Navigation – Plan du site

Chronostratigraphy of Holocene alluvial archives in the Wadi Sbeïtla basin (central Tunisia)

Chronostratigraphie des archives alluviales holocènes dans le bassin de l’oued Sbeïtla (Tunisie centrale)
Kamel Zerai
p. 271-286

Résumés

L’étude des archives alluviales holocènes dans le bassin de l’oued Sbeïtla a révélé plusieurs phases d'activité fluviale, chacune suivie d’une atténuation de l’activité morphogénique puis de la formation d’un sol. Ces alternances de périodes d’activité morphogénique et de pédogenèse sont une conséquence de l’action des fluctuations climatiques holocènes combinées avec celles des activités humaines. Des datations radiocarbone et la prise en compte de marqueurs archéologiques, nous ont permis d’établir une chronostratigraphie détaillée des archives alluviales holocènes du bassin étudié. Quatre grandes phases d’accumulation holocènes ont été datées : une phase de l'Holocène inférieur (10-7 ka cal. BP), une phase de l’Holocène moyen (6,5-4 ka cal. BP), une phase romaine tardive à post-romaine (1,6-1,1 ka cal. BP) et une phase médiévale (1-0,5 ka cal. BP). Deux périodes d'activité fluviale sont particulièrment marquées, l’une au cours de la transition Pléistocène-Holocène et l’autre au cours de la crise romaine tardive (autour de 1,6 ka cal. BP). Deux phases de pédogenèse ont également été identifiées et datées, une première qui correspond à la période capsienne (vers 7 ka cal. BP) et une deuxième qui est en phase avec la seconde moitié de l’Holocène moyen (5-4 ka cal. BP). Enfin, un paléosol peu développé est observé autour de 1,1 ka cal. BP.

Haut de page

Notes de la rédaction

Article soumis le 8 juin 2009, accepté le 2 septembre 2009

Texte intégral

I would like to thank the Laboratory for Physical Geography (LGP, UMR 8591, CNRS, Meudon) for supporting this project. I wish to thank Abderrazak Gragueb from the National Agency for the Heritage and Cultural Promotion (Tunisia) for the evaluation of the artifacts, Fethi Chelbi from Carthage Museum for the evaluation of the ceramics, Nicole Limondin-Lozouet and Jeanne Zaouali for evaluation of shells (Helix) and kamel Zouari from National Engineering School of Sfax for some radiocarbon datings. The author is also grateful to all reviewers and several individuals for their helpful comments and proof-reading especially Christoph Zielhofer, Jean-Louis Ballais, Brigitte Coque-Delhuile, Mohamed Raouf karray and Dhia Al-Bakri.

Introduction

1This paper studies the Holocene alluvial archives in central Tunisia and contributes to the establishment of a detailed chronostratigraphy for the study area. This study of the Holocene alluvial archives concerns an area sparsely studied compared to other regions of Tunisia, especially the north (Stevenson et al., 1993; Bourgou, 1993; Faust et al., 2004; Zielhofer et al., 2004; Moldenhauer et al., 2008; Zielhofer and Faust, 2008) and the south (Ballais, 1991; Petit-Maire et al., 1991; White et al., 1996; Swezey et al., 1999; Marquer et al., 2008) or compared to North Africa (Gasse and Van Campo, 1994; Ballais, 1995; Lamb and Van der kaars, 1995; Ballais and Benazzouz, 2004). Morphological and pedological approaches allowed the identification of four Holocene accumulation phases in the Wadi Sbeitla basin. These phases of accumulation correspond to sedimentary cycles, starting with coarser deposits that become fine at the top. Each phase of accumulation is often consistent with a Holocene terrace. Radiocarbon dating and an abundance of archaeological and historical data in the studied region (e.g., Vermeersch, 1973; Barbery and Delhoume, 1982; Hitchner et al., 1990) have clarified the chronology of the Holocene alluvial records in the Wadi Sbeitla basin. The Roman pottery is sometimes chronologically more precise than radiocarbon dating (Oueslati, 2005; Zerai, 2006). A synthetic profile of Holocene alluvial archives has been made from different profiles studied in the Wadi Sbeitla basin. This synthetic profile was compared with detailed studies in the north (Bourgou 1993; Faust et al., 2004; Zielhofer et al., 2004) and in the south of Tunisia (Ballais, 1991). Besides the different Holocene accumulation phases, this synthetic profile shows palaeoenvironmental changes during the last ten millennia. Several periods of geomorphic activity and stability have alternated during the Holocene. Geomorphic stability results from humid conditions and dense vegetation cover. Whereas, geomorphic activity is characterised by mobilization of soils and sediments resulting in high sedimentation rates (e.g., Faust et al., 2004). The Holocene palaeoenvironmental changes in the Wadi Sbeitla basin are presented and the findings are compared to other studies in Tunisia and in North Africa.

Geographical setting

2Located in west-central Tunisia, the catchment area of Wadi Sbeïtla is dominated in its upper part by large mountains, where Semmama Mountain reaches up to 1356 m above sea level. In the downstream, wadis are developed in alluvial plains where the altitude does not exceed 420 m above sea level. Covering an area of 702 km2, this basin is rectangular, with a NW-SE direction, perpendicular to the direction of the Atlas Mountain range (fig. 1). This position has influenced the different physical characteristics of the basin, such as slopes (fig. 2A), lithology (fig. 2B), drainage pattern (fig. 2C) and vegetation cover (fig. 2D). Mountains are covered by forest vegetation composed of Pinus halepensis, Quercus ilex, Juniperus phoenicea and Rosmarinus officinalis. In comparison, the foothills and plains have steppe vegetation such as Stipa tenacissima, Artemisia herba-alba and Artemisia campestris (Le Houérou, 1969). The climate of the studied area is semi-arid, characterised by irregular and intense rainfall, with an annual average of 273 mm. The annual average of temperature is around 17.2 °C and the prevailing wind direction is NW-SE.

Fig. 1 – Location of the Wadi Sbeïtla basin and sites of key profiles.
Fig. 1 – Localisation du bassin-versant de l’oued Sbeïtla et des coupes étudiées.

Fig. 1 – Location of the Wadi Sbeïtla basin and sites of key profiles.Fig. 1 – Localisation du bassin-versant de l’oued Sbeïtla et des coupes étudiées.

1: profile; 2: radiocarbon dating; 3: ceramics; 4: artifacts; 5: prehistoric site (Capsian); 6: key profile.
1 : coupe ; 2 : datations par le radiocarbone ; 3 : céramiques ; 4 : artéfacts ; 5 : site préhistorique (Capsien) ; 6 : code de la coupe.

Fig. 2 – Geographical settings of Wadi Sbeitla basin.
Fig. 2 – Cadre géographique du bassin de l’oued Sbeïtla.

Fig. 2 – Geographical settings of Wadi Sbeitla basin.Fig. 2 – Cadre géographique du bassin de l’oued Sbeïtla.

A: Slope classes. B: Lithology. 1: hard rock; 2: mixed rock; 3: soft rock; 4: calcareous crust; 5: Quaternary deposit. C: Drainage pattern. D: Vegetal cover.
A : Carte des pentes. B : Lithologie. 1 : roches résistantes ; 2 : roches moyennement résistantes ; 3 : roches meubles ; 4 : croûte calcaire ; 5 : dépôts quaternaires. C : Réseau hydrographique. D : Couverture végétale.

Methods

3Field investigations were conducted between 2002 and 2005 (Zerai, 2006). During this period, twenty profiles were studied and sampled for various analyses. Each of the studied profiles had a characteristic profile, allowing the presentation and correlation between various deposits. In this paper, the five profiles, most complete and rich in archaeological data are presented (fig. 1). The grain-size distribution was determined in the Laboratory of Physical Geography of Meudon (France) using a laser diffraction particle size analyser (Coulter LS 230) and the microscopic examination of the quartz grains was made using a scanning electron microscope (SEM). Some characteristics of the surfaces of the quartz grains (amorphous silica, dissolution forms, shock marks) allowed to reconstruct the palaeoenvironments of the studied area (soil formation, geomorphic stability or fluvial activity). The method of O.P. Mehra and M.L. Jackson (1960) was used to extract the free iron and total iron. The proportion of free iron and total iron provides information on the degree of weathering and pedogenesis of studied deposits, and consequently on their relative age. Analysis of samples by loss on ignition was used to determine the proportion of CaCO3, organic matter and moisture. Finally, other analysis, such as morphometry of gravels, was also carried out to understand the conditions of accumulation. Nine radiocarbon dates were obtained giving ages covering the end of the late Pleistocene and the Holocene (Tab. 1). The materials used for these datings are charcoal, bone and shell (Helix; Tab. 1). Datings on shells were avoided over concerns of reliability, especially during the Holocene (Goodfriend, 1987; Romaniello et al., 2008). Datings obtained from shells were checked by datings on other more reliable materials, mainly charcoal and bone. Four 14C samples have been conducted in the AMS 14C-Labor Erlangen, Germany (Erl-7905, Erl-7906, Erl-7907 and Erl-7908). The other five 14C samples have been performed in the National Engineering School of Sfax, Tunisia (Tab. 1). The obtained dates were converted into calibrated calendar years (cal. BP) using the program Calib Rev 5.0.1 (Stuiver and Reimer, 2005) resulting from updating of the database INTCAL98 (Stuiver et al., 1998). These dates, along with other dating from archaeological artifacts and ceramics, have provided a precise Holocene chronostratigraphic scale for the alluvial archives in the Wadi Sbeïtla basin. The abundance of archaeological material in central Tunisia was reported by several researchers (e.g., Vermeersch, 1973; Barbery and Delhoume, 1982; Hitchner et al., 1990; Bejaoui, 2003). During the field work five prehistoric sites and several Roman sites were discovered. The most important prehistoric site, has produced hundreds of artifacts that have been identified by A. Gragueb (National Agency for the Heritage and Cultural Promotion, Tunisia). The identification of Roman pottery has been entrusted to F. Chelbi (National Museum of Carthage). The dates obtained were converted into calendar years (cal. BP) to facilitate correlation with other radiocarbon dates.

