Version classiqueVersion mobile

La géoarchéologie française au xxie siècle

Nathalie Carcaud
Gilles Arnaud-Fassetta

Partie II. Les hydrosystèmes fluviaux, entre climat et anthropisation/Section 2. Fluvial hydrosystems, between climate and “anthropogenic processes”

Chapter 5. Quaternary geoarchaeology and prehistory

The model of the Somme valley (France) and the neighbouring regions

Pierre Antoine, Jean-Luc Locht, Nicole Limondin-Lozouët, Patrick Auguste, Jean-Jacques Bahain, Emilie Goval, Jean-Pierre Fagnart, Nick Debenham et Thierry Ducrocq

Texte intégral


Regional geological and geomorphological context

1The Somme basin is a hydrographic basin extending over a limited surface with a pre-Quaternary substratum made up of Upper Cretaceous chalk. The chalk is generally homogeneous and presents high porosity, which is one of the original aspects of the studied sector. The main outlines of the hydrographic network are guided by tectonic structures following a northwest/southeast axis, such as the Somme syncline. During the glacial periods, the palaeo-Somme valley was one of the main tributaries of the Channel River flowing into the Atlantic more than 1000 km to the west (Gibbard, 1994; Lericolais, 1997; Gibbard, 2007; Fig. 1). During most of the Quaternary, with a sea level as much as-120 m below the present level, the area presently corresponding to the Middle Somme Valley (area of Amiens) was located in the upstream course of the palaeo-Somme valley.

2Researches carried out these past years on the Quaternary formations of the Somme basin (fluvial terraces and loess), and on the interactions between human settlement and the evolution of the environment, are based on an interdisciplinary approach to Quaternary sequences and associated Palaeolithic sites. The emergence of new dating techniques has had a notable impact on the consolidation of the geochronological framework: Electron Spin Resonance on fluvial quartz (ESR), combined ESR/Uranium-Thorium on dental enamel (ESR/U-Th), Uranium Thorium on calcite (U/Th), Thermoluminescence (TL), Optically Stimulated Luminescence (OSL) on aeolian sediments, and TL on archaeologically heated flints. In parallel, the palaeoclimatic interpretation of the Pleistocene sequences containing Palaeolithic levels was refined by the contribution of new bioclimatic (malacology, palynology, plant macro-fossils, beetles and mammals) and sedimentological data obtained from the fluvial and loessic environments (granulometry, geochemistry, isotopic analysis of organic carbon (δ13C), magnetic susceptibility, palaeopedology, micromorphology).

History of fieldwork and research context

3Ever since the earliest prehistoric studies, the Somme valley has been a key sector in scientific debates focusing on the first European human occupation in Europe and the cultural and palaeoanthropological evolution of the first inhabitants of our continent (Antoine et al., 2003a, 2011). The archaeological importance of the Somme valley is mainly due to the fact that the geological, chronological, and palaeoenvironmental context of the region is among the most clearly established in the world over the past million years. The ongoing construction of this framework, currently carried out in close correlation with rescue archaeology, forms a solid basis for tackling the interactions between Palaeolithic occupations and the climatic and environmental parameters of the past 500 ka.

Fig. 1. The system of fluvial palaeochannels under the Channel (after Auffret et al., 1982; Lericolais, 1997) and loess extension in northwest France (after Antoine et al., 2003).

4Following the discoveries made by Boucher de Perthes (1849, 1864), which led to the invention of the term “archaeo-geology” to designate what is currently called “geoarchaeology”, and their “validation” by Prestwich (1859), the pioneering work of Ladrière (1890), Commont (1909, 1911, 1912), Breuil and Koslowski (1931), and Bordes (1954) resulted in the publication of detailed stratigraphies and the first schemas illustrating the position of the different Palaeolithic industries within the Somme Quaternary formations. The first recognised multidisciplinary research on the Quaternary and the Palaeolithic in the Somme basin was conducted by Bourdier (1969, 1974) and his collaborators. This work developed during the 1970s and the 1980s with the launch of fieldwork to re-evaluate Acheulean sites in the Somme valley (Dupuis et al., 1977; Tuffreau et al., 1981; Sommé et al., 1984). In parallel, a new geological and stratigraphic approach led by Haesaerts and Dupuis (1986) resulted in a lithostratigraphic overview of the whole terraces system and the first correlation attempts with global palaeoclimatic records. During the early 1990s, a prospection-survey programme of the Somme basin led to a total overhaul of the morphostratigraphic data (Antoine, 1990). The interdisciplinary research carried out since then indicates that the Middle Somme terrace system represents a model of response to large-scale climatic cyclicity (Milankovitch cycles) in a context of slow uplift (Antoine, 1994; Antoine et al., 1998, 2000, 2007).

5During the 1990s and the 2000s, research advanced on an unprecedented scale in relation with major development works and the rapid expansion of rescue archaeology, bringing about close research ties between AFAN-INRAP, the CNRS, the University, and the National Natural History Museum. This led to the discovery of many Palaeolithic sites in primary position, particularly for the Middle Palaeolithic in loess-palaeosols contexts (Antoine et al., 1999; Locht, 2002; Antoine et al., 2003 a and b; Locht et al., 2003). These results now form a unique database in Europe, particularly for the last climatic cycle (the past 125 ka) and have allowed us to reconstruct the geological history of the Somme valley over the course of the last million years in conjunction with the archaeological evidences. Lastly, research conducted in the Somme valley is characterised by continuous scientific monitoring by the different participants who contribute to the corpus of reference data on the Quaternary and prehistory in this valley.

Geoarchaeological approach to the prehistoric sites

6Three main types of morpho-sedimentary contexts, which incorporate the different prehistoric sites, have been distinguished on the basis of the work carried out over the past twenty years in the Somme basin (Fig. 2).

Plateau context

7The sedimentary information present on the plateaus is generally poor and is characterised by erosional processes and low aeolian sedimentation rates leading to significant gaps in the sequences, which detract from their geoarchaeological interest. Nonetheless, Palaeolithic sites (Acheulean and Middle Palaeolithic) have been preserved in loess sequences and palaeosols in dolines or sinkholes resulting from the dissolution of the chalky substratum during interglacial and early-glacial periods (depth: ~ 6-8 m). On a smaller scale, and especially in the east and southeast of the basin, a few sites were discovered in association with residual mounds of Tertiary sands (Thanetian sands). These sites are generally preserved at the surface or within aeolian sands remobilised from nearby Tertiary formations and can yield in-situ large faunal remains, as at Beauvais (Locht, 2004).

Slope context

8The loess formations covering the slopes, which are generally oriented towards the northeast (asymmetrical valleys and prevailing northwest winds), can reach a thickness of 6 to 8 m in the Somme basin and up to 10-12 m in the east of the area and in the south of Pas-de-Calais. They were mainly deposited during the second half of the Saalian (MIS 6), then especially during the Last glacial between ~ 30 and 17 ka (MIS 3-2). The majority of the Middle Palaeolithic sites were discovered on these silty slopes in association with humic soil complexes (grey forest and steppe soils) contemporaneous with the beginning of the Weichselian Early-glacial (~112-70 ka). Nonetheless, over the past few years, several rescue archaeology operations have brought to light evidence of sites from the most recent phases of the recent Middle Palaeolithic preserved in the complex of Middle Pleniglacial soils (~55-35 ka). More recently, early Upper Palaeolithic occupations (~33-35 ka) have been revealed during the rescue archaeological excavations on the future Seine-Nord-Europe canal project at Havrincourt (Goval et al., 2013).

Fig. 2. The different morpho-sedimentary contexts of prehistoric sites in the Somme basin.

