Establishing a terrestrial chronological framework as a basis for biostratigraphical comparisons

Quaternary Science Reviews 20 (2001) 1583–1592

Establishing a terrestrial chronological framework as a basis for biostratigraphical comparisons P.C. Tzedakisa,*, V. Andrieub, J.-L. de Beaulieub, H.J.B. Birksc,d, S. Crowhurste, M. Follierif, H. Hooghiemstrag,h, D. Magrif, M. Reilleb, L. Sadorif, N.J. Shackletone, T.A. Wijmstrag,h a

School of Geography, University of Leeds, Leeds LS2 9JT, UK Laboratoire de Botanique historique et Palynologie, ESA CNRS 6116, 13397 Marseille cedex 20, France c Botanical Institute, University of Bergen, Alle!gaten 41, N-5007, Bergen, Norway d Environmental Change Research Centre, University College London, WC1 0AP, UK e Department of Earth Sciences, Godwin Laboratory, Godwin Institute for Quaternary Research, New Museums Site, Pembroke Street, Cambridge CB2 3SA, UK f Dipartimento di Biologia Vegetale, Universita" di Roma ‘‘La Sapienza’’, P.le Aldo Moro 5, 00185 Roma, Italy g Hugo de Vries Laboratory, Department of Palynology and Paleo/Actuo-Ecology, University of Amsterdam, Kruislaan 318, 1098 SM Amsterdam, The Netherlands h The Netherlands Centre for Geo-ecological Research (ICG), Netherlands b

Abstract The palynological signature of interglacial deposits in the fragmentary European terrestrial record has often been used as the basis for determining their chronostratigraphical position and ultimately their age. This has placed emphasis on the presence/absence and abundance of certain characteristic taxa, but given the lack of continuous stratigraphies and independent chronologies, it has been difficult to assess the extent to which this strategy has produced reliable schemes. Here, an alternative approach is adopted whereby a chronological framework is developed for long and continuous pollen sequences from southern Europe. This in turn allows the emergence of a complete stratigraphical scheme of major vegetation events for the last 430 thousand years (ka) and the evaluation of the stage record of different taxa and their potential diagnostic value for biostratigraphical correlation. The comparison shows distinct similarities among some temperate stages of the terrestrial equivalent complexes of Marine Isotope Stages (MIS) 5 and 7 and also of MIS 9 and 11, but examination of combined records of taxa provides a possibility to differentiate between individual stages. A numerically-derived dichotomous key for the terrestrial stages based on the palynological records of 10 taxa is presented. Carpinus, Fagus, Abies, Pterocarya and Buxus emerge as the best ‘indicator pollen types’ because of their variable behaviour from one stage to the next, possibly a result of their late expansion within a temperate stage or reduced genetic variability. The analysis shows that the palynological signature of a temperate deposit can constrain the range of chronostratigraphical possibilities, but vegetation and palynological variability arising from local factors could result in difficulties in making a definite assignment at individual sites. r 2001 Elsevier Science Ltd. All rights reserved.

1. Introduction Our understanding of Pleistocene vegetation change within and between interglacials has been historically dominated by the northern European record where biostratigraphy has frequently been the basis for correlation and the development of chronostratigraphical schemes, as well as the evidence for vegetation dynamics and comparisons. The fragmentary nature of *Corresponding author. Fax: +44-113-2333-308. E-mail address: [email protected] (P.C. Tzedakis).

the northern European record means that all assessments of presence, absence, or abundance of particular taxa within any given interglacial depend upon accurate correlation of sequences. As correlation often relies on biostratigraphical evidence, there is clearly a danger of circular argument which will persist until each sequence has an independent chronology. However, in sharp contrast to the well-established marine chronology, a sufficiently precise terrestrial timescale is still far from being developed because of problems associated with direct dating of terrestrial material of Pleistocene age, which remain despite advances in geochronological

0277-3791/01/$ - see front matter r 2001 Elsevier Science Ltd. All rights reserved. PII: S 0 2 7 7 - 3 7 9 1 ( 0 1 ) 0 0 0 2 5 - 7

1584

P.C. Tzedakis et al. / Quaternary Science Reviews 20 (2001) 1583–1592

techniques. While the presence of certain datums (e.g. magnetic reversals, tephras) which can provide reliable age estimates can be of great assistance, their occurrence is often too irregular for a systematic development of terrestrial chronologies. In some cases, the presence of annual laminations along at least part of sequence combined with estimates of sediment accumulation rates can lead to the construction of detailed age-depth models (e.g. Magri, 1989; Zolitschka and Negendank, 1996; Allen et al., 1999). However, such sequences represent the exception, with most sequences containing unlaminated sediments with no structure. Here, an attempt is made to explore approaches of constructing chronological schemes for long pollen sequences from southern Europe. These sequences provide relatively complete and high-resolution records spanning multiple glacial-interglacial cycles. They represent, therefore, a convenient starting point for the systematic development of terrestrial chronologies. This can then allow comparisons of the vegetational character of coeval periods in different sequences and contribute towards our understanding of the spatial variation in the response of different taxa. It also provides a framework within which the diagnostic value of different taxa for biostratigraphical correlations can be critically examined and then perhaps applied to the fragmentary record.