Tab. 1 – Materials used for dating.
Tab. 1 – Matériaux utilisés pour les datations.

Tab. 1 – Materials used for dating.Tab. 1 – Matériaux utilisés pour les datations.

Results

Variety of Holocene deposits: significant profiles

4Safhit El-Flalig profile (SEF; fig. 3 to fig. 5) is located downstream of the Wadi Douleb. It is influenced by the lithological and structural conditions (fig. 3). This position explains the extent of deposits, especially the historical deposits, which exceed a thickness of 7 m (SEF4; fig. 4 and fig. 5). These historical deposits are filled after a major phase of fluvial incision, which reached the underlying Pleistocene deposits (SEF2) and are nested in the early Holocene deposits (SEF3; fig. 5). The historical deposits of SEF profile are composed of six layers (SEF4 a, b, c, d, e and f; fig. 5). The bottom layer (SEF4a), coarse and containing some boulders of 70 cm of diameter, were dated at the top to 1343±76 cal. BP (Erl-7908). SEF4c soil containing charcoal has been dated to 1082±105 cal. BP (Erl-7906; fig. 5). Thus, a sequence of deposits with a thickness of 4.5 m was accumulated during two centuries and a half, giving a record of the sedimentation rate of 17 mm/a. This record is explained by the position of the profile, just at the outlet of anticlinal Valley of Douleb.

Fig. 3 – Structural framework of SEF and ACH profiles.
Fig. 3 – Contexte structural des deux coupes SEF et ACH.

Fig. 3 – Structural framework of SEF and ACH profiles.Fig. 3 – Contexte structural des deux coupes SEF et ACH.

1: clay (Santonian); 2: limestone (Campanian-Maastrichtian); 3: silt (Aquitanian); 4: sandstone (Upper Langhian); 5: hogback; 6: dip; 7: slope deposit; 8: Pleistocene deposit; 9: Holocene deposit; 10: Wadi; 11: riverbank; 12: contour line; 13: Roman aqueduct; 14: road; 15: stratigraphic section; 16: Capsian site.
1 : argiles (Santonien) ; 2 : calcaires (Campanien-Maastrichtien) ; 3 : silts (Aquitanien) ; 4 : grès (Langhien supérieur) ; 5 : crêt ; 6 : pendage ; 7 : dépôts de pente ; 8 : dépôts pléistocènes ; 9 : dépôts holocènes ; 10 : oued ; 11 : berge ; 12 : courbe de niveau ; 13 : aqueduc romain ; 14 : route ; 15 : coupe stratigraphique ; 16 : site capsien.

Fig. 4 – Captions of fig. 5 to fig. 9.
Fig. 4 – Légende de la fig. 5 à la fig. 9.

Fig. 4 – Captions of fig. 5 to fig. 9.Fig. 4 – Légende de la fig. 5 à la fig. 9.

1: Miocene sandstone; 2: clayey to silty layer; 3: loam; 4: sand; 5: gravel layer; 6: humic enrichment; 7: hydromorphic features; 8: calcareous concretions; 9: prismatic structure; 10: fire-site; 11: sample; 12: sedimentary unit; 13: 14C-analysis (calibrated ka BP); 14: ceramics (calendar ka); 15: artifacts (calendar ka); 16: shell (Helix).
1 : grès miocène ; 2 : argiles et limons fins ; 3 : limons ; 4 : sables ; 5 : graviers ; 6 : sédiments organiques ; 7 : traces d’hydromorphie ; 8 : concrétions calcaires ; 9 : structure prismatique ; 10 : foyer ; 11 : échantillon ; 12 : unité chronostratigraphique ; 13 : datation 14C (ka calibrés BP) ; 14 : céramiques (ka calendaires) ; 15 : artéfacts (ka calendaires) ; 16 : coquille d’Helix.

Fig. 5 – Safhit El-Flalig (SEF). See the captions in fig. 4.
Fig. 5 – La coupe de Safhit El-Flalig (SEF). Voir la légende sur fig. 4.

Fig. 5 – Safhit El-Flalig (SEF). See the captions in fig. 4.Fig. 5 – La coupe de Safhit El-Flalig (SEF). Voir la légende sur fig. 4.

5Wadi Achihb profile (ACH; fig. 6) is located downstream of Wadi Achihb at its confluence with the Wadi Douleb. Wadi Achihb is a small tributary of Wadi Douleb which drains a small sub-basin with an area of 12 ha. This profile begins with a Capsian palaeosoil (3.7% organic matter, 1.32% of free iron, 3.5% of total iron) (ACH3; fig. 6) covering Miocene sandstones (ACH1). This palaeosoil contains a Capsian site, which has produced hundreds of artifacts in situ of which a few dozen have been identified by A. Gragueb and are attributed to typical Capsian (fig. 6A). The Capsian palaeosoil is covered by a historic deposit ACH4, containing at the base, late Roman ceramics which have been dated to 1580±50 cal. BP (fig. 6). A few yards upstream, a deposit equivalent to the historical deposit (ACH4) was blocked by a Roman aqueduct (photo 1).

Fig. 6 – Wadi Achihb profile (ACH).
Fig. 6 – La coupe de l’oued Achihb (ACH).

Fig. 6 – Wadi Achihb profile (ACH).Fig. 6 – La coupe de l’oued Achihb (ACH).

A: Capsian artifacts (ACH3). See the captions in fig. 4.
A : Artéfacts capsiens (ACH3). Voir la légende sur fig. 4.

6The Wadi El-Hammar profile (HAM; fig. 7) is one of the most interesting profiles of the Wadi Sbeïtla basin. Indeed, this profile is characterised by its wealth of chronological data (radiocarbon dating, a Capsian site and Roman pottery in situ; fig. 7). Its uniqueness is also due to its position in interfluves where colluvial deposits were accumulated by surface-wash deposition. With a total thickness of 2 m, the HAM profile begins with an indurated deposit of prismatic structure and rich in calcareous concretions (HAM1). This deposit is covered by a silty sand layer containing at the base several in situ Capsian artifacts (HAM2a). These Capsian artifacts are deposited in bulk and do not show indication of reworking (i.e., the artifacts are not ordered). These artifacts are associated with shells of Helix aperta. Specific holes on these shells suggest that they were picked up by Caspian people for food (fig. 7B). Five artifacts are identified by A. Gragueb and are attributed to typical Capsian (fig. 7C). A dark and thinner deposit (HAM3) covers previous accumulation (fig. 7). This deposit shows pedogenetic features, such as hydromorphic features and plant remains, contains about 3.8% of organic matter and a free iron/total iron report reaches 0.41 (1.27-3.02%). SEM observations showed a deposit saturated in amorphous silica of geometric shape (fig. 7D), dissolution forms (fig. 7E) and deposits of carbonate (fig. 7F). This amorphous silica on the surface of quartz grains indicates a long stability of deposits followed by an obvious pedogenesis. A radiocarbon dating from charcoal, sampled at the top of this palaeosoil, has determined its age to 4242±385 cal. BP. The deposit above it (HAM4) is rich in organic matter (4.8%). Finally, a dark deposit (HAM5a) rich in organic matter (4.8%) and in free iron (1.77%) is covered by a layer (5b) containing some Roman pottery dated to 1385±35 BP (fig. 7).

Fig. 7 – Historical deposits blocked by a Roman aqueduct.
Fig. 7 – Dépôts historiques bloqués par un aqueduc romain.

Fig. 7 – Historical deposits blocked by a Roman aqueduct. Fig. 7 – Dépôts historiques bloqués par un aqueduc romain.

