Quaternary Science Reviews 30 (2011) 2805e2814
Contents lists available at ScienceDirect
Quaternary Science Reviews journal homepage: www.elsevier.com/locate/quascirev
The pace of Holocene vegetation change e testing for synchronous developments Thomas Giesecke a, *, K.D. Bennett b, c, H. John B. Birks d, e, f, g, Anne E. Bjune g, Elisaveta Bozilova h, Angelica Feurdean i, Walter Finsinger j, Cynthia Froyd k, Petr Pokorný l, Manfred Rösch m, Heikki Seppä n, Spasimir Tonkov h, Verushka Valsecchi o, Steffen Wolters p a
Department of Palynology and Climate Dynamics, Albrecht-von-Haller-Institute for Plant Sciences, University of Göttingen, Untere Karspüle 2, 37073 Göttingen, Germany School of Geography, Archaeology and Palaeoecology, Queen’s University Belfast, Belfast BT7 1NN, Northern Ireland, UK c Department of Earth Sciences, Uppsala University, Villav. 16, SE-752 36 Uppsala, Sweden d Department of Biology, University of Bergen, Thormøhlensgate 53A, N-5006 Bergen, Norway e Environmental Change Research Centre, University College London, London WC1E 6BT, UK f School of Geography and Environment, University of Oxford, Oxford OX1 3QY, UK g Uni Bjerknes Centre and Bjerknes Centre for Climate Research, Allégaten 55, N-5007 Bergen, Norway h Sofia University St. Kliment Ohridski, Faculty of Biology, Department of Botany, Laboratory of Palynology, 8 Dragan Tzankov blvd., Sofia 1164, Bulgaria i Biodiversity and Climate Research Centre (BiK-F), Senckenberganlage 25, 60325 Frankfurt, Germany j Centre for Bio-Archaeology and Ecology (UMR 5059 CNRS), 163 Rue A. Broussonnet, F-34090 Montpellier, France k Department of Geography, College of Science, Swansea University, Swansea SA2 8PP, UK l Center for Theoretical Study, Charles University in Prague and the Academy of Sciences of the Czech Republic, Jilská 1, 110 00 Praha, Czech Republic m Landesamt für Denkmalpflege im Regierungspräsidium Stuttgart, Fischersteig 9, 78343 Gaienhofen-Hemmenhofen, Germany n Department of Geosciences and Geography, University of Helsinki, P.O. Box 64, FI 00014, Finland o Institut des Sciences de l’Evolution CC 061, UMR CNRS 5554, Université de Montpellier, Place Eugéne Bataillon, 34095 Montpellier Cedex 05, France p Lower Saxony Institute for Historical Coastal Research, Viktoriastr, 26/28, 26382 Wilhelmshaven, Germany b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 24 January 2011 Received in revised form 13 June 2011 Accepted 20 June 2011 Available online 18 July 2011
Mid to high latitude forest ecosystems have undergone several major compositional changes during the Holocene. The temporal and spatial patterns of these vegetation changes hold potential information to their causes and triggers. Here we test the hypothesis that the timing of vegetation change was synchronous on a sub-continental scale, which implies a common trigger or a step-like change in climate parameters. Pollen diagrams from selected European regions were statistically divided into assemblage zones and the temporal pattern of the zone boundaries analysed. The results show that the temporal pattern of vegetation change was significantly different from random. Times of change cluster around 8.2, 4.8, 3.7, and 1.2 ka, while times of higher than average stability were found around 2.1 and 5.1 ka. Compositional changes linked to the expansion of Corylus avellana and Alnus glutinosa centre around 10.6 and 9.5 ka, respectively. A climatic trigger initiating these changes may have occurred 0.5 to 1 ka earlier, respectively. The synchronous expansion of C. avellana and A. glutinosa exemplify that dispersal is not necessarily followed by population expansion. The partly synchronous, partly random expansion of A. glutinosa in adjacent European regions exemplifies that sudden synchronous population expansions are not species specific traits but vary regionally. Ó 2011 Elsevier Ltd. All rights reserved.
Keywords: Synchronous change Vegetation dynamics Climate change Holocene Species invasion Pollen
1. Introduction Pollen diagrams testify to the changes in vegetation composition over long periods and can give insights to past vegetation dynamics. However, they depict the result of a process, but little information of its cause. It is therefore not surprising that palaeoecologists have been arguing about the reasons for vegetation
* Corresponding author. Tel.: þ49 551 39 10675; fax: þ49 551 39 8449. E-mail address:
[email protected] (T. Giesecke). 0277-3791/$ e see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.quascirev.2011.06.014
change since the development of the subject. The comparison of pollen diagrams in space and through time offers some insights, and thus Rudolph (1930) suggested that the general succession of major forest elements in central Europe occurred synchronously. His summary of possible causes for Holocene changes in forest composition has changed little over the last 80 years and can be summarized as follows: I) an autogenous succession, controlled by competition and interaction with soil development; II) the distance to suitable Glacial maximum refugia with the delayed spread out of these areas; and III) climate change, which would determine the expansion of different species at different times. By rejecting the
2806
T. Giesecke et al. / Quaternary Science Reviews 30 (2011) 2805e2814
first two, Rudolph (1930) favoured the climate hypothesis. He assumed that all major forest species would have spread from their Glacial refugia during the Late Glacial and early Holocene before the time of the Corylus expansion (approximately 10.6 ka). With respect to Picea, Fagus and Carpinus similar ideas were already put forward by von Post (1924), who like Rudolph (1930) interpreted the expansion of different species as the result of a change in climate. Several other pioneers in the subject of vegetation history followed these ideas. For example Godwin (1975) argued that species may be present at abundances too low to be detected by pollen analysis, but may expand once the climate would become suitable. In contrast, Rudolph’s pupil Firbas (1949) stressed that seed dispersal and distance to Glacial maximum plant refugia can explain the general pattern in Holocene vegetation development. At the same time Firbas (1949) also acknowledged the importance of climate and hence followed Rudolph’s notion that the general succession is broadly synchronous, and provided an even more detailed scheme of it, which he advocated could be used for dating. In North America Davis (1976) showed how species spread during the Holocene and substantiated the hypothesis that life cycle, dispersal biology and distance to full Glacial refugia could explain the observed vegetation history. Prentice et al. (1991) condensed the three hypotheses into two, the disequilibrium and the dynamic equilibrium hypothesis with respect to climate as a variable factor, and showed that spatial patterns interpreted as range expansion of species limited by dispersal, could also be explained by directional changes in climate. Principally all climate reconstructions from pollen diagrams are based on the validity of the climate equilibrium hypothesis (Prentice et al., 1991). However, from the early simplification of climatic periods that are characterised by distinct climates with shifts in between them, like the BlytteSernander climate classification, these modern numerical reconstructions portray a general scenario of gradually changing climate parameters. Forests are rather stable systems, as trees are generally long lived, producing abundant seeds so that a gap in the canopy is immediately filled with a descendant of one of the surrounding individuals making it difficult for newcomers to establish. This inertia is broken in major disturbance events like extensive fires, which facilitate vegetation change (Green, 1982, 2006). Bradshaw and Hannon (1992) used this concept of disturbance to interpret a shift in vegetation composition from one stable state to another. Tinner and Lotter (2001, 2006) showed that the climate excursion around 8.2 ka could have acted just like a major disturbance event. By stressing the established tree populations during the climate reversal the inertia of the resident population is lowered allowing new arrivals to get established. Tinner and Lotter (2006) also mentioned that the climate excursion might have marked a shift in the moisture regime in the region and thus led to a persistent shift in vegetation composition. Many spatio-temporal patterns in pollen data-sets are often gradual in space and time and may be explained by the invasion of a species, a gradual change in climate or slowly increasing human activity. These patterns can be easily distinguished from synchronous changes in vegetation composition between distant sites. Such shifts can involve different species in different regions or constitute parallel developments in the abundance of the same species over vast areas. Synchronous vegetation change documented in many pollen diagrams from a large region may be caused by step-like changes in climate parameters or short climate excursions like the 8.2 ka event that would trigger synchronous shifts in vegetation composition. Through land-use change, human activity could also have brought about synchronous or asynchronous changes in vegetation composition, depending on how quickly a new technology or idea spread.