2. Development of terrestrial chronologies One approach is the linking of terrestrial and marine records directly through joint pollen analysis and oxygen isotope measurements on benthic foraminifera from the same sample set in marine cores. Such a coupling can then allow an in situ assessment of relative leads and lags and the use of the timescale of the marine isotopic stratigraphy for dating land palynological events. If certain major vegetation events can be identified in marine pollen sequences that can be linked to reference terrestrial records, then the timescale of the marine isotopic stratigraphy can be applied to land sequences. Given that the long sequences are located in southern Europe, it is to the Mediterranean Sea that we should be looking to generate detailed marine pollen records. At present, however, high-resolution joint pollen-isotope studies from marine cores in the Mediterranean spanning multiple glacial-interglacial cycles and located near the southern European margin are unavailable. In many of the long marine sequences, pollen is absent from pelagic sediments. Pollen is very well preserved in sapropel layers, formed during intermittent periods of anoxia and correlated with northern insolation maxima (e.g. Rossignol-Strick, 1983), but this represents only a small part of the total

sediment sequence and therefore produces an incomplete record. In certain parts of the Mediterranean Sea, such as permanently anoxic basins or high sedimentation areas, complete long pollen records have been derived, (e.g. Cheddadi et al., 1991; Cheddadi and Rossignol-Strick, 1995; Combouriet-Nebout et al., 1999), but at the moment their sampling resolution is relatively low (although this can be easily rectified in the future) and they are located far from the long terrestrial sequences, thus complicating direct palynological comparisons. Leaving the palynological issues aside, there are also complications with the oxygen-isotope records of Mediterranean marine sequences because of the discontinuous presence of benthic foraminifera. Even in cases where a complete benthic curve can be constructed, results show that they can reflect local conditions of Mediterranean deep-water formation and circulation changes which may override the global ice volume signal (e.g. Cacho et al., 2000). Isotopic records from the Mediterranean, therefore, are usually developed from planktonic foraminifera, which, however, contain a significant local temperature and salinity overprint and as such are not directly related to ice volume and sea level changes. Thus, although planktonic isotope curves can be used to develop a first-order timescale, they are not ideally-suited to resolving the precise phase relationship between continental and oceanic isotope stages. In view of the current limitations, an interim measure is the indirect linking of terrestrial and marine sequences by assuming a rapid response of vegetation to certain climate changes and by extension a small amount of diachroneity between marine and terrestrial records as regards certain specific types of transitions. Such a coupling can then allow the transfer of the marine ages for these transitions to the corresponding boundaries on land records. Tzedakis et al. (1997) used this method by aligning the four longest pollen sequences from southern Europe to the standard marine isotopic record. The terrestrial records used in that exercise are located between 391N and 451N, forming a west–east transect across southern Europe (Fig. 1): (i) Lac du Bouchet (Reille and de Beaulieu, 1990; de Beaulieu and Reille, 1995; Reille et al., 1998) (441550 N, 31470 E; 1200 m above sea level (a.s.l.)) and Praclaux (Reille and de Beaulieu, 1995) (441490 N, 31500 E; 1100 m a.s.l.), two adjacent volcanic craters in the Velay mountains, Massif Central, France. The two sequences have been linked together by the common occurrence of a trachytic tephra. A composite record was used (Bouchet D:0–110 ka before present (BP); Bouchet H: 110–287 ka BP; Praclaux: 287– 430 ka BP). (ii) Valle di Castiglione (Follieri et al., 1988; Magri, 1989) (411530 3000 N, 121450 3500 E; 44 m a.s.l.), a volcanic maar lake in central Italy. (iii) Ioannina (Tzedakis, 1993, 1994; Frogley, 1997) (391400 N,

P.C. Tzedakis et al. / Quaternary Science Reviews 20 (2001) 1583–1592

1585

Fig. 1. Location of the four pollen sequences.

