Patterns of Late Glacial vegetation in The Netherlands

: The Weichselian Late Glacial (ca. 13 ,000 -10,000 years BP) marks the transition from the cold Weichsel Late Pleniglacial to the warmer Holocene. During this period the climate rapidly changed as did the vegetation and the abiotic landscape. The vegetational development of the Weichselian Late Glacial in The Netherlands is determined firstly by the large-scale changes in climate and in the se­ cond place by local variations in lithology, geomorphology and hydrology. Pollen diagrams from different areas, em­ bracing the same time-stratigraphical interval, often show clear variations in vegetation history, which can not be ex­ plained on climatological grounds alone. In The Netherlands over 400 palynological sections, cover­ ing a part or the whole of the Weichselian Late Glacial, have been investigated by several institutes. For the compi­ lation of the data from over 250 pollen diagrams, use was made of the European Pollen Database structure. Dated shifts in the arboreal pollen content constitute the basis of a regional zonation scheme. With the help of this, iso-pol­ len maps of main taxa were constructed for different time-windows within the Weichselian Late Glacial. The dense network of palynological observation sites in The Nether­ lands permitted the drafting of high-resolution iso-pollen maps of the period considered. A clear relationship can be recognized between the iso-pollen patterns and the land­ scape type. Thus, it should be possible to distinguish more clearly between climate and other abiotic agencies of the environment which affected vegetational development,

cond place by local variations in lithology, geomorphology and hydrology.Pollen diagrams from different areas, em bracing the same time-stratigraphical interval, often show clear variations in vegetation history, which can not be ex plained on climatological grounds alone.
In The Netherlands over 400 palynological sections, cover ing a part or the whole of the Weichselian Late Glacial, have been investigated by several institutes.For the compi lation of the data from over 250 pollen diagrams, use was made of the European Pollen Database structure.Dated shifts in the arboreal pollen content constitute the basis of a regional zonation scheme.With the help of this, iso-pol len maps of main taxa were constructed for different timewindows within the Weichselian Late Glacial.The dense network of palynological observation sites in The Nether lands permitted the drafting of high-resolution iso-pollen maps of the period considered.A clear relationship can be recognized between the iso-pollen patterns and the land scape type.Thus, it should be possible to distinguish more clearly between climate and other abiotic agencies of the environment which affected vegetational development,   , 1979 andLANG, 1994).
Abb. 1: Die kontinentale Lage der Niederlande während des Weichsel Spätglazials (modifiziert nach JELGERSMA 1979und LANG 1994) However, pollen diagrams from different areas in The Netherlands show clear variations in pollen composition during the Late Glacial.

Palaeogeographical approach
For The Netherlands it is to be expected that there were only small spatial differences in climate during the Weichselian Late Glacial due to the small area and relatively large distance to the former coastline.
As sealevel was between 90 and 65 meters below the present date level (JELGERSMA, 1979), the coastline was more than 200 kilometers away and any climate gradient induced by the sea can be neglected for The Netherlands during the time under investigation.Fig ure 1 shows the relative continental position of The Netherlands during the Weichselian Late Glacial (modified after JELGERSMA, 1979 andLANG, 1994).
In the classical approach, single locations are the main basis for palaeoclimate reconstructions.It is obvious that climate parameters derived from single pollen diagrams will represent certain local influen ces.The main reason for this is the fact that not only the large scale changes in climate determined the ve getation development in for instance the Weichse lian Late Glacial.Also more local variations in lithology, geomorphology and geo-hydrological condi tions have influenced the vegetation development and patterns.Palaeoclimate reconstructions based on single pollen diagrams will therefore give a wrong picture of the regional climate.With a palaeogeographical approach the vegetation patterns and changes in the patterns can be compar ed with geological/geomorphological maps.As soon as the relations between the palaeovegetation and the abiotic components of the landscape are known, the relations between vegetation and cli mate will be more clear.This approach requires a dense network of palynological sections in an area with a well known geology and geomorphology.

