Upper Pleistocene loess-palaeosol records from Northern France in the European context: Environmental background and dating of the Middle Palaeolithic

Quaternary International xxx (2015) 1e21

Contents lists available at ScienceDirect

Quaternary International journal homepage: www.elsevier.com/locate/quaint

Upper Pleistocene loess-palaeosol records from Northern France in the European context: Environmental background and dating of the Middle Palaeolithic Pierre Antoine a, *, Sylvie coutard a, b, Gilles Guerin c, Laurent Deschodt b, Emilie Goval b, d,
ment Paris b Jean-Luc Locht a, b, Cle a UMR 8591 CNRS-Universit
es Paris I and Paris XII, Laboratoire de G
eographie Physique, Environnements quaternaires et actuels, 1 Place Aristide Briand, 92195 Meudon, France b INRAP Nord-Picardie, 518 rue Saint Fuscien, 80000 Amiens, France c ^t. 12, avenue de la Terrasse, 91198 Gif-sur-Yvette, France GEOTRAC / LSCE, Ba d UMR CNRS 7194, Mus
eum National d'Histoire Naturelle, D
epartement Pr
ehistoire 1, rue Ren
e Panhard, 75013 Paris, France

a r t i c l e i n f o

a b s t r a c t

Article history: Available online xxx

In Northern France the loess cover from the Last glacial (Weichselian) is represented by a semicontinuous mantle up to 8 m in thickness in the best localities such as leeward slopes (EeNE exposures). In this large area, pedostratigraphic sequences from the last Interglacial-glacial cycle (EemianWeichselian) have been intensively studied, especially under the auspices of active rescue archaeological programmes that have provided hundreds of individual sequences from test-pits or excavations. In spite of variations in the thickness of the different stratigraphic units, driven by differences in geomorphological contexts, the pedostratigraphic sequences from the last Interglacial-glacial cycle exhibit a particularly constant pedosedimentary pattern, including well-identified pedological and periglacial marker horizons that can be followed towards the East in Belgium, in Germany and even in Central Europe. According to the newest data, it is shown that northern France loess-palaeosol sequences are well suited to record the response of Western European environments to rapid climatic changes (DansgaardeOeschger cycles). The main objective of this paper is to present a summary of the pedostratigraphic sequence from Northern France, supplemented by the data from the new reference sequence of Havrincourt and a global correlation scheme with surrounding areas. In Northern France, the synthesis of the observations carried out on ca. 100 sequences during the last 20 years allows the establishment of a highly detailed pedostratigraphic and chronostratigraphic scheme that represents a unique database for the discussion of the relations between Palaeolithic occupation and environment. In this context, a strong relationship between the intensity of human occupation and the climatic and environmental context is demonstrated. This relationship appears to be conditioned by the relative abundance of large mammal fauna, itself linked to vegetation density, as indicated by the extremely sparse biomass and the hiatus in human occupation that characterise Upper Pleniglacial loess. © 2015 Elsevier Ltd and INQUA. All rights reserved.

Keywords: Loess Palaeosols Northern France Correlations Periglacial processes Palaeolithic

1. Introduction The great European plain Loess Belt is the most extensive and continuous continental archive of the Last Glacial period in Europe

* Corresponding author. E-mail addresses: [email protected] (P. Antoine), sylvie.coutard@ inrap.fr (S. coutard), [email protected] (G. Guerin), laurent.deschodt@inrap. fr (L. Deschodt), [email protected] (E. Goval), [email protected] (J.-L. Locht), [email protected] (C. Paris).

(Frechen et al., 2003; Haase et al., 2007; Antoine et al., 2009.). Located downwind of the North Atlantic Ocean and south of the Fenno-Scandinavian ice sheet, the western part of this great periglacial plain was ideally situated to record the impact of climatic fluctuations of the North Atlantic region during the period (Fig. 1). In this area, the loess sequences from the last climatic cycle (Eemian e Weichselian) are the most important and best preserved, reaching an average thickness of 4e8 m from NorthWestern France (Brittany) to Belgium and Western Germany. They also show much more pronounced stratigraphic contrasts

http://dx.doi.org/10.1016/j.quaint.2015.11.036 1040-6182/© 2015 Elsevier Ltd and INQUA. All rights reserved.

Please cite this article in press as: Antoine, P., et al., Upper Pleistocene loess-palaeosol records from Northern France in the European context: Environmental background and dating of the Middle Palaeolithic, Quaternary International (2015), http://dx.doi.org/10.1016/ j.quaint.2015.11.036

2

P. Antoine et al. / Quaternary International xxx (2015) 1e21

Fig. 1. Location of the study area in a palaeogeographic map of Western Europe during the Last Glacial Maximum (according to Antoine et al., 2013, modified). 1) Loess (>2 m). s-Elbeuf, V.A-Villiers-Adam, Mauq: Mauquenchy, SD: Sourdon, SSL: Saint-Sauflieu, FaV: Fresnoy-au-Val, StA: Saint-Acheul, 2) loess (<2 m). 3) Sandy loess. Sites: SPE: Saint-Pierre-le RNC: Renancourt, BsO: Bettencourt-Saint-Ouen, MT: Mautort, CS: Caours, Sav: Savy, Cbl: Combles-TGV, TL: Transloy-TGV, BG: Beugn^atre, Herm: Hermies, Havr: Havrincourt, O.T: Onnaing-Toyota, Harm: Harmignies, Rom: Romont, Kess: Kesselt, Roc: Rocourt, R.M: Remicourt-Momale, W-H: Veldwezelt-Hezerwater Garz: Garzweiller, Ind: Inden, SB.R: € nchesberg, M.W: Mainz-Weisenau, K.M.: Koblenz-Metternich, Nuss: Nussloch, Ach: Achenheim. Schwalbenberg-Remagen, T.B: To

than those from continental Central Europe (most diversified soils) and locally exhibit very rich records, such as in the Rhine Valley at Nussloch (40 units/18 m, Antoine et al., 2001, 2002) or Schwalbenberg II (13 m, ±40 units/Schirmer, 2011; Frechen and Schirmer, 2011). On the other hand, even in the best contexts, these loess sequences are definitely not continuous, since they include numerous hiatuses, the duration of which can reach several thousands of years. Within this large geographical area, numerous studies over the past three decades have focused on pedostratigraphic approaches
et al., 1980, 1986; Lautridou et al., 1985; Haesaerts et al., (Somme 1981, 1999; Haesaerts and Mestdagh, 2000; Rousseau et al., 1998;
et al., 1996; Antoine, 1989, 1991; Antoine Frechen, 1999; Juvigne et al., 1998, 1999, 2001, 2003a,b,c; Schirmer, 2000a,b; Meijs, 2002, 2011; Meszner et al., 2011; Terhorst et al., 2014, 2015). Meanwhile, major methodological developments in the field of geochronology (luminescence, 14C on loess organic matter) allowed a gradual improvement in the detailed chronology of the sequences €ller et al., 1988; Zo €ller and Wagner, 1990; Van den Haute et al., (Zo
et al., 1999, 2001; Frechen et al., 2001, 2003; Lang 1998; Hatte et al., 2003; Fuchs et al., 2007, 2013; Tissoux et al., 2009; Frechen and Schirmer, 2011; Novothny et al., 2011; Schmidt et al., 2011; Kreutzer et al., 2012). However, significant error bars, inherent to the various methods of luminescence dating, and the problems of accuracy of the ages (inversions, major underestimations) do not always lead to consistent chronostratigraphic schemes. They also make the correlations with the scale of rapid (millennial) climatic events very difficult. Finally, since the mid-1990s, the identification of abrupt climatic changes in the Greenland ice cores (Dansgaard et al., 1993; NGRIP; Members, 2004; Blockley et al., 2012) has generated a major research effort focused on their record in sedimentary archives and the clarification of their impact on the European continental en
nchez Gon ~ i et al., 1999, 2002, 2008; Müller et al., vironments (Sa

2003; Desprat et al., 2007; Seelos et al., 2009; Boch et al., 2011; Heiri et al., 2014; Moreno et al., 2014). In this context, investigations based on multi-proxy analysis and continuous highresolution sampling have been developed on the loess sequences where sedimentation rates were the highest, such as in Nussloch in Germany (Antoine et al., 2001, 2009; Gocke et al., 2014). The research on the impact of rapid climate changes on loess sequences, as well as those in archaeological context (Haesaerts et al., 2003, 2009), led to the identification of the extreme sensitivity of Euro
pean periglacial environments to millennial climatic cycles (Hatte et al., 1998; Vandenberghe et al., 1998; Antoine et al., 2001, 2009; Moine et al., 2002, 2008; Rousseau et al., 2002; Haesaerts et al., 2010). On the basis of correlations between changes in loess particle-size parameters and dust content in Greenland ice cores, the existence of a strong connection between the aeolian dynamics of North Atlantic high latitudes and Western Europe, via atmospheric circulation, has been proposed (Rousseau et al., 2007; Antoine et al., 2009; Sima et al., 2009). Meanwhile, in northern France, research in rescue archaeology experienced an unprecedented expansion in connection with the implementation of major building works and the rapid development of a national research structure dedicated to rescue archaeology (INRAP). This context has supported an active research process involving specialists in Quaternary geology, palaeontology, dating and prehistoric archaeology (Locht et al., 2014a,b, 2015; Antoine and Locht, 2014; Antoine et al., 2014b). This work has resulted in the discovery of a very large number of in situ Palaeolithic sites in loess context, particularly for the Middle Palaeolithic (Locht et al., 2002, 2003, 2006, 2013, 2014a,b; Antoine
risson et al., 2014). et al., 2003b, c; Locht, 2008; Goval et al., 2014; He The results currently represent a unique database for the last climatic cycle in Europe (Locht et al., 2014a). In this context, this article aims to provide an overview on the pedo-sedimentary budget of the last climatic cycle in the North

Please cite this article in press as: Antoine, P., et al., Upper Pleistocene loess-palaeosol records from Northern France in the European context: Environmental background and dating of the Middle Palaeolithic, Quaternary International (2015), http://dx.doi.org/10.1016/ j.quaint.2015.11.036

P. Antoine et al. / Quaternary International xxx (2015) 1e21

of France, including the latest data from research carried out on the course of the future “Canal Seine Nord Europe” (Antoine et al., 2014a). It also provides an opportunity to attempt correlations with sequences from neighbouring areas including Belgium and Western Germany. This approach, which focuses on the pedo-stratigraphic or periglacial marker horizons occurring throughout western Europe, should allow the foundation of a coherent framework at the western European scale in connection with the reference records of rapid climate variability (Rasmussen et al., 2014), as already attempted earlier (Antoine et al., 2001). Finally, this canvas, continuously updated and refined on the basis of new research, provides a solid basis for the discussion of humaneenvironment relationships during the last glacial-interglacial climatic cycle on which this article focuses. 2. Regional setting The area included in this study lies mainly in the north and north-east of the Paris Basin Cretaceous area (Fig. 1). The bedrock is there mostly made up of very homogeneous Upper Cretaceous limestone (chalk), characterised by high porosity and high sensitivity to frost action and dissolution. In its southern part, chalky bedrock is overlain by Paleogene sands (Paleocene and Eocene) occasionally preserved as relict hilltops. These sands provided a significant source of aeolian sediments easily re-mobilized by wind in the areas of the Paris Basin, located in the southern and southeastern margins of northern France loess area during at least the last two glacial periods (ex. Villiers-Adam, Antoine et al., 2003a, In the area northeast of the Artois hills, towards the Escaut (Scheldt) basin, the sandy (or) clayey Eocene formations reappear as a more continuous spread and represent the local bedrock (Deschodt, 2014). In this sector, there is a rapid transition to the northeast to the silty sands area and then to the area of cover-sands underlying the great Belgian-Dutch plain (Zagwijn and Paepe, 1968), based upon the periglacial aeolian zonation established by Paepe and et
(1970). Somme Within this geographical area, the Somme basin, which has provided the focus of the bulk of recent loess research, is a small watershed the geomorphological features of which are controlled by NW-SE-aligned axial tectonic structures, including the syncline € et al., 2000). and the fault of the Somme Valley (Van Vliet-Lanoe During the ice ages, the palaeo-Somme represented one of the main tributaries of the Channel River which flowed into the Atlantic more than 1000 km to the west (Gibbard, 1988; Lericolais, 1997). For much of the Quaternary, with a sea level that could be as low as
120 m b.s.l., the vast periglacial area thus exposed was

3

occupied by huge braided river systems that represented a major source for regional loess (Lautridou, 1985; Antoine et al., 2009). The loess formations, which cover the slopes most often orientated between East and South-east (asymmetrical valleys and prevailing winds from West to North-west), can reach 6e8 m thick in the Somme basin (Fig. 2) and up to over 10e12 m in the east and south of the Pas-de-Calais district. They were mainly deposited during the second half of Saalian (Marine Isotope Stage, MIS 6) and especially during the Weichselian Upper Pleniglacial (±MIS 2) between about 30 and 17 ka (Fig. 1). The majority of Middle Palaeolithic sites have been found on the loessic slopes in combination with humic soil complexes (grey-forest soil and steppe soils) originating during the Weichselian Early-glacial (~112e70 ka). Nevertheless, several rescue archaeological operations recently allowed the identification of deposits corresponding to the most recent phases of the Middle Palaeolithic (~50e45 ka) and to the Upper Palaeolithic (~33e35 ka, Goval et al., 2014; Antoine et al., 2014a; Paris et al., 2013). In plateau contexts, the sediment budget is generally poor (4e5 m) and is characterized by dominant erosional processes with low sedimentation rates that resulted in significant gaps in the sequence, reducing their geoarchaeological interest. Nevertheless, thanks to sinkholes resulting from the dissolution forming chalky bedrock dolines, thick loess and palaeosols sequences (~5e8 m), including several glacial-interglacial cycles, have allowed the preservation of in situ Palaeolithic sites (Acheulean and Middle Palaeolithic) as at Etricourt-Manancourt
risson et al., 2014). Finally, more marginally, especially in the (He East and South-east of the basin, a few sites have been discovered in association with residual sand mounds (Thanetian). These sites, generally preserved on the surface or within aeolian sands remobilized from the directly surrounding Paleogene formations, may locally contain well-preserved large mammal remains, such as at Beauvais (Locht et al., 1995). 3. General methodology Despite the variations in thickness of the different stratigraphic units, related to differences in geomorphological context and location of profiles with respect to local sources of material likely to be mobilized by wind (mainly sands), northern France loess sequences present a particularly constant pedosedimentary budget including numerous pedological, sedimentary and periglacial levelmarks, that can be followed for long distances (100 km). In this context, and on the basis of the study and integration of a hundred of individual profiles, it has proved possible throughout the last 20 years to build a coherent pedostratigraphic scheme for the whole area (Antoine et al., 1999, 2003b, 2014a,b).

Fig. 2. Idealised cross-section through an asymmetric valley in northern France (topography based on the area of Saint-Sauflieu, Somme). 1) Chalk bedrock with dissolution horizon at the top (clay with flints). 2) Calcareous loess (Late Saalian and Weichselian UPG). 3) Bt horizon of brown leached soil (Eemian and Holocene Interglacials). 4) Humic soil horizons and soil complexes (Weichselian Early-glacial). 5) Boreal brown soils (Weichselian MPG). 6) Colluvial deposits (Late Holocene).

Please cite this article in press as: Antoine, P., et al., Upper Pleistocene loess-palaeosol records from Northern France in the European context: Environmental background and dating of the Middle Palaeolithic, Quaternary International (2015), http://dx.doi.org/10.1016/ j.quaint.2015.11.036

4

P. Antoine et al. / Quaternary International xxx (2015) 1e21

This scheme was used for the chrono-climatic allocation of various associated Palaeolithic sites (Locht et al., 2014a,b). In northern France loess area, some major sites have allowed the study of long transects of over 50e100 m long (Fig. 3A and B). Within this article, the pedostratigraphic sequence of the Last climatic cycle of northern France is presented as a synthesis of all available data that incorporates the latest results published on the subject (Fig. 4). This presentation is structured on the basis of a chrono-climatic
, 1970), relying on the definition subdivision (Paepe and et Somme of large and globally homogeneous periods from the perspective of the type of palaeoenvironmental and pedosedimentary response Interglacial/Early-glacial (EGL)/Lower Pleniglacial (LPG), Middle Pleniglacial (MPG), Upper Pleniglacial (UPG), Lateglacial (LGL). Having been correlated for a long time with the limits of the marine isotope stages (Lisiecki and Raymo, 2005), the time limits between these major phases appeared increasingly shifted, particularly regarding the MPG-UPG boundary and that of MIS 3 and 2 (Antoine et al., 1999, 2001). In order to approach the best the complexity and richness of the loess record, the definition of these limits has been later based, in a more systematic way, on the chronology of GRIP and GISP II ice cores (Antoine et al., 2001). More recently, the development of very high resolution chronologies for the reference curves of Greenland ice cores (Rasmussen et al., 2014)

has allowed the suggestion of the more accurate limits that are those used in this article. Finally, for the termination of the Eemian Interglacial the authors have adopted the limit based on the appearance of the first major cold event centred on 110 ka in the oceanic records (MIS 5.4/ C24, Martrat et al., 2007), contemporary to Melisey I of the Grande Pile (Woillard, 1978, 1979; Rioual et al., 2001) and to the beginning of Stadial GS 25 in the Greenland d18O records (110.6 ka). Limits of large chrono-climatic phases applied throughout this paper according to the references cited above: Eemian interglacial: ± MIS 5.5/128-110.3 ka. Early-glacial (EGL): ± MIS 5.4 to 5.1/GS 25 to GI 19.2/110.3 to 70.4 ka. Lower Pleniglacial (LPG): ± MIS 4/GS 19.1 to GS 18/70.4 to 59.4 ka. Middle Pleniglacial (MPG): ± MIS 3/GI 17.2 to GI 7/59.4 to 34.8 ka. Upper Pleniglacial (UPG): ± end of MIS3 and MIS 2/GS 7 to GS 2.1.a/34.8 to 14.7 ka. However, the positioning of these limits, and especially that separating MPG and UPG, is still under discussion and ultimately depends on the resolution of the studied pedosedimentary records and prioritization of the various episodes of loess sedimentation or soil formation (see 3.2.4).

