QUATERNARY RESEARCH 29, 142-152 (1988)

Origin and Consequences of Cyclic Rafting in the Northeast Atlantic Ocean during the Past 130,000 Years

HARTMUT HEINRICH Deutsches Hydrographisches Institut, Bernhard-Nocht-Str. 78, D-2000 Hamburg 4, West Germany Received April 7, 1987 Deep-sea cores recovered from the Northeast Atlantic Ocean were examined in order to elucidate the influence of the Earth’s orbital parameters on major ice rafting. Analyses of coarse-grained ice-rafted debris and planktonic foraminifers revealed a strong reaction to the precession signal. Since 130,000 yr B.P., dropstone layers have been deposited each half period of a precessional cycle (11,000 2 1000 yr). Ice rafting occurs during times of winter minimum/summer maximum insolation and summer minimum/winter maximum insolation. In the first case, high summer insolation forces meltwater discharge from the ice sheets into the polar seas which sub- sequently enhances formation of during the winter. In the second case, growth of conti- nental ice enhances production which also leads to a salinity reduction of surface seawater. Both situations result in a southward penetration of polar . Thus, the marine record of dropstones documents ice rafting not only during Weichselian stades but also during cold events within interstades. The regularity of ice rafting yields a useful framework to calibrate and elucidate climatic changes, not only in the region of the North Atlantic Ocean but also in remote areas such as the Pacific Ocean and the Antarctic. o 1988 LJII~VCAY of Washington.

INTRODUCTION Hays et al., 1976; Berger, 1979), and is now The oceanographic and climatic history widely accepted. In order to test the orbital of the North Atlantic Ocean during the Late influence on climate, cross-spectral analy- has been fairly well known ses between astronomical and sedimentary since the publication of the CLIMAP re- parameters were carried out (e.g., Herter- sults (Cline and Hays, 1976). Extensive pet- ich and Sarnthein, 1984). Nevertheless, rographic measurements carried out on there are many attempts to correlate certain hundreds of North Atlantic sediment cores variations of climate with other terrestrial and modem data processing gave quantita- or extraterrestrial mechanisms (e.g., volca- tive information about the climatic re- nic activity and solar modulation). sponse in this area. Since that time, most The attempts to correlate insolation and efforts have focused on two tasks: to deci- Earth’s climatic responses as documented pher the climatic processes at times of dis- in deep-sea sediment from numerous petro- tinct oceanwide environmental changes like graphic parameters suffer from the lack of a glaciation and deglaciation (e.g., Ruddiman precise and continuous chronology in the et al., 1980; Ruddiman and McIntyre, sedimentary successions. Numerical ages 1981a) and to resolve the interrelationships are often scarce. Correlations with well- among the ocean, the atmosphere, and the dated sediment cores and interpolation are continents (e.g., Rind et al., 1986). often affected by various influences during The search for the causal mechanisms and after deposition, which may have al- that initiate climatic processes began much tered the climatic signal (Ruddiman and earlier with Milankovitch’s (1930) theory of Glover, 1972; Schiffelbein, 1984). Thus, climate control by insolation due to dating the climate history is primarily a changes in the geometry of the Earth’s or- question of sampling optimum sedimentary bit. His hypothesis was subsequently sup- successions. ported by a number of investigators (e.g, The verification of insolation control on

