Sediment accumulation rates as a result of environmental changes and seismic mechanisms for
turbidite triggering in the eastern Algerian margin
Pierre Giresse*, Maria-Angela Bassetti, Henri Pauc, Virginie Gaullier
Laboratoire IMAGES-EA 4218, Université de Perpignan Via Domitia, 52, Avenue Paul Alduy, 66860
Perpignan, France
Keywords: Algerian margin, Sediment accumulation rates, Turbidites, Sedimentary instability, Seismicity
Abstract.
From the analysis of seven sediment piston-cores sampled in 2005 (MARADJA-2 French-Algerian
cruise), this study aims to identify for the first time possible sedimentary instabilities controlled by
seismicity off or close to the eastern coast of Algeria. The detailed lithologic study allows us to identify
the frequency of the gravity events (turbidites, debrites, and slumping) and to discuss their geographical
origins [sources?] and their causes [triggering mechanisms?]. On the basis of a chronostratigraphy of 24
14
C AMS datings, a careful study of sediment accumulation rates at the foot of the margin (zones
extending off Bejaia and Annaba) and, in particular, a semi-quantitative observation of the biomarkers
indicating the origin of gravity flows were carried out. The statistical study of the thickness of the turbiditic
sequences compared to that of the Ta Bouma division and to basal grain-size suggests marked erosion
of the tops of the turbidite sequence. Such truncation appears more intense at the foot of the slope than
within a higher water depth. Two sediment cores located upslope, here considered as reference cores,
enable the estimation of the palaeoenvironmental parameters that controlled the sedimentation: (1) in
the prodelta of Soummam Oued, after 2285 BP, floods were less frequent and the sediment
accumulation rate decreased because of a drier climate of the Maghreb; (2) at the middle slope to the
NE of Annaba, a place shielded from the gravity mechanisms, the acceleration of the rate coincided with
warmer waters leading to maxima of carbonate fluxes (particularly Pteropods). Off Bejaia, two sediment
cores show a spectacular increase of the sediment accumulation rate between 2300 and 1145 BP while
the period of frequency of turbidites is shortened. An episode of strong instability of the slope connected
to a seismic activity is suggested. Both sediment cores of the slope of Annaba indicate an increase in
gravity flows during the high-stand interval, which is tentatively related to a general increase of seismicity
during the last thousand years. This area distribution of gravity events is clearly different from that of the
western margin where the sedimentation is slightly less perturbed by gravity events.
1. Introduction
In 2005, the MARADJA 2 oceanographic cruise has identified for the first time the main structural,
morphological and sedimentological characteristics of the slope and the deep basin of the easternmost
Algerian continental margin. Before 2005, almost no data were available for this offshore part of Algeria.
Untill now, only the morphological and structural witnesses of active deformation at the sea floor have
been explored in the area off Annaba and Skikda (Kherroubi et al., 2009). The objective of this study is
to define for the first time the sediment deposition pattern and sedimentary processes taking place along
the margin from Great Kabylia to the border between Algeria and Tunisia (Figure 1). This work is a
contribution to the more general goal of the MARADJA Project consists into trying to better assess the
role of active tectonics on depositional patterns in the Algerian Margin and deep basin. This paper
2
focuses on two portions of this eastern margin that were sampled by gravity cores : 1) the Bejaia zone
between 5° and 5°22’E and 2) the Annaba zone, between 7°40’ and 8°E. Recurrence, frequency
distribution of bed thickness, and composition of turbidite sequences are considered according to the
distance from the slope (foot of the slope and deep basin) and according to the chronostratigraphy
(lowstand system tract, transgressive system tract, highstand system tract). Various indicators of the
erosion of the last lowstand deposits lying on the edge of the shelf are identified and semi-quantified and
should improve the understanding of the origins of the different sediment sources.
A sediment core sheltered from gravity mechanisms (KMDJ38) supplies a record of about 50,000 years,
which will constitute a regional environmental reference allowing the estimation of the other variables of
the deep sedimentation such as the changes in terrigenous fluxes or in primary oceanic production. In
the same concern, a sediment core in the Soummam Oued prodelta will propose a high-resolution
record of the floods during the last three thousand years.
According to a well-established concept in the literature and in particular the extension in classical
sequence-stratigraphic models to deep-water clastic systems, all important turbidite successions are
assumed to be deposited during or shortly after an episode of lowstand (Mitchum, 1985; Mutti, 1985;
Den Hartog Jager, 1993; Pickering et al., 1995). However, this concept has been questioned because
active gravity-induced deposition coincided with high-stand conditions (e.g., Piper and Savoye, 1993;
Dennielou et al., 2003; Carvajal and Steel, 2006; Giresse et al., 2009). The seismic event is a possible
triggering mechanism during the highstand on an active margin such as the eastern part of Algeria. The
aim of this paper is to discuss the eustatic, seismogenic or random causes at the origin of sediment
gravity-flow processes on this eastern Algerian margin and to compare them with mechanisms described
in the western part of the same margin.
2. Geological, seismological and environmental backgrounds
Northern Algeria is mainly an orogen, namely the Maghrebian belt (Durand-Delga and Fonboté, 1980;
Wildi, 1983) which is composed from south to north of (1) the External Domain or Tellian units with
sedimentary units (marls and limestones); (2) the flysch nappes which thrust the External Domain and
are composed of former sediment of the Tethys ocean; and (3) the Internal Domain, composed of a
Hercynian basement, sometimes associated with its sedimentary cover (“Dorsale Kabyle“unit) (Fig. 1).
This study of the eastern margin of Algeria concerns respectively the offshore areas close to Bejaia (Fig.
2) and close to Annaba (Fig. 3). Bejaia is located between the Great Kabylia reliefs and those of Lesser
Kabylia; its hinterland belongs to the External Domain, which is bordered to the west by wide outcrops of
flysch. Annaba is situated in a Neogene-Quaternary depression dominated on the west by reliefs of the
Hercynian basement of the Kabyle Dorsal.
Seismic activity of the Algerian margin is linked to convergence between the African and European
plates. Crustal shortening is  5 mm yr-1 in a N60° W direction at the longitude of Algiers (Nocquet and
Calais, 2004). Previous MARADJA studies showed two different deep-seated tectonic styles from east to
west (Déverchère et al., 2005; Domzig et al., 2006): the eastern and central margin displays a purely
compressive finite strain and seismic activity (the last one was responsible for the 2003 Mw 6.8
3
Boumerdès earthquake), whereas the westernmost margin (from Tenes to Oran) displays a pattern of
both strike-slip and reverse faulting which is apparently seismically less active in the instrumental time
span (Buforn et al., 2004). From the slow strain rates and the distributed deformation over the Tell-Atlas
belt, it can be argued that the seismic cycle on a given active fault is probably long (i.e. ranging between
350 and more than 1500 years), as shown by palaeoseismological studies led in western Algeria
(Meghraoui et al., 1988). Recently, a reappraisal of earthquake catalogs in northern Algeria (Harbi et al.,
2010; Hamdache et al., 2010) has been performed: although it is difficult to demonstrate significant time
changes in seismic energy release since A.D. 856 (i.e. the oldest reported historical event), it appears
that within this time interval of ca. 1000 years, the eastern Algerian margin and coastal region (from
Bejaia to Annaba, longitudes 005°-008°E) have undergone a limited number of M > 6 events (only 2
events located offshore, see Figure 2 in Hamdache et al., 2010), which is significantly less than the
central margin (16 events located both onshore and offshore, longitudes 001°-005°E). Among them, the
largest event reported is the 1856 August 21-22 coastal earthquake sequence of Djidjelli
(Ambraseys,1982; Yelles et al., 2009), which magnitude has been recently revised at about 6.6
(Maouche et al., 2011).
Submarine landslides have been systematically reported along the western and central Algerian margin
(Domzig et al., 2009; Dan et al., in press). They are expressed by seafloor scars (usually of small size)
and acoustically chaotic/transparent units which may be subsurface or buried and are often interpreted
as mass transport deposits over the whole Algerian margin (Cattaneo et al., 2010). Although plausible,
the correlations between mapped seafloor instabilities and seismicity is not straightforward (e.g.
Camerlenghi et al., 2010), mainly because attributing an identified landslide to a given earthquake
requires a careful and local-scale multiproxy analysis (e.g. Nouguès et al., 2010).
Two main rivers are to be considered in the eastern part of Northern Algeria: the Soummam Oued and
the Seybouse Oued (Figs. 2 and 3). To the west, the Soummam Oued (Fig. 3) pours into a rather narrow
corridor below the reliefs of Great Kabylia and the Tellian thrust folds. The lithology of these reliefs is
made of fine-grain flyschs, or sandy marls (Domzig et al., 2009), thus the main part of the suspended
matter is composed of fine or very fine particles. The average solid discharge is 4.106 tons/year
(Leclaire, 1972) and occasionally several million tons are poured in a few months of the winter season
when exceptional suspended matter concentration (> 200 g.l-1) is recorded (Pauc et al., 1997). In the
East, the Seybouse Oued runs through a wide Neogene-Quaternary depression and drains a quite small
surface of Tellian flysch ending in a solid discharge of the order of 1.2. 106 tons/year (Leclaire, 1972). Its
mouth is located east of Annaba (Fig. 2).
The continental shelf of the eastern part of Algeria, one of the narrowest of the Maghreb, ranges in width
from 3 to 6 km (10 km in places). Its moderate slope is on average 1° and similar to that of the nearby
regions; the whole shelf and slope measuring generally less than 20 km wide. The shelf break is
indented where submarine canyon heads cut into the shelf. The shelf off Bejaia is cut between 40 and
100 m water depth by a small and specific canyon, which credibly testifies to a low-stand incision of
Soummam Oued (Leclaire, 1972). The cover of this continental shelf, frequently interrupted and not very
thick, is also highly varied.
4
Off Annaba, rocky outcrops (in particular of volcanic rocks) are frequent and surrounded by very coarse
clastic deposits; the largest part of sands and gravels consists of bioclasts: calcareous algal, bryozoans,
molluscs, sea urchins, or large benthic foraminifers (Ammonia beccarii and Elphidium crispum),
sometimes strongly oxidized. Gravels or siliceous sands are rather rare. Seybouse Oued flood deposits
accumulate in a fan, which does not exceed the middle part of the shelf (Fig.4a).
The Bejaia shelf presents a contrast between its western part, which is as varied as that of Annaba, and
its eastern part, which is dominated by silto-sandy deposits of the floods of Soummam Oued (Fig. 4b).
Further seaward, low-stand sands are both siliceous and calcareous.