The aim of this paper is to test whether synchronicity in Holocene vegetation change was significant for European forest development, and thus evaluate the importance of abrupt climate change for forest ecosystems during the Holocene. 2. Methods 2.1. Data compilation We selected pollen analytical results from 59 sites throughout Europe (Fig. 1) in two ways: Data-sets that are continuous over 9000 years or more with at least six radiocarbon dates or varve chronology were extracted from the European Pollen Database (EPD). In parallel, palynologists were approached directly and invited to contribute their data to this study. In the latter case restrictions on the number of radiocarbon dates per sequence were relaxed if the available dates and sedimentation history indicated a near linear ageedepth relationship. In some occasions the authors provided an ageedepth model based on calibrated radiocarbon dates or varve chronology, and this information was used in the analysis. In most cases a new ageedepth model was constructed through calibrating the original radiocarbon dates using Bcal or Calib with the IntCal04 (Reimer et al., 2004) curve. Depths to age relationships were constructed using the mode or weighted average of the probability distribution for the calibrated age. Age models chosen depended on the amount of available radiocarbon dates and the complexity of the sedimentation history and range from linear interpolations and linear regression models to polynomials and smoothing splines as well as combinations of models for different sections of the sequence. Ageedepth models were fitted using psimpoll (Bennett, 2007) and MatLab. Pollen percentages were calculated from the count data and based on the sum of all terrestrial pollen and spores excluding Sphagnum spores. In a few cases the authors provided pollen percentage data based on the above sum. 2.2. Zonation and analysis The zonation of stratigraphical data can be carried out in an agglomerative way e.g. using constrained cluster analysis (CONISS; Grimm, 1987) or divisively by splitting the sequence into smaller and smaller sections using binary or optimal techniques (Gordon and Birks, 1972). Binary splitting first finds the point where a division reduces the overall variance most and then looks for further splits within the new sections. Optimal splitting always considers the full sequence at once irrespectively how many divisions are carried out. Bennett (1996) showed that the broken-stick model (MacArthur, 1957) can be used to determine the number of zones that are statistically significant. The model is conceptually similar to the binary splitting approach, and is also applicable to optimal splitting and CONISS (Bennett, 1996). Gordon and Birks (1972) argued that divisive techniques may be better suited to detect gradual but stratigraphically consistent changes and the analysis presented here was therefore restricted to the two divisive techniques binary and optimal splitting. The splitting techniques require a measure of variance that is additive, however the choice influences the result of the analysis. Here we used both the information content criterion and the sum of squared deviations (Gordon and Birks, 1972). Combining the two splitting techniques with the two measures of variance we carried out the resulting four combinations on all square root transformed data-sets including all terrestrial taxa that exceeded one percent in any one sample. The number of significant zones was evaluated using the broken-stick model and the ages for the significant splits were collated. This analysis was carried out using psimpoll (Bennett, 2007).
T. Giesecke et al. / Quaternary Science Reviews 30 (2011) 2805e2814
20°W
10°W
0°
10°E
20°E
2807
30°E
40°E
1 2 3 45 6
7 60°N
27 29 31
14
28 30 32 33
8 10
13 16 15 17 18
60°N
22 23 25
9
11 12
24 26
19 21
20
34 40 38
35
50°N
37
42
39
36
45 46
47 48 50
41 43 50°N
44 55 56
51
49
53
52 54
59
57 58
40°N
40°N 0°
10°E
20°E
Fig. 1. Map indicating the location of 59 pollen diagrams used in the analysis. Numbers are linked to sites and references in the online Appendix.
The resulting ages for the time of the split are associated with large uncertainties of two main sources. They carry all the uncertainty of the age model. In addition they represent the age of the midpoint between two samples, while there is no way of knowing when this change actually occurred, and thus this uncertainty increases with decreasing sample resolution. To take these uncertainties into account the results were handled in two ways: the ages were collected into bins of 200 years width and the frequency of these bins was further analysed. Secondly, the ages were uniformly treated as the mean of a Gaussian distribution with a variance of 100 years. The probabilities of all events were added, combining the results from the four analyses. The timing of splits were considered for the time up until 13 ka, as few data-sets reached far into the Late Glacial and their age control was seldom well constrained. Single, dominant changes like the shift from the Late Glacial to the Holocene can affect the overall variance of a data-set and thus influence this analysis. Therefore, all pollen data was subsequently restricted to 11 ka and 9 ka and the splitting techniques were run on these restricted data. The resulting frequencies and probabilities were adjusted by the number of available sequences by division of the available by the maximum number of sequences. The chance that a large number of randomly occurring events would fall into one of the 65 bins can be described by the Poisson distribution. Using a Chi square test this theoretical distribution was compared to the distribution of frequencies in the individual data-sets. This test evaluates the likelihood that for example two bins were selected more than 10 times, which would be unlikely in a random process, while the 10 times selection of only one bin may occur by chance. A rigorous statistical test of significance for the selection of individual bins is not straightforward, as the 59 diagrams span different time
intervals and were divided into different numbers of zones, with a low probability of two zones from the same diagram falling into the same bin. 3. Results The highest frequency of pollen zone boundaries is seen in the early Holocene before 9 ka, regardless of the splitting technique and the variance measure. Using the sum of squares variance criterion, 32% and 28% of all available pollen data-sets show a boundary between 11.4 and 11.6 ka, corresponding to the beginning of the Holocene. Splitting procedures using the information content criterion selected zone boundaries coinciding with the onset of the Holocene less frequently. They found splits around 10.7 ka and this bin was selected by the binary splitting procedure in 25% of available data-sets (Fig. 2). However, including all splits the corresponding peak in accumulated probabilities shifted towards a younger age of 10.55 ka (Fig. 3). A distinct peak at 9.54 ka in the accumulated probabilities (Fig. 3) is clearly visible in the histograms for binary splitting with information content and optimal splitting with sum of squares (Fig. 2). A broad peak in the accumulated probabilities (Fig. 3) centres at 8.28 ka, though the histograms indicate that the frequency of splits with this approximate age varies between techniques. Also all other peaks younger than 8 ka rarely show parallels between the four histograms. The comparison with the accumulated probabilities of the restricted data-sets (Fig. 4) shows that most of these peaks remain fixed even when the overall variance of the data-set was reduced. Only one broad new peak emerged in the 9 ka restricted data-set spanning 7.2e7.5 ka, while the peak at 6 ka became more important (Fig. 4).
