Introduction
During the Early Holocene several short climatic oscillations occurred (e.g. Wijmstra & de Vin, Reference Wijmstra, de Vin, van der Hammen and Wijmstra1971; van Geel et al., Reference van Geel, Bohncke and Dee1981; Hoek, Reference Hoek1997a; Bos et al., Reference Bos, van Geel, van der Plicht and Bohncke2007). Based on changes in the forest composition (AP) and herbaceous vegetation (NAP), the Preboreal biozone in the Netherlands is subdivided into the Friesland Phase, the Rammelbeek Phase and the Late Preboreal (e.g. Behre Reference Behre1966, Reference Behre1978; Wijmstra & de Vin, Reference Wijmstra, de Vin, van der Hammen and Wijmstra1971; van Geel et al., Reference van Geel, Bohncke and Dee1981). The Friesland Phase (Behre, Reference Behre1966) was a period of rising mean summer and winter temperatures. In NW Europe January temperatures increased from values between −25°C and −15°C in the Younger Dryas to between −5°C and 5°C in the Preboreal (Renssen & Isarin, Reference Renssen and Isarin2001). During July the changes were smaller, as the July temperatures increased in NW Europe by 3–5°C from about 10–15°C in the Younger Dryas to values of 13–17°C in the Preboreal (Renssen & Isarin, Reference Renssen and Isarin2001). The Friesland Phase in the Netherlands was characterised by a strong expansion of birch woodlands (van Geel et al., Reference van Geel, Bohncke and Dee1981; Bos et al., Reference Bos, Huisman, Kiden, Hoek and van Geel2005a, Reference Bos, van Geel, van der Plicht and Bohncke2007). The Rammelbeek Phase is regarded as a dry and continental phase, with warm summers and cold winters and a temporary expansion of steppe-like grasslands (van Geel & Kolstrup, Reference van Geel and Kolstrup1978; van Geel et al., Reference van Geel, Bohncke and Dee1981). The start of the Late Preboreal was characterised in the Netherlands by a change to a wetter climate, probably forced by a sudden decline in solar activity (van der Plicht et al., Reference van der Plicht, van Geel, Bohncke, Bos, Blaauw, Speranza, Muscheler and Björck2004; Bos et al., Reference Bos, van Geel, van der Plicht and Bohncke2007), and an expansion of birch forests. In the Netherlands, the vegetation response to these Early Holocene climate changes was regionally different and dependent on the geomorphology of the landscape (Hoek, Reference Hoek1997a,Reference Hoekb; Bos et al., Reference Bos, van Geel, van der Plicht and Bohncke2007; Hoek et al., Reference Hoek, Lammertsma, Bohncke, Bos, Bunnik, Kasse, Schokker and Westerhoff2016.). Also in the Meuse River valley these climate changes had their impact on the vegetation and river course, which will be discussed in this paper.
The Meuse River valley in the Netherlands can be subdivided into northern, middle and southern parts. The northern part of the Meuse valley, or so-called Sand Meuse (‘Zandmaas’), is located between Venlo and Nijmegen. The middle part of the Meuse valley falls between Venlo and Sittard, and the southern part is located between Sittard and Maastricht (cf. Zuidhoff & Huizer, Reference Zuidhoff and Huizer2015). Teunissen (Reference Teunissen1983, Reference Teunissen1990) performed detailed palynological research on a number of channel fills from the northern Meuse valley, in order to reconstruct the vegetation development of this area during the Late-glacial and Holocene. Furthermore, over recent decades, a large number of palynological studies and prospecting scans have been performed in this area in order to date and reconstruct the development of the river valley and its vegetation history. In some cases, these studies were also supported by radiocarbon dating (Bos, Reference Bos, Zuidhoff, van Kappel and Gerrets1992; Zuidhoff & Huizer, Reference Zuidhoff and Huizer2015; Hoek et al., Reference Hoek, Lammertsma, Bohncke, Bos, Bunnik, Kasse, Schokker and Westerhoff2016). Most recently, extensive palynological studies on residual channel fills of the Meuse River were carried out in this area in combination with archaeological and geomorphological research, such as near Lomm (e.g. Bos et al., Reference Bos, Zuidhoff, van Kappel and Gerrets2012; F.S. Zuidhoff & J.A.A. Bos, in prep.) and Well–Aijen (van Haaster, Reference van Haaster2011; Bos & Zuidhoff, Reference Bos and Zuidhoff2015). At these locations large water basins are being constructed as a passage for the Meuse River during high-water levels. This made further research in the area available. At Lomm the vegetation history during the Holocene was reconstructed based on a combination of several residual channel fills that reflect shorter time periods of the Holocene (Bos et al., Reference Bos, Bouman, Moolhuizen and Prangsma2008; Zuidhoff & Bos, Reference Zuidhoff, Bos, Gerrets and de Leeuwe2011a,Reference Zuidhoff, Bos, Gerrets and Williamsb, Reference Zuidhoff and Bos2016). The Well–Aijen sequence was analysed at a high resolution and dated by five accelerator mass spectrometry (AMS) radiocarbon dates. The continuous palynological record reflects the vegetation history from c. 10,000 BP onwards and ends during the early Roman period (Zuidhoff & Bos, Reference Zuidhoff, Bos, Bouma and Müller2014; Bos & Zuidhoff, Reference Bos and Zuidhoff2015). The Well–Aijen pollen diagram may be regarded as a standard for the Meuse River valley.
From the middle Meuse River valley only a few records are available that reflect environmental response during the Early Holocene (e.g. Bohncke & Hoek, Reference Bohncke and Hoek2007; Blokker, Reference Blokker2013; Buijs, Reference Buijs2013; Otten, Reference Otten2013). A high-resolution, accurately dated palaeoecological record, however, was missing. At Haelen, the Early Holocene is reflected in the deposits of a residual channel of the river Meuse (Bos et al., Reference Bos, van Geel, van der Plicht and Bohncke2007). On the sediments of the Haelen sequence highly detailed botanical and zoological analyses (including pollen, spores, other microfossils and macroremains) and additional geochemical analyses (loss-on-ignition) were carried out at high resolution, and an accurate chronology of the record was provided by AMS wiggle-match dating (Bos et al., Reference Bos, van Geel, van der Plicht and Bohncke2007). Furthermore, finds of archaeological remains near Haelen suggest that Mesolithic people were occupying the area from the Boreal onwards (Bats et al., Reference Bats, Crombé, Devriendt, Langohr, Mikkelsen, Ryssaert and Van de Water2010; Scheurs, Reference Scheurs2013). Moreover, archaeological evidence from other sites in the Meuse River area suggests that groups of early Mesolithic people were already present in the Meuse valley during the Preboreal (Verhart, Reference Verhart, Bailey and Spikins2008; ter Wal & Tebbens, Reference ter Wal and Tebbens2012; Bouma & Müller, Reference Bouma and Müller2014). It is thus possible that Mesolithic people also inhabited the Haelen area during the Preboreal, though no archaeological evidence has yet been found. During the Mesolithic the principal means of subsistence were hunting, fishing and gathering, all activities that had a supposedly small and local impact on the vegetation, difficult to detect in pollen diagrams. High-resolution studies (e.g. Simmons et al. Reference Simmons, Turner, Innes and Bonsall1989; Day Reference Day1993; Simmons Reference Simmons and Chambers1993; Day & Mellars Reference Day and Mellars1994; Bos & Urz Reference Bos and Urz.2003; Bos et al., Reference Bos, van Geel, Groenewoudt and Lauwerier2005b) of pollen and non-pollen palynomorphs, such as at Haelen, may provide information about forest opening phases or the occurrence of human-induced fires or fire places.