Fig. 3. A e Fresnoy-au-Val (Somme): cross section of the Upper Pleistocene sequence along the slope (according to Deschodt in Locht (Dir.) 2008, modified). 1) Top soil (Bt horizon), 2) Homogeneous calcareous loess, 3) Laminated geliflucted silty deposits, 4) Boreal brown soil complex, 5) Calcareous loess, 6) Thick body of laminated colluvial deposits, 7) Brownish calcareous loess, 8) Isohumic horizons/steppe soil complex, 9) Saint-Sauflieu-1 grey forest soil, 10) Bettencourt-Saint-Ouen grey forest Soil, 11) Bt horizon of brown leached soil (Interglacial), 12) Chalk substratum with dissolution at the top (clay with flints). B e Villiers-Adam (Val d’Oise): cross-section (According to Antoine et al., 2003a). 1) Top soil (Bt horizon), 2a-2d) Calcareous loess, 3 to 8) Villiers-Adam Soil Complex, 3) Arctic brown soil, 4) Sandy loess, 5) Tundra gley, 6) Arctic meadow (lightly humic) soil, 7) Arctic brown soil, 8) Boreal brown soil, 9) Calcareous sandy silt, TN1) Heterogeneous laminated sandy silts, TN2) Humic horizon (steppe soil), TN3) Lightly humic horizon with bioturbations, 10) Humic horizon (steppe soil), 13-TN4) Grey forest soil (SS-1 Soil), TN5-6) Lower grey forest soil (BSO Soil), 14) Irregular whitish silt horizon, 15) Bt horizon of brown leached soil (Last Interglacial), 16) Laminated calcareous loess with frost cracks (Upper Saalian).

Please cite this article in press as: Antoine, P., et al., Upper Pleistocene loess-palaeosol records from Northern France in the European context: Environmental background and dating of the Middle Palaeolithic, Quaternary International (2015), http://dx.doi.org/10.1016/ j.quaint.2015.11.036

P. Antoine et al. / Quaternary International xxx (2015) 1e21

Please cite this article in press as: Antoine, P., et al., Upper Pleistocene loess-palaeosol records from Northern France in the European context: Environmental background and dating of the Middle Palaeolithic, Quaternary International (2015), http://dx.doi.org/10.1016/ j.quaint.2015.11.036

5

Fig. 4. Summary pedo-lithostratigraphic sequence for northern France, chronostratigraphy, dating and correlation with global climatic records. MIS according to Lisiecki, L.E., and Raymo, M.E., 2005. NGRIP data: according to ^tre; Herm.: Hermies; Havr.: Havrincourt; V.A: Villiers-Adam; BsO: Bettencourt-Saint-Ouen; SSL: SaintRasmussen et al., 2014. Marine Cold Stages (CS): according to Martrat et al., 2007. FaV: Fresnoy au Val; SD: Sourdon; BG: Beugna Sauflieu. Units: 1: surface soil (a: Ap. horizon; b: Bt horizon; c: banded Bt horizon); 2: homogeneous calcareous loess; 3: greyish and lightly humic tundra gley with tongued horizon (Nagelbeek Hz.); 4: carbonated laminated loess with cryo-dessication micro-cracks; 5: cryoturbated tundra gley doublet with intermediary loess (5b); 6: homogeneous calcareous loess; 7: cryoturbated tundra gley Hz. and large ice-wedges network (F-4); 8: arctic brown soil/or soil complex 9: sandy silts or calcareous loess; 10: greyish tundra gley horizon; 11: humic horizon/arctic meadow type soil; 12: brown Boreal soil complex; 13: heterogeneous bedded slope deposits and sandy silts (thermokarst gully infilling); 14: homogeneous calcareous loess; 15: thin tundra gley doublet (15a-c) including a calcareous loess unit (15b); 16: homogeneous brownish non calcareous colluvial silts/incipient soil (arctic meadow Hz.); 17: laminated colluvial deposits reworking soil lenses and soil nodules with frost cracks; 18: brownish-greyish non-calcareous loess; 19 to 22: steppe-like soils with interstratified unit of local, non-carbonated loess (MS/20); 23b: grey forest soil on colluviums; 23a: bleached horizon; 24b: clayey colluviums/grey forest soil (Bettencourt Soil); 24a: bleached horizon 25b: Bt horizon of brown leached soil (Rocourt/Elbeuf 1); 25a: bleached horizon; 26: Saalian calcareous loess; Symbols: A: large ice-wedges, F1 to F5 (main network: F-4); B: ice wedges (F-6 only); C: little frost cracks D: ice-melt channels (thermokarst); E: deep seasonal frost events; F: dissolution affecting the chalky geological substratum (pipes); G: calcareous part of the record; H: large incision features and thermokarst erosion gullies; I: archaeological level with TL dating from heated flints (Locht et al., 2014): 1) Villiers-Adam (110 ± 11) 2) Fresnoy-au-Val (106.8 ± 7.5) 3) Saint-Hilaire-sur-Helpe (98.9 ± 9.3) 4) Mauquenchy Low. (83 ± 7.6), 5); Mauquenchy Upp. (77.0 ± 7.2) 6) Beauvais (55.6 ± 4) J: ESR-U/Th on teeth and bones A) Savy (51.1 ± 3) U/Th/bone B) Savy (30 ± 2) ESR-U/Th (Locht et al., 2006).

6

P. Antoine et al. / Quaternary International xxx (2015) 1e21

Fig. 5. (A): type section of the Weichselian Early-glacial at Saint-Sauflieu (Somme): RS: Rocourt Soil (Interglacial), SS1: Saint-Sauflieu 1 soil (Grey forest soil), Whz: whitish horizon, SS2: Saint-Sauflieu 2 soil (steppe soil), SS3a: Saint-Sauflieu 3a soil (partly reworked steppe soil), MS: marker silt (non calcareous), SS3b: Saint-Sauflieu 3b soil (steppe soil). (B) e Detailed view of the whitish horizon (Whz.3) at the transition between the SS-1 grey forest soil and the SS2 steppe soil at Hermies (North). (C) e Hermies (North): top of the soil complex (Steppe soil) covered by thick laminated colluvial deposits (HCD) then by the brownish homogeneous horizon (HBS for Havrincourt brownish horizon). (D) e Hermies (North): detailed view of the laminated colluvial deposits (HCD) including reworked lenses and nodules from Early-glacial humic soils and showing periglacial deformations (cracks, faulting). (E) e Havrincourt (Northern France): general view of the humic infilling (colluvial and grey forest soil SS1) of the deep dissolution feature (sinkhole) of profile P9 (Antoine et al., 2014a,b). (F) e Havrincourt (Northern France): profile P5 section throughout units 4 to 8 (see Fig. 7): showing the large ice-wedge casts preserved at the boundary between the MPG brown soil complex and the UPG calcareous loess. (G) e Havrincourt (Northern France): aerial view of the main ice-wedge network F-4 (© Photo: D. Gliksman, INRAP).

Please cite this article in press as: Antoine, P., et al., Upper Pleistocene loess-palaeosol records from Northern France in the European context: Environmental background and dating of the Middle Palaeolithic, Quaternary International (2015), http://dx.doi.org/10.1016/ j.quaint.2015.11.036

P. Antoine et al. / Quaternary International xxx (2015) 1e21

4. The last climatic cycle (Eemian-Weichselian) in the North of France: pedostratigraphic and chrono-climatic summary; correlation with neighbouring areas The global litho-pedostratigraphic scheme illustrated by Fig. 4 is based on the combination of all the individual sequences spanning the last interglacial-glacial climatic cycle and integrating especially the reference sites of Sourdon, Saint-Sauflieu, Fresnoy-au-Val, Bettencourt-Saint-Ouen, Villiers-Adam, Hermies and Havrincourt (Fig. 1). The chronological control is provided by thermoluminescence (TL)/infra red stimulated luminescence (IRSL) on the polymineralic fraction (Engelmann et al., 1999; Antoine et al., 2003b), quartz optically stimulated luminescence (OSL) (Guerin in Antoine et al., 2014a), TL from archaeologically heated flints (Debenham, in Locht et al., 2013 and 2014a) and radiocarbon on bones and charcoals for the archaeological levels of the Upper Palaeolithic (Goval et al., 2014). 4.1. Last Interglacial (Eemian) (~128e110.3 ka) The Eemian Stage interglacial is represented in all the profiles of the studied area by a brown to red-brown clayey horizon of brown leached soil (Luvisol) of about 1 m thick, exhibiting a prismatic to polyhedral structure (Fig. 4, unit 25b, Fig. 5A, RS). This horizon, that
” of ancient authors, is especially corresponds to the “Limon fendille well preserved in plateau sites, such as in the former Sourdon brickyard (Antoine, 1990). It is generally characterized by a much larger percentage of clay (28e30%) compared to the carbonated loess in which it is developed (12e15%). This soil, later affected on 1e1.5 m deep by freeze-thaw alternations during the Weichselian Early-glacial, shows a strong blocky to lamellar structure resulting from the segregation of ice lenses induced by post-depositional seasonal freeze-thaw processes (0.5e1 cm in thickness). In thin sections, that poly-phased soil horizon shows dark brown microlaminated clayey-humic coatings, associated with at least one of

7

the Early-glacial pedogenesis that appear clearly subsequent to initial yellow orange clay coatings typical of interglacial pedogenesis (Locht et al., 2002; Coutard in; Antoine et al., 2003b, Antoine in; Antoine et al., 2014a). Given its position in the stratigraphy and its soil characteristics, this complex Bt horizon is correlated with the interglacial soils identified in Normandy (Elbeuf I soil: Lautridou, 1985), in Belgium (Rocourt soil: Gullentops, 1954 or Harmignies Soil, Haesaerts and €, 1981) and in NW Germany (Erbach Soil: Van Vliet-Lanoe € nhals et al., 1964; Schirmer, this issue). It must be emphasized Scho that no A horizon has so far been reported at the top of this interglacial soil. This is probably a consequence of the intensity of erosional processes that characterize the first episode of climatic deterioration from the beginning of the Early-glacial. From an archaeological point of view, this observation allows understanding of the apparent lack of in situ occupations attributable to the Eemian interglacial in the loess series. The Neanderthal occupations, the presence of which during the Eemian was demonstrated through the tufa sequence of Caours in the Somme basin (Antoine et al., 2006), have therefore been systematically eroded and are only represented as reworked artefacts in the first colluvial deposits of the Early-glacial (Antoine et al., 2003c; Locht et al., 2014a).

4.2. Early-glacial (110.3-70.4) In sites where the Early-glacial soil complex is the best preserved, as at Villiers-Adam, a first whitish silty horizon (Antoine et al., 2003a), combined with glosses of 20e30 cm deep (ancient roots) with greyish inner part and oxidized margins and with small frost cracks, marks the upper boundary of the Interglacial Bt and the transition with Early-glacial (Fig. 4: 25a). This unit, deposited after the truncation of the Bt horizon of the Eemian, is attributed to the first climatic deterioration subsequent to the Interglacial (MIS

Fig. 6. Correlation of the main Early-glacial soil complexes from northern France. 1) Clay with flints, 2) Calcareous loess, 3) Decalcified loess, 4) Banded Bt horizon of the Eemian Brown leached soil (Doublet horizon), 5) Main Bt horizon of the Eemian brown leached soil, 6) Whitish silty glossic horizons, 7) Lower grey forest soil (Bettencourt Soil/BSO), 8) Upper grey forest soil (Saint-Sauflieu 1/SS1), 9) Isohumic steppe soil horizons (SS2 and SS3), 10) reworked lenses of Bt horizon, 11) Lateral passage from colluvial humic material to a humic horizon, 12-Marker loess (non-calcareous aeolian silts).

Please cite this article in press as: Antoine, P., et al., Upper Pleistocene loess-palaeosol records from Northern France in the European context: Environmental background and dating of the Middle Palaeolithic, Quaternary International (2015), http://dx.doi.org/10.1016/ j.quaint.2015.11.036

8

P. Antoine et al. / Quaternary International xxx (2015) 1e21

5.4/Melisey I, C24 Marine Cold Stage, Martrat et al., 2007), GS 25 around 109e110 ka. The pedosedimentary budget of the Early-glacial is then characterized by the formation of a mainly humic soil complex well preserved in downslope morphologies (1.5e2.5 m thick, Figs. 5A and 6). It represents the result of a long and complex period (±40 ka), including up to 6 Interstadials in the NGRIP record (Fig. 4). This soil complex is determined from the superposition of 1) a soil horizon showing an intermediate facies between an Interglacial Luvisol and a grey-forest soil (BSO/Bettencourt Soil: Fig. 4, n 24b), 2) a typical grey-forest soil (SS1, base of the Saint-Sauflieu Soil Complex: Fig. 4, n 23b), and 3) a sequence of two (rarely three) steppe soil horizons (SS2 to SS3a,b of the Saint-Sauflieu Soil Complex, Antoine, 1989; Antoine et al., 1994, Fig. 4, n 22b to 19). The formation of this Early glacial soil complex reflects a stepwise evolution towards more and more continental environmental conditions, depicted by the succession of the two following phases: 1) development of grey-forest soils horizons on colluvial deposits in a context of a boreal forest with pine and birch and 2) formation of isohumic horizons in a steppe environment with rare birch (Munaut in Antoine et al., 1994). This pedosedimentary development is contemporary with a significant sea level fall (~
20 m) and an accompanying major change in the palaeogeography of the Channel and North Sea basins (definitive disappearance of the
et al., 1994). A detailed correlation oceanic influence, Somme scheme between the main Early-glacial soil complexes from northern France is proposed in Fig. 6.

Given the characteristics of the soil complex, the Early-glacial soil sequence was divided into two major periods of unequal length (Fig. 4): one grey forest soil phase (EGL A: ~ MIS 5.4-5.1/GI 24-21) and one steppe soils phase (EGL B: ~ MIS 5.1/GI 20 to 19): 4.2.1. Grey forest soils phase/early-glacial A (~108.2e77.7 ka) After the truncation of the interglacial soil around 112e110 ka, a first period of colluvial deposition began during the first cold stage directly subsequent to the Eemian (C24/Melisey I cold stage) around 110 ka. These colluvial deposits are only preserved at the base of the slopes or in localized sediment traps resulting from the dissolution of the chalky bedrock (sinkholes) as in Havrincourt or Bettencourt-Saint-Ouen (Figs. 5E and 6). 4.2.1.1. Bettencourt Soil (BSO). These first clayey colluvial deposits are then affected by a major pedogenesis (Bettencourt Soil), which exhibits an intermediate facies between that of an Interglacial brown leached soil (Bt) and that of a grey forest soil (Bth; Fig. 4, n 24b). The characteristics of this horizon, showing highly stratified clayey to silty clayey illuviations, slightly humic and many traces of bioturbation and earthworm hibernation chambers (1e1.5 cm), reflect a continental temperate climate with strong seasonal contrasts, similar to that existing today in the Greyzems zone of Central Europe. It seems to remain stable throughout the duration of the long (~20 ka) interstadial period corresponding to MIS 5.3, St. Germain I of Grande Pile and GI 22e24 of Greenland (Fig. 4).

Fig. 7. Correlation between summarized Eemian/Weichselian-Early-glacial soil complexes from Northern France (according to Antoine et al., 2013 and 2014a), Belgium (According to Haesaerts et al., 2011) and Western Germany (Lower Rhine: according to Schirmer, 2010, and Schirmer this issue). 1) Truncated Bt horizon of brown leached soil. 2) Bt to Bth horizon: intergrade between brown leached and grey forest soils. 3) Typical Bth horizon of grey forest soil. 4) Whitish bleached horizon. 5) Ah horizon of Steppe soil. 6) Weakly developed (incipient) humic horizon. 7) More or less well developed brownish soil/Cambisol horizons (Bw). 8) Non calcareous aeolian silts. 9) Tundra gley horizons (gleysols and gelic gleysols). 10) Main erosion/channels (base of laminated colluvial deposits).

