An event stratigraphy for the Last Termination in the North Atlantic region based on the Greenland ice-core record: a proposal by the INTIMATE group

Transcription

1 JOURNAL OF QUATERNARY SCIENCE (1998) 13 (4) CCC /98/ $ John Wiley & Sons, Ltd. An event stratigraphy for the Last Termination in the North Atlantic region based on the Greenland ice-core record: a proposal by the INTIMATE group SVANTE BJöRCK 1 *, MICHAEL J. C. WALKER 2, LES C. CWYNAR 3, SIGFUS JOHNSEN 4, KAREN-LUISE KNUDSEN 5, J. JOHN LOWE 6, BARBARA WOHLFARTH 7 and INTIMATE Members 8 1 Geological Institute, University of Copenhagen, Øster Voldgade 10, DK-1350 Copenhagen, Denmark 2 Department of Geography, University of Wales, Lampeter, Ceredigion SA48 7ED, Wales, UK. 3 Department of Biology, University of New Brunswick, Fredericton, New Brunswick, Canada E3B 6E1 4 Neils Bohr Institute of Astronomy, Department of Geophysics, University of Copenhagen, Haraldsgade 6, Copenhagen, Denmark 5 Department of Earth Sciences, University of Aarhus, DK-8000 Aarhus C. Denmark 6 Department of Geography, Royal Holloway, University of London, Egham, Surrey TW20 0EX, UK. 7 Department of Quaternary Geology, Lund University, Tornavägen 13, S , Lund, Sweden 8 See Table 1 Björck, S., Walker M. J. C., Cwynar, L. C., Johnsen, S., Knudsen, K-L., Lowe, J. J., Wohlfarth, B. and INTIMATE Members An event stratigraphy for the Last Termination in the North Atlantic region based on the Greenland Ice-core record: a proposal by the INTIMATE group J. Quarternary Sci., Vol. 13, pp ISSN Received 21 January 1998; revised 31 March 1998; accepted 8 April 1998 ABSTRACT: It is suggested that the GRIP Greenland ice-core should constitute the stratotype for the Last Termination. Based on the oxygen isotope signal in that core, a new event stratigraphy spanning the time interval from ca to 11.5 k GRIP yr BP (ca k 14 C yr BP) is proposed for the North Atlantic region. This covers the period from the Last Glacial Maximum, through Termination 1 of the deep-ocean record, to the Pleistocene Holocene boundary, and encompasses the Last Glacial Late-glacial of the traditional northwest European stratigraphy. The isotopic record for this period is divided into two stadial episodes, Greenland Stadials 1 (GS-1) and 2 (GS-2), and two interstadial events, Greenland Interstadials 1 (GI-1) and 2 (GI-2). In addition, GI-1 and GS-2 are further subdivided into shorter episodes. The event stratigraphy is equally applicable to ice-core, marine and terrestrial records and is considered to be a more appropriate classificatory scheme than the terrestrially based radiocarbon-dated chronostratigraphy that has been used hitherto John Wiley & Sons, Ltd. KEYWORDS: region. GRIP ice-core; stratotype; Last Termination; event stratigraphy; Greenland; North Atlantic Introduction The principles underlying the classification of the Quaternary stratigraphic record of Norden and adjacent areas of north- INTIMATE: INTegration of Ice-core, MArine and TErrestrial records is a core programme of the INQUA (International Quaternary Union) Palaeoclimate Commission. The aim is to synthesise data from the marine, terrestrial and ice-core realms for the North Atlantic region during the course of the Last Termination. * Correspondence to: Svante Björck, Geological Institute, University of Copenhagen, Øster Voldgade 10, DK-1350 Copenhagen, Denmark Contract grant sponsors: Swedish Research Council Contract grant sponsors: Royal Society, London Contract grant sponsors: INQUA Palaeoclimate Commission west Europe were first formally set out in a seminal paper by Mangerud et al. (1974). A formal international terminology was also proposed. The Quaternary was subdivided into the Pleistocene and the Holocene series, with the Pleistocene being further subdivided into provisional stages based on the sequence of glacials and interglacials, the boundaries of which were formally defined by reference to stratotypes. The Late Weichselian and the Holocene were, in turn, subdivided into chronozones with the boundaries defined in radiocarbon years before present (ad 1950). This chronostratigraphic framework has been widely adopted and the terminology, especially that relating to the Late Weichselian, has been applied not only to stratigraphical records from sites in northwest Europe, for which the classificatory scheme was originally designed, but also to records from other parts of Europe and from other areas of the world for which the classificatory scheme was never intended.