7Two profiles are situated in the downstream of Wadi Sbeïtla: SBV and SBVI (fig. 8 and fig. 9). The SBV profile has three radiocarbon dates upon which its chronology is based (fig. 8). This profile begins with a limestone crust dated at the top to 11.7 ka cal. BP (SBV1). Coarser sands, rich in calcareous concretions (SBV2), cover the deposit SBV1 and above a dark deposit rich in organic matter (SBV3). A radiocarbon dating carried out on the deposit SBV3 gave an age of 7.2 ka cal. BP (fig. 8). SBV3deposit is covered through a phase of erosion reaching a palaeochannel filled of historical deposits (SBV4; fig. 8). These historical deposits are composed of 6 layers, with different grain-size and chemical characteristics (SBV4 a, b, c, d, e and f). A coarser layer (SBV4a) is covered by a loamy soil (4b) dated on charcoal to 1.2 ka cal. BP (Erl-7907). This loamy soil contains 3.1% of organic matter, 1.12% of free iron and 2.72% of total iron (fig. 8). SEM photos revealed quartz grains covered with amorphous silica, indicating a long stabilization of deposits. Small branches of amorphous silica (fig. 8H) extend to cover the entire quartz grain (fig. 8G). Deposits in SBVI profile resemble those of the SBV profile. The layers at the bottom are disturbed by a large hole over 1 m deep and 1.5 m wide (fig. 9). At the top of this hole a large bone of 9 cm in diameter is dated to 7.5 ka cal. BP (Erl-7905) (fig. 9I). The important diameter of this bone may be indicates that it correspond to a large mammal probably an elephant or a rhinoceros, which existed in North Africa during the humid periods of the early Holocene (Jamet, 1989). The shape and the size of the hole containing the bone, indicate that this hole is not a tomb, because its filling was done naturally and slowly, unlike the artificial and immediate filling characterizing the tombs. This natural and slow filling is indicated by small sandy-clay layers that are intercalated at the base of the hole (fig. 9J).

Fig. 8 – Wadi El-Hammar profile (HAM).
Fig. 8 – La coupe de l’oued El-Hammar (HAM).

Fig. 8 – Wadi El-Hammar profile (HAM).Fig. 8 – La coupe de l’oued El-Hammar (HAM).

B: Specific holes on Helix aperta. C: Capsian artifacts (HAM2a). D: Amorphous silica of geometric shape. E: Dissolution forms. F: deposits of carbonate. See the captions in fig. 4.
B : Trous spécifiques sur des Helix aperta. C : Artéfacts capsiens (HAM2a). D : Silice amorphe de forme géométrique. E : Formes de dissolution. F : Dépôts de carbonates. Voir la légende sur fig. 4.

Fig. 9 – Wadi El-Hammar profile (HAM).
Fig. 9 – La coupe de l’oued El-Hammar (HAM).

Fig. 9 – Wadi El-Hammar profile (HAM).Fig. 9 – La coupe de l’oued El-Hammar (HAM).

G: Quartz grain covered with amorphous silica. H: Amorphous silica. See the captions in fig. 4.
G : Grains de quartz recouverts de silice amorphe. H : Silice amorphe. Voir la légende sur la fig. 4.

Chronostratigraphical correlations: a synthetic approach

8The correlation between the different studied profiles helped to establish a synthetic profile of Holocene alluvial archives in the Wadi Sbeïtla Basin (fig. 10 and fig. 11). In this synthetic profile four Holocene phases of accumulation have been identified. A probable fifth phase of Holocene accumulation characterises the study area, but its deposits are not widespread (fig. 10 and fig. 11). Deposits of this accumulation phase appear only in the two profiles of SEF (4f) and SBV (4e and 4f; fig. 5 and fig. 8). These phases of accumulation were determined according to morphological and pedological concepts. Each phase corresponds to a cycle of sedimentation, beginning with a fluvial activity and followed by a geomorphic stability favouring a fine-grained sediment and soil formation. The completion of some phases of accumulation was marked by alluvial incision, eroding the soil deposits, and marking the aggradation of the next accumulation phase. Thus, each phase of accumulation is composed of two parts, a coarser sedimentation at the base and a fine sedimentation at the top (fig. 10 and fig. 11). Each of these five phases of Holocene accumulation is not necessarily followed by a phase of incision, which explains the existence of only four Holocene terraces, of which three are historical terraces.

Fig. 10 – SBV profile.
Fig. 10 – La coupe SBV.

Fig. 10 – SBV profile.Fig. 10 – La coupe SBV.

I: Large bone of 9 cm of diametre. J: Intercalation of small sandy-clay layers at the base of the hole. See also the captions in fig. 4.
I : Grand os de 9 cm de diamètre. J : Intercalation de fines couches sablo-argileuses à la base du trou. Voir la légende sur fig. 4.

Fig. 11 – Synthetic stratigraphic section of Holocene alluvial deposits.
Fig. 11 – Coupe stratigraphique synthétique des dépôts alluviaux holocènes.

Fig. 11 – Synthetic stratigraphic section of Holocene alluvial deposits.Fig. 11 – Coupe stratigraphique synthétique des dépôts alluviaux holocènes.

1: clayey to silty layer; 2: loam; 3: sand; 4: gravel layer; 5: fire site; 6: humic enrichment; 7: hydromorphic features; 8: calcareous concretions; 9: calcareous crust; 10: 14C-analysis (calibrated ka BP); 11: ceramics (calendar ka); 12: artifact (calendar ka); 13: flooding/aggradation; 14: river incision; 15: stable fluvial dynamics; 16: active fluvial dynamics; 17: approximate transition.
1 : argiles et limons fins ; 2 : limons ; 3 : sables ; 4 : graviers ; 5 : traces de feu ; 6 : sédiments organiques ; 7 : indices d’hydromorphie ; 8 : concrétions calcaires ; 9 : croûte calcaire ; 10 : datation 14C (ka calibrées BP) ; 11 : céramiques (ka calendaires) ; 12 : artéfacts (ka calendaires) ; 13 : remblaiement ; 14 : incision ; 15 : dynamique fluviale stable ; 16 : dynamique fluviale active ; 17 : limite imprécise.

9The early Holocene accumulation phase (before 7 ka cal. BP) covers 2 to 3 thousand years and begins with coarser sedimentation (sand and gravel, SBV2; fig. 8). At the top, the sediments become finer and a soil formation is developed before its erosion by a phase of fluvial incision. These palaeosoils have been dated between 7.2 ka cal. BP (SBV3) to 7.5 ka cal. BP (SBVI3). These palaeosoils are due to increased moisture and improved environmental conditions during the Holocene optimum around the 7th to 8th millennium cal. BP (fig. 10 and fig. 11). The first Holocene phase of accumulation is superimposed on a deposit rich in carbonate content, including at the top shells dated to 11.7 ka cal. BP (SBV1; fig. 8).

10The mid-Holocene accumulation phase (6.5-4 ka cal. BP) begins with relatively coarser sediments, indicating an increase of fluvial activity (accumulation phase number 2-a; fig. 10 and fig. 11). The coarser deposits at the base become gradually finer at the top and in within which a palaeosoil was formed, characterised by 4% of organic matter and a sedimentation rate of 1 mm per year (accumulation phase 2-b; fig. 10 and fig. 11). A shell (Helix) sampled at the top of this palaeosoil was dated to 4.2 ka cal. BP. Deposits of the two previous phases of accumulation are often superposed and form the prehistoric Holocene terrace (Zerai, 2006). The absence of a significant river incision at the top of the first accumulation phase may explain the existence of a single prehistoric Holocene terrace.

11The late Roman accumulation phase (1.6-1.1 ka cal. BP) followed a major river incision period, which is responsible for the erosion of deposits of the two previous depositional phases in several profiles (SEF, SBV). Deposits accumulated between the second and third phase of accumulation are very rare in the studied area, because they were eroded by this river incision. These deposits are indicated by a question mark in fig.10. An important contrast is observed between the two periods (a and b) of the fluvial dynamics of this phase of accumulation (fig. 10 and fig. 11). The first period (a) is characterised by very coarse deposits (sand, gravel and boulder). The second period (b) is characterised by palaeosoils rich in organic matter and hydromorphic features. These palaeosoils have been dated on charcoal between 1.1 to 1.2 ka cal. BP (fig. 5 and fig. 8). The deposits of the third accumulation phase form the main historical terrace in the Sbeïtla Wadi basin (Zerai, 2006). The deposits of the third accumulation phase, which covers only 5 centuries, have almost the same thickness as the deposits of the two previous accumulation phases, covering together approximately 5 millennia (fig. 10 and fig. 11). The late Mediaeval accumulation phase (1-0.5 ka cal. BP) began after the tenth century AD. It succeeds a period of river incision, highlighting the main historical terrace (Zerai, 2006). At the upper limit of the fourth accumulation phase, dark deposit rich in organic matter was observed in the layers SEF4e and SBV4d (fig. 5 and fig. 8). Deposits of the fourth accumulation phase constitute the second historical terrace.