In this paper we discuss the environmental history of the middle Meuse River valley and inferred climate changes during the Early Holocene as recorded in the Haelen sequence. The pollen record reflects the regional vegetation development of the area, while other microfossils (including non-pollen palynomorphs), macroremains and geochemical data provide information about the development of local environmental conditions of the site and possible human presence.
Study area
The sediment sequence of Haelen (51°13ʹʺ55ʺN, 5°58ʹ51ʺE) was collected from a residual channel of the river Meuse near Roermond, SE Netherlands (Fig. 1). The former river channel had eroded westward far into the southern coversand region and was originally probably c. 10m deep and 80m wide. The channel became abandoned at the end of the Younger Dryas and sedimentation started under lacustrine conditions. Accumulation of gyttja was followed during the Boreal by formation of peat (Bos et al., Reference Bos, van Geel, van der Plicht and Bohncke2007). Accumulation ceased during the Atlantic (c. 6000 14C BP) but restarted around 4000 14C BP (Verbruggen et al., Reference Verbruggen, Van Der Putten and Warreyn2003). The top of the deposit dates at 1920±35 BP (AD 0–170). Later deposits are probably missing due to peat digging (Verbruggen et al., Reference Verbruggen, Van Der Putten and Warreyn2003).
Fig. 1. Location of the Haelen site in the middle Meuse River valley. The locations of the other sites discussed in the text are also shown, including the Younger Dryas meanders of the river Meuse of Hout-Baarlo, Boukoul, Kingbeekdal, Korbusch and Dukkelaar (e.g. Bohncke & Hoek, Reference Bohncke and Hoek2007; Buijs, Reference Buijs2013; Blokker, Reference Blokker2013; Otten, Reference Otten2013). The location of the Tungelroyse beek (Woelders et al., Reference Woelders, Bos, de Kort and Hoek2015) in the southern coversand region of the Netherlands is also shown.
Methods
Botanical analysis
Subsamples were taken every cm from the Haelen sediment core for high-resolution microfossil (pollen, spores, fungi, algae, etc.), macroremain (seeds, fruits, wood, mosses, etc.) and LOI analyses. Microfossil subsamples were collected using a small corer of defined volume (~280mm3) and prepared following Faegri & Iversen (Reference Faegri, Iversen, Faegri, Kaland and Krzywinski1989) and heavy liquid separation in addition. Sieving was carried out using a 215µm sieve. Lycopodium spores were added to calculate pollen concentrations. The residues were mounted in glycerine. A Leica light microscope (magnification 400× and 1000×) was used for microfossil analysis. Preservation in general was excellent. Pollen types were identified by comparison with modern reference material and identification keys of Moore et al. (Reference Moore, Webb and Collinson1991) and the Northwest European Pollen Flora, vols I–VIIII (Punt et al., Reference Punt1976–Reference Punt2003). Nomenclature follows these keys. Other microfossils were identified using van Geel (Reference van Geel1978, Reference van Geel, Smol, Birks and Last2001), van Geel et al. (Reference van Geel, Bohncke and Dee1981, Reference van Geel, Hallewas and Pals1983, Reference van Geel, Coope and van der Hammen1989, Reference van Geel, Buurman, Brinkkemper, Schelvis, Aptroot, van Reenen and Hakbijl2003), Francis (Reference Francis, Smol, Borks and Last2001), Komárek & Jankovská (Reference Komárek and Jankovská2001) and van Geel & Aptroot (Reference van Geel and Aptroot2006).
Macroremain subsamples were boiled in 5% KOH for c. 5min and washed over a 75–80µm sieve. Macroremains were picked out from the recovered fraction and stored at 4°C. For screening and selection a Leica dissection microscope (magnification 8–100×) was used. Plant macrofossil identifications were made by comparison with modern reference material and identification keys of Beijerinck (Reference Beijerinck1947), Nilsson & Hjelmqvist (Reference Nilsson and Hjelmqvist1967), Berggren (Reference Berggren1969, Reference Berggren1981), Schweingruber (Reference Schweingruber1978) and Anderberg (Reference Anderberg1994). Nomenclature follows van der Meijden (Reference van der Meijden1996).
The combined microfossil and macroremain diagram and a concentration diagram were constructed using the Tilia and Tg.view computer programs (Grimm, Reference Grimm1991–Reference Grimm2004). Combined arboreal pollen (AP) and non-arboreal pollen (NAP) totals were employed for percentage calculations. Pollen and spores of the local aquatic or mire vegetation (inclusive Cyperaceae) and redeposited palynomorphs of thermophilous taxa were excluded (cf. Janssen, Reference Janssen, Birks and West1973; Bos, Reference Bos1998). Redeposited palynomorphs of thermophilous taxa were mainly recorded in the more minerogenic lower part of the record. Zonation of the diagram was based on changes in the AP/NAP ratio and trends in the percentages of arboreal taxa that dominate Late-glacial – Early Holocene pollen records of NW Europe.
Chronology
In order to provide an accurate chronology for the Haelen record, samples of organic material reflecting atmospheric 14C concentrations, such as seeds and fruits from terrestrial plants, were selected for radiocarbon dating. From the Haelen site a series of 28 AMS 14C radiocarbon dates was available (Table 1; Bos et al., Reference Bos, van Geel, van der Plicht and Bohncke2007). Radiocarbon dates were converted into calendar years using visual wiggle-match dating (WMD) (e.g. van Geel & Mook, Reference van Geel and Mook1989; Blaauw et al., Reference Blaauw, Heuvelink, Mauquoy, van der Plicht and van Geel2003, Reference Blaauw, van Geel, Mauquoy and van der Plicht2004). A WMD-based chronology for the Haelen record was obtained from c. 11,520 to 10,530 cal BP and was published by Bos et al. (Reference Bos, van Geel, van der Plicht and Bohncke2007). An accumulation rate of 10.3yrcm−1 was calculated. Absolute ages in this paper are reported in cal BP, i.e. calibrated or calendar age relative to AD 1950.
Table 1. AMS radiocarbon dates of the Haelen record. GrA: Radiocarbon Laboratory of the Centre for Isotope Research, Groningen, the Netherlands. KIA: Leibniz Labor, Kiel, Germany.