2 284 JOURNAL OF QUATERNARY SCIENCE Since the publication of the Mangerud et al. classification in 1974, a large number of terrestrial and lacustrine sites containing records of the Last Termination have been described from Europe, with those spanning the Last Glacial Interglacial Transition (informally referred to as the Lateglacial, and spanning the period k 14 C yr BP) in particular forming the basis for palaeoenvironmental syntheses at the regional and national scales (e.g. Lowe and NASP Members, 1995; Berglund et al., 1996). In addition, many sites have been discovered in Atlantic Canada and along the eastern seaboard of North America (e.g. Mott et al., 1986; Cwynar et al., 1994; Peteet et al., 1994) that contain stratigraphic records showing a striking resemblance to at least parts of the northwest European bio- and climatostratigraphy upon which Mangerud et al. (1974) founded their chronostratigraphic scheme. These sites, from both sides of the North Atlantic, have yielded a wide range of proxy climatic data and many of the records are underpinned not only by radiocarbon dates but also by independent ageestimates, as in Sweden, for example, where varve chronology is widely used. Moreover, these terrestrial data, which document significant climatic shifts during the Last Termination, have been augmented by isotopic and biostratigraphic evidence from high-resolution marine cores and by the detailed isotopic and related atmospheric records obtained from the Greenland ice-cores. The latter, in particular, provide a history of climate change at a level of resolution that was unimaginable only 20 yr ago. As a consequence of all of these new data sources, we now have a much more detailed understanding of the spatial and temporal variations in environmental and climatic change around the Atlantic margins during the transition from the last glacial and into the Holocene than was the case when Mangerud et al. published their proposals in One major consequence of the more recent discoveries is that questions have been raised about the applicability of the Mangerud et al. chronostratigraphy, particularly that part of the classificatory scheme that relates to the Last Glacial Interglacial Transition (see e.g. Lowe and Gray, 1980; Björck, 1984; Lowe and NASP Members, 1995; Walker, 1995; Wohlfarth, 1996). At a meeting of the INTIMATE programme held in Sweden in September 1997 (Table 1), problems relating to the use of this chronostratigraphy were discussed in the context of the rapid proliferation of palaeoenvironmental data for the Last Glacial Interglacial Transition described above, and the greater temporal resolution of these records that can now be achieved. Arising from these discussions, and based on the collective experiences of the workshop participants, an alternative classificatory framework for the entire Last Termination was proposed. This is essentially an event stratigraphy based on the isotopic record in the GRIP Greenland ice-core. The details of this new stratigraphic scheme, and the reasons why it is preferred to the existing chronostratigraphic classification, are set out in this paper. The chronostratigraphic classification The chronostratigraphic scheme of Mangerud et al. (1974) was based largely on conventional stratigraphic procedures (i.e. Hedberg, 1970), whereby units of stage rank or higher and the boundaries of chronostratigraphic units are defined by reference to stratotypes. For the Late Weichselian and Holocene, however, the chronozone boundaries were defined directly in radiocarbon years, but without reference to type-sequences or to type-sections. Although not entirely in accord with formal stratigraphic practice, this was considered to be acceptable as it was noted that, with reference to the Late Weichselian and Holocene,...correlations within this time-span are so dependent on radiocarbon datings that if a sediment cannot be correlated to a classification defined in radiocarbon years, then it cannot be correlated to any chronostratigraphic classification defined by stratotypes (Mangerud et al., 1974, p. 114). It was also noted that in northwest Europe...no chronstratigraphic classification of the Late Weichselian and Holocene exists showing satisfactory definitions of the respective boundaries. New definitions are therefore necessary... and instead of proposing a new classification, we propose only precise definitions of already well-established terms, leaving the age of the boundaries as close as possible to the tradition of the last few decades (Mangerud et al., 1974, p. 114). In the Mangerud et al. classification, the Middle Weichselian Late Weichselian Substage boundary was defined and dated according to van der Hammen et al. s (1971) definition of the Pleniglacial Late-glacial boundary. The age of this boundary was set at 13.0 k 14 C yr BP, although Mangerud and Berglund (1978) subsequently revised the age to 25.0 k 14 C yr BP to include the last glacial maximum within the Late Weichselian, and to make it more compatible with the chronostratigraphic definitions of the Late Wisconsinan and Late Devensian. This resulted in a chronostratigraphic gap...which...would soon be filled with named units, namely the strata between 25,000 BP and the Bølling Chronozone (Mangerud and Berglund, 1978, p. 180). The latter, however, was never accomplished, and the defined chronozones of the Late Weichselian Substage remained as the Bølling Chronozone (including the Oldest Dryas interval) from 13.0 to 12.0 k 14 C yr BP, the Older Dryas Chronozone from 12.0 to 11.8 k 14 C yr BP, the Allerød Chronozone from 11.8 to 11.0 k 14 C yr BP and the Younger Dryas Chronozone from 11.0 to 10.0 k 14 C yr BP. This terminology, which had originally been developed to reflect important changes in lithostratigraphic and biostratigraphic records (principally pollen zones and macrofossil assemblages) by a number of workers (e.g. Hartz and Milthers, 1901; Jessen, 1935; Iversen, 1942, Nilsson, 1961) was therefore subsumed into the new chronostratigraphic definitions. Accordingly, the Late Weichselian Flandrian/Pleistocene Holocene boundary was fixed at 10.0 k 14 C yr BP. Problems with the chronostratigraphic classification The Mangerud et al. classification undoubtedly represented an important milestone in the history of investigation of climatic and environmental change at the end of the Last Glacial, and it was enthusiastically adopted by the majority of European Quaternary scientists. The terminology (Bølling, Allerød, Younger Dryas) has been widely used and appears not only in papers dealing with terrestrial data, but also in publications on marine and ice-core evidence. Recently, however, questions have been raised about the applicability and, more importantly, about the utility of the scheme. In this section we examine some of the problems that have arisen.