12A probable fifth Holocene accumulation phase started about five centuries ago and continued until present (fig. 10 and fig. 11). Its deposits are not very remarkable in the field. Deposits of this accumulation phase appear only in the two profiles of SEF and SBV (SEF4f et SBV4e et 4f; fig. 5 and fig. 8). Deposits of the fifth accumulation phase probably constitute the third historical terrace. The fifth accumulation phase may be related to the current development of floodplain rivers in the studied area. During this period, the environments of the studied area have been significantly degraded with the amplification of anthropogenic stress, the increase of aridity and the mobilization of aeolian sand.

Discussion: Holocene palaeoenvironmental changes, human or climate impact?

13During Late glacial-Holocene transition, an aridification and increased fluvial activity occurred. In the downstream of Wadi Sbeïtla basin, a shell collected in a limestone crust at the base of SBV profile, are dated to 11.7 ka cal. BP (fig. 8). SEM photos show traces of carbonates on the quartz grains, indicating a deposit saturated in CaCO3 (fig. 7F). This indicates that drier conditions characterised central Tunisia about 12 ka cal. BP. This drier period has also been indicated by J.-C. Fontes and F. Gasse (1989), J.-L. Ballais (1991), K. White et al. (1996), C. Swezey (2001) and C. Zielhofer et al. (2004). According to C. Swezey (2001), mobilization of aeolian deposits characterised southern Tunisia before 11 ka cal. BP. In North Africa, detailed studies by F. Gasse et al. (1990) in Sebkha Mellala in Saharan Algeria and from mountainous Mediterranean zone of Morocco (Lamb and Van der kaars, 1995) show an aridification during the same period. According to F. Gasse et al. (1990), the Sebkhet el-Mellala, Algeria, was characterised by extreme aridification to 12.5 ka cal. BP. Elsewhere in Europe, M. Magny et al. (2007) indicate a drier climate developed in southern Europe around 11.3 ka BP, relying on lake-level data in north-central Italy. Drier conditions in semi-arid Mediterranean river systems correspond with increased geomorphic activity (e.g., White et al., 1996). Coarser deposits, accumulated by a torrential fluvial process, are often observed at the base of some profiles of the studied basin (HAM1; SBV2). These coarser deposits were dated on shell at the top to 10.2 ka cal. BP (Zerai, 2006). On an average thickness of 1 m, these coarser deposits represent the beginning of the first Holocene accumulation phase, where the materials become thinner towards the top (fig. 10 and fig. 11). An increase in geomorphic activity during the same period has been shown in northern Tunisia (Bourgou, 1993; Zielhofer et al., 2004). About ten kilometres to the west of the studied basin, pollen records in the region of kasserine revealed a dry and cold period before 10.5 ka cal. BP (Medus and Laval, 1997).

14Geomorphic stability and increased humidity is demonstrated during Early Holocene. The deposits are generally characterised by a clayey to silty texture and hydromorphic features, indicating calm depositional environments and geomorphic stability (SEF, ACH, SBV and SBVI; fig. 10 and fig. 11). At the top, they are transformed into palaeosoils, rich in organic matter and containing burn marks (fig. 6 and fig. 7). These palaeosoils have been dated by two radiocarbon datings (7.2 to 7.5 ka cal. BP) and two in situ Capsian sites (ACH3, HAM2a; fig. 10 and fig. 11). According to J.-L. Ballais (1991), hydromorphic features of these palaeosoils indicate abundant hydrography during this period of the Holocene. They are explained by a high groundwater level in the Medjerda floodplain (Zeilhofer et al., 2004). The increased humidity during early Holocene in the catchment area of Sbeïtla Wadi is consistent with the reappearance of lakes in southern Tunisia (Ballais, 1991), with humid period between 10 to 7 ka BP in the region of Meknessy, central Tunisia (Ouda et al., 1998). In the southern Tunisia, N. Petit-Maire et al. (1991) have dated the reappearance of lakes to 8230±70 BP. Elsewhere in Europe, M.G. Macklin et al. (2006) indicate several periods of major flooding in Great Britain, Spain and Poland during the early Holocene. In southern Spain, eight episodes of flooding are recorded during early Holocene (Macklin et al., 2006).

15The Mid-Holocene was characterised by stable environmental conditions and soil formation. The deposits are generally fine and include several soil formation features (3.8% of organic matter, 1.2% of free iron, hydromorphic features and plant remains). SEM photos revealed quartz grains covered with amorphous silica, indicating a long stabilization of deposits (fig. 6D). The malacology analysis indicates species requiring a minimum of moisture to survive and usually thrive in open spaces, such as Helix aperta. This palaeosoil is dated at the top to 4.2 ka cal. BP. All these characteristics of mid-Holocene deposits indicate that the climate was humid and warm. The name of this period varies according between researchers, but still indicates an increase in humidity. It is termed biostasic circumstance (Coudé-Gaussen, 1984), biostasic situation (Bourgou, 1993) and mid-Holocene soil formation (Zeilhofer et al., 2004). According to C. Zielhofer et al., (2008), the average Medjerda sedimentation rate is reduced, indicating weak fluvial dynamics and favourable conditions for soil formation between 6 and 4.7 ka BP. In Mediterranean North Africa, a strong Mid-Holocene soil formation has been noted by several researchers (Molle, 1979; Scharpenseel and Zakosek, 1979; Zielhofer et al., 2002). This mid-Holocene soil formation period was further correlated with a hygric and thermal climatic optimum in the region (Zielhofer et al., 2008).

16Arid conditions occurred during the prehistoric late Holocene. A large sedimentation hiatus is well marked and is due to a major pre-Roman river incision (fig. 5 and fig. 8) leading to a discordance, which connects historical deposits with the mid-Holocene palaeosoil (fig. 8). Deposits related to this arid period of the Holocene appear in HAM profile (fig. 7). These deposits (HAM4) contain a small proportion of organic matter and free iron, compared to two soil formations, underlying (HAM3) and overlying (HAM5a). This dry condition probably agrees with a deflation period generated in pre-Saharan Tunisia before 3 ka BP (Ballais, 1991). In North Africa, J.-C. Ritchie (1984) indicates a significant decrease in pollen of Quercus and an increase in grasses in eastern Algeria after 4ka BP.

17Pre-Roman to Roman increased humidity was recorded. Considering the major phase of pre-Roman river incision, deposits related to this historical period are rare. Some researchers indicate a more rainfall during the first century AD (Poncet, 1958; Barbery and Delhoume, 1982). According to J. Barbery and J.-P. Delhoume (1982), the volume of water on the southern foothills of Mrhila Mountain is too large to have been used solely for domestic purposes. The pre-Roman to Roman increased humidity in the studied area is consistent to short wet phase, around 2400 BP, in southern Tunisia (Ballais, 1991) and to a Roman soil formation in the Medjerda Valley in northern Tunisia (Faust et al., 2004; fig. 10 and fig. 11). In eastern Algeria, the valley of Chéria-Mezeraa Wadi experienced the same humid period between 2400 and 2200 BP (Ballais and Benazzouz, 1994). Elsewhere in northwest Europe, B. Van Geel et al., (1996) indicate an abrupt climate change around 2650 BP in the Netherlands, even in the European continent and other continents. The climate changed from relatively warm and continental to oceanic. As a consequence, the ground-water table rose considerably in certain low-lying areas in The Netherlands (Van Geel et al., 1996). A change in solar activity and global changes in oceanic and atmospheric circulation could explain the observed climatic changes in northwest Europe around 2650 BP (Van Geel et al., 1996 and 2004).

18Deposits related to the Late to post-Roman crisis (1.6-1.2 ka cal. BP) are coarse and poor in organic matter. They correspond to the lower part (a) of the third Holocene accumulation phase (fig. 10 and fig. 11). In the same profile, the sedimentation rate during this period reached 17 mm/a indicating a sedimentation rate among the highest in Tunisia, because the average rate is about 7.4 mm/a (Ballais, 1991). These deposits, containing late Roman pottery, have been dated between 1.6 and 1.2 ka cal. BP (fig. 10 and fig. 11). The geomorphic activity, corresponding to the coarser deposits, is very clear, but the factors that have caused it are complex. The beginning of the post-Roman crisis has followed a rural abandonment due to the decline of Roman Empire. According to F. Decret and H. Fantar (1981) this situation continued with the arrival of the Byzantines. During this period, coarser deposits began to develop reaching final stage of development between the eighth and ninth century (1.3 to 1.2 ka cal. BP).