Please cite this article in press as: Antoine, P., et al., Upper Pleistocene loess-palaeosol records from Northern France in the European context: Environmental background and dating of the Middle Palaeolithic, Quaternary International (2015), http://dx.doi.org/10.1016/ j.quaint.2015.11.036

P. Antoine et al. / Quaternary International xxx (2015) 1e21

TL dates of 106.8 ± 7.5 and 110 ± 11 ka on heated flints, obtained respectively at Fresnoy-au-Val and Villiers-Adam in archaeological levels preserved in situ in colluvial deposits in which is developed the BSO soil and the age of 96.3 (TL ADD) from the BSO Soil at Bettencourt-Saint-Ouen are consistent with this interpretation (Figs. 4 and 6). However, the great variability of these already old dating results and the magnitude of the associated error bars (6e10 ka) do not allow us to use them as really reliable chronological milestones. In general, the first horizon of the Early-glacial soil complex appears very rarely preserved outside the Somme basin. In Belgium it was recently described in the Romont section (VSG-A horizon of Haesaerts et al., 2011) where the distinction from the interglacial Bt
horizon is however not as clear as in the North of France (Juvigne et al., 2008; Antoine, personal observation). It may also correspond to the Bt horizon described from the middle part of the soil complex of the Veldwezelt-Hezerwater sequence (Meijs, 2002, 2011) and likely to the Pesch Soil of Schirmer (2000 and this issue) (Fig. 7). In general, it is likely that the scarcity of the Bettencourt Soil is only apparent and that in fact it results from the difficulty of distinguishing this horizon from the interglacial Bt in field. This is especially difficult in plateau contexts where it forms a polygenetic horizon with the Bt of the Interglacial soil. The top of this BSO soil horizon is then affected by an important phase of waterlogging indicated by irregular greyish glosses from 20 to 50 cm deep with oxidized edges particularly well preserved at Bettencourt-Saint-Ouen and Villiers-Adam (Figs. 3B and 4, No. 24a). This episode is also marked by a deep seasonal frost indicated by well developed polyhedral to lamellar structure (1 cm) developed on approximately 1 m deep and the formation of a second bleached horizon of 5e15 cm thick (Fig. 7, Whz. 2) infilling the axis of the glosses (washed white silts related to important episodes of seasonal snowmelt). On slopes, a gravel bed (small flints) reflecting significant erosion is associated with this horizon. All the features of this level reflect a marked climatic degradation that, given its position in the sequence, could be correlated with the stage GS 22, centred on 86 ka, MIS 5.2 and cold event C22 (Martrat et al., 2007) in deep-sea cores (Fig. 4, 23b). Continued climatic degradation during this episode, a little over 2 ka, is represented by the deposition of a new generation of colluvial silts reworking the underlying soils (soil nodules, fragmented clay coatings). 4.2.1.2. Saint-Sauflieu 1 Soil (SS1). Renewed pedogenesis of the grey forest soil type then gradually affects the previous colluvial deposits reflecting a new period of relative climatic amelioration (interstadial). This soil (SS1), much more humic (Fig. 4, n 23b, TOC: ~0.4%), reflects an increase in soil dynamics. The cumulative nature of that soil formation process is suggested by: 1) the ability of such horizons to develop without significant differentiation for very large thicknesses (1.5 m) thanks to localised sediment traps, such as at Havrincourt (Figs. 5E and 7) and 2), its ability to integrate several in situ Palaeolithic levels as at Bettencourt-Saint-Ouen (Fig. 6). Moreover, this later example indicates the alternation of phases of stabilization of the surface followed by phases of accelerations of the colluvial sedimentation during the formation of SS1 soil (Fig. 6). In the Bettencourt and Villiers-Adam sequences, a differentiation into sub-horizons, based on colour and texture, was carried out in the field indicating variations in the ratio between sedimentation and pedogenesis during this period. In the type sequence of Saint-Sauflieu (Somme), which is one of the few to have yielded pollen spectra, this horizon (SS1) is characterized by a boreal forest environment (80% of tree pollen) dominated by Pinus and Betula (Munaut, in Antoine et al., 1994) and a cool temperate continental climate with strong seasonal contrasts (very marked impact of freeze-thaw processes and spring

9

snowmelt). The micromorphological study of these horizons demonstrated abundant micro-laminated brown-red to blackish clayey-silty-humic coatings, in situ in the porosity and biogalleries. These coatings are often slightly fractured by frost action. These characteristics of the infilling of the biogalleries are interpreted as resulting from brutal drainage processes affecting soil surface during spring-melt episodes in the context of a consistent snow cover (Antoine et al., 1994). According to the most recent dates obtained from Mauquenchy by TL on heated flint (83 ± 7.6 at the base and 77.0 ± 7.2 at the top; Locht et al., 2013), the grey forest SS1 soil would have been formed during MI Substage 5.1 and Interstadial GI 21 between 85 and 77.8 ka (Fig. 4). Besides, the average of the 4 TL-IRSL ages (2 Regen and 2 ADD) of about 87 ka obtained at Saint-Sauflieu at the base of SS1 Soil (Frechen et al., 1995 unpublished; Frechen et al., 2001) and the ages of the upper part of the same soil horizon at Villiers-Adam (IRSL 73.1 ± 6.8 Antoine et al., 2003a) and at Bettencourt-Saint-Ouen (2 TL: 77.8 ± 10 ka and 2 IRSL: 79.20 ± 9.9 ka, Antoine et al., in Locht et al., 2002), also support the allocation of development of this upbuilding soil to MIS 5.1. It must be stressed that SS1 grey forest soil is the most constant horizon in the Early-glacial records in northern France (Fig. 6). In Belgium it also occurs with the same characteristics at Romont and probably at Remicourt-Momalle (Haesaerts and Mestdagh, 2000). Although relatively less well recorded, it is also found in the Rhine Valley at Nussloch (Antoine et al., 2001), as well as in the profiles of the large Garzweiler open pit where it has been named Holz Soil (Schirmer, 2010) (Fig. 7). This horizon of grey forest soil SS1 is then affected by a new and more intense episode of seasonal frost (
1.5 m) which is indicated by the development of a thick polyhedral to lamellar structure (1e1.5 cm thick) underlined by abundant white silts coatings intersecting all previous soil features and infilling the biogalleries. On its surface, a bleached horizon of a few centimetres in thickness occurs incorporating reworked soil nodules from the underlying Bth horizon (Fig 5B. WHz, Fig. 4: no. 23a). On slopes, such as in Saint-Sauflieu, a small erosive flint gravel bed is associated with this level. The climatic deterioration associated to this episode is logically attributable to the short cold period (±1.4 ka), which marks the end of MI Substage 5.1 at about 77 ka (GS 21/76.4 to 77.6 ka) and to C21 event of deep-sea cores. In Belgium this horizon termed Whitish horizon of Momalle (Haesaerts et al., 1999; Haesaerts and Mestdagh, 2000) is always present at the top of the grey forest soil underlying the Humiferous Complex of Remicourt at Rocourt, Romont and Remicourt (Haesaerts et al., 1999, 2011), as well as Veldwezelt-Hezerwater (Meijs, 2002) and in the Lower Rhine area (Schirmer this issue) (Fig. 7). Although less well preserved, it has also been described in Germany in the Middle Rhine area at Nussloch (Antoine et al., 2001). Finally, based on IRSL dates, it has been suggested that in two €nchesberg and Koblenz-Metternich) Bt sites in the Rhine valley (To horizons of brown leached soils were formed during the Early€ ller et al., 1991; Boenigk et al., glacial between 70 and 110 ka (Zo 1994; Frechen, 1994; Frechen et al., 1995; Boenigk and Frechen, 2001). This interpretation, seems, however, incompatible with the characteristics of the pedosedimentary records from the other West European sites where the formation of typical Bt horizons of brown leached soil is only recorded for the Eemian Interglacial. Furthermore, it has been shown that the luminescence ages obtained especially by IRSL method on polymineral silt fraction, were subject to important underestimation resulting from anomalous fading of feldspars grains (Schmidt et al., 2011). With the exception of these two sites, the part of the sequence corresponding to grey forest soils (EGL A) may be subject to accurate correlations, at the scale of individual soil horizons, from

Please cite this article in press as: Antoine, P., et al., Upper Pleistocene loess-palaeosol records from Northern France in the European context: Environmental background and dating of the Middle Palaeolithic, Quaternary International (2015), http://dx.doi.org/10.1016/ j.quaint.2015.11.036

10

P. Antoine et al. / Quaternary International xxx (2015) 1e21

northern France to the Rhine Valley via Belgium (Fig. 7). Further to the east the comparison is complicated either by the lack of recording of the initial part of the soil complex formed on colluvial deposits (superimposition of grey forest soil horizons and interglacial Bt), as it appears to be the case in Poland (Jary and Ciszek, 2013; Antoine personal observation), or by the change to more continental facies such as chernozems occurring already during the stonice in the early stages of the Early-glacial such as at Dolní Ve Czech Republic (Kukla, 1977; Antoine et al., 2013). Besides it should be noted that in northern France, the grey forest soils phase of the Early-glacial (EGL A) has experienced particularly intense episodes of dissolution of the chalky bedrock. These dissolution features have been reported from many sites including Roisel (Tuffreau, 1987), Mautort (Antoine, 1990), Combles (Antoine, 1991), Bettencourt-Saint-Ouen (Antoine in Locht et al., 2002), Fresnoy-au-Val (Locht, 2008) and Havrincourt (Antoine et al., 2014a). This process induces the formation of localized dissolution pockets or sinkholes that can reach over 3 m deep and in which colluvium and different grey-forest soils are trapped (Figs. 5E and 7), such as at Havrincourt and Fresnoy-au-Val. At Havrincourt, the karstic origin of these structures has been clearly demonstrated in one of the profiles that allowed observation of a depression at the top of the chalk bedrock (dissolution) at about 3 m below the surface of the pocket (Fig. 5E). The basal boundary of this feature is underlined by a thick brown-red clay layer with manganese concentrations resulting from the dissolution of the chalk bedrock (blocking horizon: Bb horizon). The formation of these sinkholes began at the end of the Eemian Interglacial, as shown by the deformations of the Bt horizon collapsed on the edges of the structure and its re-deposition into depression. The structure is filled by successive phases while continuing to deepen as indicated by the collapse of its margins which include normal step-faults and the chaotic nature of the infilling composed of a mixture of heterogeneous soil nodules and soil blocks (Bt to the extreme base, then Bt and Bth mixed in the middle part). Internodule voids are infilled with light grey to white washed silt coatings indicating a very intense vertical drainage throughout the sinkhole infilling (snowmelt). These characteristics of the infilling represent the signature of the impact of the freeze-thaw and of brutal soil drainage processes during the melting of snow cover in continental climatic conditions with strong seasonal contrasts €, 1990; during the Weichselian Early-glacial (Van Vliet-Lanoe Antoine et al., 1994). The combination of acid humic soil horizons (boreal forests dominated by pine) and of intense snowmelt processes could explain the strong increase in chalk bedrock dissolution process that characterizes this period. This dynamic, which seems related to the specific climatic conditions of that period, ends with the establishment of the first steppe soils around 70 ka in a markedly more arid environment (Figs. 4 and 6). These processes are visible in the same stratigraphic position in Belgium in the
Romont sequence also developed on chalky substratum (Juvigne et al., 2008). 4.2.2. Steppe soils phase (end of MIS 5.1, GI 20 to 19/76.4-69.4 ka) After the first phase dominated by grey forest soils, the second phase of the Early-glacial soil complex (EGL B) is characterized by the systematic occurrence throughout the whole region of isohumic steppe soils without clay illuviation (SS2 and SS3b of SaintSauflieu, n 22a-b to 19). Generally the SS2 soil is always the best preserved. It has a very constant thickness (0.4e0.5 m) and occurs rather systematically in all geomorphological contexts (Fig. 6). In the type-sequence of Saint-Sauflieu (Fig. 5A), these horizons are characterized by a moderate expansion of Birch in a steppe environment with Poaceae and Asteraceae (Fig. 4, SS2 and SS3a and b), indicating a markedly more arid and increasingly more open

environment. This part of the complex also stands out clearly from the preceding part by the onset of the first aeolian deposits (Fig. 4, n 20). It is proposed that it represents the continental response to climate oscillations that characterize the Greenland climatic records between ~70 and 75 ka (Greenland Interstadials 20 and 19). Unfortunately, the lack of precision of available luminescence dates does not currently allow a sufficiently accurate age-control for these rapid events that are, however, clearly differentiated in the continental record as distinct humic soil horizons. It is notable however that the average of the four TL and IRSL ages obtained at Fresnoy-au-Val (71.8 ka), the IRSL (regenerated) ages of 72.2 ± 8.5 and 72.3 ± 7.3 for the isohumic soil SS-2 from Villiers-Adam and Bettencourt-Saint-Ouen are perfectly consistent with an allocation of this soil horizon to GI 20. In the upper part of the steppe soil complex, the first significant evidence of aeolian sedimentation is represented by a uniform greyish loessic silt unit located between SS3a and SS3b soils (Fig. 4, n 20/thickness: 15e20 cm, Fig. 7, Mk. Silt). Given its position in the sequence and its sedimentological characteristics (non-calcareous silt including particles of local origin), this deposit is close to the characteristics of the Marker Loess defined by Kukla (1977) in stonice (MS Central Europe and especially well depicted at Dolní Ve 5 of Antoine et al., 2013). It could correspond to the first major dust event (concentration > 400 ppm) centred on 73 ka in Greenland records (GS 20). It is noticed that this part of the record (steppe soil complex) occurs with the same facies, and a rather striking manner, with a very similar thickness in the Belgian sections, such as Romont or Remicourt-Momalle (Humiferous Complex of Remicourt, Haesaerts et al., 1999), where it includes the Rocourt Tephra
et al., 2008) (Fig. 7). Further east it is also found at Nus(Juvigne sloch (Antoine et al., 2001), in Mainz-Weisenau Upper Mosbacher Humus zone (Bibus et al., 2002) or in Garzweiler Holz Humus Zone (Schirmer, 2010). It therefore allows very good regional pedostratigraphic correlations (Fig. 7). Finally, in the sections of northern France, as in most of the other west European profiles, the upper limit of the Early-glacial soil complex is defined at the level of the erosion boundary underlining the top of the younger steppe soil (Antoine et al., 1999). This boundary is a fundamental level-marker, observed in all the sequences around 70 ka, according to the data and interpretations presented above (Fig. 7). In general, the main units and the bipartition of the Early-glacial humic soil complex occur rather systematically in Belgium and Germany (Fig. 7). The regularity of the pedosedimentary budget allows consideration of accurate long distance correlations within that part of the Western European loess belt. However, the chronological control of this older part of the Weichselian pedosedimentary record remains limited owing to the unreliability of IRSL, TL and OSL dates on sediments: many age inversions, limit of the method for OSL-quartz, uncontrollable underestimation related to the fading of feldspar grains (polymineralic fraction) and more generally by wide error bars inherent to luminescence methods. Given this fact, when it is possible, the most reliable chronological control is provided by TL dating of heated flint from in situ archaeological assemblages (Figs. 4 and 6). In Belgium, the boundary between Early-glacial and Lower Pleniglacial, that was formerly defined at the top of the Malplaquet soil (Haesaerts et al., 1999), has been more recently moved down to the top of the Humiferous Complex of Remicourt, this latter being correlated with Ognon I interstadial and the upper half of GI 21 (Haesaerts et al., 2011). Taking in account the richness of the records and the accuracy of litho-pedostratigraphic correlations that can be established between northern France, Belgium and Germany, at the scale of individual soil units, a common and reliable stratigraphic scheme can

Please cite this article in press as: Antoine, P., et al., Upper Pleistocene loess-palaeosol records from Northern France in the European context: Environmental background and dating of the Middle Palaeolithic, Quaternary International (2015), http://dx.doi.org/10.1016/ j.quaint.2015.11.036

P. Antoine et al. / Quaternary International xxx (2015) 1e21

now be proposed (Fig. 7). However, a systematic effort focusing on the dating of this part of the sequence, including the latest methodological developments in luminescence methods, is still necessary in the future to build a unified chronostratigraphic scheme for that part of the record that is fundamental for the understanding of the palaeoenvironment of the Middle Palaeolithic. 4.2.3. Lower Pleniglacial LPG (~70e55 ka) For a long time, in the absence of reliable dating for this part of the sequence, the pedosedimentary budget attributed to the Lower Pleniglacial (LPG), appeared quite poor and essentially characterized by the first homogeneous aeolian deposits overlying the last

11

Early-glacial isohumic humic soil (Fig. 4, n 18, Fig. 7). These first loesses were compared to those found in the same stratigraphic position in the Rhine valley where they are dated to ~65 ka at €ller et al., Achenheim (Rousseau et al., 1998) and at Nussloch (Zo 1988). These relatively thin (0.3e0.5 m maximum), and non carbonated loess (silts) are rarely preserved, since their deposition is followed by a major erosion episode which strongly incised the underlying series and can cause an important hiatus within the sequence (Fig. 7). This erosional episode is rapidly followed by the deposition of a thick set (2e3 m) of colluvial hillwashed deposits, including geliflucted soil lenses and nodules (Bt, Bth, Ah) and small

Fig. 8. Havrincourt: stratigraphy; dating and sedimentological data (according to Antoine et al., 2014a,b modified). Units: 0) Ap horizon of the topsoil, 1a-1b-1b’) Bt horizon of the topsoil, 20 ) Decalcified homogeneous loess, 2) Homogeneous calcareous loess, 3a-c) Two folded tundra gley horizon, 4) Homogeneous calcareous loess, 5aeb) Tundra gley horizon, 6) Brownish horizon of arctic brown soil to arctic meadow soil, 7) Arctic brown soil with large bioturbations (6 and 7: Havrincourt brown soil complex), 8) Homogeneous calcareous loess, 9 and 11) Thin tundra gley horizons, 10) Weakly calcareous loess, 12) Brownish horizon (Havrincourt Brown silts), 13) Stratified colluvial deposits including numerous lenses and nodules of reworked soil horizons (Bt and Bth), 14a) steppe soil horizon, 14b) Grey forest soil horizon (Bth, Saint-Sauflieu 1), 15) Bt Horizon of brown leached soil (Eemian), 16) Homogeneous calcareous loess.