3 EVENT STRATIGRAPHY FOR THE NORTH ATLANTIC REGION 285 Table 1 Participants at the INTIMATE Meeting, Höör, Sweden, September 1997 T. Andrén Quaternary Research, University of Stockholm, Sweden W. E. J. Austin Geography, University of Durham, England K. Banks Geology & Geophysics, University of Minnesota, Minneapolis, USA O. Bennike Geological Survey of Denmark, Copenhagen, Denmark H. Bergsten Oceanography, University of Gothenburg, Sweden H. H. Birks Botanical Institute, University of Bergen, Norway J. Björck Quaternary Research, University of Stockholm, Sweden S. Björck Geological Institute, University of Copenhagen, Denmark S. Bohncke Earth Sciences, Free University, Amsterdam, The Netherlands S. Bondevik Geological Institute, University of Bergen, Norway L. Brunnberg Quaternary Research, University of Stockholm, Sweden G. S. Burr NSF-Arizona AMS Laboratory, University of Arizona, Tucson, USA G. R. Coope Geography, Royal Holloway, University of London, Egham, England L. Cwynar Biology, University of New Brunswick, Fredericton, Canada F. Dobos Quaternary Geology, University of Lund, Sweden J. Eiriksson Science Institute, University of Iceland, Reykjavik, Iceland H. Haflidason Geological Institute, University of Bergen, Norway S. Hampton Quaternary Science, University of Coventry, England T. Hang Quaternary Research, University of Stockholm, Sweden C. Hawkesworth Earth Sciences, Open University, Milton Keynes, England J. Heinemeier AMS 14 C Dating Laboratory, University of Aarhus, Denmark W. Hoek Earth Sciences, Free University, Amsterdam, The Netherlands K. Hughen Arctic & Alpine Research, University of Colorado, Boulder, USA O. Ingolfsson Earth Sciences, University of Gothenburg, Sweden C. Israelson Geological Institute, Unversity of Copenhagen, Denmark S. Johnsen Physics & Geophysics, University of Copenhagen, Denmark K.-L. Knudsen Earth Sciences, Univesrity of Aarhus, Denmark N. Koç Geological Institute, University of Bergen, Norway B. Kromer Heidelberg Academy of Sciences, Heidelberg, Germany K. Lambeck Earth Sciences, Australian National University, Canberra, Australia G. Lemdahl Quaternary Geology, University of Lund, Sweden J. J. Lowe Geography, Royal Holloway, University of London, Egham, England K. Nordberg Oceanography, University of Gothenburg, Sweden S. Olsson Quaternary Geology, University of Lund, Sweden B. P. Onac Quaternary Geology & Mineralogy, University of Cluj, Romania G. Possnert Tandem Laboratory, University of Uppsala, Sweden D. Richards Geology & Geophysics, University of Minnesota, Minneapolis, USA B. Ringberg Quaternary Research, University of Stockholm, Sweden M. Rundgren Quaternary Geology, University of Lund, Sweden J. Schwander Physics, University of Bern, Switzerland N. J. Shackleton Godwin Laboratory, University of Cambridge, England G. Skog Quaternary Geology, University of Lund, Sweden I. Snowball Quaternary Geology, University of Lund, Sweden D. Subetto Quaternary Geology, University of Lund, Sweden R. Vaikmae Geology, Estonia Academy of Sciences, Tallinn, Estonia M. J. C. Walker Geography, University of Wales, Lampeter, UK S. Wastegård Geography, Royal Holloway, University of London, Egham, England M. Wastl Geography, University of Munich, Germany B. Wohlfarth Quaternary Geology, University of Lund, Sweden Terminology As noted above, the terminology used in the Mangerud et al. scheme has a long history of usage, and was derived largely from early palaeobotanical studies. The term Dryas, for example, was introduced in the last century to refer to the earliest vegetational stage in the post-glacial history of Sweden (Andersson, 1896). The Allerød warm interval was first defined on the basis of plant macrofossil evidence at the Allerød site in Denmark (Hartz and Milthers, 1901). Subsequently, following the recognition of comparable pollen zones throughout northwest Europe (Jessen, 1935, 1938), the terminology was widely applied to what were considered to be time-equivalent pollen zones as well as to geological and/or archaeological episodes. Hence, the Younger Dryas was equated with zone III and the Allerød Interstadial with zone II of the Jessen pollen zonation scheme, whereas zone I was eventually subdivided into subzones Ia, Ib and Ic, which were considered to be the equivalents of the Oldest Dryas cold episode, the Bølling Interstadial and the Older Dryas cold phase respectively (e.g. Iversen, 1954). Further, as the pollen zones were considered to be broadly timesynchronous and to reflect particular climatic events, what had originated as a biostratigraphical classification was extended to include elements of both climatostratigraphy

4 286 JOURNAL OF QUATERNARY SCIENCE and chronostratigraphy. Inevitably this led to a lack of clarity and to ambiguity in stratigraphic usage. This problem was recognised by Mangerud et al. in their 1974 paper. Their solution was to develop a stratigraphic scheme that used the traditional northwest European terminology, but in a chronostratigraphic classification in which the boundaries of units were defined using radiocarbon dates. Hence, it was anticipated that terms such as Bølling, Allerød, etc., would now be applied only in a chronostratigraphic sense, i.e. the Bølling or Allerød Chronozone. In reality, however, this procedure has proved difficult to follow. The very fact that the Mangerud et al. scheme adopted an existing terminology that had been used in different ways, created as many problems as it solved. Terms that had been introduced to designate geological climate units (e.g. Allerød Interstadial) and which, in turn, became equated with palynologically defined biozones (e.g. pollen zone II), were being applied to radiometrically determined chronozones. The difficulty, of course, is that because of the long record of usage, each of the terms Bølling, Older Dryas, etc., has its own historical associations and connotations. Hence, despite the recommendations of Mangerud et al., the terminology has not always been used in a strict chronostratigraphic sense, for terms such as Allerød Interstadial (a climatostratigraphic definition), Allerød Zone or Biozone (normally based on pollen evidence), and Allerød Chronozone still abound in the scientific literature. Indeed, because of the difficulty of applying some of these terms in a chronostratigraphic context, some researchers have advocated modifications to the present terminology. For example, as Younger Dryas is now widely used to denote a chronozone, a number of authors have referred, in a climatostratigraphic sense, to the cold episode between ca 11.0 and 10.0 k 14 C yr BP as the Late Dryas, and the term Earlier Dryas has been substituted for Older Dryas (e.g. Bohncke, 1993). Such proliferation of terminology, although understandable, serves to confuse rather than to clarify, but it is symptomatic of some of the difficulties that have arisen because of the historical legacy of the terminology that has been used. Time transgression In the stratigraphic record for northwest Europe, the chronozones, as defined in the Mangerud et al., scheme are all effectively radiocarbon-dated biozones. Biozone boundaries, however, are by their very nature time-transgressive on a fine timescale, because they reflect biological response to climate/environmental thresholds that are spatially and temporally diachronous. Boundaries of chronozones, on the other hand, are time-parallel. Yet, in discussions of the European Late-glacial, biozones and chronozones continue to be used interchangeably, and this has inevitably led to ambiguity and sometimes to confusion. For example, the Younger Dryas chronozone as formally defined is that part of the stratigraphic record spanning the time interval between 11.0 and 10.0 k 14 C yr