The outer edges of the shelf (especially off Annaba) show relict sedimentary cover at the surface, with
warm fauna of the beginning of the sea-level rise. Muds with temperate, then subarctic microfauna are
associated. Subarctic fauna supplied a 14C age of 30,000 BP, linked to MIS 3 (Leclaire, 1972).
3. Sampling and methods
The primary aim of Maradja (“MARge Active DJAzaïr”) 2 is to extend eastward (from Dellys, 80 km of
Algiers and 4° E) to Annaba Bay (8°20’E) the identification of the active faults and the sedimentary
instabilities already studied in the western part of the margin of Algeria (Kherroubi et al., 2009). During
the cruise, IFREMER R/V Le Suroît obtained continuous seafloor multibeam imagery, very high to highresolution profiling, and sediment samples using a Küllenberg piston-type corer. In this paper, we
provide a detailed sedimentological overview of the seven cores recovered during this cruise. The study
of the seismic records (Chirp in particular) based on the same data set are not discussed extensively
here but have been already used for determining fold and fault Quaternary deformation offshore
(Kherroubi et al., 2009) and will be further discussed in another paper (Gaullier et al., in preparation).
Based on the lithological changes observed, samples were taken at an average 1-5 cm, but this interval
was reduced to 0.5 or 1 cm according to colour or texture changes of the divisions of each turbidite
sequence. Samples were then analysed by standard methods including measurements of water content,
preliminary grain-size analysis by wet sieving through a 315 m and 40 m mesh, calcimetry and
microscopic study of microfaunal and mineralogical sand contents. The abundance of various tracers of
coastal lowstand deposit originating from the outer edge or the shelf break (relict coastal foraminifers as
Elphidium crispum, Ammonia beccarii, and Quinqueloculina seminulum, vegetal debris from lagoon
deposits, evolved glauconitic grains or oxidised debris) was recognised. This was followed by the
counting of these particles. The mineral composition of some intervals was studied by means of X-ray
diffraction using Co K1 radiation.
Following the approach from Talling (2001), the frequency distribution of the entire turbidite sequences
and of sandy-bed thicknesses, and the relationship between grain-size and bed thickness were used in
order to interpret the rheological changes of settling, and to distinguish preserved depositional
sequences from truncated sequences. Stable isotope analyses were carried out at the Isotope
Laboratory of the Geoscience Department of the University of Bremen. Chronostratigraphy is based on
24 AMS 14C dates measured in biogenic carbonates from planktic foraminifers collected in the
hemipelagic intervals. Measurements were carried out at Poznan Radiocarbon Laboratory. Calibration
5
into calendar scales was calculated using the northern hemisphere calibration curve (Software Calib.
5.02) (Stuiver and Reimer, 1993). Calendar ages are given with 1 standard deviation (Table 1). Given
that several ages are beyond 30,000 BP and cannot thus be the object of corrections, by concern of
homogeneity of the presentation, the conventional 14C ages will be indicated through the text.
4. Sediment facies and stratigraphy of the sediment cores
4.1.Skikda-Annaba Area
4.1.1. KMDJ38
This core is chronostratigraphically well constrained and will serve as a palaeoenvironnemental
reference for the other records of this eastern margin. It is an 8.07 m-long sediment core retrieved from
the slope east of Annaba and off the mouth of Seybouse Oued. The site is halfway up (1057 m water
depth) a slope, which is distal to the canyon heads and acts as a very wide interfluve. The highresolution seismic profile (chirp) indicates a homogeneous or very weakly layered cover of about ten
meters in thickness which overlays a rough erosional surface.
The sediment is a rather homogeneous and structureless grey-beige mud with abundant black spots. A
higher density of these spots or a darker organic pigmentation is observed at different levels (765 cm,
690 cm, 520-530 cm, 410-420 cm, 234 cm and 78 cm) suggesting marked organic terrestrial inputs. The
textural analysis presented in Fig.5 is not a representative siliciclastic contribution because only a few
grains of quartz from 300 to 400 µm-large are observed, in particular in the upper two meters. The low
concentrations in fine sand, which can reach or exceed 5%, up to 10%, correspond mainly to
accumulations of tests of pelagic microfauna, essentially pteropod debris. Together to bioclasts,
abundant plates, granules or stalks strongly epigenized by the pyrite are found in the deepest 6 meters
of the core; these diagenetic product particles participate strikingly in the weight definition of the sandy
fraction.
A 14C date at the bottom of the sediment core indicates an age beyond 50,000 years. The top of the
KMDJ38 is about 880 years old but the fact that this core has lost about 10 cm of the upper deposit
during the coring process needs to be taken into account. Three other 14C dates allow the estimation of
the sediment accumulation rate of the last six meters, which appears to be the slowest of all the
sedimentation rates calculated during studies of cores retrieved in this Algerian margin (cruises
MARADJA 1 and 2 included): it is of the order of 15 cm / 103 years. Such slow and regular accumulation
constitutes an excellent record for high-resolution oxygen isotope measurements as illustrated in Figs. 5
and 6. The glacial-interglacial 18O amplitude in this record is 4‰, largely greater than the 1.3‰
(Fairbanks, 1989) and 1‰ (Schrag et al., 1996) planetary glacial-interglacial ice volume signal. This
suggests that at least 3‰ of this long-term signal are attributed to temperature (or salinity) change. This
oxygen isotope record closely matches millennial-scale Dansgaard-Oeschger (DO) events 1-14 and
Heinrich events H1-H5 as the position of the Younger Dryas global deglacial event and of the short 8.2
ka cold event. The 8.2 ka cold event was previously evidenced in the eastern and western
Mediterranean where boreal summer insolation gradually declined (von Grafenstein et al., 1998;
6
Weninger et al., 2006; Davis and Stevenson, 2007). However, some oscillations of rather low
amplitude and short duration of the MIS 3 do not allow a very precise correlation with the indications of
temperature suggested by the isotopic curve (the positions of DO 3-4 and of H5 remain approximate).
However, the recognition of time positions of most interstadials allows us for a detailed estimate of the
successive sediment accumulation rates along this core (Figs. 6 and 7).
The sediment accumulation rates between 50,000 and 42,000 BP (DO 11-14) are moderate (12.5-18 cm
/ 103 years): they rise (35-55 cm / 103 years) only slightly later while the oceanic temperatures began a
trend to cooling. These rates decrease appreciably between DO 9 and 6 (12.5-18.3 cm / 103 years)
before reaching a maximum (55 cm / 103 years) just after DO 6. The interval between DO 5 and 1
corresponds to the irregular cooling which characterizes MIS 2 where the accumulation rates are there
relatively low (10-20 cm / 103 years). After Younger Dryas, the warming coincides with an acceleration of
the sedimentation rate (27.5-40 cm / 103 years). Finally, the first rather warm part of the wet Holocene
shows a 20 cm / 103 years rate that falls to 8.3 cm / 103 years during the second part. The general curve
of the sediment accumulation rate passes through several plateaus and several accelerations whose
amplitude can reach a factor of 5 in spite of the obvious absence of gravity mechanism transport. With
few exceptions, varied periods of latency (especially during the MIS 3), warmer waters appear more
favourable to a faster accumulation.
The carbonate contents vary between 20 and 50%. Carbonate-rich units reflect interglacial conditions as
in the Madeira abyssal plain (Weaver et al., 1986) and generally correlate with high sand content. We
can also observe minima of productivity with 20% of carbonate connected to possibly colder waters:
around 40-45,000 BP, around 35,000 BP, between 20,000 and 13,000 BP as well as during the second
half of the Holocene (Fig. 7). These minima correspond to the plateaus of the curve of accumulation
rate. A semi-quantitative estimation of the fragments of pteropods leads to the observation of a rather
regular correlation with the carbonate contents (Fig. 5), it is thus suggested that the accelerations of the
general sedimentation are mainly controlled by a more important addition of carbonate fragments
supplied by the oceanic production (in particular, by the tests of pteropods) when the temperature rises.
The scarcity of downslope processes can be evidenced through examination of the sandy fractions
under binocular microscope (Fig. 5). The tests of coastal Miliolidae such as Quinqueloculina seminulum,
altered or partially dissolved during emersion intervals, are found only in only few deposits [places?]
(around 7.70 m, between 7 m and 6.30 m, around 6 m, between 5.20 and 5 m and between 4.30 and 4
m). Among the benthic foraminifers, the vertical distribution of Hyalinea balthica depicts a similar trend.
This species is very abundant at high latitudes in cold water (9-13°C, Murray, 1991). At temperate
latitudes, it might be considered as an indicator of changing bottom temperatures. Actually, it was found
to be rather abundant in shallow water during cold periods, adapting to life in deeper settings during
warmer intervals. The input from relic deposits of the edge of the shelf is thus very limited, since littoral
species, including usual markers (i.e. A. beccarii and E. crispum coastal foraminifers) are found
extremely seldom.
In the aim to better evaluate the sediment input from the shelf edge, a ratio between allochthonous and
autochthonous species of ostracods has been calculated testifying that the occurrence of gravity
7
mechanisms seems globally or more repetitive during intervals of lowstand, especially around 20,000
and 40,000 BP. The ostracods are thought to be useful markers because the valves are lighter than
tests of big benthic foraminifers, and more easily reworked in suspension and transported. Moreover, the
presence or absence of tests of juveniles allow a fairly precise qualitative estimation of the distance in
sediment transport and discriminate the type of depositional processes. Within the reworked fauna, we
have included the littoral and circa-littoral species having a bathymetric distribution not exceeding 200 m
water depth. They include genera such as Pterigocythereis sp., Pontocythere sp., Carinocythereis sp.,
Callistocythere sp., Hemicytherura sp., Neocytherideis sp., among others. Carapaces [shells?] are fairly
well-preserved but the fragile valves of juveniles are almost absent in the shelf ostracod assemblages.
Therefore, we put forward the hypothesis of an off-shelf sediment transport possibly due to density
cascading in which individual grains are instead reworked by bottom currents rather than mobilized by
gravity processes such as turbidity currents or flows (Shanmugan, 2000). The curve of the
thanatocoenosis components anti-correlates the 18O plot and the general trend illustrates several
significant peaks of allochthonous valves (>30%) in correspondence with the H5, before and during the
H4 and during the H1. Minima are recorded during the DO interstadials with peaks ranging from 5 to
20% during the DO stadials. Almost no reworked fauna is found in the uppermost 50 cm corresponding
to the mid/late Holocene, clearly characterized by little detrital input from the shelf.
The abundance of the clusters of organic matter, indicated above, and the presence of much degraded
vegetal fibres can also indicate, in certain intervals, maxima of terrestrial flows. However the part of the
accumulation connected to gravity flows must be considered as unimportant in terms of the sedimentary
balance of KMDJ38. This core shows a history of deposition unaffected by turbidity currents: its
deposition is essentially hemipelagic and constitutes a good basis for comparison to other sectors of this
margin where various gravity mechanisms overlap.
