Calcareous nannofloral response to Termination II at 45°N, 25°W (northeast Atlantic)

ELSEVIER Marine Micropaleontology 34 (1998) 47–70

Calcareous nannofloral response to Termination II at 45ºN, 25ºW(northeast Atlantic)

Anna Lototskaya Ł, Patrizia Ziveri, Gerald M. Ganssen, Jan E. van Hinte

Geomarine Centre, Faculty of Earth Sciences, Free University, De Boelelaan 1085, 1081 HV Amsterdam, The Netherlands

Received 27 May 1997; accepted 15 December 1997

Abstract

Termination II (glacial to interglacial transition from oxygen isotope stage 6 to stage 5) is considered to be one of themost rapid and abrupt climatic changes during the late Quaternary. Our high resolution geochemical and nannofloral studyof the interval from 140 to 100 ka in the northeast Atlantic core T90-9P (45º17.50N, 25º41.30W) shows that TerminationII (135–125 ka) occurred in two steps of drastic changes, interrupted by a short period of stability, and followed byat least 25 ka of relatively stable, interglacial conditions, traditionally referred to as the Eemian age (substage 5.5). Aprimary productivity proxy based on coccolith accumulation rate (number cm�2 ka�1) shows higher values during theEemian than during glacial stage 6 and interglacial substage 5.4, with the maximum values (70 ð 109 coccoliths cm�2

ka�1) at 123–122 ka. The coccolith assemblage was dominated by two morphotypes of Gephyrocapsa alternating indominance under different climatic conditions: Gephyrocapsa muellerae was more abundant in colder periods, whereas‘small’ Gephyrocapsa dominated in the Eemian. Emiliania huxleyi, Syracosphaera pulchra and Florisphaera profundaincreased in the Eemian whereas Coccolithus pelagicus and Helicosphaera carteri decreased. ‘Small’ Gephyrocapsatogether with Emiliania huxleyi are associated with periods of higher primary productivity. The pelagic sedimentary recordof Termination II in core T90-9P includes a 5 cm thick layer enriched in ice-rafted material, a Heinrich layer. In contrast toTermination I, where such a Heinrich layer (H1) is associated with the first meltwater pulse of the deglaciation, the recordof core T90-9P suggests that the Heinrich event of Termination II (H11) occurred during the second stage of deglaciation(129.4–127.1 ka). It is characterised by accumulation of ice-rafted debris and low coccolithophore diversity. The coincidentapparent drop in coccolith accumulation rate could be explained either by changes in primary productivity due to reductionin salinity and light, or by dilution by ice-rafted material. Five models are applied for the deposition of the Heinrich layerwith changing thickness (3 cm) and duration (1000, 500 and 250 yr). These scenarios result in a lowered, constant or evenslightly increased primary productivity during this event. 1998 Elsevier Science B.V. All rights reserved.

Keywords: nannofossils; late Quaternary; North Atlantic; paleoceanography

1. Introduction

Scientific interest in the course of and processesduring the last deglaciation (Termination I) has beendocumented in numerous studies from deep-sea sed-

Ł Corresponding author. E-mail: [email protected]

iments (e.g. Duplessy et al., 1981, 1992; Weaver andPujol, 1988; Lehman and Keigwin, 1992; Sarntheinet al., 1992, 1994) whereas only in the last few yearshas attention been focused on high resolution studiesof the previous deglaciation (Termination II, transi-tion from Marine Isotope Stage (MIS) 6 to 5: 135–125 ka) (Sarnthein and Tiedemann, 1990; Cortijo et

0377-8398/98/$19.00 1998 Elsevier Science B.V. All rights reserved.PII S 0 3 7 7 - 8 3 9 8 ( 9 8 ) 0 0 0 0 5 - X

48 A. Lototskaya et al. / Marine Micropaleontology 34 (1998) 47–70

al., 1994; Oppo et al., 1997). Although this event isconsidered to be one of the most rapid and abruptterminations of the late Quaternary (Broecker andVan Donk, 1970; Sarnthein and Tiedemann, 1990), ithas been poorly studied due to the lack of appropri-ate sedimentary records and due to dating problems.Knowledge of its course will allow detailed com-parison with the last deglaciation and hence a betterprediction of the natural course of climatic variationsin Earth’s present interglacial mode.

The 6=5 stage transition (Termination II), aswell as other glacial to interglacial global climatechanges, is most clearly expressed in the North At-lantic where paleotemperature fluctuations and watermass migrations are larger than in any other area ofthe world ocean (McIntyre et al., 1972; Ruddimanand McIntyre, 1976). Nevertheless, these climaticfluctuations initiated in high northern latitudes aretransmitted to the southern hemisphere (Cortijo etal., 1995) and have a global impact.

Calcareous nannoplankton plays a significant rolein the global carbonate cycle (Bramlette, 1958;Berger and Roth, 1975; Winter et al., 1979), andcan potentially be used as an index of paleoceano-graphic and paleoclimatic conditions (McIntyre andBe, 1967; Holligan et al., 1993; Westbroek et al.,1994). Global distribution, fast evolution and sensi-tive reaction to the parameters of the euphotic zonemake them useful not only for recording long-termchange but also for high resolution studies (Weaverand Pujol, 1988; Jordan et al., 1996).

The objective of this study was to reconstruct ahigh resolution paleoceanographic record of 140–100 ka from the northeast Atlantic deep-sea pelagiccore T90-9P (45º17.50N, 25º41.30W) (Fig. 1A), andto determine the response of calcareous nanno-plankton to the glacial–interglacial transition, basedon nannopaleontological, sedimentological and geo-chemical analyses.

1.1. Heinrich events

The Pleistocene sedimentary record in the NorthAtlantic is characterised by coccolith=foram oozesand ice-rafted material, which is present through-out the section but dominates in layers resultingfrom pulses of increased iceberg discharge, calledHeinrich events. Such layers were first found in

the northeast Atlantic (45–50ºN) (Ruddiman et al.,1977; Heinrich, 1988; Broecker et al., 1992) wherethey are clearly expressed. Their major diagnos-tic features are: high ice-rafted debris (IRD) influxincluding a variable amount of detrital carbonate (re-flected in high dry bulk densities and increased mag-netic susceptibility), and the presence of dropstones.In addition, the coldwater species Neogloboquad-rina pachyderma (sinistral) dominates the plankticforaminiferal assemblage.

The upper four Heinrich layers (H1–H4) whichoriginated from a variable mix of different sourcesincluding Laurentide, Fennoscandian, and Icelandicice-sheets and the British Isles (Revel et al., 1996)are well expressed and most studied. They are almostbarren of coccoliths, and some are also free offoraminifera (Van Kreveld et al., 1996). Less isknown about the lower Heinrich layers down to MIS7, especially concerning their provenance. One of theHeinrich layers was deposited during Termination II(TII), and named by Heinrich (1988) as layer 11,H11 by Robinson et al. (1995), H7 by Broecker(1994), and h7 by Van Kreveld et al. (1996).

Decrease in foraminifera and nannoplanktonabundances in the Heinrich layers can be interpretedas a result of dilution by high detrital flux dur-ing short iceberg calving and melting events. Also,plankton productivity may have been reduced be-cause of surface water cooling (Bond et al., 1992),and=or diminished light penetration due to icebergpresence (Broecker et al., 1992) and increased sus-pended matter, and=or lowered salinity of the surfacewaters by meltwater (Van Kreveld et al., 1996).

2. Material and methods

Piston-core T90-9P (45º17.50N, 25º41.30W; corelength D 1028 cm) was recovered at 2934 m wa-ter depth from the Mid-Atlantic Ridge flank in thenortheast Atlantic (Fig. 1A) during the 1990 JGOFScruise of the R.V. Tyro (Ganssen et al., 1990). Thecore provides a good pelagic record for the past385 ka and well-expressed signals of Heinrich events(Fig. 1B, C).

One hundred and thirteen samples from the wholecore taken at 10 cm intervals were analysed for theδ18O Globorotalia inflata (d’Orbigny) record (30–50specimens per sample). The resulting curve was used

A.L

ototskayaet

al./Marine

Micropaleontology

34(1998)

47–7049

Fig. 1. Core T90-9P (45º17.50N, 25º41.30W). (A) Location in the North Atlantic. (B) The δ18O G. inflata record. (C) Magnetic susceptibility profile. Numbers indicateisotopic stages; Heinrich layers are identified with H1, H2, etc. The grey part of the curves corresponds to the studied interval.

50 A. Lototskaya et al. / Marine Micropaleontology 34 (1998) 47–70

as a framework to select the transition from isotopestage 6 to stage 5 for detailed analysis. An intervalof 81 cm (469–550 cm) was sampled for a highresolution study from the beginning of TerminationII through abrupt changes to the end of substage 5.4(Fig. 1C), 25 ka after TII. The interval (469–550 cm)contains olive-grey to brown terrigenous clays andsilts together with bioturbated white coccolith andforaminiferal oozes.

Samples for dry bulk density (DBD), carbonatecontent (CaCO3%) and δ18O analyses were takenevery centimeter but not all of them were usedfor the complete suite of analyses (some sampleswere not suitable for counts due to errors in weightestimations, or did not contain enough material). Theproduced data represent an almost continuous recordthrough the time interval.