Fluvial context: Valleys, terraces and alluvial plains

9The oldest sites (Acheulean) and some of the early Middle Palaeolithic sites are preserved in terrace contexts (Antoine et al., 2010), in association with ine textured carbonate-rich luvial formations at the top of alluvial sequences (silts/sandy silts). he bestpreserved sites are oten associated with sequences of carbonate-rich fluvial silts or interglacial calcareous tufas, like at Caours for the last interglacial (Eemian). The coarse formations (gravels and sands), which represent the main unit of the diferent terrace formations, generally contain artefacts in secondary position, strongly rolled and reworked by periglacial activity and fluvial dynamics (braided channels). Many of the patinated and rolled handaxes lining prehistory museum showcases were collected from these formations, especially at the beginning of the 20th c. he valley loor sequence is contemporaneous with the last climatic cycle (Weichselian-Holocene) and has yielded many inal Palaeolithic (Fagnart and Coudret, 2000) and Mesolithic sites (Ducrocq et al., 1991; Ducrocq, 1999), as well as several Neolithic occupations, which were extremely well-preserved by the extensive development of Holocene peats, which represents one of the main characteristics of the present day Somme Valley. hese sites are generally located at the basal concavity marking the contact between the lower slope and the alluvial plain (or a main channel). his contact zone evolved towards the slope throughout the Holocene due to rising water levels promoting aggradational and transgressive organic luvial sedimentation, in balance with the alluvial aquifer supplied by the chalk bedrock (Antoine et al., 1997).

The fluvial formations of the Somme basin, terrace systems and valley loor formations

The stepped terrace system with loessic cover: Response model to cyclical Quaternary climatic variations and the context of the first human occupations (Acheulean)

10The research conducted on the fossil luvial formations of the Somme valley has brought to light the impact of cyclical climatic modiications on luvial sedimentation and morphology and, in particular, that of the main 100 ka cycles since the end of the Lower Pleistocene, about 1 Ma ago (Antoine et al., 2007). Besides, detailed studies of the valley floor sequences deposited during the Lateglacial and the beginning of the Holocene, have yielded a fundamental model for the interpretation of Pleistocene sequences from terrace systems, which often suffer of sedimentary hiatus and less accurate geochronological reference points (Antoine et al., 1997, 2000a, 2003a; Auguste, 2009; Limondin-Lozouet, 2011).

11Research concerning the impact of climatic, tectonic, and eustatic forcings on fluvial environments allows us to insert the numerous sites of the Somme basin and the neighbouring regions into a common chronological framework. It is therefore possible (i) to advance a detailed history of valley evolution and, (ii) to address the relationships between hominins including modern humans and their environments (chronoclimatic framework, palaeo-topography, raw material sources, etc.). This work has resulted in the chronostratigraphic calibration and the reconstruction of the palaeoenvironment of many Palaeolithic and Mesolithic occupations preserved in the alluvial sequences of terraces and valley floors. It is based on a stratigraphic approach of the archaeological sequences completed by the development of large transects fundamental for (i) situating the archaeological occupations in a complete scheme of the evolution of the valley and (ii) constructing a regional chronostratigraphic reference (Fagnart, 1997; Antoine et al., 2003a, 2007).

The main features of the Pleistocene terrace system

12The fossil fluvial terrace system of the Somme (Fig. 3) is particularly well represented in the middle valley between Amiens and Abbeville where it is made up of a complex of stepped alluvial units with welldeveloped loess-palaeosol cover (Sommé et al., 1984; Haesaerts and Dupuis, 1986; Antoine et al., 1998, 2000b, 2007). In this system, each alluvial unit has been defined as a stratigraphic formation, characterised by its fluvial facies, relative altitude in relation to maximal valley depth and delimited by two major erosive contacts: (i) with the bedrock (chalk) at the base and (ii) with the overlying slope formations at the top (Antoine et al., 1990).

13Mapping shows that the different alluvial sheet accumulations are markedly asymmetrical in relation to the present valley outline and that they are mainly located at confluences with adjacent valleys due to preferential aggradation of gravely alluvial deposits and lateral migration of the system. This terrace system is made up of a complex of 10 stepped alluvial formations between + 5/6 m and + 55 m relative altitude in relation to the maximal cutting of the present valley (Fig. 3). A summary of the data shows that each alluvial formation corresponds to the morpho-sedimentary outcome of a glacial-interglacial cycle and is characterised by the following sequence:

  1. Early-glacial: Slope deposits with fluvial interstratifications preserved in places beside the outer terrace edge (i.e., beside the slope).
  2. Pleniglacial: Gravels and coarse fluvial sands corresponding to a system of periglacial-type braided channels making up most of the alluvial formation. In several cases, this gravely unit displays a lateral transition with heterogeneous solifluction deposits (clayey sandy gravels), incorporating very large blocks (several tons) of Tertiary sandstone, reworked from the slopes by frost creep and possibly transported on ice rafts during breakup. Lastly, palynological data obtained from interstratified lenses of calcareous silts indicate a very open landscape, in concordance with the periglacial nature of the deposits (Munaut, 1988).
  3. Lateglacial-interglacial: Fine carbonate-rich fluvial silts from the end of the sequence, deposited in a fluvial system with meanders (and floods) and capped in places by small organic alluvial plain soils, or more rarely by lenticular calcareous tufa deposits. The formation of these tufas is supplied by the resurgence of water from lower slope springs or directly from fluvial valley channels, rich in calcium carbonate from the chalk aquifer.

14The carbonate-rich silty sedimentation from the end of the alluvial sequence generally corresponds to the beginning or the end of interglacials, and with periods of flood intensification. The interglacial optimum is only recorded in places in calcareous tufa formations, like those of Longpré-les-Corps-Saints, Caours and Saint-Acheul (Limondin-Lozouet and Antoine, 2006). These deposits and the associated human occupations are rarely preserved due to their high susceptibility to erosional processes.

15This model, in keeping with data from the valley floor (one glacial-interglacial cycle), compels us to underline the discontinuous aspect of fluvial sedimentary terrace records. Indeed, for a given glacial-interglacial cycle, the sedimentary build-up represented by the different units constituting the alluvial sequence only represents a timescale of several thousand years and thus contains considerable gaps. Generally speaking, the data synopsis (fluvial/slope) shows that the formation of the terrace system results from a climatically controlled cyclic morpho-sedimentary evolution in a context of general upliftof the edges of the Paris basin over the past million years. The global outcome of upliftdeduced from these data is 55 m in a million years, which is comparable to upliftin the Seine or Yonne valleys (Antoine et al., 2007).

16In the downstream part of the present system (Basse-Somme), there are more gaps in the record that probably reflect the effects of recent tectonics (Beun and Broquet, 1980; Antoine et al., 2000b). The geometry of alluvial formations from this area is thus largely based on ancient data (Commont, 1910, 1911; Agache et al., 1963; Dupuis et al., 1977). However, older alluvial evidence (Lower Pleistocene: from 1.3 to 1.8 Ma) seems to be conserved in this zone, according to ESR dates on quartz and palaeomagnetic data from the Mont-Pillard (relative altitude: + 70 m) and the Bois des Prèles alluvial formations (+60 m; Laurent et al., 1994). It is likely that valley morphology was already characterised by the presence of alluvial periglacial-type formations at that time and by a braided channel flowing in a weakly incised system concentrated in the Somme syncline, in the vicinity of the present valley. Indeed, these ancient alluvial units, often referred to as “plateau gravels”, are only present at a relatively short distance from the valley (1 to 2 km) and below the chalky plateau (10 to 15 m). In the absence of detailed geometric data, it is difficult to accurately describe the Lower Pleistocene situation. Nonetheless, data obtained recently from the Amiens region appear to reveal an accentuation of incision between 0.9 and 1 Ma, before the beginning of the Middle Pleistocene, as also observed in the Seine valley (Lefèbvre et al., 1994) and in most of the fluvial systems in northwest Europe (Veldkamp and Van den Berg, 1993; Bridgland, 2000). This situation could result from more marked climatic contrasts from MIS 22 onwards and/or an acceleration of the uplift rate.

