and 5500 cal yr BP by theplanktonic foraminiferal record. This coincideswith distinct peaks in the IRD £ux in both coresat around 6000 cal yr BP (Figs. 4 and 5). This isespecially pronounced in core HM107-04, where£ux peaks in the benthic foraminiferal species Ci-bicides lobatulus and C. neoteretis are also notable(Fig. 14). A period of increased strati¢cation ofwater masses is inferred. The similarity betweenthis interval and the 8600^8000 cal yr BP coolingevent is also revealed by the less sorted sediment(increase of sand fraction) and a slightly increasedmean sortable silt size. The amount of sortable siltdecreases markedly after the 6000 cal yr BP cool-ing event. According to the foraminiferal data, thecooling event was followed by a relatively warmperiod between 5500 and 4500 cal yr BP, datawhich are in accordance with the temperature£uctuations recorded from the GISP2 ice core(O’Brien et al., 1995).

The general trend in the Holocene palaeoceano-graphic development of the area appears to becomparable in the sedimentary basins on bothsides of the Kolbeinsey Ridge, even though boththe foraminifera and the diatoms show that thein£uence of the relatively warm Irminger Currentwas higher west of the ridge than on the easternside. At core site HMl07-04, generally weak bot-tom currents are indicated by a low but variablesortable silt content and mean grain size in theHolocene part of the record, re£ecting a primarilydepositional environment. The grain size data andobserved current laminations at site HMl07-05,east of the Kolbeinsey Ridge, indicate a livelierand more e⁄cient circulation regime, accompa-nied by lower sedimentation rates. This di¡erencemay be related to the local topography.

The mean di¡erences in oxygen isotope values(planktonic as well as benthic) to the west and tothe east of the Kolbeinsey Ridge, respectively,

have been calculated for the period in which themodern oceanographic circulation system has pre-vailed in the region, i.e. since 10 200 cal yr BP.The mean planktonic value (NPS) west of theridge was 2.52x (84 measurements), while thecorresponding value east of the ridge was 2.80x(21 measurements). This di¡erence supports thefaunal indication of continuously higher sea sur-face temperatures in the western part, eventhough a di¡erence in salinity may be even moreimportant for the isotopic values. A high in£uenceof the Irminger Current in the western part of thearea would increase the oxygen isotope values dueto higher salinity and, thus, reduce the apparenttemperature di¡erence. The corresponding benthicvalues (Islandiella norcrossi) are 4.11x (99 mea-surements) west of the ridge and 4.19x (32 mea-surements) east of the ridge. Since the benthic fo-raminiferal assemblages clearly indicate highertemperatures in the western basin, these oxygenisotope values must be highly in£uenced by in-creased salinity from the Irminger Current, whichparticularly in£uences the bottom waters to thewest.

The benthic to planktonic N13C gradients are

similar on both sides of the Kolbeinsey Ridge,indicating a similar degree of surface to bottomstrati¢cation. The gradient is also relatively stablethrough the Holocene.

8. Comparison and discussion

The interval from the bottom of the record at15 740 cal yr BP to the abrupt change in forami-niferal and diatom assemblages as well as in bothoxygen and carbon isotope values at about 15 300cal yr BP (13 100 14C years BP) is correlated tothe Greenland Stadial 2 (GS-2), i.e. the pre-B=lling. At the beginning of the B=lling Intersta-dial, a high in£ux of warm and salty water intothe eastern North Atlantic presumably coincidedwith intensi¢ed deep-water formation in the Nor-dic Seas (e.g. Boyle and Keigwin, 1987; Kroon etal., 1997; Austin and Kroon, 2001), a changewhich corresponds to the end of the Heinrich 1event in the North Atlantic. Grousset et al. (2001)dated the end of the Heinrich 1 event to about

MARMIC 953 2-2-04

K.L. Knudsen et al. /Marine Micropaleontology 50 (2004) 273^305298

13 400 14C years BP (15 700 cal yr BP), whileSarnthein et al. (1999) found that the end of theHeinrich 1 melt water and IRD signals occurredbetween 13 000 and 13 400 14C years BP. Accord-ing to the GRIP ice-core record, the age of thetransition between GS-2 and GI-le (the B=llingInterstadial) is considerably younger than re-corded in the North Iceland shelf cores, i.e. at14 700 years (Bjo«rck et al., 1998; Walker et al.,1999). New evidence on very high marine reser-voir ages in the North Atlantic at the end of theHeinrich 1 event (Waelbroeck et al., 2001) mayhelp to explain this di¡erence in dating of thetransition to the B=lling Interstadial. A correla-tion of the North Icelandic shelf records withthe isotopic record of the GRIP ice core is shownin Fig. 15.

The Heinrich 1 melt water event has previouslybeen recorded o¡ North Iceland by Voelker et al.(1998) and by Sarnthein et al. (1999), and a strongpre-B=lling Irminger Current into this area hasalso previously been described by Voelker et al.(1998) and by Eir|¤ksson et al. (2000a). Thepresent study clearly supports the idea of ananti-phase temperature relationship between theNorth Icelandic shelf and the eastern North At-landic during pre-B=lling time (see also Sarntheinet al., 1995). A subsequent cooling north of Ice-land, which is especially pronounced for bottomwaters, corresponds to the onset of the Norwe-gian Sea deep water formation. This would indi-cate that high in£ux of Atlantic waters into theeastern North Atlantic led to a reduced IrmingerCurrent north of Iceland. This weakening mayre£ect an overreaction following a sudden revivalof thermohaline circulation into the eastern Nor-dic Seas after a shutdown state caused by lightsurface melt water.