2808
T. Giesecke et al. / Quaternary Science Reviews 30 (2011) 2805e2814
Fig. 2. Frequency of the timing in pollen zone boundaries collected into 200-year wide bins (light shaded); (dark shaded) show the relative frequency adjusted by the number of available pollen diagrams for each period.
The number of zone boundaries falling within each 200-year wide bin were expressed as frequency distributions and compared to Poisson distributions, which describe random encounters for the same number of observations as splits were found. The difference between the two distributions was evaluated by a Chi square test (Table 1). The observed distributions are often characterised by low and high frequency bins that occur more often than at random. The high frequency for some bins would be expected for cases where an underlying cause like the onset of the Holocene controlled a vegetation shift in many pollen diagrams. The case that average frequency bins occur more often in the observations compared to a random
process, may be caused by the fact that pollen diagrams tend to have zone boundaries more or less evenly spread over the Holocene after 9 ka. However, cases with more than expected low frequency bins indicate that vegetation composition remained stable over particular periods. The different measures of variance show a larger influence on the absolute number of zones compared to the different splitting techniques. The original 59 pollen data-sets yielded a total of 225 and 228 splits when information content was used and 260 and 264 splits with sums of squares. The resulting average of five significant zones per data-set varies between three and ten zones (Appendix). Approximately one in three pollen diagrams depicts clear changes at the Younger Dryas/Holocene transition that was picked up by the zonation techniques. It is interesting to see that this peak in coinciding zone boundaries is well defined in time and not spread
0.18
Accumulated probability
0.16 0.14
3.7
1 1.2
4.8
3.2
6
8 8.3 7.2 7.5
0.12 0.1 0.08 0.06 0.04 0.02 0
Fig. 3. Accumulated probability for the time of pollen zone boundaries, combining all results from the different techniques and variance measures. The solid line shows the original data; the dashed line is adjusted with respect to the number of pollen diagrams available; the grey line marks the proportion of diagrams covering a given point in time.
0
2000
4000
6000
8000
10000
12000
Age in years
Fig. 4. Accumulated probability for the time of pollen zone boundaries as in Fig. 3. (solid line) compared to the results of the restriction to the last 11 ka (short dashes) and to the last 9 ka (long dashes). The age of the peaks is given in thousands of years.
T. Giesecke et al. / Quaternary Science Reviews 30 (2011) 2805e2814 Table 1 Results of the Chi square test for the randomness of the frequency distribution for zone boundaries falling into a 200-year bin and the critical value for 95%. Test results exceeding the critical value are marked with an asterisk. The variance measures were abbreviated: sum of squares ¼ s.q., information content ¼ info.; adjusted refers to the frequency distribution adjusted by the availability of sites; restricted refers to the analysis of pollen data-sets restricted to the last 11 ka and 9 ka. Splitting technique
Binary s.q. Binary s.q. adjusted Binary info. Binary info. adjusted Optimal s.q. Optimal s.q. adjusted Optimal info. Optimal info. adjusted
Unrestricted
11 ka restricted
9 ka restricted
Chi
Critical value
Chi
Critical value
Chi
Critical value
5.0 5392.0* 22.2* 525.1* 20.8* 437.0* 3.4 20.6
15.5 28.9 16.9 23.7 18.3 26.3 12.6 21.0
14.2 34.3* 10.9 11.0 20.8* 37.4* 4.3 2.9
16.9 22.4 14.1 16.9 19.7 21.0 14.1 18.3
6.5 7.1 26.1* 24.3* 25.8* 26.2* 9.6 10.5
15.5 16.9 15.5 14.1 18.3 19.7 14.1 15.5
out over several hundred years, as could have been expected due to the varying dating control. This sets a benchmark for the evaluation of coinciding younger zone boundaries. Also worth noting is that this zone boundary was never the first split, indicating that this compositional change is statistically not the most important. On both sides of the Younger Dryas/Holocene boundary pollen spectra are often dominated by pine and birch pollen and this important climate change is therefore often not picked up by the statistical splitting of pollen diagrams from different European regions. 3.1. Synchronous changes involving different species Pollen zone boundaries falling into the bin around 8.3 ka and 3.7 ka were frequently selected (9 and 11 times, respectively) by the optimal splitting technique using the sum of square measure. When shifting the bins slightly eight diagrams show splits between 8.1 ka and 8.3 ka and another eight between 8.3 ka and 8.5 ka. Of the first eight diagrams four come from the British Isles while three of the splits around 8.4 ka are from Germany and two from northern Finland. The diagrams from Scotland and northern Finland show a rise of Pinus at this time, although that started somewhat earlier. We observed shifts of various pollen curves in several diagrams, but the onset of the continuous curve of a taxon (e.g. Tilia at Hockham Mere) was rare to coincide with zone boundaries around this time. Frequently diagrams show pronounced, short declines in some pollen curves (especially Corylus) dating to 8.2 ka as described by Tinner and Lotter (2001), but in many diagrams the curves quickly return to pervious values and in these cases no pollen zone boundary was set. Zone boundaries around 3.7 ka characterise a decline in Betula and Alnus in northern Norway, a decline in Pinus in northern Finland and declining Corylus curves with rising Carpinus and/or Fagus curves in Poland and Germany. While the peak around 8.3 ka is smeared over a wider time interval the peak at 3.7 is narrow and indicated mainly by the optimal splitting technique using the sum of square measure. In a back of the envelope calculation the chance of hitting a predefined 200-year wide bin between 0 and 10 ka may be compared to rolling the dice 200 times (number of splits for optimal splitting with sum of squares in this interval) yielding a probability of hitting it 9 or more times of 2%. This probability is a conservative estimate as diagrams with two zone boundaries per 200-year period are rare. Thus, while there are no parallel changes in vegetation composition around 8.3 and 3.7 ka, it is unlikely that the frequent occurrence of these particular zone boundaries was brought about solely by chance.