201510 E; 470 m a.s.l.), a tectonic basin in northwest Greece. (iv) Tenaghi Philippon (Wijmstra, 1969; Wijmstra and Smit, 1976; Wijmstra and Groenhart, 1983; van der Wiel and Wijmstra, 1987; Heijnis, 1992) (411100 N, 241200 E; 40 m a.s.l.), a tectonic basin in eastern Macedonia, Greece. Each pollen data set was aligned to a target, the SPECMAP stacked d18O record of Imbrie et al. (1984). ‘Glacial-to-deglacial’ transitions were used as the tie points on the basis that these transitions have been relatively rapid events and the response of vegetation was essentially synchronous within the error estimates for the marine chronology (sensu Martinson et al., 1987). This exercise led to the emergence of systematically-derived timescales for the four terrestrial records (Fig. 2). Despite local floristic differences, the four pollen sequences display a surprising degree of similarity in the relative expansion and contraction of forest vs. open vegetation communities and show that the many substages into which the oceanic record is divided are also reflected in the continental record. It becomes possible, therefore, to discern a one-to-one broad correspondence between terrestrial stages and marine isotopic substages, although the precise length of time represented by the two types of units may not be the same. This is especially the case with regard to ‘deglacial-to-glacial’ stage transitions where the age of the marine and terrestrial boundaries may diverge (e.g. Allen et al., 1999) and it is for this reason that this type of boundary was not used in the alignment process. The development of a timescale for the four sequences by tuning to the same standard also indirectly led to the correlation of local terrestrial stages and the emergence of a coherent stratigraphical scheme (Table 1). Also shown are the equivalent marine isotopic substages and, although the two are not strictly time parallel, for the

sake of simplification we use the marine notation when referring to specific terrestrial temperate stages. This stratigraphical scheme can provide a template of major events that should be present in records not only south but also north of the Alps. For example, it has long been appreciated that the equivalent temperate stages within the MIS 5 complex in northern Europe are represented by the Eemian, Brrup and Odderade periods (e.g. Behre, 1989; Mangerud, 1991). The registering of not only the interglacial but also the forested interstadials within the stratigraphical sequence of MIS 5 of northern Europe (e.g. Gru. ger, 1991) is significant. It suggests that temperate stages within the complexes of MIS 7 and 9 should be similarly characterised by an expansion of northern European tree populations. By extension, these periods should, in theory, be represented in the northern European stratigraphical succession as intervals of increased arboreal pollen frequencies, provided the right conditions for deposition and preservation are met (though this is not an insignificant problem (e.g. Turner, 1998)). Thus, the sequence of events shown in Fig. 2, instead of being dismissed as irrelevant for higher latitudes, might provide a framework within which northern European records can be assessed and eventually incorporated.

3. Evaluation of the stage record of different taxa In the European pollen literature, the behaviour of certain trees during particular temperate stages has often been highlighted as diagnostic for chronostratigraphical assignment. Thus, studies have tended to emphasise the characteristic high pollen values of Carpinus and the virtual absence of Fagus during the Eemian, and the importance of Abies as a telocratic tree in the

1586

P.C. Tzedakis et al. / Quaternary Science Reviews 20 (2001) 1583–1592

Fig. 2. Alignment of terrestrial pollen records to the marine isotopic sequence (from Tzedakis et al., 1997). For the terrestrial records the arboreal (AP) minus Pinus (solid line) curve has been relied for correlations and tuning, but the AP curve including Pinus (dotted line) is also shown. Marine isotope stages are indicated. Following the notation of Tzedakis et al. (1997), MIS 9a corresponds to the interval containing isotopic event 8.5 and, by extension, the other deglacial intervals of the MIS 9 complex have been redefined as 9c and 9e. Diamonds show the control points used for tuning. Table 1 Correlation scheme of temperate stages for the four pollen sequences. Also shown are the equivalent marine isotopic stages and the control points used in the age model corresponding to the ‘glacial-to-deglacial’ transitions Bouchet/Praclaux

Valle di Castiglione

Ioannina 249

Tenaghi Philippon

Marine Stratigraphy

Cotnrol points (ka BP)

Holocene St Geneys II St Geneys Ic St Geneys Ia Ribains Le Bouchet 3 Le Bouchet 2 Le Bouchet 1 Amargiers Ussel Landos Praclaux Jagonas

Holocene VdC-14 VdC-12c VdC-12a VdC-10 Roma III Roma II Roma I

Holocene Vikos Perama Thyamis Metsovon IN-26 Zitsa IN-23a Katara IN-17 Pamvotis Dodoni II Dodoni I

Holocene Elevtheroupolis Drama Doxaton Pangaion H2-3 Symvolon H1 Symvolon Strymon Kavalla Krimenes Litochoris Lekanis Lekanis