142 0033-5894/88 $3.00 Copyright 0 1988 by the University of Washington. All rights of reproduction in any form reserved. NORTH ATLANTIC ICE RAFTING 143 ice rafting requires sampling in areas with tyre, 1976). In general, the on the strong climatic contrasts and a set of sensi- deep-sea hill (the Dreizack) and in its vicin- tive indicators. The location should act as a ity consist of intercalated calcareous oozes filter; that is, it should be touched only by and poorly calcareous sandy muds. In the the stronger pulses of ice rafting. Further- Maury turbidity current channel, the pe- more, the site should be sensitive not only lagic sedimentary sequence is interrupted to cold but also to warm excursions of the by thick dark volcaniclastic turbidite layers climate. These conditions are well fulfilled (Heinrich, 1986b). Submarine landslides, in the northeastern Atlantic Ocean between consisting of more than 4 m of semiconsol- latitudes 45” and 50” (Ruddiman and idated breccias, occur on each side of the Glover, 1975; Ruddiman and McIntyre, Dreizack. 1976; Ruddiman, 1977). A set of 13 piston cores was collected from the Dreizack and the surrounding MATERIAL AND METHODS deep plain (Fig. 1; Table 1). One Kasten In order to avoid impairments from car- core was taken from an elevated plateau 25 bonate dissolution and turbidite supply nautical miles northeast of the Dreizack. from the distal Maury Channel system (Em- The cores were cut and X-ray photo- ery and Uchupi, 1984), sampling had to be graphs were made in order to avoid sub- carried out on deep-sea hills. The area of sampling in burrows of benthic organisms. investigation (Fig. 1) was carefully mapped Subsamples were taken each 3 cm and with SEA BEAM, 3.5 kHz Subbottom Pro- washed through a 180~pm sieve. Splits of filer, and KAE Parasound (Heinrich, 300 to 600 grains of the coarse fraction were 1986a, b). Water depths range from 3900 to counted for fragments of volcanic glass 4550 m. There are no obvious differences shards, quartz, volcanic rock, sedimentary from those sediments already known in the rock, and crystalline rock. Fragments Northeast Atlantic (Ruddiman and McIn- >3000 pm in diameter were neglected.

470 47O I I’ 2001 II’ 10’ ,I’ ‘0’ 35’ 30’ 25’ 29’ II’ IO’ I’ lcp I II’ 10’ FIG. 1. Bathymetric map (SEA BEAM) of the area around the Dreizack seamount (at 19”4O’W, 47”23’N) showing core locations (Table 1). The deep plain is temporarily passed by Maury Channel turbidity currents from west to east. Contour interval = 50 m. The insert shows the position of the NOAMP site in the West-European Basin. 144 HARTMUT HEINRICH

TABLE 1. BOTTOM SAMPLINGDATA Water depth Core length Core No. N Lat. W Long. (m) (cm) Corer Me68-02 47O20.8’ lY48.6’ 4510 450 PC6 Me68-09 47O26.3’ lY34.0’ 4260 434 PC6 Me68-11 47O26.6’ lY35.5’ 4470 549 PC6 Me68-14 47”28.3’ lY43.4’ 4545 379 PC6 Me68-15 47Y26.8’ lY36.1’ 4500 1149 PC12 Me69-17 47O21.4’ 19’42.9’ 3905 914 PC12 Me69-18 47O20.7’ lY43.3’ 4120 1035 PC12 Me69-19 47Ol9.4’ 19O41.3’ 4350 1080 PC12 PoOS-26 47Ol4.7’ lY37.1’ 4552 1014 PC12 PoO8-28 47”lO.l’ lY35.8’ 4557 2088 PC24 PoOS-29 47”24.7’ PY32.9’ 3900 968 PC12 PoOS-30 47O24.2’ 19O22.5’ 4550 2238 PC24 PoOS-3 1 47O22.8’ 19”38.0’ 4030 835 PC12 MOl-32 47”34.7’ 1856.0’ 4070 360 KC6

Nore. Abbreviations: Me, R.V. METEOR (old); PO, R.V. POLARSTERN; M, R.V. METEOR (new): PC, piston corer; KC, Kasten corer.

Dating was by oxygen-isotope stratigra- and paleomagnetic analyses revealed nu- phy based on measurements in core Me69- merous discordances. Because all hiatuses 17. We measured tests of Pyrgo murrhenia were older than the oxygen-isotope stage and Gfobigerina bulloides (Fig. 2). Since 6/5 boundary, investigations of ice rafting benthic foraminifers are very rare in the glacial sediments, a satisfactory deep-water record is not available. An improvement of the oxygen-isotope curve is still in prog- ress. The ages of the boundaries were taken from Herterich and Sarnthein (1984). In addition, two concentrations of sili- ceous glass shards were identified as Ash Zone I and Ash Zone II of Bramlette and Bradley (1941). Ruddiman and McIntyre (1984) dated Ash Zone I as 9800 yr B.P. Mangerud et al. (1984) correlate this zone with a Norwegian ash layer dated to 10,600 yr B.P. Ash Zone II was estimated to be 65,000 yr old (Ruddiman and Clover, 1972). RESULTS Three cores from different physiographic situations in the Dreizack region were taken as representatives (Fig. 1): steep hill