4.1.2. KMDJ39
This site is the deepest (at 2769 m water depth) and the furthest from the coast. It is situated below
several important submarine valleys perpendicular to the coastline (Kherroubi et al., 2009) and
susceptible to drain the inputs of the Seybouse Oued (Fig. 2). The echo-character of this site is typically
characterized by parallel to sub-parallel sub-bottom reflections.
The sediment core consists of a grey-brown mud, inside which we find over the entire section coarsegrained layers of some centimetres in thickness (in particular around 7.90 m, 5.0 m, 3.90 m and 0.60 m)
(Fig. 8) [je ne suis pas sûr d’avoir bien compris la phrase telle qu’elle est tournée?]. The presence of
sharp erosional contact, normal size grading, and gradational upper contact is a reliable criterion to
interpret these layers as turbidites. Many but not all turbidite beds consist of a couple of two distinctly
coloured units: a lower deep grey brown unit rich in organic matter and an upper grey unit becoming
paler further upward. The sandy division (Bouma Ta) of turbidites does not exceed 10 cm in thickness.
The first deposits, on the base of sequences, almost lack medium and coarse sands (less than 5%); on
the other hand, the contents in fine sand can reach or exceed 50%. The first centimetres of the Ta
division are often indicated by the abundance of degraded plant fragments and shelly coastal debris
8
dropped from the edge of the shelf. A double peak of carbonate and sand then indicates the beginning
of the sequence.
The sedimentary history begins at 24,050 BP and suggests a lowstand phase, that of the active
transgression and that of highstand. On the basis of three 14C dates (Table 1), we calculate a moderate
sediment accumulation rate of 29 cm / 103 years between 24,050 and 8800 BP (low-stand and
transgressive phases). This rate increases to 35 cm / 103 years between 8800 and 4635 BP, and then
rises to 66 cm / 103 years after 4635 BP where one of the most consequent gravity deposit is identified.
The global average rate of sediment accumulation is 37 cm / 103 years.
Shallow-water benthic foraminifera, A. beccarii and E. crispum are found only in two turbidites of the
lowstand interval when the erosion of the outer edge of the shelf is facilitated. On the other hand, Q.
seminulum is abundant in most of the deposits of this lowstand phase. The more or less altered rocky
fragments are especially observed in these same deposits between 6 and 10 m core depth.
Otoliths would indicate a more active oceanic productivity during the phase of active transgression and
would become seldom during the high-stand phase. Pteropods, although in a more irregular way, seem
to depict the same trend.
The long-lasting presence of highly sandy turbidites during the three phases of lowstand, transgressive
phase and highstand calls up those observed in the deepest sediment cores off Algiers (KMDJ04) and
off Oran (KMDJ07). These turbidites are expected to occur downdip of the sand-rich outer shelf.
4.1.3. KMDJ37
This 8.70 m long sediment core was sampled in a water depth of 2752 m, a value rather comparable to
that of the KMDJ39 core, but here at a shorter distance from the foot of the slope. As the previous one
(KMDJ39), this site belongs to a sector mapped in layered echo-facies with parallel and continuous
reflectors.
The grey-beige mud contains dark coarse-grained turbidite of various centimetres thickness with more
and more clear fining toward the top of the sequence, the gradation of which can be followed over
several tens of centimetres. The sandiest turbidites are rather situated in the lower half of the sediment
core (780 cm, 760 cm, 580 cm, 495 cm, 380 cm and 50 cm) (Fig. 8). The textural analysis indicates that
some Bouma Ta divisions of the turbidites are exceptionally sand-rich: about 100% of sands at 7.80 and
7.60 m, more than 50% of sands at 4.95 and 3.80 m. Described for the first time in this study, the
abundance of coarse sands in these proximal turbidites allows us to follow a very clear normal gradedbedding.
The beginning of the sedimentary history goes back to 16,080 BP. Three 14C dates allow the calculation
of successive sediment accumulation rates of 53.8 cm / 103 years between 16,080 and 8880 BP, of 47.8
cm / 103 years between 8800 and 3290 BP and of 59.2 cm / 103 years after 3290 BP; the rate for the
whole sedimentary column reaches 52.9 cm / 103 years (Table 1). We thus observe here that the rhythm
of accumulation is rather homogeneous and does not seem to have been affected by the eustatic
position of the sea level : this could be partially linked to the high recurrence of turbidites.
9
In spite of the distance of about ten kilometres that separates the sites of Cores KMDJ39 and KMDJ37,
we observe a rather comparable chronology of the major gravity events in both core sections: around
15,000 BP, a little before 9000 BP, a little after 8000 BP, and finally, after 1000 BP. The highstand phase
is characterized by a relatively quiet period of the gravity activity. The short distance in the slope of Core
KMDJ37 can explain its coarser grained lithology and thicker turbiditic sequences and, consequently, a
rate of accumulation a little higher than that of Core KMDJ39. During the phase of sea level rise and until
6-7000 BP, carbonate contents are between 25 and 35%; they then oscillate between 15 and 25%
during the highstand. The Ta division of each sequence is underlined by peaks from 60 to 70% of
carbonate, which illustrates the abundance of relict fragments of littoral shells dropped from the outer
edge of the shelf. These shell fragments are associated with the tests of E. crispum and sometimes of A.
beccarii and Q. seminulum (the latter being very corroded), the other markers of the coastal
environments of the last lowstand. We also note, especially in the lower half of the core section, the
presence of glauconite grains stemming from the overlying shelf. On the other hand, and similarly to the
KMDJ37 core, all these markers of the erosion of the shelf edge are rare or absent in turbidites
deposited during the highstand, suggesting that the gravity processes, in particular after 1000 BP, would
have initiated on the slope. It is also possible that the relict deposits of the outer edge of the shelf are
buried under more recent muds. The oxidized rocky fragments have a more irregular distribution, an
observation which concerns almost all the vertical record. Finally, coals and vegetable fibres are
observed rather frequently along the whole core section.
The oceanic production, indicated by otoliths, reaches its maximum during the first half of the Holocene
and tends to decrease after 3000 BP.
4.2. Bejaia area
4.2.1. KMDJ33
This 7.20 m sediment core was taken in 66 m water depth from the shelf, off the mouth of Soummam
Oued and within its prodelta mapped by Leclaire (1972). Overall, it is represented by an accumulation of
grey-beige sandy mud abundantly speckled with black clusters where we observe a recurrence of darker
and sandier levels of several centimetres thickness. The textural analysis indicates the presence of 20
normal size grading layers containing 40 to 70% of fine quartzeous sands; these sandy maxima
correspond to many major floods of the Soummam Oued. Medium and coarse sands are always
practically absent but vegetal fibres and charcoals are ubiquitous throughout the section (Fig. 9).
Two dates were measured and indicate 3060 BP at the base of the section and 2285 BP at about 3 m
depth. We calculate a very high sedimentation rate before 2285 BP: 536 cm / 103 years, then becoming
much more moderate after this date: 130 cm / 103 years, the global rate average being in the order of
235 cm / 103 years. It is a relative fast sedimentation, which is around three times the one recorded in
the prodelta off the Oued Cheliff, on the western margin of Algeria (Domzig et al., 2009). This high rate
can be explained by the high present-day solid discharges of Oued Soummam which are four to five
times stronger than those of Oued Cheliff (Leclaire, 1972). This globally active accumulation is credibly
10
controlled by the steep slope of the hillsides of the Soummam valley, which is dominated by reliefs at
a height of more than 1000 m. Under these conditions, the core sampler was not able to reach the
coarse shelly sands deposited during the beginning of the transgressive episode (Giresse et al., 2009);
based on the calculated global rate, these sands would be buried under a high-stand prism from 20 to 25
m thick.
These ages allow the calculation of an average frequency of 1 per 150 years for the floods of Soummam
Oued. The periodicity of the floods increases from 150 years to 250 years after 2000 BP and at the
same time as the sedimentation rate decreases. Therefore, this periodicity always represents a largely
shorter time span than the periodicity of the turbidites of the underlying basin, as we shall see in the
following cores. We also note that the sand content of these flood deposits is always markedly higher
than the one of the underlying turbidites.
Carbonates represent on average 25% wt of the sediment and exceed 30% wt, sometimes reaching
40% in the deposits between 5 and 2 m core depth. The decrease in the uppermost two meters
suggests a relative decline of the oceanic productivity. However, several carbonate peaks correspond to
the sandiest peaks of the floods indicating its partially detrital character inherited from the hillsides of
Soummam Oued where the outcrops of flysch are particularly spread. Several littoral foraminifers,
pteropods as well as the otoliths of fish tend to get seldom after 2000 BP, probably involving a general
deterioration of the conditions of oceanic production.
4.2.2. KMDJ34
This 8.68 m- long sediment core was taken in 2220 m water depth; it is approximately 5 km from the foot
of the slope and in line with the main N-S submarine valley of the Bejaia margin (Fig. 3). This site is also
approximately off the mouth of the Soummam Oued. The chirp profile indicates that this site belongs to a
vast interfluve isolated from the wide channels of transit-erosion (rough echofacies), which are mapped
to the east and to the west.
The sediment is a grey-beige mud where turbidites appear in the form of darker microbeds because they
are richer in organic matter. These turbidite beds range in thickness from 1 to 3 mm; their contents in
fine sand do not exceed 20 to 25% and are characteristic of distal facies on the side of valleys (Fig. 9).
These laminae are equivalent to Tc, Td, and in some cases Te divisions of Bouma or E division from
Piper (1978), sometimes with graded laminated basal layers (E1 and E2 divisions).
The base of the core was dated to 8260 BP, which allows us to calculate an overall sediment
accumulation rate of 103 cm / 103 years, which is the highest rate measured over the whole Algerian
margin. Four 14C ages allow the calculation of a relatively slow accumulation rate in the lowermost two
metres of the deposit, followed by a marked acceleration in the upper six metres (i.e. the last two
millenniums) (Table 1). We thus note a chronological succession of the accumulation rates, which is
roughly different from that of the prodelta of the neighbouring shelf (KMDJ33). These upper six meters
coincide with the marked recurrence of the turbidites (we count 19 microsequences in six meters).
During these maxima of gravity accumulation, the diameters of the biggest grains of quartz only
exceptionally exceed 0.5 mm.
11
The contents in carbonate are rather homogeneous and remain included between 20 and 25-wt%,
values slightly lower than those of the prodelta; they decrease a little in the upper meter. The coastal
foraminifers (A. beccarii, E. crispum), usual markers of the reworking of relict sediments of the outer
edge of the shelf, are absent. On the scale of this high-stand period post 8000 BP, the relict deposits of
the outer part of the shelf probably remained covered by muds of the Soummam Oued (Leclaire, 1972).