Fig. 3. The Middle Somme Valley terrace system: stepped fluvial sequences (alluvial formations), silty cover (loess and palaeosols), dates and location of the main Palaeolithic cultures (after Antoine et al., 2003, 2007).

The model of the valley floor sequence (Last glacial-interglacial cycle): Geoarchaeology of occupations from the final Palaeolithic to the Mesolithic (the “last hunters”)

17The interdisciplinary research carried out over the past fifteen years on the present valley floor of the Somme River and its main tributaries, such as the Selle River, has brought to light the main phases of environmental evolution during the Lateglacial and the Holocene since 13 ka and has enabled us to correlate them with global climatic variations (Limondin, 1995; Antoine, 1997; Limondin-Lozouet, 1997; Antoine et al., 2000a; Fagnart and Coudret, 2000; Limondin-Lozouet and Antoine, 2001; Ponel et al., 2005). This work indicates a progressive modification of the morphology of fluvial systems from a braided periglacial system to a meandering system initiated during the Lateglacial, and provides a model for the interpretation of terrace sequences. The main phases of this evolution can be summarised by the following succession (dates are indicated in non-calibrated 14C BP years; Fig. 4):

  • First Lateglacial climatic improvement, the Bølling oscillation (~12.6-12.1 ka BP): This period is marked by an incision and an evolution of the fluvial system towards a transitional system with stable channels and through the development of shrub vegetation with birch and willow, but still in a very open environment. The progressive re-colonisation of the environments by vegetation from 12.3 ka BP onwards is confirmed by the arrival of pioneer molluscs, then by the development and rapid diversification of malacofauna. Entomological assemblages indicate average summer temperatures similar to present temperatures (15 to 16 °C). The red deer appears for the first time since the preceding interglacial (i.e., the Eemian) some 123 ka ago. The oldest human occupations in the Somme basin are contemporaneous with this phase, at approximately 12.2 ka BP (Federmesser groups).
  • Older Dryas rapid climatic deterioration (Dryas II-Older Dryas, ~ 12.1-11.9 ka BP): This short climatic deterioration is marked by carbonate-rich silt inputs derived from the erosion of the chalk outcropping on valley slopes, the retreat of tree taxa in favour of grasses (Poaceae and Cyperaceae), the presence of several species of Boreo-alpine molluscs, and a general decrease in number of shells.
  • Allerød oscillation (~11.9-10.9 ka BP): During the course of the Allerød, oscillation the climate improves again and the fluvial system evolves towards a wide meandering pattern. The alluvial plain is progressively filled with organic overbank silts, which preserve the Federmesser type occupations. Vegetation develops with the extension of birch, then of pine. Malacofauna and mammal assemblages diversify whereas entomological data remain stable. Local bioproxies (molluscs and beetles) indicate the development of open environments on the valley floor. Forest coverage remains relatively sparse and ground vegetation is not uniformly dense. In parallel, some slope instability persists, marked by colluvial sedimentation and the formation of cumulative soils at slope bases (up-building soils). The highest intensity of human occupation for the valley floor is recorded during this period.
  • Last Lateglacial climatic deterioration, the Younger Dryas (Dryas III; ~ 10.9-10 ka BP): During the Younger Dryas, a significant reactivation of erosion processes on the slopes (freeze-thaw) and an increase in the rhythm and the intensity of floods result in the filling of the valley by carbonate-rich silts. The malacofauna reveals a marshy environment marked by a strong decrease in biotope diversity for the first phase, then by an episode of malacological re-colonisation, indicating new ground vegetation expansion and therefore a reduction in flood frequency. Several palynological records obtained at the base of the carbonate-rich silts point to a clear regression of the pine and renewed Poaceae, Artemesia, and other Asteraceae expansion. This floodplain was deserted by human groups. Valley floor human occupations only became frequent at the end of this period of repeated alluvial inputs, during the final Palaeolithic, between 10.2 and 9.8 ka BP (Epi-Ahrensburgian or Belloisian).
  • Beginning of the Holocene (Preboreal, ~ 10-9 ka BP): The abrupt Preboreal warming is reflected by an incision of the Somme River that develops a single channel with meanders and attains its maximal dimensions at this time (~100 m, 5 times wider than the present channel in the Selle valley). The development of tree vegetation and soils then provokes a rapid increase in evapotranspiration, which reduces the water balance and leads to a significant decrease in flows due to the reduction of channel size and the peaty infilling of most of the valley. This dynamics leads to a relative drying of the floodplain and entails the disappearance of the large meandering system from the end of the Preboreal onwards, towards 9 ka BP, and significant peat development. This warming is characterised by a strong change in the structure of malacological populations; the open ground species, which dominated during the Lateglacial, regress sharply whereas mesophilic taxa develop and the first thermophilous taxa appear. The fundamental change induced by this evolution towards interglacial conditions has strong repercussions on the organisation of prehistoric lifestyles. In the final Palaeolithic open-plain game-hunters with lithic production characterised by large blades, give way to Mesolithic hunter-gatherer groups with microlithic arrow points, which are more adapted to forest biotope exploitation (bow and arrow hunting).

The Middle Palaeolithic site of Caours: A unique example of Palaeolithic occupation in an interglacial context (Eemian)

18The calcareous tufa sequence at Caours was initially mentioned by Léon Auffrère, Henri Breuil, and Étienne Patte, in the 1950s-1960s. Their observations concerned a now disused quarry, which had yielded Cervid remains at the time, associated with “Levalloisian” knapped flint (Agache et al., 1963). Bourdier (1969, 1974) swiftly recognised probable evidence of the Riss-Würm interglacial. It is on the basis of these early discoveries that research was resumed in Caours in 2003 by an interdisciplinary team in the context of the SITEP programme (Eclipse II INSU-CNRS). This research resulted in the reconstruction of the complete stratigraphy of the formation, the confirmation of the chronological attribution of the sequence to the last Interglacial, the reconstruction of the palaeoenvironmental dynamics of this period and the discovery of several Palaeolithic levels associated with faunal remains (Antoine et al., 2006). Excavation supervised by Jean-Luc Locht were then conducted for six years as part of a CNRS-INRAP collaboration.

19The Caours tufa formation and the underlying fluvial levels are spread over a surface of several thousand square metres at the confluence of the Scardon and the Drucat streams, with an average thickness of 3 to 4 m. This formation represents an exceptionally well-preserved complex in comparison to the rare Pleistocene tufas known in the Somme valley, which generally only extend over a few square metres. Along a continuous transect of 170 m, a unique juxtaposition of tufa deposited in a fluvial valley floor context and tufa supplied by the activity of springs at the chalky slope base during wetter interglacial periods can be observed. The tufa lies on a Periglacial type alluvial formation corresponding to the penultimate stage of coarse infill of the valley (Étouvie Formation; Fig. 3), at a relative altitude of + 6 m in relation to present valley floor incision. This alluvial formation is attributed to marine isotopic stage (MIS) 6 (Saalian). The results from the different bioclimatic studies (molluscs, large fauna, microfauna, ostracods, foliar imprints) are extremely coherent and enable us to bring to light the transitions between the last cold Saalian stage and the Eemian interglacial (MIS 5e). The Eemian climatic optimum is recorded in the lower part of the calcareous tufa, in the small organic horizons containing the archaeological remains. The variations in spring dynamics brought to light in the tuffaceous sequence reveal marked instability in the regime of precipitations during the Eemian interglacial optimum. The chronostratigraphic attribution of the sequence is largely confirmed by a series of U/Th, ESR/U-Thand OSL dates, which yield an average age of 123± 3 ka (MIS 5e), placing the sequence within the Eemian interglacial. Towards the valley, the calcareous formation is covered by slope deposits and humic soils deposited after a major phase of erosion truncating the upper part of the tufa sequence. The facies of these soils and their stratigraphic position indicates that they are similar to the humic soil complexes from the Weichselian Early-glacial of the Somme basin, attributed to isotopic stages 5d to 5a (between 120 and 70 ka) and to the transition 5a/4 (~70 ka).