2809
3.2. Parallel patterns in pollen diagrams The individual pollen diagrams were scanned to explore which of the peaks in the timing of zone boundaries marks a similar change in vegetation composition across sites. The rise of the pollen curves of Corylus (Fig. 5) and Alnus (Fig. 6) could be associated to the peaks of 10.55 ka and 9.54 ka in the accumulated probability plot (Fig. 3). The parallel rise in the Corylus curve was found in many pollen diagrams especially from Western Europe. Noteworthy is the diagram from Dallican Water on Shetland (Bennett et al., 1992), where the Corylus curve runs parallel with several diagrams from continental Europe, even though the nearest point of the island is more than 150 km away from the Scottish mainland. The dynamics of the rising Corylus curve as well as the timing of its rise varied between sites and the latter may have been affected by the dating control. At some sites, however, the different timing for the rise of the Corylus curve could not be explained by uncertainties in the dating control. For example at Svarcenberk in the Czech Republic (Pokorný, 2002) the rise of the Corylus curve coincides approximately with the beginning of the Holocene. On the other hand, at sites in the mountains of Romania and Bulgaria the rise of Corylus displayed an altogether different pattern compared to North-West Europe (Fig. 5). Here the curves start at the same time as in diagrams towards the north-west, but the major rise and culmination date to younger ages. For all sites used in this study it can be assumed that the bulk of Corylus pollen originated from Corylus avellana. Corylus colurna occurs in Bulgaria as a rare species at elevations below 1200 m and may have contributed small and hence insignificant amounts of pollen to the two Bulgarian sites, so that also here the pollen type was interpreted to represent mainly C. avellana (Tonkov et al., 2002). The fast rise of the Alnus curve at 9.5 ka is synchronous in pollen diagrams around the Baltic Sea, while diagrams to the west and south are characterised by Alnus pollen curves that increase at different times and with varying slopes. Interesting to note is the rapid rise in the diagram from Lago Piccolo di Avigliana (Finsinger and Tinner, 2006) that almost coincides in time with that in the north-eastern diagrams, while the site is situated at the southeastern limit of the Alps. Also the Alnus curve from Meerfelder Maar (Kubitz, 2000) rises quickly, but only at 6 ka with the decline of Ulmus. In contrast many other diagrams show a gradual increase. 4. Discussion 4.1. Synchronous changes A large number of Holocene climate shifts or short-lived excursions are reported in the literature so that it seems almost possible to find one within the uncertainty of any standard radiocarbon date. However, tree-ring series can convincingly show that the Holocene had periods during which trees in specific habitats were stressed and others where growth was complacent (Leuschner et al., 2002). There are several external forcing mechanisms that work with different frequencies as well as resonating atmosphereeocean interactions, which in concert caused climate to constantly change (Wanner et al., 2008). Trees are long lived, produce abundant seeds and some are able to reproduce vegetative (e.g. Tilia). Therefore a high amplitude or long-lasting shift in climate parameters would be necessary to change the species composition of such a system. On the other hand, the effect of a small climatic excursion, like a number of dry years, could be amplified through the outbreak of a pathogen. Thus the peaks in the combined probability plot (Fig. 3) may hold some information on periods of change. In this respect the two troughs around 2.1 ka and 5.1 ka represent times that were consistently avoided by the placement of zone boundaries, indicating periods of vegetation stability.
2810
T. Giesecke et al. / Quaternary Science Reviews 30 (2011) 2805e2814
Fig. 5. Visual comparison of the Corylus pollen curve from selected pollen diagrams throughout Europe illustrating examples with broadly similar trends (1e13) and two examples with a later culmination of the curve (A, B).
The 8.2 ka event is the best described climate excursion during the Holocene (Alley and Ágústsdóttir, 2005; Seppä et al., 2007) and the various aspects of this shift may have spread over a period of up to 600 years (Rohling and Palike, 2005). The two optimal splitting techniques frequently selected zone boundaries near this time. Interestingly, the most frequent results using the information content (12 diagrams) fall into the interval 7.8e8.2 ka, while the sum of squares measure yields the highest frequency (14 diagrams) between 8.2 and 8.5 ka and there is no overlap between diagrams that were selected by either of the two measures within the combined interval. This could be caused by an initial shift in climate pattern that started 8.6 ka (Rohling and Palike, 2005) and was closely followed by the vegetation dynamics at some sites and individual species, seen in the pollen zone boundaries until 8.2 ka. Subsequently, the event also acted as a climatic disturbance reducing the inertia of established populations and thus allowing smaller but frequent shifts in the abundance of several species, highlighted by zone boundaries indicated by the optimal splitting with information content after 8.2 ka. Interesting is the fact that the decline of Ulmus, which is broadly synchronous in many pollen diagrams (Parker et al., 2002) did not give rise to a stronger peak in zone boundaries. This may be due to the relatively small abundance shift in many diagrams, eastern and northern sites, but also be partially an effect of the selection of pollen diagrams used. Also the much written about wet shift around 2.8 ka (van Geel et al., 1998) was not directly captured by the zonation techniques, but may be linked to a small peak at 2.7 ka (Fig. 4). In turn, the peak at 3.74 ka is well pronounced and based especially on
the results from the optimal splitting technique using the sum of squares measure. Based on this technique a zone boundary was placed between 3.6 ka and 3.8 ka in 11 out of 58 pollen diagrams, which is unlikely to have occurred by chance. This age coincides with one of the cold humid phases (CH-7) identified by Haas et al. (1998), which Tinner and Lotter (2006) argue to have triggered vegetation shifts like the 8.2 ka event. It also represents a time of a shift to a cooler late Holocene climate with an increased growth of Alpine glaciers especially after 3.3 ka and an advance of some smaller glaciers around 3.8 ka (Ivy-Ochs et al., 2009). Summer cooling around this time was also reconstructed from chironomid analysis (Heiri et al., 2003) and vegetation models based on this temperature reconstruction simulate the largest drop in Alpine treelines since 8.2 ka (Heiri et al., 2006). Pollen diagrams from the central European lowlands show a decline in C. avellana with a rise in Fagus sylvatica or Carpinus betulus for this time, which is the reason for the placement of the boundary in the sites from this region. This broadly parallel shift in vegetation composition was examined by Ralska-Jasiewiczowa et al. (2003), who point to human activity as the most important factor, driving the change. The authors also give minor importance to a shift to cooler and more humid climate. The three Norwegian diagrams and the north Finnish diagram that also have a zone boundary in this time window show a reduction in trees and an increase in fern spores, sedge or grass pollen which would be consistent with a general cooling. For the Scandinavian sites, the underlying changes in vegetation composition could have come about rapidly and possibly synchronous to a change in climate, while the expansion of F. sylvatica or C. betulus started often from
T. Giesecke et al. / Quaternary Science Reviews 30 (2011) 2805e2814
2811
Fig. 6. Visual comparison of the Alnus pollen curve from selected eight pollen diagrams around the Baltic Sea (1e8) with selected diagrams from central and western Europe (AeF).