The gravity transport is underlined, however, by the irregular presence of rocky microfragments or by
shell debris of coastal molluscs, which indicate varied origins for the gravity flows. The abundance of the
vegetal fibres throughout the section reveals flows stemming Oued from Soummam. However, a direct
incidence of floods originated from this Oued is not demonstrated in the deep sedimentation because the
sandy deposits of its prodelta do not seem to go beyond the middle part of the shelf (Leclaire, 1972).
The rarefaction of otoliths and of shell debris over 6 m (after 2000 BP) would mean a deterioration in
productivity which is similar to that observed on the shelf (KMDJ33). Pteropods disappear after 6000 BP.
4.2.3. KMDJ36
This 8,58 m-long sediment core is situated at ca. 15 km downdip of KMDJ34 core , which is
approximately the distance to the foot of the slope (Fig. 3). It was obtained at 2468 m water depth, from
a slightly raised zone of the basin, which is presumed to act as an interfluve.
It is essentially made of a grey-beige mud where dark turbiditic microbeds (11 microbeds) are placed
along the whole section. It is still[??]about [?? They depict???] thin-bedded distal facies of millimetrelength thickness and its content [??] in sand is lower than 10%. As in Core KMDJ34, they all belong to
the Tc, d, and e of Bouma or E divisions from Piper (1978) and generally consist of ungraded massive
silts (E3 division). However, in contrast with Core KMDJ34, these microturbidites repeat over the whole
vertical section, even if they are more abundant in the lowermost two metres where the biggest grains of
quartz reach diameters ranging from 0,2 to 1 mm (Fig. 9).
The sedimentary history begins to 15,990 BP. It allows for a comparison of the deposits of the phase of
active transgression with that of the highstand. From this date, we estimate an overall sediment
accumulation rate of 54 cm / 103 years, that is about half the rate of Core KMDJ34 which is closer to the
slope and to the continent. It results into a long periodicity of turbidites, i.e. near ten times longer than
the one of Core KMDJ34 (1 / 1450 years). Three 14C ages allow estimating an accumulation rate
moderate enough (43 cm / 103 years) during the active phase of the transgression, followed by an
acceleration during the high-level stand (69 cm / 103 years between 5105 and 2265 BP and 93 cm / 103
years after 2265 BP) (Table 1). This increase can be correlated with that observed in the lower part of
Core KMDJ34.
The contents in carbonate oscillate between 20 and 30 wt% during the phase of active transgression
and remain nearby of 20 wt% during the high stand, more exactly in the uppermost three metres of the
core. This distribution correlates with the curve of the coastal foraminifera (A. beccarii, Q. seminulum)
and especially shell debris, the concentration of which controls the importance of the carbonate
accumulation.
12
The origin of gravity mechanisms from the outer edge of the shelf is identified well in the deposits
preceding the high stand thanks to the presence of the benthic foraminifera (Q. seminulum, A. beccarii)
as well as by a bigger abundance of rocky fragments, shell debris and degraded plant fragments. All
these markers get seldom or even disappear after 6000 years BP where the entire outer shelf was
buried under prodeltaic muds of Oued Soummam (Leclaire, 1972). Here, only the rocky fragments
indicate gravity movements. Conversely,the small charcoal fragments are more abundant within the
highstand deposits where they indicate the alluvial contributions of Soummam Oued.
Considering oceanic productivity, pteropods appear towards 12,000 BP and disappear after 6000 BP
(about as in KMDJ34) while otoliths get very seldom little after 3000 BP.
4.2.4. KMDJ 35
It is an approximately 8.58 m-long sediment core taken at 2606 m water depth in a depression just below
the slope. Although at a short distance from this slope, this site is also the deepest of those studied in
the sector of Bejaia (Fig. 2). Core KMDJ35 is located in one of the rare sectors of this margin where Chirp
profiles indicate buried or superficial transparent echo facies just below a main deep valley, which is
characterised by chaotic echofacies.
The sedimentation is characterized by a disorganized accumulation of oval, circular and angular
mudstone clasts through the lowest two-thirds of the core section. Between 7.68 m and 6.10 m, blocks of
compacted mud several tens of centimetres thick are observed; their grey-blue to dark blue colour
contrasts with a10% fluid beige grey mud, i.e. a hemipelagic matrix, which generally accompanies it.
These blocks were contorted and wrinkled during the slumping and armed with thick tests of coastal
molluscs and grains of quartz or sandstone; in these blocks, structures of silty paralic laminae are
sometimes preserved. These characteristics meet the definitions of the slumps, which insist on the
rotational character of the movements along shear-surfaces ending in internal plastic deformations
(Dingle, 1977; Shanmugan, 1996). An increase in mass disaggregation [disintegration?] and mixing of
sediment with water during downslope movement can transform a slump into a debris flow.
Then, nearly one metre of a "normal" accumulation of muds and turbidites, although still including some
floating dark bluish clasts, is observed. Above, over about two metres, the accumulation of cohesive
slumps (some tens of centimetres large) and clastic debrites (some centimetres large) still occurs. All
these blue and compacted mudstones recall the facies encountered in situ on the outer edge of the shelf
and belonging to the MIS 3 (Leclaire, 1972). The same facies has only once been described at the foot
of the slope of the Algiers margin (KMDJ01) (Giresse et al., 2009). Finally, the last three meters of the
core consist of beige-grey mud where numerous (avg = 16 beds per 3 m) and thin (avg = 0.75 cm) very
sandy turbidites are included (Fig. 9).
The dating of the base of the sediment core indicates 4825 BP whereas the top of the slumps and
debrites are 4795 BP, suggesting an average age for this major gravity event. The ages of 7840 BP in
5.5 m and of 5730 BP in 3.80 m are apparent ages, which correspond to mixtures of hemipelagic
Holocene materials and allochthonous relict materials from the edge of the shelf brought by debris-flows
or slumping. Consequently, the only correct measure of sediment accumulation rate concerns the
13
uppermost 2,84 m for which we obtain a value of 59 cm / 103 years, which is, in spite of the recurrent
occurrences of turbidites, comparable to that measured in the nearby KMDJ36 core.
The disorder of the accumulation is verified by the irregularity of distribution of most of the components
of the sediment, in particular the carbonate and coarse quartz contents which increase during turbidite
and debrite intervals, but decrease during the slumped blocks intervals. Markers of the coastal relic
deposits of the edge of the shelf (E. crispum, A. beccarii and shallow water miliolidae) underline the
turbiditic episodes, but are more secluded or even absent in the dark blue masses of slumped blocks.
The long-lasting presence of these markers within the high-stand turbidites indicates erosion of the outer
edge where the relic deposits widely outcrop (Fig. 4; Leclaire, 1972).
Carbonates represent on average 20-wt% in the deposits of gravity origin with peaks between 30 and
40-wt% for clasts armed with mollusc shells. These contents decrease below 20 wt% in the hemipelagic
deposits later than 4795 BP, thus joining the trend observed in two other cores of this margin of Bejaia.
Independently from the major phenomenon of associated slumping and debris-flows, we shall note the
high recurrence of sandy turbidites in the high-stand accumulation; it is a rather rare characteristic, which
we were able to observe only in some of the deepest deposits of the basin of the western margin of
Algeria (KMDJ04 and KMDJ07) (Giresse et al., 2009).
The oceanic productivity is difficult to estimate because of the mixture of sediments from different
feeding sources through the lower two thirds of the section. We can, however, note a rarefaction of
otoliths in the deposits after 4795 BP.
5. Results of statistical analysis on turbidites
5.1. Frequency distribution of turbidite thickness
Many studies have evidenced an approximately lognormal frequency distribution of the sandy division
thickness of the turbidite sequence (Drummond and Wilkinson, 1996; Murray et al., 1996; Ishihara et al.,
1997). However, the cumulative frequency distribution of the entire turbidite thickness commonly follows
a segmented power-law relationship that is equivalent to the summation of lognormal distributions. Each
lognormal distribution is associated with a characteristic basal Bouma Ta, and Tb divisions that are
deposited by flows with high sediment concentrations, whereas Tc, Td, and Te divisions originate from
diluted flow. Such summation produces a step in the trend of the probability plot for the entire bed
population (Talling, 2001). Here, thickness data of the sandy basal interval and of the entire turbidite are
shown on probability plots with logarithmic axes (Fig. 10).
Most turbidites of KMDJ34 and KMDJ36 sediment cores are thin-bedded; similar lognormal-frequency
distribution is recorded for the entire turbidite thickness: however, two or three coarse-grained turbidites
are an exception to the rule. This means a rather homogeneous flow rheology and probably a good
preservation of the deposit (i.e. without top-truncation) (Talling, 2001). The sandy interval thickness is
observed only in some levels of KMDJ34 where we observe the same break of slope.
The graphic representation of KMDJ35 is restricted to 3 m of deposits above the slump debris-flow. From
a low statistical base which concerns only the sandy intervals, we record a distribution which is almost
14
lognormal (r2=0.876), the break being expressed is driven by the only interval above 10 cm. Debrisflow deposit is not represented because its thickness distribution deviates systematically from a
lognormal distribution.
KMDJ37 and KMDJ39 sediment cores are characterized by a clear contrast between dominant thinbedded turbidites and few thick-bedded turbidites. The sandy interval thicknesses show a typically
segmented distribution, which is driven by the four intervals thicker than 10 cm. This break of the
rheological conditions is almost identically observed in KMDJ37 and 39. However in the latter case, we
cannot preclude that the second slope corresponds to thicker layers probably truncated at the top. If we
compare the graphs of these two cores, we notice that the coefficient of correlation is better in KMDJ39
(r2=0.908) than in KMDJ37 (r2=0.712), suggesting that in the deepest water depth core (KMDJ39), the
preservation of the entire sequence is better than it is at the foot of slope (KMDJ37). Thus the
occurrence of two flows with distinct rheology is likely to be recorded here, with a good preservation of
each sequence. Most of the turbidite-bed thicknesses along the eastern Algeria margin are bimodal as
along the western Algeria margin. Very thin-bedded turbidite ( 4 to 7 cm) are represented in the Bejaia
area, whereas few thin-bedded turbidites ( 5 cm) and numerous thick-bedded turbidites ( 16 to 18 cm)
are represented in the Annaba area.
5.2. Relationship between grain-size and bed thickness
The logarithmic plots of bed thickness and basal grain-size class have an approximately lognormal
frequency distribution of bed thickness with a median thickness that increases for larger grain-size
classes. However, distinctly different slopes of the plot are observed for each grain-size class: these
differences are inferred as indicating a change from viscous settling to inertial settling as grains became
larger (Talling, 2001).