20At Caours, the synthesis of data has resulted in the reconstruction of the different depositional stages of the tufa formation and the major geomorphological modifications governing this formation from glacial/interglacial transition at the base, to incision transition and slope facies at the top (interglacial/Early glacial). The study of the slope sequence truncating and covering the Caours tuffformation towards the present valley reveals fundamental elements concerning environmental modifications of the valley and response to the climatic transition between the Eemian interglacial and the beginning of the Weichselian. The attribution of the humic units to the beginning of the Weichselian glacial implies that the incision process, which was responsible for the terracing of the Caours formation and the maximal deepening of the present valley, appears very early in the climatic cycle, just after the end of the interglacial, as previously advanced for the Somme basin by Antoine et al. (2000, 2003a, 2007).

The loessic sequences (loess-palaeosoils): Human occupationenvironment interactions during the Last climatic cycle

21Recent research underlines the extremely homogeneous character of pedo-sedimentary recording in the Somme basin and the surrounding regions. Moreover, the analysis and the correlation of the whole sequences show that loessic environments display coherent responses to millennial climatic variations, which characterise the last glacial in Western Europe (Dansgaard-Oeschger type rapid events). In this way, many pedological (soil horizons), sedimentary (specific loess facies) and/or periglacial (frost cracks, cryoturbations, thermokarst) marker horizons can be tracked on a regional, and even on a continental scale, from Normandy, through Germany and into central Europe.

22The pedo-stratigraphic sequence of the Last climatic cycle in the Somme basin and surrounding areas will be presented here in detail due to the importance of the new data concerning this period (Fig. 4; Antoine et al., 1999, 2003), and its value as a model for studying human/environment interactions (chronological resolution, data precision). In spite of variations in the thickness of the different stratigraphic units, linked to the differences in geomorphological contexts and to the localisation of the profiles in relation to sources of aeolian material, the loessic sequences of Western France exhibit a particularly constant pedo-sedimentary pattern. They contain many pedological or periglacial marker horizons that allow us to advance a detailed correlation scheme, to refine the chronoclimatic attribution of the different human occupations incorporated within them, and to compare them to other European sequences.

Pedo-sedimentary evaluation of the Last Interglacial-glacial cycle

23The synthesis of observations carried out on more than 80 sequences spanning the last Interglacialglacial cycle (Eemian-Weichselian), like those from Saint-Sauflieu or Fresnoy-au-Val, identifies five main chronostratigraphic phases (Fig. 4; chronology after Johnsen et al., 2001):

24(A) Last interglacial (Eemian; ~ 125-112 ka): The Eemian interglacial is represented in all sections by a horizon of clayey brown to brown-red leached soil, about a metre deep, with a polyhedral to prismatic structure (Fig. 4; unit 22). This horizon is especially well preserved on plateau sites, like at the Sourdon brickyard (Antoine, 1990), where it is characterised by a significant percentage of clays (28-30%) relative to that of carbonate-rich loess (12-15%). This soil was subsequently affected by freeze-thaw alternations during the course of the Weichselian early-glacial over a thickness of 1 to 1.5 m, presenting a marked centimetre-scale polyhedral to banded structure resulting from the segregation of ice lenses linked to deep seasonal frost processes.

25(B) Early-glacial (112-72 ka): After the truncation of the Eemian Bt soil horizon, which occurred during the first climatic deterioration after the interglacial (MIS 5d), the pedo-sedimentary record from the Early-glacial is characterised by the formation of humic soils that are particularly well preserved on lower slope environments (1.5 to 2.5 m thick). This complex of Early-glacial soils is made up of the superposition of: (i) a soil presenting an intermediary facies between an interglacial soil and a grey forest soil (Bettencourt soil, Fig. 4, no. 21), (ii) a typical grey forest soil (SS1, base of the Saint-Sauflieu soil complex, Fig. 4, no. 20), and (iii) a sequence of two to three steppe-like soils (SS2 to SS3 of the Saint-Sauflieu soil complex; Antoine et al., 1994). The build-up of this complex reflects a clear continentalisation of the environment, which is indicated by the development of soils formed on colluvial deposits in a Boreal forest context with pine and birch (Munaut in Antoine et al., 1994). Those two main phases are contemporaneous with a significant decrease in sea level (up to-20 m) and with major palaeogeographical change in the Channel/North Sea area leading to a definitive disappearance of the oceanic influence. Given the characteristics of this soil complex, the Early-glacial record was divided into two major periods of unequal duration: a phase with grey forest soils (~MIS 5a-5d) and a phase with steppe-like soils (end of MIS 5a):

Fig. 4. Pedo-stratigraphic reference sequence for northern France and the Somme basin, correlation with global climatic records and location of the Palaeolithic levels (after Antoine et al., 2003, modified).1: surface soil (a: Hz L; b: hz Bt; c: banded Bh hz); 2: carbonate-rich loess; 3: cryoturbated Nagelbeek horizon; 4: carbonaterich laminated loess with microcracks; 5: cryoturbated tundra gley doublet with intermediary loess; 6: carbonate-rich loess; 7: cryoturbated tundra gley/large ice-wedges; 8: Arctic brown soil; 9: sandy silts; 10: tundra gley; 11: humic Arctic meadow type soil; 12: brown Boreal soil (8 to 12: Saint-Acheul Villiers-Adam soil complex); 13: loess/sandy silts; 14: laminated colluviums with frost cracks; 15: brownish loess; 16 to 19: steppe-like soils with interstratified unit of local, non-carbonate-rich loess (17); 20: grey forest soil on colluviums (SS-1); 21: clayey colluviums/grey forest soil (16 to 21: Saint-Sauflieu soil complex); 22: horizon Bt of brown leached soil (Rocourt/Elbeuf 1); 23: Saalian carbonate-rich loess; 24: bedded carbonate-rich loess of Saalian age; 25: steppe-like humic sol [symbols: A: large ice-wedges (main network at the base of 6); B: ice wedges; C: ice melting channels (thermokarst); D: pockets of karstic dissolution affecting the chalky geologic substratum, “pipes”].

26(B1) Phase with grey forest soils (stages 5c to 5a, 112-70 ka). This phase begins with a first period of colluvial deposition during the course of the first cold stage directly after the Eemian (MIS 5d towards 112-110 ka). These colluvial deposits are then affected by a pedogenesis (Bettencourt soil), which presents an intermediary facies between an interglacial soil (Bt) and a grey forest soil (Bth; Fig. 4, no. 21). The characteristics of this horizon showing clayey to silty-clayey, strongly stratified, weakly humic illuviations, with abundant traces of bioturbation and earthworm hibernation chambers, denote a temperate continental climate with marked seasonal contrasts, similar to the present-day climate in central Europe. A significant hydromorphic phase and deep seasonal frost then affected the top of this soil during the MIS 5b/5c transition (90-88 ka), according to the TL dates obtained on heated flints from associated Palaeolithic sites. This episode was followed by the build-up of silty colluvial deposits, which rework the underlying soils and within which a, grey-forest soil type pedogenesis develops. The formation of this more humic soil (sol SS1; Fig. 4, no. 20), reflects a cumulative soil dynamics that is mainly attributable to stage 5a (85-80 ka; Locht et al., in press). This horizon is then affected by an episode of deep seasonal frost (1.5 m), then by a phase of erosion, during the intense climatic deterioration, which marks the end of stage 5a towards 78 ka (stage GS 21 – GRIP). This grey forest soil phase from the beginning of the glacial with a contrasted continental climate is characterised by a particularly intense episode of dissolution of the chalky substratum (sites of Bettencourt, Fresnoy-au-Val, and Havrincourt). This process provokes the formation of localised dissolution pockets (“pipes”), which can reach a depth of more than 3 m and which contain trapped colluviums and different grey forest soils. This dynamics appears to be linked to the very specific climatic context of this period (marked seasonal contrasts, spring nival melting, acidity of surface soils) and ends abruptly with the formation of the first steppe-like soils towards 70 ka.