The Late Glacial abiotic landscape of The Netherlands
The Late Glacial landscape is a landscape with chang ing geomorphology and vegetation.During the Weichselian Late Glacial geomorphological proces ses were active, but the abiotic changes where not as large as during the preceding Pleniglacial.Morpholo gical features related to permafrost that had existed at the end of the Pleniglacial disappeared due to the changes in climate towards the Holocene.Per mafrost disappeared, although deep seasonal frost may have occurred during the Late Glacial (VANDEN-BERGHE, 1992).The vegetation development initiated soil formation and stabilized the substratum.The Abb. 4: Regionale spätglaziale Pollendiagramme der wichtig sten Taxa für die Niederlande.
VHi.D & VAN DER SCHANS, 1961).As the ice-pushed ridges consist mainly of well-drained gravels and sands, only a few wet basins occur where organic deposits could be preserved.
3 The river region in the central Netherlands is main ly formed by the rivers Rhine and Meuse.As the ri vers changed their patterns and morphology be tween braiding and meandering due to climate change, the abandoned river channels form the basins where organic deposits could be preserved.Under dry conditions during the second phase of the Late Dryas stadial, sand was blown out of the braid ed river beds, forming riverdunes in the surrounding vegetation cover (BOHNCKE et al. 1993).The river landscape during the Late Glacial has been de scribed by KASSE et al. (1995) and BERENDSEN et al. (1995).
4 The coversand region in the eastern Netherlands is characterized by thick layers of coversand formed during the Weichselian Pleniglacial (SCHWAN, 1988).
In this region Saalian ice-pushed ridges occur also.
During the Late Dryas stadial a layer of coversand was deposited over soils and peats, indicating the landscape was more open than in the preceding Aller0d interstadial.The organic deposits in this area consist of peats and shallow lacustrine deposits formed in the depressions between coversand ridges.5 'Ehe coversand region in the southern Netherlands is characterized by a gently undulating topography.
The coversands are mainly deposited during the Weichselian Pleniglacial (SCHWAN, 1988).In the wes tern part, Early Pleistocene clayey deposits occur at shallow depth.Like in the eastern coversand region, organic deposits in this area consist of peats and shallow lacustrine deposits formed in the depres sions between coversand ridges.Some smaller pingo remnants occur (KASSE & BOHNCKE, 1992).
In figure 2a reconstruction of the landscape in The Netherlands during the Weichselian Late Glacial is given (modified after ZAGWIJN, 1986).

Available palynological data
In The Netherlands over 400 palynological sections have been investigated by several institutes during the last decades, covering part or whole of the Weichselian Late Glacial.Most of the investigated sections are unpublished and the original data are stored as counting sheets in archives of the Geologi cal Survey, Soil Survey and different universities.As a first step in this study, the pollen countings were gathered and inserted into a computer.By now 250 of these sections are available in digital format.The data are stored in a relational database, using the Eu ropean Pollen Database structure.The spatial densi ty of available pollen data decreases in westery di rection, directly related to the depth of the Late Gla cial deposits below the present day surface.In the most western part of The Netherlands the Late Gla cial deposits are covered with up to 15 meters of Holocene fluvial sediments, marine sediments and peat.Therefore, Late Glacial deposits are difficult to collect.Nevertheless, the spatial resolution is high, as can be seen in figure 3. The locations of the pol len diagrams which are stored in the database are presented as black dots, while the other locations are presented as open circles.With this dense pat tern of palynological investigated locations a recon struction of the vegetation patterns in different time windows during the Weichselian Late Glacial can be made.

Preparation of the palynological data
For the construction of the pollen diagrams from the pollen countings a uniformous pollen sum was used to calculate percentages, so the diagrams can be compared.In the pollen sum only non-thermo- philous trees, shaibs and dry herbs are included, re gional taxa sensus JANSSEN (1973).The local pollen taxa, aquatics and riparian herbs including Cyperaceae, as well as thermophilous tree pollen and spores were excluded from the pollen sum.The ma jor shifts in the main pollen taxa, radiocarbon dated in several pollen diagrams distributed over The Netherlands, are used to construct a regional zonation (HOEK, in prep.).In figure 4 the zonation is present-ed as a generalized Late Glacial pollen diagram for The Netherlands on an uncalibrated radiocarbon timescale.Based on the zonation, a zone code has been assigned to the analyzed levels from the pollen diagrams in the database.If any uncertainties ap peared, for instance in the case of a pollen sum less than 100, indications for reworking or contamina tion, no zone code was assigned to that level.In fig ure 5a-d four pollendiagrams with selected taxa are presented.The pollendiagrams are derived from small (former) lakes, pingo remnants from the north ern Netherlands till region (a), the central Nether lands ice -pushed region (b) and the eastern (c) and southern (d) coversand region.In these diagrams the differences between the percentages of Juniperus, Pinus and Ericales can be seen.The higher per centages of Pinus and Ericales at the, minerogenic, lower part of some of the diagrams are caused by reworking from older deposits.
Juniperus in zone Ic reaches the highest values up to 30% in the diagram from the eastern coversand region (c) and 15% in that from the central Netherlands (b).During the following zone 2a, the highest values for Juniperus, up to 30% are recorded in the diagrams from the central Netherlands and 10% in the southern coversand region (d).
During zone 2b Pinus has the highest value round 50% in the diagrams from the till region, the central Netherlands and the eastern coversand region.The values for Pinus in the diagram from the southern coversand region remain below 20%.The percentages of Ericales, including Empetrum ni grum, during zone 3 are the highest in the diagram from the northern Netherlands till region with values up to 25%.The diagrams from the central Nether lands and the eastern coversand region show values up to 10% and 7% respectively.For each pollen dia gram in the database the mean and maximum values of the main taxa have been computed for the distin guished zones.