crest (Me69-17), flat hill crest or plateau 5c (MOl-32), and lower slope (Me69-19). 6 1' FIG. 2. Oxygen-isotope record of core Me69-17. Stratigraphy Measurements were carried out on the benthic species Pyrgo murrhenia and the planktic Globigerina bul- Although most of the Dreizack sediment loides. No benthic foraminifers were found within the seemed to be undisturbed in the radio- dropstone layers. The ages of the stage bounderies graphs (except the breccias), nannofossil originate from Herterich and Samthein (1984). NORTH ATLANTIC ICE RAFTING 145 were restricted to the last 140,000 yr. The the isotope stage 5/4 boundary all cores oxygen-isotope record of the species P. yield six distinct peaks (Figs. 3-5). Undis- murrhenia clearly revealed stage bound- turbed core tops and additional analyses of aries 2/l, 5al4, 5e/5d, and 615 (Fig. 2). Stage precise surface samples of multicorers re- 5a/4 boundary has been assigned to the G. veal a seventh enrichment of quartz, on the inflata maximum beneath IRD peak 6 (SST order of 5 to lo%, in the surface sediment. maximum according to Ruddiman ef al., This coincides with the occurrence of large 1980). The record of stages 2 to 4 was in- faceted pebbles on the sediment surface ob- sufftcient due to the very sparse occurrence served in box cores and bottom photo- of benthic tests. Ash layers I and II were graphs. obvious in all cores. In oxygen-isotope stage 5, the number of clearly visible peaks is reduced to one in the Ice-Rafted Debris (ZRD) upper part of this stage, a slight enrichment The majority of the ice-rafted particles in at the 5e/5d boundary, and a large one at the the fraction 180 to 3000 km consists of 615 boundary. Although the cores originate transparent angular quartz grains. Few of from different physiographic positions, the these grains are well-rounded and shape of the IRD peaks is rather regular. scratched. Fragments of plutonic and vol- The uppermost four peaks are always very canic rocks are rare. Fragments of sedi- large, whereas the fifth peak is a narrow mentary rocks occur only in the fraction one. The peak between Ash Layer II and >3000 pm. the isotope stage 514 boundary has a very The record of ice-rafted debris is nearly broad base. The peaks of stage 5 are gen- uniform in all cores. From the top down to erally weaker than the younger ones. While

FORAMINIFERS(%)

FIG. 3. Core record of Me69-17 from the western peak of the Dreizack. The amount of ice-rafted debris is given as part of the total split (IRD + forams). The amount of one species is given as part of the sum of planktonic forams in a split. The right side of the IRD record shows the stratigraphic markers (Ash I and II, oxygen-isotope stage boundaries). 146 HARTMUT HEINRICH