27(B2) Phase with steppe-like soils (end of MIS 5a, 78-70 ka). The second phase of the complex differs by the appearance of isohumic steppe-like soils (without clayey illuviations). At Saint-Sauflieu, these horizons are characterised by the moderate extension of the birch in a steppe-like environment with Poaceae and Asteraceae (SS2 and SS3a and b; Fig. 4, no. 19-17-16), in a drier and increasingly open environment. This part of the complex is clearly differentiated from the preceding soils by the appearance of the first aeolian deposits (Fig. 4, units 19 to 16) and reflects an expansion of the continental response to the millennial climatic oscillations, which characterise the Greenland climatic curves between ~ 70 and 75 ka (interstadials GIS 19 and 20 of GRIP & GISP II; Fig. 4). Unfortunately, the lack of precision in the TL-IRSL dates prevents us from advancing a detailed timeframe for these rapid events, which are, however, clearly differentiated in the continental record by 20 to 30 cm thick distinct humic horizons. Lastly, the upper limit of the Early-glacial is defined by the erosive contact marking the summit of the last steppelike soil, which is a fundamental marker and is visible in all the European sequences towards 70 ka.

28(C) Lower Pleniglacial (~70-55 ka): The Lower Pleniglacial is marked by the formation of the first typical homogeneous loess deposits, which are often not carbonate-rich, in sections in the north of France and which rework local elements (Fig. 4, no. 15). This unit is dated to approximately 65± 5 ka in the Rhine valley where it is better represented and denotes the appearance of typically periglacial conditions. It constitutes a stratigraphic marker at the beginning of the loess sedimentation, which appears in parallel with the first major dust peaks recorded in Greenland ice cores. After this particularly dry phase, an intense and probably very short erosive episode is recorded by the formation of bedded colluvial units including soil nodules, cryoturbation and frost-ice cracks, which rework the underlying levels (Fig. 4, no. 14). This major erosion can entail the disappearance in places of the first underlying loess and thereby lead to a major gap in the sequence. This unit, with a maximum thickness of 2 to 3 m, represents a marker level dated between 55 and 50 ka, at the very beginning of isotopic stage 3 (beginning of the Middle Pleniglacial). The formation of these bedded deposits follows a very intense erosive episode resulting from a period of abrupt degradation of the permafrost during a phase of rapid warming, which probably marked the beginning of GIS 16-17 interstadials towards 58 ka (thermokarst). This event is the marker of the Lower Pleniglacial/Middle Pleniglacial transition at the beginning of MIS 3 from the north of France to Germany (Antoine et al., 2009).

29(D) Middle Pleniglacial/Interpleniglacial (~55-33 ka): In all the sections in Western Europe and particularly in the north of France, the Middle Pleniglacial is characterised by a drastic decrease in typical loessic contributions. At the same time, brown Boreal type to brown Arctic type soils (Bw) develop during the interstadials representative of this period (GIS 7-8 to 14, Fig. 4). Although this is a typically periglacial context, we observe the deposition of thin aeolian deposits with a high local component from the reworking of the geological substratum, particularly Tertiary sands in the Parisian basin (Fig. 4, no. 13). In the sequences in the north of France, this period is especially marked by the development of a soil, or rarely of a complex of soils (Saint-Acheul-Villiers-Adam Complex), which globally correspond to the whole of the Middle Pleniglacial or the Interpleniglacial and to most of MIS 3. In most of the sections in the Somme basin and in the Normandy region, this period is poorly represented by a single poly-phased horizon (Saint-Acheul Soil). On the other hand, in certain particularly complete sections in the Val d’Oise (Antoine et al., 2003b; Locht et al., 2003), no less than four horizons have been discovered. These horizons indicate the complexity of slope responses to climatic variations during the Middle Pleniglacial. It is thus possible to observe successively a brown Boreal soil, a humic Arctic meadow soil, a tundra gley, bedded sandy silts and a last brown Arctic soil (Fig. 4, no. 12 to 8).

30(E) Upper Pleniglacial (~30-15 ka): The Upper Pleniglacial is characterised by a very marked acceleration of the loessic sedimentation, especially between 25 and 20 ka. After a major cold event, underlined by a network of large ice-wedge casts at around 30 ka, then by significant slope erosion, it is distinguished by the development of typical carbonate-rich loess, which is, at the most, subdivided into three main units separated by Periglacial horizon markers (tundra gleys and large cracks), including the tongued (sub-divided) Nagelbeek horizon dated to ~ 22 ka 14C BP (Fig. 4, no. 6 to 2). Recent work on this loess, and namely on the malacofauna, brought to light a direct relationship between the development of tundra gleys and the short climatic warmings of the Dansgaard-Oeschger cycles in ice cores (Moine et al., 2011). The malacological observations also show particularly poorer associations compared to those from more continental environments, like the Rhine valley, which point to environments with little vegetal cover (Moine et al., 2011). This weak vegetal biomass concords with the near absence of large herbivore remains in the Upper Pleniglacial carbonate-rich loess deposits of the Somme and the north of France. This carbonate-rich loess is very thick in places, reaching 4 to 6 m on slopes facing NE to SE (downwind) and represents a major component of the present landscape. This carbonate-rich loess cover was subsequently affected by the surface soil, which began to form during the Lateglacial (Van Vliet-Lanoë, 1987), and was then strongly eroded during the course of Protohistoric and modern periods, in relation to the increase of anthropogenic pressure and the development of mechanical agriculture. The products of this erosion accumulate at slope bases, especially in dry valleys where they can reach a thickness of 4 to 5 m. However, part of this sedimentary signal is transferred in the valleys by fluvial reworking, especially upstream of the Somme tributaries where it fills in the alluvial plain and buries the peaty environments, thereby causing river entrenchment.

Interactions between the environment and human occupations

31Recent data concerning human occupations during the Last Interglacial-glacial climatic cycle show a clear concentration of recent Middle Palaeolithic sites during the Weichselian Early-glacial (Fig. 4) and more precisely during the phase of grey forest soil pedogenesis (sub-stages 5a to 5c; Antoine et al., 1994, 1998; Locht et al., 2010). During this long period marked by climatic oscillations in a cool continental context dominated by Boreal forests with pine and birch, sites were trapped by slow and progressive colluvial sedimentation, which did not interfere much with the spatial distribution of artefacts. However, this concentration of occupations during the development phase of grey forest soils must be considered in relation to the comparatively long duration of this period (~30 ka), compared to the ensuing phase of steppe-like soils (~5-7 ka). The present results underline the dramatic decrease in the occupation of the territory by Neanderthal populations during the Lower Pleniglacial. This occupation only reappears during the Middle Pleniglacial but is much more limited than during the Early-glacial.

32The end of the Middle Palaeolithic is represented by a much smaller number of sites but this deficiency is being compensated for by new discoveries as part of rescue excavations. According to these data, Neanderthals occupied the territory during the period extending from 55-40 ka. In the field, this period is characterised by the formation of the Complex of Saint-Acheul-Viliers-Adams soils in a context of very marked climatic oscillations, characterised by the alternation of very harsh stages (loess/frost cracks) and rapid climatic improvements during which brown soils develop in a clearly vegetated, steppe-type environment (Fig. 4). On-going work on the Canal Seine-Nord-Europe sequences should allow us to complete data for these periods in the near future (Goval et al., 2013).