The dry weight per wet volume (DBD, g cm�3)was determined on 4 cm3 syringe sediment samplesafter drying at 60ºC using the electronic precision bal-ance Sartorius 1465MP8 with a standard deviation ofš0.001 g. Then half of each sample was used to anal-yse the weight percent carbonate of bulk sediment(CaCO3%); the rest was wet sieved through 32 µmmesh. Each ‘coarse’ (>32 µm) and ‘fine’ (<32 µm)fraction was used for fine fraction CaCO3%, light mi-croscope and scanning electron microscope studies.

The weight percent carbonate of the bulk sampleand of the coarse fraction were measured using theScheibler gasometric technique (Bruin, 1937), withan error of š2%. The ratio between the fine andcoarse fractions was used further to calculate theweight percentage of carbonate and non-carbonatefor each fraction in all the samples.

Mass accumulation rates of bulk sediment, totalcarbonate and non-carbonate were calculated usingthe following formulas (Van Kreveld et al., 1996):

bulkMAR D R Ł DBD

cMAR D R Ł DBD Ł CaCO3% Ł 0:01

fcMAR D R Ł DBD Ł fCaCO3% Ł 0:01

nMAR D bulkMAR � cMAR

where MAR D mass accumulation rate (g cm�2

ka�1), R D average sediment accumulation rate(cm=ka), DBD D dry bulk density (g cm�3), bulkD bulk sediment, c D carbonate, CaCO3% D weight

percent carbonate of the bulk sediment, fc D fine-fraction, fCaCO3 D weight percent carbonate offine-fraction carbonate, n D non-carbonate.

Coccolith accumulation rates were calculated fol-lowing the formula:

cocAR D R Ł DBD Ł N

where cocAR D total coccolith accumulation rate(no. coccoliths cm�2 ka�1), N D number of coccol-iths per gram dry weight (no. g�1).

Oxygen stable isotope ratio’s were determinedfor the benthic foraminifera species Cibicidoideswuellerstorfi (Schwager) (2–3 specimens per sam-ple); and two planktic species, Globorotalia in-flata (5–6 specimens) and Globigerina bulloidesd’Orbigny (8–10 specimens). All specimens werepicked from the >250 µm fraction. Planktic foramswere cleaned ultrasonically to remove fine adheringparticles. We used a Finnigan MAT 251 mass spec-trometer combined with an automatic Finnigan MATcarbonate preparation line (Bremen type). Isotopedata are referred to the PDB standard notation andcalibrated via NBS 18, 19 and 20. The analyticalprecision is š0.08 (1σ).

The relative abundance of Neogloboquadrinapachyderma (sin.) was determined for 7 samplesby counting 300 specimens of planktic foraminiferafrom the >250 µm fraction.

Whole-core measurements of volume magneticsusceptibility were made using the MS2C GEOTEKMulti-Sensor Core Logger (loop sensor) connectedto a Bartington MS2 meter, with a resolution of 0.5cm. In order to preserve material for further organicchemistry sampling core logs were measured in acold state.

All measurements were done at the Faculty ofEarth Sciences of the Free University in Amsterdam.

2.1. Quantitative coccolith study technique

Analyses of coccoliths were made at a strati-graphic resolution of 1 cm. The following procedurewas used for each sample in order to estimate thenumber of coccoliths per gram sediment and therelative frequences of different coccolith species:

(1) The sample (about 1.5–2 g) was wet sievedthrough 32 µm mesh using 1 litre of tap water, andthe suspension was kept.

A. Lototskaya et al. / Marine Micropaleontology 34 (1998) 47–70 51

(2) A few drops of sodium pyrophosphate (Na4

P2O7) were added to the solution, and warmed on ahot plate until boiling to disaggregate all particles.

(3) One litre of tap water was added to the solu-tion and a magnetic stirrer applied for 15 minutes.

(4) While stirring 0.15 ml of solution was pipettedwith a high precision pipette Eppendorf 4710, andplaced in a small vial.

(5) One drop of hydrogen peroxide (H2O2) wasadded to disaggregate remaining organic matter, 70ml of tap water was added to dilute the solutionwhich was then immediately filtered on to a Mil-lipore filter (47 mm diameter, 0.45 µm pore size)using low pressure to obtain equal distribution of thematerial on the filter.

(6) The filter was oven dried at 40ºC for 1 h.A segment of each filter was used to make a

permanent slide with Canada Balsam for light mi-croscope (LM) study: a drop of balsam was placedon a warmed glass and smeared over it; then thefilter segment was placed on it and covered with acover-glass. This procedure avoided redistribution ofthe material on the filter during slide preparation.

Three hundred to 400 specimens of coccolithswere counted on an area defined by a ‘Thompson’grid (0.0144 mm2) in each filter with a polarised lightmicroscope (ð1000). An additional 100 specimenswere counted of the ‘rare species group’ (i.e. thosespecies which make up less than 10% of the totalcoccolith assemblage). Terrigenous material, detritalcarbonate and foram fragments were also quantified.The counted particles have been extrapolated for theentire sample in order to calculate the number of coc-coliths per gram dry sediment, relative abundances ofdifferent species in the coccolith assemblage and theproportion of ice-rafted material in the sediments.

The number of coccoliths per gram of dry sed-iment for each sample was calculated using thefollowing formula:

N D V1 Ł Sf Ł N 0=.V2 Ł ns Ł S0 Ł mDBD/

where N D number of coccoliths per gram dry sed-iment, V1 D volume of tap water used for sieving(ml), Sf D effective filtration area (cm2), N 0 D num-ber of counted coccoliths, V2 D volume of solutiontaken with a pipette (ml), ns D number of countedsquare fields, S0 D area of the field (cm2), mDBD Dweight of a bulk sediment sample (g).

To test the evenness of the particle distributionon the filter, three samples were checked. Threehundred fifty to 400 coccolith specimens were quan-tified from each of five transects perpendicular to thecircumference of the filter. Results show maximumerror deviation of 1.5–3.5% (Fig. 2). Also, there isno trend observed in the relationship between num-ber of coccoliths in each transect and the positionof the transect within a filter, which confirms thatparticles had a sufficiently even distribution on thefilter surfaces.

‘Small placoliths’ (small-medium size Gephyro-capsa and Emiliania huxleyi were counted separatelyon a scanning electron microscope (SEM) due todifficulties in identification of these species by LM.More than 100 ‘small placoliths’ were identified persample, and their relative frequences were then re-calculated into percentages of the total assemblage.The SEM analyses were carried out on a portion ofthe same filter used for LM analysis.

A total of 113 specimens of Gephyrocapsa ran-domly selected from 4 samples was photographedand measured with SEM in order to determinethe size distribution of G. muellerae and ‘small’Gephyrocapsa after the classification was estab-lished. The percentage of measured forms withineach species=morphotype was determined.

A coccolith diversity index was calculated forall samples using the formula developed for livingcalcareous nannoplankton communities (Honjo andOkada, 1974, after Pielou, 1966):

H 0 D �sX

iD1

.Pi ln Pi /� .S � 1/=2N

where H 0 D diversity index, Pi D proportion ofspecies i in the assemblage, S D total number ofspecies in the sample, N D number of specimenscounted in the sample.

The simple diversity (number of species) is alsocalculated for all samples.

3. Results and discussion

3.1. Isotope stratigraphy

The studied core interval belongs to the E. hux-leyi zone, NN21, of Martini’s Standard CalcareousNannofossil Zonation (Martini, 1971); its detailed

52 A. Lototskaya et al. / Marine Micropaleontology 34 (1998) 47–70

Fig. 2. Plot showing variation in number of coccoliths counted with a light microscope for 3 samples along 5 different transects over thesample preparation.

stratigraphical framework is mainly based on oxygenisotopes of the epibenthic foraminifera Cibicidoideswuellerstorfi (Fig. 3A) and the high resolution deep-sea chronostratigraphy of Martinson et al. (1987)with its assignment of isotopic events and substages.The following minima and maxima in δ18O valueswere identified: isotopic event 6.2 at 540 cm coredepth dated as 135.10 ka; event 5.53 at 519 cm depthdated as 125.19 ka; event 5.4 at 484 cm depth datedas 110.79 ka; and possibly event 5.33 at 469 cmdepth dated as 103.29 ka.

The benthic foraminiferal isotope data for eachlevel of the termination are based on two or foursubsamples of two–three specimens each. In generalwe found high reproducibility of the record, exceptfor the lower (538–531 cm) and upper (526–519cm) part of the TII interval. To compliment thebenthic isotope stratigraphy we determined the oxy-gen isotope ratios of the two planktic foraminiferaspecies Globigerina bulloides (8–10 specimens) andGloborotalia inflata (5–6 specimens) inhabiting sur-face and subsurface waters, respectively (Fig. 3B;Be, 1977). The double measurements made for theTII interval show a much higher reproducibility ex-cept for one sample at a depth of 529 cm. As a resultof their depth habitats, in stratified waters G. bul-

loides and G. inflata produce a different δ18O signalwhereas in a mixed water column they would showsimilar isotope values. Using this notion in combi-nation with the C. wuellerstorfi isotope record, weconsider the level of 519 cm core depth to representevent 5.53, i.e. the break point between non-stratifiedconditions of the glacial=termination and stratifiedwaters of the Eemian maximum.

The 6=5 stage boundary was placed at 529.5cm depth as the depth midpoint of Termination IIbetween events 6.2 and 5.53, implying a date of129.84 ka for this level (Martinson et al., 1987). The5.5 to 5.4 and 5.4 to 5.3 substage boundaries wereplaced at the benthic isotope excursions at 485 and470.5 cm, respectively.