On the Southwest Icelandic shelf, however, theB=lling Interstadial is characterised by an in£uxof relatively warm Irminger Current foraminiferalspecies (Jennings et al., 2000). Apparently, thenorthern boundary of the GI-l (B=lling^Aller=d)oceanographic warming was to the west andsouth of the North Icelandic shelf.

The transition from the B=lling^Aller=d Inter-stadial (GI-l) to the Younger Dryas Stadial (GS-1) is not clearly detectable in our data from the

cores north of Iceland. This suggests that the tem-perature change north of Iceland was only minorcompared to the abrupt temperature drop re-corded in the eastern North Atlantic at the begin-ning of the Younger Dryas by e.g. KocS Karpuzand Jansen (1992), Ha£idason et al. (1995),Kroon et al. (1997) and Klitgaard-Kristensen etal. (2001). A similar relatively warm YoungerDryas was also described by Bjo«rck et al. (2002)from Lateglacial lake sediments in southernGreenland and by Kuijpers et al. (2003) fromthe Southeast Greenland margin. Foraminiferalassemblages from a core on the Southwest Icelan-dic shelf, however, revealed a marked cooling, atleast during the early part of the Younger Dryas(Jennings et al., 2000).

The assemblages and the oxygen isotope valuesaround the level of the Vedde Ash in the twocores north of Iceland indicate a temporary highin£uence of the Irminger Current as well as a sur-face melt water peak accompanied by brine for-mation in the area. The signals are similar tothose found in the pre-B=lling, but weaker. Anincreased in£ux in boreal benthic foraminiferalspecies during the later part of the Younger Dryasis also reported from the Southwest Icelandicshelf (Jennings et al., 2000).

The abrupt cooling of both surface and bottomwaters north of Iceland at the beginning of thePreboreal (11 500 cal yr BP) occurs at exactlythe same time as a pronounced warming inthe eastern North Atlantic (e.g. Hald and Hagen,1998; Klitgaard-Kristensen et al., 2001; at 11 400and 11 500 cal yr BP, respectively) and increaseddeep water formation in the Nordic Seas. Again,this shows that a strong in£ow of Atlantic waterinto the eastern Atlantic apparently coincides witha reduced Irminger Current around Iceland and astrong in£uence of Norwegian Sea Deep Water onthe benthic assemblages north of Iceland. Amarked sea surface cooling at around 10 700 calyr BP on the North Icelandic shelf might corre-spond to the 11 000 cooling event as known fromthe Greenland ice cores (e.g. Bjo«rck et al., 1998;Lowe et al., 2001).

The general consensus seems to be that the es-tablishment of modern sea surface circulation inthe northern North Atlantic occurred at about

MARMIC 953 2-2-04

K.L. Knudsen et al. /Marine Micropaleontology 50 (2004) 273^305 299

10 000 cal yr BP. For instance, Bauch et al. (2001)concluded that the freshwater input to the centralNordic Seas from melting icebergs completelyceased at that time. Supported by a strong atmo-spheric warming, the improved sea surface condi-tions persisted until about 6000 cal yr BP. Similarages for the Holocene optimal conditions (9000^5000 14C years BP) were recorded from the Nor-dic Seas by KocS Karpuz and Schrader (1990),KocS Karpuz and Jansen (1992) and KocS et al.(1993).

The results from the Nordic Seas correspondgenerally to those from the North Icelandic shelf,which indicate a Holocene climatic optimum be-tween 10 200 and 7000^6000 cal yr BP. Bauch etal. (2001), however, found that the maximumtemperatures were reached at about 7000 cal yrBP in the central Nordic Seas. The planktonicforaminiferal record o¡ North Iceland would sug-gest a general temperature decrease between10 200 and 7000 cal yr BP, while the diatomdata show fairly consistently high temperaturesbetween 10 000 and 7000 cal yr BP. Results of astudy of the coccolith distribution in a near-shorecore north of Iceland (Andrews and Giraudeau,2002) also suggest that the Holocene climatic op-timum in that area occurred between 10 000 and7000^6000 cal yr BP. This was indicated by thein£ux of North Atlantic Drift coccolith species tothe region, with the highest input during the earlypart of that period.

It is interesting to note that a Holocene max-imum in the northward position of the Intertrop-ical Convergence Zone (titanium and iron concen-tration data as proxy for on-shore precipitation)was reached in the Cariaco Basin as early as10 500 cal yr BP (Haug et al., 2001). The ‘thermalmaximum’ lasted until 5400 cal yr BP in that area,but with a decreasing trend since about 7000 calyr BP, indicating similarities between the tropicsand the Iceland region. A palaeotemperature re-construction from the GRIP ice-core record inGreenland, however, indicates optimum temper-atures for the Holocene between about 8600 and4300 yr BP (Johnsen et al., 2001), which is con-siderably later than indicated by sea surface tem-perature reconstructions. The ice-core data sug-gest a temperature drop by 4^8‡C during the

8200 cal yr BP event above central Greenland(Johnsen et al., 2001). Planktonic foraminiferaldata north of Iceland indicate surface water cool-ing of about 3‡C, while transfer function calcula-tion based on diatoms indicate a summer seasurface temperature drop of about 2‡C (Jiang etal., in prep.). The actual background for the 8200cal yr BP cooling event is not fully understood,but it has been suggested by many authors thatfresh melt water input from the Laurentide icesheet may have in£uenced the thermohaline sys-tem at that time (Klitgaard-Kristensen et al.,1998,2001; Bianchi and McCave, 1999; Bauch etal., 2001).