19Soil Formation occurred from 1.1 ka cal. BP. The coarser deposits of the late to post-Roman crisis are superimposed by soil formation (amorphous silica in SEM photos, 2.7% of organic matter, hydromorphic features and plant remains; SEF, SBV and SBVI profiles). This soil corresponds to the upper part (b) of the third Holocene accumulation phase (fig. 10 and fig. 11). All these chemical and sedimentological characteristics indicate a short wet period. Two radiocarbon datings on charcoal indicated that the age of this wet period ranges between 1.1 and 1.2 ka cal. BP. This humid period lasted about two centuries between the ninth and tenth centuries AD, roughly during the Aghlabid Dynasty (AD 800-909). During the short humid period from 1.1 ka cal. BP, the climatic factor had an important role in the geomorphic stability. The same wet phase was reported in northern Tunisia by D. Faust et al. (2004), indicating a soil formation in the Medjerda Valley around 1.3 ka cal. BP, roughly during the eighth century. Central Tunisia experienced a period of prosperity and expansion of trees until the eleventh century (Despois, 1955; Fehri et al., 2007).

20Progressive degradation of environments was recorded between the tenth and the nineteenth century. Except during the colonial and postcolonial period, the dominant landscape in central Tunisia during the last ten centuries is typical steppe (Ballais, 2000; Fehri et al., 2007). Since the nomad invasion of Beni Hilal, coming from Upper Egypt in the mid-eleventh century (Ibn Khaldoun, 1968; Ballais, 2000) the steppes of retam, alfa and jujube have expanded after the decline of trees, although they were prosperous before the eleventh century (Despois, 1955; Fehri et al., 2007). This change of landscape and the prevalence of insecurity have led to the decline of trees and expansion of nomadism. The transition to transhumance necessitated the adoption of livestock, which became intense (Abdu-Wahab, 1954). A. Brun (1983) indicates grazing around the Gulf of Gabes in the same period. According to K.W. Butzer (1961), this degradation of the environment coincided with a tendency towards dry climate during the late 11th century. From the beginning of the tenth century the studied area showed coarser deposits, low in organic matter (fig. 10 and fig. 11). These coarser deposits constitute the second historical terrace (Zerai, 2006).

21From the end of the nineteenth century (e.g., since the beginning of colonial period), the Wadi Sbeïtla basin experienced a considerable expansion of agriculture, especially planting trees. The decree of February 8, 1892 facilitates to farmers the appropriation of land, including public lands (Fehri et al., 2007). The downstream part of the Wadi Sbeïtla basin (about 8000 ha), has been developed for the cultivation of trees from 1919 ha in 1948 to 3550 ha in 1988. This expansion of trees was done mainly at the expense of rangelands, which decreased during the same period from 3819 ha to 2639 ha (Zerai, 2006). The clearing, grazing and mechanization of agriculture have increased the effect of wind activity in the region. The same downstream part of the Wadi Sbeïtla basin has witnessed the expansion of 100 ha of dunes between 1948 and 1988. A comparative study of aerial photographs revealed a general trend towards river incision and landward erosion during the twentieth century (Zerai and Coque, 2004). This fluvial incision is represented by the third historical terrace and a system of two benches in the convex parts of meanders.

Conclusions

22Detailed study of Holocene alluvial records in Wadi Sbeïtla indicates multiple phases of accumulation and fluvial incision and succession of several humid and dry periods. The different phases of accumulation are not necessarily followed by phases of fluvial incision, which explains the existence of a single prehistoric terrace and three historical terraces (fig. 10 and fig. 11). The number of Holocene terraces, relatively important geomorphic features compared to the southern and northern Tunisia, a specific to central Tunisia. According to J.-L. Ballais (1993), the number of Holocene terraces in Tunisia varies in accordance with the bioclimatic zones; a single terrace in the Saharan stage, two terraces in the arid stage, three terraces in the semi-arid stage and one terrace in subhumid stage. The multiple phases of accumulation and incision indicate several major changes in geomorphic dynamics during the Holocene (fig. 10 and fig. 11). Geomorphic stability results from humid conditions and dense vegetation cover, allowing soil formation. Whereas, geomorphic activity is characterised by soil-sediment erosion, leading to high sedimentation rates. On the other hand, the succession of humid and dry periods reveals several Holocene climatic fluctuations. Two long humid periods characterise the early and mid-Holocene. A short humid phase occurred around 1.1 ka cal. BP (fig. 10 and fig. 11). These humid periods were separated by dry phases, especially during the late Holocene. Succession of humid and dry periods during the Holocene in Wadi Sbeïtla is due to major climatic oscillations. The large-scale study of these major climatic oscillations can show several other minor variations in climate. For example, M. Magny (2004) indicates 15 phases of higher level in 26 lakes studied in the Jura Mountains, the northern French Pre-Alps and the Swiss Plateau. During the early and mid-Holocene, the main factor influencing geomorphic dynamics was the climate followed by human impact. During historical periods, the role of anthropogenic impacts is important, but climate has intensified or attenuated these human impacts.

Haut de page

Bibliographie

Abdul-Wahab H.H. (1954) – Les steppes tunisiennes (région de Gammouda) pendant le Moyen Age. Les Cahiers de Tunisie, 5, 5-16.

Ballais J.-L. (1991) – Evolution holocène de la Tunisie saharienne et présaharienne. Méditerranée, 4, 31-38.

Ballais J.-L. (1993) – Morphogenèse fluviatile holocène en Tunisie. Travaux URA 903 du CNRS, 22, 63-78.

Ballais J.-L. (1995) – Alluvial Holocene terraces in eastern Maghreb: climate and anthropogenic controls. In Lewin J., Macklin M.G.,Woodward J.C. (Ed.): Mediterranean Quaternary river environments. Balkema, Rotterdam, 183-194.

Ballais J.-L. (2000) – Conquests and land degradation in the eastern Maghreb during classical antiquity and the Middle Ages. In Barker G., Gilbertson D. (Ed.): The Archaeology of Drylands, One World Archaeology, 39, Routledge, London, 125-136.

Ballais J.-L., Benazzouz M.T. (1994) – Données nouvelles sur la morphogenèse et les paléoenvironnements tardiglaciaires et holocènes dans la vallée de l'oued Chéria-Mezeraa (Nemencha, Algérie orientale). Méditerranée, 80, 3-4, 59-71.

Barbery J., Delhoume J.-P. (1982) – La voie romaine de piémont, Suffetula-Masclianae (dj. Mrhila, Tunisie centrale). Antiquités africaines, 18, 27-43.

Bejaoui F. (2003) Histoire des hautes steppes, Antiquité-Moyen Age. Actes du colloque de Sbeitla, République tunisienne, Ministère de la Culture, Institut national du Patrimoine, 232 p.

Bourgou M. (1993)Le bassin-versant du Kébir-Miliane (Tunisie Nord-orientale) : étude géomorphologique. Publications de l’Université de Tunis, 2, 33, 435 p.

Brun A. (1983) – Etude palynologique des sédiments marins holocènes de 5000 B.P. à l’actuel dans le golfe de Gabès (Mer Pélagienne). Pollen et Spores, 25, 3-4.

Butzer K.W. (1961) – Les changements climatiques dans les régions arides depuis le Pliocène. Recherche sur la zone aride, UNESCO, 17, 35-64.

Coudé-Gaussen G. (1984) – Mise en place des basses terrasses holocènes dans les Matmata et leurs bordures (Sud tunisien). Bulletin de l’Association Française pour l’Etude du Quaternaire, 1-2-3, 173-180.

Decret F., Fantar H. (1981)L’Afrique du Nord dans l’Antiquité : des origines au Ve siècle. Payot, 391 p.

Despois J. (1955) La Tunisie orientale, Sahel et Basse Steppe. Étude géographique. PUF, Paris (2ème édition), 554 p.

Faust D., Zielhofer C., Baena R., Diaz del Olmo F. (2004) – High-resolution fluvial record of late Holocene geomorphic change in northern Tunisia: climatic or human impact ? Quaternary Science Reviews 23, 1757-1775.

Fehri N., Ballais J.-L., Bonifay M. (2007) – Morphogenèse, climat et société dans la plaine de Sfax (Tunisie) depuis le Pléistocène supérieur: l’exemple du bassin-versant de l’oued Châal-Tarfaoui. Physio-Géo, 1, 61-77.

Fontes J.-C., Gasse F. (1989) – On the age of the humid Holocene in late Pleistocene phases in North Africa: remarks on “Late Quaternary climatic reconstitution on the Maghreb (North Africa)” by P. Rognon. Palaeogeography, palaeoclimatology, palaeoecology 70, 393-398.

Gasse F., Van Campo E. (1994) – Abrupt post-glacial climate events in West Asia and North Africa monsoon domains. Earth and Planetary Science Letters 126, 435-456.

Gasse F., Téhet R., Durand A., Gibert E., Fontes J.-C. (1990) – The arid-humid transition in the Sahara and the Sahel during the last deglaciation. Nature 346, 141-146.