small populations and thus a delay between cause and effect may have to be anticipated here. However, this delay may not be as long as during the early Holocene when populations started to develop from single individuals. During the late Holocene pollen of F. sylvatica and C. betulus are frequently found several hundred or thousand years before the curves rise distinctly and here the change in the slope of population increase should be the time to investigate when looking for causes of this expansion (Giesecke et al., 2007). The here reviewed data are inconclusive with respect to the cause of the change, but the coincidence of zone boundaries at this time is intriguing and should stimulate further ideas and studies. 4.2. Parallel changes The most widespread parallel shift in Holocene vegetation composition was the expansion of C. avellana populations, which occurred broadly synchronous at far apart sites regardless of their distance to possible LGM refugia, ice cover during the glacial or separation by water (Fig. 5; Tallantire, 2002; Finsinger et al., 2006). The trigger for the expansion of the populations must have operated on a large scale and is thus likely to have been climatic. A short-term fluctuation in climate influencing the competitive balance is not a possible mechanism. At the northern sites populations of the light demanding C. avellana expanded in an open forest dominated by Pinus sylvestris and Betula species (Firbas, 1949; Tallantire, 2002) and thus interspecies competition was probably of little importance. Therefore, a continued shift in climate parameters or patterns may have been the trigger for the synchronous expansion of C. avellana at far apart sites in western Europe. In his review of the hypothesis for
the rapid early Holocene spread of C. avellana Huntley (1993) concluded that the unique combinations of climate parameters favoured the rapid expansion. However, he did not address the synchronous population increase at far apart sites and pointed to plateaux in the radiocarbon timescale that in his view lead to the impression of spuriously rapid migration rates. Unlike Huntley (1993), Tallantire (2002) came to the conclusion that C. avellana must have been widespread in Europe before the rise of the pollen curve, but he argued that the populations were already large and their flowering suppressed under early Holocene climate. Although, C. avellana has been reported to reproduce by layering in marginal situations (Persson et al., 2004) propagation through seed is necessary for populations to expand and the pollen produced in the process should be visible in sediments from that time. However, the spread of this large pollen producer is not apparent palynologically and thus most likely occurred at low population density, possibly including frequent long distance dispersal (Giesecke, 2005c). Assuming that the rapid increase in pollen curves represents the increase of populations (Bennett, 1983, 1988) these populations must have started their growth before it is visible in the pollen diagrams (Bennett, 1988). Based on pollen accumulation rate estimates from Hockham Mere, Bennett (1983) estimated a doubling time for C. avellana populations between 32 and 53 years. Extrapolating the exponential population increase at Hockham Mere back in time would yield a supposed initial growth of the first scattered individuals to have started at about 11.2 ka, while the data from Meerfelder Maar point to a somewhat earlier initiation of growth. In time this corresponds to the end of the Preboreal oscillation at 11.25 ka and a proposed shift to more humid and warmer conditions
2812
T. Giesecke et al. / Quaternary Science Reviews 30 (2011) 2805e2814
(Bos et al., 2007; Magny et al., 2007) could have favoured if not triggered the expansion of C. avellana over a wide area. Considering the uncertainties of dating and early growth rates, it may not be necessary to invoke a climatic shift at the end of the Preboreal oscillation, but it could also be argued that the onset of the Holocene itself set the start for the expansion of C. avellana populations throughout north-western Europe. The timing of the rise in Alnus at sites around the Baltic Sea appears more confined compared to the C. avellana rise, but was spatially more restricted (Fig. 6). Although this expansion of Alnus can be ascribed mainly to Alnus glutinosa, Alnus incana probably played a part in this population expansion, at least at some sites (Giesecke, 2005a). In this respect it is interesting that the synchronous population expansion occurred largely in the area where both species are present today and hybridisation between the two species could have influenced the pattern. For example the hybrid A. incana glutinosa grows faster than A. glutinosa (Mejnartowicz, 1999) and the different preference of A. incana for climate and soils could have assisted in the spread of A. glutinosa. While this gives room for speculation it does not explain the synchronous expansion as such. Moreover, the regional synchronicity of the species expansion in one region is contrasted by the clear individual and site-specific development in the west, which has been shown for the British Isles (Bennett and Birks, 1990). As in the case of C. avellana, if climate was a trigger for this expansion, then this change must have preceded the rapid rise in the curve by a few hundred years. The central Swedish sites Holtjärnen and Abbortjärnen (Giesecke, 2005a,b) show doubling times for pollen accumulation rates between 80 and 130 years and the increase can be extrapolated backwards to about 10.3 ka (Bond event 7, Bond et al., 1997). C. avellana and Alnus are not the only two species for which a parallel increase in population between distant sites has been reconstructed. Also, for example, the Quercus curves show a parallel increase in many pollen diagrams, but with a lower rate, which is not giving rise to large changes in the pollen composition of adjacent samples. Apart from 10.6 ka and 9.5 ka all other peaks in zone boundaries are mainly caused by shifts in frequencies of pollen types that were present long before and after the change. These do not necessarily involve a long time lag as underlying changes in plant abundance are often smaller than a factor of two and can therefore take place within decades. However, where the change in pollen composition was brought about by the expansion of a population with a slow intrinsic growth rate (e.g. Fagus, Bradshaw et al., 2010) the trigger for the shift in vegetation composition may be found several hundred or a thousand years earlier. 4.3. Implications and open questions This study highlights that shifts in Holocene vegetation composition have often occurred at the same time, which corroborates the earlier notion by Tinner and Lotter (2006) that climate events like that at 8.2 ka triggered the expansion of F. sylvatica in different European regions. A climate trigger coincides with changes in vegetation composition around 8.2 ka, while the causes for some of the other periods of change identified here are less certain and warrant further investigation. The synchronous population peaks of Populus in North America bracketing the Younger Dryas cold phase were reactions to the effects that the changing climate had on its competitors (Peros et al., 2008). In contrast, the synchronous expansions of C. avellana and A. glutinosa, highlighted in this study, are direct responses to climate change. This is especially interesting in the case of C. avellana. Here we either have to assume extremely fast rates of immigration, in the order of 2000 m yr1, or a spread during the Late Glacial with survival during the Younger Dryas cold