The oxygen isotope curve shows 2‰ change inδ18O values of C. wuellerstorfi between 540 and 519cm recording a considerable climatic shift during the10 ka of Termination II. The isotope shift showstwo major steps of deglaciation (Fig. 3A). Two parts(538–531 and 526–519 cm), with a major shift in theδ18O C. wuellerstorfi record and low reproducibil-ity, are separated by an interval with little changeand good reproducibility (a four data points plateau,‘Younger Dryas’ style event).

A bioturbation depth of 6–7 cm may be seen in

A. Lototskaya et al. / Marine Micropaleontology 34 (1998) 47–70 53

Fig. 3. Stratigraphic framework for the studied interval in core T90-9P. (A) δ18O C. wuellerstorfi record. Black curve represents meanvalues, grey dots show individual measurements. Numbers indicate isotopic events. (B) δ18O G. inflata and G. bulloides record. (C) Agesestimated from isotope chronostratigraphy of Martinson et al. (1987), sediment accumulation rates and isotope substage boundaries.

the second step of the termination as an intervalwhere light benthic δ18O values from above 520 cmmixed downwards, with the heavier values of thebase of the second deglaciation step. In the sedimentthe time interval of TII is represented by 21 cmof brown calcareous muds. Sedimentation rates arecalculated for the first part of the termination as 2.0cm ka�1, and 2.3 cm ka�1 for the second.

Low oxygen isotope variations (<1‰ in the ben-thic and <0.45‰ in the planktic isotope records)(Fig. 3A, B) are observed in the interval deposited

during the Eemian (substage 5.5) from the 5.53 eventon. This time is characterised by stable climate con-ditions. The average sediment accumulation rate dur-ing the Eemian was 2.4 cm ka�1. Sediments of MIS5.4 accumulated at the average rate of 2.0 cm ka�1.Taking into account the errors of age estimates (Mar-tinson et al., 1987), the differences in calculated sedi-ment accumulation rates in different parts of the stud-ied interval (Fig. 3C) are not significant, and the sed-iment accumulation rate could be considered as ap-proximately constant throughout the whole interval.

54 A. Lototskaya et al. / Marine Micropaleontology 34 (1998) 47–70

Fig. 4. Physical and compositional properties of the studied interval of core T90-9P. Heinrich layer is indicated by grey area. (A) Drybulkdensity. (B) Magnetic susceptibility. (C) Weight percentage of carbonate (in the fine and coarse fractions) and of non-carbonate. (D)Percentage of ice-rafted debris in the fine fraction (detrital carbonate particles and total amount of debris grains), and relative abundanceof N. pachyderma (sin.) in the planktic foraminifera assemblage (fraction >250 µm).

The petrophysical and sedimentary data showconsiderable variation in the lower part of the studiedinterval (Fig. 4): the section between 540 and 520cm is characterised by peaks in dry bulk density,magnetic susceptibility, non-carbonate fraction (upto 76%), fine fraction IRD (up to 20%), detrital car-bonate, and % of N. pachyderma (sin.) in the plankticforaminifera assemblage. This interval is consideredto contain a Heinrich layer (Fig. 4), its strong sig-nal being spread upwards and downwards by bio-turbation (conforming to the bioturbation model ofManighetti et al., 1995). We have defined the ‘Hein-rich layer’ as the interval corresponding to maximalchanges (523–528 cm; Fig. 4). This particular event

coincides with the second, and largest, deglaciationstep as recognised in the benthic isotope record, andended about 2000 yr before event 5.53.

In contrast to Termination I where the Heinrichlayer (H1) is associated with the first meltwaterpulse of the deglaciation (Broecker, 1994), the Hein-rich event of Termination II (H11) seems to haveoccurred during its second part. Yet, it appears to beof much lower intensity than other Heinrich eventsin the record of the same core: Fig. 1C shows here aconsiderably smaller shift in magnetic susceptibilitythan for the other well-studied Heinrich layers. Thecore is located at the edge of the IRD depositionbelt (Ruddiman, 1977; Grousset et al., 1993) where

A. Lototskaya et al. / Marine Micropaleontology 34 (1998) 47–70 55

Fig. 5. Mass accumulation rates for the studied interval in core T90-9P. Heinrich layer is indicated by grey area. (A) Bulk sediment,carbonate, fine fraction carbonate and non-carbonate accumulation rates. (B) Coccolith accumulation rate.

magnetic susceptibility is a key feature of Heinrichevents. Increased amount of IRD material below theHeinrich layer (Fig. 4) during the first part of thedeglaciation is probably a result of a fast temperaturerise followed by intensive iceberg discharge.

Comparison of mass accumulation rates calcu-lated for total bulk sediment, total carbonate, fine-fraction carbonate, non-carbonate, and coccoliths(Fig. 5) shows a higher detrital input during theHeinrich event of about 1.7 g non-carbonate cm�2

ka�1, and low carbonate and coccolith accumulationrates. Assuming a constant sediment accumulationrate between the 6=5 stage boundary and event 5.53,

the age of the Heinrich layer has been calculatedto be 129.4–127.1 ka (Fig. 5). This is a maximumduration, since it does not include any correction forenhanced sedimentation rate (see further).

In the Eemian the mass accumulation rate is gen-erally higher, especially the carbonate flux which iscalculated to reach 1.8–1.9 g carbonate cm�2 ka�1.Increased bulk and carbonate accumulation rates anda minimum in IRD associated with substage 5.5 indi-cate higher biological productivity (Van Kreveld et al.,1996). Estimated SSTs by MAT (modern analog tech-nique) and TFT (transfer function technique) (Oppoet al., 1997) and european pollen records (Guiot et al.,

56 A. Lototskaya et al. / Marine Micropaleontology 34 (1998) 47–70

1989) indicate a gradual warming in the early Eemianfollowed by cooling in the late Eemian. The amplitudeof oscillations superimposed on the gradual SST trendduring MIS 5.5 were only 1–2ºC (Oppo et al., 1997).

3.2. Coccolith accumulation rate as a potentialprimary productivity proxy?

Today biogenic carbonate dominates the annualmass flux in the northeast Atlantic (Samtleben andBickert, 1990; Honjo and Manganini, 1993; New-ton et al., 1994), with the coccolithophores as ma-jor contributors (Broerse et al., submitted; Ziveri etal., submitted). Coccolithophores were also the ma-jor carbonate producers during the last interglacial(McIntyre et al., 1972). At 45ºN, 25ºW the fine-fraction carbonate, mainly composed of coccoliths,dominates the total carbonate in weight through-out the studied interval (Fig. 4C). Nearly the sameweight percent of coccolith carbonate was producedduring the interglacial (average value is 74.1% of thetotal carbonate) as during the glacial (around 71%).

Sediment trap studies from the North Atlantic (at48ºN, 21ºW) show that the carbonate composes 55–59% of the total mass flux (Honjo and Manganini,1993). Biogenic silica (opal) is here a minor compo-nent (18–21% of the total mass flux), and is positivelycorrelated with variations in the carbonate flux.

A positive correlation has also been observedbetween present day carbonate and organic carbonfluxes in sediment traps when normalised to 3200m water depth (Brummer and Van Eijden, 1992).We considered that coccolith accumulation rate canbe used as a proxy of primary paleoproductivity inthis region, because: (1) coccolithophores here weremajor primary and export producers; and (2) bio-genic carbonate preservation throughout the wholeinterval is relatively good. However, coccolith ac-cumulation rate estimates are not reliable for theHeinrich event where sedimentation rates could havebeen much higher than the calculated average as willbe discussed later.

The coccolith primary paleoproductivity proxyshows (Fig. 5B):

(1) Low values during glacial stage 6 (¾ 20ð 109

coccoliths cm�2 ka�1).(2) A pronounced minimum at 128–125 ka (down

to 4 ð 109 coccoliths cm�2 ka�1) which is of the

same duration as the Heinrich event but lags it byabout 1 ka (the major drop corresponds to the secondhalf of the Heinrich event and to 1600 yr afterwards).The reduction in coccolithophore productivity couldbe the result of a considerable decrease in light pene-tration and salinity. The decrease of light could haveslowed down coccolithophore growth and photosyn-thesis rates leading to a reduction in productivity.Nevertheless, coccoliths still dominate the carbonateproduction during the event even though the detritalcarbonate content is higher (Fig. 5A).

(3) A sharp peak of 70 ð 109 coccoliths cm�2

ka�1 corresponding to an age of 123–122 ka. Thislags the Eemian maximum (event 5.53) by 2000 yr.

(4) Highly fluctuating values during the Eemian(¾ 45 ð 109 coccoliths cm�2 ka�1) decreasing in itssecond part (after ¾117 ka) towards MIS 5.4 with aminimum at the end of MIS 5.5 (112 ka). When com-pared to fine fraction carbonate accumulation rates(Fig. 5A), the glacial=interglacial transition in thecoccolith flux shows the similar shift of 2.3 times.Unfortunately, it is not possible to usefully convertcoccolith accumulation from numbers into weight dueto the lack of mass measurements of many species inthe sediment, discrepancies in the available mass esti-mates, and a non-absolute preservation of coccoliths.

(5) Generally low values during substage 5.4 (¾20ð 109 coccoliths cm�2 ka�1).