The present Holocene data show a persistentdi¡erence between the surface water masses aswell as the bottom water masses on each side ofthe Kolbeinsey Ridge, which has formed a partialbarrier between di¡erent water masses in the out-er shelf area since the last glaciation. The coreHM107-05 record on the east side is highly in£u-enced by cold water masses, while the coreHM107-04 site west of the ridge has continuouslybeen in£uenced by an input of Atlantic waterfrom the Irminger Current.

In a study of modern distribution of benthicforaminifera, Rytter et al. (2002) demonstratedhow the Kolbeinsey Ridge also acts as an ocean-ographic barrier for the present benthic assem-blages. A similar di¡erence was, however, notseen in the modern diatom distibution in thearea (Jiang et al., 2001). Both the modern forami-niferal data and the modern diatom data showedthe importance of the position of the oceanic Po-lar Front for the assemblage distributions. Thetemporal variations in sea surface temperaturesthrough the Holocene records are supposed tobe caused by £uctuations in the position of themarine Polar Front across the area.

9. Summary and conclusions

The data from the North Icelandic shelf allowus to reconstruct the environmental changes sinceabout 15 740 cal yr BP. Signi¢cant changes arefound through the Lateglacial record, with £uctu-ating in£uence of the Atlantic water masses of the

MARMIC 953 2-2-04

K.L. Knudsen et al. /Marine Micropaleontology 50 (2004) 273^305300

Irminger Current and Polar water masses derivingfrom the East Greenland Current.

The isotopic signal reveals that a marked fresh-water event, presumably with pronounced brineformation, occurred in the area north of Iceland,being pronounced in pre-B=lling time (before15 300 cal yr BP, corresponding to the GreenlandStadial 2, GS-2), but continuing into the B=llingInterstadial (until about 14 300 cal yr BP, i.e. intothe Greenland Interstadial 1, GI-1). The forami-niferal and diatom assemblages re£ect cold seasurface water masses and a mixing with underly-ing Atlantic water masses from the Irminger Cur-rent during this period. The in£uence of the Ir-minger Current was especially strong before15 300 cal yr BP, but there was a continuous,but less strong in£uence during the entireB=lling^Aller=d (GI-1) and the Younger Dryas(GS-1).

A less pronounced melt water event with brineformation, spanning a few hundred years, oc-curred in the Younger Dryas. The Vedde Ashwas deposited as primary tephra within this inter-val. This melt water event also appears to havebeen accompanied by a temporary increase in thein£uence of the Irminger Current in subsurfacewaters on the North Icelandic shelf. There is noindication of signi¢cant cooling in the area duringthe Younger Dryas.

Very low surface water temperatures and astrati¢ed water column characterise the Preboreal,though with gradually ameliorating conditionsboth in the surface waters and in the bottomwaters towards the end of the Preboreal. A pro-nounced cooling event is recorded in the diatomdata at around 10 700 cal yr BP.

The present study supports previous results of astrong palaeo-Irminger Current, which carriedwarm water masses around Iceland during pre-B=lling time (corresponding to the Heinrich 1event) and again, but less pronounced, duringpart of the Younger Dryas. This corresponds toperiods of reduced in£ow of Atlantic waters intothe eastern North Atlantic. Correspondingly,there was a reduced in£ow of Irminger Currentwaters to the North Icelandic shelf during theB=lling Interstadial and during the Preboreal,when, on the other hand, a strong Atlantic Cur-

rent brought warm water masses into the easternNorth Atlantic.

The Lateglacial and Preboreal anti-phase rela-tionship between the environmental indicationsnorth of Iceland compared to those in the easternAtlantic appears to change to an in-phase rela-tionship during the remaining part of the Holo-cene, at least with respect to the long-term andmillennial scale changes.

Planktonic foraminifera show that the modernoceanographic sea surface circulation was estab-lished just prior to deposition of the Saksunarvatntephra at 10 200 cal yr BP. This interpretation issupported by a decrease in benthic N

13C values, adrop in IRD £uxes and the ¢rst immigration ofwarm water diatoms.

Generally warm, but £uctuating, sea surfaceconditions prevailed through the Holocene cli-matic optimum (between 10 200 and 7000^6000cal yr BP). This was interrupted by a major cool-ing event between 8600 and 8000 cal yr BP, cover-ing a time span of about 600 years. The cooling iscorrelated to the 8200 cal yr BP cooling eventrecorded in the Greenland ice-core records. An-other distinct cooling, culminating around 6000cal yr BP, is followed by a relatively long warmperiod between about 5500 and 4500 cal yr BP,which is again succeeded by a general, but £uctu-ating, cooling leading to the present-day condi-tions. The maximum Holocene foraminiferal£uxes are found in the time interval betweenabout 9000 and 6000 cal yr BP.

Through the entire period since 10200 cal yr BP,the planktonic oxygen isotope values remainedlower on the western side of the Kolbeinsey Ridgethan on the eastern side. This supports the faunaland £oral indication of an eastward decreasingin£uence of the Irminger Current in that area.