Construction of the iso-pollen maps
For three zones within the Late Glacial iso-pollen maps were constructed showing the highest percen tages of the distinctive taxa in that zone.Zone Ic and the base of zone 2a are characterized by high values of Juniperus communis (Juniper), a species spread which show large scale patterns in pollen these maps can not be used for regional analyses, percentage over Europe.The spatial resolution For the construction of the iso-pollen maps in this of these maps of Europe is unavoidably low and study, the locations with their maximum value for the specific taxa within their distinct zone were re trieved from the database.Iso-pollen maps were constructed using a squared inverse distance inter polation with a search radius of 20 kilometers.The search radius is in accordance with the possible source area of the regional pollen record.In areas where no data were available within the search ra dius the outcome has automatically been blanked.Thus no extrapolations towards areas without data have been made.The data points used in the inter polation are displayed as black dots.
For zone Ic and the base of zone 2a, the time-win dow from 12,100 -11,500 BP, the highest percen tages of Juniperus are plotted in figure 6.The presence of Juniperus pollen during these zones is recorded in 82 pollen diagrams in the database.For zone 2b, the time-window from 11,250 -10,950 BP, the ltighest percentages of Pinus are plotted in figure 7. The presence of Pinus pollen during pollen zone 2b is recorded in 113 pollen diagrams in the database.For zone 3, the time-window from 10,950 -10,150 BP, the highest percentages of Ericaes (mainly Empet rum nigrum) are plotted in figure 8.The pre sence of Ericales pollen during pollen zone 3 is recorded in 114 pollen diagrams in the database.

Relationship between the iso-pollen patterns and the abiotic landscape
The highest percentages Juniperus during zone Ic and 2a with values up to 30 percent are related to the ice-pushed ridges in the central and eastern Nether lands and the southern Netherlands coversand re gion.These areas consist of the more sandy welldrained sediments.The growth oi Juniperus commu nis is favoured by a bare sandy substratum and is therefore likely to have been growing in the coversand areas.Lower percentages are recorded in the river val leys and the northern Nedterlands till region, areas with higher groundwater levels and a less sandy substratum.The highest values of Pinus during zone 2b are re corded in the central Netherlands river region.The lowest values are recorded in the eastern part of the southern Netherlands coversand region.There seems to be no south-north gradient in the percen tage of Pinus as suggested by several authors.As Pi nus sylvestris is at present growing on drier loca tions, it is supposed that Pinus was not inhabiting the river valleys but grew on the higher parts of the terraces along the river valleys.The high percen tages of Pinusin the central Netherlands river region may also be a result of a more open herbaceous vegetation type, suggesting Pinus is overrepresented clue to king distance transport.
The iso-pollen map for the maximum values of Eri cales during zone 3 (figure 8), shows the high val ues, over 20%, linked to the poorly drainage and leached soils in tills situated in the northern Nether lands.In the ice-pushed region of the central Nether lands and western part of the southern Netherlands coversand area percentages above 10% occur, pre sumably related to the occurrence of a clayey sub stratum at shallow depth.Pollen diagrams from the coversand area and the nutrient rich river area show values below 5% for Empetrum during zone 3. The high occurrence of Empetrum nigrum is often used as an indicator for oceanity, based on higher preci pitation rates.At present, a higher occurrence of Em petrum indicates a low nutrient availability or acid soils, a situation occurring already in the till region during the time under investigation.Empetrum is able to grow in areas with an active aeolian sedimen tation, a situation that occurred during the second phase of the Younger Dryas (BOHNCKE et al., 1993).

Conclusions
Not only climatic changes (temperature and precipi tation) influenced the vegetation development.Also more local variations in lithology, geomorphology and geo-hydrological conditions influenced the ve getation and especially the vegetation patterns.As the vegetation in The Netherlands, and other areas, will not have been uniformous during the Late Gla cial one has to be careful with deriving the climate signal from single pollen diagrams.As The Netherlands occupied a relative continental position during the Weichselian Late Glacial, it is not feasible that differences in the pattern of Ericales dur ing the Late Dryas stadial are caused by a climatic gradient over The Netherlands.There will, however have occurred a climatic event causing the great ex pansion of Empetrum nigrum in the areas favour able for this species.

Figure 2 .
Figure 2. Reconstruction of the landscape types in The Netherlands during the Weichselian Late Glacial (modified afterZagwijn, 1986).Abb.2: Rekonstruktion der Landschaftstypen in den Niederlanden während des Weichsel Spätglazials (modifiziert nach ZAGWIJN 1986) main landscape types that existed during the Late Glacial in The Netherlands are formed by glacial, flu vial and aeolian processes.For The Netherlands in general five larger landscape regions existed during the Late Glacial.1 The till region in the northern Netherlands was form ed as a result of the Saalian glaciation.Glacial tills form the substratum in a gently undulating land scape.In this region hundreds of Pleniglacial pingo remnants occur(DE GANS, 1981).From these pingo remnants, formed after melting of the pingos at the end of the Pleniglacial, many pollendiagrams have been obtained.2The ice-pushed region in the central Netherlands

Figure 6 :Figure 7 :Figure 8 :
Figure 6: Iso-pollen map for the maximum values of Juniperus between 12,100-11,500 BP (zone Ic/2al).Abb.6: Iso-Pollen-Kaite für Maximalwerte von Juniperus zwischen 12.100 und 11.500 BP (Zone Ic/2al).by birds and favoured by the presence of a bare ed around 11,250 BP from the south-east, presumsandy substratum.Zone 2b is characterized by high ably distributed along the river Rhine course.Zone 3 values of Pinus sylvestris (Scots pine) which migrat-is characterized by high values of Ericales, especial-