Loss-on-ignition and tephra analysis
Loss-on-ignition (LOI) analysis was carried out to provide a measure of the organic C content of the sediments. Sediment samples of a few cm3 were dried overnight in crucibles at 105°C to remove any moisture, cooled in a desiccator and weighed. The sample plus crucible were then put into the furnace at 550°C for 4 hours for ashing of the organic matter. Once ashing was completed the crucibles were cooled in a desiccator and re-weighed to provide a measure of LOI.
From the Haelen sequence (355–395cm depth) also 1cm thick samples were checked for the presence of microtephra. Microscopic tephra layers can be used as chronological marker horizons, which may enable correlation of the Haelen record with other European records (Davies et al., Reference Davies, Brancj, Lowe and Turney2002, Reference Davies, Hoek, Bohncke, Lowe, Pyne O'Donnell and Turney2005) and the Greenland ice-core records (e.g. Dansgaard et al., Reference Dansgaard, Johnsen, Clausen, Dahl-Jensen, Gundestrup, Hammer, Hvidberg, Steffensen, Sveinbjörndottir, Jouzel and Bond1993; Johnsen et al., Reference Johnsen, Clausen, Dansgaard, Gundestrup, Hammer, Andersen, Andersen, Hvidberg, Dahl-Jensen, Steffensen, Shoji, Sveinbjörnsdöttir, White, Jouzel and Fisher1997; NGRIP members, 2004; Rasmussen et al., Reference Rasmussen, Andersen, Svenson, Steffensen, Vinther, Johnsen, Clausen, Larsen, Bigler, Röthlisberger, Andersen, Fisher, Ruth, Goto-Azuma and Hansson2006, Reference Rasmussen, Vinther, Clausen and Andersen2007, Reference Rasmussen, Abbott, Blunier, Bourne, Brook, Buchardt, Buizert, Chappellaz, Clausen, Cook, Dahl-Jensen, Davies, Guillevic, Kipfstuhl, Laepple, Seierstad, Severinghaus, Steffensen, Stowasser, Svensson, Vallelonga, Vinther, Wilhelms and Winstrup2013). However, although a few possible tephra particles were found, no distinct layers were present (S. Davies, pers. comm., 2012).
Results and interpretation
The Haelen microfossil and macroremain diagram (Fig. 2a–e) covers the late Younger Dryas to early Boreal biozones (cf. van Geel et al., Reference van Geel, Bohncke and Dee1981; Hoek, Reference Hoek1997a). A WMD chronology is given for the period 11,520–10,530 cal BP (Bos et al., Reference Bos, van Geel, van der Plicht and Bohncke2007). The zone boundaries (in cal BP) for the Haelen record are displayed in Table 2. Furthermore a concentration diagram is presented showing a selection of curves (Fig. 3). The concentration diagram demonstrates that the high concentrations of all taxa during the Younger Dryas are in strong contrast to the much lower values during the Early Holocene. This indicates that the accumulation rate during the Younger Dryas was low but strongly increased during the Early Holocene as a result of an increase in biological productivity. The regional development around Haelen in the middle Meuse valley, and the local vegetation succession in the Haelen residual channel, are discussed below.
Fig. 2. Combined microfossil and macroremain diagram of the Haelen record. Microfossils (pollen, spores etc.) are shown as curves (%), macroremains (fruits, seeds, wood, etc.) as histograms giving total amounts. Exaggeration of microfossil curves 5×. (a) Trees, shrubs and Ericales; (b) upland herbs and fire indicators; (c) reed swamp and damp woodland soil taxa, HdV-types and coprophilous taxa; (d) aquatic botanical and zoological taxa; (e) reworked taxa.
Table 2. WMD dated zone boundaries (in cal BP) in the Haelen record.
Fig. 3. Concentration diagram of the Haelen record.
Regional vegetation development
Younger Dryas
Due to the absence of datable botanical material, the start of the Haelen record could not be radiocarbon-dated. However, the pollen assemblage clearly points to a late Younger Dryas age. Around Haelen the late Younger Dryas was characterised by an open landscape with herbaceous vegetation and dwarf shrubs such as Betula nana (dwarf birch), Salix (dwarf willow) and Juniperus communis (juniper). Furthermore, taxa such as Ribes nigrum (blackcurrant) and Ephedra distachya-type (sea grape) were probably present, but records of Ephedra pollen may also be the result of long-distance transport. However, Hoek (Reference Hoek1997b) suggests that sea grape was probably locally present in the more sandy regions of the Netherlands. During this period, the number of trees was low, although macrofossils of Betula sect. Albae (tree birch) suggest that birch trees were growing in the woodland vegetation around the site. Pinus sylvestris (pine) probably remained present at sheltered locations (so-called local refugia).
In the Netherlands, especially during the later part of the Younger Dryas, Empetrum nigrum (crowberry) but also some Calluna vulgaris (heather) were important species in the open vegetation on the more acid and poor sandy soils (cf. Hoek, Reference Hoek1997a,b). Sandy soils however, were not so frequently available in the Meuse River area around Haelen. During the Younger Dryas, these were more common in the SW and NE coversand area of the Netherlands (Hoek, Reference Hoek1997b; Bos et al., Reference Bos, van Geel, van der Plicht and Bohncke2007). Grasslands and tall-herb vegetation were probably abundant on the Meuse River floodplain near Haelen and on the slopes of the nearby terraces with, apart from various Poaceae (grasses), also herbs such as Sanguisorba officinalis (great burnet), Polemonium caeruleum (Jacob's ladder), Helianthemum (rock rose), Saxifraga granulata (meadow saxifrage), Campanula (bellflower) and various Asteraceae (composites), Caryophyllaceae (carnations) and Apiaceae. Amongst the Apiaceae, various taxa were recorded, such as Anthriscus sylvestris (cow's parsley), Pimpinella major-type (greater burnet-saxifrage), Bupleurum falcatum (sickle hare's ear) and Heracleum sphondylium (hogweed). Many of these taxa prefer lime-rich soils.
On the higher terraces the vegetation was rather scarce. Here probably some shrubs of juniper and willow, Poaceae and herbaceous vegetation with Botrychium, Chenopodiaceae (goosefoots), Asteraceae and Caryophyllaceae were present (cf. Bos et al., Reference Bos, Bohncke and Janssen2006). Saxifraga spp. (saxifrages), Polygonum viviparum (alpine bistort) and Dryas octopetala (mountain avens) were probably also found on these open, often calcareous-rich soils.
On the river floodplains, birch woodlands were present in the wet to moist zones around depressions and (oxbow) lakes. In the understorey of these forests and on forest edges, Chaerophyllum temulum (rough chervil), Anthriscus sylvestris, Anemone nemorosa (wood anemone), Poaceae and some Polypodiaceae (ferns) could be found, while taxa such as Polygonum aviculare (pigweed), Rumex acetosella (sheep's sorrel) and Artemisia (mugwort) indicate the presence of trampled areas and open, disturbed grounds.