Only minor changes have occurred in the gen-eral sea surface circulation system around Icelandduring the late Holocene. The time resolution isnot high enough in this study to pick up details inthis part of the record.

Acknowledgements

This paper is a contribution to the PANIS proj-

MARMIC 953 2-2-04

K.L. Knudsen et al. /Marine Micropaleontology 50 (2004) 273^305 301

ect (‘Palaeoenvironments on the North IcelandicShelf’). We would like to thank members of theBIOICE project (‘Benthic Invertebrates of Icelan-dic Waters’) for support during the sampling andRune S=raas and Odd Hansen for mass spectrom-eter operation. Financial support has been ob-tained from the Icelandic Research Council (thePANIS project to J.E.), the Danish Natural Sci-ence Research Council (the PALINAS project toK.L.K. and the Ole R=mer Stipendium to M.-S.S.) and the National Natural Science Foundationof China (Grant No. 40131020 to H.J.).

Appendix A. List of foraminiferal taxa

The original descriptions of the foraminiferaltaxa cited in the text are reported in Ellis andMessina (1949)

Planktonic taxa:Globigerina bulloides d’Orbigny, 1826Globigerinita glutinata (Egger) = Globigerina

glutinata Egger, 1895Neogloboquadrina pachyderma (Ehrenberg) =

Aristerospira pachyderma Ehrenberg, 1861Turborotalita quinqueloba (Natland) = Globi-

gerina quinqueloba Nat1and, 1938Benthic taxa:Cassidulina neoteritis Seidenkrantz, 1995 (previ-

ously often referred to Cassidulina teretis) Tap-pan, 1951

Cibicides lobatulus (Walker and Jacob) = Nau-tilus lobatulus Walker and Jacob, 1798

Islandiella norcrossi (Cushman) = Cassidulinanorcrossi Cushman, 1933

Miliolida

Appendix B. List of diatom taxa

(H and S = Hasle and Syvertsen, 1997)Actinocyclus curvatulus Janish, in A. Schmidt,

1878 (H and S, pl. 21^122)Fragilariopsis cylindrus (Grunow) Krieger, in

Helmcke and Krieger, 1954 (H and S, pp. 302,305)

Fragilariopsis oceanica (Cleve) Hasle, 1965 (Hand S, pp. 299, 305)

Odontella aurita (Lyngbye) Agardh, 1832 (Hand S, p. 239)

Paralia sulcata (Ehrenberg) Cleve, 1873 (H andS, p. 91; Hendey, 1964, p. 73)

Rhizosolenia borealis Sundstro«m, 1986 (H andS, pl. 47)

Rhizosolenia hebetata Bailey, 1856 (H and S, pl.49^150; Hendey, 1964, pp. 150^151)

Thalassionema nitzschioides (Grunow) Meresch-kowsky, 1902 (H and S, pp. 257^262; Hendey,1964, pl. 65)

Thalassiosira antarctica Comber, 1896 (H andS, pp. 66^68, 71)

Thalassiosira nordenskioeldii Cleve, 1873 (H andS, p. 56, 59; Hendey, 1964, p. 85)

Thalassiosira oestrupii (Ostenfeld) Hasle, 1972(H and S, pp. 83, 86)

Thalassiosira paci¢ca Gran and Angst, 1931 (Hand S, pp. 57, 59)

Thalassiothrix longissima Cleve and Grunow,1880 (H and S, pp. 263^267)

References

Andersen, G.J., Heinemeier, J., Nielsen, H.L., Rud, N., John-sen, S., Sveinbjo«rnsdo¤ttir, AŁ .E., Hjartarson, AŁ ., 1989. AMS14C dating on the Fossvogur sediments, Iceland. Radiocar-bon 31, 592^600.

Andrews, J.T., Hardardo¤ttir, J., Helgado¤ttir, G., Jennings,A.E., Geirsdo¤ttir, AŁ ., Sveinbjo«rnsdo¤ttir, AŁ .E., School¢eld,S., Kristiansdo¤ttir, G.B., Smith, L.M., Thors, K., Syvitski,J.P.M., 2000. The N and W Iceland Shelf: Insights into LastGlacial Maximum ice extent and deglaciation based onacoustic stratigraphy and basal radiocarbon AMS dates.Quat. Sci. Rev. 19, 619^631.

Andrews, J.T., Kristja¤nsdo¤ttir, G.B., Geirsdo¤ttir, AŁ ., Hardar-do¤ttir, J., Helgado¤ttir, G., Sveinbjo«rnsdo¤ttir, AŁ .E., Jennings,A.E., Smith, L.M., 2001a. Late Holocene (V5 cal ka) trendsand century-scale variability of N. Iceland marine records:Measures of surface hydrography, productivity, and land/ocean interactions. In: Seidov, D., Maslin, M., Haupt, B.(Eds.), The Oceans and Rapid Climate Change: Past,Present and Future. Geophysical Monograph 126, AmericanGeophysical Union, pp. 69^81.

Andrews, J.T., Helgado¤ttir, G., Geirsdo¤ttir, AŁ ., Jennings, A.E.,2001b. Multicentury-scale records of carbonate (hydro-graphic?) variability on the Northern Iceland Margin overthe last 5000 yrs. Quat. Res. 56, 199^206.