Goodfriend G. (1987) – Radiocarbon age anomalies in shell carbonate of land snails from semi-arid areas. Radiocarbon, 29, 2, 159-167.

Hitchner R.B., Ellis S., Graham A.D., Mattingly L. (1990) – The kasserine Archaeological Survey. Antiquités Africaines, 26, 231-260.

Ibn Khaldoun A. (1968) Muqqadima. Beirut, Commission Libanaise pour la traduction des Chefs d’œuvre, translated by V. Monteil.

Jamet P. (1989) – Carte paléoenvironnementale du Sahara holocène. Annales des Mines et de la Géologie, 34, 59-66.

Lamb H.F., Van der kaars S. (1995) – Vegetational response to Holocene climatic change: pollen and palaeolimnological data from the Middle Atlas, Morocco. The Holocene 5, 400-408.

Le Houérou H.N. (1969) – La végétation de la Tunisie steppique: avec références au Maroc, à l'Algérie et à la Libye. Annales de l'Institut National de la Recherche Agronomique, 42, 5, 622 p.

Macklin M.G., Benito G., Gregory K.J., Johnstone E., Lewin J., Michczyńska D.J., Soja R., Starkel L., Thorndycraft V.R. (2006) – Past hydrological events reflected in the Holocene fluvial record of Europe. Catena 66, 1-2, 145-154.

Magny M. (2004) – Holocene climate variability as reflected by mid-European lake-level fluctuations and its probable impact on prehistoric human settlements. Quaternary International 113, 65-79.

Magny M., Vannière B., De Beaulieu J.-L., Bégeot C., Heiri O., Millet L., Peyron O., Walter-Simonnet A.-V. (2007) – Early-Holocene climatic oscillations recorded by lake-level fluctuations in west-central Europe and in central Italy. Quaternary Science Reviews 26, 1951-1964.

Marquer L., Pomel S., Abichou A., Schulz E., kaniewski D., Van Campo E. (2008) – Late Holocene high resolution palaeoclimatic reconstruction inferred from Sebkha Mhabeul, southeast Tunisia. Quaternary Research 70, 2, 240-250.

Medus J., Laval H. (1997) – Transition palynologique tardiglaciaire-holocène dans un site de Tunisie méridionale. Archives des Sciences et Compte Rendus des Séances de la Société de Physique et d’Histoire Naturelle de Genève, 50, 17-26.

Mehra O.P., Jackson M.L. (1960) – Iron oxide removal from soil clays by dithionite-citrate system buffered with sodium bicarbonates. Clays and Clay Minerals, 317 p.

Moldenhauer K.M., Zielhofer C., Faust F. (2008) – Heavy metals as indicators for Holocene sediment provenance in a semi-arid Mediterranean catchment in northern Tunisia. Quaternary International 189, 1, 129-134.

Molle H.G. (1979) – Untersuchungen zur Entwicklung der vorzeitlichen Morphodynamik im Tibesti-Gebirge (Zentralsahara) und in Tunesien. Berliner Geographische Abhandlungen 25, 96 p.

Ouda B., Zouari K., Ben Ouezdou H., Chkir N., Causse C. (1998) – Nouvelles données paléoenvironnementales pour le Quaternaire récent en Tunisie centrale (Bassin de Maknassy). Comptes Rendus de l’Académie des Sciences de Paris, II a, 326, 12, 855-861.

Oueslati A. (2005) – The evolution of valley floors and ancient societies adaptation to natural constraints: A geoarcheological approach from the Majerda catchment (North Tunisia). PAGES Second Open Science Meeting, 10-12 August 2005, Beijing, China, Abstracts, 1 p.

Petit Maire N., Burollet P.-F., Ballais J.-L., Fontugne M., Rosso J.-C., Lazaar A. (1991) – Paléoclimats holocènes du Sahara septentrional: dépôts lacustres et terrasses alluviales en bordures de Grand Erg Oriental à l’extrême sud de la Tunisie. Comptes Rendus de l’Académie des Sciences de Paris, 312, 13, 1661-1666.

Poncet J. (1958)Rapport entre les modes d'exploitation agricole et l'érosion des sols en Tunisie. Thèse de doctorat, Paris, 245 p.

Ritchie J.-C. (1984) – Analyse pollinique de sédiments holocènes supérieurs des hauts plateaux du Maghreb oriental. Pollen et Spores, 24, 3-4, 489-496.

Romaniello L., Quarta G., Mastronuzzi G., D’Elia M., Calcagnile L. (2008)14C age anomalies in modern land snails shell carbonate from Southern Italy. Quaternary Geochronology 3, 1-2, 68-75.

Scharpenseel H.W., Zakosek H. (1979) – Phasen der Bodenbildung in Tunesien. Zeitschrift für Geomorphologie N.F. 33 (Supplement), 118-126.

Stevenson A.C., Phethean S.J., Robinson J.E. (1993) – The palaeosalinity and vegetational history of Garaet el Ichkeul, northwest Tunisia. The Holocene 3, 3, 201-210.

Stuiver M., Reimer P.J. (2005) – Radiocarbon Calibration Program Calib Rev 5.0 (http://calib.qub.ac.uk/calib/calib.html).

Stuiver M., Reimer P.J., Bard E., Beck J.W., Burr G.S., Hughen K.A., Kromer B., McCormac F.G., Plicht J., Spurk M. (1998) – INTCAL 98 radiocarbon age calibration, 24,000 to 0 cal. BP. Radiocarbon 40, 1041-1083.

Swezey C. (2001) – Aeolian sediment responses to late Quaternary climate changes: temporal and spatial patterns in the Sahara. Palaeogeography, Palaeoclimatology, Palaeoecology 167, 119-155.

Swezey C., Lancaster N., Kocurek G., Deynoux M., Blum M., Price D., Pion J.C. (1999) – Response of aeolian systems to Holocene climatic and hydrologic changes on the northern margin of the Sahara: a high-resolution record from the Chott Rharsa basin,Tunisia. The Holocene 9, 141-147.

Van Geel B., Buurman J., Waterbolk H.T. (1996) – Archaeological and palaeoecological indications for an abrupt climate change in the Netherlands and evidence for climatological teleconnections around 2650 BP. Journal of Quaternary Science 11, 451-460.

Van Geel B., Bokovenko N.A., Burova N. D., Chugunov K.V., Dergachev V.A., Dirksen V.G., Kulkova M., Nagler A., Parzinger H., Van der Plicht J., Vasiliev S.S., Zaitseva G.I. (2004) – Climate change and the expansion of the Scythian culture after 850 BC: a hypothesis. Journal of Archaeological Science 31, 12, 1735-1742.

Vermeersch P. (1973) – Résultats d’une prospection préhistorique dans le bassin de kasserine (Tunisie Steppique). Annales des Mines et de la géologie, Tunis, 26, 607-620.

White K., Drake N., Millington A., Stokes S. (1996) – Constraining the timing of alluvial fan response to Late Quaternary climatic changes, southern Tunisia. Geomorphology 17, 295-304.

Zerai K. (2006) Les environnements holocènes et actuels dans le bassin-versant de l’oued Sbeïtla (Tunisie centrale). Thèse de doctorat, université Paris-Diderot (Paris 7), 333 p.

Zerai K., Coque B. (2004) – Evolution de l’érosion hydrique dans le bassin-versant de l’oued Sbeïtla (Tunisie centrale) au cours de la seconde moitié du 20ème siècle. Etude quantitative et modélisation à partir d’un SIG. Mosella, XXIX, 3-4, 271-295.

Zielhofer C., Faust D. (2008) – Mid- and Late Holocene fluvial chronology of Tunisia. Quaternary Science Reviews 27, 5-6, 580-588.

Zielhofer C., Faust D., Diaz del Olmo F., Baena R. (2002) – Sedimentation and soil formation phases in the Ghardimaou Basin (Northern Tunisia) during the Holocene. Quaternary International 93-94, 109-125.

Zielhofer C., Faust D., Baena R., Diaz del Olmo F., kadereit A., Moldenhauer K.M., Porras A.(2004) – Centennial-scale late Pleistocene to mid-Holocene synthetic profile of the Medjerda floodplain (Northern Tunisia). The Holocene 14, 851-861.

Zielhofer C., Faust D., Linstadter J. (2008) – Late Pleistocene and Holocene alluvial archives in the Southwestern Mediterranean: Changes in fluvial dynamics and past human response. Quaternary International 181, 1, 39-54.