BP, whereas the Younger Dryas biozone is that section of the stratigraphic record that is characterised by a fossil assemblage considered to be indicative of colder climatic conditions. At some sites, the onset of the Younger Dryas biozone may be dated to around, or before, 11.0 k 14 C yr BP, whereas at other sites the onset may be considerably later. Chronostratigraphically, therefore, the boundaries of the two types of zone will often not coincide (Walker, 1995). The problem, of course, is that formal stratigraphical procedures are being applied to a part of the geological record for which they were never originally intended. In older strata, lithostratigraphic boundaries, although usually time-transgressive, appear to be synchronous when set against the vast span of geological time. By contrast, ultra-rapid climatic shifts over time-scales of half a century or less can now be detected in many palaeoenvironmental records from the Last Termination. Furthermore, detailed analyses of the GISP2 core (Taylor et al., 1997) show that the time-lag, between low and high latitudes, for major climate events is less than a few decades. On the other hand, the spatial evolution of climate is complex. Many proxy records may not register a hemispherically strong signal because, at that locality, a threshold that causes change is not crossed. The response to large-scale climate shifts may therefore often seem diachronous (e.g. Koç Karpuz and Jansen, 1992; Walker et al., 1993; Coope and Lemdahl, 1995). Indeed, in view of what is now known about the spatial and also the temporal variations in climate change in northwest Europe during the Last Glacial Interglacial Transition (Lowe et al., 1994; Walker, 1995), it seems questionable whether the 1974 chronostratigraphic scheme is even appropriate within the Scandinavian region for which it was originally intended (Wohlfarth, 1996). It follows, therefore, that the application in a chronostratigraphic sense, of such terms as Bølling or Allerød, to regions of Europe outside Scandinavia, or to North America for example, has little (if any) validity. Radiocarbon dating Over the last 20 yr or so, it has become increasingly apparent that there are problems with the radiocarbon time-scale, the nature and extent of which were not fully appreciated at the time of the Mangerud et al. publication in These include both long- and short-term variations in atmospheric 14 C, as well as problems associated with radiocarbon measurement. Comparisons between radiocarbon-dated wood samples and dendrochronological records show an increasing divergence between radiocarbon and calendar years prior to 10.0 k 14 C yr BP (e.g. Kromer and Becker, 1993), and paired measurements of U-series and radiocarbon dates from fossil corals (Bard et al., 1990) and from radiocarbon-dated laminated marine sediments (Hughen et al., 1998) suggest age differences of up to 2.5 kyr by 16.0 k 14 C yr BP. Superimposed on this long-term trend are a number of radiocarbon plateaux of constant 14 C age, which appear to have durations of 100 to several 100 yr. So far five such events have been identified within the period of the Last Glacial Interglacial Transition, at ca , , and and k 14 C yr BP (Amman and Lotter, 1989; Zbinden et al., 1989; Goslar et al., 1995; Björck et al., 1996; Hughen et al., 1998; Gulliksen et al., 1998; Kitagawa and van der Plicht, 1998). These are often followed by a rapid decline in 14 C ages, reflecting abrupt changes in atmospheric 14 CO 2 content. As a result, certain time intervals during the Last Termination cannot be dated precisely using the radiocarbon method alone. This has major implications for the chronostratigraphic definition of, for example, the Younger Dryas Chronozone, because the upper boundary of this chronozone coincides with a 14 C plateaux, and the lower boundary is characterised by an exceptional rise in 14 C (Björck et al., 1996), which is reflected in a rapid decrease in radiocarbon age (from ca to ca k 14 CyrBPin

5 EVENT STRATIGRAPHY FOR THE NORTH ATLANTIC REGION 287 less than 100 calendar yr). If the boundaries of the chronozones cannot be dated precisely by radiocarbon, this raises fundamental questions about the validity and applicability of the entire chronostratigraphic classificatory scheme. Temporal distortions in the radiocarbon time-scale resulting from both short- and long-term variations in the atmospheric 14 C/ 12 C ratio, mean that it is necessary to calibrate radiocarbon dates in order to establish their calendar-age equivalence. This has become increasingly important as terrestrial records from the Last Termination are now available at much higher levels of resolution, and are being dated by methods other than radiocarbon (dendrochronology, varve chronology, U-series dating, etc). However, it has not always proved easy to reconcile the time-scales based on these different techniques (e.g. Wohlfarth et al., 1993). Moreover, there is an increasing imperative to establish a temporal link between radiocarbondated terrestrial and marine sequences, and the independently dated, high-resolution ice-core records from Greenland (e.g. Haflidason et al., 1995; Björck et al., 1996). Calibration programs are available to convert radiocarbon years into sidereal years (e.g. Stuiver and Reimer, 1993), but both the early Holocene calibrations, which use tree-ring data (Spurk et al., 1998), and the Late-glacial calibrations, which are based on direct comparisons of U Th and 14 C dates, are still being revised. An additional concern is that the radiocarbon dates themselves may be in error because of a range of site-, sampleor laboratory-specific factors (Mangerud, 1970). In the dating of limnic sediments, for example, there is the recurrent problem of isotopic fractionation, and of contamination by older or younger materials (e.g. Lowe and Walker, 1980; Björck and Håkansson, 1982; Lowe, 1991; Björck et al., 1998), and although AMS radiocarbon dating may offer a solution to some of these problems, difficulties are now beginning to emerge with this approach also (Colman et al., 1996; Wohlfarth et al., 1998). Marked spatial and temporal variations of the marine reservoir age (e.g. Mangerud and Gulliksen, 1975; Bard et al., 1994; Haflidason et al., 1995; Björck et al., 1998) are a further complication. These sampleor method-specific problems impose additional limitations on a chronostratigraphic scheme that relies entirely on radiocarbon dating for the provision of a time-scale. Overall, therefore, it is clear that with alternative chronologies now available, a chronostratigraphy for the Last Termination that is based exclusively on radiocarbon no longer has the validity that it did some 25 yr ago, and an alternative classificatory scheme is therefore required. The Greenland ice-core record The recovery of continuous cores from the polar ice sheets and the extraction from these cores of a range of atmospheric proxy climate records has been one of the major scientific achievements of the last decade. In the Northern Hemisphere, five deep cores have now been obtained from the Greenland ice sheet, and these provide a record of climate change extending back to the last interglacial. Of particular significance have been the two most recent cores, GRIP and GISP2, which were