1 5a 5c 5c 5e 7a 7c 7e 9a 9c 9e 11a 11c

11.5 84 F 103 128 200 222 245 293 F 339 372 423

Holsteinian (Phillips 1974; Turner, 1975; Watts 1988). Such considerations have often underpinned correlations in the fragmentary northern European record to the extent that isolated deposits would sometimes be

assigned to the above mentioned interglacials on the basis of the relative abundances of these taxa. Here the four long pollen sequences provide an opportunity to examine the stage record of these taxa from the

P.C. Tzedakis et al. / Quaternary Science Reviews 20 (2001) 1583–1592

1587

Fig. 3. Comparison of the Carpinus record in the four sequences. Temperate stages are highlighted.

viewpoint of the complete stratigraphical scheme constructed above. Figs. 3–5 show the frequencies of Carpinus, Fagus and Abies pollen at each sequence plotted against time. Inspection of the records provides support for the important roles of Carpinus and Abies during MIS 5e and 11c, respectively. However, it also clearly shows that the taxa display similar behaviour patterns during other stages (high Carpinus in MIS 7c and 7e and high Abies in MIS 9e) and raises some doubts over the ‘uniqueness’ of the Eemian and Holsteinian pollen signatures. This, in turn, implies that a number of deposits in the fragmentary record whose chronostratigraphy has relied primarily on the identification of a characteristic taxon may have been assigned to the wrong stage. By extension, it may also provide a partial explanation for the ‘shortage’ of temperate stages between the Eemian and Holsteinian in the northern European record, in addition to a geological explanation (i.e. lack of suitable basins for deposition and erosional activity (e.g. Turner, 1998)). Instead of relying on the presence/absence of a characteristic taxon, an attempt is made here to explore

the differences and similarities of individual temperate stages by considering the combined records of several taxa which form the main components of forest succession and occur at all sites (Table 2). The stages shown represent the main periods of forest expansion of the last 430 ka. Although the small number of sequences available here precludes any definite conclusions, coeval palynological similarities among these different records, situated in very different environments, may provide some indication of the presence of some pattern characterizing a particular stage. Examination of Table 2 reveals that most of the additional taxa (Quercus, Ulmus, Tilia, Corylus) show a consistent pattern of presence from one stage to the next. This underlines further the variable behaviour of Carpinus, Fagus and Abies and illustrates the reason they have been traditionally used for biostratigraphical correlation. The question, however, is why these differences in behaviour arise. We propose here two complementary explanations: (1) Differences in the relative timing of a taxon’s expansion. Despite compositional variations from one stage to another, vegetation development during the

1588

P.C. Tzedakis et al. / Quaternary Science Reviews 20 (2001) 1583–1592

Fig. 4. Comparison of the Fagus record in the four sequences. Temperate stages are highlighted.

course of a temperate stage generally shows a recurrent pattern of succession with certain taxa (Quercus, Ulmus, Tilia, Corylus) expanding early, while others (Carpinus, Fagus, Abies) do so later. It is possible that small variations in the pattern of vegetation succession arising from slight differences in climate or historical accidents of founding and occurring early in the course of a temperate stage could be magnified through time and influence the character of the ensuing succession and the chances for establishment and expansion of latecomers (e.g. Bennett, 1993). In other words, the later a taxon normally expands, the higher the probability that it may not be able to attain its usual niche and therefore on separate occasions Carpinus, Fagus and Abies (and also Picea in the Massif Central (Reille and de Beaulieu, 1995)) may have been at a disadvantage. (2) Differences in genetic variability. The inconsistent presence of Carpinus, Fagus and Abies may be associated with their low species: genus ratio and reduced variability in comparison with that of taxa like Quercus and Ulmus. Adverse climatic conditions or the spread of disease during a temperate stage or the preceding cold

stage have a higher probability of influencing taxa of reduced diversity. The effect of differences in taxonomic status was already recognised by West (1980) who drew attention to the fact that most of the Tertiary genera which became extirpated during the Pleistocene in Europe are monotypic and have restricted present ranges, while those with large variability have been consistently successful in surviving the environmental oscillations. In addition to these taxa, Olea is consistently present during all stages in the Greek and Italian sites (in the Massif Central its record is more variable and probably represents long-distance pollen transport) but shows significantly higher values during MIS 5e. This is probably a reflection of the climatic signature of this interval with increased summer radiation and associated higher temperature and evaporation regimes providing a competitive advantage for summer-drought resistant taxa. The inconsistent stage record of Buxus (B. sempervirens), on the other hand, is probably associated with reduced diversity as with the other monospecific taxa. This is also seen with Pterocarya (P.

1589

P.C. Tzedakis et al. / Quaternary Science Reviews 20 (2001) 1583–1592

Fig. 5. Comparison of the Abies record in the four sequences. Temperate stages are highlighted.