Please cite this article in press as: Antoine, P., et al., Upper Pleistocene loess-palaeosol records from Northern France in the European context: Environmental background and dating of the Middle Palaeolithic, Quaternary International (2015), http://dx.doi.org/10.1016/ j.quaint.2015.11.036

12

P. Antoine et al. / Quaternary International xxx (2015) 1e21

frost cracks, and which rework the whole underlying soil sequence (Fig. 5C and D, HCD). This unit, termed the Hermies laminated colluvial deposits according to the site where it was originally described in this area (Antoine et al., 2014a), represents a major level mark for loess stratigraphy of northern France. It is also found fairly consistently in the same stratigraphic position in Belgium in the slope profiles as in
Remicourt, Romont and Harmignies (Haesaerts et al., 1999; Juvigne et al., 2008). It also appears locally at Nussloch (Antoine et al., 2001) and in other German profiles (Niedereschbacher Zone Semmel, 1969, Bibus et al., 2007; Niedereschbacher Layer, Schirmer, this issue) (Fig. 7). However, due to its sedimentation dynamics, its dating by luminescence methods is extremely delicate, as shown by the many inversions and very high dispersion of results obtained from these deposits at Harmignies and Remicourt (P. Haesaerts oral communication). Recently (2012), new research associated with a campaign of OSL dating on quartz conducted on the sequences of Havrincourt (Pas-de-Calais), along the course of the future Seine Nord Europe Canal came significantly to complete the pattern of LPG (Fig. 8, units 12 to 8). If, in this site, is also observed a significant erosion of the humic soil complex followed by the deposition of heterogeneous bedded colluvial deposits (Fig. 8, unit 13), there are then to the top a more homogeneous and mainly non-calcareous brown silt incorporating a strong aeolian component (Fig. 8, units 12). This unit, dated to ±65 ka by OSL (Antoine et al., 2014a) is differentiated by a larger percentage of coarse silt, a drastic decrease in magnetic susceptibility values and a slight increase in CaCO3 in its upper 10 cm (Fig. 8, unit 12). This brown horizon of fairly uniform thickness (0.3e0.50 m), weakly weathered and in which is preserved the archaeological level Hav.2-N1 (Middle Palaeolithic) seems to be affected by a weak pedogenesis (thin soil of brown boreal or arctic meadow type). Nevertheless, the magnetic susceptibility values and the percentages of clay and total organic carbon (TOC) (Fig. 8) show decreasing values towards the top contrary to what is generally observed in a soil profile. In addition, the uppermost part of this unit shows a weak enrichment in CaCO3 indicating fresh carbonate inputs of aeolian origin. It is not clear whether this horizon, named the Havrincourt brown silt (HBS) has recorded a significant climatic improvement or not, even if its age is close to that of the short interstadial GI 18. Besides OSL ages of 67.6 ± 3.9 ka and 65.0 ± 3.8 ka, obtained respectively in the middle of this unit and just above (unit 10), are clearly in agreement with an allocation to LPG (Fig. 4). At a regional level, this facies has probably an equivalent in the Hermies sequence, located a few kilometres to the south-west along the “Canal du Nord”. At that site it occurs at the end of the sequence of bedded colluvial deposits, including reworked soil nodules and soil lenses (HCD) and potentially in the profile of Combles in the same stratigraphic position (Antoine, unpublished, location: Fig. 1). The top of the Havrincourt brown silt is then emphasised by the development of a very thin hydromorphic horizon (Fig. 4, microgley n 15c, Fig. 8, unit 11) related to the increase of surface moisture and to freeze-thaw processes. The sequence then shows the deposition of clearly loessic material (15b) dated at 65.0 ± 3.8 ka but still very poor in CaCO3 (Fig. 8, unit 10). Loess sedimentation is then reinforced by the deposition of a loess deposit showing a constant thickness in all profiles of Havrincourt and which reflects a very significant increase in wind dynamics with the sedimentation of the first typical carbonated loess (CaCO3: 10e15%) of the Last Glacial period (Fig. 8, unit 8; Fig. 4, n 14). Finally a first level of frost cracks opening at the base of this unit is observed (F-6). The characteristics of these small cracks with calcareous loess infilling (width: 3e5 cm) do not indicate the occurrence of a permafrost.

Comparison over longer distances with Belgian profiles as Harmignies, Romont or Remicourt (Haesaerts and Van Vliet, 1974;
et al., Haesaerts et al., 1981, 1999; Frechen et al., 2001; Juvigne 2008), leads to the identification of strong similarities for that part of the sequence spanning units 18 to 16 (Figs. 7 and 9). Indeed, in these three profiles, the Early-glacial humic soil complex is overlain by a thick and discordant unit made of heterogeneous stratified silts and soil lenses fed by the reworking of the underlying deposits and soils, then capped by a homogeneous brown loamy horizon (Malplaquet Soil of Belgian sequences). The first typical calcareous loess, including small tundra gley horizons and dated from 66 to 60 ka (Frechen et al., 2001), always appear above this brown horizon. The same observations have been reported for the Lower Rhine area at Garzweiler and Rheindhalen by Schirmer (2010 and this issue). According to this scheme, the calcareous loess that represent the first true marker of loess sedimentation in the sequence, could be compared with the LPG calcareous loess of Harmignies and Nussloch and dated at c.a. 65 ka (Antoine et al., 2001; Frechen et al., 2001). Given the OSL dating results, field observations (significant erosion, colluvial facies, frost cracks) and sedimentology (increasing wind dynamics), units 17 to 15 (Fig. 4) are allocated to the initial phases of LPG; the establishment of loess unit 14 then indicating a more advanced and intense phase of this period. This aeolian deposit is considered as a true marker of LPG at a regional scale (Fig. 9) and is allocated to the upper half part of MIS4 (Fig. 4). Further to the south of the loess area, this loess deposit can locally be replaced by an aeolian sand facies, as at Beauvais, where it is dated at 55.6 ± 4 ka by TL on heated archaeological flints (Locht et al., 1995). Hypothetically, the underlying thick layered colluvium series (Fig. 4, unit 17) could be related to GS 19 between and 64 and 69 ka and the Havrincourt brownish horizon (unit 16) to GI 18 at about 64 ka. The detailed correlation of this part of the sequence with the Greenland ice records however remains highly uncertain owing to the wide error bars that affect OSL dates from this level. 4.2.4. Middle Pleniglacial MPG/Inter-pleniglacial (~59e33 ka) In the North of France, as in all the Western European profiles, the MPG is characterized by a drastic reduction of typical loess deposition (Antoine et al., 2001). Meanwhile, boreal brown soils (Bw horizons) to arctic brown soils formed during the main interstadials that characterize this period (GI 17-16 to 8-7, Fig. 4). However, in some profiles, such as at Villiers-Adam, the base of this part of the sequence is first marked by an intense erosion phase followed by the deposition of a thick body of heterogeneous bedded colluvial and hillwashed deposits (Fig. 4, unit 13). Given the observations at Villiers-Adam (Antoine et al., 2003a), this phase is related to a major thermokarst erosion event resulting from the brutal degradation of a permafrost during a rapid warming phase that probably marks the beginning of the first interstadial of the MPG (±60e59 ka). In Germany, the same situation has been described from the base of the Nussloch profiles (thermokarst TK-2, Fig. 9). This large-scale erosional structure clearly incises across the whole underlying sequence (soil complex and LPG loess). The same type of feature is also present in Belgium at Veldwezelt-Hezerwater (Meijs, 2002) and in Garzweiler (Schirmer, 2000b), but the lack of dating at these sites prevents the comparison to be pushed further. In an environment that remains typically periglacial, the end of this episode sees the return of dry conditions and of typical aeolian sedimentation processes but its composition attests to a strong local component originating from the reworking of Paleogene

Please cite this article in press as: Antoine, P., et al., Upper Pleistocene loess-palaeosol records from Northern France in the European context: Environmental background and dating of the Middle Palaeolithic, Quaternary International (2015), http://dx.doi.org/10.1016/ j.quaint.2015.11.036

P. Antoine et al. / Quaternary International xxx (2015) 1e21

13

Fig. 9. Attempt at correlation of the sequences from northern France, Belgium and western Germany. See Fig. 4 for unit symbols. Belgium: according to: Haesaerts et al., 1999,
et al., 2008; Germany (Nussloch): according to Antoine et al., 2001, 2009; dating: Bibus et al., 2007, Tissoux et al., Haesaerts and Mestdagh 2000, Haesaerts et al., 2011; Juvigne 2009, Antoine et al., 2009). A: large ice-wedges (main network/base of 6); B: ice wedges; C: ice-melt channels (thermokarst); D: frost-creep, solifluction, gelifluction E: location of the main incision features in the sequences F: Large incision features (thermokarst erosion gullies) G: Calcareous part of the records H: Dissolution features (sinkholes).

sands in the Paris Basin (Fig. 4, n 13) or of fluvial sands in the Rhine Valley. In Nussloch, the final phase of the infilling of the thermokarst feature TK-2 is represented by aeolian sands dating from ~55 ka (Antoine et al., 2009). The oldest boreal brown soils of the

MPG soil complex are developed at the top of these sandy deposits €selberg Soil, Antoine et al., 2009). (e.g. the Lower Gra Finally, according to the data from Villiers-Adam, reinforced by those from Nussloch, it is proposed that the first witness of MPG

Please cite this article in press as: Antoine, P., et al., Upper Pleistocene loess-palaeosol records from Northern France in the European context: Environmental background and dating of the Middle Palaeolithic, Quaternary International (2015), http://dx.doi.org/10.1016/ j.quaint.2015.11.036

14

P. Antoine et al. / Quaternary International xxx (2015) 1e21

rapid warming, around 59e58 ka is represented by a strong erosion event linked to thermokarst processes rather than by the formation of a soil. 4.2.4.1. MPG soil complex: lower part (Fig. 4, n 12 to 10). In northern France, the older boreal brown soils (Bw horizon) typical of the MPG then develop on top of LPG loess or aeolian sands deposits. In most of the sequences, the MPG is characterized by a very poor pedosedimentary budget corresponding to a single brown soil horizon (Saint-Acheul Soil). However, in some especially complete profiles, such as at Villiers-Adam, a 3 m thick soil complex (Villiers-Adam Soil Complex) corresponding to the entire MPG or Inter-pleniglacial and most of MIS 3 has been described (Locht et al., 2003; Antoine et al., 2003b). At that place, no less than four horizons have been identified. They reflect the complexity of the response of the slope environments to climate and environmental changes during the MPG. They can be described successively as: a boreal brown soil, a slightly humic arctic meadow horizon, a tundra gley, laminated sandy loams and lastly an arctic brown soil (Fig. 4, n 12-8). Besides, in the North of France, the site of Henin-sur-Cojeul showed evidence of Middle Palaeolithic occupation associated with large mammals remains dated at about 42 ka (35.6 ± 1.1 and 37.9 ± 1.8 14C BP, Marcy et al., 1995) preserved in a brown soil horizon that can be related to the younger part of the MGP. At the regional scale, recent data derived from the study of the sequences of Havrincourt (Antoine et al., 2014a) allow significantly to complete this pattern. At this site, the uppermost LPG loess is affected by a clear decalcification linked to the development of the overlying boreal brown soil (compact, brown-orange, silty clayey horizon, with thin lamellar to polyhedral structure). The development of this soil horizon (Fig. 8, unit 7, Fig. 4, n 12) is emphasised by higher percentages of clay and TOC reaching respectively 28 and 0.5%, the leaching of the carbonates and strong ferro-manganic precipitation processes (nodules and coatings on biotubules). This unit corresponds to a Bw horizon of boreal brown soil, which exhibits the same characteristics as that of the lower horizon of the Villiers-Adam soil complex (Antoine et al., 2003a), the Les Vaux Soil €selberg soil of Nussloch in Belgium (Haesaerts, 1985), the Gra (Antoine et al., 2009) or the Lower Remagen soils (Remagen 3 and 2: Schirmer, 2000a, this issue). However the record of Nussloch shows that, in the contexts where sedimentation rates are the highest, this soil which corresponds to the balance of the most important part of the MPG, can be subdivided into two sub-units €seldeveloped between ~58 and 48.5 ka (Lower and Upper Gra berg soils). The OSL dates of 57.00 ± 3.6 and 51.5 ka ± 3.2 ka, obtained respectively at the base and top of the brown soil of Havrincourt (Fig. 4, unit 12; Fig. 8, unit 7), confirm its allocation to the lower half of the MPG. According to this result, the various cold episodes dated around 48 and 58 ka in NGRIP would not have any impact on the loess environments of north-western France, likely because the loess dynamic was too weak to allow the separation of the different phases of weathering and soil formation. This pattern, in which the development of the main MPG boreal brown soil is assigned to the long time span from GI interstadials 14 to 12 is reinforced by: 1) the ages of 48.7 þ 8.2/e6.9 and 51 ± 3 ka obtained at the base of the boreal brown soil in the site of Savy using respectively ESR-U/Th (on tooth and bones) and TL from sediments (Locht et al., 2006), and 2) the TL age of 55.6 ± 4 ka from the aeolian sands directly underlying this soil at Beauvais (Oise) (Locht et al., 2014a). In addition, this brown soil is characterized by numerous large rounded or oval bioturbations (krotovinas: 5e15 cm in diameter) resulting from burrowing small mammals (marmots, ground squirrels, arvicolidae, steppe polecat). Besides, the large mammal

assemblages found at Havrincourt in this horizon include large herbivores such as horse, the woolly rhinoceros and mammoth (Auguste in Antoine et al., 2014a). These data both indicate the presence of a vegetal biomass, much greater than during the deposition of the underlying loess. Note that these bioturbations structures, which deeply penetrate into the underlying calcareous loess (Fig. 7, unit 8), appear in the same stratigraphic position in several northern sequences of France (Antoine et al., 2003a), and more generally in other profiles of north-western Europe as in Nussloch (Germany). The abundance of these burrows reflects the development of a singular steppe environment particularly favourable to colonisation by these rodents (stabilized and densely vegetated surface). According to 14C ages obtained from rodent bones from krotovinas, the surface of the soil where these mammals have been living is dated at about 45e46 ka and thus can be allocated to GI 12. It is possible that this phase of the MPG characterised by more vegetal biomass corresponds to the slightly more humic horizon observed in Nussloch on top of the Upper €selberg soil dated at about 45 ka (Antoine et al., 2001) and to Gra the humic horizon of the Villiers-Adam MPG soil complex (Fig. 4, n 11). In this context, the sudden disappearance of these rodents and their trapping deep in the burrow system observed at Havrincourt could result from one directly posterior and especially intense cold event with permafrost (between ~ 44 and 40 ka) that likely corresponds to the Hasselo Stadial defined in the Netherlands Coversands region (Ran et al., 1990). The brown soil horizon is indeed affected by a fine lamellar structure resulting from the multiple freeze-thaw alternations post-dating the soil development and indicating at least one episode of permafrost formation. This observation is completed by the discovery of a first network of Vshaped frost cracks of the ice-wedge type (~1.5 m deep) that open at the top of the unit (Fig. 4, network F-5). These periglacial features are unfortunately badly preserved and often largely masked by the opening of the large ice-wedge casts from the overlying main level F-4 that are systematically superimposed to F-5 (Figs. 4 and 8). Given the OSL ages of the succession of the pedosedimentary stages observed in the units that surround it, it is possible that the first icewedge network developed around 44e43 ka during GS 11. These structures would thus predate the Heinrich 4 event that is however distinguished by particularly cold conditions in the North Atlantic and in the bordering areas around 40 ka (Bond et al., 1993; Scourse et al., 2009; Marcott et al., 2011). According to their position in the stratigraphy and the OSL data, the thaw structures, which strongly degrade this first generation of ice-wedge that have developed at Havrincourt, could correspond to those described in the Villiers-Adam soil complex (Fig. 4, top of unit 10) (Antoine et al., 2003a) and to the oldest 14C ages on wood (±40 ka cal.) obtained at the base of thermokarst infilling TK-1 at Nussloch (Antoine, unpublished, Antoine, 2012). The evidencing of these thawing features at Havrincourt, in association with a first network of large ice-wedges is very important because it confirms the registration of a permafrost episode, followed by a degradation event (rapid warming) within the MPG soil complex. The sedimentary hiatus between units 10 and 9/8 (Fig. 4) is also emphasized in the analytical data from Havrincourt by a decline in total organic carbon concentration (0.32%) and a first peak of CaCO3 indicating a new phase of loess deposition to the base of units 9/8. However, the break in the curve of the TOC concentration at the transition between units 6 and 7 (Fig. 8) suggests that the brown soil of unit 7 has undergone a significant truncation related to the phase of permafrost degradation associated with the F-5 ice-wedge network. This erosion could explain the absence of a humic horizon at Havrincourt at the top of the lower brown soil as in Villiers-Adam (Fig. 4, n 11).