drilled within 28 km of each other at the Summit of the ice sheet. Proxy climate records from the cores include snow accumulation rates (Alley et al., 1993; Kapsner et al., 1995), dust content (Taylor et al., 1993a,b), and oxygen isotope ( 18 O) profiles (Dansgaard et al., 1993; Grootes et al., 1993). The last-named have proved to be especially valuable as indicators of abrupt climatic fluctuations, as a basis for correlating between individual ice-core records (Johnsen et al., 1992a), and also as a means of correlating the Greenland climatic record with that in Atlantic marine (e.g. Bond et al., 1993; Hughen et al., 1996; Rasmussen et al., 1996) and European terrestrial sequences (Lotter et al., 1992; Goslar et al., 1995; Björck et al., 1996, 1997). The similarity between the isotopic profiles from different parts of the ice sheet is particularly striking, and underlines the value of the 18 O record as a climatic proxy for the Greenland area. What is equally apparent, however, is its potential as a basis for correlation within the wider realm of the North Atlantic region. The ice-core data provide a continuous, sensitive and high-resolution record of atmospheric changes, not only in the vicinity of the Greenland ice sheet but in other areas of the Northern Hemisphere as well (Hughen et al., 1996; Taylor et al., 1997), and the sequence of climatic fluctuations during, and at the end of, the last glaciation are therefore reflective of climatic changes throughout the North Atlantic province. Consequently, the Greenland 18 O profile (Fig. 1) can be used as a stratigraphic template for the Last Termination in the North Atlantic region. The most detailed isotopic profile so far available is that from the GRIP ice-core (Johnsen et al., 1992b; Dansgaard et al., 1993), and it is suggested that this should constitute the type profile for this time period. No terrestrial stratotype has ever been established for the Last Termination. Moreover, the ice-core record can be dated independently. The preliminary time-scale we use here for the GRIP core has been developed by counting annual ice layers down from the surface back to 14.5 k GRIP yr BP, and below that by the use of a steadystate ice-flow model (Dansgaard et al., 1993). There is, however, an extended alternative stratigraphic chronology (Hammer et al., 1997), as well as the published GISP2 chronology (Bender et al., 1994; Meese et al., 1997), both of which give older ages prior to 14.5 kyr BP. The chronology used here for the older part of the Last Termination (ca k GRIP yr BP) therefore must be considered as tentative. An event stratigraphy based on the GRIP Greenland ice-core record We propose that the high-resolution oxygen isotope record from the GRIP ice-core be used as a basis for an event stratigraphy for the Last Termination that is applicable to the entire North Atlantic region. Events are short-lived occurrences that have left some trace in the geological record, and which therefore may be used as a means of correlation (Whittaker et al., 1991). The oxygen isotope profile from the Greenland ice-cores therefore can be divided into a series of isotopic events, following the count-fromtop procedure that has been used successfully in the development of the marine oxygen isotope record. These events can be dated in GRIP ice-core years (Table 2). A preliminary Greenland isotopic climatostratigraphy has already been outlined, based on the recognition of interstadial events (Johnsen et al., 1992b). Here we suggest extending and formalising this sequence to include not only interstadial, but also stadial episodes, back to Greenland Isotope Interstadial 2 (GI-2, dated to k GRIP yr BP (Fig. 1). However, we also recommend the use of Johnsen et al. s (1992b) interstadial

6 288 JOURNAL OF QUATERNARY SCIENCE Figure 1 The 18 O record ( SMOW) from the GRIP deep ice-core (Johnsen et al., 1992a; Dansgaard et al., 1993) between 11.0 and 23.0 k GRIP yr BP. Johnsen et al. s (1992b) division of the isotope stratigraphy into interstadials (and subinterstadials) has been extended by defining intervening stadials (and substadials). Table 2 Depths and preliminary ages (ad 1950) of the onset of events and episodes in the GRIP ice-core, including the Holocene epoch. It should be noted that the older part of the chronology is still regarded as uncertain Events Episodes Ice-core depth Ice-core age (m) for the (yrs) for the onset onset Holocene epoch ,500 GS ,650 GI-1a ,900 GI-1b ,150 GI-1c ,900 GI-1d ,050 GI-1 GI-1e ,700 GS-2a ,900 GS-2b ,500 GS-2 GS-2c ,200 GI ,800 event stratigraphy, including the addition of intervening stadials, for the main part of the last glacial period. We have not attempted a stratigraphic subdivision of the Holocene, because our concern is solely with the Last Termination. Hence, working down the GRIP isotopic trace (Fig. 1), we begin at the first marked cold episode between ca and k GRIP yr BP, hitherto widely referred to as the Younger Dryas. We propose that this interval henceforth be termed Greenland (Isotope) Stadial 1 (GS-1). The preceding interstadial (which equates with the Bølling- Allerød interval of the Mangerud et al. scheme) can then be designated Greenland (Isotope) Interstadial 1 (GI-1), dated to k GRIP yr BP, and is readily subdivisible into three warmer episodes GI-1a, 1c and 1e with the intervening colder periods GI-1b and 1d. This follows the convention and practice of numbering warm substages in the deep ocean isotope record (e.g. marine oxygen isotope stages 5a to 5e). The preceding long cold period is designated Greenland (Isotope) Stadial 2 (GS-2) and is subdivided into two distinctly colder episodes, GS-2a and 2c, dated to k GRIP yr BP and k GRIP yr BP respectively, with an intervening less cold period, GS-2b, dated to k GRIP yr BP. Finally, the short interstadial before that is named Greenland (Isotope) Interstadial 2 (GI-2) and is dated to k GRIP yr BP. It should be noted that the corresponding event in the GISP2 core is dated to k GISP yr BP (Bender et al., 1994) The advantages of this scheme over the existing terrestrially based chronostratigraphy are as follows. 1. The scheme is based on a continuous, high-resolution, proxy climatic record that spans the entire period from the Last Glacial Maximum through Termination 1 of the marine isotope sequence to the Pleistocene Holocene boundary (Fig. 1). 2. It is based on the record from a type sequence (the GRIP Summit Ice-core) and therefore conforms to formal stratigraphic procedures. 3. The scheme is underpinned by an independent chronology (GRIP ice-core years: Table 2) and therefore it does not have the problems associated with radiocarbon dating (plateaux, calibration). Furthermore, comparison between the GRIP record, northwest European lake sediment records and the recently revised German oak-pine dendrochronology (Björck et al., 1996; Spurk et al., 1998) suggests that synchronisation between GRIP ice years and treering years is within the stated uncertainty of 30 yr for the GRIP record. Our recommendation, therefore, is that the chronology for the Last Termination should be expressed in GRIP Ice-core Years Before Present (AD 1950), preferably shortened to GRIP yr BP, in order to be compatible with the tree-ring-based calibration timescale (cal. yr BP, ad 1950).