Records of ascospores of coprophilous fungi such as Sordaria-type and Apiosordaria verruculosa (HdV-55A and HdV-169; van Geel et al., Reference van Geel, Buurman, Brinkkemper, Schelvis, Aptroot, van Reenen and Hakbijl2003; van Geel & Aptoot, Reference van Geel and Aptroot2006; Fig. 2c) may suggest that large herbivores were present in the area and used the lakeshores as a watering place.
Preboreal: Friesland Phase
At Haelen, the start of the Holocene was dated around 11,520 cal BP (Table 2). This transition is characterised by abrupt warming and an increase in precipitation. Based on changes in the forest composition (AP) and herbaceous vegetation (NAP), the Preboreal biozone in the Netherlands is subdivided into the Friesland Phase, the Rammelbeek Phase and the Late Preboreal (e.g. Behre Reference Behre1966, Reference Behre1978; Wijmstra & de Vin, Reference Wijmstra, de Vin, van der Hammen and Wijmstra1971; van Geel et al., Reference van Geel, Bohncke and Dee1981). Also in the Haelen sequence this subdivision of the Preboreal is clearly recognisable (Fig. 2).
The Friesland Phase (Behre, Reference Behre1966) was a period of rising mean summer and winter temperatures (Renssen & Isarin, Reference Renssen and Isarin2001), characterised in the Netherlands by a strong expansion of woodlands with birch, i.e. mainly Betula pubescens (van Geel et al., Reference van Geel, Bohncke and Dee1981; Bos et al., Reference Bos, Huisman, Kiden, Hoek and van Geel2005a, Reference Bos, van Geel, van der Plicht and Bohncke2007). Around the Haelen site, birch trees were forming woodlands. In the river valleys of the Meuse and Scheldt (Bos et al., Reference Bos, van Geel, van der Plicht and Bohncke2007) the expansion of birch is accompanied by an increase in Populus (poplar) (Figs 2a and 3). During the Friesland Phase, woodlands with birch and poplar replaced the grasslands on the floodplains of the Meuse River and on the slopes of the river terraces. Also Sorbus, probably S. aucuparia (mountain ash), was growing in these forests. Grasses, heather, hogweed, rough chervil and cow parsley were forming the understorey or could be found along forest edges.
The open herbaceous vegetation with dwarf shrubs and Ericales that was so typical for the Younger Dryas largely disappeared during the Friesland Phase. On the nearby sandy and gravelly terraces and slopes, some open herbaceous vegetation, grasses and juniper shrubs remained present, while on sandy, nutrient-rich beaches, along the river, Plantago maritima type (likely P. arenaria, branched plantain) was probably growing.
Some pine trees could be found on the higher terraces. Pine had suffered considerably from the colder, wetter climate during the Younger Dryas, but finds of Pinus macroremains at several sites indicate that the species was still present at sheltered locations in the Netherlands during the early Preboreal (Bos et al., Reference Bos, van Geel, van der Plicht and Bohncke2007).
On the slopes of the terraces grasslands with Daucus carota (wild carrot), Centaurea scabiosa (greater knapweed), Sanguisorba minor (salad burnet), cow parsley and Ranunculus acris group (buttercup) were present. Trollius europaeus (globeflower) was probably growing in damp meadows, and in grasslands less rich in nutrients Ophioglossum vulgatum (adder's-tongue) appeared. Although Trollius is poisonous for large herbivores, some of the plants mentioned above suggest that grazing of the grasslands was taking place. Furthermore, records of Polygonum aviculare, sheep's sorrel and mugwort suggest the presence of tramped areas or disturbed soils, but no coprophilous fungi were recorded. Trampling and disturbance of soils may also have been related to the presence of Mesolithic people. During the Mesolithic, the principal means of subsistence were hunting, fishing and gathering, all activities that had a small, local impact on the surrounding vegetation (e.g. Bos & Urz, Reference Bos and Urz.2003; Bos et al., Reference Bos, van Geel, Groenewoudt and Lauwerier2005b; Woelders et al., Reference Woelders, Bos, de Kort and Hoek2015). However, the presence of Mesolithic people may also have been reflected in the occurrence of fire indicators. During the Friesland Phase, large numbers of fire indicators (i.e. charred plant tissue, ascospores of Gelasinospora sp. and G. cf. retispora, charred epidermis of Poaceae and Cyperaceae, remains of charred leaves of birch and pollen of Rumex acetosella and Epilobium angustifolium, fireweed; Fig. 2b) were found, indicating the occurrence of fires in the shore/forest vegetation and possible presence of fire places of Mesolithic people. Both Epilobium angustifolium and Rumex acetosella can be pioneers in localities where the (forest) vegetation is burned or cut down (van der Hammen, Reference van der Hammen1951; van Geel et al., Reference van Geel, Bohncke and Dee1981; www.soortenbank.nl). However, there is no direct archaeological evidence for the presence of Mesolithic people in the area during the Friesland Phase.
Preboreal: Rammelbeek Phase
The Rammelbeek Phase (Behre Reference Behre1966, Reference Behre1978; Wijmstra & de Vin, Reference Wijmstra, de Vin, van der Hammen and Wijmstra1971; van der Plicht et al., Reference van der Plicht, van Geel, Bohncke, Bos, Blaauw, Speranza, Muscheler and Björck2004) started in the Haelen sequence around 11,420 cal BP (Table 2). During this phase, the pollen diagram shows a maximum in the Poaceae percentages and a minimum in the values of birch pollen and macroremains (Figs 2 and 3). This suggests that forest expansion was temporarily interrupted and grasses dominated the regional vegetation. Also other taxa indicating open grounds or short grasslands, such as Artemisia, Rumex acetosella (also seeds were recorded), Polygonum viviparum, Helianthemum and Dianthus (pink), (re)appear or show higher values again.
On the river floodplain the birch woodlands diminished and were largely replaced by grassland vegetation. Also, on the slopes of the nearby terraces, grasslands appeared. In these grasslands many herbaceous species were present, such as Sanguisorba officinalis and S. minor, Polemonium, Helianthemum, Ranunculus, Campanula, Anthriscus sylvestris, Bupleurum falcatum, Gentianella (dwarf gentian), Plantago media (hoary plantain) and various Asteraceae and Caryophyllaceae. On the sandy and gravelly terraces open herbaceous vegetation, some juniper and Hippophae rhamnoides (sea-buckthorn) shrubs and pines remained present.
The Rammelbeek Phase was a dry, continental phase with relatively warm summers and cold winters (van Geel & Kolstrup, Reference van Geel and Kolstrup1978; van Geel et al., Reference van Geel, Bohncke and Dee1981). Given the drier and warmer climate, fires might have been expected to occur more regularly. However, only few fire indicators were recorded, far fewer than during the previous Friesland Phase. This may suggest that many of the fires during the earlier Friesland Phase were indeed related to the presence of Mesolithic people, who occupied the area only briefly and may not have left archaeological traces.