In Figure 11, the logarithmic plots of the entire turbidite thickness and sandy-bed thickness are
presented against total sand content (>40 m). Several general characteristics are evidenced:
- The positive correlation of sand percentage with sandy-bed thickness, in each case, is better than
for sand percentage and entire turbidite thickness.
- The positive correlation of sand percentage with entire turbidite thickness is poor or even
[sometimes] absent.
Such results suggest rather marked erosion of the tops of the entire turbidite sequence and, in different
degrees, the top of basal interval (Ta). Such irregular erosion of the top of the sequence is controlled by
recurrent turbiditic events.
In the Bejaia sector, thin-bedded turbidites show basal intervals that are difficult to determine in term of
Bouma classification, especially in the case of Core KMDJ36. Only Core KMDJ34 could be correctly
analyzed from this point of view. The positive correlation of sand percentage with entire turbidite
thickness is statistically driven by only one point with thickness greater than 100 mm. Core KMDJ35 is
analysed only in the upper 3 m where thick-bedded turbidites are observed: the relationship is poor.
The area of Annaba shows basal divisions Ta of more than 10 cm in thickness (KMDJ37 and 39 cores).
This area also supports the view that thick sediment sequences were commonly truncated. This process
15
appears more intense at the foot of the slope (KMDJ37) than with a higher water depth and a remote
location from the slope (KMDJ39). A mean regression coefficient was calculated for the sandy base, the
values being lower in KMDJ37 (r2=0.498) than in KMDJ39 (r2=0.618) where the upper beds have
undergone less severe truncation. In some places, sharp contrast or reworking would indicate these
erosions. As in the western margin (Giresse et al., 2009), the deepest site generally indicates that
accumulation processes appear to take precedence over erosion processes.
6. Relationships between eustasy, emplacement frequency of turbidites and fluxes
The significant number of 14C datings obtained in this study authorizes a general balance of the
accumulation rate and that of the repetition of turbidites. However, this balance will sometimes be
constrained by the chronological differences of the records obtained, which still do not allow for
comparisons on the same time intervals (Fig. 12).
6.1. Annaba area
The KMDJ38-core site, isolated for at least 50,000 years from the major axes of the sedimentary transit,
presents fluxes that are about third of those generally observed at the foot of the slope. [In nearly
constant hemipelagic conditions of sedimentation, the accumulation rate depends both on direct
terrigenous inputs and on the carbonate oceanic production. Cette phrase est-elle réellement utile ici?
Mon avis est que non] We observe that the sedimentation rate decreases significantly after 7120 BP, i.e.
during highstand. A similar trend was observed from the analyses of the relationship between the
accumulation rate and the distance of the coastline from the shelf break in the western margin of Algeria
(KMDJ04, 06, 07 and 08) (Domzig et al., 2009; Giresse et al., 2009).
Amplitudes of the curve of the sediment accumulation rate (Fig. 6) can reach a factor of 5: warmer
waters generally seem more favourable to a faster accumulation linked to higher biogenic carbonate
fluxes. This relationship is particularly clear at the scale of the last climatic oscillation, i.e. the scale of
most of the records in this study. We note, as in the prodelta of Soummam, a decrease of this rate
during the second half of the Holocene. In the Western Mediterranean, Ramrath et al. (2000) depict the
mid-Holocene hydro-climatic switch as a climatic deterioration with cooler and drier conditions and
widespread aeolian mobilization in the Sahara (Swezey, 2001). After 3500 BC, in extreme conditions,
such as in the Lybian Desert, even in ecological niches, the rains ceased completely (Kuper, 2006). The
northern African region was characterized by mid- to late-Holocene alternation of soil formation and
flooding (Zielhofer et al., 2008): the fluvial dynamics of Moulouya Oued (NE Morocco) indicates stable
and more humid environmental conditions around 6.7 ka and an enduring change in flood dynamics
somewhat before 4.1 ka. Overbank deposits of Medjerda Oued (northern Tunisia and eastern Algeria)
indicate a period of mid-Holocene stability, then around 4.7 ka an enhancement of fluvial dynamics and
around 2.0 ka, morphodynamic stability and soil formation may be observed.
16
The two cores KMDJ37 and KMDJ39 are about ten kilometres apart and show rather comparable
successions of lithological characteristics and fluxes. The same turbiditic paroxysms can be correlated,
at least at the scale of the last 16,000 years.
The KMDJ 37 record, beginning around 16,080 BP, shows a rather active and regular accumulation,
which does not seem directly influenced by eustasy. Closer to the slope, the fluxes at KMDJ37 site are
slightly higher than ththose at KMDJ39 site. Coarser grained turbiditic beds, especially in the lowermost
part of the core, result into slightly higher fluxes. However, the main information is the evidence for an
acceleration of the sedimentation which takes place at the end of the highstand, as in Core KMDJ39.
The irregularities of the frequency of occurrenceof turbidites in this period could be compensated by
intrinsic fluctuations in terrigenous flows or by erosions in every deposition of sandy turbidite. The
statistical study of the thickness of sequences tends to favor the second explanation.
The site of the KMDJ39 core, as the previous one, is located on the route of channel gravity transport:
consequently, it is the place of sandy and graded turbidites accumulations, which succeed one another
over 24,050 years. The fluxes are moderate, however, and increase gradually from the interval of
lowstand to that of highstand while conversely, the period between two turbidites tends rather to
decrease during the highstand. However, the main information is the evidence for an acceleration of the
sedimentation, which takes place at the end of highstand but not during lowstand as in KMDJ37. It is
suggested that the fluctuations in the gravity process are powerful enough to be sufficient to balance the
fluctuations in terrigenous flows. We can also suggest here that the scouring exerted in particular during
each of these sandy turbiditic accumulations weigh strongly in the integrated balance of the
accumulation rate.
Turbidites are mainly coarse-grained and intervene on almost all the vertical section. Especially, in Core
KMDJ37, the periodicity of the appearance of turbidites is shorter (more repetitive) in phases of lowstand
and the beginning of sea level rise (1/248 years) than in phase of active transgression (1/508 years) and
highstand (1/365 years), suggesting a stronger instability at the top of the slope during the low-stand
period. Over a slightly longer time scale, KMDJ39 also shows an active transgression phase
characterized by moderate fluxes and turbidites spaced out in time during the last 5000 years of the
highstand. There is once more a simultaneous acceleration of the rate of the accumulation and of the
recurrence of turbidites [pas sûr de bien comprendre cette phrase – reformuler?]. The phase of lowstand
does not allow a direct comparison with that of KMDJ37 [pourquoi?].
In these two sediment cores, in spite of the long-lasting repetition of the gravity processes over most of
the vertical section, we can distinguish:
- Turbidites of lowstand and active transgression interval where the reworking of the coast relict deposit
of the outer shelf regularly allows to identify the point of departure of the mechanism;
- High-stand turbidites (after 8 or 9000 years BP) where the markers of the littoral relict deposits are rare
or absent, giving evidence either for a deeper initiation of the mechanism, near the heads of canyons or,
more probably, the burial of these relict deposits under prodeltaic Holocene muds.
The two cores KMDJ37-39 show marked coarse-grained turbidites (with more than 50 % of sand and
coarse sands in the very beginning of the sequence), which are characteristic of the corridor of transit
17
drawn on the echofacies map (Gaullier et al., 2006). These repetitive accumulations of sandy
turbidites are situated upstream of a sector of the basin where a wide field of sedimentary ridges has
been mapped, which could result from the same mechanics of gravity transfer. Here, this field should not
be attributed to contour-current activity unless additional supporting evidence is found (sediments,
current measurements, and so forth). This regional association suggests that the ridges represent bed
forms created by turbidity flows that spread across the slope and occasionally outside the channel axis
as in the Monterey-fan valley (Normark et al., 1980), the northern South China Basin (Damuth, 1979) or
in the western Mediterranean, as in the Pyreneo-Languedocian Sedimentary Ridge in the deep western
Gulf of Lions (Jallet and Giresse, 2005).
6.2. Bejaia area
KMD33, in the Soummam prodelta, is the site of an accumulation rate three times faster than that of the
Cheliff prodelta on the western margin of Algeria. This character is related to the ascendancy of the
flysch outcrops on its hillsides. After 2285 BP, and in connection with a climatic degradation, the fluxes
decrease by a factor of 5. This trend is not directly expressed in the other three deep sedimentary
successions analyzed in the same area. Moreover, these flood deposits are sharply coarser than the
turbidites at the bottom of the slope and with 20 maxima during 3000 years, and their recurrence is much
higher. The map of the prodelta of the shelf (Leclaire, 1972) indicates rather SW-NE transit of silty
suspended matter, which does not affect the outer edge of the shelf, much of the coarsest discharged
sediment remaining in the prodelta and inner shelf. Finally, if suspended matters were sometimes able to
pass the break of slope, they would rather have been transferred in the direction of the canyons situated
further east.
Cores KMDJ34 and 36 are situated in the same transect on the west of the deep-sea fan of Bejaia and
sheltered from the accumulations of proximal turbidites (the contents in sand are generally below 20 %).
KMDJ34, closer to the slope, logically shows fluxes around double that of Core KMDJ36. These two core
sections show the same flux increments during the highstand (multiplied by 5 - 10 in KMJDJ34,
multiplied by 2 in KMDJ36).
Core KMDJ34, at the foot of the slope, shows a clear acceleration of the rate of accumulation between
2300 and 1145 BP. After 2300 BP, we move from a gravity event each 350 to 400 years to a gravity
event each ~ 30 years. This change occurs while the eustatic level remained stable and the terrigenous
fluxes of the Soummam Oued tended to decrease at the same time as the frequency of its floods. We
thus have here the evidence for a gravity instability independent from any eustatic factor or terrestrial
(environmental) change and credibly enhanced by a local intensification of the seismic activity spreading
from 2300 to ~ 500 BP.
In Core KMDJ36, the rather long emplacement frequency of turbidites (more than 700 years) between
16,000 and 6,000 years became shorter by half during the high-stand interval. This evolution has a direct
incidence on the acceleration of the sedimentation after 5105 BP and still more after 2265 BP. If we take
into account its longer time scale, this evolution was rather similar to that of Core KMDJ34, namely the
rate of recurrence of the turbidites controls the rate of the accumulation. We thus have here another
18
example of accumulation, which does not seem driven by the eustatic position of the sea level, but
rather by the instability of the top of the slope. In coherence with the record of KMDJ34, this instability
could credibly be related to seismic activity.