Haut de page

Annexe

  

Version française abrégée

Cet article porte sur l’étude des archives alluviales holocènes dans le bassin de l’oued Sbeïtla et l’établissement d’une échelle chronostratigraphique détaillée. Le bassin de l’oued Sbeïtla se situe au contact entre la Dorsale tunisienne et les hautes steppes. Il est dominé dans sa partie amont par des volumes montagneux importants où culmine le djebel Semmama à 1356 m. La forme du bassin, perpendiculaire à la direction atlasique des alignements montagneux (fig. 1), a influencé les différentes caractéristiques physiques du bassin, tels que les pentes (fig. 2A), la lithologie (fig. 2B), le réseau hydrographique (fig. 2C) et le couvert végétal (fig. 2D). Le climat du secteur étudié est semi-aride. Il tombe, de manière très irrégulière et avec une forte intensité, 273 mm de précipitations moyennes annuelles, la température moyenne annuelle est de l’ordre de 17,2 °C et la direction des vents dominants est NW-SE.

Les cinq coupes présentées dans cet article ont été choisies parmi une vingtaine de coupes étudiées dans tout le bassin-versant de l’oued Sbeïtla. Pour caractériser les unités stratigraphiques, on a utilisé la sédimentologie, les analyses chimiques et des datations radiocarbone (Tab. 1). Neuf datations ont été effectuées sur des charbons de bois, des ossements et des coquilles d’hélix. Ces dernières ont été évitées au maximum, étant donné leur manque de fiabilité. Quatre datations radiocarbone ont été effectuées au laboratoire AMS 14C-Labor Erlangen, Allemagne (Erl-7905, Erl-7906, Erl-7907 and Erl-7908) et cinq autres datations classiques à l’Ecole Nationale d’Ingénieurs de Sfax (ENIS), Tunisie (Tab. 1). Les différentes dates obtenues ont été converties en âge calendaire calibré (cal. BP) en utilisant le programme Calib Rev 5.0.1 (Stuiver et Reimer, 2005). Ces datations ont été complétées par l’identification de matériaux archéologiques, très abondants en Tunisie centrale et très bien calés dans le temps. La méthode de O.P. Mehra et M.L. Jackson (1960) a permis d’extraire le fer libre et le fer total des formations étudiées. Ces analyses nous renseignent sur le degré d’altération et de pédogénéisation de ces formations et par conséquent sur leurs âges relatifs. On a aussi utilisé la perte au feu pour déterminer la proportion de la matière organique du CaCO3 et de l’eau dans les mêmes formations étudiées.

Les cinq coupes étudiées présentent des dépôts alluviaux holocènes accumulés sous des conditions environnementales différentes. La corrélation chronostratigraphique entre les différentes coupes a permis d’établir une coupe synthétique des archives alluviales holocènes (fig. 10 and fig. 11). Cette coupe synthétique a été comparée à des travaux détaillant la morphogenèse holocène au nord (Bourgou, 1993 ; Faust et al., 2004 ; Zielhofer et al., 2004) et au sud de la Tunisie (Ballais, 1991). Nous identifions quatre phases d’accumulation holocènes. Chaque phase d’accumulation correspond à un cycle de sédimentation commençant par une dynamique active de remblaiements suivie par une atténuation de la morphogenèse, donnant des dépôts plus fins, souvent pédogénéisés et couronnés par une formation de sol. Ainsi, chacune de ces phases d’accumulation est constituée de deux stades, une phase rhexistasique à la base et une phase biostasique au sommet (fig. 10 and fig. 11). Chaque phase d’accumulation n’est pas forcément suivie d’une phase d’incision, ce qui explique l’existence de quatre terrasses holocènes seulement, dont trois sont historiques. La première phase d’accumulation est d’âge Holocène inférieur (avant 7ka cal. BP). Dans sa partie sommitale se développe un paléosol daté entre 7,5 ka cal. BP (fig. 9) et 7,2 ka cal. BP (fig. 8). Les dépôts de la première phase d’accumulation surmontent une formation riche en CaCO3, datés au sommet à 11,7 ka cal. BP (fig. 10 and fig. 11). La deuxième phase d’accumulation remonte à l’Holocène moyen (6,5-4 ka cal. BP). Les dépôts de cette phase d’accumulation s’affinent progressivement de la base au sommet, où un paléosol bien développé (phase d’accumulation 2-b ; fig. 10 and fig. 11) a été daté au radiocarbone à 4,2 ka cal. BP. Les deux phases d’accumulation précédentes sont généralement superposées et constituent la terrasse holocène antéhistorique (Zerai, 2006). L’absence d’une phase d’incision importante à la fin de la de cette première phase d’accumulation explique l’existence d’une seule terrasse holocène antéhistorique. La troisième phase d’accumulation est fini-romaine à post-romaine (1,6-1,1 ka cal. BP). Elle a été précédée par une grande phase d’incision, responsable de l’érosion des dépôts des deux phases précédentes dans plusieurs coupes (SEF, SBV ; fig. 5 et fig. 8). La partie biostasique (b) de cette phase d’accumulation est caractérisée par des paléosols très fins, riches en matière organique, qui renferment des traces de feu et datés de 1,1 à 1,2 ka cal. BP (fig. 10 and fig. 11). La troisième phase d’accumulation correspond à la terrasse historique principale datée de la période islamique. Les remblaiements de cette phase d’accumulation ont commencé peu après le Xè siècle ap. J.-C. et ont continué probablement jusqu’à la transition Moyen Âge-Epoque moderne (vers 0,5 ka cal. BP). Pour cette période, l’ébauche d’un sol très peu développé peut être observée dans les coupes SEF4d et SBV4e (fig. 5 et fig. 8). La quatrième phase d’accumulation holocène correspond à la terrasse historique intermédiaire. Une cinquième phase d’accumulation holocène a probablement commencé au début de l’époque moderne et a continué jusqu’à l’actuel (après 0,5 ka cal. BP ; fig. 10 and fig. 11). Les dépôts relatifs à cette période sont rares sur le terrain. La cinquième phase d’accumulation peut être rattachée à la construction actuelle du lit majeur des cours d’eau.

Les différentes phases d’accumulation et d’incision présentées, corrélées entre elles (fig. 10 and fig. 11) sont en phase avec des alternances de phases humides et arides qui se succèdent tout au long de l’Holocène : une phase d’assèchement à la fin du Tardiglaciaire ; une activité torrentielle des cours d’eau à la transition Tardiglaciaire-Holocène ; une atténuation de la morphogenèse et croissance de l’humidité au cours de l’Holocène inférieur ; une atténuation de la morphogenèse et formation de sol à l’Holocène moyen ; un retour à l’aridité au cours de l’Holocène supérieur antéhistorique ; un climat relativement humide au début de l’Antiquité ; une crise morphogénique fini à post-romaine (1,6-1,2 ka cal. BP) ; une courte pulsation humide et formation de sol vers 1,1 ka cal. BP ; une dégradation progressive des environnements entre le Xè et le XIXè siècle ; l’extension des terres agricoles et l’incision des cours d’eau depuis le début de la période coloniale.

En conclusion, la complexité des enregistrements holocènes dans le bassin de l’oued Sbeïtla s’explique par la variété des modelés, la multiplicité des phases d’accumulation et d’incision et la succession de plusieurs périodes humides et sèches (fig. 10 and fig. 11). Les différentes phases d’accumulation holocènes n’ont pas été forcément suivies par des phases d’incision, ce qui explique que ne soient observées qu’une seule terrasse holocène antéhistorique et trois terrasses historiques (fig. 10 and fig. 11). Le nombre de terrasses holocènes relativement élevé par rapport à la Tunisie méridionale et septentrionale constitue une spécificité de la Tunisie centrale. Selon J.-L. Ballais (1993), le nombre de terrasses holocènes en Tunisie varie selon les étages bioclimatiques ; une seule terrasse dans l’étage saharien, deux dans l’étage aride, trois dans l’étage semi-aride et une dans l’étage subhumide. La multiplicité des phases d’accumulation et d’incision holocènes indiquent des fluctuations climatiques majeures engendrant des phases de stabilité et d’activité morphogéniques (fig. 10 and fig. 11). Les périodes de stabilité résultent de conditions humides et d’un couvert végétal dense permettant la formation d’un sol. Les périodes d’activité correspondent à des conditions arides favorisant la déflation éolienne. La morphogenèse holocène antéhistorique a été largement déterminée par le climat. Au cours des temps historiques, le facteur anthropique a été beaucoup plus présent, mais il est resté néanmoins subordonné au facteur climatique. Ce n’est que récemment (époques moderne et contemporaine) que le facteur anthropique devient morphogénétiquement déterminant.