Previous primary productivity estimates for theopen northeast Atlantic were of lower resolution butshow similar glacial=interglacial fluctuations. McIn-tyre and McIntyre (1970) found ¾2 times higherproductivity in the Recent sediments than in theGlacial based on weight coccolith concentrations inthe sediment. Estimated primary productivity for thepast 130 ka derived from biogenic carbonate massaccumulation rates shows a north–south decrease(from 53 to 45ºN) both in absolute values, and in thecontrast between ‘warm’ and ‘cold’ productivity (Ot-tens et al., 1995). Generally higher paleoproductivityis observed during interglacials, and at 48ºN MIS5 is characterised by lower values than in isotopestages 1 and 7 (Van Kreveld et al., 1996).

3.3. Coccolith assemblage

The assemblage of calcareous nannofossils in-cludes 17 species from 15 genera. Gephyrocapsa ac-

A. Lototskaya et al. / Marine Micropaleontology 34 (1998) 47–70 57

Fig. 6. Changes in the cumulative coccolith assemblage of the studied interval of core T90-9P. Numbers correspond to isotopic stages.

counts for 59–89% of the total flora throughout theinterval (Fig. 6). Emiliania huxleyi follows Gephyro-capsa spp. in relative abundance forming 2–25% ofthe total coccolith assemblage. The rest of the floramakes up 8–27%.

3.3.1. DiversityThe coccolith diversity index .H 0/ does not

change significantly through the interval (Fig. 7A),and reveals values between 0.9 and 1.4. The simplediversity shows values of 8–14. Coccolithophore di-versity indices calculated for the surface sedimentsfrom the same area (45–47ºN) show comparablevalues varying from 1.0 to 1.5 (Ziveri et al., sub-mitted). During the Eemian (substage 5.5) both thediversity index and the simple diversity are gener-ally higher than during MIS 6 and 5.4, which can

be explained by a retreat of low diversity waters ofthe subpolar front (modern high-latitude coccolithassemblages are characterised by a considerably re-duced diversity, Honjo and Okada, 1974). A shortdecrease in diversity value is evident correspondingto the Heinrich layer.

3.3.2. Gephyrocapsa spp.Major changes in coccolith assemblage are ob-

served in the ratio of two Gephyrocapsa forms:‘small’ Gephyrocapsa and G. muellerae. They showchanges in dominance in the relative abundances dur-ing TII and at transition 5.5–5.4 (Fig. 7A). Gephyro-capsa muellerae is considered to prefer cool surfacewater conditions (SST below 18ºC, Samtleben et al.,1995) and is known to presently dominate NorthAtlantic waters in the Azores Front (Jordan, 1988;

58 A. Lototskaya et al. / Marine Micropaleontology 34 (1998) 47–70

A. Lototskaya et al. / Marine Micropaleontology 34 (1998) 47–70 59

Winter et al., 1994). In T90-9P G. muellerae consid-erably decreased in relative abundance in the Eemianalthough in absolute numbers per gram sediment itshows less variation (Fig. 7B). In the Eemian ‘small’Gephyrocapsa became very important (up to 60%of the coccolith assemblage producing a maximumof 15 ð 109 coccoliths g�1 sed.) forming an acmezone. A similar pattern was described from ODP Site658 (off West Africa) with the ‘small’ Gephyrocapsa(G. aperta) being more numerous than G. muelleraein MIS 5.1 and 5.5 (Jordan et al., 1996), and fromthe Eastern Mediterranean at TII, sapropel S5 ofcore TTR3-80G (Lototskaya, 1994). ‘Small’ Gephy-rocapsa (Gephyrocapsa ericsonii) was also found todominate the Eemian in the Red Sea (Winter, 1982).

Living G. ericsonii preferentially inhabit waterswith a temperature range of 13–22ºC (Okada andMcIntyre, 1979). Discrepancies exist on the optimaltemperature for this species: 19ºC is observed byMcIntyre et al. (1970), and 14ºC by Samtleben etal. (1995). Pujos and Giraudeau (1993) observe thepredominance of ‘small’ Gephyrocapsa during MIS7 in the North Atlantic north of 45ºN, and alsoduring isotope stages 6 and 5 — between 45 and20ºN. Hine and Weaver (1998) report an acme zoneof Gephyrocapsa aperta in the northeast Atlanticfrom the lower part of MIS 6 downwards to the topof MIS 8 and refer to the interval between this G.aperta Acme zone and the E. huxleyi Acme of MIS4 as ‘Transitional Zone’, with lower frequences ofG. aperta and G. muellerae, in which G. muelleraeis dominant in the north but is replaced by G. apertatoward the south.

Our results also show that ‘small’ Gephyrocapsafollows the total flora absolute abundance distribu-tion curve .R2 D 0:734/ (Fig. 7B). This suggests thatit is related to interval of higher productivity. Gartner(1988) also associated ‘small’ Gephyrocapsa withhigher nutrient concentrations in the photic zone.He reports an episode of overwhelming dominanceof ‘small’ Gephyrocapsa at 930–125 ka in tropi-cal and subtropical oceans referring to E. huxleyi

Fig. 7. Frequency distribution of coccoliths in core T90-9P. The grey band marks a Heinrich layer. (A) Coccolith diversity (diversityindex and simple diversity) and relative distribution of G. muellerae, ‘small’ Gephyrocapsa, E. huxleyi and C. leptoporus. (B) Numberof coccoliths per gram of dry sediment for the total flora, G. muellerae, ‘small’ Gephyrocapsa, E. huxleyi and C. leptoporus. Note scaledifferences.

as to its modern analog. In these regions ‘small’Gephyrocapsa is replaced by G. oceanica during theEemian and is thus an indicator of relatively de-pressed temperatures, while in the temperate north-east Atlantic the temperature relation is reversed.Here, the ‘warmest’ morphotype, G. oceanica, is ab-sent, and the temporal distribution of G. muelleraeand ‘small’ Gephyrocapsa clearly shows that G.muellerae dominates in colder waters and is replacedby ‘small’ Gephyrocapsa during the Eemian withhigher temperature regimes and higher productivity.These observations imply that the acme zones of dif-ferent Gephyrocapsa species in the Pleistocene arenot just the product of Gephyrocapsa evolution butare also related to environmental change.

3.3.3. Emiliania huxleyiThe studied interval was deposited before the

E. huxleyi Acme zone, 73–0 ka for the temperateregions (Thierstein et al., 1977). Emiliania huxleyihere is relatively low in numbers. However, at 517cm this species reaches 25% of the total assemblage(Fig. 7A), and increased E. huxleyi relative abun-dances and absolute numbers (interval of 522–484cm) are found throughout the part of MIS 5.5 where‘small’ Gephyrocapsa dominates over G. muellerae(Fig. 7A, B) suggesting a positive correlation withproductivity (correlation with the total flora absoluteabundance distribution R2 D 0:757). The studied in-terval could include different morphotype(s) but theidentified E. huxleyi are low calcified forms (Plate I,7) which are different from the morphotype A dom-inating present waters (Young and Westbroek, 1991;Ziveri et al., 1995). Its poor preservation does notallow for assignment to the other described extantmorphotypes (low calcified forms B and C) or todetermine whether the observed variations in fre-quency are due to environmental changes or reflectan evolutionary trend.

The relative and particularly the absolute abun-dance distribution of ‘small’ Gephyrocapsa and E.huxleyi (Fig. 7A, B) suggest that the gradual reduc-

60 A. Lototskaya et al. / Marine Micropaleontology 34 (1998) 47–70

A. Lototskaya et al. / Marine Micropaleontology 34 (1998) 47–70 61

tion toward substage 5.4 started by 117 ka suggestingan early decrease in productivity and possibly in sur-face water temperatures as seen from the plankticforaminifera (G. inflata and G. bulloides) oxygenisotope record (Fig. 3B). An implication for predic-tion of the present interglacial natural course can bemade when more information is available.

3.3.4. Calcidiscus leptoporusThe distribution of Calcidiscus leptoporus, the

most abundant coccolith species after Gephyrocapsaand E. huxleyi (Fig. 7A, B), does not show a cleartrend. Its absolute frequency curve is generally pos-itively correlated to the coccolith number per gramsediment (i.e., indicates higher primary productiv-ity). Highest abundances of this species are recordedduring the Eemian in agreement with the abundancepeaks during MIS 1 and 5.5 reported by Gard (1989)for the North Atlantic.

3.3.5. Rare speciesThe whole group of rare species makes up only

¾10% of the total nannoflora (Fig. 6) and is repre-sented by 12 species. Although we counted upto 100specimens additionally, only species present at 7%can be detected on the 95% probability level. Thus,abundance fluctuations of different species withinthe assemblage are statistically reliable only on thebase of the general trends. The relative frequency ofeach of the rare species is presented in Fig. 8A as apercentage of the total ‘rare species group’.