Andrews, J.T., Caseldine, C., Weiner, N., Hatton, J., 2001c.Late Holocene (ca. 4 ka) marine and terrestrial environmen-tal change in Reykjarfjo«rdur, north Iceland: Climate and/orsettlement? J. Quat. Sci. 16, 133^143.

MARMIC 953 2-2-04

K.L. Knudsen et al. /Marine Micropaleontology 50 (2004) 273^305302

Andrews, J.T., Giraudeau, J., 2002. Multi-proxy records show-ing signi¢cant Holocene environmental variability: The in-ner N. Iceland shelf (Hu¤na£o¤ i). Quat. Sci. Rev. 22, 175^193.

Austin, W.E.N., Bard, E., Hunt, J.B., Kroon, D., Peacock,J.D., 1994. The 14C age of the Icelandic Vedde Ash: Impli-cations for Younger Dryas Marine Reservoir Age Correc-tions. Radiocarbon 37, 53^62.

Austin, W.E.N., Kroon, D., 2001. Deep sea ventilation of thenortheastern Atlantic during the last 15,000 years. GlobalPlanet. Change 30, 13^31.

Bard, E., 1990. Calibration of the 14C timescale over the past30,000 years using mass spectrometric U^Th ages from Bar-bados corals. Nature 345, 405^410.

Bard, E., Arnold, M., Mangerud, J., Paterne, M., Labeyrie, L.,Duprat, J., Me¤lieres, M.-A.M., S=nstegaard, E., Duplessy,J.-C., 1994. The North Atlantic atmosphere^sea surface 14Cgradient during the Younger Dryas climatic event. EarthPlanet. Sci. Lett. 126, 275^287.

Bauch, H.A., Erlenkeuser, H., Spielhagen, R.F., Struck, U.,Matthiessen, J., Thiede, J., Heinemeier, J., 2001. A multi-proxy reconstruction of the evolution of deep and surfacewaters in the subarctic Nordic seas over the last 30,000 yr.Quat. Sci. Rev. 20, 659^678.

Bianchi, G.G., McCave, I.N., 1999. Holocene periodicity inNorth Atlantic climate and deep-ocean £ow south of Ice-land. Nature 397, 515^517.

Bjo«rck, S., Walker, M.J.C., Cwynar, L.C., Johnsen, S., Knud-sen, K.-L., Lowe, J.J., Wohlfarth, B., the INTIMATE Mem-bers, 1998. An event stratigraphy for the Last Terminationin the North Atlantic region based on the Greenland ice-core record: A proposal by the INTIMATE Group. J. Quat.Sci. 13, 283^292.

Bjo«rck, S., Bennike, O., Rose¤n, P., Andresen, C.S., Bohncke,S., Kaas, E., Conley, D., 2002. Anomalously mild YoungerDryas summer conditions in southern Greenland. Geology90, 427^430.

Bondevik, S., Mangerud, J., Gulliksen, S., 2001. The marine14C age of the Vedde Ash Bed along the west coast of Nor-way. J. Quat. Sci. 16, 3^7.

Boyle, E.A., Keigwin, L., 1987. North Atlantic Thermohalinecirculation during the last 20,000 years linked to high-lati-tude surface temperatures. Nature 330, 35^40.

De Se¤ve, M.A., 1999. Transfer function between surface sedi-ment diatom assemblages and sea-surface temperature andsalinity of the Labrador Sea. Mar. Micropaleontol. 36, 249^267.

Dokken, T., Jansen, E., 1999. Rapid changes in the mode ofdeep and intermediate water formation linked with atmo-spheric and ice sheet variations during the Last Glacial.Nature 401, 458^461.

Eir|¤ksson, J., Knudsen, K.L., Ha£idason, H., Henriksen, P.,2000a. Late-glacial and Holocene palaeoceanography of theNorth Icelandic shelf. J. Quat. Sci. 15, 23^42.

Eir|¤ksson, J., Knudsen, K.L., Ha£idason, H., Heinemeier, J.,2000b. Chronology of Late Holocene climatic events in thenorthern North Atlantic based on AMS 14C dates and teph-

ra markers from the volcano Hekla, Iceland. J. Quat. Sci.15, 573^580.

Ellis, B.F., Messina, A., 1949. Catalogue of Foraminifera(Supplements, including 1998). Am. Mus. Natl. Hist. Micro-paleontpl. Press, New York.

Fairbanks, R., 1989. A 17,000-year glacio-eustatic sea levelrecord: In£uence of glacial melting rates on the YoungerDryas event and deep-ocean circulation. Nature 342, 637^642.

Feyling-Hanssen, R.W., J=rgensen, J.A., Knudsen, K.L., An-dersen, A.-L.L., 1971. Late Quaternary Foraminifera fromVendsyssel, Denmark and Sandnes, Norway. Bull. Geol.Soc. Den. 21, 67^317.

Fronval, T., Jansen, E., 1997. Eemian and early Weichselian(140^60 ka) paleoceanography and paleoclimate in the Nor-dic seas with comparisons to Holocene conditions. Pale-oceanography 12, 443^462.

Grimm, E., 1987. CONISS: A Fortran 77 program for strati-graphically constrained cluster analysis by the method ofincremental sum of squares. Comput. Geosci. 13, 13^35.

Grousset, F.E., Cortijo, E., Huon, S., Herve¤, L., Richter, T.,Burdlo¡, D., Duprat, J., Weber, O., 2001. Zooming in onHeinrich layers. Paleoceanography 16, 240^259.