ICE-RAFTED PEERIS (X) 0 FORAM15~IFERS (%) 0 I I , I 1 __ -___------_--~--______-AAs). 1 / I 3

-N. pochydermofs I -9% ,“,,a,0 l -+G.,r”nc.tulr”o~d.s

FIG. 4. Core record of Me69-19 (from the southwestern foot of the Dreizack). See Figure 2 for modes of calculation. the peak in the upper part of stage 5 is al- abundant. G. inflata culminates between ways easy to recognize, there is only a very the N. pachyderma (s.) peaks. The highest weak enrichment of quartz grains above the abundances occur shortly before substage 5e/5d boundary. 5e/5d boundary. Subtropical to tropical species, especially G. truncatulinoides, are Foraminifera common in the Holocene and throughout Seven species of planktic foraminifera isotope stage 5. They culminate in the mid- were used to characterize surface cold wa- Holocene and in early substage 5e. The pre- ter: N. pachyderma (s.) and G. bulloides; cise surface samples reveal a strong de- surface temperate water: G. inflata; and crease since the mid-Holocene maximum. surface warm water: G. truncatulinoides, It is worthy of note that subtropical species G. scitula, G. ruber (w), and G. hirsuta. occurred in the time shortly after the depo- The results based on these species are sition of Ash layer II. given in Figures 3-5. The foraminiferal rec- The sediment accumulation rates of the ord does not present such a clear picture as cores (Fig. 6) have been rather constant the IRD. In general, the most abundant spe- since the isotope stage 6/5 boundary: 2 cm/ cies in the grain-size fraction >180 urn are 1000 yr in Me69-17 and in Me69-19 and 3 G. bulloides, N. pachyderma (d.), and G. cm/1000 yr in MOl-32, although sedimenta- infla la. tion varied repeatedly from calcareous Although these curves are rather saw- ooze to clayey quartz sand. toothed in shape, there are clear corre- sponding features in all cores. N. pachy- DISCUSSION derma (s.) matches nearly perfectly with The positions of the IRD peak maxima in the IRD record. Where the amount of IRD the time/depth graphs show that the respec- is high, the polar representative is also tive peaks correlate fairly closely between NORTH ATLANTIC ICE RAFTING 147 the cores (Fig. 6). Since all cores were re- 270”, the summer insolation is elevated and covered from the same area of the north- the winter insolation is reduced (Vernekar, east Atlantic Ocean, each set of peaks 1972; Berger, 1978). Thus, there are two probably has the same age. Thus, the age opportunities of enhanced ice rafting in the differences between the respective peaks northeast Atlantic Ocean during one cycle observed originate from local variations in of the Earth’s precession (i.e., every 11,000 accumulation rate and from uncertainties in +- 1000 yr). the exact depths of the IRD maxima in the Two insolation minima are not represent- cores. ed by dropstone layers. Their estimated po- The search for true ages of the IRD lay- sitions in the core records are marked by ers, and for the trigger of the respective ice italic numbers 7 and 9 in Figures 3-5. rafting, leads to a comparison of the esti- These two peaks belong to the precessional mated ages deduced from the accumulation angle II = 270 at 83,000 and 103,000 yr B .P. rate and of the variations of the solar radi- Assuming that the mean temperature was ation received on Earth. As can be seen in relatively warm during those times (as indi- Figure 7, phases of intense ice rafting cated by the occurrence of G. truncatuli- match perfectly with the precessional influ- nodes), it seems that ice rafting was re- ence on Northern Hemisphere insolation. stricted to the polar seas. Nevertheless, the At times of a precessional angle of II = 90”, diminished insolation might have led to a the Northern Hemisphere receives a re- cooling of Northeast Atlantic surface wa- duced insolation during summer and a ters, which would have caused the retreat slightly increased insolation during winter. of temperate planktonic species in the fo- At times of a precessional angle of II = raminiferal core records. The decrease of G.

_I FIG. 5. Core record of MOl-32 (from the elevated plateau northeast of the Maury turbidite channel). See Figure 2 for modes of calculation. 148 HARTMUT HEINRICH infuta, which is best illustrated in core the winter insolation minima due to a strong Me69-17, could be related to this cooling. recessional response of the ice sheets upon Although ice rafting is obviously linked the controlling reduction of insolation to the precession, it is difficult to estimate (Hyde and Peltier, 1987). the exact dates and duration of these In contrast, the temporal linking of ice phases. If the modes of “ice-sheet decay” rafting and insolation at times of a preces- and “ice-sheet growth” are correct (Ruddi- sional angle of II = 90” is more difficult. man and McIntyre, 198 1b), enhanced sum- There has to be a certain delay between mer insolation initiates the decay of the ice ice-sheet construction and Northeast At- sheets (at times of II = 270). Large melt- lantic ice rafting because both can hardly water supply to the ocean reduces the sa- occur at the same time. linity of the surface water and favors freez- Ruddimen et al. (1980) infer that ice- ing during the long, cold winters. Thus, construction is initiated, among other which floats across the Northeast Atlantic reasons, by an intensified supply of warm during times of ice-sheet decay is mainly surface water into the high-latitude North sea ice containing little continental debris, Atlantic. During these times, with a strong with a subordinate influx of material trans- North Atlantic Drift, ice rafting across the ported by . This is emphasized by Northeast Atlantic Ocean seems rather im- the reduced width of the peaks with odd probable. In this area, ice rafting may occur numbers (Figs. 3-5). It seems reasonable to when enhanced discharges of cold, less- place these peaks very close to the times of saline surface water from the northern

AGE f103vr B.?)