Glacial isotope stage 6, and, to a lesser de-gree, interstadial substage 5.4, are characterised byhigher frequences of Coccolithus pelagicus and He-licosphaera carteri (Fig. 8A). C. pelagicus is known

Plate IScale bars D 1 µm. Note the same scale for 2–7.1. Gephyrocapsa oceanica Kamptner, 1943. Core T90-9P, depth506 cm.2, 3. ‘Small’ Gephyrocapsa. Core T90-9P, depth 506 cm.4. Gephyrocapsa muellerae Breheret, 1978 form A. Core T90-9P,depth 506 cm.5. Gephyrocapsa caribbeanica Boudreaux and Hay, 1967. CoreT90-9P, depth 489 cm.6. Gephyrocapsa muellerae Breheret, 1978 form B. Core T90-9P,depth 504 cm.7. Emiliania huxleyi (Lohmann, 1902) Hay and Mohler, 1967.Core T90-9P, depth 506 cm.

to be a polar species occupying cold waters downto a temperature as low as 1ºC (Okada and McIn-tyre, 1979). C. pelagicus also shows a remarkablepeak in relative abundances in the Heinrich layer(Fig. 8A), suggesting both southward penetration ofsubpolar waters and possible redeposition of north-ern contemporaneous sediments during the event. Itmight be used as potential biotic marker for Hein-rich events. C. pelagicus was almost absent duringmost of the Eemian (Fig. 8). Similar fluctuations areobserved in the distribution of H. carteri (Fig. 8A,B) which also was more abundant in the glacial andduring the termination. In contrast, this species is ob-served to increase in abundance with increasing seasurface temperatures (Okada and McIntyre, 1979;Gard, 1989). According to Castradori (in Violantiet al., 1991) H. carteri is less light dependent thanother coccoliths and does relatively well under turbidconditions.

Florisphaera profunda and Syracosphaera pul-chra have comparable no.=g sediment with highervalues in MIS 5.5 (Fig. 8B). Florisphaera profundais restricted to the lower euphotic zone at temper-atures >10ºC and inhabits present low latitude re-gions (Okada and Honjo, 1973; Okada and McIntyre,1977). Through time the relative frequency of F. pro-funda in coccolith assemblages permits monitoringof the nutricline depth (Molfino and McIntyre, 1990).Off west Africa (ODP site 658) it is very abundantduring substage 5.5 and represents the warm wa-ter flora (Jordan et al., 1996). Low frequences ofF. profunda are due to the northern position of ourcore site at 45ºN, where even during the Eemianlow water temperatures constitute the major limitingfactor to F. profunda production. As a result, there isno correlation observed between the abundance of F.profunda and water column stratification defined bythe oxygen isotope record of planktic foraminiferaG. bulloides and G. inflata. Syracosphaera pulchrais common in water samples and sediments fromthe tropics and subtropics as a part of the wholeSyracosphaera group (McIntyre et al., 1972; Rothand Berger, 1975; Jordan et al., 1996). However, re-cently a S. pulchra June bloom has been recorded ina North Atlantic sediment trap (47ºN, 20ºW) (Knap-pertsbusch and Brummer, 1995).

Gephyrocapsa oceanica is virtually absentthrough the whole interval. This can be explained

62 A. Lototskaya et al. / Marine Micropaleontology 34 (1998) 47–70

A. Lototskaya et al. / Marine Micropaleontology 34 (1998) 47–70 63

by the preference of this species for equatorial(Okada and Honjo, 1975; Geitzenauer et al., 1977)and tropical=subtropical (McIntyre et al., 1970) re-gions, and marginal seas (Okada and Honjo, 1975).

All other species make up less than 5% of the‘rare species group’.

3.4. Primary productivity during the Heinrich event.Production or dilution?

The coccolith accumulation rate shows a consid-erable decrease within the Heinrich layer (Fig. 5B).This might be a result of a drop in productivity orsimply a dilution effect. IRD discharges in the NorthAtlantic have been suggested as near-instantaneousevents (Heinrich, 1988; Broecker et al., 1992, andothers) but the resolution of age dating is insufficientto determine the exact duration of Heinrich layerdeposition and is hampered by bioturbation and re-suspension of sediments. Also, Heinrich events ofthe past 400 ka could have been different in origin,duration and intensity. MacAyeal (1993) estimatesthe average continuation of iceberg calving pulsesto be 150–500 yr based on his oscillation model;Dowdeswell et al. (1995) propose Heinrich events tolast 250–1250 yr by using a mass-balance approach.The duration of the event studied in the present paperis estimated at 2300 yr, as calculated by interpolationof the average sediment accumulation rate based onoxygen isotope chronostratigraphy (ages 129.84 and125.19 ka at 529 and 519 cm, respectively). We useda 5 cm interval, corresponding to maximal changesin sediment composition, as the thickness of theHeinrich layer. This duration can only be a rough es-timate because of errors in the two age assignmentsand in the estimated thickness of the layer, as well asin the fact that the accumulation rate of the intervalwill not have been constant.

We present here five models for paleoproductiv-ity (coccolith accumulation rate) changes during theHeinrich event (Fig. 9). Firstly, we assume shorterdurations than the 2300 yr derived from the sedimentaccumulation rate of 2.3 cm ka�1, i.e. 1000, 500 and

Fig. 8. Frequency distribution of coccoliths in core T90-9P. The grey band mark a Heinrich layer. (A) Relative abundance of C. pelagicus,H. carteri, F. profunda and S. pulchra, as a percentage of the total rare species amount. (B) Absolute abundance distribution of rarespecies (C. pelagicus, H. carteri, F. profunda and S. pulchra) in the studied interval. Note scale differences.

250 yr. Second, it is assumed that the original layerwas thinner than our estimated 5 cm (528–523 cm),i.e. 3 cm (527–524 cm). Scenarios for less than 3cm thick Heinrich layer are not considered, as an in-terval of IRD distribution below and above the layer(Fig. 4D) would have then been much higher thanpreviously estimated maximum bioturbation depth of7 cm.

The originally calculated coccolith accumulationrates during the Heinrich event (¾ 4 ð 109 coccol-iths cm�2 ka�1) is 5 times lower than the precedingTermination II values (¾ 20 ð 109 coccoliths cm�2

ka�1) (Fig. 9A). Smaller drop in productivity duringthe Heinrich event would also be observed in thesituation of 1000 yr duration of the event if we as-sume a layer thickness of 3 cm (Fig. 9B). Models forthe Heinrich event for 500 yr=3 cm and 1000 yr=5cm show values which would be expected during thenormal course of the deglaciation averaging betweenthe glacial and the Eemian (Fig. 9C, D). These twoscenarios imply dilution effect and no paleoproduc-tivity changes. A productivity increase would haveoccurred in the scenarios of 250 yr duration of theHeinrich event deposited as a 3 cm layer, and of 500yr=5 cm (Fig. 9E, F).

The models described above (Fig. 9) attempt tosimulate processes that happened during the depo-sition of the Heinrich event. However, factors suchas completeness of sediment mixing (Ruddiman andGlover, 1972) and selectiveness of burrowing organ-isms (McCave, 1988) during bioturbation are nottaken into account. A minimum in coccolith abun-dance is observed at the interval of 520–525 cm(Fig. 7B) while the interpretation of the Heinrichlayer position based on sedimentological parame-ters is 523–528 cm (Fig. 4). These intervals couldbe different ‘events’ explained by a shift in coccol-ith accumulation rate which decreases only duringthe second half of the Heinrich event and stayslow for another 1600 yr (Fig. 5). That ‘delay’ inprimary productivity could be a result of the sig-nificant switch in dominance between G. muelleraeand ‘small’ Gephyrocapsa at the end of the Heinrich

64 A. Lototskaya et al. / Marine Micropaleontology 34 (1998) 47–70

Fig. 9. Different scenarios for coccolith accumulation rate changes (paleoproductivity proxy) during the Heinrich event assuming itsshorter duration and that the original layer was thinner. Grey area indicates the Heinrich layer. (A) Thickness 5 cm, duration 2300 yr(originally estimated values). (B) Thickness 3 cm, duration 1000 yr. (C) Thickness 3 cm, duration 500 yr. (D) Thickness 5 cm, duration1000 yr. (E) Thickness 3 cm, duration 250 yr. (F) Thickness 5 cm, duration 500 yr.

event when climatic factors were not yet favourablefor either of these gephyrocapsids. Alternatively, the1600 yr delay in coccolith accumulation rate reduc-tion (expressed in a 3 cm interval) may be an artefactcaused by selective bioturbation of fine particlesby burrowing animals, resuspension of the surfacesediment, or sinking of heavy IRD particles in thesediment suspension.

4. Summary and conclusions

High resolution studies of an 80 cm inter-val in the North Atlantic core T90-9P (45º17.50N,25º41.30W) show a good record of the penultimateglacial=interglacial transition (Termination II) fromMIS 6 to 5 (140–100 ka). Coccolithophores were themajor carbonate producers through the whole inter-val. The diversity calculated for the coccolith assem-blages is comparable to surface sediment values, andis slightly higher during the Eemian (substage 5.5)than during glacial stage 6 and substage 5.4, whichcan be explained by a retreat of low diversity watersof the subpolar front.

The floral assemblage is dominated by two mor-photypes of Gephyrocapsa replacing each other indominance under different climatic conditions: G.muellerae was more abundant in colder periods,whereas ‘small’ Gephyrocapsa dominated in timesof warmer climate. Emiliania huxleyi, with a dif-ferent morphotype to that which dominates presentwaters, was also recorded in high abundance duringMIS 5.5. Both ‘small’ Gephyrocapsa and E. huxleyimay be associated with higher temperature and=orhigher primary productivity.

The benthic δ18O C. wuellerstorfi record duringTermination II (135–125 ka) suggests that the transi-tion occurred in 2 steps of drastic change interruptedby more stable conditions that could be interpreted asa ‘Younger Dryas’ type event. A combination of theoxygen isotope record of G. bulloides and G. inflatamark the 5.53 isotope event (Eemian maximum) asthe transition from a mixed to a well stratified upperwater column.