Approximately 20 km to the west of the KMDJ34-KMDJ36 transect, and situated close to the foot of
slope, Core KMDJ35 was the site of a gravity “cataclysm” at around 4700 BP, allowing the accumulation
over at least 5 meters of debrites and convoluted layers. In the uppermost three meters of this core, i.e.
during approximately the last two thousand years, the rate of repetition of turbidites and the fluxes are
roughly the same for the same period as those recorded in Core KMDJ36.
In summary, the rate of the turbiditic accumulations may be related to two main trigger mechanisms:
1) At the top of slope and on the edge of the shelf, gravity processes can be induced by the littoral
turbulences during the episodes of lowstand. This relation is weakening consequently in high-stand
intervals. In the eastern Algerian margin, this process would apply only to Core KMDJ 37 (and possibly
39) while it is common on the western Algerian margin.
2) In stable and prolonged high-stand episodes, “accelerations” restricted in time (from 2300-1100 BP)
and controlled by the occurrence of the turbiditic sequences can be linked either to paroxysms of the
seismic activity or to exceptional river floodings.
We favor here the first explanation, because such “accelerations” are localized specifically in this eastern
part of the Algerian margin, and more particularly off Bejaia, and because in the record of the main
floods in the prodelta of Soummam Oued, neither the coarse nature of the deposits nor the period of the
frequencies correspond to those of the turbidites. Lastly, seismically triggered turbidites have volumes
that are several fold larger than the “flood-generated” turbidites (Gorsline et al., 2000). Such
mechanisms {lesquels? “seismically triggered’ ou “flood-triggered”? Pas clair dans la tournure de la
phrase –je comprends que c’est le premier] seem to have occurred during the deposition of the thickbedded and coarse-grained turbidites, which are correlated over more than 10 km distance through ca.
16,000 years (KMDJ37-39) on the Annaba margin.
7. Comparison of sedimentation patterns on the western (MARADJA 1) and eastern (MARADJA 2)
margins of Algeria
7.1. Comparison of sediment accumulation rates on the scale of the records of the entire core sections
Accumulation rates calculated on the western part of the margin of Algeria (MARADJA 1, see Giresse et
al., 2009) are relatively homogeneous [over which time scale?] : they depict similar mean rates ranging
between 39.5 and 64.2 cm / 103 years. Slight differences are associated with the more or less remote
positions with regard to the channels of the submarine valleys. In this western part, the periodicity of
turbidites is included between 299 and 646 years. Only the sediment core KMDJ02 makes an exception
with a periodicity of 1071 years, the site corresponding to a sector of interfluve at the middle-slope where
only some fine-grained microturbidites settled (Giresse et al., 2009).
Accumulation rates calculated on the eastern margin (MARADJA 2) are very heterogeneous, varying
from 15.5 cm / 103 years (KMDJ38) to 105 cm / 103 years (KMDJ34). The site of Core KMDJ38 at the
19
middle-slope cannot be taken into account because it is about an almost fully hemipelagic deposit
corresponding to the rate of the slowest accumulation of the margin whereas the high rate of Core
KMDJ34 is consequent to a proximal situation right at the foot of the slope, depicting a strong recurrence
of turbidites. The other rates of this eastern margin are included between 35.5 and 57.5 cm / 103 years,
which is an order of value that is then comparable with those obtained on the western margin. In this
eastern margin, the average periodicity of turbidites is included between 137 and 534 years, i.e.
comparable to those found in the western margin.
It is worth to compare these turbidite periodicities in the deeper domains, which we suspect to be
seismically triggered, with the return period of major earthquakes on the fault close to the shoreline.
Although the completeness of the seismicity catalogue over the last 1000 years of northern Algeria is
questionable (Harbi et al., 2010), we are aware that M> 6.5 earthquakes indeed occur and are able to
trigger large turbidity currents (see 1954 and 2003 events), provided that the margin has been loaded
enough by sediment accumulation. About 20 M > 6.5 earthquakes have shaked the Algerian margin
since A.D. 856, i.e. over ca. 1200 years, most of them being located over central Algeria (1°-4°E),
whereas only two earthquakes of significant magnitude occurred in the western and eastern margins
(Hamdache et al., 2010). This observation supports the view that margin shaking by earthquakes is not
occurring continuously in space and time, and suggests that for a given area (“segment”) of the margin,
typical time spans between two earthquakes are of the order of 500 years, i.e. very similar to the
periodicities of turbidites found in the deeper domains in this study and in Giresse et al. (2009) over the
last highstand.
The longer time series available for a given fault system close to the shoreline is the El Asnam fault,
which ruptures a > 100 km long fault in 1980 (Meghraoui et al., 1988), i.e. 26 years after the Ms 6.7
Orléansville event in the same area. It has been shown that significant co-seismic displacements
occurred on this fault over the last 7000 years, with returned periods of 350-500 years (“cluster” periods)
and also longer, “quiet” periods of ca. 1000-1700 years. Considering that the margin is likely to be
shaked by other large events from other sources in the area (as shown by the occurrence of the 1954
and 1980 earthquakes for instance), we again find a typical periodicity (over the highstand period) that
agrees with periodicities of turbidites found in the deep Algerian basin.
We can conclude that on the basis of the analyses of about fifteen sediment cores of the Algerian margin
and by discarding here two only examples of submarine avalanches with accumulation of debrites or
slumps (KMDJ01 and KMDJ35), we therefore infer rather comparable rates of accumulation in all the
bottoms [??] below the Algerian slope, in good agreement with known earthquake return periods over
the Algerian margins. A comparison between the two parts of the Algerian margin also shows
periodicities of the gravity processes, which are appreciably equivalent on both parts. In a time scale
which generally does not exceed the last eustatic cycle, we cannot define a privileged [peculiar?] sector
in term of coarse terrigenous fluxes because these, for the main part, accumulate on the shelf: sandy
fluxes resulting from the erosion of the flysch of the hillsides of Soummam Oued condition an active
accumulation at the level of the pro-delta, but without marked incidence in the underlying deep basin. On
the other hand, we can envisage that a seismicity control could have been determining in deeper and
20
more remote bottoms from the basin if we take into account the observations of cable breakages
consequent to the 2003 Mw 6.8 Boumerdès earthquake (Ayadi et al., 2003) as well as the 1954 Ms 6.7
Orléansville earthquake (Heezen and Ewing, 1955).
7.2. Sediment accumulation rate and periodicity of turbidites at the scale of the eustatic stages In the
western Algerian margin, we distinguish the conditions of sedimentation near the foot of the slope from
those in the deeper basin. All the records at the foot of the slope indicate a decreasing frequency of
turbidite beds from the transgressive interval to the highstand system interval. In the same way, off-shelf
sediment fluxes were lower during the current highstand (last ca. 6000 years). This decrease was
strong, even reduced by half, during the transgressive interval; this trend is also observed off the Gulf of
Lions (Dennielou et al., 2003). Here [c’est-à-dire? Pour le cas du golfe du Lion, pour l’ouest de l’Algérie
dont il est question en début de phrase, ou pour celui de notre étude? Ambigü – à préciser], it is
suggested that the frequency of the sediment gravity flows is mainly controlled by eustatic changes.
However, in the eastern Algerian margin and especially in the area off Bejaia, the flux calculations
indicate a frequent increase during the highstand and more precisely between ca. 2300 and 1100 BP:
this short-time increase is tentatively assigned to a seismically induced signal that was not recorded in
the western part.
In the two deeper sediment cores of the western margin of Algeria (KMDJ04, 07), turbidite sequences
are more complete than at the foot of the slope, the greater thickness of each sequence being probably
linked to a better preservation. The turbidite beds are continuously observed through the entire sediment
core, suggesting that the frequency was not related to eustatic changes, because they appear partly
independent from sea level position. However, their emplacement frequency decreased with rising sea
level. As at the foot of the slope, this decrease is commonly associated with a lower sediment
accumulation rate. The normal log-frequency of the sandy component is the same power law distribution
as the magnitude of earth tremor (Talling, 2001). Consequently, these two deeper sites would reflect the
seismicity of the margin whereas the deposition at the foot of the slope is more strongly controlled by the
eustatic trend, probably due to the short distance separating the high-energy shoreline from the shelf
break.
In the eastern Algerian margin, accumulation in the deeper sites does not seem driven by the eustatic
position of the sea level, but rather by the instability of the top of the slope. Core KMDJ37 establishes a
slightly particular case because turbidites are distributed over almost the whole vertical section, but their
emplacement frequency is shorter in phases of lowstand and the beginning of sea-level rise than in
phases of active transgression and highstand.
8. Discussion and conclusions
This study of sedimentary records from the eastern half of the Algerian margin allows us to differentiate:
(1) environments shielded from any seismic incidence and controlled by oceanic or climatic forcing; (2)
by comparison, environments at the foot of the slope and in the deep basin, where gravity mechanisms
randomly intervene and are likely to be seismically triggered in many cases.
21
(1) The record of KMDJ33 in the pro-delta of Soummam indicates an active sedimentation fed by the
erosion of the sandstone outcrops of flysch in the watershed basin. The accumulation, three times faster
than that of the Cheliff pro-delta, in western Algeria, allows us to study the deposit of the last three
millenniums. The very high sedimentation rate before 2285 BP decreased later by a factor of 5. This
evolution is directly connected to the periodicity of the floods, which are less frequent after 2285 BP
where the period lengthened from 150 years to 250 years because of a drier trend of the climate of the
Maghreb, which is attested in particular on the margins of the Sahara. As these prodeltaic deposits do
not reach the outer border of the shelf, their influence on the slope deposits is unlikely. Moreover, the
period of the floods is shorter than that of the turbidites at the foot of the neighbouring slope. Both
processes are clearly independent. At the same time, during the last two millenniums, a decline of the
content in carbonate and a relative rarefaction of pteropods and otoliths of fish indicate a decrease of the
oceanic productivity in colder waters. We thus have here, at about 2000 BP, the evidence for a double
environmental change, both continental and oceanic.
The sedimentation at the middle-slope to the NE of Annaba (KMDJ38) is almost totally shielded from the
gravity mechanisms; the rhythm of the accumulation reaches only the third of that recorded on the foot of
the slope. In spite of the apparent homogeneity of the sedimentation, we observe, with the detailed
chronology of DO events, that it consists of a succession of decelerations and accelerations in
accumulation rate, the amplitude reaching a factor of 5. The accelerations coincided with warmer waters
where there was an increase of the oceanic productivity leading to a surplus of carbonate fluxes (as
particularly indicated by pteropods). Higher accumulation rates are noted after 12,000 BP (27.5, 40, 20
cm / 103 years), then after 5000 BP a slowing down (8.3 cm / 103 years) occurred and integrated the
decrease of the continental fluxes observed after 2285 BP in the pro-delta of Soummam Oued, even if
here we detect no direct sign of periodic floods of the closest river (Seybouse Oued). A double influence,
oceanic and continental, is possible, even if the oceanic factor is probably dominant.