Preboreal: Late Preboreal
The Rammelbeek Phase was followed at the Haelen site around 11,270 cal BP by the Late Preboreal (Table 2). In the Netherlands the start of the Late Preboreal was characterised by a change to a wetter climate (van der Plicht et al., Reference van der Plicht, van Geel, Bohncke, Bos, Blaauw, Speranza, Muscheler and Björck2004; Bos et al., Reference Bos, van Geel, van der Plicht and Bohncke2007). At Haelen, the Late Preboreal is a phase during which birch woodlands expanded again on the river floodplains and slopes of the nearby terraces. Also, poplar became more abundant in these woodlands (Figs 2a and 3). The local presence of poplar is indicated not only by records of macrofossils (both catkin and bud scales), but also by numerous finds of ascospores of the parasitic fungus Amphisphaerella dispersella (Type HdV-310; van Geel & Aptroot, Reference van Geel and Aptroot2006). Macrofossils of birch and poplar suggest that birch and poplar swamp forests were locally present around the Haelen residual channel. Furthermore, shrubs such as Sorbus aucuparia, Rhamnus cathartica (buckthorn) and Viburnum opulus (guilder rose) and herbs like cow parsley, rough chervil, wood anemone and grasses were growing in the understorey or at the edges of these forests, while Humulus lupulus (hop) was climbing in the trees.
Pine expanded on the higher sandy and gravelly terraces around the Haelen site during the later part of the Late Preboreal, around 11,170 cal BP (Table 2; Figs 2a and 3). In the diagram pine pollen becomes dominant over birch during the Late Preboreal, but pine values remain only slightly higher than birch values. The high pollen values and presence of pine macroremains (i.e. periderm, scale leaves, seeds, needles and anthers; Fig. 2a) suggest that pine was growing close to the sampling site, probably on a higher terrace to the west. Due to the prevailing NW winds, pine pollen and macrofossils could spread easily into the Haelen residual channel.
As both pine and birch forests became more abundant in the landscape around Haelen, herbaceous communities slightly decreased. On the slopes of the river terraces, however, grass and tall-herb vegetation with Daucus carota, Centaurea scabiosa, Sanguisorba minor, Campanula, Heracleum sphondylium, Anthriscus sylvestris, Ranunculus, Plantago media and various Asteraceae remained present. At the end of the Late Preboreal some taxa such as Solanum nigrum (black nightshade), Pulsatilla vulgaris (pasqueflower) and Chenopodiaceae (re)appear. The presence of Rumex acetosella and Artemisia indicates that some disturbed grounds and dry and open sandy soils, poor in nutrients, were present. These may be related to trampling or disturbances by animals or by early Mesolithic people. Coprophilous fungi nevertheless were absent from the record, and there is no archaeological evidence for the presence of Mesolithic people in the area during the Late Preboreal. The presence of numerous fire indicators, however, may suggest otherwise. During the Late Preboreal quite a few fire indicators were recorded (Fig. 2b). In the lower part of the Late Preboreal charred plant tissue, pollen of fireweed, Gelasinospora ascospores and charred epidermis of grasses were found. In the upper part of the Late Preboreal some charred remains of pine and poplar or willow wood were present. A rather exceptional find was a charred seed of Nymphaea alba (white water lily) and charred plant tissue of Nymphaeaceae. This may suggest that natural or man-made fires or fire places occurred during the Late Preboreal. The presence of charred remains of water lily may be explained by the fact that seeds and tubers of water lily were consumed and probably roasted during the Mesolithic (Out, Reference Out, Baker, Allen, Middle and Poole2008).
Boreal
The onset of the Boreal was characterised by the immigration of Corylus (hazel) and dates around 10,710 cal BP (Table 2; Figs 2a and 3). Corylus is shortly followed by Quercus (oak) and Ulmus (elm). Later also Tilia (lime) appears in the forests around Haelen (Verbruggen et al., Reference Verbruggen, Van Der Putten and Warreyn2003). In this period, dense forests started to expand in the middle Meuse River valley with thermophilous deciduous shrubs and trees such as hazel, oak, elm and lime. Initially, hazel shrub developed in and at the edges of open forests on the river floodplain and on the slopes of the terraces. Later in the Boreal period, oak and elm spread in these forests. Sambucus nigra (black elder), Melampyrum (cow wheat) and grasses were growing in the understorey of oak forests. Birch and poplar forests were present on the wetter parts of the floodplains, and around the Haelen residual channel they probably formed a zone behind the shoreline or were partly intermixed with the shore vegetation. In the undergrowth of these swamp forests, Scirpus sylvaticus (wood club rush), Lythrum salicaria (purple loosestrife), Filipendula ulmaria (meadow sweet), Valeriana officinalis (common valerian), Urtica dioica (common nettle), Solanum dulcamara (bittersweet), Hippuris vulgaris (mare's-tail), Lycopus europaeus (gypsywort) and ferns could be found. Tilia expanded slightly later during the Boreal on the nearby higher terraces.
As the quantity of forests increased, grasslands became less abundant. However, some herbaceous vegetation remained present, with grasses such as Festuca (fescue), and with Daucus carota, Centaurea scabiosa, Sanguisorba minor, Campanula, Anthriscus sylvestris, Ranunculus and various Asteraceae. Also Rumex acetosella, Artemisia, Dianthus and Jasione (sheep's bit scabious) suggest that some open, disturbed soils remained present on the higher terraces.
Finds of charred plant tissue, charred grass epidermis, charred wood and leaf remains indicate the occurrence of (natural or man-made) fires during the early Boreal. The Haelen pollen diagram presented here ends around 10,530 cal BP. The first archaeological evidence for the presence of early Mesolithic people near Haelen dates around 10,250–10,150 cal BP (Bats et al., Reference Bats, Crombé, Devriendt, Langohr, Mikkelsen, Ryssaert and Van de Water2010). However, early Mesolithic people may have been present earlier, leaving no archaeological traces, or such traces were simply not recovered yet.
Vegetation succession in the residual channel
During the last part of the Younger Dryas, the former channel near Haelen was abandoned and an oxbow lake was formed. Aquatic plant communities developed with predominantly submerged taxa such as Utricularia (bladderwort) and Myriophyllum spicatum (spiked water milfoil), Ranunculus aquatilis-group (water crowfoot), Potamogeton (pondweed), Nitella sp. (stoneworts) and many different algae (e.g. Botryococcus braunii, Pediastrum spp., Mougeotia spp., Spirogyra, Zygnema, Scenedesmus, Tetraedon minimum and HdV-128) were abundant. The presence of Myriophyllum spicatum may suggest that the water depth in the deepest parts of the channel would have been up to 5m (cf. Hannon & Gaillard, Reference Hannon and Gaillard1997).
In the former channel also a rich water fauna was present with waterfleas (Daphnia and other Cladocera), mites (Acari), non-parasitic flatworms (Rhabdocoela), Caddis flies (Trichoptera), bryozoa (Cristatella mucedo), seed shrimps (Ostracoda), rotifers (Filinia), non-biting midges (Chironomidae) and beetles (Coleoptera). The aquatic flora and fauna suggest that the water in the residual channel was calcareous and nutrient-rich and probably was (temporarily) still flowing slowly. A light brown gyttja was deposited in the residual channel, in which the organic component must have been rather low, as LOI values of c. 22% are recorded.