The Eemian (substage 5.5) is characterised byhigher carbonate mass and coccolith accumulationrates compared to glacial stage 6 and substage 5.4.

A. Lototskaya et al. / Marine Micropaleontology 34 (1998) 47–70 65

The same glacial=interglacial shift of about 2.3 timesis observed in both fine carbonate and coccolithfluxes. Primary productivity of coccolithophores wasmaximal (70ð109 coccoliths cm�2 ka�1) at 123–122ka. The Eemian is further marked by a dominance of‘small’ Gephyrocapsa in the calcareous nannoplank-ton assemblage, while in MIS 6 and 5.4 G. muelleraewas dominant. Of the rare species, S. pulchra and F.profunda increase in substage 5.5 whereas C. pelag-icus and H. carteri decrease. C. pelagicus might beused as a potential biotic marker for Heinrich events.

Gradual reduction of ‘small’ Gephyrocapsa andE. huxleyi abundance from the Eemian maximum(isotopic event 5.53) toward substage 5.4 occurredalready after 117 ka suggesting an early decreasein productivity and possibly in surface water tem-peratures as seen from the planktic foraminifera (G.inflata and G. bulloides) oxygen isotope record. Animplication for prediction of the present interglacialnatural course can be made when more informationis available.

In contrast to Termination I where the Heinrichlayer (H1) is associated with the first meltwaterpulse of the deglaciation, the Heinrich event of Ter-mination II (H11) occurred during its second part(129.4–127.1 ka). It is characterised by higher ac-cumulation of ice-rafted debris and slightly lowercoccolithophore diversity. The accompanied appar-ent drop in coccolith accumulation rate may be theresult of dilution by ice-rafted material and=or oflow primary productivity due to reduction in lightpenetration and salinity. It also could be an artifactof low dating precision and incorrect thickness of theHeinrich layer. Assuming a constant accumulationrate for the interval 519–529 cm that contains the 5cm thick layer, it was deposited over 2300 yr.

To simulate processes that might have happenedduring the Heinrich event different scenarios of pri-mary productivity changes are made for the timeinterval by assuming shorter duration of the event(1000, 500 and 250 yr), and an originally thinnerlayer (3 cm). A productivity reduction during theHeinrich event would have occurred if the eventlasted 1000 yr and produced a layer of 3 cm; sce-narios for the event for 500 yr=3 cm and 1000 yr=5cm show values which would be expected during thenormal course of the deglaciation averaging betweenthe glacial and the Eemian. A slight productivity

increase would have occurred in the scenarios of 250yr duration of the Heinrich event deposited as a 3 cmlayer, and of 500 yr=5 cm.

At present, a number of genus Gephyrocapsaclassifications exist based on the combination ofmorphological characteristics such as size, bridgeangle, roundness, pore width and other descriptivefeatures (an overview is given by Bollmann, 1997).However, there is no generally accepted classifica-tion.

In our study, coccoliths of the genus Gephyro-capsa constitute up to 89% of the total coccolithassemblage. During the LM analyses of our ma-terial we could distinguish G. oceanica Kamptner(Table 1; Plate I, 1), a form >4 µm, and 2 mor-photypes of Gephyrocapsa of smaller size (<4 µm)which largely dominate the coccolith assemblagethroughout the studied interval.

SEM measurements on the <4 µm forms showa roughly bimodal size distribution in two samplesrelated to interglacial and glacial intervals (smallerplacoliths dominate in one sample, and larger onesin the other) (Fig. 10) which suggests the presence oftwo discrete morphological groups within Gephyro-capsa <4 µm, with a size boundary varying between2.7 and 2.9 µm (a size-interval where the coccolithnumber is very low). The larger group of Gephy-rocapsa (>2.7 µm) may be directly related to G.muellerae Breheret (Table 1; Plate I, 4, 6), while thesmaller ones represent a plexus of species or forms.Various species names exist in the literature forthe small-size Gephyrocapsa; their distinctions havebeen made on bridge characteristics and the size ofthe distal shield and the central area (e.g. McIntyreand Be, 1967; Hay and Beaudry, 1973; Samtleben,1980). Most smaller Gephyrocapsa in our materialare represented by coccoliths of Gephyrocapsa cf.aperta, in a size range <2.7 µm. The rest of thegroup includes morphologically distinct coccoliths(thick bridge, wide distal shield with low number ofelements) of size up to 2.9 µm. However, morpho-logical criteria did not allow us to make consistentdistinctions in our material, and all smaller placoliths<2.9 µm excluding G. muellerae are referred to as‘small’ Gephyrocapsa (Table 1; Plate I, 2, 3). Thisclosely parallels Gartner (1988) who arbitrarily sep-arated Gephyrocapsa by size into ‘large’ (>3 µm)and ‘small’ (<3 µm) groups.

66 A. Lototskaya et al. / Marine Micropaleontology 34 (1998) 47–70

Table 1List of species=morphotypes and their distinctive features used for the present study. The bridge angle is expressed relative to the longaxis of a placolith. The bar-charts represent the size distribution per defined species=morphotype. Drawings 1, 3–7 are reproduced afterBreheret (1978); 2 is modified after Breheret (1978)

All Gephyrocapsa forms with a completely orpartially closed central area, irrespective of size, arereferred to G. caribbeanica (Table 1; Plate I, 5; Hayet al., 1967).

To summarize, in the present study we recognisethe following forms: G. oceanica, G. muellerae, G.caribbeanica, and the ‘small’ Gephyrocapsa mor-photype (Table 1). This distinction of Gephyrocapsaforms is close to the routine use by many researchers.If compared with the classification suggested byBollmann (1997) based on the two parameters (sizeand bridge angle), G. muellerae could be placed

within three morphotypes (>2.4 µm with angle<27º, 2.4–3.1 µm with angle 27–56º, >3.1 µm withangle 27–56º), G. oceanica is related to the Gephy-rocapsa ‘Big’ morphotype (>3.9 µm with angle>56º), ‘small’ Gephyrocapsa is associated with threedifferent morphotypes (<2.4 µm, >2.4 µm with an-gle <27º, and 2.4–3.1 µm with angle 27–56º).

Acknowledgements

We thank P.H. Willekes and R. van Elsas for theirassistance in analysing the carbonate content of the

A. Lototskaya et al. / Marine Micropaleontology 34 (1998) 47–70 67

Fig. 10. Size distribution of Gephyrocapsa species in two sam-ples from core T90-9P.

samples and determining the magnetic susceptibility.We are also greatful to S. Kars for help with theSEM study and for taking photographs. R. Kaandorpis thanked for his contribution to the whole-coreoxygen isotope analysis. We thank Nicky Hine andtwo anonymous reviewers who commented on a

previous draft of the manuscript. This study is fundedby the program “Disturbances of the Earth system”as a part of the Dutch National Scientific Foundation(grant no. 77018242). This is publication no. 970174of the Netherlands Research School of SedimentaryGeology, Free University, Amsterdam.

Appendix A. Taxonomic appendix

References for the original descriptions of the followingspecies, except for Kleijne (1992), can be found in Perch-Nielsen(1985).

Acanthoica Lohmann emend. Schiller, 1913,emend. Kleijne, 1992

Calcidiscus leptoporus (Murray and Blackman, 1898)Loeblich and Tappan, 1978

Calciosolenia murrayi Gran, 1912Coccolithus pelagicus (Wallich, 1877) Schiller, 1930Emiliania huxleyi (Lohmann, 1902) Hay and Mohler, 1967Florisphaera profunda Okada and Honjo, 1973Gephyrocapsa caribbeanica Boudreaux and Hay, 1967Gephyrocapsa muellerae Breheret, 1978Gephyrocapsa oceanica Kamptner, 1943‘small’ GephyrocapsaHelicosphaera carteri (Wallich, 1877) Kamptner, 1954Neosphaera coccolithomorpha Lecal-Schlauder, 1950Pontosphaera Lohmann, 1902Rhabdosphaera clavigera Murray and Blackman, 1898Scyphosphaera apstenii Lohmann, 1902Syracosphaera Lohmann, 1902Syracosphaera pulchra Lohmann, 1902Umbellosphaera tenuis (Kamptner, 1937) Paasche, 1955Umbilicosphaera sibogae (Weber-Van Bosse, 1901)

Gaarder, 1970

References

Be, A.W.H., 1977. An ecological, zoogeographic and taxo-nomic review of recent planktonic foraminifera. In: Ramsey,A.T.S. (Ed.), Oceanic Micropaleontology. Academic Press,New York, NY, pp. 1–100.

Berger, W.H., Roth, P.H., 1975. Oceanic Micropaleontology:Progress and Prospect. Rev. Geophys. Space Phys. 13 (3),561–585.

Bollmann, J., 1997. Morphology and biogeography of the genusGephyrocapsa in Holocene sediments. Mar. Micropaleontol.29, 319–350.

Bond, G., Heinrich, H., Broecker, W., Labeyrie, L., McManus,J., Andrews, J., Huon, S., Jantschik, R., Clasen, S., Simet, C.,Tedesco, K., Klas, M., Bonani, G., Ivy, S., 1992. Evidence formassive discharges of icebergs into the North Atlantic oceanduring the last glacial period. Nature 360, 245–249.

Bramlette, M.N., 1958. Significance of coccolithophorids in cal-cium-carbonate deposition. Bull. Geol. Soc. Am. 69, 121–126.