33Human occupation in the Somme basin is discontinuous during the Upper Palaeolithic. Until recently, the early Upper Palaeolithic (38-28 ka BP) was only represented by a few surface sites with no stratigraphic context, indicating brief Aurignacian and Gravettian incursions. However, rescue works conducted by the INRAP led to the recent discovery of occupations dating from 27 to 28 ka 14C BP, which are currently being studied: Havrincourt (Nord, excavations E. Goval, 2011) and Renancourt les Amiens (Somme, excavations, C. Paris, 2012). During the Upper Pleniglacial, characterised by the build-up of thick loessic deposits, hundreds of archaeological survey pits carried out in the region indicate that large mammals were extremely rare whereas conservation conditions are optimal in the carbonate-rich loess. The final Gravettian level of Renancourt-lès-Amiens attests the incursion of human groups in the region towards 23 ka BP (Fagnart et al., 2013).

34Following an occupation gap of at least 10 ka, the Somme basin and northern France were occupied once again during the Lateglacial (13-12 ka BP) by horse and reindeer hunters, during the Upper and Final Magdalenian in a steppe-like and continental climate. Given the low density of sites, the region appears to have been episodically occupied or traversed by hunters and was probably on the fringes of more significant occupation centres (Parisian basin, Belgian Ardennes, Rhineland).

35The Federmesser tradition appeared slightly before the Allerød oscillation, at about 12.2 ka BP, in a bioclimatic environment where birch, then pine forests gradually replaced steppe-like areas. Alongside these changes, the cold Magdalenian fauna made way for a more temperate biocenose adapted to forest environments (red deer, roe deer, aurochs). These new conditions were fully attained during the Allerød oscillation between 11.8 and 10.8 ka BP, which represents the most intense period of occupation of the Weichselian Lateglacial. Climatic and environmental changes led to significant modifications in technical traditions and the social structure of hunters. Federmesser groups represent the adaptation of final Palaeolithic populations to the new conditions of the environment. The intense human occupation during the Allerød oscillation declined considerably during the Younger Dryas, during the course of which no archaeological evidence has been found in the Somme basin in spite of 20 years of research and intensive prospections. Towards 10 ka BP, at the limit between the Younger Dryas and the Preboreal, a new phase of occupation occurred in the region, characterised by industries with “pièces mâchurées” (Long Blade Technology), recently attributed to the final Ahrensburgian. This ultimate Palaeolithic tradition is characterised by (i) numerous sites specialised in raw material procurement (Belloy-sur-Somme) and the production of blades and (ii) a large fauna associating the red deer, the aurochs and the horse. Lastly, the first Mesolithic groups with micro-lithic armatures appeared in the region between 9.8 and 9.5 ka BP, during the development of the Holocene forest (appearance of the hazel tree) and the considerable expansion of the boar and the red deer.


36Researches carried out in the Somme basin and in the neighbouring regions over the past twenty years have been based on a unique integration of geological, geomorphological, palaeobiological, geochronological and archaeological approaches, resulting in a reference database for the study of interactions between humans and the environment for more than 500 ka. The Somme basin is characterised by the coexistence of two types of complementary sedimentary contexts: fluvial valley sequences and pedo-sedimentary slope sequences that are inextricably associated in stepped terraces systems with loess cover of this area. Alluvial sequences are characterised by abundant palaeontological remains allowing a detailed reconstruction of the palaeo-biocenoses, whereas slope sequences contain numerous marker horizons facilitating the correlation of the various pedo-sedimentary records. By correlating all of these records on a regional scale, and more broadly with neighbouring regions, we can place the human occupations within an accurate chronoclimatic time frame. The original geoarchaeological approach is characterised by the overlap of temporal scales and investigation methods concerning an extensive chronological period: from the first human occupation in north western Europe to the last Mesolithic post-glacial hunter-gatherers. The main conclusions are the following:

  1. Although the Somme appears to be a reference region for prehistory, the results of the multiple investigations carried out over the past twenty years (survey pits and excavations) compel us to point out the extreme rarity of ancient human occupations attributable to the Acheulean. Indeed, no new Acheulean site in primary position was discovered in fluvial levels in spite of multiple surveys. The confrontation of results from modern excavations, geochronological data and earlier discoveries shows that the first major occupations in the Somme basin are linked to the 450-500 ka period at the earliest (beginning of MIS 12) and that they are represented by evolved Acheulean industries from the outset. However, new research on the “historic” levels from Abbeville/Carpentier quarry could indicate that Palaeolithic populations inhabited the region during an earlier phase, about 550-600 ka ago, and that these first occupations would thus be contemporaneous with Mauer Man (Homo Heidelbergensis).
  2. New data concerning the Last climatic cycle (Eemian-Weichselian) show that the human occupation of the Somme basin is clearly discontinuous with maximum remains during transition phases with Early-glacial to continental type climate and a Boreal to forest-steppe forest context (~112-70 ka). Only a few occupations were brought to light during the Lower Pleniglacial (~60 ka) and during the course of the Middle Pleniglacial (~ 40-50 ka). A total abandonment of the region is revealed between about 27 and 15 ka during the period of maximal loess sedimentation. The re-colonisation of the valley is then observed at the beginning of the Lateglacial climatic warming by final Palaeolithic, then by Mesolithic human groups. Given the results concerning the Last climatic cycle, there appears to be a proven relationship between the intensity of human occupation and the climatic and especially environmental context. This relationship appears to be conditioned by the relative abundance of large fauna, an essential source of subsistence for prehistoric groups. In turn, faunal presence is linked to vegetation density, as indicated by the extremely sparse biomass contemporaneous with the Upper Pleniglacial loess.



Agache R., Bourdier F. et Petit R., «Le Quaternaire de la Basse Somme : tentative de synthèse», Bulletin de la Société Géologique de France, no 7 (V), 1963, p. 422-442.

Antoine P., Chronostratigraphie et environnement du Paléolithique du bassin de la Somme, Publications du Centre d’Études et de Recherches Préhistoriques, Université de Lille, 2, 1990.

Antoine P., «The Somme Valley terrace system (Northern France); a model of river response to quaternary climatic variations since 800 000 BP», Terra-Nova, no 6, 1994, p. 453-464.

Antoine P., «Modifications des systèmes fluviatiles à la transition Pléniglaciaire-Tardiglaciaire et à l’Holocène: l’exemple du bassin de la Somme (Nord de la France)», Géographie Physique et Quaternaire, no 51 (1), 1997, p. 93-106.

Antoine P., Lautridou J.-P., Sommé J., Auguste P., Auffret J.-P., Baize S., Clet-Pellerin M., Coutard J.-P., Dewolf Y., Dugué O., Joly F., Laignel B, Laurent M., Lavollé M., Lebret P., Lécolle F., Lefebvre D., Limondin-Lozouet N., Munaut A.-V., Ozouf J.-C., Quesnel F. et Rousseau D.-D., «Le Quaternaire de la France du Nord-Ouest: limites et corrélations», Quaternaire, no 9, 1998, p. 227-241.

Antoine P., Rousseau D.-D, Lautridou J.-P. et Hatté C., «Last Interglacial-Glacial climatic cycle in loess-palaeosol successions of north-western France», Boreas, no 28, 1999, p. 551-563.

Antoine P., Fagnart J.-P., Limondin-Lozouet N. et Munaut A.-V., «Le Tardiglaciaire du Bassin de la Somme», Quaternaire, no 11 (2), 2000a, p. 85-98.

Antoine P., Lautridou J.-P. et Laurent M., «Long-Term Fluvial archives in NW France: Response of the Seine and Somme Rivers to Tectonic movements, Climatic variations and Sea level changes», Geomorphology, no 33, 2000b, p. 183-207.

Antoine P., Limondin-Lozouet N., Auguste P., Locht J.-L., Ghaleb B., Reyss J.-L., Escudé E., Carbonel P., Mercier N., Bahain J.-J., Falguères C. et Voinchet P., « Le tuf de Caours (Somme, France): mise en évidence d’une séquence éémienne et d’un site paléolithique associé », Quaternaire, no 17 (4), 2006, p. 281-320.