Gro«nvold, K., OŁ skarsson, N., Johnsen, S.J., Clausen, H.B.,Hammer, C.U., Bond, C., Bard, E., 1994. Ash layers fromIceland in the Greenland GRIP ice core correlated with oce-anic and land sediments. Earth Planet. Sci. Lett. 135, 149^155.

Gulliksen, S., Birks, H.H., Possnert, G., Mangerud, J., 1998. Acalendar age estimate of the Younger Dryas^Holoceneboundary at Krafikenes, western Norway. Holocene 8, 249^259.

Ha£idason, H., Sejrup, H.P., Kristensen, D.K., Johnsen, S.,1995. Coupled response of the Late Glacial climatic shifts ofnorthwest Europe re£ected in Greenland ice cores: Evidencefrom the northern North Sea. Geology 23, 1059^1062.

Hald, M., Hagen, S., 1998. Early Preboreal cooling in theNordic seas region triggered by meltwater. Geology 26,615^618.

Hansen, B., Msterhus, S., 2000. North Atlantic^Nordic Seasexchanges. Prog. Oceanogr. 45, 109^208.

Hasle, G.R., Syvertsen, E.E., 1997. Marine diatoms. In:Tomas, C.R. (Ed.), Identifying Marine Phytoplankton. Aca-demic Press, CA, pp. 5^385.

Haug, G.H., Hughen, K.A., Sigman, D.M., Peterson, L.C.,Ro«hl, U., 2001. Southward migration of the IntertropicalConvergence Zone through the Holocene. Science 293,1304^1308.

Hayward, B.W., Neil, H., Carter, R., Grenfell, H.R., Hay-ward, J.J., 2002. Factors in£uencing the distribution pat-terns of Recent deep-sea benthic foraminifera, east of NewZealand, Southwest Paci¢c Ocean. Mar. Micropaleontol. 46,139^176.

Hendey, N.I., 1964. An introductory account of the smalleralgae of British coastal waters, Part V. Bacillariophyceae(diatoms). Fish. Invest., Lond. Ser. 4, 1^317.

Henrich, H., Wagner, T., Goldschmidt, P., Michels, K., 1995.

MARMIC 953 2-2-04

K.L. Knudsen et al. /Marine Micropaleontology 50 (2004) 273^305 303

Depositional regimes in the Norwegian^Greenland Sea: Thelast two glacial to interglacial transitions. Geol. Rundsch.84, 28^48.

Hurdle, B.G. (Ed.), 1986. The Nordic Seas. Springer, NewYork, 777 pp.

Hustedt, F., 1930^1960. Die Kieselalgen Deutschlands, Oº ster-reichs und der Schweiz unter Beru«cksichtung der u«brigenLa«nder Europas sowie der angrenzenden Meeresgebiete.In: Rabenhorsts, L. (Ed.), Kryptogamen-Flora 7. Akad.Verlag, Leipzig.

Jennings, A., Syvitski, J., Gerson, L., Gro«nvold, K., Geirsdo¤t-tir, AŁ ., Hardardo¤ttir, J., Andrews, J., Hagen, S., 2000. Chro-nology and paleoenvironments during the late Weichseliandeglaciation of the southwest Iceland shelf. Boreas 29, 167^183.

Jiang, H., 1996. Diatoms from the surface sediments of theSkagerrak and the Kattegat and their relationship to thespatial changes of environmental variables. J. Biogeogr. 23,129^137.

Jiang, H., Bjo«rck, S., Knudsen, K.-L., 1997. A palaeoclimaticand palaeoceanographic record of the last 11 000 14C yearsfrom the Skagerrak^Kattegat, northeastern Atlantic margin.Holocene 7, 301^310.

Jiang, H., Seidenkrantz, M.-S., Knudsen, K.L., Eir|¤ksson, J.,2001. Diatom surface sediment assemblages around Icelandand their relationships to oceanic environmental variables.Mar. Micropaleontol. 41, 73^96.

Jiang, H., Seidenkrantz, M.-S., Knudsen, K.L., Eir|¤ksson, J.,2002. Late-Holocene summer sea surface temperatures basedon a diatom record from the north Icelandic shelf. Holocene12, 137^147.

Johannesen, O.M., 1986. Brief overview of the physical ocean-ography. In: Hurdle, B.G. (Ed.), The Nordic Seas. Springer,New York, pp. 103^127.

Johannesen, T., Jansen, E., Flat=y, A., Ravelo, A.C., 1994.The relationship between surface water masses, oceano-graphic fronts and paleoclimatic proxies in surface sedimentsof the Greenland, Iceland, Norwegian Seas. In: Zahn, R.,Kaminski, M. (Eds.), Carbon Cycling in the Glacial Ocean.NATO ASI Series I, 17, Springer, Berlin, pp. 61^85.

Johnsen, S.J., Clausen, H.B., Dansgaard, W., Fuhrer, K.,Gundestrup, N., Hammer, C.U., Iversen, P., Jouzel, J.,Stau¡er, B., Ste¡ensen, J.P., 1992. Irregular glacial intersta-dials recorded in a new Greenland ice core. Nature 359,311^313.

Johnsen, S., Dahl-Jensen, D., Dansgaard, W., Gundestrup, N.,1995. Greenland palaeotemperatures derived from GRIPbore hole temperature and ice core isotope pro¢les. Tellus47B, 624^629.