68 A. Lototskaya et al. / Marine Micropaleontology 34 (1998) 47–70

Breheret, J.G., 1978. Formes nouvelles quaternaires et actuellesde la famille des Gephyrocapsaceae (Coccolithophorides). C.R. Acad. Sci. Paris 287, 447–450.

Broecker, W.S., 1994. Massive iceberg discharges as triggers forglobal climate change. Nature 372, 421–424.

Broecker, W.S., Van Donk, J., 1970. Insolation changes, icevolumes, and the δ18O record in deep-sea cores. Rev. Geophys.Space Phys. 8, 169–198.

Broecker, W., Bond, G., Klas, M., Clark, E., McManus, J., 1992.Origin of the northern Atlantic’s Heinrich events. Clim. Dyn.6, 265–273.

Broerse, A.T.C., Ziveri, P., Van Hinte, J.E., Honjo, S., submitted.Coccolithophore export production, seasonal species compo-sition, and coccolith–CaCO3 fluxes in the NE Atlantic (34ºN21ºW and 48ºN 21ºW). Deep-Sea Res. II, Topical Studies inOceanography, “Particle flux and its preservation in deep-seasediments”.

Bruin, P., 1937. Enige ervaringen bij de bepaling van het gehaltevan grond aan koolzure kalk volgens de methode Scheibler.Chem. Weekbl. 34, 755–759.

Brummer, G.J.A., Van Eijden, A.J.M., 1992. ‘Blue-ocean’ paleo-productivity estimates from pelagic carbonate mass accumula-tion rates. Mar. Micropaleontol. 19, 99–117.

Cortijo, E., Duplessy, J.C., Labeyrie, L., Leclaire, H., Duprat, J.,Van Weering, T.C.E., 1994. Eemian cooling in the NorwegianSea and North Atlantic ocean preceding continental ice-sheetgrowth. Nature 372, 446–449.

Cortijo, E., Reynaud, J.Y., Labeyrie, L., Paillard, D., Lehman,B., Cremer, M., Grousset, F., 1995. Etude de la variabilite cli-matique a haute resolution dans des sediments de l’AtlantiqueNord. C. R. Acad. Sci. Paris Ser. IIa 321, 231–238.

Dowdeswell, J.A., Maslin, M.A., Andrews, J.T., McCave, I.N.,1995. Iceberg production, debris rafting, and the extent andthickness of Heinrich layers (H-1, H-2) in North Atlanticsediment. Geology 23 (4), 301–304.

Duplessy, J.-C., Delibrias, G., Turon, J.L., Pujol, C., Duprat, J.,1981. Deglacial warming of the northeastern Atlantic Ocean.Correlation with the paleoclimatic evolution of the Europeancontinent. Palaeogeogr., Palaeoclimatol., Palaeoecol. 35, 121–144.

Duplessy, J.C., Labeyrie, L., Arnold, M., Paterne, M., Duprat, J.,Van Weering, T.C.E., 1992. Changes in surface salinity of theNorth Atlantic Ocean during the last deglaciation. Nature 358,485–487.

Ganssen, G., the shipboard party, 1990. Shipboard report JGOFSLeg 4, M.S. Tyro, Reykjavik–Galway, 2–29 June 1990. Inter-nal Report, Free Univ., Amsterdam, 164 pp.

Gard, G., 1989. Variations in coccolith assemblages during thelast glacial cycle in the high and mid-latitude Atlantic andIndian oceans. In: Crux, J.A., Van Heck, S.E. (Eds.), Nanno-fossils and their Applications. British MicropalaeontologicalSociety Series. Ellis Horwood, pp. 108–121.

Gartner, S., 1988. Paleoceanography of the Mid-Pleistocene.Mar. Micropaleontol. 13, 23–46.

Geitzenauer, K.R., Roche, M.B., McIntyre, A., 1977. Coccolithbiogeography from North Atlantic and Pacific surface sedi-

ments. In: Ramsey, A.T.S. (Ed.), Oceanic Micropaleontology.Academic Press, New York, NY, pp. 973–1008.

Grousset, F.E., Labeyrie, L., Sinko, J.A., Cremer, M., Bond, G.,Duprat, J., Cortijo, E., Huon, S., 1993. Patterns of ice-rafteddetritus in the glacial North Atlantic (40–55ºN). Paleoceanog-raphy 8 (2), 175–192.

Guiot, J., Pons, A., De Beaulieu, J.L., Reille, A., 1989. A140,000 year continental climate reconstruction from two Eu-ropean pollen records. Nature 338, 309–313.

Hay, W.W., Beaudry, F.M., 1973. Calcareous Nannofossils Leg15. Init. Rep. DSDP 15, 625–683.

Hay, W.W., Mohler, H.P., Roth, P.H., Schmidt, R.R., Boudreaux,J.E., 1967. Calcareous nannoplankton zonation of the Ceno-zoic of the Gulf Coast and Caribbean–Antillean area, andtransoceanic correlation. Gulf Coast Assoc. Geol. Soc. Trans.17, 428–480.

Heinrich, H., 1988. Origin and consequences of cyclic ice raftingin the northeast Atlantic ocean during the past 130,000 Years.Quat. Res. 29, 142–152.

Hine, N.M., Weaver, P.P.E., 1998. Quaternary, In: Brown, P.R.(Ed.), Calcareous nannofossil biostratigraphy. Chapman andHall, pp. 266–283.

Holligan, P.M., Fernandez, E., Aiken, J., Balch, W.M., Boyd, P.,Burkill, P.H., Finch, M., Groom, S.B., Malin, G., Muller, K.,Purdie, D.A., Robinson, C., Trees, C.C., Turner, S.M., Van derWal, P., 1993. A biogeochemical study of the coccolithophore,Emiliania huxleyi, in the North Atlantic. Global Biogeochem.Cycles 7 (4): 879–900.

Honjo, S., Manganini, S.J., 1993. Annual biogenic particle fluxesto the interior of the North Atlantic Ocean; studied at 34ºN21ºW and 48ºN 21ºW. Deep-Sea Res. 40 (1–2), 587–607.

Honjo, S., Okada, H., 1974. Community structure of coccol-ithophores in the photic layer of the mid-Pacific. Micropaleon-tology 20 (2), 209–230.

Jordan, R.W., 1988. Coccolithophorid communities in the north-east Atlantic. Ph.D. Thesis, Univ. Surrey (unpubl.).

Jordan, R.W., Zhao, M., Eglinton, G., Weaver, P.P.E., 1996.Coccolith and alkenone stratigraphy and palaeoceanographyat an upwelling site off NW Africa (ODP 658C) during thelast 130,000 years. In: Moguilevsky, A., Whatley, R. (Eds.),Microfossils and Oceanic Environments. Univ. Wales, Aberys-twyth Press, UK, pp. 111–130.

Kamptner, E., 1943. Zur Revision der Coccolithineen-SpeziesPontosphaera huxleyi Lohm. Akad. Wiss. Wien, Math. Nat.Kl. Anz. 80, 43–49.

Kleijne, A., 1992. Extant Rhabdosphaeraceae (coccolithophorids,class Prymnesiophyceae) from the Indian Ocean, Red Sea,Mediterranean Sea and North Atlantic Icean. Scr. Geol. 100,1–63.

Knappertsbusch, M., Brummer, G.-J.A., 1995. A sediment trapinvestigation of sinking coccolithophorids in the North At-lantic. Deep-Sea Res. 42 (7), 1083–1109.

Lehman, S.J., Keigwin, L.D., 1992. Sudden changes in NorthAtlantic circulation during the last deglaciation. Nature 356,757–762.

Lototskaya, A., 1994. Late Quaternary calcareous nannoplanktonfrom the Alboran Sea and Olympi mud-diapiric area (the

A. Lototskaya et al. / Marine Micropaleontology 34 (1998) 47–70 69

Mediterranean). Diploma Thesis, Moscow State Univ., 89 pp.(in Russian).

MacAyeal, D.R., 1993. Binge=purge oscillations of the Lauren-tide ice sheet as a cause of the North Atlantic’s Heinrichevents. Paleoceanography 8 (6), 775–784.

Manighetti, B., McCave, I.N., Maslin, M., Shackleton, N.J.,1995. Chronology for climate change: Developing age modelsfor the Biogeochemical Ocean Flux Study cores. Paleoceanog-raphy 10 (3), 513–525.

Martini, E., 1971. Standard Tertiary and Quaternary calcare-ous nannoplankton zonation. In: Farinacci, A. (Ed.), Proc. IIPlanktonic Conference, Roma 1970, 2, pp. 739–785.

Martinson, D.G., Pasias, N.G., Hays, J.D., Imbrie, J., Moore,T.C., Shackleton, J.J., Shackleton, N.J., 1987. Age datingand the orbital theory of the ice ages: Development of ahigh-resolution 0 to 300,000-year chronostratigraphy. Quat.Res. 27, 1–29.

McCave, I.N., 1988. Biological pumping upwards of the coarsefraction of deep-sea sediments. J. Sediment. Petrol. 58 (1),148–158.

McIntyre, A., Be, A.W.H., 1967. Modern Coccolithophoridae ofthe Atlantic Ocean — I. Placoliths and cyrtoliths. Deep-SeaRes. 14, 561–597.