event. The latter would be in line with Rudolph’s (1930) ideas and may be important for some other tree species as well. The synchronous expansion of C. avallana populations also suggests that a climatic threshold for the species was reached at the same time from Scotland to the Alps and at more continental sites in the East. A climatically controlled population expansion is exemplified by the parallel shape of the curves from north-western Italy (Finsinger and Tinner, 2006) and Shetland (Bennett et al., 1992). The example of A. glutinosa shows that the same species can show a different behaviour with respect to its regional expansion. The synchronous expansion in formally glaciated areas around the Baltic Sea shows that dispersal cannot have been limiting here and its spatially erratic expansion in the west is likely to reflect site specific developments (Bennett and Birks, 1990). This study has made use of statistical zonation techniques to identify the timing of significant vegetation change in different European regions. This composite record can be compared to the rate of change analysis of 18 North-East American pollen diagrams by Grimm and Jacobson (1992). Despite the different methods used, the general trends of both analyses are similar, as both highlight the beginning of the Holocene as a period of pronounced and frequent change in vegetation composition. Vescovi et al. (2007) showed that statistical zonation is a useful tool evaluating climatically induced parallel vegetation changes in several diagrams from the same region. Already the comparison of two pollen diagrams may be assisted by statistical zonation of pollen diagrams (e.g. Giesecke, 2005b). These statistically defined zone boundaries often differ from visually defined zones and thus the results of this analysis differ from Gajewski et al. (2006), who analysed the temporal distribution of pollen zone boundaries summarized by Berglund et al. (1996). The latter analysis had the additional problem that not all diagrams included in the book edited by Berglund et al. (1996) were well dated by independent means and ages for zone boundaries were often derived by correlation to distant diagrams with some independent dating, assuming that changes in vegetation pattern were synchronous. 5. Conclusion This analysis shows that tree-dominated vegetation in Europe has often changed synchronously over the last 11 500 years. The only driver that can synchronize vegetation change on a subcontinental scale is climate, both as abrupt step-like change, but also as climate excursions that work like large-scale disturbance events. The parallel population expansions of C. avellana and A. glutinosa represent a special case of synchronous developments. In both cases populations built up from palynologically undetectable occurrences of these large pollen producers. This implies that the species were widespread at low abundances prior to population expansions and a climate trigger initiated the growth of these populations over a wide area. The case of A. glutinosa is of particular interest as the species shows a synchronous expansion only around the Baltic Sea but a site specific, erratic behaviour in Western Europe. For both species the palynologically visible time of expansion does not hold any information to the time of their spread or even rates of spread. What is true for these two species may also play a role for many other trees that spread early without leaving evidence of their presence and only expanded when the climate became suitable (Rudolph, 1930; Godwin, 1975). Acknowledgements We wish to thank Anneli Poska and Ricarda Voigt for making their data available and commenting on the manuscript. Thanks are due to Judy Allen, Brigitta Ammann, Gunnar Digerfeldt, Hans Göransson,
T. Giesecke et al. / Quaternary Science Reviews 30 (2011) 2805e2814
Beate Kubitz, Andy Lotter, Andrzej Obidowicz, the late Winifred Pen_ nington, Reet Pirrus, Magdalena Ralska-Jasiewiczowa, Grazyna MiotkSzpiganowicz, Alan Smith, Kazimierz Tobolski, Irmeli Vuorela and Clare Watson, who submitted their data to the European pollen database and thus made this meta-analysis possible. We are also grateful to Willy Tinner and Simon Brewer for providing constructive comments on an earlier version of the manuscript. T. Giesecke acknowledges funding from the German Research Foundation through a personal grant GI 732/1-1. K.D. Bennett gratefully acknowledges a Royal Society e Wolfson Merit Award. Pollen data from the Czech Republic resulted from the project IAAX00020701. This is publication no. A337 from the Bjerknes Centre for Climate Research. Appendix. Supplementary material Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.quascirev.2011.06.014. References Alley, R.B., Ágústsdóttir, A.M., 2005. The 8 k event: cause and consequences of a major Holocene abrupt climate change. Quaternary Science Reviews 24, 1123e1149. Bennett, K.D., 1983. Postglacial population expansion of forest trees in Norfolk, UK. Nature 303, 164e167. Bennett, K.D., 1988. Holocene geographic spread and population expansion of Fagus grandifolia in Ontario, Canada. Journal of Ecology 76, 547e557. Bennett, K.D., 1996. Determination of the number of zones in a biostratigraphical sequence. New Phytologist 132, 155e170. Bennett, K.D., 2007. Psimpoll and Pscomb Programs for Plotting and Analysis. http:// www.chrono.qub.ac.uk/psimpoll/psimpoll.html. Bennett, K.D., Birks, H.J.B., 1990. Postglacial history of alder (Alnus glutinosa (L.) Gaertn.) in the British Isles. Journal of Quaternary Science 5, 123e133. Bennett, K.D., Boreham, S., Sharp, M.J., Switsur, V.R., 1992. Holocene history of environment, vegetation and human settlement on Catta Ness, Lunnasting, Shetland. Journal of Ecology 80, 241e273. Berglund, B.E., Birks, H.J.B., Ralska-Jasiewiczowa, M., Wright, H.E. (Eds.), 1996. Palaeoecological Events during the Last 15 000 Years. J Wiley, Chichester, p. 764. Bond, G., Showers, W., Cheseby, M., Lotti, R., Almasi, P., de Menocal, P., Priore, P., Cullen, H., Hajdas, I., Bonani, G., 1997. A pervasive millennial-scale cycle in North Atlantic Holocene and glacial climates. Science 278, 1257e1266. Bos, J.A.A., van Geel, B., van der Plicht, J., Bohncke, S.J.P., 2007. Preboreal climate oscillations in Europe: wiggle-match dating and synthesis of Dutch highresolution multi-proxy records. Quaternary Science Reviews 26, 1927e1950. Bradshaw, R., Hannon, G., 1992. Climatic-change, human influence and disturbance regime in the control of vegetation dynamics within Fiby forest, Sweden. Journal of Ecology 80, 625e632. Bradshaw, R.H.W., Kito, N., Giesecke, T., 2010. Factors influencing the Holocene history of Fagus. Forest Ecology and Management 259, 2204e2212. Davis, M.B., 1976. Pleistocene biogeography of temperate deciduous forests. Geoscience and Man 13, 13e26. Finsinger, W., Tinner, W., 2006. Holocene vegetation and land-use changes in response to climatic changes in the forelands of the southwestern Alps, Italy. Journal of Quaternary Science 21, 243e258. Finsinger, W., Tinner, W., van der Knaap, W.O., Ammann, B., 2006. The expansion of hazel (Corylus avellana L.) in the southern Alps: a key for understanding its early Holocene history in Europe? Quaternary Science Reviews 25, 612e631. Firbas, F., 1949. Spät- und nacheiszeitliche Waldgeschichte Mitteleuropas nördlich der Alpen, Allgemeine Waldgeschichte. Gustav Fischer, Jena. Gajewski, K., Viau, A.E., Sawada, M., Atkinson, D.E., Fines, P., 2006. Synchronicity in climate and vegetation transitions between Europe and North America during the Holocene. Climatic Change 78, 341e361. Giesecke, T., 2005a. Holocene forest development in the central Scandes Mountains, Sweden. Vegetation History and Archaeobotany 14, 133e147. Giesecke, T., 2005b. Holocene dynamics of the southern boreal forest in Sweden. Holocene 15, 858e872. Giesecke, T., 2005c. Moving front or population expansion: how did Picea abies (L.) Karst. become frequent in central Sweden? Quaternary Science Reviews 24, 2495e2509. Giesecke, T., Hickler, T., Kunkel, T., Sykes, M.T., Bradshaw, R.H.W., 2007. Towards an understanding of the Holocene distribution of Fagus sylvatica L. Journal of Biogeography 34, 118e131. Godwin, H., 1975. History of the British Flora. Cambridge University Press, Cambridge. Gordon, A.D., Birks, H.J.B., 1972. Numerical methods in Quaternary palaeoecology. I. Zonation of pollen diagrams. New Phytologist 71, 961e979. Green, D.G., 1982. Fire and stability in the postglacial forests of southwest Nova Scotia. Journal of Biogeography 9, 29e40.