Table 2 Stage record of taxa forming the main components of forest succession and occurring at all four sitesa Stages

Olea

Quercus

Ulmus

Tilia

Corylus

Carpinus

Fagus

Abies

Buxus

Pterocarya

1 5a 5c 5e 7a 7c 7e 9a 9c 9e 11a 11c

+ + + ++ + + + + (+) + + +

++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++

+ + + + + + + + (+) + + +

++ + + + + + + + (+) + + +

++ + + + + + + (+) (+) + + +

(+) + + ++ + ++ ++ + (+) + + +/

+ (+) + + + (+) (+)  +  /+

+ + + + + + (+) (+)  ++ + ++

  + +  +  +  +  +

          + +

a

Legend: ‘‘’’ complete absence or presence at one site only. ‘‘(+)’’ low presence at two or more sites. ‘‘+’’ presence at two or more sites. ‘‘++’’ dominance in the assemblage or unusually increased abundances at two or more sites. ‘‘+/’’ presence at Greek sites and absence at French site. ‘‘/ +’’ absence at Greek sites and presence at French site. Note that Valle di Castiglione extends only to 300 ka BP. In addition, new evidence from the Ioannina basin (Frogley, 1997), suggests that MIS 7e in core 249 may not be fully represented and therefore the scheme for this stage only includes the record of the other three sites.

1590

P.C. Tzedakis et al. / Quaternary Science Reviews 20 (2001) 1583–1592

carpinifolia) which has disappeared from Europe and is presently confined to the Caucasus (e.g. Godwin, 1975). Against this background, the degree of differentiation between stages is assessed in terms of their palynological signature as defined by the combined record of 10 taxa shown in Table 2. A two-way indicator species analysis (TWINSPAN; Hill, 1979) was made of the palynological record of 10 taxa in 12 terrestrial stages (Table 2) to find potential ‘indicator pollen types’ for the different stages that could provide the basis for a simple key to identify the stages palynologically. The records in Table 2 were coded numerically as explained in Table 3 and the computations were made with TWINSPAN version 2.2a with strict convergence criteria. Identical results were obtained with three different numerical codings. A simple hierarchical dichotomous key was constructed (Table 3) on the basis of the indicator taxa identified for each division in the TWINSPAN polythetic divisive hierarchical classification. Pterocarya pollen serves to distinguish palynologically MIS 11a and 11c from the other stages, whereas the occurrence of Buxus, Fagus and Abies pollen separates MIS 11a from 11c. The relative abundance of Carpinus pollen distinguishes between MIS 1 and 9c, and MIS 5a, 5c, 5e, 7a, 7c, 7e, 9a and 9e. The relative importance of Fagus pollen discriminates between MIS 1 and 9c pollen records. MIS 5c, 7a, 7c and 9e are distinguished from MIS 5a, 5e, 7e and 9a on the basis of the occurrence and relative abundance of Fagus pollen. Subsequent divisions into MIS 5c, 7a, 7c and 9e are made on the basis of differences in the occurrences and representation of Buxus (absent in 7a), Carpinus (dominant in 7c) and Abies (dominant in 9e) pollen. MIS 5a, 5e, 7e and 9a are distinguished on the basis of the almost complete absence of Fagus pollen (5e), the low presence of Abies pollen (7e), and the presence of Buxus pollen (9a) or Abies pollen (5a). The basis of the dichotomous key (Table 3) is, of course, an over-simplification of complex and highly variable pollen-stratigraphical data (Table 2). Nonetheless, the key serves to illustrate possible ways of defining a palynological signature for individual temperate stages, which could be used as a basis for determining the chronostratigraphical position of isolated deposits. It is, however, important to emphasise the limitations of this approach. Thus, while Tables 2 and 3 may indicate that separation between stages is possible, great caution should be exercised as individual sites may show local particularities, which do not follow the general pattern. For example, the presence of Carpinus, Fagus and Abies during all temperate stages of MIS 5 and 7 at Valle di Castiglione, makes their differentiation difficult. Maximum expansion of Olea sets MIS 5e apart, but the rest of the stages may look very similar, so much so that if they occurred as isolated deposits in the Lazio region without a clear chrono-

Table 3 Hierarchical dichotomous key for the twelve terrestrial stages and ten taxa of Table 2 based on the indicator taxa identified in a TWINSPAN of Table 2 using the numerical coding described at the base of this tablea I. –

Pterocarya pollen present at two or more sites Pterocarya pollen completely absent or present at one site only

II III

II.