7 EVENT STRATIGRAPHY FOR THE NORTH ATLANTIC REGION The scheme follows the established stratigraphic practice of the marine sequence in that alternate warm and cold episodes are numbered downcore on a count from the top to the basis. As in the marine oxygen isotope sequence, it embodies sufficient flexibility to denote stages and substages. As a consequence, there will now be an ice-core oxygen isotope sequence to parallel the marine oxygen isotope record. 5. The scheme is an event stratigraphy. As such, boundaries between the events are not specifically designated, and problems of time-transgression that have arisen in applications of the terrestrial chronostratigraphy are no longer encountered. 6. The scheme uses a totally new terminology, and therefore the historical baggage associated with the use of such terms as Bølling, Allerød and Older Dryas no longer exists. 7. As multidisciplinary records in terrestrial/limnic sequences from both the European and North American seaboards show remarkably similar events to those in the Greenland ice-cores, the scheme proposed here should be applicable throughout the North Atlantic region, and possibly in more continental areas also. 8. Moreover, as the climatic signal in the Greenland (GRIP) oxygen isotope profile finds close parallels in the palaeoceanographic record from deep-ocean sediments, the scheme can be applied readily to North Atlantic marine sequences from the Last Termination. 9. The scheme proposed allows for local climate variations to change the key events, both in terms of their timing and intensity, without disruption to the existing stratigraphical nomenclature. Application of the event stratigraphy to the North Atlantic region Figure 2 shows five palaeoenvironmental records spanning the Last Termination from sites around the North Atlantic region. These have been obtained from different geological contexts (Greenland ice sheet, deep ocean floor, lake basins, glacial sediments) and are based on different environmental proxies (fossil insects; stable isotope profiles; laminated sediments; marine molluscs). The vertical axes vary, with two plotted in radiocarbon years, two in sediment thickness, and one in ice-core years. Neverthless, the curves can be compared directly using the event stratigraphy described above. The stadial GS-2 is recorded at the base of each of the sequences, followed by the abrupt warming event at the beginning of GI-1. The gradual deterioration throughout GI- 1 is also evident, leading into the cold stage of GS-1. Whether the boundaries between GS-2, GI-1 and GS-1 are precisely synchronous remains to be established, but the signature of the major climatic events registers clearly in each of the records, and it is these and not the boundaries (or horizons) that form the basis for classification and correlation. Some or all of the substages within GI-1 also can be Figure 2 Four sediment-based records from the North Atlantic region related to the GRIP 18 O record (extreme right) between 10.5 and 16.0 k GRIP yr BP using the proposed event stratigraphy. From the left these are: reconstructed mean temperatures for the warmest month in the British Isles during the period k 14 C yr BP, based on coleopteran data (after Atkinson et al., 1987); ice-front oscillations in southwest Iceland relative to the inner part of the Borgarfjördur (0 km point is located arbitrarily) during the period k 14 C yr BP, based on till stratigraphy and radiocarbon-dated glaciomarine sediments (Ingólfsson et al., 1997); the 18 O record ( PDB) in lacustrine carbonates in Lake Gerzensee, Switzerland (Eicher and Siegenthaler, 1976); and greyscale variations (reflecting biological productivity) in laminated marine sediments from the Cariaco Basin, Venezuela (Hughen et al., 1996, 1998). Note that the last-named records are related to sediment depths, but that an independent annual chronology, correlated closely with the ice-core and tree-ring records, exists for the Cariaco sediments (Hughen et al., 1998). The dashed lines indicate tentative correlations between the events discussed in the text and defined on the basis of the GRIP record. Hypothetical correlations are shown by a question mark.

8 290 JOURNAL OF QUATERNARY SCIENCE identified. For example, each of the records shows the clearly defined initial warm event (GI-1e), followed by the cooling into GI-1d. The last warm event of the interstadial GI-1a is also clearly defined, and the preceding cold substage GS-1b may be detected as a discrete event in most of the profiles. As stratigraphical investigations become increasingly more sophisticated, this will inevitably lead to even finer subdivision of stratigraphical records. One of the main advantages of the event stratigraphy as set out here is that the alphanumeric system that has been devised can readily incorporate such developments, without the need for proliferation of new terms. For example, a two-fold division of terrestrial records assigned to event GS-1 (currently termed the Younger Dryas ) has been recognised in a number of sites in northwest Europe (Lowe et al., 1994; Walker, 1995); significantly, a two-fold subdivision is evident in the GRIP icecore record. A nomenclature could easily be devised for these or for any new substages within the existing alphanumeric framework. Local stratigraphic terms may continue to be used in the construction of local stratigraphical schemes, especially where the records are perhaps not directly comparable with the GRIP ice-core stratigraphy. In such cases, however, it should be made clear that these are local schemes only, and the stratigraphic basis of the terminology used, as well as the geographical extent of the region to which they are meant to be applied, should be clearly explained. For broader, interregional comparisons within the North Atlantic province, however, the event stratigraphy set out above offers both the simplicity and stratigraphic rigour that are inherent in the now universally accepted marine oxygen isotope stratigraphy. Conclusion It is the proposal of the INTIMATE group that the Greenland GRIP Ice-core