In and along the shallow water of the shores a vegetation type was present with willow, Carex spp. (sedges, also indicated by Clasterosporium caricinum or HdV-126, which uses Carex as a host plant) and herbs such as Valeriana officinalis, horsetail (Equisetum), meadow-rue (Thalictrum), Filipendula, Urtica dioica, burreed (Sparganium), bulrush (Typha spp.), bogbean (Menyanthes trifoliata), crucifers (Sinapis and Hornungia-type) and Apiaceae (e.g. Oenanthe and/or Cicuta, Peucedanum). Many of these taxa (e.g. Thalictrum, Filipendula, Valeriana officinalis and Carex spp.) may also have grown in grasslands on the wetter parts of the river floodplains. On peaty patches peat moss (Sphagnum) could be found.
At the start of the Friesland Phase the LOI values show a strong increase to a maximum of almost 80%, indicating a change to more organic-rich deposits (i.e. detritus-rich grey/brown gyttja) as a result of an increase in biological productivity. The increase is followed by a major drop in the values at 387cm, where a sandy layer is present. After this drop, the LOI values return to earlier levels.
Amongst the aquatics, most-demanding taxa remain present that were present earlier, but many (especially Myriophyllum spicatum and various algae) show much lower values. New taxa such as rigid hornwort (Ceratophyllum demersum), cyanobacteria (Gloeotrichia-type) and freshwater sponges appear, while some algae such as Pediastrum boryanum var. boryanum and Tetraedon minimum show maximal values. The aquatic taxa indicate that during the early Preboreal a gradual increase in water level occurred. Higher lake water levels between 11,450 and 11,400 cal BP have also been reported from elsewhere in the Netherlands and west-central Europe (Hoek & Bohncke, Reference Hoek and Bohncke2002; Magny et al., Reference Magny, Vannière, de Beaulieu, Bégeot, Heiri, Millet, Peyron and Walter-Simonnet2007), suggesting that the increase in temperature at the start of the Holocene coincided with an increase in precipitation. The presence of the above-mentioned taxa furthermore suggests that the water in the residual channel was nutrient-rich and may have been streaming from time to time. Cyanobacteria of the Gloeotrichia-type are often among the first colonisers in a new lake, and by fixing nitrogen they can open local conditions for other aquatic plants (van Geel et al., Reference van Geel, Coope and van der Hammen1989). Their abundance, however, may also have been a reflection of the relatively warmer climatic conditions at the start of the Holocene (van Geel et al., Reference van Geel, Coope and van der Hammen1989).
Along the shores vegetation was present with less willow and sedges than during the Younger Dryas, but with more different herbs (e.g. Filipendula, Urtica dioica, Menyanthes trifoliata, Equisetum, Solanum dulcamara, Hippuris vulgaris, Lycopus europaeus and Comarum palustris (marsh cinquefoil)). Taxa such as Menyanthes trifoliata, Hippuris vulgaris and Lycopus europaeus are typical for seepage situations (Weeda et al., Reference Weeda, Westra, Westra and Westra1987, Reference Weeda, Westra, Westra and Westra1988). Their presence may suggest that seepage occurred during the Friesland Phase, which is probably related to the high terrace directly west of the Haelen residual channel.
During the Rammelbeek Phase, aquatics with floating leaves, such as white water lily, became abundant in the aquatic vegetation of the residual channel, which may suggest that the water in the residual had become stagnant. The white water lily vegetation remained present far into the Late Preboreal. Brown gyttja was deposited during the Rammelbeek Phase, which became more dark brown and peaty at the transition to the Late Preboreal. LOI values drop from c. 76% to c. 66% during the second part of the Rammelbeek Phase, which is related to a higher minerogenic content. This may have had several causes, namely a decrease in biological productivity, a lowering of the water table and/or an increase in eolian activity due to a less forested, more open landscape with soils not completely covered by vegetation. At Haelen, shifts in the aquatic and telmatic taxa suggest that a lowering of the water table occurred. However, upland surface erosion due to the more open landscape may also have caused an increased influx of minerogenic particles.
The composition of the shore vegetation remains more or less the same as during the Friesland Phase. New taxa such as Phragmites australis (reed) and Caltha palustris (marsh marigold) are occasionally recorded and suggest that these species also became a component of the shore vegetation.
From the Late Preboreal onwards, a genuine reed swamp with Thelypteris palustris (marsh ferns) and tall plants such as Phragmites australis, Sparganium, Typha angustifolia and T. latifolia formed along the shores of the residual channel. At the shores and in the swamp forests with birch and poplar also various herbs (e.g. Lysimachia (loosestrife), Solanum dulcamara, Filipendula, Thalictrum, Equisetum, Valeriana officinalis, Urtica dioica, Hydrocotyle vulgaris (marsh pennywort) and Apiaceae) were growing. Dark brown gyttja was deposited in the residual channel, which is also reflected in high LOI values of c. 80% on average. During the Late Preboreal, only a few minor oscillations are recorded in the LOI values.
During the later part of the Late Preboreal, Lemna (duckweed), Myriophyllum verticillatum (whorl-leaf watermilfoil) and Chara (stone worts) (re)appear, which suggests slightly higher water tables. It also indicates that due to seepage the water was slightly calcareous and rich in nutrients during the later part of the Late Preboreal. The (re)appearance of Scirpus sylvaticus and Lycopus europaeus suggests that seepage of groundwater probably continued during the Boreal. During the early Boreal, aquatic vegetation of species with floating leaves (i.e. Nymphaea alba and Nuphar lutea (yellow water lily) and Potamogeton) and plants with submerged leaves (e.g. Myriophyllum verticillatum and Ceratophyllum demersum) remained present in the residual channel. White water lilies were probably present at more shallow, sheltered locations near the shores, while submerged Myriophyllum verticillatum and Ceratophyllum demersum were more abundant in the deeper parts of the residual channel. The abundance of Nymphaea alba indicates that the water at the sampling site was rather shallow, probably 1–1.5m deep (cf. Hannon & Gaillard, Reference Hannon and Gaillard1997). The abundance of bryozoans (Plumatella cf. repens) may suggest an increase of shade around the residual channel caused by the surrounding forest vegetation. During this period the shores were fringed by reed swamps in which Lythrum salicaria and Alisma plantago-aquatica (common water plantain) were also growing. On peaty spots along the shores Viola palustris (marsh violet) and Sphagnum were also found.