Buxus pollen with increased abundances at two or more sites. Fagus pollen present at the French sites, and Abies pollen dominant at two or more sites Buxus and Fagus pollen absent or present at one site only, and Abies pollen Present at two or more sites

MIS 11c

–

III. –

Carpinus pollen present or dominant at two or more sites Carpinus pollen with low presence at two or more sites

MIS 11a

V IV

IV. –

Fagus pollen present at two or more sites Fagus pollen completely absent or present at one site only

MIS 1 MIS 9c

V. –

Fagus pollen present at two or more sites Fagus pollen completely absent or present at one site only, or low presence at two or more sites

VI IX

VI. –

Buxus pollen present at two or more sites Buxus pollen completely absent or present at one site only

VII MIS 7a

VII.

Carpinus pollen dominant or with unusually increased abundances at two or more sites Carpinus pollen present at two or more sites

MIS 7c

–

VIII

VIII. Abies pollen dominant or with unusually increased abundances at two or more sites – Abies pollen present at two or more sites

MIS 5c

IX.

X

–

Fagus pollen with low presence at two or more sites Fagus pollen completely absent or present at one site only

MIS 9e

MIS 5e

X. –

Buxus pollen present at two or more sites Buxus pollen completely absent or present at one site only

MIS 9a XI

XI. –

Abies pollen present at two or more sites Abies pollen with low presence at two or more sites

MIS 5a MIS 7e

a

The categories of presence and abundance follow Table 2 and were coded numerically for the TWINSPAN as follows: ‘‘’’=0, ‘‘(+)’’=1, ‘‘+/’’ and ‘‘/+’’=2, ‘‘+’’=3, ‘‘++’’=5. Four pseudovariable cut levels were used in the TWINSPAN, namely 1, 2, 3 and 5.

stratigraphical context they might be easily and erroneously correlated. Another local variation is the reversal of roles of Fagus and Carpinus during 11c between Praclaux and the Greek records. In addition, although Pterocarya is supposed to make its last appearance during MIS 11, it is found during MIS 9e at Tenaghi Philippon and during MIS 7e and 7c at very low abundances at Valle di Castiglione. Finally,

P.C. Tzedakis et al. / Quaternary Science Reviews 20 (2001) 1583–1592

although the importance of Carpinus during MIS 5e has been emphasised here, even within Greece there is considerable regional variability from high values at Ioannina, to intermediate at Tenaghi Philippon (Fig. 3) to negligible presence in the last interglacial interval of a new long sequence from the Kopais basin, central Greece (Tzedakis, 1999). These differences between sites underline the importance of local factors (climate, topography, soils and historical aspects) in producing unique palynological signatures that will at one point or another complicate attempts to draw general rules.

4. Concluding remarks A timescale for the four longest pollen sequences in Europe was developed by tuning the terrestrial records to the marine isotopic stratigraphy using glacial-tointerglacial transitions as the tie points. Although the assumption of regular sediment accumulation rates between control points is certainly an over-simplification, this simple step of placing the records on a common timescale has produced a terrestrial scheme of events for the last 430 ka that can be compared with marine and ice-core records in terms of the amplitude and structure of changes (Tzedakis et al., 1997). Given the preliminary nature of the age model discussed here, there is still significant scope in developing improved terrestrial chronologies through a more precise alignment of the terrestrial sequences to the marine stratigraphy by taking into account the exact phase relationship between continental and oceanic stages. This can be achieved in carefully selected long marine sequences where both a continuous benthic isotopic stratigraphy of global ice volume changes and a highresolution pollen record representing an integrated picture of the adjacent landmass are generated as, for example, off the Portuguese margin (Shackleton et al., 2001). In addition, new evidence suggesting that certain recurring vegetation patterns may be a result of climate changes linked to specific orbital geometries (Magri and Tzedakis, 2000), could provide an opportunity to develop astronomically-calibrated timescales for pollen sequences. Although these improved chronologies are expected to produce readjustments in the precision of the timing of terrestrial stage boundaries, nevertheless the original timescale shown in Fig. 2 will remain fundamentally correct, in terms of the average age of the different temperate stages. It thus represents a reliable platform on which to build a terrestrial stratigraphical framework, which can be of value in assessing not only southern but also northern European pollen records. The development of a correlation scheme provided an opportunity to examine the behaviour of different tree taxa during successive stages and evaluate their diag-

1591

nostic value for chronostratigraphical purposes. Of the taxa examined, Carpinus, Fagus, Abies, Pterocarya and Buxus showed the greatest potential because of their variable behaviour from one stage to the next, possibly a result of their late expansion within a temperate stage or reduced genetic variability. Nevertheless, reliance on any single one of these could potentially lead to erroneous correlations as highlighted by similarities between the pollen signatures of certain stages of MIS 5 and 7 and of MIS 9 and 11. Use of combined records of several taxa suggested that it is possible to differentiate between stages and that therefore palynological characterization of a temperate deposit may provide a way of constraining the number of chronostratigraphical solutions. Schemes of the type presented in Tables 2 and 3, could certainly be expanded to include more sites with independent chronologies and more taxa and the dichotomous key may eventually find a wider application. However, ultimately the limitation of this approach appears to lie in the inherent variability of vegetation patterns and hence palynological records from one area to another as determined by local factors. This underlines the difficulties in attempting correlations relying on biostratigraphy alone in the absence of a continuous record and emphasises the importance of using several lines of evidence (e.g. lithostratigraphy, morphostratigraphy). Once this is in place, the pollen signature of the deposit can come into play and provide a more specific solution.