FIG. 6. Age/depth correlation of the dropstone peaks. The apparent ages of the peaks are obtained by interpolation between the stratigraphic markers indicated by arrows. OISB, oxygen-isotope stage boundary. NORTH ATLANTIC ICE RAFTING 149 oceans bend the North Atlantic Drift to parameters (Saltzman et al., 1984), incep- more southerly directions. Indeed, in the tion and duration of a single ice-rafting stage 5/4 transition, the warm North Atlan- phase should be variable. Nevertheless, the tic Drift water rich in G. ittjlata was re- mean ages of the dropstone layers derived placed by cold Arctic water which brought from accumulation rates in relation to or- ice-rafted detritus and polar foraminifers bital data reveal a considerable coincidence down to lower latitudes (Ruddiman ef al., of insolation minimum and maximum ice 1980). rafting in the northeastern Atlantic Ocean A possible mechanism which might have (Fig. 7). initiated increased formation of cold sur- face water results from decreasing stabili- SUMMARY AND CONCLUSIONS ties of continental ice sheets during en- hanced ice accumulation (Reeh, 1985). This Since the Eemian interglaciation, major leads to a faster lateral progression of the Northeast Atlantic ice rafting has been pri- continental ice sheets and, subsequently, to marily controlled by the precession of the a greater discharge of ice into the ocean. Earth’s axis. It is linked to (a) summer and The fresh water supply from large numbers (b) winter insolation minima which occur of melting icebergs reduces the salinity of two times during one precessional cycle (at the oceanic surface water. Therefore, the precession angles (a) II = 90” and (b) Kl = oxygen-isotope ratios of G. bulloides tests 270”). Thus, the mean period of enhanced became low in the youngest Weichselian ice rafting is about 11,000 2 1000 yr. Incep- dropstone layers when ice rafting reached tion, length, and intensity of an ice-rafting its greatest extent (Fig. 2). phase depend on the interactions with such If the magnitude of ice-sheet growth is other basic parameters as obliquity and ec- primarily a function of interacting orbital centricity. The dropstone record indicates

INSOL$;ION M;oNIMA h&yr) ,20 0 20 40 0

FIG. 7. Relationship of dropstone ages (from Fig. 6) and times of insolation minima (from Vemekar, 1972). HARTM”T ----. .- - --- 150 HEINKICH ice rafting not only during stadial intervals with climatic situations during II = 270”. It but also more or less in the middle of inter- seems reasonable to correlate the cold stades. Younger Dryas oscillation with the preces- At times when the precessional angle II sional winter insolation minimum (Fig. 7). = 270”, Northern Hemisphere ice rafting is Warm climatic events (61,000 yr B.P.) re- probably coupled with enhanced formation corded in the Spitsbergen area off northern of winter sea ice which benefits from a Norway (Gard, 1986) can be attributed to forced meltwater supply into the ocean due the precessional phase with increased sum- to high summer insolation. At times when mer insolation. Mangerud et al. (1979) re- the precessional angle II = 90”, ice rafting port interstadial ice-sheet retreats in Scan- across the northeastern Atlantic Ocean dinavia between 38,000 and 28,00 yr B.P. seems to be a consequence of increased ice which correlate well with the period of en- discharge which results from continental hanced summer insolation at that time. ice-sheet growth due to reduced summer in- As can be seen in Figure 7, the distance solation in the higher latitudes. In both from the last major IRD deposition to the cases, the crucial point for major ice rafting present occupies one interval between two is the intensified formation of cold, less- phases of ice rafting. Indeed, the preces- saline surface water which favors the trans- sional angle II = 90” was reached in the port of ice down to the latitude of the Ibe- middle of the 14th century (Bouvier, 1983). rian Peninsula (McIntyre et al., 1972). This Therefore, it is more likely that the Little occurs essentially during times when II = (1350-1900 A.D.) is a very cool 90”. The southward penetration of polar part of the whole phase of reduced summer water then keeps warm North Atlantic Drift insolation than a consequence of enhanced water away from the northern part of the volcanic activity (Hammer et al., 1980; ocean and reduces the transport of mois- Dansgaard, 1984; Porter, 1986). ture to the ice sheets. The systematic changes of major North- ACKNOWLEDGMENTS ern Hemisphere ice rafting since the Ee- mian correspond with many other climate I thank the Masters and crews of both RVs Meteor records from different fields and regions. and RV Polarstern for their help. D. Meischner (Get- As illustrated by the New Guinea sea-level tingen) and C. Hemleben (Tiibingen) supported the in- curve (Bloom et al., 1974; Chappell and vestigation with fruitful discussions. H. Oberh$insli Veeh, 1978), relevant drops correlate with (Ziirich) carried out stable oxygen- and carbon-isotope analyses. R. Jantschik and R. Lohoff (Giittingen) car- the dropstone layers of the precessional an- ried out sampling and core descriptions. The NOAMP gle II = 90”. The winter insolation mini- study is funded by the Bundesminister fiir Forschung mum (precessional angle II = 270”) causes u. Technologie, No. KWA 5310. obviously minor sea-level drops because these phases of enhanced ice rafting coin- REFERENCES cide precisely with dated relative high sea Berger, A. L. (1978). Long-term variations of caloric stands which are, for example, indicated by insolation resulting from the Earth’s orbital eie- coral terraces on Barbados (Broecker et ments. Quaternary Research 9, 139-167. al., 1968). Berger, A. (1979). Spectrum of climatic variations and Furthermore, the improved dropstone their causal mechanisms. Geophysical Surveys 3, stratigraphy explains certain climatic 351402. events better than before. Two intervals of Bloom, A. L., Broecker, W. S., Chappell, J. A. M., Matthews, R. K., and Mesolella, K. J. (1974). Qua- enhanced “Be deposition in Antarctic ice ternary sea level fluctuations on a tectonic coast: at 60,000 and 35,000 yr B.P., which were New 230Th/234U dates from the Huon Peninsula, interpreted as changes in solar modulation New Guinea. Quaternary Research 4, 185-205. (Raisbeck et al., 1987), match perfectly Bouvier, J.-M. (1983). An astronomical approach to NORTH ATLANTIC ICE RAFTING 151