Please cite this article in press as: Antoine, P., et al., Upper Pleistocene loess-palaeosol records from Northern France in the European context: Environmental background and dating of the Middle Palaeolithic, Quaternary International (2015), http://dx.doi.org/10.1016/ j.quaint.2015.11.036

P. Antoine et al. / Quaternary International xxx (2015) 1e21

4.2.4.2. MPG soil complex: upper part (Fig. 4, units 8 and 9). The formation of the upper part of the Havrincourt soil complex (Fig. 8, unit 6, Fig. 4, n 8) begins with a new thin loess deposit, mostly trapped in the ice-wedge casts of F-5 level. It is followed by at least two phases of soil formation whose level of development remains significantly lower than for the underlying brown soil. This unit (Fig. 8, unit 6) is distinguished in the field by its compact clayey facies with a strong lamellar structure (2e4 mm in thickness) covered with orange iron oxides and black manganese coatings. It corresponds to an hydromorphic arctic brown to arctic meadow soil horizon (TOC: 0.4% clay: 29%) strongly structured by the freezethaw processes connected with the ice-wedge level F-4. The structure in two peaks of the curve of the clay percentage, as well as the presence of an in situ archaeological level included in its upper third (level Hav.2-N2), suggest that unit 6 corresponds to a polygenetic unit composed at least by two stacked soil horizons. The originality of this horizon is to present a gradual increase in CaCO3 percentage from the base to the top (2e8%) showing that the weak pedogenesis is progressively drown by new calcareous loess input. Given the OSL (~35 ka) and 14C dates obtained on large mammals remains associated to the archaeological level included in its upper part (~33e34 ka cal. BP), this horizon could represent the balance of interstadials GI 8 and 7 or 8 to 6 (~38.5e33 ka). According to available data, this horizon can be correlated with the uppermost part of the Villiers-Adam soil complex (upper brown soil, Fig. 4, n 8) and with the Lohne soil of German profiles as Nussloch (Antoine et al., 2009). It does not appear to have an equivalent in the Belgian records where it is either eroded or included within the MPG Les Vaux Soil. Data from Havrincourt are here considered very important because they provide fundamental information for the dating and interpretation of the youngest brown soil horizon that, in all European loess sections, marks the end of the MPG and the dating of which is debated, in particular concerning the reference sequence of Nussloch (Antoine et al., 2009; Kadereit et al., 2013). In the field, this stratigraphic limit represents a fundamental marker for correlations at the European level and therefore it is important to attempt to clarify it. In general, it results in the abrupt transition from a long period (~20 ka) characterised by very weak loess sedimentation (0.05e0.1 mm/year) and the formation of boreal brown soils to a phase where the environment is definitely dominated by the deposition of thick calcareous loess covers, very high sedimentation rates and an outstanding deposition volume (1 mm/year, Frechen et al., 2003; Antoine et al., 2009). However loess records never being continuous, a variable hiatus, which can reach several thousand years (5 ka) characterises this limit. Indeed, given the geochronological data obtained from Havrincourt and in the reference sequences like that of Nussloch, the strong increase in loess sedimentation rates starts around 31.5 to 30 ka while the youngest brown soils are dated to ~34e38 ka. At Havrincourt the upper brown soil horizon occurs at the hinge that marks the extreme end of the MPG. Compared with sequences like Nussloch, this unit should be allocated to a final MPG for its base (Lohne Soil) and to an initial UPG for its more hydromorphic uppermost level containing the Early Upper Palaeolithic level N2 (Nussloch G1?). Nevertheless, the pedosedimentary budget of this part of the record being weakly developed, for obvious reasons of clarity in the field approach, this unit is still assigned to the MPG (Antoine et al., 2014a). In general the issue of the boundary between MPG and UPG is important and especially regarding Havrincourt where an archaeological level dating from the Older Upper Palaeolithic (Hav.2-N2) is contemporaneous of this hinge time period (Fig. 8, unit 6). Finally, it should be noted that a new Older Upper Palaeolithic level has been discovered recently in the loess sequence Renancourt-2, close

15

to Amiens (Fig. 1), within a unit of compact brownish loess with abundant secondary CacO3 coatings on biogalleries, underlying the deposition of the laminated loess of the UPG (Paris et al., 2013). The large mammal fauna associated with this level is dominated by horse, reindeer and bison. Based on radiocarbon dates obtained from these bones (average: 2,8 ,000 14C BP/33e34 cal. BP) this Palaeolithic level (Gravettian) and associated mammal remains are contemporaneous of Havrincourt Hav.2-N2. 4.2.5. Upper Pleniglacial/UPG (~30e15 ka) In general, this phase of the Last Glacial is characterized in all regional profiles by a marked acceleration of loess sedimentation rates (Fig. 4, n 6-2). Locally very thick, 4e5 m on the leeward slopes facing from NE to SE, these calcareous loess represent a major component of the current landscape (Fig. 2). The UPG calcareous loess cover is divided up into three main units separated by periglacial marker horizons (tundra gleys and large ice-wedge cast networks) (Fig. 4). Within these loesses, recent investigations, particularly on molluscan assemblages, allowed demonstration of a direct relationship between the development of tundra gley horizons and millennial climate improvements recorded in the Greenland ice cores (Antoine et al., 2009; Moine et al., 2011). The malacological observations lead on UPG sequences of Northern France, however, show especially poor assemblages compared to the more continental environments, like those in the Rhine Valley, and which reflect poorly vegetated areas (Moine et al., 2011). This low vegetal biomass is consistent with the near absence of large herbivores remains in the calcareous loess deposits of the UPG in Northern France. The new data from Havrincourt allow general completion of the scheme of UPG for the North of France, notably regarding associated periglacial features (Antoine et al., 2014a). In this site, the first unit allocated to the UPG is a tundra gley (Fig. 8, unit 5, Fig. 4, n 7) that represents a fundamental limit versus the underlying soil complex. Indeed, regardless of its facies related to waterlogging, this unit shows already from its basal part, sedimentological characteristics of typical loess (extremely low TOC, high CaCO3%, very high coarse silt %). This observation indicates that this horizon has developed upon a ‘fresh’ aeolian input strongly contrasting with the weathered clayey loessic silt, partially gleyed, forming the top of the underlying Havrincourt soil complex. The formation of this tundra gley horizon results from seasonal water saturation of the active layer of a permafrost indicated by its connection with the large ice-wedge casts of the main network (F-4). The cryoturbation of the horizon likely relates to thawing of the ice lenses in the upper part of the permafrost as proposed by Vandenberghe and Nugteren (2001). It is underlain by an irregular greyish strip, a few centimetres thick (unit 5a), enriched in secondary CaCO3 (concretions and roots hypo-coatings) and including blackish organic debris (vegetal remains, fungal and lichen crust?) that likely represents the surface of a former tundra soil. The curves of the malacological abundance seem also to show a separation of the tundra gley into two sub-units (5ae5b). The two peaks of abundance, associated to increases in the proportion of juveniles in the genus Pupilla, indicate a warming and lengthening of the summer season consistent with the thawing of the ice at the top of the permafrost and with the cryoturbation of tundra gley horizons characterizing the interstadials (Moine et al., 2008, 2011). The OSL results date the formation of the tundra gley of unit 5 from Havrincourt to about 31e32 ka (31.2 ± 2.1 and 31.4 ± 2.0 ka). They are consistent with the dates of underlying units, 14C ages of archaeological level Hav.2-N2, as well with an allocation to GI 5 or 5 and 6 (~32e33.5 ka). The age of this gley corresponds to that of tundra gley G2 from Nussloch (Antoine et al., 2009) and horizon 8a8b at Savy (30 ± 2 ka ESR-U/Th, Locht et al., 2006). This horizon has also been observed in the North at Curgies (Deschodt, unpublished)

Please cite this article in press as: Antoine, P., et al., Upper Pleistocene loess-palaeosol records from Northern France in the European context: Environmental background and dating of the Middle Palaeolithic, Quaternary International (2015), http://dx.doi.org/10.1016/ j.quaint.2015.11.036

16

P. Antoine et al. / Quaternary International xxx (2015) 1e21

and Bourlon (Moine et al., 2011) in the same stratigraphic position and in both sites in connection with the main ice-wedge cast level F-4. The two-folding of this horizon, highlighted in the sites of Havrincourt, Nussloch and Bourlon, thus would characterize the response of Western European palaeoenvironments to millennial climatic warming phases in a context of weak loess sedimentation rates (Moine et al., 2011). Finally, given its stratigraphic position and its facies, this tundra gley can be compared to the MesnilEsnard hydromorphic soil described from Normandy (Lautridou, 1985) where it is associated with a network of large ice-wedge casts comparable to that of F-4 network. In Belgium sequences it could match with the small hydromorphic horizon figured at the top of the Vaux Soil (Haesaerts, 1985). In the field, the base of the UPG sequence is thus differentiated by the formation of a network of large ice-wedge casts with loess infilling (2e3 m depth/opening ~ 0.5 m) with a decametres-scale mesh (12e14 m), indicating a particularly intense climatic degradation (Fig. 5F). This network of tundra polygons (F-4), the horizontal extension of which has been monitored and photographed in a unique way over more than 4000 m2 at Havrincourt (Fig. 5G), represents the original signature of a continuous permafrost that is the marker of the beginning of the UPG across the whole northern France loess area (Fig. 4). At the same time, during the Late Pleniglacial (UPG), continuous permafrost has been demonstrated from braided fluvial sequences (Kasse et al., 2003). The large epigenetic V-shaped ice-wedge structures (Haesaerts €, 1987; and Van Vliet, 1973, 1981; Pissart, 1987; Van Vliet-Lanoe French, 2007; French and Shur, 2010; Matsuoka, 2011) observed in Northern France are particularly well preserved at the interface between units 6 and 7 (Fig. 4). Their visibility is accentuated by major lithological and colour contrasts that distinguish the parent material (greyish-brown clayey silt) from the infilling of these structures (pale yellow calcareous loess), as in other profiles from
ger-TGV, Ploisy, CourNorthern France (Savy, Curgies, Saint-Le melles…) (Sellier and Coutard, 2007). The formation of large ice-wedge casts, the opening of which reaches ~0.5e0.6 m at Havrincourt or Sourdon (Antoine, 1990), indicates a gap of in the sequence. However annual opening rates of these structures are still unclear and partly related to the type of substratum in which they grow and its ice content (Black, 1976; Mackay, 2000; Murton, 2007) and thus the duration of this gap is very difficult to determine. The sediment infilling, which allowed the preservation of these ice-wedge casts after the thawing of the ice core, corresponds to two distinct phases, clearly visible at Havrincourt for example (Fig. 5F): 1) Fossilisation of the borders (banks) of the wedge during permafrost degradation (thawing of the ice core) by the greyish hydromorphic material produced by the slow creep of the active layer of the permafrost (tundra gley). The lack of stratification in the infilling of wedges of the main network F-4 involves a slow and gradual decline of the ice core with a good drainage of meltwater in depth without hillwash processes, contrary to what was observed for level F-5. 2) Infilling by homogeneous calcareous loess in arid and cold environment, and final fossilisation of the structure. This succession, which has been studied in detail for several years in the sequences of north-western Europe, reflects the rapid alternation of two extreme climatic phases (Moine et al., 2008; Antoine et al., 2009). 1) Development of ice-rich permafrost and of a network of large ice-wedges during a period characterized by extremely cold

winters likely to initiate the process. The environment was however sufficiently wet for the soil to be rich in ice (snow) and aeolian sedimentation is in strong reduction. 2) Rapid degradation of permafrost causing the fusion of ice wedges, deepening of the active layer (thickening of tundra gley facies) and its creep around wedges features released by the melting of the ice (Fig. 5F). In case of particularly intense and rapid warming, and in the presence of a significant slope, melt water released by ice-wedges melting causes runoff process (meltwater channels) that erode the initial wedges structures (thermal erosion) (see F-2 level). In slope contexts, this process can lead to the progressive enlargement of initial channels guided by the polygonal structure and the setting of a drainage network linked to thermokarst erosion processes. At the final stage, the coalescence of these channels induced the formation of thermokarst gullies able to deeply incise the slopes as it was observed in Villiers-Adam (Antoine et al., 2003a), in Bettencourt-Saint-Ouen (Antoine in Locht et al., 2002) or in Nussloch (Antoine, 2012 Antoine et al., 2001). These processes, which repeat several times during the UPG, could explain the origin of the often extremely important hiatuses (and erosional facies and gravel beds) which appear in these sequences in slope positions in particular in the Somme basin (case of SaintSauflieu). Later on, the fossilization of the tundra gley and of the network of large ice-wedge casts associated with a loess deposition occurs in an environment that becomes again very dry and cold. Nevertheless, the interweaving of the tongues of the geliflucted gley of unit 7 and of the overlying loess (Fig. 8) shows the high speed of the change between the final phase of creep of the tundra gley and the sedimentation of the first loess deposits. The climatic degradation of the UPG goes on with the deposition of the homogeneous calcareous loesses of unit 6 (Fig. 4), on a particularly constant thickness in the various plateau sites of northern France (~0.6e0.8 m). These loesses, which were observed in the same stratigraphic position in a large number of profiles of the North of France, in particular along the track of the TGV North (High Speed Train) (Antoine, 1991), constitute a marker of the beginning of the UPG the dating of which was so far rather weak (Locht et al., 2006). These new data from Havrincourt and the OSL dates now allows them to be placed in parallel with the major climatic degradation centred on 30.5 ka in the reference palaeoclimatic records (Fig. 4). It is notable that this period corresponds to the Heinrich 3 event described in the marine cores of the North Atlantic (Bond et al., 1993; Marcott et al., 2011). As in the case of the network F-5, the formation of the permafrost associated with the major network of ice-wedges F-4 would be older of a few thousand years to the following Heinrich event. This first loess unit is contemporaneous of an unprecedented increase of the values of dust concentration in Greenland ice cores (Rasmussen et al., 2014) (Fig. 4) reinforcing the link that has been evidenced between Greenland ice-cores and western European loess records (Rousseau et al., 2007; Antoine et al., 2009). The peak of this first loess unit of the UPG is then underlain by a thick and complex tundra gley (Fig. 4, n 5a-c; Fig. 8, n 3a-c) which is subdivided into two sub-horizons separated by a thin bed of calcareous loess mainly preserved within an ice-wedge cast (network F-3, Fig. 4). The infilling of this F-3 network shows locally laminations and cross stratifications indicating a phase of permafrost thawing during an episode of interstadial warming. This subdivision of the gley of unit 5, often difficult to observe in the field in the absence of ice-wedge casts, is however clearly depicted by the analytical data, especially by clay percentage, TOC and malacology (Fig. 8). Indeed, although less clear than in the

Please cite this article in press as: Antoine, P., et al., Upper Pleistocene loess-palaeosol records from Northern France in the European context: Environmental background and dating of the Middle Palaeolithic, Quaternary International (2015), http://dx.doi.org/10.1016/ j.quaint.2015.11.036

P. Antoine et al. / Quaternary International xxx (2015) 1e21

underlying gley, a bimodal profile characterizes the curve of the total abundance in molluscs concerning all taxa throughout this gley (Moine in Antoine et al., 2014a,b). The upper horizon of the tundra gley is associated with a large ice-wedge casts network with homogeneous loess infilling ~0.3  1.2 m deep (network F-2) (without thawing structures), which is systematically superimposed on the previous level (F-3). The succession of facies 3a-b (Fig. 8) corresponds to the same dynamics as the one who was highlighted for units 4-3c and detailed above. The size of the network could not be accurately identified, because of the outcrop conditions (no horizontal exposure); however it is noted that level F-2 seems to be systematically superimposed in the level F-4 (as well as the level F-3), indicating a decametric polygonal network of the same magnitude. One OSL date (28.4 ± 1.8 ka) is available from the loess that fossilized this level. This result shows that this part of the sequence (Fig. 8, units 5 to 2) accumulated very rapidly, possibly during a time span of ±2 to 3 ka maximum. The analysis of this part of the sequence at Havrincourt therefore shows similar characteristics to those of the complex gley formerly observed in northern French sequences and called the Santerre cryoturbated Horizon (Antoine, 1991). Until now the age of this layer has been unclear, however, given the timing provided by the new OSL and climatic history recorded in this part of the sequences (quick succession of two interstadials separated by a short stadial), it is possible to allocate it to the succession of interstadials GI 3 and GI 4, respectively centred around 27.4 and 28.5 ka. Finally, it should be noted that the age of about 27 ka obtained recently on the site of Amiens-Renancourt-1 (Final Gravettian) associated with a small tundra gley (Paris et al., 2013) is also consistent with an allocation of this tundra gley to one of the two short interstadials GI 3 or GI 4. In western Europe, this tundra gley doublet was identified and dated around 28 ka in the reference sequence of Nussloch (G3eG4, Antoine et al., 2009) and could correspond to the gley HC6 in Belgium sequences as Harmignies, although TL ages obtained there are slightly younger (mean age: ~26 ka, Frechen et al., 2001). The overlying unit, that represents the thickest loess deposit in slope position throughout the whole northern France area (3e4 m), is represented by finely laminated loess showing syn-sedimentary cryo-dessication micro-cracks (polygonal network: 0.2e0.4 m), typical of niveo-aeolian deposition processes (Fig. 4, n 4). The laminations are composed by alternating millimetric layers of clayey silts and of coarser material showing a fine wavy pattern. The coarser layers frequently include medium to coarse sands and chalk grains or even small gelifracted flint fragments in the sequences of the Somme basin where the chalky bedrock outcrops in the immediate vicinity. It must be underlined that this loess, explored in hundreds of test-pits, has never provided any Palaeolithic artefact and no large mammal remains. Besides, molluscs are extremely rare or even absent in this unit indicating a fully open environment and very cold climatic conditions during its deposition (Moine et al., 2011). Taking in account the age of the underlying tundra gley, the average of TL-IRSL dates from Saint-Sauflieu (Frechen et al., 1995 unpublished) and the recent OSL results from Amiens-Renancourt 2 (Paris, 2015), the deposition of this loess unit can be allocated to the cold stage between about ~27 and ~23 ka, to H2 event and to Greenland stadial 3. Those calcareous laminated loess represent a marker facies of the UPG that corresponds to the Hesbayan loess of Belgium and Lower Rhine area (Fig. 9). Tis unit can be followed on the long distance until western Ukraine or in Czech Republic as in stonice (Antoine et al., 2013). Dolní Ve A the top of these loess a last greyish tundra gley horizon is observed in northern France sometime associated with the youngest ice-wedge casts network as in Havrincourt (level F-1: V-