Acknowledgements We are grateful to R.C. Preece and S. Leroy for comments on the manuscript. PCT acknowledges a NERC Advanced Fellowship (GT5/ES/11) and a Fellowship from Robinson College, Cambridge. He thanks the Greek Institute of Geology and Mineral Exploration for their co-operation. Work on the French sites was funded by EU grants Euromaar and EVSV CT92-0133. MF, DM and LS acknowledge support from CNR project ‘Sedimentazione lacustre, paleoambiente e paleoclima’. HH thanks G. van Reenen for support in data handling.

References Allen, J.R.M., Brandt, U., Brauer, A., Hubberten, H.-W., Huntley, B., Keller, J., Kraml, M., Mackensen, A., Mingram, J., Negendank, J.F.W., Nowaczyk, N.R., Oberh.ansli, H., Watts, W.A., Wulf, S., Zolitschka, B., 1999. Rapid environmental changes in southern Europe during the last glacial period. Nature 400, 740–743. Behre, K.-E., 1989. Biostratigraphy of the last glacial period in Europe. Quaternary Science Reviews 8, 25–44.

1592

P.C. Tzedakis et al. / Quaternary Science Reviews 20 (2001) 1583–1592

Bennett, K.D., 1993. Holocene forest dynamics with respect to southern Ontario. Review of Palynology and Palaeobotany 79, 69–81. Cacho, I., Grimalt, J.O., Sierro, F.J., Shackleton, N.J., Canals, M., Evidence for enhanced Mediterranean thermohaline circulation during rapid climatic coolings. Earth and Planetary Science Letters, 183, 417–429. Cheddadi, R., Rossignol-Strick, M., Fontugne, M., 1991. Eastern Mediterranean palaeoclimates from 25 to 5 ka B.P. documented by pollen and isotopic analysis of a core in the anoxic Bannock Basin. Marine Geology 100, 53–566. Cheddadi, R., Rossignol-Strick, M., 1995. Eastern Mediterranean Quaternary paleoclimates from pollen and isotope records of marine cores in the Nile cone area. Paleoceanography 10, 291–300. Combouriet-Nebout, N. Londeix, L., Baudin, F., Turon, J-L., von Grafenstein, R., Zahn, R., 1999. Quaternary marine and continental palaeoenvironments in the western Mediterranean (SITE 976, Alboran Sea): palynological evidence. In: Zahn, R., Comas, M.C., Klaus, A. (Eds), Proceedings of the Ocean Drilling Program, Scientific Results 161, pp. 457–468. de Beaulieu, J.-L., Reille, M., 1995. Pollen records from the Velay craters: a review and correlation of the Holsteinian Interglacial with isotopic stage 11. Mededelingen Rijks Geologische Dienst 52, 59–70. Follieri, M., Magri, D., Sadori, L., 1988. –250,000-year pollen record from Valle di Castiglione (Roma). Pollen et Spores 30, 329–356. Frogley, M.R., Biostratigraphy, Palaeoecology and Geochemistry of a Long Lacustrine Sequence from NW Greece. Ph.D. thesis, University of Cambridge, 1997. Godwin, H., 1975. History of the British Flora. Cambridge University Press, Cambridge. Gru. ger, E., 1991. Late Quaternary biostratigraphy in northern Germany years B P 150,000–15,000. Striae 34, 7–14. Heijnis, H., Uranium/Thorium Dating of Late Pleistocene Peat Deposits in NW Europe. Ph.D. Thesis, Rijksuniversiteit Groningen, 1992. Hill, M.O., TWINSPANFA FORTRAN program for arranging multivariate data in an ordered two-way table by classification of individuals and attributes. Cornell University, Ithaca, NY, 1979. Imbrie, J., Hays, J.D., Martinson, D.G., McIntyre, A., Mix, A.C., Morley, J.J., Pisias, N.G., Prell, W.L., Shackleton, N.J., 1984. The orbital theory of Pleistocene climate: support from a revised chronology of the marine dO18 record. In: Berger, A.L., Imbrie, J., Hays, J., Kukla, G., Saltzman, B. (Eds.), Milankovitch and Climate. D. Reidel, Dordrecht, pp. 269–306. Magri, D., 1989. Interpreting long-term exponential growth of plant populations in a 250,000-year pollen record from Valle di Castiglione (Roma). New Phytologist 112, 123–128. Magri, D., Tzedakis, P.C., 2000. Orbital signatures and long-term vegetation patterns in the Mediterranean. Quaternary International 73–74, 69–78. Mangerud, J., 1991. The last interglacial/glacial cycle in northern Europe. In: Shane, L., Cushing, J. (Eds.), Quaternary Landscapes. University of Minnesota Press, Minneapolis, pp. 38–75. Martinson, D.G., Pisias, N.G., Hays, J.D., Imbrie, J., Moore Jr., T.C., Shackleton, N.J., 1987. Age dating and the orbital theory of ice ages: development of a high-resolution 0 to 300,000-year chronostratigraphy. Quaternary Research 27, 1–29.