Discussion
Comparison with other records in the middle Meuse River valley
The trends in the Haelen pollen diagram are highly comparable with those in the pollen diagram of the slightly more eastern-located Boukoul site (Buijs, Reference Buijs2013). The Boukoul pollen diagram is less detailed and lacks radiocarbon dates. Other pollen diagrams representing this period in the middle Meuse valley are present from locations slightly further south, such as Korbusch (S.J.P. Bohncke, unpub. data), Kingbeekdal (Bohncke & Hoek, Reference Bohncke and Hoek2007) and Dukkelaar (Otten, Reference Otten2013), and further north, such as Hout-Baarlo (Blokker, Reference Blokker2013) and Blerick-Koelbroek (Bos, Reference Bos1992). Furthermore, a detailed pollen diagram is present from the Tungelroyse Beek valley, a tributary of the river Meuse, located in the southern coversand region of the Netherlands (Woelders et al., Reference Woelders, Bos, de Kort and Hoek2015).
The meanders at Dukkelaar, Kingbeekdal and Korbusch in the lower part of the middle Meuse River valley were all abandoned during the Younger Dryas when the Maas shifted eastwards (Otten, Reference Otten2013) and probably all represent the same meander generation. The meanders of Haelen in the west and Boukoul in the east are situated opposite each other and may also be of the same meander generation. From the more upstream part of the middle Meuse valley only the meander near Hout-Baarlo is known as being abandoned during the Younger Dryas. All Younger Dryas records show that the middle Meuse River valley during this period was characterised by an open landscape with heliophilous herbs and dwarf shrubs, while grasslands were also rather abundant. This suggests that the vegetation development during the Younger Dryas at Haelen is representative for the middle Meuse valley. Higher grass values were also recorded in the Scheldt River area and can thus be regarded as typical for the river region during the Younger Dryas period in the Netherlands (Bos et al., Reference Bos, van Geel, van der Plicht and Bohncke2007). For example, at Tungelroy (Woelders et al., Reference Woelders, Bos, de Kort and Hoek2015), a river setting within the eastern coversand region, grass vegetation was much less abundant during this period.
The Friesland Phase can easily be recognised in all records by a strong increase in birch values. The grass-dominated Rammelbeek Phase varies and may be easily recognisable (i.e. Hout-Baarlo, Boukoul, Korbusch and Kingbeekdal), recognisable only with rather low grass values (i.e. Tungelroy) or not recognisable at all (i.e. Dukkelaar). In Hout-Baarlo the grass maximum is very high, which may be a reflection of the local vegetation around the channel. In the pollen diagrams of the northern Meuse valley (Hoek et al., Reference Hoek, Lammertsma, Bohncke, Bos, Bunnik, Kasse, Schokker and Westerhoff2016) the Rammelbeek Phase cannot always be distinguished. This may be related to the lower groundwater levels that occurred during the Rammelbeek Phase in the Netherlands, but it may also be connected with river incision during this period (Hoek, Reference Hoek1997a; Hoek & Bohncke, Reference Hoek and Bohncke2002; Bohncke & Hoek, Reference Bohncke and Hoek2007; Bos et al., Reference Bos, van Geel, van der Plicht and Bohncke2007). The start of the Rammelbeek Phase is coeval with the coldest part of the Preboreal oscillation (PBO; Björck et al., Reference Björck, Kromer, Johnsen, Bennike, Hammarlund, Lehmdahl, Possnert, Rasmussen, Wohlfarth, Hammer and Spurk1996, Reference Björck, Rundgren, Ingólfsson and Funder1997) as observed in the δ18O record of the Greenland ice-core records (e.g. Dansgaard et al., Reference Dansgaard, Johnsen, Clausen, Dahl-Jensen, Gundestrup, Hammer, Hvidberg, Steffensen, Sveinbjörndottir, Jouzel and Bond1993; Johnsen et al., Reference Johnsen, Clausen, Dansgaard, Gundestrup, Hammer, Andersen, Andersen, Hvidberg, Dahl-Jensen, Steffensen, Shoji, Sveinbjörnsdöttir, White, Jouzel and Fisher1997; NGRIP members, 2004; Rasmussen et al., Reference Rasmussen, Andersen, Svenson, Steffensen, Vinther, Johnsen, Clausen, Larsen, Bigler, Röthlisberger, Andersen, Fisher, Ruth, Goto-Azuma and Hansson2006, Reference Rasmussen, Vinther, Clausen and Andersen2007, Reference Rasmussen, Abbott, Blunier, Bourne, Brook, Buchardt, Buizert, Chappellaz, Clausen, Cook, Dahl-Jensen, Davies, Guillevic, Kipfstuhl, Laepple, Seierstad, Severinghaus, Steffensen, Stowasser, Svensson, Vallelonga, Vinther, Wilhelms and Winstrup2013; sensu Bos et al., Reference Bos, van Geel, van der Plicht and Bohncke2007).
During the Late Preboreal, birch values increase again in all records of the middle Meuse River valley. In the records downstream (Kingbeekdal, Korbusch and Dukkelaar) pine is already high or dominant from the start of the Late Preboreal onwards. In the more central and northern sites of the middle Meuse valley (Haelen, Boukoul, Hout-Baarlo and Tungelroy), pine increases and becomes dominant during the later part of the Late Preboreal. The increase in pine thus occurs earlier in the records downstream. Although pine was already locally present at sheltered locations, this may suggest that pine was using the terraces of the Meuse River valley as a (re)immigration route, similar to the river Rhine during the Late-glacial (sensu Hoek, Reference Hoek1997a; Bos, Reference Bos2001). Despite numerous macrofossils, the pine pollen values in the Haelen record, however, remain rather low and pine and birch remain subdominant. This was probably caused by an overrepresentation of birch that was growing in the local vegetation around the Haelen residual channel.
The early Boreal can be recognised in all these palynological studies by a strong increase in hazel, followed by a increase in oak, elm and later also in lime. In some records (Hout-Baarlo, Dukkelaar and Tungelroy), lime does not appear earlier than the Atlantic period.
The Haelen pollen zones can also be compared with the radiocarbon-dated Pollen Assemblage Zones (PAZ) established for the Late-glacial and Early Holocene of the Netherlands (Hoek, Reference Hoek1997a). These PAZ are also applicable on the pollen diagrams of the northern part of the Meuse River valley (=Sand Maas) (Hoek et al., Reference Hoek, Lammertsma, Bohncke, Bos, Bunnik, Kasse, Schokker and Westerhoff2016) and are based on some 35 pollen diagrams and >50 radiocarbon dates. A highly detailed and well-dated reconstruction of the vegetation and terrace formation during the Late-glacial and Early Holocene was made (Hoek et al., Reference Hoek, Lammertsma, Bohncke, Bos, Bunnik, Kasse, Schokker and Westerhoff2016). Pollen zone PAZ 4 (Betula-Poaceae PAZ) can be linked with the Friesland and Rammelbeek Phases, while PAZ 5 (Pinus PAZ) reflects the Late Preboreal. Further north of Haelen, there are also the archaeological sites of Lomm and Well–Aijen (F.S. Zuidhoff & J.A.A. Bos, in prep.; Bos & Zuidhoff, Reference Bos and Zuidhoff2015). The residual channels discussed in these archaeological studies, however, start accumulating during the early Boreal, which means that there is only a small overlap between those records and Haelen. The records of Haelen, Well–Aijen and Lomm all show a strong increase in hazel during the early Boreal, followed by an increase in oak, elm and later also lime. This suggests that hazel occurred almost simultaneously at all these sites.