is designated as the stratotype for the time period of the Last Termination (ca k GRIP yr BP), and that the oxygen isotope profile can be used as the basis for an event stratigraphy that can be divided into isotopically defined stadials (GS-1 and GS-2) and interstadials (GI-1 and GI-2). Greenland Interstadial 1 (GI-1) is further subdivided into alternate warm and cold episodes, 1a to 1e, following the practice adopted for the marine oxygen isotope sequence (e.g. MIS 5 and 7), and Greenland Stadial 2 (GS-2) is subdivided into three episodes, 2a to 2c, of which 2b is the distinctly warmer one. This event stratigraphy is underpinned by an independent chronology based on GRIP ice-core years. The classificatory scheme is equally applicable to marine and terrestrial records. It is the proposal of the INTIMATE group that it should replace the former terrestrially based chronostratigraphy, which has been used hitherto not only to subdivide records from the North Atlantic region, but has been extended to other parts of the world. We recommend that the terms Bølling, Older Dryas, Allerød, and Younger Dryas be discontinued in favour of the new terminology described above. Acknowledgements BW acknowledges a grant from the Swedish Research Council, which supported the INTIMATE Workshop at Höör. JJL and MJCW are grateful to the Royal Society, London, for travel grants, which enabled them to attend the Workshop, and to the INQUA Palaeoclimate Commission for their financial support of the INTIMATE programme. We should like to thank Hilary Birks, Russell Coope, Richard Preece and James Scourse for their helpful comments on an earlier draft of this paper. References ALLEY, R. B., MEESE, D. A., SHUMAN, C. A., GOW, A. J., TAYLOR, K. C., GROOTES, P. M., WHITE, J. W. C., RAM, M., WADDING- TON, E. D., MAYEWSKI, P. A. and ZIELINSKI, G. A Abrupt increase in Greenland snow accumulation at the end of the Younger Dryas event. Nature, 362, AMMANN, B. and LOTTER, A. F Late-glacial radiocarbon and palynostratigraphy on the Swiss Plateau. Boreas, 18, ANDERSSON, G Svenska växtvärldens historia. P. A. Norsted & B. Söner, Stockholm, 136 pp. ATKINSON, T. C., BRIFFA, K. R. and COOPE, G. R Seasonal temperatures in Britain during the past 22,000 years reconstructed using beetle remains. Nature, 352, BARD, E., HAMELIN, B. and FAIRBANKS, R. G Calibration of the 14 C timescale over the past 30,000 years using mass spectrometric U Th ages from Barbados. Nature, 345, BARD E., ARNOLD, M., MANGERUD, J., PATERNE, M., LABEYRIE, L., DUPRAT, J., MÉLIÈRES. M-A., SøNSTEGAARD, E. and DUPLESSY, J-C The North Atlantic atmosphere sea surface 14 C gradient during the Younger Dryas climatic event. Earth and Planetary Science Letters, 126, BENDER, M., SOWERS, T., DICKSON, M., ORCHADO, J., GROOTES, P., MAYEWSKI, P. and MEESE, D Climatic correlations between Greenland and Antarctica during the past 100,000 years. Nature, 372, BERGLUND, B. E., BIRKS, H. J. B., RALSKA-JASIEWICZOWA, M. and WRIGHT, H. E. (eds) Palaeoecological Events during the last 15,000 Years. Wiley, Chichester. BJÖRCK, S Bio- and chronostratigraphic significance of the Older Dryas Chronozone on the basis of new radiocarbon dates. Geologiska Föreningens i Stockholm Förhandlingar, 106, BJÖRCK, S. and HAKANSSON, S Radiocarbon dates from Late Weichselian lake sediments in South Sweden as a basis for chronostratigraphic subdivision. Boreas, 11, BJÖRCK, S., KROMER, B., JOHNSEN, S., BENNIKE, O., HAMMAR- LUND D., LEMDAHL, G., POSSNERT, G., RASMUSSEN, T. L., WOHLFARTH, B., HAMMER, C. U. and SPURK, M Synchronised terrestrial atmospheric deglacial records around the North Atlantic. Science, 274, BJÖRCK, S., RUNDGREN, M., INGOLFFSON, O. and FUNDER, S The Preboreal oscillation around the Nordic Seas: terrestrial and lacustrine respoonses. Journal of Quaternary Science, 12, BJÖRCK, S., BENNIKE, O., POSSNERT, G., WOHLFARTH, B. and DIGERFELDT, G A high-resolution 14 C dated sediment sequence from southwest Sweden: age comparisons between different components of the sediment. Journal of Quaternary Science, 13, BOHNCKE, S. J. P Lateglacial environmental changes in the Netherlands: spatial and temporal patterns. Quaternary Science Reviews, 12, BOND, G., BROECKER, W., JOHNSEN, S., McMANUS, J., LABEY- RIE, L., JOUZEL, J. and BONANI, G Correlation between climate records from North Atlantic sediments and Greenland ice. Nature, 365, COLMAN, S., JONES, G. A., MEYER RUBIN, KING, J., PECK, J. A. and OREM, W. H AMS radiocarbon analyses from Lake Baikal, Siberia: challenges of dating sediments from a large oligotrophic lake. Quaternary Science Reviews (Quaternary Geochronology), 15, COOPE, G. R. and LEMDAHL, G Regional differences in the Lateglacial climate of northern Europe based on coleopteran analysis. Journal of Quaternary Science, 10, CWYNAR, L. C., LEVESQUE, A. J., MAYLE, F. E. and WALKER, I Wisconsinan Late-glacial environmental change in New

9 EVENT STRATIGRAPHY FOR THE NORTH ATLANTIC REGION 291 Brunswick a regional synthesis. Journal of Quaternary Science, 9, DANSGAARD, W., JOHNEN, S. J., CLAUSEN, H. B., DAHL-JENSEN, D., GUNDESTRUP, N. S., HAMMER, C. U., HVIDBERG, C. S., STEFFENSEN, J. P., SVEINBJÖRNSDOTTIR, A. E., JOUZEL, J. and BOND, G Evidence for general instability of past climate from a 250-kyr ice-core record. Nature, 364, EICHER, U. and SIEGENTHALER, U Palynological and oxygen isotope investigations on Late-glacial cores from Swiss lakes. Boreas, 5, GOSLAR, T., ARNOLD, M., BARD, E., KUC, T., PAZDUR, M. F., RALSKA-JASIEWICZOWA, M., ROZANSKI, K., TISNERAT, N., WALANUS, A., WICIK, B. and WIECKOWSKI, K High concentration of atmospheric 14 C during the Younger Dryas cold episode. Nature, 377, GROOTES, P. M., STUIVER, M., WHITE, J. W. C., JOHNSEN, S. and JOUZEL, J Comparison of oxygen isotope records from the GISP2 and GRIP Greenland ice cores. Nature, 366, GULLIKSEN, S., BIRKS, H. H., POSSNERT, G. and MANGERUD, J The calendar age of the Younger-Dryas Holocene transition at Krakenes, western Norway. The Holocene, 8, HAFLIDASON, H., SEJRUP, H. P., KRISTENSEN, D. K. and JOHNSEN, S Coupled response of the late glacial climate shifts of