2813
Grimm, E.C., 1987. CONISS: a FORTRAN 77 program for stratigraphically constrained cluster analysis by the method of incremental sum of squares. Computers and Geosciences 13, 13e35. Grimm, E.C., Jacobson, G.L., 1992. Fossil-pollen evidence for abrupt climate changes during the past 18 000 years in eastern North America. Climate Dynamics 6, 179e184. Haas, J.N., Richoz, I., Tinner, W., Wick, L., 1998. Synchronous Holocene climatic oscillations recorded on the Swiss Plateau and at Timberline in the Alps. The Holocene 8, 301e309. Heiri, O., Lotter, A.F., Hausmann, S., Kienast, F., 2003. A chironomid-based Holocene summer air temperature reconstruction from the Swiss Alps. Holocene 13, 477e484. Heiri, C., Bugmann, H., Tinner, W., Heiri, O., Lischke, H., 2006. A model-based reconstruction of Holocene treeline dynamics in the Central Swiss Alps. Journal of Ecology 94, 206e216. Huntley, B., 1993. Rapid early Holocene migration and highabundance of hazel (Corylus avellana L.): alternative hypotheses. In: Chambers, F.M. (Ed.), Climate Change and Human Impact on the Landscape. Chapman & Hall, London. Ivy-Ochs, S., Kerschner, H., Maisch, M., Christl, M., Kubik, P.W., Schluchter, C., 2009. Latest Pleistocene and Holocene glacier variations in the European Alps. Quaternary Science Reviews 28, 2137e2149. Kubitz, B., 2000. Die holozäne Vegetations- und Siedlungsgeschichte in der Westeifel am Beispiel eines hochauflösenden Pollendiagrammes aus dem Meerfelder Maar. Dissertationes Botanicae 339, 1e106. Leuschner, H.H., Sass-Klaassen, U., Jansma, E., Baillie, M.G.L., Spurk, M., 2002. Subfossil European bog oaks: population dynamics and long-term growth depressions as indicators of changes in the Holocene hydro-regime and climate. Holocene 12, 695e706. MacArthur, R.H., 1957. On the Relative Abundance of Bird Species. In: Proceedings of the National Academy of Sciences of the United States of America, 43 293e295. Magny, M., Vanniere, B., de Beaulieu, J.L., Begeot, C., Heiri, O., Millet, L., Peyron, O., Walter-Simonnet, A.V., 2007. Early-Holocene climatic oscillations recorded by lake-level fluctuations in west-central Europe and in central Italy. Quaternary Science Reviews 26, 1951e1964. Mejnartowicz, L., 1999. Evidence for long-term heterosis phenomenon in the Alnus incana glutinosa F-1 hybrids. Silvae Genetica 48, 100e103. Parker, A.G., Goudie, A.S., Anderson, D.E., Robinson, M.A., Bonsall, C., 2002. A review of the mid-Holocene elm decline in the British Isles. Progress in Physical Geography 26, 1e45. Peros, M.C., Gajewski, K., Viau, A.E., 2008. Continental-scale tree population response to rapid climate change, competition and disturbance. Global Ecology and Biogeography 17, 658e669. Persson, H., Widen, B., Andersson, S., Svensson, L., 2004. Allozyme diversity and genetic structure of marginal and central populations of Corylus avellana L. (Betulaceae) in Europe. Plant Systematics and Evolution 244, 157e179. Pokorný, P., 2002. A high-resolution record of Late-Glacial and Early-Holocene climatic and environmental change in the Czech Republic. Quaternary International 91, 101e122. Prentice, I.C., Bartlein, P.J., Webb, T., 1991. Vegetation and climate change in eastern North-America since the Last Glacial Maximum. Ecology 72, 2038e2056. Ralska-Jasiewiczowa, M., Nalepka, D., Goslar, T., 2003. Some problems of forest transformation at the transition to the oligocratic/Homo sapiens phase of the Holocene interglacial in northern lowlands of central Europe. Vegetation History and Archaeobotany 12, 233e247. Reimer, P.J., Baillie, M.G.L., Bard, E., Bayliss, A., Beck, J.W., Bertrand, C.J.H., Blackwell, P.G., Buck, C.E., Burr, G.S., Cutler, K.B., Damon, P.E., Edwards, R.L., Fairbanks, R.G., Friedrich, M., Guilderson, T.P., Hogg, A.G., Hughen, K.A., Kromer, B., McCormac, G., Manning, S., Ramsey, C.B., Reimer, R.W., Remmele, S., Southon, J.R., Stuiver, M., Talamo, S., Taylor, F.W., van der Plicht, J., Weyhenmeyer, C.E., 2004. IntCal04 terrestrial radiocarbon age calibration, 0e26 cal kyr BP. Radiocarbon 46, 1029e1058. Rohling, E.J., Palike, H., 2005. Centennial-scale climate cooling with a sudden cold event around 8,200 years ago. Nature 434, 975e979. Rudolph, K., 1930. Grundzüge der nacheiszeitlichen Waldgeschichte Mitteleuropas. Beihefte Botanisches Centralblatt 47, 11e176. Seppä, H., Birks, H.J.B., Giesecke, T., Hammarlund, D., Alenius, T., Antonsson, K., Bjune, A.E., Heikkila, M., MacDonald, G.M., Ojala, A.E.K., Telford, R.J., Veski, S., 2007. Spatial structure of the 8200 cal yr BP event in northern Europe. Climate of the Past 3, 225e236. Tallantire, P.A., 2002. The early-Holocene spread of hazel (Corylus avellana L.) in Europe north and west of the Alps: an ecological hypothesis. Holocene 12, 81e96. Tinner, W., Lotter, A.F., 2001. Central European vegetation response to abrupt climate change at 8.2 ka. Geology 29, 551e554. Tinner, W., Lotter, A.F., 2006. Holocene expansions of Fagus silvatica and Abies alba in Central Europe: where are we after eight decades of debate? Quaternary Science Reviews 25, 526e549. Tonkov, S., Panovska, H., Possnert, G., Bozilova, E., 2002. The Holocene vegetation history of Northern Pirin Mountain, southwestern Bulgaria: pollen analysis and radiocarbon dating of a core from Lake Ribno Banderishko. Holocene 12, 201e210. van Geel, B., Van der Plicht, J., Kilian, M.R., Klaver, E.R., Kouwenberg, J.H.M., Renssen, H., Reynaud-Farrera, I., Waterbolk, H.T., 1998. The sharp rise of Delta C14 ca. 800 cal BC: possible causes, related climatic teleconnections and the impact on human environments. Radiocarbon 40, 535e550. Vescovi, E., Ravazzi, C., Arpenti, E., Finsinger, W., Pini, R., Valsecchi, V., Wick, L., Ammann, B., Tinner, W., 2007. Interactions between climate and vegetation during
2814
T. Giesecke et al. / Quaternary Science Reviews 30 (2011) 2805e2814
the Lateglacial period as recorded by lake and mire sediment archives in Northern Italy and Southern Switzerland. Quaternary Science Reviews 26, 1650e1669. von Post, L., 1924. Ur de sydsvenska skogarnas regionala historia under postarktisk tid. Geologiska Föreningen Förhandlingar 46, 83e128. Wanner, H., Beer, J., Butikofer, J., Crowley, T.J., Cubasch, U., Fluckiger, J., Goosse, H., Grosjean, M., Joos, F., Kaplan, J.O., Kuttel, M., Muller, S.A., Prentice, I.C., Solomina, O., Stocker, T.F., Tarasov, P., Wagner, M., Widmann, M., 2008. Mid- to Late Holocene climate change: an overview. Quaternary Science Reviews 27, 1791e1828.