Antoine P., Auguste P., Bahain J.-J., Coudret P., Depaepe P., Fagnart J.-P., Falguères C., Fontugne M., Frechen M., Hatté C., Lamotte A., Laurent M., Limondin-Lozouet N., Locht J.-L., Mercier N., Moigne A.-M., Munaut A.-V., Ponel P. et Rousseau D.-D., «Paléoenvironnements pléistocènes et peuplements paléolithiques dans le bassin de la Somme (nord de la France)», Bulletin de la Société préhistorique française, no 100, 2003a, p. 5-28.

Antoine P., Bahain J.-J., Debenham N., Frechen M., Gauthier A., Haesaerts P., Hatté C., Limondin-Lozouet N., Locht J.-L., Raymond P. et Rousseau D.-D., « Nouvelles données sur le Pléistocène du Nord du Bassin Parisien: les séquences loessiques de Villiers-Adam (Val d’Oise, France) », Quaternaire, no 14, 2003b, p. 219-235.

Antoine P., Limondin-Lozouet N., Chaussé C., Lautridou J.-P., Pastre J.-F., Auguste P., Bahain J.-J., Falguères C. et Ghaleb B., «Pleistocene fluvial terraces from northern France (Seine, Yonne, Somme): synthesis, and new results from interglacial deposits», Quaternary Science Reviews, no 26, 2007, p. 2701-2723.

Antoine P., Auguste P., Bahain, J.-J., Chaussé C., Falguères C., Ghaleb B, Limondin-Lozouet N., Locht J.-L. et Voinchet, P., «Chronostratigraphy and palaeoenvironnement of Acheulean occupations in Northern France (Somme, Seine and Yonne River valleys)», Quaternary International, no 223-224, 2010, p. 456-461.

Antoine P., Bahain J.-J., Auguste P., Fagnart J.-P., Limondin-Lozouet N. et Locht J.-L., « Quaternaire et Préhistoire dans la vallée de la Somme : 150 ans d’histoire commune », in Hurel A. and Coye N. (éds.), Dans l’épaisseur du temps. Archéologues et géologues inventent l’histoire, Publications du Muséum national d’Histoire naturelle, Paris, 2011.

Auguste P., « Évolution des peuplements mammaliens en Europe du Nord-Ouest durant le Pléistocène moyen et supérieur, Le cas de la France septentrionale », Quaternaire, no20 (4), 2009, p. 527-550.

Beun N. et Broquet P., « Tectonique quaternaire (holocène?) dans la plaine littorale picarde des Bas-Champs de Cayeux et de leurs abords orientaux. Incidences possibles sur le réseau hydrographique régional », Bulletin de l’Association Française pour l’Étude du Quaternaire, no 1-2, 1980, p. 47-52.

Bordes F., Les limons quaternaires du bassin de la Seine, Archives de l’Institut de Paléontologie Humaine, Mémoire no 26, 1954.

Boucher de Perthes J., Antiquités celtiques et antédiluviennes: Mémoire sur l’industrie primitive et les arts à leur origine, vol. 1, Paris, Treuttel & Wurtz, Derache, Dumoulin & Didron, 1849.

Boucher de Perthes J., De la machoire humaine de Moulin-Quignon: nouvelles découvertes en 1863 et 1864, Jung-Treuttel, Derache, Dumoulin, Victor Didron, Paris, 1864.

Bourdier F., « Étude comparée des dépôts quaternaires des bassins de la Seine et de la Somme », Bulletin d’Information des Géologues du Bassin de Paris, no 21, 1969, p. 169-231.

Bourdier F. (dir.), Quaternaire et Paléolithique des bassins de la Somme et de la Basse Seine, Bulletin de l’Association Française pour l’Étude du Quaternaire, no 3-4, 1974.

Breuil H. et Koslovski L., « Études de stratigraphie paléolithique dans le Nord de la France, la Belgique et l’Angleterre », L’Anthropologie, no XLI, 1931, p. 449-488.

Bridgland D.R., «River terrace systems in north-west Europe: an archive of environmental change, upliftand early human occupation», Quaternary Science Reviews, no 19, 2000, p. 1293-1303.

Commont V., « Saint-Acheul et Montières. Notes de Géologie, de Paléontologie et de Préhistoire », Mémoire de la Société géologique du Nord, no 6 (3), 1909, p. 1-68.

Commont V., « Les terrasses fluviatiles de vallée de la Somme », Bulletin archéologique, 1911, p. 173-195.

Commont V., « Notes sur le Quaternaire du Nord de la France, de la vallée du Rhin et de la Belgique », Bulletin de la Société géologique du Nord, no 41, 1912, p. 12-52.

Commont V., Les Hommes contemporains du renne dans la vallée de la Somme, Mémoire de la Société des Antiquaires de Picardie, no 37, 1913.

Ducrocq T., Le Mésolithique du bassin de la Somme. Insertion dans un cadre morpho-stratigraphique, environnemental et chronoculturel, Thèse de Doctorat, Université des Sciences et Technologies de Lille, 1999.

Ducrocq T., Bridault A. et Munaut A.V., « Un gisement mésolithique exceptionnel dans le Nord de la France: Le Petit-Marais de la Chaussée-Tirancourt. (Somme) », Bulletin de la Société Préhistorique Française, no 88, 1991, p. 272-276.

Fagnart J.-P., La fin des temps glaciaires dans le Nord de la France. Approches archéologique et environnementale des occupations humaines du Tardiglaciaire, Mémoire de la Société Préhistorique Française, Tome 24, 1997.

Fagnart J.-P. et Coudret P., «Le Tardiglaciaire dans le Nord de la France», in Valentin B., Bodu P. et Christensen M. (dir.), L’Europe centrale et septentrionale au Tardiglaciaire: confrontation des modèles régionaux de peuplement, Mémoires du Musée de Préhistoire, no 7, Nemours, Ed. de l’APRAIF, 2000, p. 111-128.

Fagnart J.-P., Coudret P., Antoine, P., Vallin L., Sellier N. et Masson B., « Le Paléolithique supérieur ancien dans le Nord de la France », Mémoire LVI de la Société préhistorique de France, 2013, p. 197-214.

Gibbard P.L., The Pleistocene history of the Lower Thames Valley, University Press, Cambridge, 1994.

Gibbard P.L., «Europe cut adrift», Nature, no 448, 2007, p. 259-260.

Goval E. (dir.), Antoine P., Auguste P., Bahain J.-J., Claud E., Coutard S., Font C., Guerrin C., Hérisson D., Hulin G., Jamet G., Moine O., Prilaux G., Schmidt E., Valladas E., Canal Seine-Nord Europe, fouille 22, Nord-Pas-de-Calais, Pas-de-Calais, Havrincourt. Des derniers Néandertaliens aux premiers Homo Sapiens: les occupations paléolithiques d’Havrincourt, «Les Bosquets» (Pas-de-Calais). Rapport de Fouille INRAP, Coordination Canal Seine-Nord-Europe, 2013.

Haesaerts P. et Dupuis C., «Contribution à la stratigraphie des nappes alluviales de la Somme et de l’Avre dans la région d’Amiens», in Chronostratigraphie et faciès culturels du Paléolithique inférieur et moyen dans l’Europe du Nord-Ouest, Supplément au Bulletin de l’Association Française pour l’étude du Quaternaire, no 26, 1986, p. 171-186.

Ladrière J., « Étude stratigraphique du terrain Quaternaire du Nord de la France », Annales de la Société Géologique du Nord, no 8, 1890, p. 93-276.

Laurent M., Falguères C., Bahain J.-J. et Yokoyama Y., « Géochronologie du système de terrasses fluviatiles quaternaires du bassin de la Somme par datation RPE sur quartz, déséquilibres des familles de l’uranium et magnétostratigraphie », Comptes-Rendus de l’Académie des Sciences de Paris, no 318, série II, 1994, p. 521-526.