Haut de page

Table des illustrations

Titre Fig. 1 – Location of the Wadi Sbeïtla basin and sites of key profiles.Fig. 1 – Localisation du bassin-versant de l’oued Sbeïtla et des coupes étudiées.
Légende 1: profile; 2: radiocarbon dating; 3: ceramics; 4: artifacts; 5: prehistoric site (Capsian); 6: key profile.1 : coupe ; 2 : datations par le radiocarbone ; 3 : céramiques ; 4 : artéfacts ; 5 : site préhistorique (Capsien) ; 6 : code de la coupe.
URL http://geomorphologie.revues.org/docannexe/image/7737/img-1.jpg
Fichier image/jpeg, 616k
Titre Fig. 2 – Geographical settings of Wadi Sbeitla basin.Fig. 2 – Cadre géographique du bassin de l’oued Sbeïtla.
Légende A: Slope classes. B: Lithology. 1: hard rock; 2: mixed rock; 3: soft rock; 4: calcareous crust; 5: Quaternary deposit. C: Drainage pattern. D: Vegetal cover.A : Carte des pentes. B : Lithologie. 1 : roches résistantes ; 2 : roches moyennement résistantes ; 3 : roches meubles ; 4 : croûte calcaire ; 5 : dépôts quaternaires. C : Réseau hydrographique. D : Couverture végétale.
URL http://geomorphologie.revues.org/docannexe/image/7737/img-2.jpg
Fichier image/jpeg, 520k
Titre Tab. 1 – Materials used for dating.Tab. 1 – Matériaux utilisés pour les datations.
URL http://geomorphologie.revues.org/docannexe/image/7737/img-3.jpg
Fichier image/jpeg, 172k
Titre Fig. 3 – Structural framework of SEF and ACH profiles.Fig. 3 – Contexte structural des deux coupes SEF et ACH.
Légende 1: clay (Santonian); 2: limestone (Campanian-Maastrichtian); 3: silt (Aquitanian); 4: sandstone (Upper Langhian); 5: hogback; 6: dip; 7: slope deposit; 8: Pleistocene deposit; 9: Holocene deposit; 10: Wadi; 11: riverbank; 12: contour line; 13: Roman aqueduct; 14: road; 15: stratigraphic section; 16: Capsian site.1 : argiles (Santonien) ; 2 : calcaires (Campanien-Maastrichtien) ; 3 : silts (Aquitanien) ; 4 : grès (Langhien supérieur) ; 5 : crêt ; 6 : pendage ; 7 : dépôts de pente ; 8 : dépôts pléistocènes ; 9 : dépôts holocènes ; 10 : oued ; 11 : berge ; 12 : courbe de niveau ; 13 : aqueduc romain ; 14 : route ; 15 : coupe stratigraphique ; 16 : site capsien.
URL http://geomorphologie.revues.org/docannexe/image/7737/img-4.jpg
Fichier image/jpeg, 820k
Titre Fig. 4 – Captions of fig. 5 to fig. 9.Fig. 4 – Légende de la fig. 5 à la fig. 9.
Légende 1: Miocene sandstone; 2: clayey to silty layer; 3: loam; 4: sand; 5: gravel layer; 6: humic enrichment; 7: hydromorphic features; 8: calcareous concretions; 9: prismatic structure; 10: fire-site; 11: sample; 12: sedimentary unit; 13: 14C-analysis (calibrated ka BP); 14: ceramics (calendar ka); 15: artifacts (calendar ka); 16: shell (Helix). 1 : grès miocène ; 2 : argiles et limons fins ; 3 : limons ; 4 : sables ; 5 : graviers ; 6 : sédiments organiques ; 7 : traces d’hydromorphie ; 8 : concrétions calcaires ; 9 : structure prismatique ; 10 : foyer ; 11 : échantillon ; 12 : unité chronostratigraphique ; 13 : datation 14C (ka calibrés BP) ; 14 : céramiques (ka calendaires) ; 15 : artéfacts (ka calendaires) ; 16 : coquille d’Helix.
URL http://geomorphologie.revues.org/docannexe/image/7737/img-5.jpg
Fichier image/jpeg, 60k
Titre Fig. 5 – Safhit El-Flalig (SEF). See the captions in fig. 4.Fig. 5 – La coupe de Safhit El-Flalig (SEF). Voir la légende sur fig. 4.
URL http://geomorphologie.revues.org/docannexe/image/7737/img-6.jpg
Fichier image/jpeg, 492k
Titre Fig. 6 – Wadi Achihb profile (ACH).Fig. 6 – La coupe de l’oued Achihb (ACH).
Légende A: Capsian artifacts (ACH3). See the captions in fig. 4.A : Artéfacts capsiens (ACH3). Voir la légende sur fig. 4.
URL http://geomorphologie.revues.org/docannexe/image/7737/img-7.jpg
Fichier image/jpeg, 452k
Titre Fig. 7 – Historical deposits blocked by a Roman aqueduct. Fig. 7 – Dépôts historiques bloqués par un aqueduc romain.
URL http://geomorphologie.revues.org/docannexe/image/7737/img-8.jpg
Fichier image/jpeg, 340k
Titre Fig. 8 – Wadi El-Hammar profile (HAM).Fig. 8 – La coupe de l’oued El-Hammar (HAM).
Légende B: Specific holes on Helix aperta. C: Capsian artifacts (HAM2a). D: Amorphous silica of geometric shape. E: Dissolution forms. F: deposits of carbonate. See the captions in fig. 4.B : Trous spécifiques sur des Helix aperta. C : Artéfacts capsiens (HAM2a). D : Silice amorphe de forme géométrique. E : Formes de dissolution. F : Dépôts de carbonates. Voir la légende sur fig. 4.
URL http://geomorphologie.revues.org/docannexe/image/7737/img-9.jpg
Fichier image/jpeg, 584k
Titre Fig. 9 – Wadi El-Hammar profile (HAM).Fig. 9 – La coupe de l’oued El-Hammar (HAM).
Légende G: Quartz grain covered with amorphous silica. H: Amorphous silica. See the captions in fig. 4.G : Grains de quartz recouverts de silice amorphe. H : Silice amorphe. Voir la légende sur la fig. 4.
URL http://geomorphologie.revues.org/docannexe/image/7737/img-10.jpg
Fichier image/jpeg, 484k
Titre Fig. 10 – SBV profile.Fig. 10 – La coupe SBV.
Légende I: Large bone of 9 cm of diametre. J: Intercalation of small sandy-clay layers at the base of the hole. See also the captions in fig. 4.I : Grand os de 9 cm de diamètre. J : Intercalation de fines couches sablo-argileuses à la base du trou. Voir la légende sur fig. 4.
URL http://geomorphologie.revues.org/docannexe/image/7737/img-11.jpg
Fichier image/jpeg, 408k
Titre Fig. 11 – Synthetic stratigraphic section of Holocene alluvial deposits.Fig. 11 – Coupe stratigraphique synthétique des dépôts alluviaux holocènes.
Légende 1: clayey to silty layer; 2: loam; 3: sand; 4: gravel layer; 5: fire site; 6: humic enrichment; 7: hydromorphic features; 8: calcareous concretions; 9: calcareous crust; 10: 14C-analysis (calibrated ka BP); 11: ceramics (calendar ka); 12: artifact (calendar ka); 13: flooding/aggradation; 14: river incision; 15: stable fluvial dynamics; 16: active fluvial dynamics; 17: approximate transition.1 : argiles et limons fins ; 2 : limons ; 3 : sables ; 4 : graviers ; 5 : traces de feu ; 6 : sédiments organiques ; 7 : indices d’hydromorphie ; 8 : concrétions calcaires ; 9 : croûte calcaire ; 10 : datation 14C (ka calibrées BP) ; 11 : céramiques (ka calendaires) ; 12 : artéfacts (ka calendaires) ; 13 : remblaiement ; 14 : incision ; 15 : dynamique fluviale stable ; 16 : dynamique fluviale active ; 17 : limite imprécise.
URL http://geomorphologie.revues.org/docannexe/image/7737/img-12.jpg
Fichier image/jpeg, 384k
Haut de page

Pour citer cet article

Référence papier

Kamel Zerai, « Chronostratigraphy of Holocene alluvial archives in the Wadi Sbeïtla basin (central Tunisia) », Géomorphologie : relief, processus, environnement, vol. 15 - n° 4 | 2009, 271-286.

Référence électronique

Kamel Zerai, « Chronostratigraphy of Holocene alluvial archives in the Wadi Sbeïtla basin (central Tunisia) », Géomorphologie : relief, processus, environnement [En ligne], vol. 15 - n° 4 | 2009, mis en ligne le 01 janvier 2012, consulté le 27 avril 2017. URL : http://geomorphologie.revues.org/7737 ; DOI : 10.4000/geomorphologie.7737

Haut de page

Auteur

Kamel Zerai

University of Monastir, Higher Institute of Applied Studies in Humanities of Mahdia, Route Rajich, B.P. 22, 5112 Mahdia (Kamel.Zerai@cnrs-bellevue.fr)

Haut de page

Droits d’auteur

© Groupe français de géomorphologie

Haut de page
  • Logo CNRS - Institut des sciences humaines et sociales
  • Logo Groupe français de géomorphologie
  • Revues.org