17

shaped features of ~0.3 m opening and 70e80 cm deep). In slope profiles, as in St Sauflieu, a gravel bed (gelifracted flints) testifies of a strong erosion event responsible for the truncation of the gley and of the upper part of the underlying laminated loess. Generally, the gley horizon (Fig. 4, n 3) and the associated ice-wedge casts are however fully integrated within the clayey Bt horizon of the surface soil and strongly masked by pedogenesis, and thus often difficult to demonstrate. This configuration is found in many northern French profiles because of the weak thickness of the uppermost loess units (1.5e2 m, up to 2.5 m thick in the best locations), which are generally fully decalcified and enriched in clay. No dating is currently available for this gley horizon and for the homogenous loess representing the end of the UPG (Fig. 4, n 2), equivalent to the Brabantian of Belgium profiles. However, given the sequence of events described in the underlying deposits, it is likely that this horizon was formed during the last short interstadial occurring at the end of Last Glacial (GI 2 ~23 ka). From a stratigraphic point of view, it could match the Nagelbeek €, 1992), whose Horizon (Haesaerts et al., 1981; Van Vliet-Lanoe equivalent was identified in the north of France in the OnnaingToyota section (Antoine, unpublished), or the gley G7 of Nussloch directly overlying a loess dated at ~23 ka OSL and including the Eltville Tephra (Antoine et al., 2009). At the Onnaing-Toyota site an AMS radiocarbon date of about 21.4 ka from loess organic matter has been obtained from this layer (18,031 ± 184 BP, Gif 80229), but this result has likely underwent rejuvenation due to the proximity of the surface soil. Until now, only one large mammal remain (fragment of a reindeer antler) has been discovered in this unit at Mautort (Somme) but no human artefacts. 5. Humans and environment during the Last climatic cycle: some reflections Recent data concerning human occupation during the Last InterglacialeGlacial cycle allow recognition of a strong concentration of Middle Palaeolithic sites during the Weichselian Earlyglacial (Fig. 4) and more specifically during the phase characterised by grey-forest soils (Antoine et al., 1994, 2003c; Locht et al., 2014a,b). During this long period, marked by climatic oscillations in a cool continental climate dominated by boreal forest with pine and birch, the sites are trapped by a slow and gradual colluvial sedimentation which allow the spatial distribution of artefacts to be preserved in situ. However, this concentration of occupation is relative because of the long duration of this period of grey-forest soils development (~30 ka) compared to that of the following steppe soils phase (~5e7 ka). The current results emphasize the sharp reduction of the occupation by Neanderthals during the Lower Pleniglacial, a period during which some Neanderthal human settlements, however, were evidenced such as at Beauvais or Havrincourt. A denser occupation reappear only during the Middle Pleniglacial but with a much more restricted importance than during the Early-glacial. The youngest phase of the Middle Palaeolithic is represented by a much smaller number of sites, although the time-gap appears to fill increasingly with new discoveries arising from rescue archaeological investigations. Based on these data, the last Neanderthals occupied the region during the formation of the lower half of the Villiers-Adam soils Complex. During this period the climatic context was characterised by the alternation of extremely rigorous stages (loess/ice-wedges) and of rapid climatic improvements in which brown soils are developed in a steppe environment (Fig. 4). The work undertaken on the sequences of the Canal Seine-Nord Europe have contributed to the strengthening of the data for these periods (Goval et al., 2014). In particular they showed that human occupation allocated to the Late Middle Palaeolithic, dated

Please cite this article in press as: Antoine, P., et al., Upper Pleistocene loess-palaeosol records from Northern France in the European context: Environmental background and dating of the Middle Palaeolithic, Quaternary International (2015), http://dx.doi.org/10.1016/ j.quaint.2015.11.036

18

P. Antoine et al. / Quaternary International xxx (2015) 1e21

at around 65e60 ka, were settled in a steppe environment where vegetal biomass was important enough to support a rich herbivorous fauna combining mammoth, rhinoceros, bison and horse (Auguste in Antoine et al., 2014a). At the Havrincourt site the level of Early Upper Palaeolithic, located in the upper part of the Middle Pleniglacial soil complex, differs significantly with the appearance of reindeer the signs of which indicate a markedly poorer vegetal cover, closer to the tundra type during this period. During the Upper Palaeolithic, human occupation in the area was discontinuous. During the early phases of the Upper Palaeolithic (38e28 ka BP), only a few and brief Aurignacian and Gravettian incursions, usually out of stratigraphic context, were known until recently. However, recent work has allowed the discovery of in situ occupation dating from the early Upper Palaeolithic in the loess sequence of Havrincourt (Goval, in Antoine et al., 2014a; Goval et al., 2014) and Amiens Renancourt-2 (Paris et al., 2013). Both human occupations, dated around 33e34 ka, correspond to the latest incursion of Palaeolithic hunters onto the great plains of northern France before the Upper Pleniglacial. During the Upper Pleniglacial, characterized by the deposition of large loess blankets (units 6-2, Fig. 4), hundreds of archaeological test-pits performed at the regional level have shown that large mammal remains were extremely rare in the area, while the preservation of bones is optimal in calcareous loess. Only one Final Gravettian level recently found at Amiens-Renancourt-1 (Fig. 1) demonstrates the presence of human groups in the region around 27 ka (Fagnart et al., 2013; Paris et al., 2013). This exceptionally rich archaeological level is preserved in situ within a small tundra gley horizon clearly underlying the laminated calcareous loess with cryo-dessication cracks that marks the thickest loess sedimentation event at the regional level (Fig. 4, n 4). Following a settlement hiatus of at least 10 ka, northern France area was again occupied during the Lateglacial by the Upper to final Magdalenian hunters of reindeer and horses in a steppic and continental environment. The low density of the sites implies that the region was a territory occasionally occupied or travelled by hunters at the margins of more important population centres (Paris Basin, Belgian Ardennes, Rhineland). Finally a markedly denser human occupation occurs during the Allerød Interstadial (~13.4e13.6 ka) with the rapid spread of Final Palaeolithic groups (Federmesser) throughout the great northern Europe Plain (Fagnart, 1997).

3) In addition, new developments in the field of geochronology now allow quite precise dating of the main units forming that sequence and consideration of correlations with the reference palaeoclimatic records like those in Greenland. Indeed, based on the dating and high-resolution sedimentological analysis it appears that northern French sequences, like those of neighbouring regions, recorded the impact of millennial climatic variations that characterize the Last glacial in Western Europe (DansgaardeOeschger cycles) in the form of loess-soil or loesstundra gley successions (doublets). However, pedostratigraphic and sedimentological analysis and dating clearly show the actual limit of this approach. Indeed, many of the short periods of climatic degradation that separate Greenland interstadials have not generated sufficiently significant loess deposition phases to be preserved as independent units in the continental record. The consequence of this is the stacking of a many interstadials within a small thickness, as it is the case during the Middle Pleniglacial between about 55 and 35 ka (formation of polygenetic soil horizons and of soil complexes). 4) This work has also recently contributed to highlighting a unique succession of periglacial horizons of regional value, including four main levels of large ice-wedge casts between ±45 and 28 ka. The main level F-4, connected with the calcareous loess that rapidly infill it, represent a marker of the base of the Upper Pleniglacial in the whole area. 5) Finally, the renewal of the data shows that the human occupation is significantly discontinuous with a very clear maximum in the density of sites during the Early-glacial with continental climate and a context of boreal forest to forest steppe (~112e70 ka). However, only a few occupations occurred during the Lower Pleniglacial (around 60 ka) and during the Middle Pleniglacial (around 40e50 ka). A total abandonment of the region is observed between ~27 and 15 ka. Given these results, a proven relationship between the density of human population and climate and especially environmental context appears to exist. This relationship seems to be conditioned by the relative abundance of a large mammal fauna, an essential source of livelihood for prehistoric populations, itself related to the type and density of vegetation. This apparently reflects the extreme poverty of the biomass during the deposition of the Upper Pleniglacial loess cover.

6. Conclusion

Acknowledgments

Over the past twenty years, the investigations undertaken on the last Interglacial-glacial loess-palaeosols sequences in northern France have been based on the integration of geological, geomorphological, palaeo-biological, archaeological and geochronological approaches, in a context where rescue archaeology generated an unprecedented increase in new data. This multidisciplinary work has led to the establishment of a reference litho-pedostratigraphic and chrono-climatic sequence and database for the study of human-climate-environment relationships in Western Europe and to the following conclusions:

The authors warmly thank the Pr. Philip L. GIBBARD for the review and the correction of the English version of this manuscript and the French Institute for Rescue Archaeological Research (INRAP) for the financial support to OSL dating of the Havrincourt sequence.

1) From a pedostratigraphic perspective, the analysis and correlation of all sequences have resulted in a rich and complex loesssoil reference sequence in which many marker horizons allow to draw accurate correlations and to build a synthetic record of regional value. 2) This work shows that the pedostratigraphic sequence of northern France exhibits a strong coherence at the regional level over more than 20 000 km2 between Normandy and the north of France and that the outlines of this system are perfectly reflected in the sequences of Belgium and western Germany.

References Antoine, P., 1989. Le complexe de sols de Saint-Sauflieu (Somme), micro
but Weichse
lien. Publ. du morphologie et stratigraphie d'une coupe type du De
historiques 1, 51e59. Centre d’Etudes et de Recherches Pre
olithique du bassin Antoine, P., 1990. Chronostratigraphie et environnement du Pale

historiques de la Somme. Publications du Centre d'Etudes et de Recherches pre 2, 1e231.
es sur la stratigraphie du Ple
istoce ne supe
rieur de Antoine, P., 1991. Nouvelles donne s les sondages effectue
s sur le trace
du TGV la France septentrionale, d'apre
historiques 3, 9e20. Nord. Publ. du Centre d'Etudes et de Recherches Pre
, J., 1994. Re
ponse des environnements a  Antoine, P., Munaut, A.V., Somme
volution climatique du De
but Glaciaire weichse
lien: donne
es de la France du l'e Nord Ouest. Quaternaire 5, 151e156.
, J., Auguste, P., Auffret, J.P., Baize, S., CletAntoine, P., Lautridou, J.-P., Somme
, O., Joly, F., Laignel, B., Laurent, M., Pellerin, M., Coutard, J.P., Dewolf, Y., Dugue
, M., Lebret, P., Le
colle, F., Lefe bvre, D., Limondin-Lozouet, N., Munaut, A.Lavolle V., Ozouf, J.C., Quesnel, F., Rousseau, D.-D., 1998. Le Quaternaire de la France du
lations. Quaternaire 9 (3), 227e241, 1 pl. HT. Nord-Ouest: Limites et Corre

Please cite this article in press as: Antoine, P., et al., Upper Pleistocene loess-palaeosol records from Northern France in the European context: Environmental background and dating of the Middle Palaeolithic, Quaternary International (2015), http://dx.doi.org/10.1016/ j.quaint.2015.11.036

P. Antoine et al. / Quaternary International xxx (2015) 1e21
, C., 1999. Last Interglacial-Glacial Antoine, P., Rousseau, D.D., Lautridou, J.P., Hatte climatic cycle in loess-palaeosol successions of north-western France. Boreas 28, 551e563. €ller, L., Lang, A., Munaut, A.V., Hatte
, C., Fontugne, M., Antoine, P., Rousseau, D.D., Zo 2001. High resolution record of the last Interglacial-glacial cycle in the Nussloch loess-palaeosol sequences, Upper Rhine Area Germany. Quaternary International 76e77, 211e229.
, C., Zo € ller, L., Lang, A., Fontugne, M., Moine, O., Antoine, P., Rousseau, D.D., Hatte
e
nements e
oliens rapides en contexte loessique: l'exemple de la 2002. Ev
quence du Ple
niglaciaire supe
rieur weichselien de Nussloch (Valle
e du Rhinse Allemagne). Quaternaire 13 (3e4), 199e208.
, C., LimonAntoine, P., Bahain, J.-J., Debenham, N., Frechen, M., Gauthier, A., Hatte din-Lozouet, N., Locht, J.-L., Raymond, P., Rousseau, D.D., 2003a. Nouvelles
es sur le Ple
istoce ne du Nord du Bassin Parisien: les se
quences loessiques donne de Villiers-Adam (Val d'Oise, France). Quaternaire 14, 219e235.
, J., 2003b. The loess and coversands of Antoine, P., Catt, J., Lautridou, J.P., Somme Northern France and Southern England. Journal of Quaternary Sciences 18, 309e318. res, C., Antoine, P., Auguste, P., Bahain, J.J., Coudret, P., Depaepe, P., Fagnart, J.P., Falgue
, C., Lamotte, M., Laurent, M., LimondinFontugne, M., Frechen, M., Hatte Lozouet, N., Locht, J.L., Mercier, N., Moigne, A.M., Munaut, A.V., Ponel, P.,
oenvironnements ple
istoce nes et peuplements Rousseau, D.D., 2003c. Pale
olithiques dans le bassin de la Somme (Nord de la France). Bulletin de la pale
te
pre
historique Française 100 (1), 5e28. Socie Antoine, P., Limondin-Lozouet, N., Auguste, P., Locht, J.L., Ghaleb, B., Reyss, J.L.,
, E., Carbonel, P., Mercier, N., Bahain, J.-J., Falgue res, C., Voinchet, P., 2006. Escude
vidence d'une se
quence Le tuf de Caours (Somme, Nord de la France) : mise en e
olithique associe
. Quaternaire 17 (4), de tufs calcaires eemiens et d'un site pale 281e320.
, C., Lang, A., Zo €ller, L., Antoine, P., Rousseau, D.-D., Moine, O., Kunesch, S., Hatte 2009. Evidence of rapid and cyclic eolian deposition during the Last Glacial in European loess series (Loess Events): the high-resolution records from Nussloch (Germany). Quaternary Science Reviews 28, 2955e2973. Antoine, P., 2012. Thermokarst processes and features from west-european loess series: new evidences for rapid climatic warming events during the Last Glacial. Quaternary International 21, 279e280. Antoine, P., Rousseau, D.-D., Degeai, J.-P., Moine, O., Lagroix, F., Kreutzer, S.,
, C., Gauthier, C., Svoboda, J., Lisa, L., 2013. High-resolution reFuchs, M., Hatte cord of the environmental response to climatic variations during the last interglacial-glacial cycle in Central Europe: the loess-palaeosol sequence of stonice (Czech Republic). Quaternary Science Reviews 67, 17e38. Dolní Ve
risson, D., Auguste, P., Antoine, P., Goval, E., Jamet, G., Coutard, S., Moine, O., He
rin, G., Lagroix, F., Schmidt, E., Robert, V., Debenham, N., Meszner, S., Gue
quences loessiques ple
istoce ne supe
rieur d’HavrinBahain, J.-J., 2014a. Les se
oenvironnement, court (Pas-de-Calais, France): stratigraphie, pale
ochronologie et occupations pale
olithiques. Quaternaire 25 (4), 321e368. ge
oenvironnements et Antoine, P., Locht, J.-L., 2014. Chronostratigraphie, pale
olithique moyen : les donne
es du nord de la France. Les peuplements au Pale
olithique moyen? In: plaines du Nord-Ouest : carrefour de l’Europe au Pale
moire 59 de la Depaepe, P., Goval, E., Koehler, H., Locht, J.-L., dir. (Eds.), Me
te
Pre
historique Française, pp. 11e23. Socie Antoine, P., Limondin-Lozouet, N., Auguste, P., Locht, J.L., Debenham, N., Bahain, J.J., Goval, E., Fagnart, J.P., Ducrocq, T., 2014b. Quaternary geoarchaeology and prehistory: the model of the Somme valley (France) and the neighbouring regions. In: Arnaud-Fassetta, G., Carcaud, N. (Eds.), French Geoarchaeology in the
ditions 2014. 21st Century, pp. 71e86. CNRS e Bibus, E., R€ ahle, W., Wedel, J., 2002. Profilaufbau, Molluskenführung und Paral€glichkeiten des Altwürmabschnitts im Lo €ssprofil Mainz-Weiselelisierungsmo nau. Eiszeitalter und Gegenwart 51, 1e14. € sel, M., Ra €hle, W., 2007. Das jungpleistoza €ne Lo €ssprofil von Bibus, E., Frechen, M., Ko Nussloch (SW-Wand) im Aufschluss der Heidelberger Zement AG. Eiszeitalter und Gegenwart 56 (4), 227e255. Black, R.-F., 1976. Periglacial features indicative of permafrost: ice and soils wedges. Quaternary Research 6, 3e26. Blockley, S.P.E., Lane, C.S., Hardiman, M., Rasmussen, S.O., Seierstad, I.K., Steffensen, J.P., Svensson, A., Lotter, A.F., Turney, C.S.M., Ramsey, C.B., INTIMATE members, 2012. Synchronisation of palaeoenvironmental records over the last 60,000 years, and an extended INTIMATE1 event stratigraphy to 48,000 b2k. Quaternary Science Reviews 36, 2e10. Bond, G., Broecker, W., Johnsen, S., Mc Manus, J., Labeyrie, L., Jouzel, J., Bonani, G., 1993. Correlations between climate records from North Atlantic sediments and Greenland ice. Nature 365 (6442), 143e147. €tl, R., Edwards, L., Wang, X., Ha €uselmann, Ph, 2011. NALPS: a Boch, R., Cheng, H., Spo precisely dated European climate record 120e60 ka. Climate of the Past 7, 1247e1259. €ne Boenigk, W., Frechen, M., Weidenfeller, M., 1994. Die mittelund oberpleistoza Deckschichtenfolge im “Naturschutzgebeit Eiszeitlisches Loess profil” in Koblenz-Metternich. Mainzer geowiss. Mitt 23, 287e320. Boenigk, W., Frechen, M., 2001. The loess record at Koblenz-Metternich and €nchesberg in the Middle Rhine area. Quaternary international 76e77, To 201e209. Dansgaard, W., Johnsen, S.J., Clausen, H.B., Dahl-Jensen, D., Gundestrup, N.S., €rnsdottir, A.E., Jouzel, J., Hammer, C.U., Hvidberg, C.S., Steffensen, J.P., Sveinbjo Bond, G., 1993. Evidence for general Instability of past climate from a 250-kyr ice-core record. Nature 364, 218e220.