human prehistoric ecology: Quantitative insolation Mangerud, J., Larsen, E., Longva, O., and Sonste- at Late Wihm in south-west France. Actes Coll. gaard, E. (1979). Glacial history of Western Norway AGSO Bordeaux, Mai 1983, Bulletin de l’lnstitut de 15,000-10,000 B.P. Boreus 8, 179-187. Geologie du Bassin d’dquitaine, 34, and CNRS Ca- Mangerud, J., Lie, S. E., Fumes, H., Kristiansen, hiers du Quaternaire, Special edition, pp. 27-35. I. L., and Lomo, L. (1984). A Younger Dryas ash Bramlette, M. N., and Bradley, W. H. (1941). “Ge- bed in Western Norway, and its possible correlation ology and Biology of North Atlantic Deep-Sea with tephra in cores from the Norwegian Sea and the Cores between Newfoundland and Ireland: I. Li- North Atlantic. Quaternary Research 21, 85-104. thology and Geologic Interpretation.” U.S. Geolog- McIntyre, A., Ruddiman, W. F., and Jantzen, R. ical Survey Professional Paper, 196-A, pp. l-34. (1972). Southward penetration of the North Atlantic Broecker, W. S., Thurber, D. L., Goddard, J., Ku, Polar Front: Fauna1 and floral evidence of large- T.-L., Matthews, R. K., and Mesolella, K. J. scale surface water mass movements over the last (1968). Milankovitch hypothesis supported by pre- 225,000 years. Deep-Sea Research 19, 61-77. cise dating of coral reefs and deep-sea sediments. Milankovitch, M. (1930). Mathematische Klimalehre Science 159, 297-300. und astronomische Theorie der Klimaschwankun- Chappell, J., and Veeh, H. H. (1978). 230Th/234U age gen. In “Handbuch der Klimatologie” (W. Koppen support of an interstadial sea level of -40 m at and R. Geiger, Eds.), I(A). Bomtrager, Berlin. 30,000 yr B.P. Nature (London) 276, 602-603. Porter, S. C. (1986). Pattern and forcing of Northern Cline, R. M., and Hays, J. D., Eds. (1976). Investiga- Hemisphere variations during the last millen- tion of Late Quatemary paleoceanography and pa- nium. Quaternary Research 26, 27-48. leoclimatology. Geological Society of America Raisbeck, G. M., Yiou, F., Bourles, D., Lorius, C., Memoir 145. Jouzel, J., and Barkov, N. I. (1987). Evidence for Dansgaard, W. (1984). The future degree of glaciation. two intervals of enhanced IO-Be deposition in Ant- In “The Climate of Europe: Past, Present and arctic ice during the last glacial period. Nature (Lon- Future” (H. Flohn and R. Frantechi, Eds.), pp. 225- don) 326, 273-277. 248. Reidel, Dordrecht. Reeh, N. (1985). Was the Greenland ice sheet thinner Emery, K. D., and Uchupi, E. (1984). “The Geology in the late Wisconsin than now? Nature (London) of the Atlantic Ocean.” Springer-Verlag. New 317, 797-799. York. Rind, D., Peteet, D., Broecker, W. S., McIntyre, A., Gard, G. (1986). Calcareous nannofossil biostratigra- and Ruddiman, W. F. (1986). The impact of cold phy of Late Quatemary Arctic sediments. Boreas North Atlantic sea surface temperatures on climate: U(3), 217-230. Implications for the Younger Dryas cooling (1 l-10 Hammer, C. U., Clausen, H. B., and Dansgaard, W. k). Climate Dynamics l(l), 3-33. (1980). Greenland ice sheet evidence of post-glacial Ruddiman, W. F. (1977). Late Quatemary deposition volcanism and its climatic impact. Nature (London) of ice-rafted sand in the subpolar North Atlantic (lat 288, 230-235. 