Further reading1 Allen, J.R.M., Huntley, B., Watts, W.A., 1996. The vegetation and climate of northwest Iberia over the last 14,000 years. Journal of Quaternary Science 11, 125e147. Ammann, B., 1985. Lobsingensee e Late Glacial and Holocene environments of a lake on the central Swiss plateau. Dissertationes Botanicae 87, 127e134. Bigler, C., Larocque, I., Peglar, S.M., Birks, H.J.B., Hall, R.I., 2002. Quantitative multiproxy assessment of long-term patterns of Holocene environmental change from a small lake near Abisko, northern Sweden. Holocene 12, 481e496. Birks, H.H., Birks, H.J.B., 2003. Reconstructing Holocene climates from pollen and plant macrofossils. In: Mackay, A., Battarbee, R.W., Birks, H.J.B., Oldfield, F. (Eds.), Global Change in the Holocene. Arnold, London, pp. 342e357. Bjune, A.E., Birks, H.J.B., Seppä, H., 2004. Holocene vegetation and climate history on a continentaleoceanic transect in northern Fennoscandia based on pollen and plant macrofossils. Boreas 33, 211e223. Bjune, A.E., 2005. Holocene vegetation history and tree-line changes on a northsouth transect crossing major climate gradients in southern Norway e evidence from pollen and plant macrofossils in lake sediments. Review of Palaeobotany and Palynology 133, 249e275. Bozilova, E., Tonkov, S., 2000. Pollen from Lake Sedmo Rilsko reveals southeast European postglacial vegetation in the highest mountain area of the Balkans. New Phytologist 148, 315e325. Digerfeldt, G., 1977. The Flandrian Development of Lake Flarken. Regional Vegetation History and Palaeoclimatology. University of Lund Department of Quaternary Geology, Report 13, 101 pp. Digerfeldt, G., 1982. The Holocene development of Lake Sambosjon. 1. The regional vegetation history. Lundqua Report 23, 1e24. Eide, W., Birks, H.H., Bigelow, N.H., Peglar, S.M., Birks, H.J.B., 2006. Holocene forest development along the Setesdal valley, southern Norway, reconstructed from macrofossil and pollen evidence. Vegetation History and Archaeobotany 15, 65e85. Feurdean, A., 2005. Holocene forest dynamics in northwestern Romania. Holocene 15, 435e446. Froyd, C.A., 2005. Fossil stomata, reveal early pine presence in Scotland: implications for postglacial colonization analyses. Ecology 86, 579e586. Froyd, C.A., 2006. Holocene fire in the Scottish Highlands: evidence from macroscopic’ charcoalrecords. Holocene 16, 235e249. Giesecke, T., 2001. Pollenanalytische und sedimentchemische Untersuchungen zur natürlichen und anthropogenen Geschichte im Schlaubetal. In: Sitzungsberichte der Gesellschaft Naturforschender Freunde zu Berlin, 39 89e112. Göransson, H., 1991. Vegetation and Man Around Lake Bjärsjöholmssjön during Prehistoric Time, Lundqua Report 31, 44 pp. Heikkila, M., Seppä, H., 2003. A 11,000 yr palaeotemperature reconstruction from the southern boreal zone in Finland. Quaternary Science Reviews 22, 541e554. Heinrichs, M.L., Peglar, S.M., Bigler, C., Birks, H.J.B., 2005. A multi-proxy palaeoecological study of Alanen Laanijarvi, a boreal-forest lake in Swedish Lapland. Boreas 34, 192e206. Lotter, A.F., 1988. Paläoökologische und paläolimnologische Studie des Rotsees bei Luzern. Pollen-, grossrest-, diatomeen- und sedimentanalytische Untersuchungen. Dissertationes Botanicae 124, 187. Miotk-Szpiganowicz, G., 1992. The history of vegetation of Bory Tucholskie and the role of man in the lighht of palynological investigations. Acta Palaeobotanica 32, 39e122.
1
References to pollen diagrams cited only in Supplementary Information.
Obidowicz, A., 1989. Type region P-a: inner West Carpathians e Nowy Targ Basin. Acta Palaeobotanica 29, 11e15. Pawlikowski, M., Ralska-Jasiewiczowa, M., Schönborn, W., Stupnicka, E., Szeroczynska, K., 1982. Woryty near Gietrzwald, Olsztyn Lake District, NE Poland e vegetational history and lake development during the last 12,000 years. Acta Palaeobotanica 22, 85e116. Pennington, W., 1975. A chronostratigraphic comparison of late Weichselian and late Devensian subdivisions illustrated by two radiocarbon dated profiles from western Britain. Boreas 4, 157e171. Pennington, W., Haworth, E.Y., Bonny, A.P., Lishman, J.P., 1972. Lake sediments in northern Scotland. Philosophical Transactions of the Royal Society B 264, 191e294. Pirrus, R., Róuk, A.-M., Liiva, A., 1987. Geology and stratigraphy of the reference site of lake Raigastvere in Saadjärv drumlin field. In: Raukas, A.S., Saarse, L. (Eds.), Palaeohydrology of the Temperate Zone II, Lakes. Valgus, Tallinn, pp. 101e122. Pokorný, P., 2005. Role of man in the development of Holocene vegetation in Central Bohemia. Preslia 77, 113e128. Poska, A., Saarse, L., 2002. Biostratigraphy and C-14 dating of a lake sediment sequence on the north-west Estonian carbonaceous plateau, interpreted in terms of human impact in the surroundings. Vegetation History and Archaeobotany 11, 191e200. Ralska-Jasiewiczowa, M., Goslar, T., Madeyska, T., Starkel, L., 1998. Lake Gosciaz, Central Poland. A Monographic Study. Szafer Institute of Botany, Kraków. Rankama, T., Vuorela, I., 1988. Between inland and coast in Metal Age Finland e human impact on the premeval forests of Southern Häme during the Iron Age. Memoranda Societatis Pro Fauna et Flora Fennica 64, 25e34. Rösch, M., 1989. Pollenprofil Breitnau-Neuhof: Zum zeitlichen Verlauf der holozänen Vegetationsentwicklung im südlichen Schwarzwald. Carolinea 47, 15e24. Rösch, M., 2000. Long-term human impact as registered in an upland pollen profile from the southern Black Forest, south-western Germany. Vegetation History and Archaeobotany 9, 205e218. Saarse, L., Poska, A., Kaup, E., Heinsalu, A., 1998. Holocene environmental events in the Viitna area, north Estonia. Proceedings of the Estonian Academy of Sciences, Geology 47, 31e44. Seppä, H., Weckström, J., 1999. Holocene vegetational and limnological changes in the Fennoscandian tree-line area as documented by pollen and diatom records from Lake Tsuolbmajavri, Finland. Ecoscience 6, 621e635. Seppä, H., Nyman, M., Korhola, A., Weckstrom, J., 2002. Changes of treelines and alpine vegetation in relation to post-glacial climate dynamics in northern Fennoscandia based on pollen and chironomid records. Journal of Quaternary Science 17, 287e301. Smith, A.G., Gobbard, I.C., 1991. A 12500 years record of vegetational history at Sluggan Bog, County Antrim, Northern Ireland (incorporating a pollen zone scheme for the non-specalist). New Phytologist 118, 167e187. Tobolski, K., 1990. Paläoökologische Untersuchungen des Siedlungsgebietes im Lednica Landschaftspark (Nordwestpolen). Offa 47, 109e131. Valsecchi, V., Finsinger, W., Tinner, W., Ammann, B., 2008. Testing the influence of climate, human impact and fire on the Holocene population expansion of Fagus sylvatica in the southern Prealps (Italy). Holocene 18, 603e614. Velle, G., Larsen, J., Eide, W., Peglar, S.M., Birks, H.J.B., 2005. Holocene environmental history and climate of Ratasjoen, a low-alpine lake in south-central Norway. Journal of Paleolimnology 33, 129e153. Voigt, R., 1996. Paläolimnologische und vegetationsgeschichtliche Untersuchungen an Sedimenten aus Fuschlsee und Chiemsee. Dissertationes Botanicae 270, 1e303. Voigt, R., 2006. Settlement history as reflection of climate change: the case study of Lake Jues (Harz Mountains, Germany). Geografiska Annaler Series A-Physical Geography 88, 97e105. Watson, C.S., 1996. The vegetational history of the northern Apennines, Italy: information from three new sequences and a review of regional vegetational change. Journal of Biogeography 23, 805e841. Wolters, S., 2002. Vegetationsgeschichtliche Untersuchungen zur spätglazialen und holozänen Landschaftsentwicklung in der Döberitzer Heide (Brandenburg). Dissertationes Botanicae 366, 1e157.