19


oenvironenments des fonds de valle
e Deschodt, L., 2014. Chronostratigraphie et pale du bassin française de l'Escaut/Chronostratigraphy and palaeoenvironments in the valley bottoms of the french Schedt river catchment. PhD Thesis. University Paris I, 2 vol., 581 pp., 356 fig.
nchez-Gon ~ i, M.F., Naughton, F., Turon, J.-L., Duprat, J., Malaize
, B., Desprat, S., Sa Peypouquet, J.-P., 2007. Climate variability of the last five isotopic interglacials: direct LandeSeaeIce correlation from the Multiproxy analysis of North
nchez Western Iberian Margin Deep-Sea Cores. In: Sirocko, F., Claussen, M., Sa ~ i, M.F., Litt, T. (Eds.), The Climate of Past Interglacials. Elsevier, pp. 375e386. Gon Engelmann, A., Frechen, M., Antoine, P., 1999. Chronostratigraphie frühweichselselzeitlicher kolluvialer Sedimente von Bettencourt-Saint-Ouen (Nord-Frank€rgeologie. reich). In: Becker-Hauman, Frechen, M. (Eds.), Terrestrische Quata Logabook, Koln, pp. 12e22. Fagnart, J.-P., 1997. La fin des temps glaciaires dans le nord de la France. Approches
ologique et environnementale des occupations humaines au cours du arche
m. Socie
te
pre
historique française 24, p. 270. Tardiglaciaire. Paris. Me Fagnart, J.-P., Coudret, P., Antoine, P., avec la collaboration de Vallin, L., Sellier, N.,
olithique supe
rieur ancien dans le Nord de la France. Masson, B., 2013. Le Pale
moire LVI de la Socie
te
pre
historique française, pp. 198e214. Me € nchesberg Frechen, M., 1994. Thermolumineszenz-Datierungen an Lossen des To aus der Osteifel. Eiszeitalter und Gegenwart 44, 79e93. Frechen, M., Boenigk, W., Weidenfeller, M., 1995. Chronostratigraphie des “Eiszei€ss profils” in Koblenz-Metternich. Mainzer Geowiss. Mitt 24, 155e180. tlichen Lo Frechen, M., 1999. Upper Pleistocene loess stratigraphy in Southern Germany. Quaternary Geochronology 18, 243e269. €, B., van den Haute, P., 2001. The Upper Pleistocene Frechen, M., van Vliet-Lanoe loess record at Harmignies/Belgium- high resolution terrestrial archive of climate forcing. Paleogeography, Paleoclimatology, Paleoecology 173, 175e195. Frechen, M., Oches, E.A., Kohfeld, K.E., 2003. Loess in Europe-mass accumulation rates during the Last Glacial Period. Quaternary Science Reviews 22, 1835e1857. Frechen, M., Schirmer, W., 2011. Luminescence chronology of the Schwalenberg II loess in the middle Rhine Valley. Eiszeitalter und Gegenwart 60 (1), 78e89. French, H.M., 2007. The Periglacial Environment, third ed. John Wiley and Sons Ltd., Chichester, U.K., p. 458 French, H.-M., Shur, Y., 2010. The principles of cryostratigraphy. Earth-Science Reviews 101, 190e206.
, C., Gauthier, C., 2007. Chronology of the Fuchs, M., Rousseau, D.D., Antoine, P., Hatte Last Climatic Cycle (Upper Pleistocene) of the Surduk loess sequence, Vojvodina, Serbia. Boreas 10, 1e8.
, C., Lagroix, F., Moine, O., Fuchs, M., Kreutzer, S., Rousseau, D.D., Antoine, P., Hatte
, L., 2013. The loess sequence of Dolní Ve stonice Gauthier, C., Svoboda, J., Lisa (Czech Republic): a new OSL based chronology of the Last Climatic Cycle. Boreas 42, 664e677. Gibbard, P.-L., 1988. The history of the great northwest European rivers during the past three million years. Philosophical Transactions of the Royal Society of London, Series B 318, 559e602. €ller, L., Fuchs, M., Lo €scher, M., Gocke, M., Hambach, U., Eckmeier, E., Schwark, L., Zo Wiesenberg, G.L.B., 2014. Introducing an improved multi-proxy approach for paleoenvironmental reconstruction of loess-paleosol archives applied on the Late Pleistocene Nussloch sequence (SW Germany). Palaeogeography, Palaeoclimatology, Palaeoecology 410, 300e315.
risson, D., Claud, E., Locht, J.L., Antoine, P., Coutard, S., 2014. Hunting Goval, E., He Camp at the end of the Middle Palaeolithic at Havrincourt «Les Bosquets» (Pasde-Calais, France). In: Delagnes, A., Conard, N. (Eds.), Settlement Dynamics of the Middle Paleolithic and Middle Stone Age IV, pp. 317e334.
istoce ne et des formes de Gullentops, F., 1954. Contribution  a la chronologie du Ple
moire de l’Institut de Ge
ologie. Universite
de Louvain relief en Belgique. Me XVIII, pp. 125e252.
csi, M., Richter, H., Altermann, M., J€ Haase, D., Fink, J., Haase, G., Ruske, R., Pe ager, K.D., 2007. Loess in Europe e its spatial distribution based on a European Loess Map, scale 1:2,500,000. Quaternary Science Reviews 26, 1301e1312. €, B., 1974. Compte-rendu de l’excursion du 25 mai 1974 Haesaerts, P., Van Vliet-Lanoe
e a  la stratigraphie des limons aux environs de Mons. Annales de la consacre
te
Ge
ologie et de Pale
ontologie 978, 547e560. Socie €, B., 1973. Evolution d’un permafrost fossile dans les Haesaerts, P., Van Vliet-Lanoe limons du Dernier Glaciaire  a Harmignies (Belgique). Bulletin de l’Association française pour l’Etude du Quaternaire 1973 (3), 151e164. €, B., 1981. Phe
nome nes pe
riglaciaires et sols fossiles Haesaerts, P., Van Vliet-Lanoe
s a  Maisie res-Canal, a  Harmignies et  observe a Rocourt. Biultyn peryglacalny 28, 291e324.
, E., Kuyl, O., Mücher, H., Roebroeks, W., 1981. Compte rendu de Haesaerts, P., Juvigne
erlandais, consacre
e l'excursion du 13 Juin 1981, en Hesbaye et au Limbourg Ne 
istoce ne supe
rieur. Annales de la a la chronostratigraphie des loess du Ple
te
Ge
ologique de Belgique 104, 223e240. Socie
istoce ne supe
rieur en Belgique. Comparaisons Haesaerts, P., 1985. Les loess du Ple
quences d'Europe centrale. Bulletin de l'Association Française pour avec les se l'Etude du Quaternaire 22e23 (2e3), 105e115. Haesaerts, P., Mestdagh, H., Bosquet, D., 1999. The sequence of Remicourt (HesbayeBelgium): new insights on the pedo and chronostratigraphy of the Rocourt Soil. Geologica Belgica 2 (3/43), 5e27. Haesaerts, P., Mestdagh, H., 2000. Pedosedimentary evolution of the last interglacial and early glacial sequence in the European loess belt from Belgium to central Russia. Geologie en Mijnbouw 79, 313e324. Haesaerts, P., Borziak, I., Chirica, V., Damblon, F., Koulakovska, L., Van der Plicht, J., 2003. The East-Carpathian loess record: a reference for the Middle

Please cite this article in press as: Antoine, P., et al., Upper Pleistocene loess-palaeosol records from Northern France in the European context: Environmental background and dating of the Middle Palaeolithic, Quaternary International (2015), http://dx.doi.org/10.1016/ j.quaint.2015.11.036

20

P. Antoine et al. / Quaternary International xxx (2015) 1e21

and Late Pleniglacial stratigraphy in Central Europe. Quaternaire 14 (3), 163e188. Haesaerts, P., Borziac, I., Chekha, V., Chirica, V., Damblon, F., Drozdov, N., Orlova, L., Pirson, S., van der Plicht, J., 2009. Climatic signature and radiocarbon chronology of Middle and Late Pleniglacial loess from Eurasia: comparison with the marine and Greenland records. Radiocarbon 51 (1), 301e318. Haesaerts, P., Borziac, I., Chekha, V.P., Chirica, V., Drozdov, N.I., Koulakovska, L., Orlova, L.A., van der Plicht, J., Damblon, F., 2010. Charcoal and wood remains for radiocarbon dating Upper Pleistocene loess sequences in Eastern Europe and Central Siberia. Palaeogeography, Palaeoclimatology, Palaeoecology 29 (1e2), 106e127.
olithique de la Sablie re Haesaerts, P., Di Monica, K., Pirson, S., 2011. Le gisement pale ge)Le Pale
olithique moyen en Belgique. Gritten  a Rocourt (province de Lie
langes Marguerite Ulrix-Closset. In: Toussaint, M., Di Modica, K., Pirson, S. Me
te
belge d'e
tudes Ge
ologiques et (Eds.), pp. 359e374. Bulletin de la Socie
ologiques «Les Chercheurs de la Wallonie», H.S. n 4 et Etudes et Arche
ologiques de l'Universite
de Liege, 128. Recherches arche
, C., Fontugne, M., Rousseau, D.D., Antoine, P., Zo €ller, L., Tisne
ra-Laborde, N., Hatte Bentaleb, I., 1998. 13C Variations of loess organic matter as a record of the vegetation response to climatic changes during the Weichselian. Geology 26 (7), 583e586.
, C., Antoine, P., Fontugne, M., Rousseau, D.-D., Tisne
ra-Laborde, N., Zo € ller, L., Hatte 1999. New chronology and organic matter d13C paleoclimatic significance of Nussloch loess sequence (Rhine Valley, Germany). Quaternary International 62, 85e91.
, C., Pessenda, L.C., Lang, A., Paterne, M., 2001. Development of accurate and Hatte reliable 14C chronologies for loess deposits: application to the loess sequence of Nussloch (Rhine Valley, Germany). Radiocarbon 43 (2B), 611e618. €tl, C., Barrett, S., Brauer, A., Drescher-Schneider, R., Gaar, D., Heiri, O., Koinig, K.A., Spo Ivy-Ochs, Kerschner, H., Luetscher, M., Moran, A., Nicolussi, K., Preusser, F., Schmidt, R., Schoeneich, P., Schworer, C., Sprafke, T., Terhorst, B., Tinner, W., 2014. Palaeoclimate records 60e8 ka in the Austrian and Swiss Alps and their forelands. Quaternary Science Reviews 106, 190e205.
risson, D., Goval, E., Antoine, P., Bahain, J.-J., Balescu, S., Chantreau, Y., Claud, E., He Coutard, S., Debenham, N., Font, C., Locht, J.-L., Pirson, S., Spagna, P., 2014. Canal Seine-Nord Europe, fouille 49, Picardie, Somme, Etricourt-Manancourt «Du
olithique infe
rieur a  moyen, les occupations d’Etricourt-Manancourt au Pale «Moulin  a Vent». Rapport de Fouille de sauvetage INRAP, 224 pp, 120 Fig. Jary, Z., Ciszek, D., 2013. Late Pleistocene loessepalaeosol sequences in Poland and western Ukraine. Quaternary International 296, 37e50.
, E., Haesaerts, P., Metsdagh, H., Balescu, S., 1996. Re
vision du stratotype Juvigne
mie des loessique de Kesselt (Limbourg, Belgique). Compte Rendu de l’Acade
rie IIa 323, 801e807. Sciences, Paris, se
Tallier, E.,
Haesaerts, P., Pirson, S., 2008. Un nouveau stratotype du
, E., Juvigne
phra de Rocourt dans la carrie re de Romont (Eben/Bassenge, Belgique). Te Quaternaire 19 (2), 133e139. Kadereit, A., Kind, C.J., Wagner, G.A., 2013. The chronological position of the Lohne Soil in the Nussloch loess section e re-evaluation for a European loess-marker horizon. Quaternary Science Reviews 59, 67e86. Kasse, C., Vandenberghe, J., Van Huissteden, J., Bohncke, S.J.P., Bos, J.A.A., 2003. Sensitivity of Weichselian fluvial systems to climate change (Nochten mine, eastern Germany). Quaternary Science Reviews 22, 2141e2156. Kreutzer, S., Fuchs, M., Meszner, S., Faust, D., 2012. OSL chronostratigraphy of a loess-palaeosol sequence in Saxony/Germany using quartz of different grain sizes. Quaternary Geochronology 10, 102e109. Kukla, G.J., 1977. Pleistocene land-sea correlations I. Europe. Earth Science Review 13, 307e374.
, C., Rousseau, D.-D., Antoine, P., Fontugne, M., Zo €ller, L., Hambach, U., Lang, A., Hatte 2003. High-resolution chronologies for loess: comparing AMS 14C and optical dating results. Quaternary Science reviews, Quaternary Geochonology 22, 953e959.
riglaciaire ple
istoce ne en Europe du Nord-Ouest et Lautridou, J.-P., 1985. Le cycle pe rement en Normandie. The se Lettres. Univ. Caen, 908 pp. plus particulie
, J., Heim, J., Puisse
gur, J.J., et Rousseau, D.D., 1985. La stratiLautridou, J.P., Somme graphie des lœss et formations fluviatiles d'Achenheim (Alsace): nouvelles
es bioclimatiques et corre
lations avec les se
quences ple
istoce nes de la donne France du Nord-Ouest. Bulletin de l’Association Française pour l’Etude du Quaternaire 22, 125e132. Lericolais, G., 1997. Evolution Plio-quaternaire du Fleuve Manche: Stratigraphie et
omorphologie d’une plateforme continentale en re
gime pe
riglaciaire. Ph.D. Ge
de Bordeaux I, 265 pp. Thesis. Universite Lisiecki, L.E., Raymo, M.E., 2005. A Pliocene-Pleistocene stack of 57 globally distributed benthic d18O records. Paleoceanography 20 (1). Locht, J.-L., Swinnen, C., Antoine, P., Auguste, P., Patou-Mathis, M., Depaepe, P., res, C., Laurent, M., Bahain, J.-J., avec la collaboration de MATHYS, P. 1995. Falgue
olithique moyen de Beauvais (Oise-France). Bulletin de la Le gisement pale
te
Pre
historique française, 92e2, 213e226. Socie
villon, S., Depaepe, P., 2002. Le Gisement Locht, J.L., Swinnen, C., Antoine, P., Re
olithique moyen de Bettencourt-Saint-Ouen (Somme). L'Acheule
en dans la pale
e de la Somme: donne
es re
centes. Publications du Centre d’Etudes et de valle
historiques 6, 199e237. Recherches Pre Locht, J.L., Antoine, P., Bahain, J.J., Dwrila, G., Raymond, P., Limondin-Lozouet, N.,
, C., Haesaerts, P., Gauthier, A., Debenham, N., Frechen, M., Rousseau, D.-D., Hatte
olithique moyen et les se
quences Metsdagh, H., 2003. Le gisement pale