(2) In the same canyon alignment to the north of Bejaia, sediment cores KMDJ34 and 36 are both
characterized by one rather fast accumulation where thin-bedded turbidites succeeded one another,
especially during the second half of the Holocene. In KMDJ 34, the accumulation rate shows a
spectacular acceleration (×5-×10) between 2300 and 1145 BP while the period of frequency of turbidites
decreased from 350-400 years to approximately 30 years. This episode occurred at a moment when the
fluxes of Soummam Oued tended to decrease and where the eustatic highstand was stabilized. We
proposed an episode of strong instability of this slope of Bejaia connected to a “clustered-type” period of
seismic activity, which lasted more than one thousand years. With a little slower accumulation rate,
KMDJ36 does not allow such precise resolution. However, we also observe an increase of the rate
during the highstand with a maximum during the last two thousand years, which integrated the same
episode of instability as Core KMDJ34. The very marked increase of the frequency of turbidites was
controlled by the same instability down from slope.
At the bottom of the slope to the NE of Annaba, both sediment cores KMDJ37 and 39 are characterized
by the interstratifications of the coarse-grained and thick-bedded turbidites. Their frequency was rather
low and the important planning of the top of sequences results in quite moderate accumulation rates. In
22
Core KMDJ37, the rather higher accumulation rate during the lowstand decreased during the sea level
rise interval and the highstand according to alternately higher and lower frequencies of turbidites. In
these cases, the eustatic control might relate to the triggering of the gravity movements, as on the slopes
of the western part of the Algerian margin. Then, the highstand showed a trend to the acceleration of the
accumulation rate linked to a shorter period of turbidites. In Core KMDJ39, with a longer space of time,
we cannot discriminate between phases of lowstand and sea-level rise: the set, in spite of a rather
strong repetition of turbidites, shows moderate accumulation rates connected also to the erosions of the
top of sequences. On the other hand, during the highstand, we observe, as in Core KMDJ37, an
increase of the accumulation rate during the second part of the Holocene where turbidites rapidly
succeed one another. Both sediment cores indicate an increase in gravity flows during the highstand,
which is connected to a general seismicity, and which became intensified during the last thousand years.
Earthquakes and consequent liquefaction of potential weak layers acted as a triggering mechanism for
gravity-driven processes. As main basin-wide turbidites can be identified through KMDJ37-39
correlation, this areal distribution indicates that these gravity flows were probably seismically generated.
Acknowledgments
This project has received support from the following programmes: GDR Marges “Instabilités”, ACI
(Action Concertée Incitative) “Algérie” 2003-2006, the French-Algerian Cooperation Project CMEPTassili 2004-2007 N° 041MDU619, and ANR (Agence Nationale de la Recherche) projects ISIS and
DANACOR. We wish to thank the crew of the R.V. Suroit during the Maradja 2005 cruise, Ifremer Brest
Centre technicians for assistance in the core opening operations and Gary Fowler for linguistic
improvements.
23
References
Ambraseys, N.N., 1982. The seismicity of North Africa: the earthquake of 1856 at Jijeli, Algeria. Boll.
Geofis. Teor. Appl., XXIV(93), 31–37.
Ayadi, A., Maouche, C., Harbi, A., Meghraoui, M., Beldjoudi, H., Oussadou, F., Mahsas, A. Benouar, D.,
Heddar, A., Rouchiche, Y., Khroubi, A., Frogneux, M., Lammali, K., Benhamouda, F., Sebaï, A.,
Bourouis, S., Alasset, P.J., Aoudia, A., Cakir, Z., Merahi, M., Nouar, O., Yelles, A., Bellik, A., Briole,
P., Charade, O., Thouvenot, F., Semmane, F., Ferkoul, A., Deramchi, A., Haned, S.A., 2003. Strong
Algerian earthquake strike near capital city. Eos Trans. AGU 84 (50), 561-568.
Buforn, E., Bezzeghoud, M., Udias, A., and Pro, C., 2004. Seismic sources on the Iberia-African plate
boundary and their tectonic implications. Pure Appl. Geophys., 161, doi:10.1007/s00024-003-2466-1.
Butzin, M., Prange, M., Lohman, G., 2005. Radiocarbon simulations for the glacial ocean: the effects of
wind stress, Southern Ocean sea ice and Heinrich events. Earth Planet. Sci. Lett., 235, 45-61.
Camerlenghi, A., Urgeles, S., and Fantoni, L., 2010. A Database on Submarine Landslides of the
Mediterranean Sea. In D.C. Mosher et al. (eds), Submarine mass movements and their
consequences, Advances in Natural and Technonological Hazards Research, 28, Springer Sciences,
28, 503-513.
Carvajal, C.R., Steel, R.J., 2006. Thick turbidite successions from supply-dominated shelves during sealevel highstand. Geology, 34(8), 655-668.
Cattaneo, A., Babonneau, N., Dan, G., Déverchère, J., Domzig, A., Gaullier V., Lepillier, B., G , de
Lépinay, B.M., Nouguès, A., Strzerzynski, Sultan, N., Yelles, K. , 2010. Submarine landslides along
the Algerian Margin: a review of their occurrences and potential link with tectonic structures. In D.C.
Mosher et al. (eds), Submarine mass movements and their consequences, Advances in Natural and
Technonological Hazards Research, 28, Springer Sciences, 28, 515-525.
Damuth, J.E., 1979. Migrating sediment waves created by turbidity currents in the northern South China
Basin. Geology, 7, 520-523.
Dan, G., Savoye, B., Gaullier, V., Cattaneo, A., Déverchère, J., Yelles, K., & the MARADJA 2003
Scientific Party, in press. Algerian margin sedimentation patterns (Algiers area, south-western
Mediterranean). AAPG-SEPM Special Publication, Mass-Transport deposits in deepwater settings,
Shipp R.C., Weimer P. & Posamentier H.W. (Eds.), 95.
Davis, B., and Stevenson, A., 2007.The 8.2 ka event and Early-Mid Holocene forests, fires and flooding
in the Central Ebro Desert, NE Spain. Quaternary Science Reviews, 26, 1695-1712.
Den Hartog Jager, D.R., Giles, M.R., Griffiths, G.,1993. Evolution of Paleogene submarine fans of the
North Sea in space and time. In: Parker, J.R. (Ed.) Petroleum Geology of Northwest Europe,
Proceedings of the Fourth Conference, Geological Society, London, 1, 59-71.
Dennielou, B., Bonnel, C., Sultan, N., Voisset, M., Berné, S., Droz, L., 2003. Sand transfer into the deep
basin of the Gulf of Lions: even during the sea level high stand…! In Ocean Margin Conference,
Paris, p.129.
Déverchère, J., Yelles, K., Domzig, A.,, Mercier de Lépinay, B., Bouillin, J.-P., , Gaullier, V., Bracène, R.,
Calais, E., Savoye, B., Kherroubi, A., Le Roy, P., Pauc., H, Dan, G., 2005. Active thrust faulting
offshore Boumerdes, Algeria and its relation to the 2003Mw 6.9 eartquake. Geophy. Res. Lett., 32,
L04311, doi: 10229/2004GL021646.
Dingle, R.V., 1977. The anatomy of a large submarine slump on a sheared continental margin (SE
Africa). Geological Society of London Journal, 134, 293-310.
Domzig, A., Yelles, K., Le Roy, C., Déverchère, J., Bouillin, J.-P., Bracene, R., Calais, E., Kherroubi, A.,
Gaullier, V., Savoye, B., Pauc, H., 2006. Searching for the Africa-Eurasia Miocene boundary offshore
western Algeria (MARADJA’03 cruise). C.R. Geoscience, 338, 80-91.
Domzig, A., Gullier, A., Giresse, P., Pauc, H., Déverchère, J., Yelles, K., 2009. Deposition processes
from echo-character mapping along the western Algerian margin (Oran-Tenes), Western
Mediterranean. Marine and Petroleum Geology, 26, 673-694.
Drummond, C.N, Wilkinson, B.H., 1996.Stratal thickness frequencies and the prevalence of orderedness
in stratigraphic sections. J. Geol., 104, 1-18.
Durand-Delga, M., Fonboté, J.M., 1980. Le cadre structural de la Méditerranée occidentale. In: Aubouin,
J., Debelmas, J., Latreille, M. (Eds), Géologie des chaînes alpines issues de la Thétis. Coll. n°5, 26ème
Congrès Géologique International, Paris. Mém. BRGM, 67-85.
Fairbanks, R.G., 1989. A 17,000-year glacio-eustatic event and deep ocean circulation. Nature, 342,
637-642.
Gaullier V., Cattaneo A., Déverchère J., Yelles K., Bracène R., Domzig A., Graindorge D., Kherroubi A.,
Mercier de Lépinay B., 2006. Submarine mass movements on the eastern Algerian margin :
24
preliminary results from the MARADJA 2 – Leg 2 cruise (2005). European Geosciences Union
General Assembly 2006, Vienna, Austria, 02 - 07 April, Geophysical Research Abstracts (GRA CDRom), Volume 8, ISSN : 1029-7006, SRef-ID: 1607-7962/gra/ EGU06-A-08833.
Giresse, P., Pauc, H., Déverchère, J., the Maradja Shipboard Scientific Part, 2009. Sedimentary
processs and origin of sediment gravity-flow deposits on the western Algerian margin during late
Pleistocene and Holocene. Marine and Petroleum Geology, 26, 695-710.
Gorsline, D.S., De Diego, T., Nava-Sanchez, E.H., 2000. Seismically triggered turbidites in small margin
basins: Alfonso Basin, Western Gulf of California and Santa Monica Basin, California Borderland.
Sedimentary Geology, 21-35.
Hamdache, M., Pelàez, J.A., Talbi, A., and Lopez Casado, C., 2010. A unified catalog of main
earthquakes for Northern Algeria from A.D. 856 to 2008. Seism. Res. Lett., 81-5), 732-739,
doi:10.1785/gssrl.81.5.732.
Harbi, A., Peresan A, and Panza, G.F, 2010. Seismicity of Eastern Algeria: a revised and extended
earthquake catalogue. Nat. Hazards, 54, 725-747, doi:10.1007/s11069-009-9497-6.
Harbi, A., Meghraoui, M., and Maouche, S., 2011. The Djidjelli (Algeria) earthquakes of 21 and 22
August 1856 (I-0 VIII, IX) and related tsunami effects revisited. J. Seismology, 15(1), 105-129.