northwest Europe reflected in Greenland ice cores: evidence from the northern North Sea. Geology, 23, HAMMER, C. U., ANDERSEN, K. K., CLAUSEN, H. B., DAHL- JENSEN, D., SCHØTT HVIDBERG, C. and IVERSEN, P The Stratigraphic Dating of the GRIP Ice Core. Special Report, Geophysical Department, Niels Bohr Institute for Astronomy, Physics and Geophysics, University of Copenhagen, Copenhagen. HARTZ, N. and MILTHERS, V Det senglaciale Ler i Allerød Teglv rksgrav. Meddelelser Dansk Geologisk Foreningen, 8, HEDBERG, H. D. (ed.) Preliminary report on stratotypes. International Subcommission on Stratigraphic Classification, Report 4, 39 pp. 24th International Geological Congress, Montreal, Canada. HUGHEN, K. A., OVERPECK, J. T., PETERSON, L. C. and TRUM- BORE, S Rapid climate changes in the tropical North Altantic region during the last deglaciation. Nature, 380, HUGHEN, K. A., OVERPECK, J. T., LEHMAN, S. J., KASHGARIAN, M., SOUTHON, J., PETERSON, J. C., ALLEY, R. and SIGMAN, D. M Deglacial changes in ocean circulation from an extended radiocarbon calibration. Nature, 391, INGÓLFSSON, Ó., BJÖRCK, S., HAFLIDASON, H. and RUNDGREN, M Glacial and climatic events on Iceland reflecting regional North Atlantic climatic shifts during the Pleistocene Holocene transition. Quaternary Science Reviews, 16, IVERSEN, J En pollenanalytisk Tidsf telse af Ferskvandslagene ved Nørre Lyngby. Meddelelser Dansk Geologiska Föreningens, 10, IVERSEN, J The late-glacial flora of Denmark and its relation to climate and soil. Danmarks Geologiske Undersøgelser, II, R ke 80, JESSEN, K Archaeological dating in the history of North Jutland s vegetation. Acta Archaeologica, 5, JESSEN, K Some West Baltic pollen diagrams. Quartär, 1, JOHNSEN, S., CLAUSEN, H. B., DANSGAARD, W., GUNDESTRUP, N., HANSSON, M., JONSSON, P., STEFFENSEN, J. P. and SVEINBJøRNSDOTTIR, A. E. 1992a. A deep ice core from East Greenland. Meddelelser om Grønland, Geoscience, 29, 22 pp. JOHNSEN, S. J., CLAUSEN, H. B., DANSGAARD, W., FUHRER, K., GUNDESTRUP, N., HAMMER, C. U., IVERSEN, P., JOUZEL, J., STAUFFER, B. and STEFFENSEN, J. P. 1992b. Irregular interstadials recorded in a new Greenland ice core. Nature, 359, JOHNSEN, S. J., DAHL-JENSEN, D., DANSGAARD, W. and GUN- DESTRUP, N Greenland palaeotemperatures derived from GRIP borehole temperature and ice core profiles. Tellus, 47B, KAPSNER, W. R., ALLEY, R. B., SHUMAN, C. A., ANANDAKRISH- NAN, S. and GROOTES, P. M Dominant influence of atmospheric circulation on snow accumulation in Greenland over the past 18,000 years. Nature, 373, KITAGAWA, H. and VAN DER PLICHT, J Atmospheric radiocarbon calibration to 45,000 yr BP: Late Glacial fluctuations and cosmogenic isotope production. Science, 279, KOÇ KARPUZ, N. and JANSEN, E A high-resolution diatom record of the last deglaciation from the SE Norwegian Sea: documentation of rapid cliamtic changes. Paleoceanography, 5, KROMER, B. and BECKER, B German oak and pine 14 C calibration, BC. Radiocarbon, 35, LOTTER, A., EICHER, U., BIRKS, H. J. B. and SIEGENTHALER, U Late-glacial climate oscillations as recorded in Swiss lake sediments. Journal of Quaternary Science, 7, LOWE, J. J Stratigraphic resolution and radiocarbon dating of Late Devensian Lateglacial sediments. IN: Radiocarbon Dating: Recent Applications and Future Potential. Quaternary Proceedings, 1. Quaternary Research Association, Cambridge. LOWE, J. J. and GRAY, J. M The stratigraphic subdivision of the Lateglacial of NW Europe: a discussion. IN: Lowe, J. J., Gray, J. M. and Robinson, J. E. (eds), Studies in the Lateglacial of North- West Europe, Pergamon Press, Oxford. LOWE, J. J. and NASP Members Palaeoclimate of the North Atlantic seaboards during the Last Glacial/Interglacial Transition. Quaternary International, 28, LOWE, J. J. and WALKER, M. J. C Problems associated with radiocarbon dating the close of the Lateglacial period in the Rannoch Moor area Scotland IN: Lowe, J. J., Gray, J. M. and Robinson, J. E. (eds), Studies in the Lateglacial of North-West Europe, Pergamon Press, Oxford. LOWE, J. J., AMMANN, B., BIRKS, H. H., BJÖRCK, S., COOPE, G. R., CWYNAR, L., DE BEAULIEW, J-L., MOTT, R. J., PETEET, D. M. and WALKER, M. J. C Climatic changes in areas adjacent to the North Atlantic during the Last Glacial-interglacial transition (14 9 ka BP). Journal of Quaternary Science, 9, MANGERUD. J Late Weichselian vegetation and icefront oscillations in the Bergen district, western Norway. Norsk Geografisk Tidsskrift, 24, MANGERUD, J. and BERGLUND, B. E The subdivision of the Quaternary of Norden: a discussion. Boreas, 7, MANGERUD, J. and GULLIKSEN, S Apparent radiocarbon ages of recent marine shells from Norway, Spitsbergen and Arctic Canada. Quaternary Research, 5, MANGERUD, J., ANDERSEN, S. T., BERGLUND, B. E. and DONNER, J. J Quaternary stratigraphy of Norden, a proposal for terminology and classification. Boreas, 4, MEESE, D. A., GOW, A. J., ALLEY, R. B., ZIELINSKI, G. A., GROOTES, P. M., RAM, M., TAYLOR, K. C., MAYEWSKI, P. A. and BOLZAN, J. F The Greenland Ice Sheet Project 2 depth-age scale: methods and results. Journal of Geophysical Research, 102 (C12), MOTT, R. J., GRANT, D. R., STEA, R. and OCHIETTI, S Late-glacial climatic oscillation in Atlantic Canada equivalent to the Allerød/Younger Dryas event. Nature, 323, NILSSON, T Ein neues Standardpollendiagramm aus Bjärsjöholmssjön in Schonen. Lunds Universitet Årskr N.F., 2 (44), 80 pp. PETEET, D. M., DANIELS, R., HEUSSER, L. E., VOGEL, J. S., SOU- THON, J. R. and NELSON, D. E Wisconsinan Late-glacial environmental change in southern New England: a regional synthesis. Journal of Quaternary Science, 9, RASMUSSEN, T. L., THOMSEN, E., LABEYRIE, L. and VAN WEER- ING, T. C. E Circulation changes in the Faroe Shetland Channel correlating with cold events during the last glacial period (58 10 ka). Geology, 24, SPURK, M., HOFMANN, J., FRIEDRICH, M., REMMELE, S., LEUSCHNER, H. H. and KROMER, B Revision and extension of the Hohenheim oak and pine chronologies new evidence about the timing of the Younger Dryas/Preboreal transition. Radiocarbon, in press. STUIVER, M. and REIMER, P. J Extended 14 C data base and