Johnsen, S.J., Dahl-Jensen, D., Gundestrup, N., Ste¡ensen,J.P., Clausen, H.B., Miller, H., Masson-Delmotte, V.,Sveinbjo«rnsdo¤ttir, A.E., White, J., 2001. Oxygen isotopeand palaeotemperature records from six Greenland ice-corestations: Camp Century, Dye-3, GRIP, GISP2, Renlandand NorthGRIP. J. Quat. Sci. 16, 299^307.

Klitgaard-Kristensen, D., Sejrup, H.P., Ha£idason, H., John-sen, S., Spurk, M., 1998. A regional 8200 cal. yr BP cooling

event in northwest Europe, induced by ¢nal stages of theLaurentide ice-sheet deglaciation. J. Quat. Sci. 13, 165^169.

Klitgaard-Kristensen, D., Sejrup, H.P., Ha£idason, H., 2001.The last 18 kyr £uctuations in the Norwegian Sea surfaceconditions and implications for the magnitude of climaticchange: Evidence from the North Sea. Paleoceanography16, 455^467.

Knudsen, K.L., 1998. Foraminiferer i Kvartar stratigra¢:Laboratorie- og fremstillingsteknik samt udvalgte eksempler.Geol. Tidsskr. 3, 1^25.

Knudsen, K.L., Eir|¤ksson, J., 2002. Application of tephrochro-nology to the timing and correlation of palaeoceanographicevents recorded in Holocene and Lateglacial shelf sedimentso¡ North Iceland. Mar. Geol. 191, 165^188.

KocS Karpuz, N., Jansen, E., 1992. A high-resolution diatomrecord of the last deglaciation from the SE Norwegian Sea:Documentation of rapid climatic changes. Paleoceanography7, 499^520.

KocS Karpuz, N., Schrader, H., 1990. Surface sediment diatomdistribution and Holocene paleotemperature variations inthe Greenland, Iceland and Norwegian Sea. Paleoceanogra-phy 5, 557^580.

KocS, N., Jansen, E., Ha£idason, H., 1993. Paleoceanographicreconstructions of surface ocean conditions in the Green-land, Iceland and Norwegian seas through the last 14 Kabased on diatoms. Quat. Sci. Rev. 12, 115^140.

Kroon, D., Austin, W.E.N., Chapman, M.R., Ganssen, G.M.,1997. Deglacial surface circulation changes in the northeast-ern Atlantic: Temperature and salinity records o¡ NW Scot-land on a century scale. Paleoceanography 12, 755^763.

Kuijpers, A., Troelstra, S.R., Prins, M.A., Linthout, K., Akh-metzhanov, A., Bouryak, S., Bachmann, M.F., Lassen, S.,Rasmussen, S., Jensen, J.B., 2003. Late Quaternary sedimen-tary processes and ocean circulation changes at the South-east Greenland Margin. Mar. Geol. 195, 109^129.

Lackschewitz, K.S., Dehn, J., Wallrabe-Adams, H.-J., 1994.Volcaniclastic sediments from mid-oceanic KolbeinseyRidge, north of Iceland: Evidence for submarine volcanicfragmentation processes. Geology 22, 975^978.

Lassen, S., Jansen, E., Knudsen, K.L., Kuijpers, A., Kristen-sen, M., Christensen, K., in press. Bottom water variabilityin the Faeroe area, NE Atlantic, 30^10 14C kyr BP ^ Benthicforaminiferal evidence. Mar. Micropaleontol.

Lowe, J.J., Hoek, W.Z., INTIMATE Group, 2001. Inter-re-gional correlation of palaeoclimatic records for the LastGlacial^Interglacial transition: A protocol for improvedprecision recommended by the INTIMATE project group.Quat. Sci. Rev. 20, 1175^1187

Malmberg, S.A., Jonsson, S., 1997. Timing of deep convectionin the Greenland and Iceland Seas. J. Mar. Sci. 54, 300^309.

Mangerud, J., Andersen, S.T., Berglund, B., Donner, J.J.,1974. Quaternary stratigraphy of Norden, a proposal forterminology and classi¢cation. Boreas 4, 109^128.

McBride, E.F., 1971. Mathematical treatment of size distribu-tion data. In: Carver, R.E. (Ed.), Procedures in SedimentaryPetrology. Wiley, New York, pp. 109^127.

McCave, I.N., Manighetti, B., Robinson, S.G., 1995. Sortable

MARMIC 953 2-2-04

K.L. Knudsen et al. /Marine Micropaleontology 50 (2004) 273^305304

silt and ¢ne sediment size/composition slicing: Parametersfor paleocurrent speed and palaeoceanography. Paleocean-ography 10, 593^610.

O’Brien, S.R., Mayewski, P.A., Meeker, L.D., Meese, D.A.,Twickler, M.S., Whitlow, S.L., 1995. Complexity of Holo-cene climate as reconstructed from a Greenland ice core.Science 270, 1962^1964.

Rasmussen, T.L., Thomsen, E., Labeyrie, L., van Weering,T.C.E., 1996. Circulation changes in the Faeroe^ShetlandChannel correlating with cold events during the last glacialperiod (58^10 ka). Geology 24, 937^940.