Heezen, B.C., and Ewing, M., 1955. Orleansville earthquake and turbidity currents. AAPG Bulletin,
39(12), 2505-2514.
Ishihara, Y., Miiyataa, Y., Tokuashi, S., 1997. Time series analysis of turbidite sequence in the upper
part of the Awa Group in the Boso Peninsula, central Japan. J. Geol. Soc. Jpn 103, 307-326.
Jallet, L. and Giresse, P., 2005. Construction of the Pyreneo-Languedocian Sedimentary Ridge and
associated sediment waves in the deep western Gulf of Lions, western Mediterranean. Marine and
Petroleum Geology, 22, 865-888.
Kherroubi, A., Déverchère, J., Yelles, K., Mercier de Lépinay, B., Domzig, A., Cattaneo, A., Bracene, R.,
Gaullier, V., and Graindorge, D., 2009. Recent and active deformation pattern off the easternmost
Algerian margin, Western Mediterranean Sea: New evidence for contractional tectonic reactivation.
Marine Geology, Special Issue on EUROMARGINS: Imaging, monitoring, and modelling the physical,
chemical and biological processes in the European passive continental margins, 261(1-4), 17-32,
doi:10.1016/j.margeo.2008.05.016.
Kuper, R., 2006. After 5000 BC: The Lybian desert in transition. Comptes Rendus Palevol., 5, 1-2, 409419.
Leclaire., L., 1972. La sédimentation holocène sur le versant méridional Algéro-Baléares (précontinent
Algérien). Mém. Mus. Natl. Hist. Nat. Paris, CXXIV, 1-392.
Meghraoui, M., Jaegy, R., Lammali, K., and Albarède, F., 1988. Late Holocene earthquake sequences
on the El Asnam (Algeria) thrust fault, Earth Planet. Sci. Lett., 90, 187-203.
Mitchum, R.M., 1985. Seismic stratigraphic expression of submarine fans. In: Berg, O.R., Woolverton,
D.G. (Eds), Seismic stratigraphy 2: an Integrated Approach to Hydrocarbon Exploration. AAPG
Memoir, 39, 117-138.
Murray, J.W., 1991. Ecology and distribution of benthic foraminifera. In: J.J. Lee and O.R. Anderson,
Editors, Biology of Foraminifera, Academic Press, New York (1991), pp. 221–253.
Murray, C.J., Lowe, D.R., Graham, H.S., S.A., Martinez, P.A., Zeng, J., Caroll, A., Cox, R., Hendrix, M.,
Heubeck, C., Miller, D., Moxon, I.W., Sobel, E., Wendebourg, E., Williams, T., 1996. Statistical
analysis of bed thickness patterns in a turbidite section from the Great Valley Sequence, Cache
Creek California. J. Sediment. Petrol., 66, 900-908.
Mutti, E.,1985.Turbidite system and their relations to depositional sequences. In: Zuffa et al. (Eds),
Provenance of arenites. D. Reidel Publishing, Dordrecht, pp. 65-93.
Nocquet, J.-M., Calais, E., 2004. Geodetic measurements of crustal deformation in the Western
Mediterranean and Europe. Pure Appl.Geophys., 161, 661-681.
Normark, W.R., Hess, G.R., Stow, D.A.V. Bowen, A.J., 1980. Sediment waves on the Monterey fan
levee: A preliminary physical interpretation. Marine Geology, 37, 1-2, 1-18.
Nouguès, A., Sultan, N., Cattaneo, A., Dan, G., and Yelles, K., 2010. Detailed Analysis of a Submarine
Landslide (SAR-27) in the Deep Basin Offshore Algiers (Western Mediterranean). In D.C. Mosher et
al. (eds), Submarine mass movements and their consequences, Advances in Natural and
Technonological Hazards Research, 28, Springer Sciences, 28, 541-552.
Pauc, H., Benslama, L., Maouche-Barkane, S., 1997. Les apports fluviatiles sur la marge algérienne,
aspects qualitatifs et quantitatifs. Leur rôle dans la pollution du basin méditerranéen occidental. In:
8èmes Rencontres de l’Agence Régionale pour l’Environnement P.AC.A, Coll. Les apports fluviatiles en
Méditerranée, leurs charges polluantes, leurs effets sur le milieu, Nice.
25
Pickering K.T., Clark, J.D., Smith, R.D.A., Hiscott, R.N., Ricci Lucci, F., Kenyon, N.H., 1995.
Architectural element analysis of turbidite systems, and selected topical problems for sand-prone
deep-water systems. In: Pickering, K.T., Hiscott, R.N., Ricci Lucci, F. (Eds), Atlas of Deep-Water
Environments, Chapman & Hall, London, pp. 1-11.
Piper, D.J.W., 1978. Turbidite muds and silts on deep sea fans and abyssal plains. In: Stanley, D.J. and
Kelling, G. (Eds) : Sedimentation in submarine canyons, fans and trenches, Dowden, Hutchinson and
Ross, Stroudsborg, Pa, 163-176.
Piper, D.J.W., Savoye, B., 1993. Process of late Quaternary turbidites current flow and deposition on the
Var deep-sea fan, north-western Mediterranean. Sedimentology 40, 557-582.
Ramrath, A., Sadori, L., Negendank, F.W., 2010. Sediments from Lago di Mezzano, central Italy: a
record of Late Glacial/ Holocene climatic variations and anthropogenic impact. The Holocene, 10, 8795.
Schrag, D.P., Hampt, G., Murray, D.W., 1996. The temperature and oxygen isotopic composition of the
glacial ocean. Science, 272, 1930-1931.
Shanmugan, G., 1996. High-density currents: are they sandy debris flows ? Journal of Sedimentary
Research, 66, 1, 2-10.
Shanmugan, G., 2000. 50 years of the turbidite paradigm (1950s-1990s): deep-water process and facies
models- a critical perspective. Marine and Petroleum Geology, 17, 285-342.
Stuiver, M., Reimer, P.J., 1993. Extended 14C data base and CALIB 3.0 14C Age calibration program.
Radiocarbon 35 (1), 215-230.
Swezey, C., 2001. Eolian sediment responses to late Quaternary climate changes: temporal and spatial
patterns in the Sahara. Palaeogeography, Palaeoclimatology, Palaeoecology, 167, 1-2, 119-155.
Talling, P.J., 2001. On the frequency distribution of turbidite thickness. Sedimentology, 48, 1297-1329.
von Grafenstein, H., Erlenkeuser, H., Muller, J., Jouzel, J. and Johnsen, S., 1998. The cold event 8200
years ago documented in oxygen isotope records of precipitation in Europe and Greenland. Climate
dynamics, 14, 73-81.
Weninger, B., Alram-Stern, E., Bauer, E., Clare, L., Danzeglocke, U., Jöris, O., Kubatzki, C., Rollefson,
G., Todorova, H., van Andel, T., 2006. Climate forcing due to the 8200 cal yr BP event observed at
Early Neolithic sites in the eastern Mediterranean. Quaternary Research, 66, 401-420.
Wildi, W., 1983. La chaîne tello-rifaine (Algérie-Mroc-Tunisie). Structure, Stratigraphie et Évolution du
Trias au Miocène . Rev. Géol. Dyn. Géogr. Phys., 24, 201-297.
Yelles-Chaouche, A.K., Roger, J., Déverchère, J., Bracene,R.,Domzig, A., Hebert, H., and Kherroubi, A.,
2009. The 1856 Tsunami of Djidjelli (Eastern Algeria): Seismotectonics, modelling and hazard
implications for the Algerian coast. Pure Applied Geophys., 166(1-2), 283-300.
Zielhofer, C., Faust, D., Linstädter, J., 2008. Late Pleistocene and Holocene alluvial archives in the
Southwestern Mediterranean: Changes in fluvial dynamics and past human response. Quaternary
International, 18, 39-54.
Captions
Fig. 1. Main geological units of the Maghrebian Chain (after Domzig et al., 2006) and study zone of
MARADJA 2.
Fig. 2. Shaded bathymetry (50 m resolution DEM) of the Bejaia area and positions of four Küllenberg
cores (KMDJ33-36) are shown.
Fig. 3. Shaded bathymetry (50 m resolution DEM) of the Annaba area and positions of three Küllenberg
cores (KMDJ37-38) are shown.
Fig. 4. Simplified sedimentological maps of the continental shelf of the sectors of Annaba and of Bejaia
(redrawn according to Leclaire,1972).
Fig. 5. Vertical profile of KMDJ38 sediment core; basic sedimentological parameters are shown with
relative abundance of bio- and geo-indicators of the sediment’s provenance. Ages are calendar 14C
ages.
Fig. 6. Oxygen isotopic curve of the tests of planktic foraminifers enabling the location of the complete
succession of the Dansgaard-Oeschger and Heinrich events. On the basis of this chronostratigraphy,
sediment accumulation rates were calculated for each interval.
Fig. 7. Accumulation sediment rate of KMDJ38 sediment core (bold solid green line) deduced from 14C
ages (red squares), Dansgaard-Oeschger (black circles) and Heinrich events (blue arrows). The
accumulation rate curve closely matches the linear regression line; the convex parts of the curves
indicate acceleration intervals attributed to higher bioclastic carbonate accumulation (additional
accumulation)
26
Fig. 8.Vertical profiles of sediment cores from the Annaba Area: a: KMDJ37, b: KMDJ39. The number
of coastal lowstand deposit markers originating from the outer edge is plotted on a 10-g fraction
reference.
Fig. 9. Vertical profiles of sediment cores from the Bejaia area: a: KMDJ33, b: KMDJ34, c: KMDJ36, d:
KMDJ35. The number of coastal lowstand deposit markers originating from the outer edge is plotted on
a 10-g fraction reference.
Fig. 10. Probability plots of entire turbidite thickness and sandy basal interval (Ta). Curves are related to
y = ax + b calculation; grey lines show segmented frequency distributions.
Fig. 11. Logarithmic plots of entire turbidite thickness and of sandy interval thickness (Ta) presented
against total sand content (> 40 µm). Curves are related to y = ax + b calculation.
Fig. 12. Successive fluxes in KMDJ 34, 36, 35, 38, 37 and 39 expressed in mg cm-2 y-1calculated on the
basis of 14C ages (upper axis from left to right). In KMDJ35, debris flow thickness was not taken into
account. Average time (years) The number of coastal lowstand deposit markers originating from
between two turbidites was deduced during the same intervals (lower axis from right to left).
Table 1. Chronology of the seven cores. Calendar ages are given with 1 standard deviation and does not
include reservoir age correction because on the Algerian margin, there are only four measures all
located in the same zone close to the Bay of Algiers, values scattered enough going from 376 to 516
(Butzin et al., 2005).