Lefèbvre D., Antoine P., Auffret J.-P., Lautridou J.-P. et Lécolle F., « Rythme de réponse des environnements fluviatiles aux changements climatiques en France du Nord-Ouest », Quaternaire, no 5, (3-4), 1994, p. 165-172.

Lericolais G., Évolution Plio-quaternaire du Fleuve Manche: Stratigraphie et Géomorphologie d’une plateforme continentale en régime périglaciaire, Thèse de Doctorat de Géologie marine, Université de Bordeaux I, 1997.

Limondin N., «Late-glacial and Holocene Malacofaunas from Archaeological Sites in the Somme Valley (North France)», Journal of Archaeological Science, no 22, 1995, p. 683-698.

Limondin-Lozouet N., « Les successions malacologiques du Tardiglaciaire et du début de l’Holocène dans la vallée de la Somme », 119e congrès national des sociétés historiques et scientifiques, Amiens 1994, Pré-et Protohistoire, 1997, p. 39-46.

Limondin-Lozouet N., « Successions malacologiques à la charnière Glaciaire/Interglaciaire: du modèle Tardiglaciaire-Holocène aux transitions du Pléistocène », Quaternaire, no 22 (3), 2011, p. 211-220.

Limondin-Lozouet N. et Antoine P., «Lateglacial and Early Holocene palaeoenvironmental changes inferred from malacofaunas at Conty (Northern France)», Boreas, no 30, 2001, p. 148-164.

Limondin-Lozouet N. et Antoine P., «A new Lyrodiscus (Mollusca, Gastropoda) assemblage at Saint-Acheul (Somme Valley): reappraisal of Stage 11 malacofaunas from Northern France», Boreas, no 35, 2006, p. 622-633.

Locht J.-L., Le gisement paléolithique moyen de Beauvais. Contribution aux modalités de subsistance des chasseurs de rennes du Paléolithique moyen, Thèse de Doctorat de l’Université des Sciences et Technologies de Lille, 2004.

Locht J.-L., Antoine P., Bahain J.-J., Dwrila G., Raymond P., Limondin-Lozouet N., Gauthier A., Debenham N., Frechen M., Rousseau D.-D., Hatté C., Haesaerts P. et Metsdagh H., « Le gisement paléolithique moyen et les séquences pléistocènes de Villiers-Adam (Val d’Oise, France) : chrono-stratigraphie, environnement et implantations humaines», Gallia Préhistoire, no 45, 2003, p. 1-111.

Locht J.-L. (dir.), Bettencourt-Saint-Ouen (Somme), cinq occupations du Paléolithique au début de la dernière glaciation, Documents d’Archéologie Française (DAF), 2002.

Locht J.-L., Goval E. et Antoine, P., «Reconstructing Middle Palaeolithic hominids behaviour during OIS 5 in Northern France», in Conard N. et Delagnes A.(eds.), Settlements dynamics of the Middle Palaeolithic and Middle Stone Age, volume 3, Kerns Verslag ed., Tubingen, 2010, p. 329-355.

Locht J.-L., Sellier N., Antoine P., Koehler H. et Debenham N., « Mauquenchy (Seine Maritime): mise en évidence de deux niveaux d’occupation paléolithique dans un sol gris-forestier daté du SIM 5a (Début-glaciaire weichselien) », Quaternaire (sous presse).

Moine O., Antoine P., Deschodt L. et Sellier-Segard N., «Enregistrements malacologiques à haute résolution dans les lœss et les gleys de toundra du Pléniglaciaire weichselien supérieur: premiers exemples du nord de la France», Quaternaire, no 22 (4), 2011, p. 307-325.

Munaut A.-V., «L’environnement végétal de quelques dépôts quaternaires du bassin de la Somme», in Cultures et industries paléolithiques en milieu loessique, Revue Archéologique de Picardie, no 1-2, 1988, p. 45-56.

Ponel P., Coope G.-R., Antoine P., Limondin-Lozouet N., Leroyer C., Munaut A.-V., Pastre J.-F. et Guiter F., «Late-Glacial palaeoenvironments and palaeoclimates from Conty and Houdancourt Northern France, reconstructed from Beetle remains», Journal of Quaternary Science, no 24, 2005, p. 2449-2465.

Prestwich J., « Sur la découverte d’instruments en silex associés à des restes de mammifères d’espèces perdues dans des couches non remaniées d’une formation géologique récente », Comptes Rendus de l’Académie des Sciences, Paris, no 49, 1859, p. 634-636.

Sommé J. et Tuffreau A., « Historique des recherches sur le Quaternaire de la région du Nord de la France », Bulletin de l’Association Française pour l’étude du Quaternaire, no 15, 1978, p. 5-13.

Sommé J., Fagnart J.-P., Léger M., Munaut A.-V., Puisségur J.-J. et Tuffreau A., « Terrasses fluviatiles du Pléistocène moyen en France septentrionale : signification dynamique et climatique », Bulletin de l’Association Française pour l’étude du Quaternaire, no 21, 1984, p. 52-58.

Tuffreau A., Munaut A.-V., Puissegur J.-J. et Sommé J., « Les basses terrasses dans les vallées du Nord de la France et de la Picardie: Stratigraphie et Paléolithique», Bulletin de la Société Préhistorique de France, no 78, 1981, p. 291-305.

Van Vliet-Lanoë B., Le rôle de la glace de ségrégation dans les formations superficielles de l’Europe du Nord-Ouest, Thiesis (« Thèse d’État », Université Paris 1, 1987.

Veldkamp A. et Van Den Berg M.W., «Three-dimentional modelling of Quaternary fluvial dynamics in a climo-tectonic dependant system. A case study of the Maas record (Maastricht, The Netherlands)», Global and Planetary Change, no 8, 1993, p. 203-218.

Table des illustrations

Légende Fig. 1. The system of fluvial palaeochannels under the Channel (after Auffret et al., 1982; Lericolais, 1997) and loess extension in northwest France (after Antoine et al., 2003).
Fichier image/jpeg, 99k
Légende Fig. 2. The different morpho-sedimentary contexts of prehistoric sites in the Somme basin.
Fichier image/jpeg, 257k
Légende Fig. 3. The Middle Somme Valley terrace system: stepped fluvial sequences (alluvial formations), silty cover (loess and palaeosols), dates and location of the main Palaeolithic cultures (after Antoine et al., 2003, 2007).
Fichier image/jpeg, 521k
Légende Fig. 4. Pedo-stratigraphic reference sequence for northern France and the Somme basin, correlation with global climatic records and location of the Palaeolithic levels (after Antoine et al., 2003, modified).1: surface soil (a: Hz L; b: hz Bt; c: banded Bh hz); 2: carbonate-rich loess; 3: cryoturbated Nagelbeek horizon; 4: carbonaterich laminated loess with microcracks; 5: cryoturbated tundra gley doublet with intermediary loess; 6: carbonate-rich loess; 7: cryoturbated tundra gley/large ice-wedges; 8: Arctic brown soil; 9: sandy silts; 10: tundra gley; 11: humic Arctic meadow type soil; 12: brown Boreal soil (8 to 12: Saint-Acheul Villiers-Adam soil complex); 13: loess/sandy silts; 14: laminated colluviums with frost cracks; 15: brownish loess; 16 to 19: steppe-like soils with interstratified unit of local, non-carbonate-rich loess (17); 20: grey forest soil on colluviums (SS-1); 21: clayey colluviums/grey forest soil (16 to 21: Saint-Sauflieu soil complex); 22: horizon Bt of brown leached soil (Rocourt/Elbeuf 1); 23: Saalian carbonate-rich loess; 24: bedded carbonate-rich loess of Saalian age; 25: steppe-like humic sol [symbols: A: large ice-wedges (main network at the base of 6); B: ice wedges; C: ice melting channels (thermokarst); D: pockets of karstic dissolution affecting the chalky geologic substratum, “pipes”].
Fichier image/jpeg, 515k


© CNRS Éditions, 2015

Conditions d’utilisation :

Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search