istoce nes de Villiers-Adam (Val d’Oise, France): chronostratigraphie, enviple
histoire 45, 1e111. ronnement et implantations humaines. Gallia Pre res, C., Locht, J.-L., Antoine, P., Auguste, P., Bahain, J.-J., Debenham, N., Falgue
quence loessique ple
istoce ne supe
rieur de Savy Farkh, S., Tissoux, H., 2006. La se
olithiques. Qua(Aisne France): stratigraphie, datations et occupations pale ternaire 17 (3), 269e275.
olithique moyen de Fresnoy-au-Val (Somme, (Dir.) Locht, J.L., 2008. Le gisement pale se, INRAP, 129 pp, 77 fig. 10 Tabl. France), Document final de Synthe Locht, J.-L., Sellier, N., Antoine, P., Koehler, H., Debenham, N., 2013. Mauquenchy
vidence de deux niveaux d’occupation (Seine Maritime) : mise en e
olithique dans un sol gris-forestier date
du SIM 5a (De
but-glaciaire pale weichselien). Quaternaire 24, 247e257.
risson, D., Locht, J.-L., Goval, E., Antoine, P., Coutard, S., Auguste, P., Paris, C., He 2014a. Palaeoenvironments and prehistoric Interactions in Northern France from the Eemian Interglacial to the End of the Weichselian Middle Pleniglacial. In: Foulds, F.W.F., Drinkall, H.C., Perri, A.R., Clinnick, D.T.G., Walker, J.W.P. (Eds.), «Wild Things» Recent Advances in Palaeolithic and Mesolithic Research. Oxbow Books Oxford and Philadelphia, 2014, pp. 70e78. Locht, J.-L., Coutard, S., Antoine, P., Sellier, N., Ducrocq, T., Paris, C., Guerlin, O.,
es Kiefer, D., Defaux, F., Deschodt, L., Limondin-Lozouet, N., 2014b. Donne
dites sur le Quaternaire et le Pale
olithique du Nord de la France. Revue ine
ologique de Picardie (3/4), 5e70. Arche
risson, D., Goval, E., Cliquet, D., Huet, B., Coutard, S., Antoine, P., Locht, J.-L., He Feray, P., 2015. Timescales, space and culture during the Middle Palaeolithic in north-western France. Quaternary International. http://dx.doi.org/10.1016/j. quaint.2015.07.053. Mackay, J.-R., 2000. Thermally induced movements in ice-wedge polygons, western
ographie Physique et Quaternaire 54, 41e68. Arctic coast: a long-term study. Ge Marcott, S.A., Clark, P.U., Padman, L., Klinkhammer, G.P., Springer, S.R., Liu, Z., OttoBliesner, B.L., Carlson, A.E., Ungerer, A., Padman, J., He, F., Cheng, J., Schmittner, A., 2011. Ice-shelf collapse from subsurface warming as a trigger for Heinrich events. Proceedings of the National Academy of Sciences of the United States of America 108 (33), 13415e13419. €, B., 1995. Le gisement Marcy, J.L., Auguste, P., Fontugne, M., Munaut, Van Vliet-Lanoe
rien d He
nin-sur-Cojeul (Pas-de-Calais). Bulletin de la Socie
te
mouste
historique française 90 (4), 251e256. pre Martrat, B., Grimalt, J.O., Shackleton, N.J., de Abreu, L., Hutterli, M.A., Stocker, T.F., 2007. Four climate cycles of recurring deep and surface water destabilizations on the Iberian Margin. Science 317, 502e507. Matsuoka, N., 2011. Climate and material controls on periglacial soil processes: toward improving periglacial climate indicators. Quaternary Research 75, 356e365. Meijs, E.P.M., 2011. The Veldwezelt site (province of Limburg, Belgium), environmental and stratigraphical interpretations. In: Jagt, J.W.M., Jagt-Yazykova, E.A., Schins, W.J.H. (Eds.), A Tribute to the Late Felder Brothers-Pioneers of Limburg Geology and Archaeology, Netherland Journal of Geosciences, 90 (2e3), pp. 73e94. Meijs, E.P.M., 2002. Loess stratigraphy in Dutch and Belgian Limburg. Eiszeitalter und Gegenwart 51, 114e130. Meszner, S., Fuchs, M., Faust, D., 2011. Loess-paleosol-sequences from the loess area of saxony (Germany). E and G. Quaternary Science Journal 60, 47e65.
, C., 2002. Mise en e
vidence Moine, O., Rousseau, D.-D., Antoine, P., Hatte
ve
nements climatiques rapides par les faunes de mollusques terrestres des d’e loess weichseliens de Nussloch (Allemagne). Quaternaire 13 (3e4), 209e218. Moine, O., Rousseau, D.-D., Antoine, P., 2008. Abrupt malacological and lithological changes along the Upper Weichselian loess sequence of Nussloch (Rhine Valley, Germany). Quaternary Research 70, 91e104. Moine, O., Antoine, P., Deschodt, L., Sellier, N., 2011. Enregistrements mala
solution dans les loess et les gleys de toundra du ple
nicologiques  a haute re
rieur: premiers exemples du nord de la France. glaciaire weichselien supe Quaternaire 22 (4), 307e325. Moreno, A., Svensson, A., Brooks, S.J., Connor, S., Engels, S., Fletcher, W., Genty, D., Heiri, O., Labuhn, I., Persoiu, A., Peyron, O., Sadori, L., Valero-Garces, B., Wulf, S., Zanchetta, G., data contributors, 2014. A compilation of Western European terrestrial records 60e8 ka BP: towards an understanding of latitudinal climatic gradients. Quaternary Science Reviews 106, 167e185. Müller, U.C., Pross, J., Bibus, E., 2003. Vegetation response to rapid climate change in central Europe during the past 140,000 yr based on evidence from the Füramos pollen record. Quaternary Research 59, 235e245. Murton, J.B., 2007. Periglacial landforms/ice wedges and ice-wedges casts. In: Elias, S.A. (Ed.), Encyclopedia of Quaternary Science. Elsevier, pp. 2153e2170. Members, N.G.R.I.P., 2004. High resolution climate record of the Northern Hemisphere reaching into the last glacial interglacial period. Nature 431, 147e151.
Frechen, M., Horvath, E., Wacha, L., Rolf, C., 2011. Investigating the Novothny, A., penultimate and last glacial cycles of the Sütto loess section (Hungary) using luminescence dating, high-resolution grain size, and magnetic susceptibility data. Quaternary International 234 (1e2), 75e85.
, J., 1970. Les lœss et la stratigraphie du Ple
istoce ne re
cent dans Paepe, R., et Somme
te
le Nord de la France et en Belgique, pp. 191e201. Annales de la Socie
ologique du Nord, XC. Ge Paris, C., Fagnart, J.P., Coudret, P., 2013. Du Gravettien final dans le Nord de la France
es a  Amiens-Renancourt (Somme, France). Bulletin de la ? Nouvelles donne
te
pre
historique française 110 (1), 123e126. Socie

Please cite this article in press as: Antoine, P., et al., Upper Pleistocene loess-palaeosol records from Northern France in the European context: Environmental background and dating of the Middle Palaeolithic, Quaternary International (2015), http://dx.doi.org/10.1016/ j.quaint.2015.11.036

P. Antoine et al. / Quaternary International xxx (2015) 1e21
olithique supe
rieur ancien dans le dir. Paris, C., 2015. Renancourt 2: un site du Pale
ration, mars 2015. INRAP Nord de la France. Fouille 2012. Rapport final d'ope Nord-Picardie, 250 pp.
omorphologie Pe
riglaciaire e Texte des leçons de la Chaire Pissart, A., 1987. Ge
omorphologie et Ge
ologie du Francqu belge 1987. Edition du laboratoire de Ge
de Lie ge, 135 p. Quaternaire de l'Universite Ran, E.T.H., Bohncke, S.J.P., Van Huissteden, K.J., Vandenberghe, J., 1990. Evidence of episodic permafrost conditions during the Weichselian Middle Pleniglacial in the Hengelo Basin (The Netherlands). Geologie en Mijnbouw 69, 207e218. Rasmussen, S.O., Bigler, M., Blockley, S.P., Blunier, T., Buchardt, S.L., Clausen, H.B., Cvijanovic, I., Dahl-Jensen, D., Johnsen, S.J., Fischer, H., Gkinis, V., Guillevic, M., Hoek, W.Z., Lowe, J., Pedro, J.B., Popp, T., Seierstad, I.K., Steffensen, J.P., Svensson, A.M., Vallelonga, P., Vinther, B.M., Walker, M.J.C., Wheatley, J.J., Winstrupa, M., 2014. A stratigraphic framework for abrupt climatic changes during the Last Glacial period based on three synchronized Greenland ice-core records: refining and extending the INTIMATE event stratigraphy. Quaternary Science Reviews 106, 14e28. Rioual, P., Andrieu-Ponel, V., Rietti-Shati, M., Battarbee, R.W., de beaulieu, J.L., Cheddadi, R., Reille, M., Svobodova, H., Shemesh, A., 2001. High-resolution record of climate stability in France during the last interglacial period. Nature 413, 293e296.
, C., Lang, A., Zo €ller, L., Fontugne, M., Ben Rousseau, D.-D., Antoine, P., Hatte Othman, D., Luck, J.M., Moine, O., Labonne, M., Bentaleb, I., Jolly, D., 2002. Abrupt millennial climatic changes from Nussloch (Rhine Valley, Germany) eolian (loess) records during the Last Glaciation. Quaternary Science Reviews 21, 1577e1582. €ller, L., Hradilova, J., 1998. Early Weichselian dust storm Rousseau, D.-D., Kukla, G., Zo layer at Achenheim in Alsace, France. Boreas 27 (3), 200e207.
, C., Lang, A., Zo €ller, L., 2007. Link beRousseau, D.-D., Sima, A., Antoine, P., Hatte tween European and North-Atlantic abrupt climate changes over the last glaciation. Geophysical Research Letters 34, 22.
nchez Goni, M.F., Eynaud, F., Turon, J.L., Shackleton, N.J., 1999. High resolution Sa palynological record off the Iberian margin: direct land-sea correlation for the Last Interglacial complex. Earth and Planetary Science Letters 171, 123e137. ~ i, M.F., Cacho, I., Turon, J.-L., Guiot, J., Sierro, F.J., Peypouquet, J.-P., S
anchez Gon Grimalt, J., Shackleton, N.J., 2002. Synchronicity between marine and terrestrial responses to millenial scale climatic variability during the last glacial period in the Mediterranean region. Climate Dynamics 19, 95e105. S
anchez Goni, M.F., Landais, A., Fletcher, William J., Naughton, F., Desprat, S., Duprat, J., 2008. Contrasting impacts of Dansgaard-Oeschger events over a western European latitudinal transect modulated by orbital parameters. Quaternary Science Reviews 27 (11), 1136e1151. Schirmer, W., 2000a. Rhein loess, ice cores and deep-sea cores during MIS 2e5. Zeitschrift der Deutschen Geologischen Gesellschaft 151 (3), 309e332. €ns aus dem rheinischen Schirmer, W., 2000b. Eine Klimakurve des Oberpleistoza €ss. Eiszeitalter und Gegenwart 50, 25e49. Lo Schirmer, W., 2010. Interglacial complex and soil complex. Central European Journal of Geosciences 2 (1), 32e40. Schirmer, W., 2011. Rhine loess at Schwalbenberg II MIS 4 and 3. Eiszeitalter und Gegenwart 61 (1), 32e47. Schirmer, W., 2015. Late Pleistocene loess of the Lower Rhine. Quaternary international (in press). Schmidt, E.D., Frechen, M., Murray, A.S., Tsukamoto, S., Bittman, F., 2011. Lumines€nchesberg section: a comcence chronology of the loess record from the To parison of using quartz and feldspar as dosimeter to extend the age range beyond the Eemian. Quaternary International 234, 10e22. €nhals, E., Rohdenburg, H., Semmel, A., 1964. Ergebnisse neuerer UntersuScho € ss-Gliederung in Hessen. Eiszeitalter und Gegenwart 15, chungen zur Würmlo 199e206. Scourse, J.D., Haapaniemi, A.I., Colmenero-Hidalgo, E., Peck, V.L., Hall, I.R., Austin, W.E.N., Knutz, P.C., Zahn, R., 2009. Growth, dynamics and deglaciation of the last British-Irish ice sheet: the deep-sea ice-rafted detritus record. Quaternary Science Reviews 28, 3066e3084. Seelos, K., Sirocko, F., Dietrich, S., 2009. A continuous high-resolution dust record for the reconstruction of wind systems in central Europe (Eifel, Western Germany) over the past 133 ka. Geophysical Research Letters 36. http://dx.doi.org/ 10.1029/2009GL039716.
es re
centes sur le Pale
olithique moyen de l'Aisne. Sellier, N., Coutard, S., 2007. Donne  Courmelles. Revue Arche
ologique de Une occupation du Weichselien ancien a Picardie 3/4, 5e16.

21

€ ßgliederung im Rhein-Main-Gebiet. e Semmel, A., 1969. Bemerkungen zur Würmlo Notizbl. hess. L-Amt Bodenforsch. 97, Wiesbaden, pp. 395e399. Sima, A., Rousseau, D.-D., Kageyama, M., Ramstein, G., Schulz, M., Balkanski, Y.,
, C., 2009. Imprint of North-Atlantic abrupt climate Antoine, P., Dulac, F., Hatte changes on western European loess deposits as viewed in a dust emission model. Quaternary Science Reviews 28 (25), 2851e2866.
, J., Paepe, R., Lautridou, J.-P., 1980. Principes, me
thodes et syste me de la Somme stratigraphie du Quaternaire dans le Nord-Ouest de la France et la Belgique. In: mes de stratigraphie quaternaire en France et dans les pays limitrophes, Proble
ment au Bulletin de l’Association Française pour l’Etude du pp. 148e162. Supple Quaternaire NS-1.
, J., Lautridou, J.P., Heim, J., Maucorps, J., Puisse
gur, J.-J., Rousseau, D.-D., Somme
venin, A., Van Vliet-Lanoe €, B., 1986. Le cycle climatique du Ple
istoce ne The
rieur dans les loess d'Alsace 
ment au Bulletin de supe a Achenheim. Supple l’Association Française pour l’Etude du Quaternaire 23, pp. 97e104.
, J., Munaut, A.V., Etmontspohl, A.-F., Limondin, N., Lefebvre, D., CunatSomme
, N., Mouthon, J., Gilot, E., 1994. The Watten boring: an Early Weichselian Boge and Holocene climatic and paleoecological record from the French North-Sea coastal plain. Boreas 23, 231e243. Terhorst, B., Kühn, P., Damm, B., Hambach, U., Meyer-Heintze, S., Sedov, S., 2014. Paleoenvironmental fluctuations as recorded in the loess-paleosol sequence of the Upper Paleolithic site Krems-Wachtberg. Quaternary International 351, 67e82. Terhorst, B., Sedov, S., Sprafke, T., Peticzka, R., Meyer-Heintze, S., Kühn, P., Solleiro Rebolledo, E., 2015. Austrian MIS 3/2 loessepalaeosol records-Key sites along a westeeast transect. Palaeogeography, Palaeoclimatology, Palaeoecology 418, 43e56. res, C., Bahain, J.J., Tissoux, H., Valladas, H., Voinchet, P., Reyss, J.L., Mercier, N., Falgue €ller, L., Antoine, P., 2009. OSL and ESR studies of aeolian quartz sediments Zo from the Upper Pleistocene loess sequence of Nussloch (Germany). Quaternary Geology, Special Issue LED 5 (2e3), 131e136.
olithique infe
rieur et moyen du Nord de la France (Nord, Tuffreau, A., 1987. Le Pale se de doctorat Pas-de-Calais, Picardie) dans son cadre stratigraphique. The d'Etat. Univ. Sciences et Techniques de Lille Flandres-Artois, p. 609. Vandenberghe, J., Huijzer, B., Mücher, H., Laan, W., 1998. Short climatic oscillations in a western European loess sequence (Kesselt,Belgium). Journal of Quaternary Science 13, 471e485. Vandenberghe, J., Nugteren, G., 2001. Abrupt climatic changes recorded in loess sequences. Global and Planetary Change 28, 1e9. Van den haute, P., Vancraeynest, L., De Corte, F., 1998. The Late Pleistocene loess deposits and palaeosols of eastern Belgium: new TL age determinations. Journal of Quaternary Science 13 (5), 487e497. €, B., 1987. Le ro ^le de la glace de se
gre
gation dans les formations Van Vliet-Lanoe se de Doctorat d’Etat. Universite
superficielles de l'Europe du Nord-Ouest. The Paris I, p. 864. €, B., 1990. Le Pe
docomplexe de Warneton. Ou en est-on?, Bilan Van Vliet-Lanoe
ope
dologique et micro-morphologique. Quaternaire 1, 65e76. pale €, B., 1992. Le niveau  re de la Van Vliet-Lanoe a langues de Kesselt, horizon repe
rieur
en: stratigraphie du Weichselien supe europe signification
oenvironnementale et pale
oclimatique. Me
moires de la Socie
te
Ge
ologique pale de France 160, 35e44. €, B., Laurent, M., Everaerts, M., Mansy, J.L., Manby, G., 2000. EvoVan Vliet-Lanoe
oge ne et quaternaire de la Somme: une flexuration tectonique active. lution ne
mie des Sciences. Sciences de la Terre et des Plane tes Comptes rendus de l'Acade 331, 151e158. Woillard, G.M., 1979. The Last Interglacial-Glacial Cycle at Grande Pile in north
te
Belge de Ge
ologie 88, 51e69. eastern France. Bulletin de la Socie Woillard, G.M., 1978. Grande Pile Peat Bog: a continuous pollen record for the last 140,000 years. Quaternary Research 9, 1e21. Zagwijn, W., Paepe, R., 1968. Die Stratigraphie der weichselzeitlichen Ablagerungen der Niederlande und Belgiens. Eiszeitalter und Gegenwart 19, 129e146. €ller, L., Stremme, H.E., Wagner, G.A., 1988. Thermolumineszenz-Datierung an Zo € ss-Pal€ Lo aoboden-Sequenzen von Nieder-, Mittle-und Obberhein. Chemical Geology, Isot. Geosc. Sect 73, 39e62. €ller, L., Conard, N.J., Hahn, J., 1991. Thermoluminescence dating of Middle Zo Palaeolithic open air sites in the Middle Rhine Valley-Germany. Naturwissenschaften 78, 408e410. €ller, L., Wagner, G., 1990. Thermoluminescence dating of loess-recent deZo velopments. Quaternary International 7/8, 119e128.

Please cite this article in press as: Antoine, P., et al., Upper Pleistocene loess-palaeosol records from Northern France in the European context: Environmental background and dating of the Middle Palaeolithic, Quaternary International (2015), http://dx.doi.org/10.1016/ j.quaint.2015.11.036