McIntyre, A., McIntyre, R., 1970. Coccolith concentrations anddifferential solution in oceanic sediments. In: Funnel, B.M.,Riedel, W.R. (Eds.), Micropaleontology of the Oceans. Cam-bridge Univ. Press, Cambridge, pp. 253–261.

McIntyre, A., Be, A.W.H., Roche, M.B., 1970. Modern PacificCoccolithophorida: a paleontological thermometer. Trans. N.Y. Acad. Sci. Ser. II 32 (6), 720–731.

McIntyre, A., Ruddiman, W.F., Jantzen, R., 1972. Southwardpenetrations of the North Atlantic Polar Front: faunal andfloral evidence of large-scale surface water mass movementsover the last 225,000 years. Deep-Sea Res. 19, 61–77.

Molfino, B., McIntyre, A., 1990. Precessional forcing of nutri-cline dynamics in the Equatorial Atlantic. Science 249, 766–769.

Newton, P.P., Lampitt, R.S., Jickells, T.D., King, P., . Boutle, C.,Deep-Sea Res. 41 (11/12), 1617–1642.

Okada, H., Honjo, S., 1973. The distribution of oceanic coccol-ithophorids in the Pacific. Deep-Sea Res. 20, 355–374.

Okada, H., Honjo, S., 1975. Distribution of coccolithophores inmarginal seas along the western Pacific Ocean and in the RedSea. Mar. Biol. 31, 271–285.

Okada, H., McIntyre, A., 1977. Modern coccolithophores of thePacific and North Atlantic oceans. Micropaleontology 23 (1),1–55.

Okada, H., McIntyre, A., 1979. Seasonal distribution of moderncoccolithophores in the western North Atlantic Ocean. Mar.Biol. 54, 319–328.

Oppo, D.W., Horowitz, M., Lehman, S.J., 1997. Marine coreevidence for reduced deep water production during Termina-tion II followed by a relatively stable substage 5e (Eemian).Paleoceanography 12 (1), 51–63.

Ottens, J., Van Kreveld, S., Ganssen, G., Van Hinte, J.E., 1995.Evaluating the role of the biological pump in the NortheastAtlantic through paleo primary productivity reconstruction. In:

Zwerver, S., Van Rompaey, R.S.A.R. et al. (Eds.), ClimateChange Research: Evaluation and Policy Implications. (Stud-ies in Environmental Science, 65A.) Elsevier, pp. 331–338.

Perch-Nielsen, K., 1985. Cenozoic calcareous nannofossils. In:Bolli, H.M., Saunders, J.B., Perch-Nielsen, K. (Eds.), PlanktonStratigraphy. Cambridge Univ. Press, Cambridge, pp. 427–554.

Pielou, E.C., 1966. The measurement of diversity in differenttypes of biological collections. J. Theor. Biol. 13, 131–144.

Pujos, A., Giraudeau, J., 1993. Repartition des Noelaerhab-daceae (nannofossiles calcaires) dans le Quaternaire moyenet superieur des oceans Atlantique et Pacifique. Oceanol. Acta16 (4), 349–362.

Revel, M., Sinko, J.A., Grousset, F.E., Biscaye, P.E., 1996. Srand Nd isotopes as tracers of North Atlantic lithic particles:Paleoclimatic implications. Paleoceanography 11 (1), 95–113.

Robinson, S.G., Maslin, M.A., McCave, I.N., 1995. Magneticsusceptibility variations in Upper Pleistocene deep-sea sedi-ments of the NE Atlantic: Implications for ice rafting andpaleocirculation at the last glacial maximum. Paleoceanogra-phy 10 (2), 221–250.

Roth, H.P., Berger, W.H., 1975. Distribution and dissolution ofcoccoliths in the South and Central Pacific. In: Silter, W.V.,Be, A.W.H., Berger, W.H. (Eds.), Dissolution of Deep-seaCarbonates. Cushman Found. Spec. Publ. 13, 87–113.

Ruddiman, W.F., 1977. Late Quaternary deposition of ice-raftedsand in the subpolar North Atlantic (lat 40º to 65ºN). Geol.Soc. Am. Bull. 88, 1813–1827.

Ruddiman, W.F., Glover, L.K., 1972. Verical mixing of ice-raftedvolcanic ash in North Atlantic sediments. Geol. Soc. Am. Bull.83, 2817–2836.

Ruddiman, W.F., McIntyre, A., 1976. Northeast Atlantic paleo-climatic changes over the past 600,000 Years. Geol. Soc. Am.Mem. 145, 111–146.

Ruddiman, W.F., Sancetta, C.D., McIntyre, A., 1977.Glacial=interglacial response of subpolar North Atlantic watersto climatic change: the record in oceanic sediments. Philos.Trans. R. Soc. London Bull. 280, 119–142.

Samtleben, C., 1980. Die Evolution der Coccolithophoriden-Gat-tung Gephyrocapsa nach Befunden im Atlantik. Palaontol. Z.54 (1–2), 91–127.

Samtleben, C., Bickert, T., 1990. Coccoliths in sediment trapsfrom the Norwegian Sea. Mar. Micropaleontol. 16, 39–64.

Samtleben, C., Baumann, K.-H., Schroder-Ritzrau, A., 1995.Distribution, composition and seasonal variation of coccol-ithophore communities in the northern North Atlantic. In: Flo-res, J.A., Sierro, F.J. (Eds.), Proceedings 5th INA Conference,Salamanca, Univ. Salamanca, pp. 219–235.

Sarnthein, M., Tiedemann, R., 1990. Younger Dryas-style cool-ing event at glacial terminations I–VI at ODP site 658: asso-ciated benthic δ13C anomalies constrain meltwater hypothesis.Paleoceanography 5 (6), 1041–1055.

Sarnthein, M., Jansen, E., Arnold, M., Duplessy, J.-C., Er-lenkeuser, H., Flatoy, A., Veum, T., Vogelsang, E., Weinelt,M.S., 1992. δ18O time-scale reconstructions of meltwateranomalies at Termination I in the North Atlantic between50 and 80ºN. In: Bard, E., Broecker, W.S. (Eds.), The Last

70 A. Lototskaya et al. / Marine Micropaleontology 34 (1998) 47–70

Deglaciation: Absolute and Radiocarbon Chronologies. NATOASI Ser. I, 2, 183–200.

Sarnthein, M., Winn, K., Jung, S.J.A., Duplessy, J.-C., Labeyrie,L., Erlenkeuser, H., Ganssen, G., 1994. Changes in east At-lantic deepwater circulation over the last 30,000 years: Eighttime scale slice reconstructions. Paleoceanography 9 (2), 209–267.

Thierstein, H.R., Geitzenauer, K.R., Molfino, B., Shackleton,N.J., 1977. Global synchroneity of late Quaternary coccolithdatum levels: validation by oxygen isotopes. Geology 5, 400–404.

Van Kreveld, S., Knappertsbusch, M., Ottens, J., Ganssen, G.M.,Van Hinte, J.E., 1996. Biogenic carbonate and ice-rafted debris(Heinrich layers) accumulation in deep-sea sediments from aNortheast Atlantic piston core. Mar. Geol. 131, 21–46.

Violanti, D., Grecchi, G., Castradori, D., 1991. Paleoen-vironmental interpretation of core BAN88-11GC (EasternMediterranean, Pleistocene–Holocene) on the grounds ofForaminifera, Thecosomata and calcareous nannofossils. Qua-ternario 4 (1a), 13–39.

Weaver, P.P.E., Pujol, C., 1988. History of the last deglaciationin the Alboran Sea (Western Mediterranean) and adjacentNorth Atlantic as revealed by coccolith floras. Palaeogeogr.,Palaeoclimatol., Palaeoecol. 64, 35–42.

Westbroek, P., Van Hinte, J.E., Brummer, G.-J., Veldhuis, M.,

Brownlee, C., Green, J.C., Harris, R., Heimdal, B.R., 1994.Emiliania huxleyi as a key to biosphere-geosphere interactions.In: Green, J.C., Leadbeater, B.S.C. (Eds.), The HaptophyteAlgae. Syst. Assoc. Spec. Vol. 51, 321–334.

Winter, A., 1982. Paleoenvironmental interpretation of Quater-nary coccolith assemblages from the Gulf of Aqaba (Elat),Red Sea. Rev. Esp. Micropaleontol. XIV, 291–314.

Winter, A., Reiss, Z., Luz, B., 1979. Distribution of livingcoccolithophore assemblage in the Gulf of Elat (’Aqaba). Mar.Micropaleontol. 4, 197–223.

Winter, A., Jordan, R.W., Roth, P.H., 1994. Biogeography of liv-ing coccolithophores in ocean waters. In: Winter, A., Siesser,W. (Eds.), Coccolithophores. Cambridge Univ. Press, pp. 161–178.

Young, J.R., Westbroek, P., 1991. Genotypic variation in the coc-colithophorid species Emiliania huxleyi. Mar. Micropaleontol.18, 5–23.

Ziveri, P., Thunell, R.C., Rio, D., 1995. Seasonal changes incoccolithophore densities in the Southern California Bightduring 1991–1992. Deep-Sea Res. 42 (11/12), 1881–1903.

Ziveri, P., Broerse, A.T.C., Van Hinte, J.E., Honjo, S., submitted.The fate of the coccoliths at 48ºN 21ºW, northeast Atlantic.Deep-Sea Res. II, Topical Studies in Oceanography, “Particleflux and its preservation in deep-sea sediments”.