Archaeology
Finds of archaeological remains indicate that people were present in the Haelen area from the Boreal onwards (Bats et al., Reference Bats, Crombé, Devriendt, Langohr, Mikkelsen, Ryssaert and Van de Water2010; Scheurs, Reference Scheurs2013). Flint artefacts and AMS 14C dates on fragments of hazelnut shells date the first habitation phase near Haelen around 10,250–10,150 cal BP (i.e. 8300–8200 cal BC; Bats et al., Reference Bats, Crombé, Devriendt, Langohr, Mikkelsen, Ryssaert and Van de Water2010). The concentration of flint artefacts was probably linked to a short-lived habitation phase of hunter-gatherers (Bats et al., Reference Bats, Crombé, Devriendt, Langohr, Mikkelsen, Ryssaert and Van de Water2010). The Haelen pollen diagram shows that the early Mesolithic people during the Boreal were living in a forested landscape. On the lower parts of the river floodplains and around former channels birch forest was present, while hazel shrubs and deciduous mixed forests expanded on the higher parts of the floodplains and edges of terraces. On the sandy Meuse River terraces, pine woodlands remained present as a relict from the Late-glacial. Along the residual channel tall reed swamps had developed, which may have restricted the view to open water with aquatic vegetation in the residual channel.
In the Haelen record there are indications that people temporarily inhabited the area during the Preboreal. During the Friesland Phase, Late Preboreal and early Boreal, a large number of fire indicators was recorded, indicating the occurrence of fires in the shore/forest vegetation and possible presence of fire places of Mesolithic people. Also herbs such as Epilobium angustifolium, Rumex acetosella, Artemisia and Polygonum aviculare may indicate the presence of trampled areas, disturbed grounds or localities where the (forest) vegetation was burned or cut down. Furthermore, the presence of a charred seed of white water lily and presence of charred plant tissue of Nymphaeaceae suggests that seeds and tubers of these plants were consumed. The tubers of white water lily were probably used as a root vegetable during the Mesolithic (Out, Reference Out2009). The charred plant tissue suggests that tubers were roasted. Roots and tubers likely formed the most abundant and readily available source of starch in the increasingly forested environment of Mesolithic temperate Europe (Kubiak-Martens, Reference Kubiak-Martens, Bescherer-Methèny and Beaudry2015). Other species that may have been consumed as starchy food during the early Mesolithic are the inner bark tissue of birch and pine and tubers of horsetail species. In the vegetation around Haelen furthermore a wide range of edible plants was present for gathering (based on Renfrew Reference Renfrew1973; Körber-Gröhne Reference Körber-Grohne1987; Oberdorfer, Reference Oberdorfer1990; Kubiak-Martens Reference Kubiak-Martens1996, Reference Kubiak-Martens1999, Reference Kubiak-Martens, Bescherer-Methèny and Beaudry2015; Out, Reference Out2009); for example, knotgrass (Polygonum), dock (Rumex), common nettle and goosefoots (Chenopodiaceae) were available as green vegetables. The latter could also have been used for their edible seeds. Tubers of wild carrot, fruits of mountain ash and rhizomes of hop, knotgrass, bulrush and ferns may also have been used as vegetables, and Filipendula, Valeriana officinalis, Solanum dulcamara, Menyanthes trifioliata, Nymphaea alba and Polygonum aviculare as medicinal plants. During the Boreal also, acorns and hazelnuts, fruits of guilder rose and black elder and seeds and tubers of yellow water lily were available.
Despite the botanical signals that early Mesolithic people may have occupied the Haelen area during the Friesland Phase, Late Preboreal and early Boreal, no archaeological evidence for their presence was found before 10,250–10,150 cal BP (see Bats et al., Reference Bats, Crombé, Devriendt, Langohr, Mikkelsen, Ryssaert and Van de Water2010). Early Mesolithic people may only have occupied the area briefly, leaving no archaeological traces, or such traces may not yet have been recovered. However, archaeological evidencefrom other sites in the Meuse and the more easterly Roer River valley (e.g. Verhart, Reference Verhart, Bailey and Spikins2008; ter Wal & Tebbens, Reference ter Wal and Tebbens2012; Bouma & Müller, Reference Bouma and Müller2014) suggests that groups of early Mesolithic hunter-gatherers-fishers were already present in this area during the Preboreal.
Conclusions
-
(1) The detailed multi-proxy, high-resolution analyses (including microfossils, macroremains, LOI measurements and AMS 14C wiggle-match dating) of the Haelen site provide valuable and detailed information about the vegetation and climatic history of the middle Meuse River valley during the Early Holocene.
-
(2) The Haelen record, covering the Late-glacial/Holocene transition and Preboreal, reflects a >1000-year period during which major climate changes occurred on a northern Atlantic scale. The Late-glacial/Early Holocene transition is characterised by a period of abrupt warming and expansion of birch forests, the Friesland Phase. This is followed by two short climatic oscillations, the dry, continental Rammelbeek Phase with open grassland vegetation and the wetter Late Preboreal during which birch forest and later pine forest expanded. These oscillations could be recognised in the detailed regional pollen diagram of the Haelen sequence.
-
(3) The Haelen record furthermore compares well with other pollen diagrams of the northern and middle Meuse River valley. All records show a strong increase in the pollen representation of birch during the Friesland Phase, but in some records the Rammelbeek Phase is not recognisable or is absent due to a hiatus. During the Late Preboreal, the pine values increase. This occurs earlier in the records of the downstream part of the middle Meuse River valley and may suggest that the terraces of the Meuse River valley formed a migration route for pine.
-
(4) The microfossil record of the Haelen site suggests that the landscape was already densely forested when Mesolithic people arrived in the area c. 10,250–10,150 cal BP. The Preboreal part of the Haelen record shows some indications that people already temporarily inhabited the area before this period. Despite the lack of archaeological evidence, there are several indications for the presence of (natural or man-made) fires in the forest and grass vegetation. Nymphaeaceae seeds and tubers may have been consumed by early Mesolithic people.
Acknowledgements
The investigation at the Haelen record was supported by the Research Council for Earth and Life Science (ALW) with financial aid from The Netherlands Organisation for Scientific Research (NWO, project No. 8811.32.002). We would like to thank Anne-marie Philip for preparing the microfossil samples and Nico de Wilde de Ligny for the LOI analyses. Cyriel Verbruggen and Nathalie van der Putten are thanked for coring the Haelen sequence and enabling the subsampling of the lower part of the Haelen sequence. Frieda Zuidhoff and Eric Drenth are thanked for useful suggestions, and Annette Botman for the construction of Figure 1. We also thank Lineke Woelders and an anonymous referee for constructive comments on an earlier version of this paper.