Rasmussen, T.L., van Weering, T.C.E., Labeyrie, L., 1997.Climatic instability, ice sheets and ocean dynamics at highnorthern latitudes during the Last Glacial period (58^10 kaBP). Quat. Sci. Rev. 16, 71^80.

Rytter, F., Knudsen, K.L., Seidenkrantz, M.-S., Eir|¤ksson, J.,2002. Modern distribution of benthic foraminifera on thenorth Icelandic shelf and slope. J. Foraminif. Res. 32,217^244.

Sancetta, C., 1981. Oceanographic and ecological signi¢canceof diatoms in surface sediments of the Bering and Okhotskseas. Deep-Sea Res. 28, 789^817.

Sarnthein, M., Jansen, E., Weinelt, M., Arnold, M., Dup1essy,J.-C., Erlenkeuser, H., Flat=y, A., Johannessen, G., Johan-nessen, T., Jung, S., KocS, N., Labeyrie, L., Maslin, M.,P£aumann, U., Schulz, H., 1995. Variations in Atlantic sur-face paleoceanography, 50‡^80‡N: A time-slice record of thelast 30,000 years. Paleoceanography 10, 1063^1094.

Sarnthein, M., Stattegger, K., Dreger, D., Erlenkeuser, H.,Grootes, P., Haupt, B., Jung, S., Kiefer, T., Kuhnt, W.,P£aumann, U., Scha«fer-Neth, C., Schulz, H., Schulz, M.,Seidov, D., Simstich, J., van Krevelt, S., Vogelsang, E.,Vo«lker, A., Weinelt, M., 1999. Fundamental modes andabrupt changes in North Atlantic circulation and climateover the last 60 ky ^ Concepts, reconstruction and numericalmodelling. In: Scha«fer, P., Ritzrau, W., Schu«ter, M., Thiede,J., (Eds.), The Northern North Atlantic: A Changing Envi-ronment. Springer, Berlin, pp. 45^66.

Shackleton, N.J., 1974. Attainment of isotopic equilibrium be-tween ocean water and the benthonic foraminifera genusUvigerina : Isotopic changes in the ocean during the lastglacial. Colloq. Cent. Natl. Rech. Sci. 219, 203^209.

Stefansson, U., 1962. North Icelandic Waters. Rit Fiskideild.3, 1^269.

Stuiver, M., Reimer, P.J., Bard, E., Beck, J.W., Burr, G.S.,Hughen, K.A., Kromer, B., McCormack, G., van der Plicht,J., Spurk, M., 1998a. INTCAL98 Radiocarbon Age Calibra-tion, 24,000^0 cal BP. Radiocarbon 40, 1041^1083.

Stuiver, M., Reimer, P.J., Braziunas, T.F., 1998b. High-preci-sion radiocarbon age calibration for terrestrial and marinesamples. Radiocarbon 40, 1127^1151.

Swift, J.H., 1986. The Arctic Waters. In: Hurdle, B.G. (Ed.),The Nordic Seas. Springer, New York, pp. 129^153.

van Andel, T.H., Hearth, G.H., Moore, T.C., 1975. Cenozoichistory and paleoceanography of the central equatorial Pa-ci¢c. Mem. Geol. Soc. Am. 143, 1^134.

Veum, T., Jansen, E., Arnold, M., Beyer, I., Duplessy, J.C.,1992. Water mass exchange in the North Atlantic and theNorwegian Sea during the past 28,000 years. Nature 356,783^785.

Vidal, L., Labeyrie, L., van Weering, T.C.E., 1998. BenthicN

180 records in the North Atlantic over the last glacial peri-od (60^10kyr): Evidence for brine formation. Paleoceanog-raphy 13, 245^251.

Voelker, A.H.L., Sarnthein, M., Grootes, P.M., Erlenkeuser,H., Laj, C., Mazaud, A., Nadeau, M.-J., Schleicher, M.,1998. Correlation of marine 14C ages from the NordicSeas with the Gisp2 isotope record: Implications for 14Ccalibration beyond 25 ka BP. In: Mook, W.G., Plicht, J.van der (Eds.), Proceedings of the 16th International 14CConference, Radiocarbon 40, pp. 517^534.

Vogelsang, E., 1990. Pala«o-Ozeanographie des Europa«ischenNordmeeres an Hand stabiler Kohlensto¡- und Sauersto¡-isotopen. Ber. Sonderforsch.bereich 313, 1^136.

Waelbroeck, C., Duplessy, J.-C., Michel, E., Labeyrie, L., Pail-lard, D., Duprat, J., 2001. The timing of the last deglacia-tion in North Atlantic climate records. Nature 412, 724^727and Nature 414, 470 (correction).

Walker, M.J.C., Bjo«rck, S., Lowe, J., Cwynar, L.C., Johnsen,S., Knudsen, K.-L., Wohlfarth, B., the INTIMATE Group,1999. Isotopic ‘events’ in the GRIP ice core: A strato-type for the Late Pleistocene. Quat. Sci. Rev. 18, 1143^1150.

Zielinski, G.A., Mayewski, P.A., Meeker, L.D., Gro«nvold, K.,Germani, M.S., Whitlow, S., Twickler, M.S., Taylor, K.,1997. Volcanic aerosol records and tephrochronology ofthe Summit, Greenland, ice cores. J. Geophys. Res. 102,26625^26640.

MARMIC 953 2-2-04

K.L. Knudsen et al. /Marine Micropaleontology 50 (2004) 273^305 305