Phillips, L., 1974. Vegetational history of the Ipswichian/Eemian interglacial in Britain and continental Europe. New Phytologist 73, 589–604. Reille, M., Andrieu, V., de Beaulieu, J.-L., Guenet, P., Goeury, C., 1998. A long pollen record from Lac du Bouchet, Massif Central, France for the period 325 to 100 ka (OIS 9c to OIS 5e). Quaternary Science Reviews 17, 1107–1123. Reille, M., de Beaulieu, J.-L., 1990. Pollen analysis of a long Upper Pleistocene continental sequence in a Velay maar (Massif Central, France). Palaeogeography Palaeoclimatology Palaeoecology 80, 35–48. Reille, M., de Beaulieu, J.-L., 1995. Long Pleistocene pollen records from the Praclaux Crater, south-central France. Quaternary Research 44, 205–215. Rossignol-Strick, M., 1983. African monsoons, an immediate climate response to orbital insolation. Nature 304, 46–49. Shackleton, N.J., Sanchez-Go*ni, M.F., Pailler, D., Lancelot, Y., Marine Isotope Substage 5e and the Eemian Interglacial. Global and Planetary Change in press. Turner, C., 1975. The correlation and duration of Middle Pleistocene interglacial periods in northwest Europe. In: Butzer, K., Isaac, G.L. (Eds.), After the Australopithecines. Mouton, The Hague, pp. 259–308. Turner, C., 1998. Volcanic maars, long Quaternary sequences and the work of the INQUA Subcomission on European Quaternary Stratigraphy. Quaternary International 47/48, 41–49. Tzedakis, P.C., 1993. Long-term tree populations in northwest Greece through multiple Quaternary climatic cycles. Nature 364, 437–440. Tzedakis, P.C., 1994. Vegetation change through glacial-interglacial cycles: a long pollen sequence perspective. Philosophical Transactions Royal Society London B 345, 403–432. Tzedakis, P.C., 1999. The last climatic cycle at Kopais, central Greece. Journal of the Geological Society London 156, 425–434. Tzedakis, P.C., Andrieu, V., de Beaulieu, J.-L., Crowhurst, S., Follieri, M., Hooghiemstra, H., Magri, D., Reille, M., Sadori, L., Shackleton, N.J., Wijmstra, T.A., 1997. Comparison of terrestrial and marine records of changing climate of the last 500,000 years. Earth and Planetary Science Letters 150, 171–176. van der Wiel, A.M., Wijmstra, T.A., 1987. Palynology of the lower part (78–120) of the core Tenaghi Philippon II, Middle Pleistocene of Macedonia, Greece. Review of Palaeobotany and Palynology 52, 73–88. Watts, W.A., 1988. Europe. In: Huntley, B., Webb III, T. (Eds.), Handbook of Vegetation Science 7 Vegetation History. Kluwer, Dordrecht, pp. 155–192. West, R.G., 1980. Pleistocene forest history in East Anglia. New Phytologist 85, 571–622. Wijmstra, T.A., 1969. Palynology of the first 30 m of a 120 m deep section in Northern Greece. Acta Botanica Neerlandica 18, 511–527. Wijmstra, T.A., Groenhart, M.C., 1983. Record of 700,000 years vegetational history in Eastern Macedonia (Greece). Revista de la Academia Colombiana Ciencias Exactas, F!ısicas y Naturales 15, 87–98. Wijmstra, T.A., Smit, A., 1976. Palynology of the middle part (30 to 78 m) of the 120 m deep section in Northern Greece (Macedonia). Acta Botanica Neerlandica 25, 297–312. Zolitschka, B., Negendank, J.F.W., 1996. Sedimentology, dating and palaeoclimatic interpretation of a 76.3 ka record from Lago Grande di Monticchio, southern Italy. Quaternary Science Reviews 15, 101–112.