40 to 65). Geological Society of America Bulletin 88, Hays, J. D., Imbrie, J., and Shackleton, N. J. (1976). 1813-1827. Variations in the earth’s orbit: Pacemaker of the Ice Ruddiman, W. F., and Glover, L. K. (1972). Vertical Ages. Science 194, 1121-l 132. mixing of ice-rafted volcanic ash in North Atlantic Heinrich. H. (1986a). Bathymetrie und Geomorpholo- sediments. Geological Society of America Bulletin gie des NOAMP-Gebietes, Westeuropaisches 83(9), 2817-2835. Becken (17W bis 22W, 46N bis 49N). Deutsche Hy- drogruphische Zeitschrif 39, 183-196. Ruddiman, W. F., and Glover, L. K. (1975). Subpolar Heinrich, H. (1986b). A comparison of conventional North Atlantic circulation at 9300 yr BP: Fauna1 ev- ship-installed 3.5 kHz Sub Bottom Profiler and the idence. Quaternary Research 5, 361-389. new KAE “Parasound” illustrated by an acoustic Ruddiman, W. F., and McIntyre, A. (1976). Northeast mapping of a deep-sea meander. Deutsche Hydro- Atlantic paleoclimatic changes over the past 600,000 gruphische Zeitschrift 39, 255-262. years. Geological Society of America Memoir 145, Herterich, K., and Samthein, M. (1984). Brunhes time 111-146. scale: Tuning by rates of calcium-carbonate disso- Ruddiman, W. F., and McIntyre, A. (198la). The lution and cross spectral analyses with solar insola- North Atlantic during the last deglaciation. Pulueo- tion. In “Milankovitch and Climate” (A. Berger, et geography, Paleoclimatology, Palueoecology 35, al., Eds.), Part 1. Vol. 126, pp. 447466. NATO AS1 145-214. Series C, Reidel, Dordrecht. Ruddiman, W. F., and McIntyre, A. (1981b). Oceanic Hyde, W. T., and Peltier, W. R. (1987). Sensitivity mechanisms for ampliiication of the 23,000-year ice- experiments with a model of the Ice Age cycle: The volume cycle. Science 212, 617-627. response to Milankovitch Forcing. Journal of the Ruddiman, W. F., and McIntyre, A. (1984). Ice-age Atmospheric Sciences 44(10), 1351-1374. thermal response and climatic role of the surface 152 HARTMUT HEINRICH

Atlantic Ocean, 4ON” to 63”N. Geological Society of Earth-orbital forcing. Journal of the Atmospheric America Bulletin 95, 381-3%. Sciences 41, 338&3389. Ruddiman, W. F., McIntyre, A., Niebler-Hunt, V., Schiffelbein, P. (1984). Effect of benthic mixing on the and Durazzi, J. T. (1980). Oceanic evidence for the information content of deep-sea stratigraphical sig- mechanism of rapid northern hemisphere glaciation. nals. Nature (London) 311, 651-653. Ouaternarv Research 13, 33-64. Szzman, R., Hansen, A. R., and Maasch, K. A. Vemekar, A. D. (1972). Long-period global variations (1984). The late Quatemary glaciations as the re- of incoming solar radiation. Meteorological Mono- sponse of a three-component feedback system to graphs 12, 1-21.